Newly Discovered Neutral Glycosphingolipids in Aureobasidin A-resistant Zygomycetes IDENTIFICATION OF A NOVEL FAMILY OF GALA-SERIES GLYCOLIPIDS WITH CORE Gal (cid:1)

We found for the first time that Zygomycetes species showed resistance to Aureobasidin A, an antifungal agent. A novel family of neutral glycosphingolipids (GSLs) was found in these fungi and isolated from Mucor hiemalis , which is a typical Zygomycetes species. Their structures were completely determined by composi-tional sugar, fatty acid, and sphingoid analyses, methylation analysis, matrix-assisted laser desorption ionization time-of-flight/mass spectrometry, and 1 H NMR spectroscopy. They were as follows: Gal (cid:1) 1-6Gal (cid:1) 1-1Cer (CDS), Gal (cid:2) 1-6Gal (cid:1)

Sphingolipids are essential membrane components of both mammalian and fungal cells. The early steps in their biosynthetic pathways up to the formation of sphingosine are the same, but the subsequent pathways are very different in both cells (1). In mammalian cells, sphingosine is attached to fatty acids to yield ceramide. In fungi, phytoceramide is produced from a phytosphingosine having an additional hydroxylation on C-4 and 2-hydroxy fatty acid. The phytoceramide gives rise to inositol-containing sphingolipids such as inositol phosphorylceramide (IPC), 1 mannose-IPC (MIPC), and inositol phosphor-yl-MIPC (2,3). This step involves the transfer of the phosphoinositol group from phosphatidylinositol to the 1-hydroxy group of the phytoceramide to yield IPC, which is then mannosylated to yield MIPC (4). Because sphingolipid synthesis is essential for the growth and viability of fungi, it is likely that a blocking of the synthesis would efficiently inhibit cell growth. Therefore, the enzymes catalyzing the synthesis of inositol-containing sphingolipids that are present in fungi but absent in humans have been focused as targets for antifungal agents (5). One of these enzymes is IPC synthase, which catalyzes the transfer of the inositolphosphate from phosphatidylinositol to ceramide to give IPC (6).
Aureobasidin A is well known and widely used as an antifungal agent for Eumycetes including yeasts and fungi. It exhibits strong fungicidal activity against many pathogenic fungi, including Candida albicans, Cryptococcus neoformans, and Aspergillus fumigatus (7,8). Recent studies (6) have shown that this antifungal agent inhibits IPC synthase in fungal cells. The inhibition of this enzyme causes the depletion of essential sphingolipids in the fungal cells. Because it is recognized that all fungi have this enzyme, Aureobasidin A is potentially a broad spectrum antifungal (9).
We found that all Zygomycetes species tested were resistant to the antifungal agent Aureobasidin A, a cyclic depsipeptide produced by Aureobasidium pullulans (10). The Zygomycetes species do not have inositol-containing sphingolipids but contain novel neutral glycosphingolipids (GSLs) consisting glucose or galactose as sugar constituents. This suggests a remarkable difference in GSLs between Zygomycetes species and other fungi. We supposed that the lack of a synthetic pathway for inositol-containing sphingolipids in their cells might be the cause of the resistance of Zygomycetes species to Aureobasidin A. We also suggested a new synthetic pathway for GSLs containing galactooligosaccharides with Gal␤1-6Gal and Gal␣1-6Gal residues in Zygomycetes species such as Mucor and Rhizopus species. A novel family of neutral GSLs found from Zygomycetes species belongs to a homologous series with a phytoceramide consisting of phytosphingosine and 2-hydroxy C24 -C26 fatty acids. This is the first report that Zygomycetes species are resistant to the antifungal agent Aureobasidin A and that they contain a novel family of galactose-containing GSLs but not phosphoinositol-containing sphingolipids.
Materials-QAE-Sephadex A-25, DEAE Sephadex A-25, and D-[U- 14 C]glucose were purchased from Amersham Biosciences. Iatrobeads 6RS-8060 was obtained from Iatron Laboratories Inc. Silica gel 60 precoated plates were from Merck, magnesium silicate (Florisil) was from Nacalai Tesque, green coffee bean ␣-galactosidase was from Sigma, and jack bean ␤-galactosidase was from Seikagaku Co. Aureobasidin A was obtained from Takara Bio Inc. All other reagents used were of best grade available commercially.
Extraction and Purification of Sphingolipids-Sphingolipids were prepared from mycelia by consecutive extractions, as described elsewhere (14). Lipid extracts were saponified with 0.5 M KOH in methanolwater (95:5, v/v) at 37°C for 6 h. The hydrolysate was acidified to pH 1.0 with concentrated HCl and then dialyzed against tap water for 2 days followed by concentration and precipitation with acetone. The sphingolipids were fractionated on a QAE-Sephadex A-25 column (20 ϫ 300 mm, OH Ϫ form). The neutral fraction was further purified by silica gel chromatography (column, 15 ϫ 600 mm) with a linear gradient elution system of chloroform-methanol-water (400 ml of 90:10:0.5 by volume to 420 ml of 40:60:10 by volume). The polar fraction was then applied to a column of DEAE-Sephadex A-25 (20 ϫ 200 mm, acetate form), as described elsewhere (14).
Carbohydrate and Fatty Acid Composition Analyses-For determination of the compositions of the fatty acids and sugars in GSLs, 100 -200 g of GSLs were methanolyzed in thick glass test tubes with 200 l of freshly prepared 1 M anhydrous methanolic HCl using a microwave oven (14,15). After methanolysis, the fatty acid methyl esters were extracted three times with 400 l of n-hexane and then analyzed by capillary gas-liquid chromatography (GLC)/MS (14,15). The remaining methanolic phase was evaporated to dryness for deacidification under a nitrogen stream. The residue containing methylglycosides was trimethylsilylated and then analyzed by GLC. Sphingoids prepared from GSLs by methanolysis with 1 M aqueous methanolic HCl at 70°C for 18 h were converted to their O-trimethylsilyl (N-free) derivatives and then analyzed by GLC/MS (14,15).
Methylation for Sugar Linkage Analysis-For determination of the sugar linkages of oligosaccharides in GSLs, 300 g of a purified GSL was partially methylated with NaOH and CH 3 I in Me 2 SO (16). The permethylated GSL was acetolyzed and hydrolyzed with 300 l of a mixture of HCl-water-acetic acid (0.5:1.5:8 by volume.) by exposure to the maximum power of the microwave oven for 1 min and then was reduced with NaBH 4 and acetylated with a mixture of acetic anhydridepyridine (1:1, v/v) at 100°C for 15 min. The partially methylated alditol acetates thus obtained were analyzed by GLC and GLC/MS (14,15).
Labeling Studies of GSLs-Fungal cells were grown on YPG liquid medium at 28°C for 48 h, and then mycelia were collected and washed with distilled water. Mycelia were incubated with 20 l of [ 14 C]glucose (7.4 MBq/ml) at 28°C. Incubation was stopped at the appropriate times, and lipids were extracted from the mycelia with a solvent mixture of chloroform-methanol-water (30:30:10, by volume). They were separated by TLC and visualized with an imaging analyzer (Fujifilm, BAS2000).
Cleavage of Sugar Linkages by Exoglycosidases-␣-Galactosidase from green coffee beans and ␤-galactosidase from jack beans were used for exoglycosidase cleavage of the sugar linkages of oligosaccharides in GSLs. Samples (10 -30 g) were suspended in 0.1 ml of 50 mM Tris-HCl buffer (pH 6.5) for ␣-galactosidase treatment and 50 mM citrate buffer (pH 3.5) for ␤-galactosidase treatment, respectively, in the presence of 0.1 mg of sodium taurodeoxycholate. Each reaction was carried out with 0.25 units of ␣-galactosidase and 0.5 units of ␤-galactosidase, respectively, at 37°C for 12 h and was stopped by adding 0.5 ml of chloroformmethanol (2:1, v/v). The hydrolysate, after extraction into the lower phase, was dried under a nitrogen stream and then analyzed by TLC.
1 H NMR Spectroscopy-NMR spectra of the purified neutral GSLs were obtained with a JEOL A-500 500 MHz 1 H NMR spectrometer at 60°C as the operating temperature. Each purified GSL was dissolved in 0.6 ml of dimethyl sulfoxide-d 6

Growth of Fungi Resistant to
AbA-During studies on GSLs of fungi, we found that all Zygomycetes species examined were resistant to AbA and grew in the medium containing AbA (0.1-10 g/ml). As shown in Fig. 1A, AbA strongly inhibits the growth of filamentous fungi such as A. oryzae, P. oxalicum, and Acremonium sp., previously reported for the yeast Saccharomyces cerevisiae (W303-1A) and other fungi (7)(8)(9)(10)13). However, growth inhibition by AbA was not observed for any Zygomycetes species such as M. hiemalis, R. microsporus, R. pusillus, and A. corymbifera (Fig. 1B). These results suggested that the GSLs in Zygomycetes species were very different from those in other fungi. Therefore, we analyzed the GSLs in Zygomycetes species to investigate the AbA resistance mechanism.
GSLs of Various Fungi Zygomycetes Species-GSLs of fungi were separated into neutral, acidic, and zwitterionic fractions by ion-exchange column chromatography based on their polarities. Each fraction was analyzed by TLC with a chloroformmethanol-water system. The GSLs of all Zygomycetes species examined were recovered in the neutral fraction but not in the acidic and zwitterionic fractions. In general, acidic GSLs are found in all fungal cells and have been reported to be inositolphosphate-containing sphingolipids such as glycosylinositolphosphoceramides (1-3). Surprisingly, they were not found in Zygomycetes species (Fig. 2A), and only neutral GSL (NGLs) being present. A TLC of the NGLs of all Zygomycetes species gave the same pattern but they were apparently different from those of other fungi (Fig. 2B). The NGLs of all Zygomycetes species comprised five components, which were identified to be ceramide mono-, di-, tri-, tetra-, and pentasaccharides (tentatively named as CMS, CDS, CTS, CTeS, and CPS, respectively). On the other hand, for the other fungi, only ceramide monosaccharide was found in the neutral fraction. Then these NSLs of M. hiemalis number 314 were further purified by silica gel column chromatography and confirmed by TLC, as shown in  Table I. The fatty acid composition of CMS comprised of C14h:0 (10.5%) and C16h:0 (89.5%) fatty acids, but CDS-CPS were composed of long chain fatty acids such as C24h:0, C25h:0, and C26h:0. The sphingoid components of CMS were 9-methyl-octadeca-4,8-sphingadienine (d19:2), eicosasphingenine (d20:1), and the unknown (18), whereas those of CDS-CPS were entirely 4-hydroxyoctadecasphinganine (phytosphingosine, t18:0). Because the ceramide compositions of CDS-CPS were entirely the same, these NGLs were supposed to be a series of intermediates of GSL biosynthesis. Such phytoceramides consisting of a 2-hydroxy fatty acid and phytosphingosine are generally found in fungal cells as the major aliphatic component of glycosyl- inositolphosphoceramides (1-3).  the positive ion mode, as shown in Fig. 3. They had different pseudomolecular ions because of different ceramide species, which were in agreement with the mass values calculated from the proposed structures; the major [MϩH] ϩ ions of CMS at m/z 698.7 and 726.3 (Fig. 3A) coincided with the mass value of 1 mol each of Glc, fatty acid (2-hydroxy C14:0 or C16:0), and sphingoid (d19:2 or d20:1) (see Table I (Fig. 5D), respectively. Enzymatic hydrolysis of the above NGLs with ␣and ␤-galactosidase also revealed the presence of ␣and ␤-galactose residues. As a result, CDS was degraded to ceramide monosaccharide (galactosylceramide (GalCer)) by ␤-galactosidase, and CTS, CTeS, and CPS were also hydrolyzed to GalCer through the sequential actions of ␣and ␤-galactosidase (data not shown).
Analyses of GSL Synthesis-To investigate the biosynthetic pathway for GSLs, fungal cells were incubated with [ 14 C]glucose, and then lipids extracted from mycelia were analyzed by TLC (Fig. 6). Our preliminary results showed that CMS-CPS were found to be labeled by 14 C on incubation within 1 h (Fig. 6A). However, GalCer was not detected on TLC even after incubation for 12 h (Fig. 6B). We supposed that the metabolic process yielding a digalactosylceramide from a phytoceramide might be rapid.

DISCUSSION
Sphingolipids have been established to be essential components of eukaryotic cells, where they are predominantly found on the plasma membrane. Inhibitors of specific steps of sphingolipid biosynthesis have proven useful for understanding sphingolipid functions, and some of them have been used as fungicidal agents. One of the more useful agents for blocking sphingolipid synthesis in fungi is AbA, which inhibits IPC synthase in fungi cells. As all fungi and plants seem to synthesize IPC as an intermediate of the biosynthetic pathway for GSLs (Scheme 1), AbA exhibits strong fungicidal activity against many pathogenic fungi. However, we found that this agent was not entirely effective for the Zygomycetes species examined. We also found that Zygomycetes species did not have IPC but did have a novel family of neutral GSLs that are not found in other fungi. These GSLs isolated from M. hiemalis, which is a typical Zygomycetes species, were determined as Gal␤1-6Gal␤1-1Cer (CDS), Gal␣1-6Gal␤1-6Gal␤1-1Cer (CTS), Gal␣1-6Gal␣1-6Gal␤1-6Gal␤1-1Cer (CTeS), and Gal␣1-6Gal-␣1-6Gal␣1-6Gal␤1-6Gal␤1-1Cer (CPS). Their aliphatic components were the same phytoceramides consisting of phytosphingosine and C24 -C26 2-hydroxy fatty acids, which were bound through amide linkages. These ceramide moieties substantially differ from that of glucosylceramide (CMS) ( Table I). The only glucosylceramide detected was ceramide monosaccharide, i.e. we did not identify a galactosylceramide with phytoceramide, which is supposed to be the precursor of a series of galactosecontaining glycosphingolipids (CDS-CPS). It seemed that the enzymatic reaction to form CDS from galactosylceramide might proceed rapidly. In fact, there are some preliminary data about the existence of a very little amount of GalCer, which was found by means of sensible method using borated thin layer plate (data not shown). However, we could not know whether this GalCer is an intermediate of metabolic process of these novel NGLs or a degraded product from digalactosylceramide produced. Moreover, it could be speculated that digalactosylceramide is directly formed from phytoceramide by the addition of disaccharide from nucleotide diphosphate sugars. Such investigation is carrying out at present.
The biosynthesis of GSLs in Zygomycetes species seemed to be different from that described for other fungal species. In most fungi, sphingolipid synthesis begins in the endoplasmic reticulum, where phytoceramide is converted to IPC before transport to the Golgi apparatus for further glycosylation (1-3). Our results indicated that two independent ceramide groups existed in the Zygomycetes species, and the fungal cells synthesized neutral GSLs of both glucosylceramide and galactosecontaining glycosphingolipids from different ceramide pools, because the ceramide structures of the two types of GSLs were significantly different. Although glycosylinositolphosphoceramides have been detected in many fungi as important constituents of cells, we could not obtain evidence of their presence in Zygomycetes species, nor could we detect inositolphosphatecontaining sphingolipids. Surprisingly, Zygomycetes species showed strong resistance to AbA, and the above fact seems to be the reason why Zygomycetes species are resistant to AbA.
The roles of fungi in infections have been considered to be of lesser important, because only 5% of fungi have been found to be infectious. It has already been reported that aspergillosis (55%) is the most common invasive fungal disease, followed by mucormycosis (zygomycosis) (15%), fusariosis (15%), and acremoniosis (10%) (25). The pathogenic fungi responsible for these disease were not considered previously to be important human pathogens but are widely present in soil, plants, and elsewhere in the environment. Aspergillus spp. and Mucor spp. have been shown recently to be human pathogens (26). In particular, Mucor spp. cause many diseases, and other members of the Mucorales family act as opportunistic human pathogens (27). Mucorales infections are observed in a variety of disease states that cause immunosuppression associated with leukemia (28), aplastic anemia (29), organ or bone marrow transplantation (28), renal disease (30), and asthma (31). Therefore, new effective drugs for mucormycosis are required immediately. In this point, an inhibitor of the synthesis of galactose-containing GSLs might be useful. Although the functional roles of these GSLs have not been elucidated, our finding may facilitate the development of new antifungal agents for Mucorales.