HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 276, Issue 36, 33621-33629, September 7, 2001

This Article Abstract Full Text (PDF) All Versions of this Article: 276/36/33621    most recent M104952200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Leipelt, M. Articles by Heinz, E. Search for Related Content PubMed PubMed Citation Articles by Leipelt, M. Articles by Heinz, E. Social Bookmarking                      What's this?

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

Martina Leipelt, Dirk Warnecke^§, Ulrich Zähringer^¶, Claudia Ott, Frank Müller, Bernhard Hube^**, and Ernst Heinz

From the  Institut für Allgemeine Botanik, University of Hamburg, Ohnhorststr. 18, 22609 Hamburg, Germany, ^¶ Zentrum für Medizin und Biowissenschaften, Parkallee 22, 23845 Borstel, Germany, and ^** Robert Koch-Institut, Nordufer 20, 13353 Berlin, Germany

Received for publication, May 30, 2001, and in revised form, July 6, 2001

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

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 pastoris resulted 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 plant Gossypium arboreum (cotton), the nematode Caenorhabditis 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.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

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 (1), plants (2-4), fungi (5-8), and bacteria (9) 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 (10). Recently, the first cDNA coding for a human GCS has been cloned (11). This success has opened new possibilities for analyzing GlcCer functions (12, 13), particularly by studying the phenotypes resulting from gene deletions in knock-out mice (14). 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 (15-17). In addition, studies on the intracellular location and transport of GlcCer contribute to the understanding of the apical/basolateral asymmetry of epithelial cells (18).

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 (19)).

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 (6-8, 20). 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 (21, 22). 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.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Bacterial and Yeast Strains and Plasmids

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 in Saccharomyces cerevisiae, P. pastoris, and C. albicans was performed with pVT-U (23), pYES2, pGAPZ, pPIC3.5 (Invitrogen), pBI-1 (24). Yeast strains used in this study were S. cerevisiae UTL-7A (MATa, ura3-52, trp1, leu2-3, 112), UTL-7Augt51 (MATa, ura3-52, trp1, leu2-3, 112, ugt51::kanMX4; Ref. 25), and Sc334 (26); P. pastoris GS115 (Invitrogen); and C. albicans SC5314 CAI4 ura3::imm434/ura3::imm434, congenic to SC5314 (27).

Expression Vectors for the Human GCS

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 for S. cerevisiae and P. pastoris (pVT-U  pVHs, pGAPZ  pGHs, pPIC3.5  pPHs).

Cloning Novel GCS

C. elegans-- PCR with a -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 GenBank^TM data base under GenBank^TM accession number AF364401. The cloned cDNA corresponded to parts of a genomic sequence with the GenBank^TM 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).

C. albicans-- 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 CDS HSX11 was generated with genomic DNA of C. albicans as template. This PCR fragment was cloned into expression vectors for S. cerevisiae, P. pastoris, and C. albicans (pYES2  pYCa, pVT-U  pVCa, pPIC3.5  pPCa, pGAPZ  pGCa, pBI-1  pBICa).

P. pastoris-- 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 (28). The insert of a positive clone was sequenced, and these data have been deposited in the GenBank^TM data base under GenBank^TM 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).

M. grisea-- We obtained a cDNA clone, mgae5aF11f, corresponding to the expressed sequence tag with the GenBank^TM accession number AI069020, from the Clemson University Genomics Institute. Both strands of the cDNA were sequenced, and these data have been deposited in the GenBank^TM data base under GenBank^TM accession AF364402. 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).

Gossypium arboreum-- We obtained a cDNA clone GA_Ea25B13 from cotton, corresponding to the expressed sequence tag with the GenBank^TM accession number AW729468, from the Clemson University Genomics Institute. Both strands of the cDNA were sequenced, and these data have been deposited in the GenBank^TM data base under GenBank^TM 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).

Synechocystis sp. (strain PCC 6803)-- A PCR fragment corresponding to the CDS BAA18121 (1170 bp, 389 amino acids, 43.6 kDa, equivalent to slr0813, GenBank^TM accession number D90911) was generated with genomic DNA as template. This PCR fragment was cloned into expression vectors for S. cerevisiae and P. pastoris (pYES2  pYS, pGAPZ  pGS).

Deletion of the P. pastoris Sterol Glucosyltransferase Gene (UGT51B1)

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 a KAN^R cassette resulted in high background as well. Therefore, the entire CDS of the sterol glucosyltransferase gene UGT51B1 (25) was replaced by the HIS4 gene from P. pastoris and the KAN^R gene. Transformants were subsequently screened for histidine prototrophy and Geneticin (G418) resistance. The construct for disruption comprised the HIS4 gene and the KAN^R 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 by NotI/EcoRI from the plasmid pHIL-D2 (Invitrogen) and cloned into the vector pBluescript NotI/EcoRI, resulting in pMF1. The 3'-region of UGT51B1 was excised by EcoRI from pPCR2Pp and cloned into pMF1 EcoRI resulting in pMF2. The 5'-region of UGT51B1 was excised by NotI from pPCR1Pp and cloned into pMF2 NotI resulting in pMF3. The kanamycin cassette was excised from the plasmid pKRP11 (29) 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).

Disruption of the C. albicans and P. pastoris GCS Genes

C. albicans-- 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 (27, 30). 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 by BglII/PstI from the vector pMB-7 (27) and cloned into the vector pUPCR2Ca BglII/PstI, resulting in pML1. The 5'-region of HSX11 was excised by PstI/HindIII from pPCR1Ca and cloned into pML1 PstI/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 (ura3::imm434/ura3::imm434/hsx11::hisG/hsx11::hisG). Gene disruptions were monitored by Southern analysis after digestion of genomic DNA with XbaI and HindIII (data not shown).

P. pastoris-- To construct a GCS null mutant, a fragment of the CDS was replaced by the Sh ble cassette, which confers resistance to Zeocin^TM (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 by XbaI/EcoRI from pPCR2Pp and was cloned into pBluescript XbaI/EcoRI, resulting in pBPCR2Pp. The Sh ble cassette was excised by EcoRI/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 by XbaI/KpnI. Transformations of P. pastoris were performed by electroporation. Zeocin^TM -resistant transformants were selected by growth on YPD plates containing 100 mg/liter Zeocin^TM. Replacement of the wild-type GCS gene was monitored by PCR and Southern analysis (data not shown).

Lipid Extraction and Analysis

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. (25). 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.

Combined Gas-Liquid Chromatography/Mass Spectrometry Analysis

Fatty acid and sugar analysis by gas-liquid chromatography/mass spectrometry was performed as described previously (31). 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).

Proton (¹H) NMR Spectroscopy

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 (_H = 0.0 ppm) using standard Bruker software (XWINNMR, version 2.6).

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Cloning of Putative Glucosylceramide Synthases-- 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 from Homo sapiens (11). A BLAST data base search (32) 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 (Table I). 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.

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 (33, 34) (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) (35). 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 (36). Within the family 21, three mammalian sequences from humans (11), mice (37), and rats (38) and one from C. elegans (C. elegans 1; Ref. 37) 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 from Drosophila melanogaster had 46% identity, three different proteins from C. elegans had 30% identities, three proteins from fungi (P. pastoris, C. albicans, and M. grisea) had 16-21% identities, three proteins from bacteria had 18-20% identities (not shown), and one protein from a plant (G. arboreum) had only 9% identity to the human counterpart (Fig. 1). Most sequences contain a single putative transmembrane domain at the N terminus and a segment of mainly hydrophobic amino acids at the C terminus, which may interact with the membrane. These features are similar to the mammalian GCS, which are integral membrane proteins with their active site and the C terminus on the cytosolic face of the Golgi membrane (39, 40). Thus, the orientation of these putative membrane proteins seems to be similar, but the enzymatic function of most members of this protein family is unknown.

View larger version (111K): [in this window] [in a new window]   Fig. 1.   ClustalX alignment of glucosylceramide synthases, all of which are members of the glycosyltransferase family 21 (33). The amino acid sequences are from G. arboreum (G.a.), H. sapiens (H.s.), D. melanogaster (D.m.), C. elegans (C.e.1, C.e.2, and C.e.3), P. pastoris (P.p.), C. albicans (C.a.), and M. grisea (M.g.). Dark gray regions indicate identical amino acids in all nine sequences. Gray regions indicate identical amino acids in 5-8 sequences. All sequences have been identified by functional expression as glucosylceramide synthases (this study and Ref. 11), except for D. melanogaster and C. elegans 3. Despite functional identity, the sequence identities are rather low and vary between 9 and 46%. The putative N-terminal transmembrane domains are boxed. Hypothetical "nucleotide recognition domains" (NRD2L and NRD2S) described by Kapitonov and Yu (35) are underlined. An arrowhead indicates histidine 193 of the mammalian GCS, which is conserved within glycosyltransferase family 21 (38). The other arrowheads indicate amino acids that are conserved between glycosyltransferase family 2 and 21 (36). Several of these amino acids, including those of the indicated D1,D2,D3,Q/RXXRW motif, are essential for enzyme activity as demonstrated by site-directed mutagenesis of mammalian GCS (36, 38). The C-terminal ends of the sequences were omitted to save space.

In order to examine whether the identified nucleotide sequences encode GCS, we selected at least one candidate from each kingdom for a systematic functional analysis of this glycosyltransferase family. CDS from putative GCS genes of G. arboreum, C. elegans, M. grisea, C. albicans, P. pastoris, and Synechocystis were amplified from either a cDNA clone or from genomic DNA by PCR and cloned into expression vectors for S. cerevisiae, P. pastoris, and C. albicans, respectively. The GlcCer accumulated in these organisms as a result of the expression of the different sequences were analyzed and compared with those resulting from the expression of the GCS from H. sapiens as a control. In addition, we generated null mutants of P. pastoris and C. albicans.

Only Glucosylceramide Synthases from Humans, C. elegans, and C. albicans Can Be Functionally Expressed in S. cerevisiae-- When S. cerevisiae was transformed with expression vectors containing the putative glycosyltransferases from G. arboreum, C. elegans, P. pastoris, C. albicans, M. grisea, and Synechocystis and the GCS from H. sapiens, only expression of the sequences from H. sapiens, C. elegans, and C. albicans resulted in de novo biosynthesis of GlcCer (which is absent in the parental S. cerevisiae strain). Only one novel type of GlcCer was accumulated after the expression of the gene from C. albicans and C. elegans, whereas two different GlcCer were detected after expression of the GCS from H. sapiens (Fig. 2). These glycolipids were purified and subjected to structural analysis as discussed below. From these expression studies, we conclude that the sequences from C. elegans and C. albicans encode GCS. Since the expression of putative GCS genes from P. pastoris, M. grisea, G. arboreum, and Synechocystis did not result in the biosynthesis of detectable quantities of GlcCer in S. cerevisiae (data not shown), we chose P. pastoris and C. albicans as alternative systems that may be more suitable for expressing GCS. In contrast to S. cerevisiae, these yeasts contain significant amounts of GlcCer (6, 41) and should therefore be able to provide suitable sugar acceptors for heterologously expressed GCS. However, for unambiguous results, the elimination of endogenous GCS activity by gene deletion was a prerequisite before expressing the heterologous genes.

View larger version (65K): [in this window] [in a new window]   Fig. 2.   Expression of putative glucosylceramide synthases from different organisms in S. cerevisiae resulted in glucosylceramide biosynthesis. Lipid extracts were separated by thin layer chromatography. Lane 1, the host strain was devoid of GlcCer; lane 2, a strain expressing the human GCS as a control produced two novel GlcCer (GlcCer 1' and GlcCer 3'); lanes 3 and 4, strains expressing putative GCS from C. elegans and C. albicans produced one single GlcCer. Lane 4 also contains sterol glucoside (SG), which is usually absent in S. cerevisiae.

Deletion of Glucosylceramide Synthase Genes in P. pastoris and C. albicans Resulted in the Loss of Glucosylceramide Synthesis but Had No Significant Effect on Vegetative Growth-- We generated GCS null mutants of P. pastoris and C. albicans. The gene disruptions resulted in the complete loss of GlcCer in the mutant cells (Fig. 3). Therefore, these results confirm that the gene of C. albicans indeed encodes a GCS and that the homologous gene of P. pastoris has the same function, although it was not active in S. cerevisiae. This conclusion is further supported by plasmid-borne homologous expression of these genes in the respective null mutants, which restored GlcCer biosynthesis (Figs. 3 and 4). Both null mutants were still viable and grew like the parental strains on complex and minimal medium. The C. albicans null mutant was able to grow in both yeast and filamentous forms. These results show that GlcCer do not play essential roles during growth of P. pastoris and C. albicans.

View larger version (43K): [in this window] [in a new window]   Fig. 3.   Generation of P. pastoris and C. albicans mutants deficient in glycolipid synthesis. Lane 1, the parental P. pastoris strain GS115 contained sterol glucoside (SG) and glucosylceramide; lane 2, deletion of the sterol glucosyltransferase gene resulted in the complete loss of sterol glucoside; lane 3, deletion of the glucosylceramide synthase gene caused the loss of GlcCer (the faint spot with the same Rf value as GlcCer, which is still visible, is not a glycolipid); lane 4, the C. albicans parental strain CAI4 contains GlcCer but no sterol glucoside; lane 5, deletion of the GCS gene resulted in the complete loss of GlcCer and the compensatory biosynthesis of a small quantity of sterol glucoside; lane 6, transformation of the null mutant for the expression of the homologous GCS restored GlcCer synthesis in C. albicans.

View larger version (71K): [in this window] [in a new window]   Fig. 4.   Expression of glucosylceramide synthases from different organisms in a GCS null mutant of P. pastoris resulted in glucosylceramide biosynthesis. Lane 1, the GCS null mutant contained sterol glucosides (SG) but no GlcCer; lanes 2-5, expression of GCS from P. pastoris, C. albicans, M. grisea, and G. arboreum resulted in the biosynthesis of two main GlcCer (GlcCer 1 and GlcCer 2); lane 6, expression of the human GCS resulted in the synthesis of five GlcCer and of glucosyldiacylglycerol (MGlcD). The two most apolar GlcCer overlap with sterol glucosides (see Fig. 5).

Besides the GCS null mutants, we generated a sterol glucosyltransferase null mutant of P. pastoris. This strain was devoid of sterol glucoside (Fig. 3), a glycolipid that is difficult to separate by silica gel chromatography from some molecular species of GlcCer. Thus, use of this sterol glucosyltransferase null mutant strain facilitated the purification and identification of the overlapping GlcCer molecular species resulting from the heterologous expression of putative GCS genes (see below).

Expression of Glucosylceramide Synthases in P. pastoris Resulted in the Biosynthesis of a Series of Different GlcCer-- We expressed glycosyltransferases from H. sapiens, G. arboreum, P. pastoris, C. albicans, M. grisea, and Synechocystis in the GlcCer-deficient P. pastoris GCS null mutant strain. All expression experiments except for the putative GCS sequence of Synechocystis led to the biosynthesis of GlcCer (Fig. 4). Thus, we identified another two novel GCS from G. arboreum and M. grisea. We do not know why expressions of GCS from G. arboreum, P. pastoris, and M. grisea in P. pastoris were successful but failed in S. cerevisiae. On the other hand, these results demonstrated the usefulness of different expression hosts. Interestingly, the expression of the heterologous genes resulted in all cases in the appearance of at least two GlcCer bands. These GlcCer differ in their R_f values depending on the origin of the expressed gene. Expression of the human GCS resulted in the synthesis of five different GlcCer (Fig. 4, 1-5). The GlcCer molecular species 4 and 5 showed R_f values similar to that of sterol glucoside, which hampered their purification for structural analysis. Therefore, the human GCS was also expressed in the sterol glucoside-deficient P. pastoris strain (Fig. 5). As expected, this expression resulted in the appearance of several GlcCer bands, which gave insight into the substrate specificity of the enzyme (see below).

View larger version (31K): [in this window] [in a new window]   Fig. 5.   Expression of the human glucosylceramide synthase in a sterol glucoside-deficient P. pastoris strain facilitated the purification of different glucosylceramides. A, lane 0, total lipid extract of the transgenic P. pastoris cells; lanes 1-5, the different GlcCer were purified for subsequent structural analysis. Straight phase (B) and reversed phase (C) high performance liquid chromatography of the individual GlcCer show that the lipids 1, 3, and 5 are nearly homogenous, whereas 2 and 4 consist of five and two different molecular species, respectively. See Table II for identification of individual species.

Fig. 4 shows that the human GCS expressed in P. pastoris synthesized an additional glycolipid, which was identified as 1,2-diacyl-3-[O--D-glucopyranosyl]-sn-glycerol, also known from our previous study on diacylglycerol glucosyltransferases (31). The formation of this new glycolipid as a consequence of GCS expression may be ascribed to the structural similarity of 1,2-diacyl-sn-glycerol and ceramide as discussed before (31). It results in the use of both acceptors by diacylglycerol and ceramide glycosyltransferases, which normally prefer only one of the two substrates. This conclusion is supported by the simultaneous loss of galactosylceramide and galactosyldiacylglycerol in knock-out mice lacking the ceramide galactosyltransferase gene (42).

Novel GlcCer from Transgenic Yeasts Contained Very Long Chain Fatty Acids Characteristic for Inositol-containing Sphingolipids-- We separated the different GlcCer resulting from the heterologous expression of the GCS by TLC (Figs. 4 and 5). The purified glycolipids were acetylated and subjected to structural analysis by mass spectrometry and NMR spectroscopy. In transgenic S. cerevisiae expressing the GCS from C. albicans or H. sapiens, we found two molecular species of GlcCer with 26:0(2-OH)-t18:0 and 26:0-t18:0 ceramide backbones (1' and 3' in Table II; in the text, we use the abbreviated names of ceramide backbones, whereas systematic names are listed in Table II). In contrast, the expression of the human GCS in P. pastoris resulted in the biosynthesis of five different GlcCer (spots 1-5 in Figs. 4 and 5). From these glycolipids, GlcCer 2 and 4 could be resolved by HPLC into five and two different molecular species, respectively (2a, 2a', 2b, 2c, 2d; 4a, 4b; Fig. 5C; Table II). Thus, human GCS synthesized 10 different molecular species of GlcCer, which contained different sphingobases and either long chain or very long chain fatty acids (LCFA and VLCFA): 18:0-18:0, 18:0-18:1⁴, 18:0-18:2^4,8, 16:0(2-OH)-18:0, 18:0(2-OH)-18:0, 18:0(2-OH)-18:1⁴, 18:0(2-OH)-18:2^4,8, 18:0(2-OH)-18:2^4,89m, 24:0(2-OH)-t18:0, and 24:0-t18:0 (see Table II). TLC and HPLC analysis revealed that the expression of the GCS from P. pastoris, C. albicans, M. grisea, and G. arboreum in P. pastoris resulted in the biosynthesis of GlcCer containing also both LCFA or VLCFA in their ceramide backbones. Remarkably, those ceramide backbones that contained phytosphingosine and a VLCFA (1, 1', 3, 3'; Table II) usually do not occur in fungal GlcCer but are typical for inositol-containing sphingolipids of yeasts and fungi.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

A BLAST data base search with the human GCS revealed a protein family that was described as NDP-sugar hexosyltransferase family 21 (33) or NRD2 glycosyltransferase group 9 (35). This protein family consists of a few well characterized GCS of mammalian origin (11, 37) and of many putative glycosyltransferases from animals, plants, fungi, and bacteria with unknown function. For a more comprehensive characterization of this protein family, we expressed six members of unknown function from different kingdoms in the yeasts S. cerevisiae and P. pastoris to determine their activity as GCS.

The expression experiments resulted in the identification and characterization of novel GCS from plants (G. arboreum), animals (C. elegans), and fungi (M. grisea, P. pastoris, C. albicans). Therefore, the systematic functional analysis of the glycosyltransferase family 21 that was performed in this study discovered GCS from all eukaryotic kingdoms. Only the bacterial members of this protein family seem not to be GCS, since we were not able to detect GCS activity with the slr0813 gene from Synechocystis, and thus the functions of these bacterial proteins remain to be characterized.

Despite the fact that GlcCer ubiquitously occur in the plant kingdom, the GCS cDNA from G. arboreum described in the present paper is the first plant GCS nucleotide sequence to be cloned. We ascribe this to the very low amino acid sequence similarity of the plant GCS to the other GCS including the lack of some of the amino acids of the D1,D2,D3,Q/RXXRW motif. But it should be mentioned that for plants two different pathways for GlcCer synthesis have been suggested that involve the transfer of a glucosyl residue either from UDP-glucose (43) or from sterol glucoside (44) to ceramide. P. pastoris contains considerable amounts of sterol glucoside (6), which could serve as alternative glucosyl donor for a hypothetical sterol glucoside:ceramide glucosyltransferase. The generation of a P. pastoris double mutant deficient in both sterol glucosides and GlcCer and subsequent expression of the GCS from cotton will show whether it codes for a UDP-glucose-independent transglucosylase or a UDP-glucose-dependent GCS.

In addition to the sequences from C. elegans and plants, we were able to identify novel GCS from the fungi M. grisea, P. pastoris, and C. albicans. In this context, it should be pointed out that S. cerevisiae does not contain a homologue GCS of this protein family, which contrasts with the rare reports on GlcCer isolated from S. cerevisiae (45). However, we were not able to isolate GlcCer from a number of different strains of S. cerevisiae under various culture conditions. Thus, we conclude that S. cerevisiae may synthesize at most minor amounts of this glycolipid under particular growth conditions and by a mechanism or enzyme that is different from that used by other fungi.

The expression studies of GCS from animals, plants, and fungi did not only reveal the enzymatic function of the novel GCS. In addition, the structural analysis of the different GlcCer isolated from the transgenic yeasts provides new insights into the biosynthesis of sphingolipids. (i)The expression of the human GCS in P. pastoris resulted in the most complex series of GlcCer. The ceramide backbones, 18:0-18:0, 18:0(2-OH)-18:0, 18:0-18:1⁴, 18:0-18:2^4,8, 18:0(2-OH)-18:1⁴, and 18:0(2-OH)-18:2^4,8 (Table II, compounds 5, 2a, 4a, 4b, 2c, and 2d) may all be regarded as biosynthetic precursors of 18:0(2-OH)-18:2^4,89m (Table II, 2b), which is the major ceramide moiety in the GlcCer of P. pastoris and many other fungi (6). The different intermediates found suggest a sequential modification of the sphingobase starting with the introduction of the 4-double bond followed by the 8-double bond and a final methylation at C9 (boxed in Fig. 6). However, it remains unclear whether these modifications occur at the level of the free sphingobase, the ceramide, or the GlcCer. (ii) It has been stated before that some fungi synthesize both GlcCer and inositol-containing sphingolipids (5, 7, 8) including free glycosylinositol phosphorylceramides and glycosylphosphatidylinositol-anchored proteins (46). We expect that further investigations will show that this is also true for many more yeasts and fungi. Interestingly, these two classes of sphingolipids contain completely different ceramide backbones. The inositol-containing sphingolipids consist of phytosphingosine (4-hydroxysphinganine) and a VLCFA (5, 20, 46, 47), whereas the GlcCer contain the characteristic 4,8-diunsaturated, C9-methyl-branched sphingobase and a saturated or 3-trans-unsaturated C16 or C18 -hydroxy LCFA (this study and Refs. 6, 7, 8, 21, 41, 48, and 49). This situation requires the partitioning of two distinct sphingolipid pools resulting either from compartmentalization or from distinct substrate specificities of enzymes operating subsequent to ceramide assembly downstream in glycosphingolipid synthesis (Fig. 6) (5, 7, 8).

View larger version (23K): [in this window] [in a new window]   Fig. 6.   Separation of ceramide pools for glycosphingolipid biosynthesis in yeasts and fungi. Metabolites are shown in boldface type, and genes are shown in italics. Gene symbols are according to the S. cerevisiae nomenclature. The scheme of inositol phosphoryl sphingolipid biosynthesis on the left side is similar to the pathway proposed for S. cerevisiae (5). The boxed reactions on the right occur in the given sequential order, whereas some of the other steps may be carried out in reversed sequence. Ceramide C was the main ceramide backbone that we found in very long chain base glucosylceramides, but ceramides A and B may serve as substrates for GCS and inositolphosphorylceramide (IPC) synthase as well. It is not known whether SCS7 exclusively hydroxylates only VLCFA or both VLCFA and LCFA. LCB, long chain base.

Our data show that the expression of different GCS in S. cerevisiae resulted in the biosynthesis of GlcCer containing phytosphingosine and a VLCFA (Fig. 2, spots 1' and 3'; Table II, 1' and 3'). These ceramide backbones are usually restricted to inositol-containing sphingolipids. GlcCer with such VLCFA have never been isolated from yeasts or fungi before, whereas some plants contain such GlcCer (3). The expression of different GCS in the P. pastoris GCS null mutant resulted in the synthesis of the normal 18:0(2-OH)-18:2^4,89m GlcCer as well as of the new very long chain GlcCer (Fig. 4). Thus, all GCS were able to accept very long chain ceramide as substrate. This lipid is also present in C. albicans and P. pastoris and is usually used for synthesis of inositol-containing sphingolipids but not for GlcCer synthesis. Therefore, we conclude that the separation of two distinct ceramide pools in C. albicans and P. pastoris cannot be attributed to the substrate specificity of their GCS. The differences in the ceramide backbone of inositol-containing sphingolipids and GlcCer may rather originate from compartmentalization of both the ceramides and the downstream enzymes GCS and inositolphosphorylceramide synthase (Fig. 6).

However, it is a puzzling result that the endogenous GCS of P. pastoris synthesizes exclusively C18 GlcCer, whereas the strong expression of the homologous or heterologous genes resulted in the additional use of the very long chain ceramide species. A possible explanation for this phenomenon could be a disturbance of the protein targeting machinery by flooding the system with GCS molecules including the overloading of the Golgi/ER retrieval capacity. Additional impairment of a lateral restriction of the enzyme to membrane areas such as sphingolipid/sterol domains may contribute to this effect. Thus, basic mechanisms of the creation and separation of two obvious glycosphingolipid pools, their trafficking, and their role in the formation of lipid domains in fungal membranes remain to be elucidated.

Biosynthesis of Sphingolipids-- Our results suggest that the partitioning of the two ceramide pools in fungi results from compartmentalization of lipids and enzymes of sphingolipid synthesis. This hypothesis requires further experimental verification. Analysis of the intracellular location of the ceramide pools should be possible by membrane fractionation of P. pastoris cells with subsequent lipid analysis including in situ GlcCer localization by newly developed anti-GlcCer antibodies (50-52). In addition, specific antibodies against the P. pastoris GCS will be generated to study its intracellular location.

Biological Function of Glucosylceramides-- The identification and characterization of GCS from nonanimal origin enables studies of the function of GlcCer in organisms such as plants, fungi, and single cell organisms like yeasts. The generation of viable GCS null mutants shows that GlcCer, in contrast to inositol-containing very long chain sphingolipids, are not essential for normal growth of P. pastoris and C. albicans. Further studies of the yeast null mutants will show whether phenotypes will show up under different and extreme culture conditions. Discovery of phenotypes, especially those that result from ceramide overaccumulation, will enable a screening for suppressor mutants, which in turn would be powerful tools to study the function and regulation of GlcCer synthesis. These studies with P. pastoris should take S. cerevisiae as a model, since suppressor mutants have been successfully generated and used to examine the glycosylinositol phosphorylceramide synthesis (53). In addition, the generation of GlcCer-deficient mutants of phytopathogenic fungi will enable studies on the functions of GlcCer in host/pathogen interactions.

	ACKNOWLEDGEMENTS

We thank S. Ichikawa and Y. Hirabayashi for providing the cDNA clone of the human GCS, J. M. Cregg for a DNA library from P. pastoris, R. D. Walter for the cDNA library of C. elegans, J. F. Ernst for the plasmid pBI-1, W. Hellmeyer for skillful technical assistance, and the Clemson University Genomics Institute for the cDNA clones from M. grisea and G. arboreum.

	FOOTNOTES

^* This work was supported by the Deutsche Forschungsgemeinschaft, Sonderforschungsbereich 470, the BMBF grant NAPUS 2000, and Fonds der Chemischen Industrie. Sequencing of C. albicans was accomplished with the support of the National Institute of Dental and Craniofacial Research and the Burroughs Wellcome Fund.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The nucleotide sequence(s) reported in this paper has been submitted to the GenBank^TM/EMBL Data Bank with accession number(s) AF364401 (C. elegans), AF364403 (P. pastoris), AF367245 (G. arboreum), and AF364402 (M. grisea).

^§ To whom correspondence should be addressed: Institut für Allgemeine Botanik, University of Hamburg, Ohnhorststr. 18, 22609 Hamburg, Germany. Tel.: 49 40 42816 364; Fax: 49 40 42816 254; E-mail warnecke@botanik.uni-hamburg.de.

Present address: Institut für Biologie I, RWTH Aachen, Worringerweg 1, 52056 Aachen, Germany.

Published, JBC Papers in Press, July 6, 2001, DOI 10.1074/jbc.M104952200

	ABBREVIATIONS

The abbreviations used are: GlcCer, glucosylceramide(s); CDS, coding region; GCS, glucosylceramide synthase; PCR, polymerase chain reaction; bp, base pairs; HPLC, high performance liquid chromatography; NRD, nucleotide recognition domain; contig, group of overlapping clones; LCFA, long chain fatty acid(s); VLCFA, very long chain fatty acid(s).

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				R. Rhome, T. McQuiston, T. Kechichian, A. Bielawska, M. Hennig, M. Drago, G. Morace, C. Luberto, and M. Del Poeta Biosynthesis and Immunogenicity of Glucosylceramide in Cryptococcus neoformans and Other Human Pathogens Eukaryot. Cell, October 1, 2007; 6(10): 1715 - 1726. [Full Text] [PDF]

				E. Arigi, S. Singh, A. H Kahlili, H. C Winter, I. J Goldstein, and S. B Levery Characterization of neutral and acidic glycosphingolipids from the lectin-producing mushroom, Polyporus squamosus Glycobiology, July 1, 2007; 17(7): 754 - 766. [Abstract] [Full Text] [PDF]

A.-H. Lebrun, C. Wunder, J. Hildebrand, Y. Churin, U. Zahringer, B. Lindner, T. F. Meyer, E. H