Portal de Periódicos da CAPES

Glucosylceramide Synthases, a Gene Family Responsible for the Biosynthesis of Glucosphingolipids in Animals, Plants, and Fungi

2001; Elsevier BV; Volume: 276; Issue: 36 Linguagem: Inglês

10.1074/jbc.m104952200

1083-351X

Martina Leipelt, Dirk Warnecke, Ulrich Zähringer, Claudia Ott, Frank Müller, Bernhard Hube, Ernst Heinz,

Polysaccharides and Plant Cell Walls

Glucosylceramides are membrane lipids in most eukaryotic organisms and in a few bacteria. The physiological functions of these glycolipids have only been documented in mammalian cells, whereas very little information is available of their roles in plants, fungi, and bacteria. In an attempt to establish appropriate experimental systems to study glucosylceramide functions in these organisms, we performed a systematic functional analysis of a glycosyltransferase gene family with members of animal, plant, fungal, and bacterial origin. Deletion of such putative glycosyltransferase genes in Candida albicans and Pichia pastorisresulted in the complete loss of glucosylceramides. When the corresponding knock-out strains were used as host cells for homologous or heterologous expression of candidate glycosyltransferase genes, five novel glucosylceramide synthase (UDP-glucose:ceramide glucosyltransferase) genes were identified from the plantGossypium arboreum (cotton), the nematodeCaenorhabditis elegans, and the fungi Magnaporthe grisea, Candida albicans, and P. pastoris. The glycosyltransferase gene expressions led to the biosynthesis of different molecular species of glucosylceramides that contained either C18 or very long chain fatty acids. The latter are usually channeled exclusively into inositol-containing sphingolipids known from Saccharomyces cerevisiae and other yeasts. Implications for the biosynthesis, transport, and function of sphingolipids will be discussed. Glucosylceramides are membrane lipids in most eukaryotic organisms and in a few bacteria. The physiological functions of these glycolipids have only been documented in mammalian cells, whereas very little information is available of their roles in plants, fungi, and bacteria. In an attempt to establish appropriate experimental systems to study glucosylceramide functions in these organisms, we performed a systematic functional analysis of a glycosyltransferase gene family with members of animal, plant, fungal, and bacterial origin. Deletion of such putative glycosyltransferase genes in Candida albicans and Pichia pastorisresulted in the complete loss of glucosylceramides. When the corresponding knock-out strains were used as host cells for homologous or heterologous expression of candidate glycosyltransferase genes, five novel glucosylceramide synthase (UDP-glucose:ceramide glucosyltransferase) genes were identified from the plantGossypium arboreum (cotton), the nematodeCaenorhabditis elegans, and the fungi Magnaporthe grisea, Candida albicans, and P. pastoris. The glycosyltransferase gene expressions led to the biosynthesis of different molecular species of glucosylceramides that contained either C18 or very long chain fatty acids. The latter are usually channeled exclusively into inositol-containing sphingolipids known from Saccharomyces cerevisiae and other yeasts. Implications for the biosynthesis, transport, and function of sphingolipids will be discussed. glucosylceramide(s) coding region glucosylceramide synthase polymerase chain reaction base pairs high performance liquid chromatography nucleotide recognition domain group of overlapping clones long chain fatty acid(s) very long chain fatty acid(s) Glycosylceramides are present in almost all eukaryotic organisms and also in a few bacteria. The structures of the sugar headgroups and the ceramide backbones of many different glycosylceramides from animals (1Stults C.L. Sweeley C.C. Macher B.A. Methods Enzymol. 1989; 179: 167-214Crossref PubMed Scopus (234) Google Scholar), plants (2Cahoon E.B. Lynch D.V. Plant Physiol. 1991; 95: 58-68Crossref PubMed Scopus (86) Google Scholar, 3Lynch D.V. Moore Jr., T.S. Lipid Metabolism in Plants. CRC Press, Inc., Boca Raton, FL1993: 285-308Google Scholar, 4Heinz E. Christie W.W. Advances in Lipid Methodology Three. The Oily Press, Dundee, UK1996: 211-332Google Scholar), fungi (5Dickson R.C. Lester R.L. Biochim. Biophys. Acta. 1999; 1426: 347-357Crossref PubMed Scopus (171) Google Scholar, 6Sakaki T. Zähringer U. Warnecke D.C. Fahl A. Knogge W. Heinz E. Yeast. 2001; 18: 679-695Crossref PubMed Scopus (88) Google Scholar, 7Toledo M.S. Levery S.B. Straus A.H. Suzuki E. Momany M. Glushka J. Moulton J.M. Takahashi H.K. Biochemistry. 1999; 38: 7294-7306Crossref PubMed Scopus (102) Google Scholar, 8Toledo M.S. Levery S.B. Suzuki E. Straus A.H. Takahashi H.K. Glycobiology. 2001; 11: 113-124Crossref PubMed Scopus (46) Google Scholar), and bacteria (9Kawahara K. Moll H. Knirel Y.A. Seydel U. Zähringer U. Eur. J. Biochem. 2000; 267: 1837-1846Crossref PubMed Scopus (99) Google Scholar) have been analyzed in detail. The biosynthesis of glucosylceramides (GlcCer)1 is catalyzed by a UDP-glucose:ceramide glucosyltransferase (glucosylceramide synthase (GCS), EC 2.4.1.80), which was originally found in animal tissues (10Basu S. Kaufman B. Roseman S. J. Biol. Chem. 1968; 243: 5802-5804Abstract Full Text PDF PubMed Google Scholar). Recently, the first cDNA coding for a human GCS has been cloned (11Ichikawa S. Sakiyama H. Suzuki G. Hidari K.I. Hirabayashi Y. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 4638-4643Crossref PubMed Scopus (221) Google Scholar). This success has opened new possibilities for analyzing GlcCer functions (12Liu Y.Y. Han T.Y. Giuliano A.E. Cabot M.C. FASEB J. 2001; 15: 719-730Crossref PubMed Scopus (255) Google Scholar, 13Tepper A.D. Diks S.H. van Blitterswijk W.J. Borst J. J. Biol. Chem. 2000; 275: 34810-34817Abstract Full Text Full Text PDF PubMed Scopus (52) Google Scholar), particularly by studying the phenotypes resulting from gene deletions in knock-out mice (14Yamashita T. Wada R. Sasaki T. Deng C. Bierfreund U. Sandhoff K. Proia R.L. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 9142-9147Crossref PubMed Scopus (407) Google Scholar). These studies address different aspects of GlcCer functions: (i) GlcCer are membrane lipids and contribute to the physical properties and physiological functions of membranes; (ii) GlcCer serves as basic precursor for over 300 species of glycosphingolipids found in different mammalian cell types; and (iii) GlcCer synthesis and degradation are believed to contribute to the control of the level of ceramide, which is regarded as a second messenger involved in many biological processes such as heat stress response and apoptosis (15Huwiler A. Kolter T. Pfeilschifter J. Sandhoff K. Biochim. Biophys. Acta. 2000; 1485: 63-99Crossref PubMed Scopus (383) Google Scholar, 16Dickson R.C. Nagiec E.E. Skrzypek M. Tillman P. Wells G.B. Lester R.L. J. Biol. Chem. 1997; 272: 30196-30200Abstract Full Text Full Text PDF PubMed Scopus (220) Google Scholar, 17Jenkins G.M. Richards A. Wahl T. Mao C. Obeid L. Hannun Y. J. Biol. Chem. 1997; 272: 32566-32572Abstract Full Text Full Text PDF PubMed Scopus (261) Google Scholar). In addition, studies on the intracellular location and transport of GlcCer contribute to the understanding of the apical/basolateral asymmetry of epithelial cells (18VanMeer G. Holthuis J.C.M. Biochim. Biophys. Acta. 2000; 1486: 145-170Crossref PubMed Scopus (135) Google Scholar). In contrast, very little is known about the functions and intracellular location of GlcCer in nonanimal organisms such as plants, fungi, and bacteria. Progress in this field is hampered by the lack of a genetic approach, since no genes or cDNAs coding for GCS have been cloned or identified from these organisms so far (except for our preliminary communication on a putative GCS from Candida albicans(19Leipelt M. Warnecke D.C. Hube B. Zähringer U. Heinz E. Biochem. Soc. Trans. 2000; 28: 751-752Crossref PubMed Scopus (15) Google Scholar)). The aim of the present work was the cloning of GCS from nonanimal organisms and establishment of novel model systems suitable for alterations in GlcCer content. These genetically modified organisms will enable investigations on various aspects of GlcCer synthesis, transport, and functions, which should extend and complement the studies limited so far to mammalian cells. Apart from this general approach, we address two aspects of sphingolipid metabolism, which are characteristic for fungi. First, there is growing evidence that these organisms maintain two separate pools of ceramides to be used for the synthesis of different sphingolipids. Ceramide backbones with C16 or C18 fatty acids linked to a 4,8-diene-9-methyl-sphingobase are exclusively precursors for GlcCer synthesis, whereas ceramide backbones with very long chain C24 and C26 fatty acids bound to phytosphinganine are restricted to the synthesis of the inositol-containing phosphosphingolipids (6Sakaki T. Zähringer U. Warnecke D.C. Fahl A. Knogge W. Heinz E. Yeast. 2001; 18: 679-695Crossref PubMed Scopus (88) Google Scholar, 7Toledo M.S. Levery S.B. Straus A.H. Suzuki E. Momany M. Glushka J. Moulton J.M. Takahashi H.K. Biochemistry. 1999; 38: 7294-7306Crossref PubMed Scopus (102) Google Scholar, 8Toledo M.S. Levery S.B. Suzuki E. Straus A.H. Takahashi H.K. Glycobiology. 2001; 11: 113-124Crossref PubMed Scopus (46) Google Scholar, 20Toledo M.S. Levery S.B. Straus A.H. Takahashi H.K. FEBS Lett. 2001; 493: 50-56Crossref PubMed Scopus (31) Google Scholar). The mechanism of this separation of two sphingolipid pools is not understood. In the present study, we report the heterologous and homologous expression of various GCS in different yeasts and demonstrate that the enzymes can accept both ceramide pools as substrates. Furthermore, GlcCer of fungal origin have been found to act as elicitors of a plant defense response involving the transcription of specific genes. The structural features required to induce this effect have been identified in studies with different molecular species of GlcCer from the phytopathogen Magnaporthe grisea (21Koga J. Yamauchi T. Shimura M. Ogawa N. Oshima K. Umemura K. Kikuchi M. Ogasawara N. J. Biol. Chem. 1998; 273: 31985-31991Abstract Full Text Full Text PDF PubMed Scopus (135) Google Scholar, 22Umemura K. Ogawa N. Yamauchi T. Iwata M. Shimura M. Koga J. Plant Cell Physiol. 2000; 41: 676-683Crossref PubMed Scopus (71) Google Scholar). The identification of downstream members of this signaling cascade is a prerequisite for understanding the molecular mechanism of this plant-pathogen interaction. Therefore, we also carried out a detailed analysis of the GlcCer molecular species accumulating in P. pastoris as a consequence of the heterologous expression of GCS. Escherichia coli strain XL1-Blue (MRF′) and the vector pBluescript (Stratagene, La Jolla, CA) were used for cloning of the putative GCS genes and cDNAs. Expression of these sequences inSaccharomyces cerevisiae, P. pastoris, andC. albicans was performed with pVT-U (23Vernet T. Dignard D. Thomas D.Y. Gene (Amst.). 1987; 52: 225-233Crossref PubMed Scopus (465) Google Scholar), pYES2, pGAPZ, pPIC3.5 (Invitrogen), pBI-1 (24Stoldt V.R. Sonneborn A. Leuker C.E. Ernst J.F. EMBO J. 1997; 16: 1982-1991Crossref PubMed Scopus (517) Google Scholar). Yeast strains used in this study wereS. cerevisiae UTL-7A (MATa, ura3–52, trp1, leu2–3, 112), UTL-7AΔugt51 (MATa, ura3–52, trp1, leu2–3, 112, ugt51::kanMX4; Ref. 25Warnecke D. Erdmann R. Fahl A. Hube B. Müller F. Zank T. Zähringer U. Heinz E. J. Biol. Chem. 1999; 274: 13048-13059Abstract Full Text Full Text PDF PubMed Scopus (102) Google Scholar), and Sc334 (26Hovland P. Flick J. Johnston M. Sclafani R.A. Gene (Amst.). 1989; 83: 57-64Crossref PubMed Scopus (143) Google Scholar);P. pastoris GS115 (Invitrogen); and C. albicansSC5314 CAI4 Δura3::imm434/Δura3::imm434, congenic to SC5314 (27Fonzi W.A. Irwin M.Y. Genetics. 1993; 134: 717-728Crossref PubMed Google Scholar). A cDNA clone of the human GCS was a gift from Dr. S. Ichikawa and Dr. Y. Hirabayashi (Institute of Chemical and Physical Research, Saitama, Japan). A PCR fragment corresponding to the coding sequence (CDS) of this clone was cloned into expression vectors forS. cerevisiae and P. pastoris (pVT-U → pVHs, pGAPZ → pGHs, pPIC3.5 → pPHs). PCR with a λ-ZAP cDNA library of C. elegans (a gift from R.D. Walter, Bernhard-Nocht-Institut für Tropenmedizin, Hamburg) was performed with degenerate primers: 5′-GAYCCNAAYYTNMWNMAYAAYYTNGARACNTTYTT-3′ and 5′-TANCCNGGVWKNADRTTRTTDATYTTNGGRTT-3′. The resulting PCR fragment was used to synthesize a digoxigenin-labeled probe for the screening of the same cDNA library. In vivo excision of a positive clone resulted in the plasmid pBCe2 that had an insert of 1746 bp. It contains a CDS of 1332 bp encoding a polypeptide of 443 amino acids with a calculated molecular mass of 50.1 kDa. These sequence data have been deposited in the GenBankTM data base under GenBankTM accession number AF364401. The cloned cDNA corresponded to parts of a genomic sequence with the GenBankTM accession number U58735. However, it should be pointed out that the polypeptide deduced from the CDS of the isolated cDNA clone is not identical to the hypothetical protein AAC48147 (GenPep; synonymous Q19624, TrEMBL), which was deduced from the genomic fragment. A PCR fragment corresponding to the CDS was cloned into expression vectors for S. cerevisiae and P. pastoris (pYES2 → pYCe2, pGAPZ → pGCe2). Sequence data for C. albicans were obtained from the Stanford DNA Sequencing and Technology Center Web site (www-sequence.stanford.edu/group/candida). Sequencing of C. albicans was accomplished with the support of the NIDR and the Burroughs Wellcome Fund. The CDS HSX11, on contig 6–2503, of 1635 bp encodes a polypeptide of 544 amino acids with a calculated molecular mass of 62.5 kDa. A PCR fragment corresponding to the CDSHSX11 was generated with genomic DNA of C. albicans as template. This PCR fragment was cloned into expression vectors for S. cerevisiae, P. pastoris, andC. albicans (pYES2 → pYCa, pVT-U → pVCa, pPIC3.5 → pPCa, pGAPZ → pGCa, pBI-1 → pBICa). We cloned a GCS from P. pastoris by a PCR-based strategy with degenerate oligonucleotide primers: 5′-GGIGSIRRRYTNGANGARATGTT-3′ and 5′-YTTICKIACNCKNARCCA-3′. The resulting PCR fragment was used to synthesize a digoxigenin-labeled probe for the screening of a genomic DNA library of P. pastoris strain GS115 (28Cregg J.M. Barringer K.L. Hessler A.Y. Madden K.R. Mol. Cell. Biol. 1985; 5: 3376-3385Crossref PubMed Scopus (457) Google Scholar). The insert of a positive clone was sequenced, and these data have been deposited in the GenBankTM data base under GenBankTM accession number AF364403. It contains a CDS of 1530 bp encoding a polypeptide of 509 amino acids with a calculated molecular mass of 57.2 kDa. A PCR fragment corresponding to the CDS was cloned into expression vectors for S. cerevisiae and P. pastoris (pYES2 → pYPp, pVT-U → pVPp, pPIC3.5 → pPPp). We obtained a cDNA clone, mgae5aF11f, corresponding to the expressed sequence tag with the GenBankTM accession number AI069020, from the Clemson University Genomics Institute. Both strands of the cDNA were sequenced, and these data have been deposited in the GenBankTM data base under GenBankTM accessionAF364402. It contains a CDS of 1485 bp encoding a polypeptide of 494 amino acids with a calculated molecular mass of 55.1 kDa. A PCR fragment corresponding to the CDS was cloned into expression vectors for S. cerevisiae and P. pastoris (pVT-U → pVMg, pGAPZ → pGMg, pPIC3.5 → pPMg). We obtained a cDNA clone GA_Ea25B13 from cotton, corresponding to the expressed sequence tag with the GenBankTM accession number AW729468, from the Clemson University Genomics Institute. Both strands of the cDNA were sequenced, and these data have been deposited in the GenBankTM data base under GenBankTM accession number AF367245. It contains a CDS of 1563 bp encoding a polypeptide of 520 amino acids with a calculated molecular mass of 58.6 kDa. A PCR fragment corresponding to the CDS was cloned into expression vectors for S. cerevisiae and P. pastoris (pYES2 → pYGa, pGAPZ → pGGa, pPIC3.5 → pPGa). A PCR fragment corresponding to the CDS BAA18121 (1170 bp, 389 amino acids, 43.6 kDa, equivalent to slr0813, GenBankTM accession numberD90911) was generated with genomic DNA as template. This PCR fragment was cloned into expression vectors for S. cerevisiae andP. pastoris (pYES2 → pYS, pGAPZ → pGS). To generate a ugt51b1 null mutant, we made use of a double selection strategy. The use of a HIS4 cassette alone for transformation caused high background. Similarly, plating of more than 50,000 cells on Geneticin plates after transformation with aKANR cassette resulted in high background as well. Therefore, the entire CDS of the sterol glucosyltransferase geneUGT51B1 (25Warnecke D. Erdmann R. Fahl A. Hube B. Müller F. Zank T. Zähringer U. Heinz E. J. Biol. Chem. 1999; 274: 13048-13059Abstract Full Text Full Text PDF PubMed Scopus (102) Google Scholar) was replaced by the HIS4 gene fromP. pastoris and the KANR gene. Transformants were subsequently screened for histidine prototrophy and Geneticin (G418) resistance. The construct for disruption comprised theHIS4 gene and the KANR gene flanked by 0.5-kb homology regions to the UGT51B1 3′- and 5′-noncoding regions. These regions were amplified by PCR with a genomic fragment of P. pastoris DNA (AF091397). The primers were as follows: FM16 (5′-AAA GCG GCC GCC CGG GGT CCC CAT CAC AAG CAA-3′) and FM17 (5′-AAA GCG GCC GCC CTT CAG AAC CCC CCT TAG AG-3′) for the 5′-region (PCR1), and FM18 (5′-AAA GAA TTC ATT TTG TGT AGC TTT TCT TTT TTT TTT TTC-3′) and FM19 (5′-AAA GAA TTC CCG GGT TGA ACT TGG AGT AAG TGG AG-3′) for the 3′-region (PCR2). The two PCR fragments were cloned into the vector pGEM-T (Promega), resulting in pPCR1Pp and pPCR2Pp. The HIS4 gene was excised byNotI/EcoRI from the plasmid pHIL-D2 (Invitrogen) and cloned into the vector pBluescriptNotI/EcoRI, resulting in pMF1. The 3′-region ofUGT51B1 was excised by EcoRI from pPCR2Pp and cloned into pMF1 EcoRI resulting in pMF2. The 5′-region ofUGT51B1 was excised by NotI from pPCR1Pp and cloned into pMF2 NotI resulting in pMF3. The kanamycin cassette was excised from the plasmid pKRP11 (29Reece K.S. Phillips G.J. Gene (Amst.). 1995; 165: 141-142Crossref PubMed Scopus (135) Google Scholar) by SphI and cloned into pMF3 SphI, resulting in pMF3KR. This plasmid was digested by SmaI and introduced into P. pastoris GS115. Transformants were screened for histidine prototrophy on minimal media lacking histidine. Colonies were rinsed off with water, and 50,000 cells were plated on YPD plates (13.3 cm2) containing Geneticin (250 μg/ml). The replacement of UGT51B1 was confirmed by PCR (data not shown). Gene disruption in the diploid yeast C. albicans requires the successive elimination of both alleles. Disruption of the GCS gene (HSX11) in the strain CAI4 was performed using the ura-blaster protocol (27Fonzi W.A. Irwin M.Y. Genetics. 1993; 134: 717-728Crossref PubMed Google Scholar, 30Gow N.A. Robbins P.W. Lester J.W. Brown A.J. Fonzi W.A. Chapman T. Kinsman O.S. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 6216-6220Crossref PubMed Scopus (125) Google Scholar). The plasmid generated for gene disruption comprised a hisG::URA3::hisG cassette flanked by 0.6-kb homology regions to the GCS 3′- and 5′-non-coding regions. These regions were amplified by PCR with genomic DNA of C. albicans as template. The primers were as follows: 5′-GAGCTCATGGTTCAAGAAGAATTA-3′ and 5′-AGATCTTTGTCCCATTTTTCTCGA-3′ for the 5′-region (PCR1) and 5′-CTGCAGAAGACCCTAAAGTGAAA-3′ and 5′-AAGCTTTCACATTTCTTCAGCAGT-3′ for the 3′-region (PCR2). The two PCR fragments were cloned into the vector pBluescript EcoRV, resulting in pPCR1Ca and pPCR2Ca. The 3′ region of HSX11 was excised from pPCR2Ca by SacI/SalI and cloned into the vector pUC18 SacI/SalI, resulting in pUPCR2Ca. The hisG::URA3::hisG cassette was excised byBglII/PstI from the vector pMB-7 (27Fonzi W.A. Irwin M.Y. Genetics. 1993; 134: 717-728Crossref PubMed Google Scholar) and cloned into the vector pUPCR2Ca BglII/PstI, resulting in pML1. The 5′-region of HSX11 was excised byPstI/HindIII from pPCR1Ca and cloned into pML1PstI/HindIII, resulting in pML2. This gene disruption construct was linearized with PvuII and used for transformation of C. albicans. Successive disruptions of the wild-type HSX11 alleles resulted finally in the strain ML4 (Δura3::imm434/Δura3::imm434/Δhsx11::hisG/Δhsx11::hisG). Gene disruptions were monitored by Southern analysis after digestion of genomic DNA with XbaI and HindIII (data not shown). To construct a GCS null mutant, a fragment of the CDS was replaced by the Sh ble cassette, which confers resistance to ZeocinTM (Invitrogen). The construct for disruption comprised the Sh ble gene flanked by 0.7-kb homology regions to the CGS 3′- and 5′-noncoding regions. These regions were amplified by PCR with a genomic fragment of P. pastoris DNA. The primers were as follows: 5′-TCTAGAATGATAATGCAGCTTGGA-3′ and 5′-GAATTCAGGTACATGATGTACAAG-3′ for the 5′-region and 5′-CTCGAGACTACAGAGTGCCTGTTA-3′ and 5′-GGTACCTACAGCTTCTCAGTCTCC-3′ for the 3′-region. The two PCR fragments were cloned into the vector pBluescript/EcoRV resulting in pPCR1Pp and pPCR2Pp. The Sh ble cassette was amplified by PCR with the vector pGAPZC as a template and the primers 5′-GAATTCGATCCCCCACACACCATA-3′ and 5′-CTCGAGAACGCCAGCAACGCGGCC-5′. The PCR fragment was cloned into the vector pBluescript/EcoRV, resulting in pBShble. The 3′ region of the GCS was excised byXbaI/EcoRI from pPCR2Pp and was cloned into pBluescript XbaI/EcoRI, resulting in pBPCR2Pp. The Sh ble cassette was excised byEcoRI/XhoI from pBShble and cloned into the vector pBPCR2Pp EcoRI/XhoI, resulting in pML3. The 5′ region of the GCS was excised by XhoI/KpnI from pPCR1Pp and cloned into pML3 XhoI/KpnI, resulting in pML4. A linear fragment containing the gene disruption construct was obtained by digestion of pML4 byXbaI/KpnI. Transformations of P. pastoris were performed by electroporation. ZeocinTM-resistant transformants were selected by growth on YPD plates containing 100 mg/liter ZeocinTM. Replacement of the wild-type GCS gene was monitored by PCR and Southern analysis (data not shown). Cells were harvested by centrifugation, the sedimented cells were boiled for 10 min in water, and lipid extraction was performed as described by Warnecke et al. (25Warnecke D. Erdmann R. Fahl A. Hube B. Müller F. Zank T. Zähringer U. Heinz E. J. Biol. Chem. 1999; 274: 13048-13059Abstract Full Text Full Text PDF PubMed Scopus (102) Google Scholar). Straight phase high performance liquid chromatography (HPLC) of GlcCer species was performed on a Lichrosorb 60 Si 7 column (125 × 3 mm) with a linear gradient from solvent A to 25% solvent B in A (where A is chloroform and B is methanol/water (95:5, v/v)) in 15 min at 40 °C. Molecular species of GlcCer were separated by isocratic reversed phase HPLC on a Multospher 100 RP18–5 column (250 × 4.6 mm) in chloroform/methanol (60:40, v/v) at 24 °C. Elution was monitored by a Sedex light scattering detector operated at 50 °C. The glycolipids were acetylated for NMR spectroscopy and mass spectrometry. Fatty acid and sugar analysis by gas-liquid chromatography/mass spectrometry was performed as described previously (31Jorasch P. Warnecke D.C. Lindner B. Zähringer U. Heinz E. Eur. J. Biochem. 2000; 267: 3770-3783Crossref PubMed Scopus (66) Google Scholar). Peracetylated derivatives of GlcCer and glycosyldiacylglycerols were analyzed by mass spectrometry on an HP 5989A instrument (Hewlett-Packard) using the direct insertion probe mode and heating the sample by a temperature gradient starting from 80 °C (3 min) → 325 °C at 30°/min. Electron impact mass spectra were measured at 70 eV, and chemical ionization mass spectra were recorded with ammonia as reactant gas (0.1 megapascals). 1H NMR spectra were recorded on a 600-MHz spectrometer (Avance DRX 600, Bruker, Rheinstetten, Germany) equipped with an inverse probe head using capillary microtubes with 3-mm outer diameter (Kontes Glass Company). The peracetylated and purified monoglycosyl diacylglycerol (∼100 μg) was dissolved in 200 μl of CDCl3 (99.96%; Cambridge Isotope Laboratories, Andover, MA), and spectra were recorded at 300 K with reference to internal tetramethylsilane (δH = 0.0 ppm) using standard Bruker software (XWINNMR, version 2.6). In order to discover putative GCS genes or cDNAs from animals, fungi, plants, and bacteria, sequence candidates were identified using a sequence alignment approach based on the amino acid sequence of the GCS fromHomo sapiens (11Ichikawa S. Sakiyama H. Suzuki G. Hidari K.I. Hirabayashi Y. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 4638-4643Crossref PubMed Scopus (221) Google Scholar). A BLAST data base search (32Altschul S.F. Gish W. Miller W. Myers E.W. Lipman D. J. Mol. Biol. 1990; 215: 403-410Crossref PubMed Scopus (71457) Google Scholar) revealed a number of sequence similarities to deduced amino acid sequences of cDNAs, genomic DNA fragments, and expressed sequence tags from animals, plants, fungi, and bacteria (TableI). We obtained cDNA clones corresponding to two of the expressed sequence tags from G. arboreum and M. grisea from Clemson University Genomics Institute. These clones were sequenced, and their deduced amino acid sequences showed similarities to the human GCS.Table IMembers of a family of glucosylceramide synthases and sequences of unknown functionSourceGenBank™/GenPeptSwissProt/TrEMBLGCSType1-aG, genomic DNA; EST, expressed sequence tag.ReferenceAnimals H. sapiensD50840/BAA09451.1Q16739YescDNA11 Mus musculusD89866/BAA28782.1O88693YescDNA37 Rattus norvegicusAF047707/CAA11853.1O55149YescDNA38 Cricetulus griseus (Chinese hamster)AF351131?cDNA D. melanogaster (fruit fly)AE003456/AAF46795.1?G C. elegans(C.e.1)U53332/AAK31533/AAC71156Q21054YesG37 C. elegans(C.e.2)U58735/AF364401/AAC48147Q19624YesG + cDNAThis study C. elegans(C.e.3)Z81116/CAB03296O18037?G Xenopus laevis (African clawed frog)BE190123?EST Danio rerio (zebrafish)AI545863?EST Bombyx mori (silkworm)AV403938?ESTYeasts and fungi C. albicans—1-b—, not available from GenBank™, gene HSX11, on contig 6–2503; information available on the World Wide Web at www-sequence.stanford.edu/group/candida.YesGThis study P. pastorisAF364403YesGThis study M. griseaAF364402YescDNAThis study Neurospora crassaAL353820/CAB88601?G Pneumocystis cariniiAP338415?cDNA Pichia angustaAL435585?GPlants Ceratopteris richardii(fern)BE643203?EST Arabidopsis thaliana (thale cress)AC005169/AAC62128.1?G G. arboreum(cotton)AF367245YescDNAThis studyBacteria Synechocystis sp. (blue-green alga)slr0813,D90911/BAA18121.1P74046?GThis study Zymomonas mobilisAF193753?G Thermoplasma acidophilumAL445064/CAC11705?G Mesorhizobium lotiNC002678/NP10627?G1-a G, genomic DNA; EST, expressed sequence tag.1-b —, not available from GenBank™, gene HSX11, on contig 6–2503; information available on the World Wide Web at www-sequence.stanford.edu/group/candida. Open table in a new tab In a second approach, degenerate primers deduced from conserved regions of GCS were used to clone novel putative GCS sequences by a PCR-based strategy with subsequent screening of cDNA or genomic libraries with specific probes. Thereby, we cloned another two putative GCS sequences from C. elegans and P. pastoris. The cloning procedures, the features of the DNA fragments, and the availability of these data in public data bases are described under “Experimental Procedures” and in Table I. When deduced protein sequences were aligned, all sequences listed in Table I fell into a previously identified glycosyltransferase family: family 21 of NDP-sugar hexosyltransferases (33Campbell J.A. Davies G.J. Bulone V. Henrissat B. Biochem. J. 1997; 326: 929-939Crossref PubMed Scopus (630) Google Scholar, 34Coutinho P.M. Henrissat B. Gilbert H.J. Davies G. Henrissat B. Svensson B. Recent Advances in Carbohydrate Bioengineering. The Royal Society of Chemistry, Cambridge1999: 3-12Google Scholar) (Carbohydrate-Active Enzymes Server, available on the World Wide Web at afmb.cnrs-mrs.fr/∼pedro/CAZY/db.html). The same family has also been described as group 9 of the NRD2 glycosyltransferases, which were grouped with respect to different types of “nucleotide recognition domains” (NRDs) (35Kapitonov D. Yu R.K. Glycobiology. 1999; 9: 961-978Crossref PubMed Scopus (139) Google Scholar). This family shares few but significant similarities to the glycosyltransferase family 2. These similarities have been described as the D1,D2,D3,Q/RXXRW motif (36Marks D.L. Dominguez M. Wu K. Pagano R.E. J. Biol. Chem. 2001; 276: 26492-26498Abstract Full Text Full Text PDF PubMed Scopus (44) Google Scholar). Within the family 21, three mammalian sequences from humans (11Ichikawa S. Sakiyama H. Suzuki G. Hidari K.I. Hirabayashi Y. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 4638-4643Crossref PubMed Scopus (221) Google Scholar), mice (37Ichikawa S. Hirabayashi Y. Trends Cell Biol. 1998; 8: 198-202Abstract Full Text Full Text PDF PubMed Scopus (183) Google Scholar), and rats (38Wu K. Marks D.L. Watanabe R. Paul P. Rajan N. Pagano R.E. Biochem. J. 1999; 341: 395-400Crossref PubMed Google Scholar) and one from C. elegans (C. elegans 1; Ref. 37Ichikawa S. Hirabayashi Y. Trends Cell Biol. 1998; 8: 198-202Abstract Full Text Full Text PDF PubMed Scopus (183) Google Scholar) have previously been identified experimentally as GCS. However, based on their similarity to the mammalian enzymes, some of the other sequences have been automatically annotated as GCS in the data bases without supporting experimental evidence. Fig. 1 shows a sequence alignment of selected members of this enzyme family. Remarkably, there are only a few conserved amino acids, and the overall similarity between the enzymes from species with remote evolutionary relationship is rather low. The identities to the human GCS are as follows: one protein fromDrosophila melanogaster had 46% identity, three different proteins from C. elegans had 30% identities, three proteins from fungi (P. pastoris, C. albicans, andM. grisea) had 16–21% id