Reconstitution in vitro of the GDP-fucose biosynthetic pathways of Caenorhabditis elegans and Drosophila melanogaster Simone Rhomberg 1 , Christina Fuchsluger 1 , Dubravko Rendic ´ 1 , Katharina Paschinger 1 , Verena Jantsch 2 , Paul Kosma 1 and Iain B. H. Wilson 1 1 Department fu ¨ r Chemie, Universita ¨ tfu ¨ r Bodenkultur, Vienna, Austria 2 Abteilung fu ¨ r Chromosomenbiologie, Vienna Biocenter II, Austria Fucose is a key component of many oligosaccha- rides involved in recognition events and therefore has roles in disease and development [1]. For instance, Notch, a protein involved in developmental processes in animals, is modified with fucose O-linked to the protein backbone [2], and a defect in the relevant O-fucosyltransferase (POFUT1)is lethal in mice [3], whereas the orthologous gene, nti, is necessary for normal development of Drosophila melanogaster [4,5]. A second O-fucosyltransferase (POFUT2) is also known, and RNAi in Caenorhab- ditis elegans of the relevant gene, pad-2, results in severe body malformations [6]. Less drastic are the effects of ablation of the FUT7 gene required for biosynthesis of certain fucose-containing selectin lig- ands; a lack of the encoded enzyme results in defi- cient leukocyte trafficking [7]. On the other hand, certain fucosylated glycans are immunogenic and allergenic, an example of such a structure being the modification of the N-glycan core by a1,3-linked fucose [8]. This feature is recognized by, e.g. anti- (horseradish peroxidase), which is used to stain Keywords Caenorhabditis; Drosophila; GDP-fucose biosynthesis; GDP-keto-6-deoxymannose 3, 5-epimerase ⁄ 4-reductase; GDP-mannose dehydratase Correspondence I. B. H. Wilson, Department fu ¨ r Chemie, Universita ¨ tfu ¨ r Bodenkultur, Muthgasse 18, A-1190 Vienna, Austria Fax: +43 1 36006 6059 Tel: +43 1 36006 6541 E-mail: iain.wilson@boku.ac.at Database The nucleotide sequences of C. elegans and D. melanogaster gmd and ger cDNA have been submitted to the EMBL database under accession numbers AM231683, AM231684, AM231685, AM231686, AM231687 and AM231688 (Received 30 November 2005, revised 17 February 2006, accepted 20 March 2006) doi:10.1111/j.1742-4658.2006.05239.x The deoxyhexose sugar fucose has an important fine-tuning role in regula- ting the functions of glycoconjugates in disease and development in mam- mals. The two genetic model organisms Caenorhabditis elegans and Drosophila melanogaster also express a range of fucosylated glycans, and the nematode particularly has a number of novel forms. For the synthesis of such glycans, the formation of GDP-fucose, which is generated from GDP-mannose in three steps catalysed by two enzymes, is required. By homology we have identified and cloned cDNAs encoding these two pro- teins, GDP-mannose dehydratase (GMD; EC 4.2.1.47) and GDP-keto-6- deoxymannose 3,5-epimerase ⁄ 4-reductase (GER or FX protein; EC 1.1.1.271), from both Caenorhabditis and Drosophila. Whereas the nema- tode has two genes encoding forms of GMD (gmd-1 and gmd-2) and one GER-encoding gene (ger-1), the insect has, like mammalian species, only one homologue of each (gmd and gmer). This compares to the presence of two forms of both enzymes in Arabidopsis thaliana. All corresponding cDNAs from Caenorhabditis and Drosophila, as well as the previously uncharacterized Arabidopsis GER2, were separately expressed, and the encoded proteins found to have the predicted activity. The biochemical characterization of these enzymes is complementary to strategies aimed at manipulating the expression of fucosylated glycans in these organisms. Abbreviations GER, GDP-keto-6-deoxymannose 3,5-epimerase ⁄ 4-reductase; GMD, GDP-mannose dehydratase. 2244 FEBS Journal 273 (2006) 2244–2256 ª 2006 The Authors Journal compilation ª 2006 FEBS neural tissue and cells in many invertebrates, inclu- ding Caenorhabditis and Drosophila [9–11]. Not only have the relevant fucosyltransferases and fucose-containing glycans been the object of study, but also the proteins required for the biosynthesis and transport of the fucose donor, GDP-Fuc, have been examined. This nucleotide sugar is generated de novo from GDP-Man in three steps catalysed by two cytoso- lic enzymes: GDP-mannose dehydratase (GMD; EC 4.2.1.47) and GDP-keto-6-deoxymannose 3,5-epi- merase ⁄ 4-reductase (GER, otherwise known as GDP- Fuc synthase; other synonyms include the FX protein in man, the P35B tumour rejection antigen in mice, as well as fcl or wcaG in bacteria and GMER in Drosophila; EC 1.1.1.271) [12]. In eukaryotes, GDP-Fuc is then transported into the Golgi [13], the site of action of, at least the majority of, the fucosyltransferases. To date, sequences encoding GMD and GER have been cloned and expressed from man [14–16], Arabid- opsis thaliana [17–19], Escherichia coli [20], Helicobact- er pylori [21,22] and Paramecium bursaria Chlorella virus 1 [23]. Indeed Arabidopsis has two GMD genes, one of which corresponds to the MUR1 ⁄ GMD2 gene, a defect in which results in deficiencies in cell wall bio- synthesis [24]. GMD is also defective in the Chinese hamster ovary Lec13 and murine lymphoma PL R 1.3 mutant cell lines, and this absence results in resistance to fucose-specific lectins [25,26]. Mice defective in GER suffer from postnatal failure to thrive and an absence of leukocyte selectin ligand expression [27], whereas mutant strains of both the intestinal symbiont Bacteriodes and the nodulation symbiont Sinorhizo- bium fredii unable to produce GDP-Fuc display reduced colonization competitiveness in the presence of wild-type strains [28,29]. There also exists a Dicytos- telium discoideum (slime mould) strain (HL250) with a genetically undefined defect in the conversion of GDP- Man into GDP-Fuc and a resultant reduced germina- tion efficiency for older spores, suggesting that, as for the aforementioned bacterial symbionts, the presence of fucose may confer a selective advantage under natural conditions [30]. However, although early stud- ies were taken to suggest that GMD may be defective in patients with leukocyte adhesion deficiency II (OMIM 266265) [31,32], it now appears to be accepted that mutations in the GDP-Fuc transporter are the reason for the observed reduction in fucosylation [33,34]. On the other hand, the high level of fucosyla- tion in human hepatocellular carcinoma has been cor- related, at least in part, with high expression of GER and increased concentrations of GDP-fucose [35]. Considering that the enzymes involved in GDP-Fuc biosynthesis in the two model invertebrates C. elegans and D. melanogaster have not been studied to date, even though fucosylation appears to be important for their development [4–6], we sought to clone cDNAs predicted to encode GMD and GER genes in these two organ- isms, using previously characterized A. thaliana homo- logues as controls; indeed the encoded proteins were successfully expressed in bacteria and found, in concert, to direct the synthesis of GDP-Fuc in vitro. The two Drosophila enzymes GMD and GMER were also puri- fied; the GDP-Fuc product of these two enzymes was also characterized by NMR and by a functional assay. Results Cloning and expression of Caenorhabditis and Drosophila GMD and GER cDNAs Homologues of the human GMD protein were identified from Caenorhabditis and Drosophila, and the relevant cDNAs cloned. Whereas Drosophila has, as previously determined in a theoretical study [36], one gmd gene (CG8890), Caenorhabditis has, like Arabidopsis [18], two gmd genes (gmd-1 and gmd-2 corresponding to the C53B4.7 and F56H6.5 Wormbase entries), which encode proteins that are 88% identical with each other (see Fig. 1 for alignment). The Caenorhabditis gmd-1 gene is transcribed in two different forms resulting from use of different 5¢ exons (the second and smaller form, C53B4.7a, which is designated gmd-1a in this study); both 5¢ -end gmd-1a EST clones in the databases contain an SL1 spliced leader. In the case of the second worm gene, encoding GMD-2, RT-PCR using a forward pri- mer containing the predicted start codon was unsuccess- ful, as was PCR using forward primers corresponding to the SL1 or SL2 spliced leaders and gmd-2-specific reverse primers. Finally, gmd-2 was cloned in an incom- plete form starting with the second exon, which, how- ever, still contains the first region (Gly-Leu-Glu) conserved in comparison with the gmd-1 cDNAs. As for GMD, homologues of the human GER pro- tein were identified from Caenorhabditis and Droso- phila, and the relevant cDNAs cloned; we also cloned both Arabidopsis homologues. The Drosophila homo- logue has already been named gmer (CG3495) [36], whereas the Caenorhabditis ger-1 corresponds to the R01H2.5 reading frame. As for the GMD enzymes, alignments show a high degree of conservation between GER homologues (Fig. 2). Enzymatic activity of GMD and GER proteins All Arabidopsis, Caenorhabditis and Drosophila GMD and GER homologues were expressed using the S. Rhomberg et al. GDP-fucose biosynthesis in invertebrates FEBS Journal 273 (2006) 2244–2256 ª 2006 The Authors Journal compilation ª 2006 FEBS 2245 pET30a system in the presence of kanamycin and chlo- ramphenicol; in addition, a pCRT7-NT vector carrying Caenorhabditis gmd-1 was also coexpressed with the pET30a clone of Caenorhabditis ger-1 in the presence of ampicillin, kanamycin and chloramphenicol. Western blotting with an antibody to His showed for Fig. 1. Alignment of GMD sequences. The following GMD protein sequences were aligned: C1 (C. elegans GMD-1); C1a (C. elegans GMD-1 alternatively spliced form, first 56 residues only); C2 (C. elegans GMD-2); Dm (D. melanogaster GMD); Hs (Homo sapiens GMD); Sj (Schisto- soma japonica GMD); Ec (E. coli GMD); Pb (P. bursaria Chlorella virus 1 GMD); A1 (A. thaliana GMD1); A2 (A. thaliana MUR1 ⁄ GMD2). Resi- dues conserved in comparison with the fly and worm sequences are highlighted, whereas key conserved residues noted in the Discussion (GXXGXXG as well as the Ser ⁄ Thr residue and YXXXK motif catalytically important for SDR family members) are marked underneath with an asterisk. GDP-fucose biosynthesis in invertebrates S. Rhomberg et al. 2246 FEBS Journal 273 (2006) 2244–2256 ª 2006 The Authors Journal compilation ª 2006 FEBS GMD homologues the expression of proteins of  50 kDa, whereas in the case of the different forms of GER the proteins were  35 kDa (Fig. 3). The Caenor- habditis and Drosophila GMD and GER proteins were tested for activity in coupled enzyme assays. The Ara- bidopsis GMD2 (MUR1) and GER1 proteins were also Fig. 2. Alignment of GER sequences. The following GER protein sequences were aligned: Ce (C. elegans GER-1); Dm (D. melanogaster GMER); Hs (Homo sapiens GER ⁄ FX); Sj (Schistosoma japonica GMD); Ec (E. coli GMD ⁄ wcaG); Pb (P. bursaria Chlorella virus 1 GER); A1 (A. thaliana GER1); A2 (A. thaliana GER2). Residues conserved in comparison with the fly and worm sequences are highlighted, whereas key conserved residues noted in the Discussion (GXXGXXG as well as the Ser ⁄ Thr residue and YXXXK motif catalytically important for SDR family members) are marked underneath with an asterisk. Fig. 3. Western blots of expressed GMD and GER isoforms. GMD and GER proteins from A. thaliana (MUR1, GMD1, GER1 and GER2), C. elegans (GMD-1, GMD-1a, GMD-2, GER-1 and coexpressed GMD-1 and GER-1) and D. melanogaster (GMD and GMER) were expressed in E. coli BL21(DE3)pLysS cells for 2 h, and the soluble fractions of the bacterial proteins (equal lysate equivalents) were subjected to SDS ⁄ PAGE and western blotting using a primary antibody to His. The sizes of the molecular mass standards are shown in kDa. S. Rhomberg et al. GDP-fucose biosynthesis in invertebrates FEBS Journal 273 (2006) 2244–2256 ª 2006 The Authors Journal compilation ª 2006 FEBS 2247 tested as positive controls, as these have been previ- ously shown to be enzymatically active when expressed in E. coli [17,19]. The assays were performed using GDP-Man as sub- strate; incubations were performed with extracts con- taining either of the Arabidopsis, Caenorhabditis or Drosophila enzymes alone or with both enzymes from the various species together. The incubations were then analysed by RP-HPLC, using authentic GDP-Man and GDP-Fuc as external standards. Initially, 0.5 m KH 2 PO 4 was used as eluent [37], but, analogous to the use of ammonium formate buffers for the purification of UDP-xylose [38], it was then decided to examine the use of the formate buffer. As the results with the two buffers were comparable, all subsequent analyses were performed with the volatile formate buffer. Further- more, it was not absolutely necessary to perform the GMD reaction before boiling and then adding GER; such a procedure, though, has been described for assays with E. coli K12 Gmd and WcaG [39]. When GMD ⁄ GER ‘pairs’ of any one of the three species were present, a component that was coeluted with standard GDP-Fuc was produced (Fig. 4). In the case of the Arabidopsis MUR1 and GER1 enzymes, the putative GDP-Fuc product was shown to be a donor substrate in fucosyltransferase assays (data not shown). GDP-Fuc synthesis was also observed when either the Arabidopsis MUR1 or the Caenorhabditis GMD-1a iso- form were incubated with Caenorhabditis GER-1 and GDP-Man (data not shown). In the absence of any GER enzyme, but in the presence of any GMD, a broad peak of intermediate retention time was observed, as shown for Caenorhabditis GMD-1a and Drosophila GMD (Fig. 4A,D); this presumably corres- ponds to the previously observed ketone and hydrate forms of GDP-4-keto-6-deoxymannose [40]. No inter- mediate product was formed in the absence of any GMD enzyme, and no GDP-Fuc was formed in the absence of either GMD or GER, nor with the empty vector control, showing that the strain of E. coli used has no detectable GDP-Fuc synthesis system. The chro- matograms also indicate that the amount of remaining intermediate product was generally low or nonexistent compared with the amount of GDP-Fuc, even though GDP-Man was still present, and that the concentration of GDP-4-keto-6-deoxymannose produced in the pres- ence of GMD isoforms alone was greater than the con- version of GDP-Man into products in the presence of both enzymes (e.g. with Drosophila GMD alone the conversion of GDP-Man into the intermediate was  80%, whereas in the presence of both enzymes the conversion into GDP-Fuc was only  50%; compare Fig. 4A with 4C). This would indeed be compatible with the feedback inhibition by GDP-Fuc previously shown for other forms of GMD also occurring to some extent with the fly and worm enzymes [14,39]. For the Caenorhabditis and Drosophila enzymes, expression at room temperature was necessary to detect activity: for the Drosophila enzymes, no activity was detected on expression at 37 °C, whereas for the Caenor- habditis enzymes, only minimal activity was found on expression at 16 °C. GDP-Fuc synthesis on coexpres- sion of Caenorhabditis GMD-1 and GER-1 was margin- ally less efficient (12%) than synthesis in the presence of both separately expressed enzymes (15–20%) assayed under the same conditions; thus, there is no obvious requirement to coexpress GMD and GER. This is unlike the situation with expression of the Arabidopsis MUR1 in yeast, as in this system MUR1 was susceptible to degradation when not coexpressed with GER1 [37]. Fig. 4. Activity of expressed GMD and GER isoforms. The soluble fractions of lysates (equal lysate equivalents) of bacteria expressing GMD and GER enzymes were incubated overnight with GDP-Man and subjected to RP-HPLC. The chromatograms of the following combinations are shown: (A) Drosophila GMD alone; (B) Drosophila GMER alone; (C) Drosophila GMD and GMER; (D) Caenorhabditis GMD-1a; (E) Caenorhabditis GMD-1 and GER-1; (F) Caenorhabditis GMD-2 and GER-1; (G) Arabidopsis MUR1 (GMD2) and GER1; (H) Arabidopsis GMD1 and GER2. The elution positions of GDP-Man and GDP-Fuc standards are indicated. In the experiments shown, the GMD and GER isoforms were expressed separately. GDP-fucose biosynthesis in invertebrates S. Rhomberg et al. 2248 FEBS Journal 273 (2006) 2244–2256 ª 2006 The Authors Journal compilation ª 2006 FEBS In addition to using MUR1 and GER1 as controls, we also examined the other Arabidopsis homologues of these enzymes, respectively, GMD1 and GER2. Whereas GMD1 has been previously shown to be active [18], GER2 was only identified in silico as a putative epimerase-reductase [19]. The assay data showed that, as for MUR1 and GER1, incubations with both GMD1 and GER2 also resulted in synthesis of GDP-Fuc, con- firming the activity of GER2 for the first time (Fig. 4H). Properties of GMD and GER proteins from different species Initially, the composition of the assay mixture used was based on previously published procedures [14]. There- fore, to test the limits of the system, we performed assays with the Arabidopsis enzymes in the absence of one of each of the nonenzymatic components. Compar- able to previous reports, we found that the synthesis by GER1 of GDP-Fuc from the intermediate product was absolutely dependent on NADPH, whereas reducing the NADPH concentration by half, or increasing it twofold, had no influence on the yield of GDP-Fuc. Furthermore, reduced conversion of the intermediate was observed in the absence of dithiothreitol. On the basis of these data, we did not alter the assay mixture composition for the later assays. However, to optimize the preparation of GDP-Fuc in a ‘one-pot’ method, we also examined the pH and temperature requirements for its production from GDP-Man. The Drosophila enzymes, taken together, displayed a relatively broad pH optimum (pH 5–8), resulting, under the conditions used, in 50% conversion of GDP-Man into GDP-Fuc. Caenorhabditis GMD-1 and GER showed, in combination, optimal activity at pH 8–9 (15–20% conversion using the same amount of soluble bacterial extract as for the Drosophila enzymes); similarly, an optimum at pH 8 was reported for the synthesis of GDP-Fuc by Aerobacter aerogenes and CHO cell extracts [25,41], whereas both the sepa- rately assayed GMD and GER from porcine thyroid show optima at pH  7 [42,43] and recombinant forms of human and E. coli GMD have optima of pH 7.5– 8.0 [32,44]. The recombinant E. coli K-12 GER enco- ded by the wcaG gene was most active in the range pH 6–7 [45]. As regards temperature, the Drosophila enzymes were most active at temperatures of 16–30 °C, whereas the Caenorhabditis enzymes (specifically GMD-1 and GER-1) displayed a temperature optimum of 23–37 °C (Fig. 5). Assays with recombinant GDP-mannose dehydratases alone showed that both Caenorhabditis GMD-1a and Drosophila GMD had temperature optima  30 °C, whereas Caenorhabditis GMD-2 was most active at 16–23 °C (data not shown). Purification of Drosophila GMD and GMER In the preceding studies, the identity of the GDP-Fuc product was based on HPLC retention time; thus, it was decided to purify the product of the fruitfly pro- teins for further analysis. Thus Drosophila GMD and GMER were subjected to nickel-chelation chromato- graphy either separately or together and isolated after elution with 250 mm imidazole (Fig. 6, upper panel). The dominant bands (35 kDa and 50 kDa, corres- ponding to GMER and GMD, respectively) eluted with the latter buffer reacted with an antibody to His (Fig. 6, lower panel), and their identity was verified by MALDI-TOF tryptic peptide mapping. Protein assays indicated that the yields of individually purified Drosophila GMD and GMER were  0.5 mg from a 50-mL culture; when purified together, the total pro- tein yield was 1 mg. Under the conditions used, the yield of GDP-Fuc with the separately purified enzymes was comparable to that using enzymes purified together; it also appeared that, after purification, GMER was more stable than GMD (data not shown). Using the purified forms of Drosophila GMD and GMER, scaled-up incubations were performed, prepu- rified by passage over a small Lichroprep column and subjected to RP-HPLC to yield an estimated total of  1 mg GDP-Fuc (20% yield after purification). This material was lyophilized and used successfully as a fuc- osyltransferase substrate when an extract of Sf9 cells transfected with Drosophila core fucosyltransferase 120 100 80 60 40 20 0 16 23 30 37 Ara Ce Dm Temperature [ºC] % Relative yield Fig. 5. Relative yield of GDP-Fuc with respect to incubation tem- perature. Assays of Arabidopsis MUR1 and GER1, Drosophila GMD and GMER, and Caenorhabditis GMD-1 and GER-1 were performed at different temperatures, and the relevant RP-HPLC peaks were integrated. The data were then recalculated individually for each enzyme pair relative to the respective activity at 23 °C. S. Rhomberg et al. GDP-fucose biosynthesis in invertebrates FEBS Journal 273 (2006) 2244–2256 ª 2006 The Authors Journal compilation ª 2006 FEBS 2249 FucTA was used as an enzyme source as judged by the conversion of the dabsyl-GnGnF 6 glycopeptide sub- strate into a species with an m ⁄ z 146 higher (data not shown). Furthermore, the compound was subjected to NMR, which confirmed its identity as GDP-Fuc (Table 1), the data matching those reported for syn- thetic GDP-Fuc [46]. Developmental expression profile in Caenorhabditis Considering the multiplicity of genes and transcripts encoding GDP-mannose dehydratase in C. elegans, semi-normalized RT-PCR was performed using cDNA from L1 larvae, L2 ⁄ 3 larvae (combined as these are difficult to distinguish), L4 larvae and adults. The results (Fig. 7) would suggest minor variations in the concentrations of the gmd-2 and ger-1 transcripts dur- ing worm development. A peak of gmd-1 transcription may be occurring in the L2 ⁄ 3 stage, but transcripts of this form are seemingly under-represented in adults. Fig. 6. Purification of recombinant Drosophila GMD and GMER. GMD and GMER expressed separately were subjected to nickel chelation chromatography; fractions marked ‘co’ are from the purif- ication of GMD and GMER from mixed lysates. Fractions (wash, 20 m M imidazole and 250 mM imidazole) were then electrophor- esed and stained using Coomassie blue (upper panel) or transferred to nitrocellulose and probed with an antibody to His (lower panel). In the case of copurification, some GMER, but no GMD, was elut- ed with 20 m M imidazole (lane 4). The sizes of the molecular mass standards are shown in kDa. Table 1. NMR data of GDP-b-L-fucose prepared using recombinant Drosophila enzymes. ND, not determined. Further signals at 3.76 p.p.m. in the proton spectrum and at 60.16 in the carbon spectrum are from residual Tris buffer. Atom H ⁄ C ⁄ P (p.p.m.) 1 2 3 4 5 6 b-Fuc1 fi 1 H 4.94 3.58 3.69 3.72 3.77 1.25 J (Hz) 6.5 10.0 3.5 ND 6.6 13 C 99.19 71.81 73.30 72.24 71.98 16.22 b-Rib1 fi 1 H 5.95 4.84 4.56 4.37 4.25–4.22 (2H) J (Hz) 6.4 5.2 3.5 ND ND 13 C 87.57 74.27 71.32 84.76 66.17 b-Fuc1-P- 31 P ) 12.65 J (Hz) J PP 20.5, J HP 8.0 P-5-Rib 31 P ) 10.8 Guanine 1 H 8.13 13 C 138.48 Fig. 7. Development RT-PCR profile for GMD and GER transcripts in Caenorhabditis. RT-PCR was performed using RNA isolated from L1, L2 ⁄ L3, L4 and adult C. elegans using primers specific for gmd-1, gmd-1a (alternatively spliced form of GMD-1), gmd-2 and ger-1. The amounts of cDNA used in the PCRs were normalized on the basis of the intensity of actin transcripts. GDP-fucose biosynthesis in invertebrates S. Rhomberg et al. 2250 FEBS Journal 273 (2006) 2244–2256 ª 2006 The Authors Journal compilation ª 2006 FEBS On the other hand, the alternatively spliced gmd-1a transcript is present at its lowest concentrations in L1 larvae and is relatively more abundant in the later sta- ges. The expression of the GDP-Fuc biosynthesizing enzymes throughout development is compatible with the rich variety of fucosylated N-glycans and O-gly- cans in this species [47,48]. Discussion GDP-fucose was first found in 1958 [49], and its bio- synthesis is a process present in many life forms, from bacteria through to plants, invertebrates and verte- brates. There appear to be two basic strategies for the formation of GDP-Fuc: either the route through GDP-Man, using GMD and GER enzymes, or the ‘salvage’ pathway through fucose 1-phosphate. The first route was first found in A. aerogenes and shown to be dependent on the presence of NADPH (then called TPNH) [50]. GMD was first isolated from Phaseolus vulgaris [51], whereas GER was initially purified from porcine thyroid [43]. The presence of a fucose salvage pathway was first suggested in 1964 because of the ability to radiolabel glycoproteins after administration of [ 14 C]Fuc to rats [52] and confirmed by detection of l-fucose kinase and GDP-l-Fuc pyro- phosphorylase [53,54] in porcine liver. More recent studies indicate that there are varying levels of fuco- kinase activity in different rat tissues [55]. Some organisms have both pathways, as shown by biochemical work; initially both routes were consid- ered to be only present in mammals, but now the two pathways have been demonstrated in Bacteriodes [28]. The genomic ‘revolution’, however, means that further phylogenetic analyses can now be performed. By this approach, it can be seen that the GMD ⁄ GER route is probably present in all organ- isms known to produce fucose-containing glycoconju- gates; on the other hand, as noted previously, Drosophila has no genetically detectable ‘salvage’ pathway. Plants and mammals do have relevant homologues, although the putative plant ‘salvage’ pathway is seemingly closer to that of Bacteriodes, as plant genomes contain homologues of the fkp gene from Bacteriodes, a gene that encodes a protein with both fucokinase and GDP-Fuc phosphorylase activities [28]. In mammals, however, these activities are encoded by separate genes. Caenorhabditis appears, on the other hand, only to have an obvious fucokinase homologue (C26D10.4). In addition, GMD is also required for the de novo synthesis of GDP-Rha in Ps. aeruginosa, as the product of GMD, GDP-4-keto-6-deoxy-d-mannose, can also be acted on by a reductase [56], whereas the GMD of the P. bursaria Chlorella virus 1 can directly convert GDP-4-keto-6-deoxy-d-mannose into GDP-Rha [23]. Thus it is conceivable that GMD is more ancient than GER. The GMD and GER sequences across the various kingdoms of life are remarkably highly conserved; both proteins are members of the short chain dehy- drogenase (SDR) family and display homologies to other enzymes of sugar nucleotide metabolism, such as dTDP-glucose dehydrogenase, UDP-Gal epimerase and UDP-GlcA decarboxylase. Phylogenetic trees (not shown) suggest that the plant enzymes are closer to the bacterial, than to the animal, ones; regardless of this, however, residues found by crystallographic or mutagenesis studies to be important for binding or catalysis are identical across all sequences. A Ross- mann motif (Gly-Xaa-Xaa-Gly-Xaa-Xaa-Gly), which is a common feature of nucleotide-binding proteins, is, for instance, conserved in all the GMD and GER sequences from worm and fly. Furthermore, the resi- dues corresponding to Gln39, Asp40, Ser117 and Arg220 in MUR1 3D structure, which form hydrogen bonds with the NADPH cofactor, and the residues that correspond to Asn214, Lys228, Arg253, Arg314 and Glu317 of the MUR1 sequence and form hydro- gen bonds with the GDP moiety [57] are retained in the worm and fly GMD enzymes. The Ser ⁄ Thr residue and Tyr-Xaa-Xaa-Xaa-Lys motif catalytically import- ant for SDR family members are also conserved in all sequences. If the corresponding Ser ⁄ Thr, Tyr and Lys residues of the E. coli GMD or GER are separately subjected to site-directed mutagenesis, then either activity is abolished or the k cat drastically decreased [58,59]. It is also noteworthy that some organisms have mul- tiple GMD or GER genes. In particular, Arabidopsis has two proven GMD enzymes (GMD1 and MUR1; At5g66280 and At3g51160), displaying differential expression [18], as well as the previously proven GER1 and now, by us, proven GER2 (respectively, At1g73250 and At1g17890): in both cases the genes are in, or at least close to, regions that have putatively been duplica- ted during the evolution of Arabidopsis [60,61]. Further- more, the presence of duplicated genes means that knocking-out one GMD, i.e. MUR1, does not totally diminish the fucose content of Arabidopsis glycoconju- gates [24]. Any strategy to abolish all fucosylation in plants is possibly also complicated by the presence of the aforementioned fkp homologue. On the other hand, Drosophila has only one GMD and one GER homo- logue; indeed, a GMD mutation has been isolated and is lethal at the third larval stage [62], commensurate S. Rhomberg et al. GDP-fucose biosynthesis in invertebrates FEBS Journal 273 (2006) 2244–2256 ª 2006 The Authors Journal compilation ª 2006 FEBS 2251 with the putative key role for peptide O-fucosyl- transferases in development and the probable lack of any salvage pathway. C. elegans, however, is somewhere between these extremes, as it has two GMD enzymes (although the related nematode Caenorhabditis briggsae appears to have only one gmd gene, suggesting that the duplication of gmd genes is an evolutionarily relatively recent event), whose activities were proven in the course of our studies, but only one GER isoform. Suggestive of functional degeneracy are RNAi data on the two Caenorhabditis GMD homologues: at least when performed individu- ally, as part of a large-scale screen, RNAi of gmd-1, gmd-2 and ger-1 resulted in no obvious associated lethal- ity. However, in another large-scale RNAi screen with the hypersensitive rrf-3 worm strain, various defects were indeed reported upon knock-down of gmd-2 [63]; no data, however, on gmd-1 or gmd-1a were reported in the study using rrf-3 worms, so neither the relative importance of the two genes nor the biological signifi- cance of the alternative splicing of gmd-1 can be sur- mised at present. Owing to the previous report as to the effect of mutations in the C. elegans POFUT2 homo- logue pad-2 [6], the Drosophila orthologue of which modifies thrombospondin repeats [64], it is quite prob- able that any effects of RNAi targeting of GDP-Fuc biosynthesizing enzymes will be due to peptide O-fuco- sylation defects. There may also be tissue-specific expression of the two gmd genes. Apparently, gmd-1 is expressed in body wall muscle and head neurons. (For summaries of the various RNAi and expression data, see: http://www.wormbase.org/db/gene/gene?name ¼ C53B4.7 for gmd-1, http://www.wormbase.org/db/gene/ gene?name ¼ F56H6.5 for gmd-2 and http:// www.wormbase.org/db/gene/gene?name ¼ R01H2.5 for ger-1.). Our own developmental RT-PCR profile data would suggest that there is no major developmental regulation of the transcription of either GMD-encoding gene, although there appears to be a peak of gmd-1 and gmd-1a expression at the L2 ⁄ L3 stage; such results, however, are not incompatible with variation of expres- sion within tissues and may reflect a requirement for higher concentrations of GDP-Fuc at certain times or in certain tissues during development (e.g. for Notch signaling). In summary, we have shown for the first time that the GMD and GER homologues of Caenorhabditis and Drosophila, as well as the Arabidopsis GER2, are indeed functional enzymes, which can work together to reconstitute GDP-Fuc synthesis in vitro. Biochemical characterization of these enzymes lends confidence to any subsequent reverse genetic or phylogenetic studies or in the use of conditional mutants and lays the foun- dation for future work on the role of fucose in the biology of these model organisms. Experimental procedures Cloning of GMD and GER cDNAs RNA was extracted from A. thaliana (Columbia), C. elegans (N2) or D. melanogaster (Canton S) using Trizol reagent (Invitrogen, Paisley, UK). Two-step RT-PCR was performed using Superscript III reverse transcriptase (Invitrogen) and Table 2. Primers used in this study. AtGMD1 AtGMD1 ⁄ 1 ⁄ NcoI, CATGCCATGGCCTCCAGATCTCTC (fwd) AtGMD1 ⁄ 2 ⁄ EcoRI, CGGAATTCAAGGTCGTGCTGAGCTC (rev) AtMUR1 AtMUR1 ⁄ 1 ⁄ NcoI, CATGCCATGGCGTCAGAGAACAACGG (fwd) AtMUR1 ⁄ 2 ⁄ XhoI, ACCCTCGAGTCAAGGTTGCTGCTTAGC (rev) AtGER1 AtGER1 ⁄ 1 ⁄ NcoI, CATGCCATGGCTGACAAATCTGCC (fwd) AtGER1 ⁄ 2 ⁄ XhoI, ACCCTCGAGTTATCGGTTGCAAACATTCTT (rev) AtGER2 AtGER2 ⁄ 1 ⁄ NcoI, CATGCCATGGAATCAGGTTCGTTTATGTTA (fwd) AtGER2 ⁄ 2 ⁄ XhoI, CCGCTCGAGTTACTGCTTCTTCTGCACAA (rev) CeGMD-1 CeGMD1 ⁄ 1 ⁄ NcoI, CATGCCATGGCAACCGGCAAGTCTG (fwd), CeGMD ⁄ 1 ⁄ BamHI, CGGGATCCAATGCCAACCGGCAAGTCTG (fwd), or CeGMD1a ⁄ 1 ⁄ NcoI, CATGCCATGGCTGATCAAAATGCGAA (fwd) CeGMD1 ⁄ 2 ⁄ HindIII, CCCAAGCTTAAGCCATTGGATTGGACTTC (rev) CeGMD-2 CeGMD2 ⁄ 3 ⁄ NcoI, CATGCCATGGGTCTCGAATCATGTATTGA (fwd) CeGMD2 ⁄ 1 ⁄ BamHI, CGGGATCCTAAGCCATTGGATCTGCC (rev) CeGER-1 CeGER ⁄ 1 ⁄ NcoI, CATGCCATGGCTAAAACTATTCTAGTTACT (fwd) CeGER ⁄ 2 ⁄ EcoRI, CGGAATTCTTATTTTCTAGCCGTCTCATAA (rev) DmGMD DmGMD ⁄ 1 ⁄ BamHI, CGGGATCCATGCTAAATACCCGGC (fwd) DmGMD ⁄ 2 ⁄ XhoI, CCGCTCGAGTTAAGCGATTGGATTTTTCCT (rev) DmGMER DmGMER ⁄ 1 ⁄ BamHI, CGGGATCCATGAAGAAGGTTCTGGTTA (fwd) DmGMER ⁄ 2 ⁄ XhoI, CCGCTCGAGTTACTTTCTAGCCTGGTCG (rev) GDP-fucose biosynthesis in invertebrates S. Rhomberg et al. 2252 FEBS Journal 273 (2006) 2244–2256 ª 2006 The Authors Journal compilation ª 2006 FEBS Expand polymerase (Roche, Vienna, Austria) using the pri- mer pairs listed in Table 2. The PCR fragments were cut and ligated into either pET30a (Novagen, Merck Biosciences, Darmstadt, Germany) or pCRT7-NT (Invitrogen) digested with the relevant restriction enzyme(s). DNA sequencing was performed using the BigDye kit (Applera, Norwalk, CT, USA). In the case of the Caenorhabditis gmd-1 and ger-1 pET clones, the second codon encodes alanine (respectively a replacement codon or an additional one) in order to accom- modate an NcoI site. Expression of GMD and GER proteins Plasmids were used to transform BL21(DE3)pLysS Gold cells (Stratagene, Amsterdam, the Netherlands), which were grown overnight in 10 mL Luria–Bertani medium containing kanamycin and chloramphenicol (also containing ampicillin in the case of double transformation). In the case of trial expression, 1 mL (or 2.5 mL for larger-scale cultures) was taken from the overnight culture to inoculate 20 mL (or 50 mL) Luria–Bertani medium containing the relevant anti- biotics. After the A 600 had reached  0.6 at 37 °C, small-scale cultures were split, and to one half was added isopropyl b-d-thiogalactoside to a concentration of 1 mm; in the case of larger-scale cultures, isopropyl b-d-thiogalactoside was added to the entire culture. The growth was continued at 23 °C for up to three hours. Cells were resuspended in 500 lL (small-scale) or 5 mL (large-scale) lysis buffer containing 50 mm Tris, 400 mm NaCl, 100 mm KCl, 10% glycerol, 0.5% Triton X-100, 10 mm imidazole, pH 7.8, and lysed by performing repeated freeze–thaw cycles, using alternately a methanol bath and a 42 °C water bath. DNase I was added, and the lysates were incubated for 10 min at 37 ° C before centrifugation for 1 min (small-scale) or 20 min (large-scale) at 14 000 g, 4 °C. The supernatant was taken for assays or, in the case of large-scale cultures, purification. For the presented data, the cells were always grown and lysed under the same con- ditions (i.e. same initial cell density, temperature, time of induction and concentration of isopropyl b-d-thiogalacto- side). Aliquots of these lysates stored at )80 °C still dis- played activity after 1 year of storage. Purification by nickel-chelation chromatography The supernatants from lysed cells were incubated with 2 mL Ni ⁄ nitrilotriacetate resin (Qiagen, Vienna, Austria) for at least 1 h at 4 °C. The lysate ⁄ resin mixture was poured into a column at room temperature and washed twice with 1 mL lysis buffer, before further washing twice with 4-mL aliquots of a lysis buffer containing 20 mm imidazole. Elution was performed using four 0.5-mL aliquots of a lysis buffer con- taining 250 mm imidazole. All fractions were collected on ice. Protein assays were performed using the modified Lowry kit (Sigma, Vienna, Austria). Western blotting Aliquots of the soluble fractions of lysed bacteria or of affinity chromatography fractions (20 lL) were precipitated with a fivefold excess of cold methanol and, after 1 h at )20 °C, centrifuged (14 000 g, 5 min). After removal of residual methanol at 65 °C, the samples were resuspended in Laemmli sample buffer (20 lL) and denatured at 95 °C for 5 min; 5 lL of these samples were subjected to SDS ⁄ PAGE with subsequent blotting on to nitrocellulose. Recombinant His-tagged forms of GMD and GER were then detected using antibody to His (HIS-1; Sigma; 1 : 3000 dilution) followed by anti-mouse IgG (Fc or c-specific) con- jugated with alkaline phosphatase (Sigma; 1 : 10 000 dilu- tion) and use of SigmaFAST BCIP ⁄ NBT substrate. Assay of GMD and GER activity To determine the enzymatic activity, aliquots of crude sup- ernatants of lysed cells or of purified proteins (2 lL) were typically incubated at room temperature in the presence of 20 mm Tris ⁄ 5mm EDTA ⁄ 10 mm dithiothreitol ⁄ 1mm GDP-mannose ⁄ 5mm NADPH ⁄ 1mm NAD + , pH 7.4 (final volume 10 lL). RP-HPLC was then performed using a Hypersil column with isocratic elution using 600 mm ammonium formate, pH 3.2. Peak integrations were used to estimate the yield of either GDP-4-keto-6-deoxymannose (GMD assays) or GDP-Fuc (combined GMD ⁄ GER assays). NMR analysis Approximately 1 mg of the HPLC-purified reaction product of GDP-Man with purified Drosophila GMD and GMER was lyophilized twice and taken up in D 2 O before NMR analysis. Spectra were recorded at 300 K at 300.13 MHz for 1 H, at 75.47 MHz for 13 C, and at 121.49 MHz for 31 P with a Bruker AVANCE 300 spectrometer equipped with a 5-mm QNP-probehead with z gradients. Data acquisition and processing were performed with the standard xwinnmr software (Bruker BioSpin GmbH, Rheinstetten, Germany). 1 H-NMR spectra were referenced to 2,2-dimethyl-2-silapen- tane-5-sulfonic acid (d ¼ 0), 13 C-NMR spectra were refer- enced externally to 1,4-dioxane (d ¼ 67.40), and 31 P-NMR spectra were referenced externally to H 3 PO 4 (d ¼ 0). HMQC and HMBC spectra were recorded in the phase- sensitive mode using TPPI and pulsed field gradients for coherence selection. Developmental transcript analysis A total of 120 individual Caenorhabditis L1 larvae and 60 individuals of three other stages (L2 ⁄ L3 larvae, L4 larvae and adults) were picked; the RNA, extracted using S. Rhomberg et al. GDP-fucose biosynthesis in invertebrates FEBS Journal 273 (2006) 2244–2256 ª 2006 The Authors Journal compilation ª 2006 FEBS 2253 [...]... of 10 lL in order to judge relative intensities For the subsequent experiments, the volumes of cDNA used were corrected so as to attain approximately the same intensities for the actin PCR products Seminormalized RT-PCR was then performed using the primer pairs listed in Table 2 for gmd-1 (and its alternatively spliced form), gmd-2, ger-1 and actin using an annealing temperature of 58 °C and 1 lL diluted... We also thank Sandra Drozd, a previous project student in the laboratory, for the initial cloning of the Caenorhabditis gmd-1 and ger-1 cDNAs, and Dr Andreas Hofinger for recording the NMR spectra References 1 Becker DJ & Lowe JB (2003) Fucose: biosynthesis and biological function in mammals Glycobiology 13, 41R– 53R 2 Haltiwanger RS (2002) Regulation of signal transduction pathways in development by... Synthesis of GDP-L-fucose by the human FX protein J Biol Chem 273, 27274–27279 17 Bonin CP, Potter I, Vanzin GF & Reiter W-D (1997) The MUR1 gene of Arabidopsis thaliana encodes an isoform of GDP-D-mannose-4,6-dehydratase, catalyzing the first step in the de novo synthesis of GDP-L-fucose Proc Natl Acad Sci USA 271, 2085–2090 18 Bonin CP, Freshour G, Hahn MG, Vanzin GF & Reiter W-D (2003) The GMD1 and. . .GDP-fucose biosynthesis in invertebrates S Rhomberg et al Trizol reagent in the presence of glass beads, was taken up in 10 lL nuclease-free water After reverse transcription with Superscript III of each entire RNA preparation, the cDNAs were diluted and subjected to a preliminary PCR screen using 1 lL of a 1 : 10 000 cDNA dilution and Expand polymerase with actin-specific primers in a final volume of. .. synthesis of GDP-Lfucose in Arabidopsis Plant J 21, 445–454 Andrianopoulos K, Wang L & Reeves PR (1998) Identification of the fucose synthetase gene in the colanic acid gene cluster of Escherichia coli K-12 J Bacteriol 180, 998–1001 Jarvinen N, Maki M, Rabina J, Roos C, Mattila P & ¨ ¨ ¨ ¨ Renkonen R (2001) Cloning and expression of Helicobacter pylori GDP-L-fucose synthesising enzymes (GMD and GMER) in. .. and the requisite fucosyltransferase in Drosophila melanogaster Potential basis of the neural anti-horseradish peroxidase epitope J Biol Chem 276, 28058–28067 10 Paschinger K, Rendic D, Lochnit G, Jantsch V & Wilson IBH (2004) Molecular basis of anti-horseradish peroxidase staining in Caenorhabditis elegans J Biol Chem 279, 49588–49598 11 Rendic D, Linder A, Paschinger K, Borth N, Wilson IBH & Fabini... course and steady state mechanism of the reaction catalysed by GDP-fucose synthetase from Escherichia coli J Biol Chem 274, 26743–26750 46 Gokhale UB, Hindsgaul O & Palcic MM (1990) Chemical synthesis of GDP-fucose analogs and their utilization by the Lewis a (1 fi 4) fucosyltransferase Can J Chem 68, 1063–1071 47 Haslam SM & Dell A (2003) Hallmarks of Caenorhabditis elegans N-glycosylation: complexity and. .. A & Wandrey C (2004) Kinetic examination and simulation of GDP-b-L-fucose synthetase reaction using NADPH or NADH Biocatalysis and Biotransformation 22, 49–56 Ginsburg V (1960) Formation of guanosine diphosphate l-fucose from guanosine diphosphate d-mannose J Biol Chem 235, 2196–2201 Broschat KO, Chang S & Serif G (1985) Purification and characterisation of GDP-D-mannose 4,6-dehydratase from porcine... J, Chang XJ, Boodhoo A, Potvin B & Cumming DA (1998) Molecular cloning of human GDP mannose 4,6-dehydratase and reconstitution of GDP-fucose biosynthesis in vitro J Biol Chem 273, 8193–8202 15 Ohyama C, Smith PL, Angata K, Fukuda MN, Lowe JB & Fukuda M (1998) Molecular cloning and expression of GDP-D-mannose-4,6-dehydratase, a key enzyme for fucose metabolism defective in Lec13 cells J Biol Chem 273,... (1968) The metabolism of lfucose III The biosynthesis of guanosine diphosphate l-fucose in porcine liver J Biol Chem 243, 1110–1115 54 Ishihara H, Massaro DJ & Heath EC (1968) The metabolism of 1-fucose IV The enzymatic synthesis of b-lfucose 1-phosphate J Biol Chem 243, 1103–1109 55 Miller EN, Rupp AL, Lindberg MK & Wiese TJ (2005) Tissue distribution of l-fucokinase in rodents Comp Biochem Physiol B . Reconstitution in vitro of the GDP-fucose biosynthetic pathways of Caenorhabditis elegans and Drosophila melanogaster Simone Rhomberg 1 , Christina. in 10 mL Luria–Bertani medium containing kanamycin and chloramphenicol (also containing ampicillin in the case of double transformation). In the case of