Sphingolipidomics: An Important Mechanistic Tool for Studying Fungal Pathogens

Sphingolipids form of a unique and complex group of bioactive lipids in fungi. Structurally, sphingolipids of fungi are quite diverse with unique differences in the sphingoid backbone, amide linked fatty acyl chain and the polar head group. Two of the most studied and conserved sphingolipid classes in fungi are the glucosyl- or galactosyl-ceramides and the phosphorylinositol containing phytoceramides. Comprehensive structural characterization and quantification of these lipids is largely based on advanced analytical mass spectrometry based lipidomic methods. While separation of complex lipid mixtures is achieved through high performance liquid chromatography, the soft – electrospray ionization tandem mass spectrometry allows a high sensitivity and selectivity of detection. Herein, we present an overview of lipid extraction, chromatographic separation and mass spectrometry employed in qualitative and quantitative sphingolipidomics in fungi.

Characterization of various fungal sphingolipids requires the understanding of three key components (Figure 1). These are: the pathways of biosynthesis and degradation; structures of various sphingolipids being synthesized; understanding the ways for efficient extraction of sphingolipids from the cell and the precise methods for their analysis. A tremendous amount of literature in the field of fungal lipid metabolism allows us to categorically understand these components. In the sections below, we have described a literature based review of these key components of sphingolipid characterization, with an emphasis on mass spectrometry based structural and functional characterization of sphingolipids from pathogenic fungi.
The ability of fungal cells to produce sphingolipids with different sphingoid backbone structures adds to the complexity of the lipid mixtures (Guimarães et al., 2014). In addition to the backbone structure, the degree of unsaturation, fatty acyl chain lengths, methylation and hydroxylation modifications, all make it quite difficult to analyze these lipids using classical techniques like in vivo labeling (Chigorno et al., 2006), TLC (Urban et al., 1980) and GC (Cacas et al., 2012). Therefore, advanced analytical methods for the analyses of these lipids are employed to characterize their structures.
In last two decades, several mass spectrometry techniques have been used to identify and characterize the total sphingolipid pool or the "sphingolipidome." A wide variety of mass spectrometry based methods are available in literature that allows accurate analyses of these complex sphingolipid mixtures (Merrill et al., 2005(Merrill et al., , 2009Haynes et al., 2009;Wenk, 2010;Köfeler et al., 2012). Below we describe methods of extraction, chromatographic separation and the mass spectrometry based strategies to analyze sphingolipids.

LIPID EXTRACTION
Extraction of lipids is the most crucial step for lipid analysis by both classical and high throughput techniques. Currently several modified adaptations of Folch method (Folch et al., 1957) and Bligh and Dyer method (Bligh and Dyer, 1959) are being employed to extract lipids from fungal cells (Prasad and Ghannoum, 1996;Schneiter and Daum, 2006;Ejsing et al., 2009;Haynes et al., 2009). In our laboratory and others, we use the method described by Mandala et al. (1995) for lipid extraction. This method has shown a good extraction efficiency and reproducibility for lipid analyses. The scheme of lipid extraction from fungal cells is shown in Figure 3. Fungal cells are extracted in ethanol: dH 2 O:diethylether:pyridine:NH4OH (15:15:5:1:0.018; v/v) at 60 • C for 1 h as described previously (Hanson and Lester, 1980). Lipid extract is then subjected to a solution of methanol:chloroform (2:1; v/v) followed by addition of 1/3rd volume chloroform and 1/3rd volume dH 2 O and the lower organic phase reserved (Bligh and Dyer, 1959). Organic phase is dried in SpeedVac, flushed with N 2 and stored in −20 • C. At this stage, these lipid extracts can be used for the estimation of inorganic phosphate (Pi) content, dry lipid weight or for the isolation and purification specific lipid classes (Merrill et al., 2009;Rana et al., 2015). Both Pi content and dry lipid weight have been used to normalize the lipid amounts. Several groups have employed SPE techniques to further purify specific sphingolipid groups (Bodennec et al., 2000;Barreto-Bergter et al., 2004). Extensive lipid purification by SPE is cumbersome and time consuming, and is usually not required for routine lipid analysis; however, is preferred for a thorough structural characterization of complex glycosphingolipids (Barreto-Bergter et al., 2004. Glycerolipids are the major lipid contaminants that are coextracted with sphingolipids (Bligh and Dyer, 1959). A mild alkaline methanolysis (Clarke and Dawson, 1981), usually for 60 min at room temperature is sufficient to hydrolyze the ester linkages of fatty acyls of glycerolipids. This allows a Subsequently, a second Bligh and Dyer extraction is performed using methanol and chloroform. Samples at this stage can be used to determine the Pi content, lipid dry weight or to purify specific sphingolipid components using solid phase extraction. Finally, samples are base hydrolyzed using mild alkaline base hydrolysis (0.6 M KOH in methanol) to remove the glycerol backbone containing lipid contaminants. For certain fungal masses it is required to first crush the sample using glass beads in the lipid extraction buffer itself or to make the sample into powder using pestle and mortar in liquid N 2 prior to extraction.
clean extraction of alkaline-stable components, which are highly enriched in sphingolipids.
For accurate quantification of lipids by mass spectrometry, internal standards must be added prior to lipid extraction (Rana et al., 2015). Although, absolute quantification of each lipid species by mass spectrometry requires the use of isotopelabeled standard for that species (Ecker and Liebisch, 2014); however, presently this is not possible due to high synthesis costs and the large number of lipids being analyzed. Addition of one internal standard per lipid class being analyzed is widely accepted (LIPID MAPS Consortium). This is primarily because the ionization of lipid species is largely dependent upon the specific head group rather than the attached fatty acyls (Köfeler et al., 2012). The C17-sphingoid backbone lipids (Avanti Polar Lipids Inc., Alabaster, AL, USA; Matreya Inc., Pleasant Gap, PA, USA) are routinely used as internal standards as these closely resemble C18 lipids in physicochemical properties and ionization efficiencies (Rana et al., 2015). Quantification of sphingolipid species with different chain lengths can be achieved using the calibration curves of closest chain length standards (Rana et al., 2015).

METHODS OF LIPID ANALYSIS
In Vivo Labeling, Thin-layer Chromatography and Gas Chromatography Mass Spectrometry For several decades, in vivo labeling has been used to characterize the sphingolipid metabolic pathways in fungi (Chigorno et al., 2006). The technique involves the uptake of a radiolabeled precursor by cells. The radiolabeled precursor then gets incorporated into the complex lipid structures. Heavy labeled [ 14 C]-palmitate and [ 14 C]-serine are the most commonly used labeled precursors to follow sphingolipid synthesis (Chigorno et al., 2006). However, these labels get incorporated into other untargeted lipids and give high background signals.
[ 3 H]-dihydrosphingosine provides a more sphingolipid specific labeling (Karashima et al., 2013). [ 3 H]-inositol has been used to focus on IPC derivatives (Haak et al., 1997) and [ 3 H]or [ 14 C]-glucose or galactose have been used to focus on glycosphingolipids (Sasaki, 1981). Unfortunately, the very long half-life of these radiolabeled precursors presents a serious risk to health if exposed and is a challenge during disposal (Ecker and Liebisch, 2014).
Structural composition of the purified sphingolipids can be determined using GC-EI-MS (Cacas et al., 2012). Only the volatile components that can be carried by the carrier gas (He) can be analyzed by GCMS. Most sphingolipid structures are non-volatile. For this reason, sphingolipids are hydrolyzed to release the fatty acid, sugar, and LCB components, which are then derivatized into methyl esters or TMS derivatives and analyzed by GC-EI-MS (Figure 4A; Christie, 1989;Matsubara and Hayashi, 1991;Johnson and Brown, 1992; Merkle and Poppe, 1994). GC-EI-MS is not a preferred technique to analyze sphingolipids because fragmentation obtained in EI-MS is not very consistent as high energy electrons are used for fragmentation, strong ionization may completely destroy the molecular ion and extensive derivatization may generate complicated spectral patterns with poor resolution (Cacas et al., 2012).

High-performance Liquid Chromatography (HPLC)
Various sphingolipid classes can readily be separated by HPLC on both the normal phase C18 and reverse phase C8 columns. Normal phase HPLC utilizes the head group properties to obtain separation (for example, ceramides and hexosylceramides; Merrill et al., 2005). However, the reverse phase HPLC using C8 column is capable of resolving sphingolipid classes based on their hydrophobicity of carbon backbone and degree of unsaturation (for example, sphingosine and dihydrosphingosine; Rana et al., 2015). During most of the routine sphingolipid analysis the reverse phase HPLC when coupled with mass spectrometry is a powerful tool for simultaneous separation and detection of sphingolipid species. The main goal of HPLC is to chromatographically separate the sphingolipid species that cannot be resolved based on m/z (mass-to-charge) by the mass spectrometers, like isomeric and isobaric species, and to improve the sensitivity of detection. Several different buffer systems have been described as mobile phases (Merrill et al., 2005). The binary buffer system most frequently used as the mobile phase in sphingolipid analysis is: 2 mM ammonium formate + 0.2% formic acid in H 2 O (Buffer 1) and 1 mM ammonium formate + 0.2% formic acid in methanol (Buffer 2). A gradient elution of analytes using buffers 1 and 2 allows complete separation of sphingolipid species. Additional separation on HPLC may be achieved by lowering the flow rate of the mobile phase and changing the gradient conditions (Rana et al., 2015).

Advanced Mass Spectrometry
A wide variety of mass spectrometry platforms are now available that can be used to analyze the molecular composition of complex lipid mixtures (Köfeler et al., 2012). Mass spectrometry is a structure-based analysis of biological molecules (Wenk, 2010), and has been extensively used to characterize various sphingolipid species (Sugawara et al., 2010;Brügger, 2014;Li et al., 2014;Ogiso et al., 2014).
For sphingolipidomic purposes, ESI-MS/MS approach is most extensively employed. Coupled with reverse phase HPLC, ESI-MS/MS provides a simple, sensitive, and structure specific quantification of sphingolipids. Although the setup of other lipidomic platforms may be different but the basic concepts of mass spectrometry remain the same.
The primary requisition of any analytes to be analyzed by mass spectrometry depends on the fact that whether they can be readily ionized in gaseous phase without in source fragmentation (Han and Gross, 1994). ESI is a soft-ionization technique that allows the formation of intact positively or negatively charged ions that are then transmitted to the mass analyzers. In a typical ESI source, analytes in solvents are infused into the ion source via a narrow capillary needle. A high voltage, positive or negative depending upon the lipids being targeted, is applied between the tip of the needle and the inlet of mass analyzer. This results in the formation of a charged droplet which gets desolvated under high voltage, vacuum and a drying N 2 gas, leading to the production of charged molecular ions in gaseous phase (Brügger, 2014). Sphingolipids are readily ionized in ESI source to form [M + H] + molecular ions for long chain sphingoid bases, ceramides, and glycosphingolipids, or to from [M−H] − molecular ions for IPC derivatives (Shaner et al., 2009).
A scheme for HPLC-ESI-MS/MS is summarized in Figure 4B. Different scanning approaches have developed around this triple quadrupole set up (Han and Gross, 1994;Merrill et al., 2005;Bielawski et al., 2006;Guan and Wenk, 2006;Brügger, 2014;Rana et al., 2015), and these are: 1) Full scan (Figure 4C): In a full scan, the m/z of molecular ions generated in ESI are recorded in the first quadrupole or the first mass analyzer. Full scan approach is widely used in non-targeted lipidomics. However, no structural information can be deduced from these scanning. 2) Product ion scanning ( Figure 4D): In this mode, the first mass analyzer allows a molecular ion of specific m/z to pass to the collision cell. Based on the applied CE the molecular ion is fragmented into product ions using an inert collision gas (N 2 or Ar) via RF; this process is referred as CID. The m/z of these fragments is recorded in the third quadrupole or the second mass analyzer. For sphingolipids, the product ion analysis provides valuable structural information about the sphingoid backbone, polar head groups and the fatty acyl attached. For example, in positive ion mode, at high CE, glucosylceramide is fragmented into a characteristic ion representing the loss of double dehydration product of sphingoid backbone ( Figure 4D, Table 1). Similar ions characteristic to specific sphingolipid structures are identified and further used for lipid detection. Optimization of CE is important to obtain correct fragmentation of the molecular ions into desired daughter ions. CE required for achieving CID of different sphingolipid molecular species  may vary depending upon hydroxylations, unsaturations, and carbon chain length. 3) Parent-or precursor-ion scanning (PREIS; Figure 4E): Here, the mass analyzer records molecular ions based on their m/z, these ions undergo CID, and only a select daughter ion is passed into the third quadrupole. This mode is very useful in the analysis of sphingolipid species which posses a common backbone or headgroup fragment. 4) Single-reaction monitoring (SRM; Figure 4F): Only a select parent ion is recorded in the first quadrupole and undergoes CID and m/z of a select daughter ion is recorded in the third quadrupole. SRM's are highly selective and very sensitive methods. 5) Multiple-reaction monitoring (MRM; Figure 4G): Several SRM reactions can be simultaneously recorded by the mass spectrometer, however, this largely depends upon the scanning speed of the instrument. MRM scanning represents a targeted lipidomics approach and is quite successfully used for the analysis of sphingolipids. 6) Neutral loss (NL) scanning ( Figure 4H): Both first and third quadrupole records the m/z, with a constant mass offset between them. For example, a NL of 18 amu is set, which represents a loss of water molecule from the molecular ion and is characteristic of sphingolipid structures (Rana et al., 2015).
A large amount of literature containing the fragmentation pattern of different sphingolipid classes is now available. Most of these fragmentation data (product ions) were acquired using direct infusion ESI-MS/MS approaches. This, however, requires extensive offline purification of sphingolipid of interest, usually using SPE (Barreto-Bergter et al., 2004). Fortunately a large amount of literature containing the fragmentation patterns of various fungal sphingolipids is available and sphingolipid classspecific fragment and precursor ions are known for most classes (Guan and Wenk, 2006). In targeted lipidomics, the unique precursor and parent ions are selected for each sphingolipid species and a MRM based method is employed for analyses (Bielawski et al., 2006).

ANALYSIS OF FUNGAL SPHINGOLIPIDS
By using a precursor m/z and product m/z ion pairs of molecular species of interest, an accurate detection and quantification can be done. Although the complete sphingolipidome for most fungi is yet to be determined, many sphingolipid specific backbones and polar head groups have been characterized (summarized in Table 1). Targeted methods using PREIS and MRM scanning methods have been developed using these fragments to trace the sphingolipid metabolism of many fungi (Bielawski et al., 2006;Guan and Wenk, 2006;Ejsing et al., 2009;Sugawara et al., 2010;Brügger, 2014;Li et al., 2014;Ogiso et al., 2014).
For example, dihydrosphingosine (d18:0) and sphingosine (d18:1) can be identified using precursor ions of m/z 302 and 300, and product ions of m/z 266 and 264, respectively, in the positive ion mode. Ions of m/z 266 and 264 are a result of loss of two water molecules from the sphingoid backbone. Loss of one water molecule from the backbone results ions of m/z 284 and 282. However, the lipid extracts may contain a mixture of d18:1 backbone with a difference in the position of double bonds, like 4-Sphingenine (major species) and 8-Sphingenine. In this situation, a prior chromatographic separation of these species is important and can be achieved by optimizing the mobile phase. The d18:0 backbone fragment ion m/z 266 is used to analyze dihydroceramide species while the d18:1 backbone fragment ion m/z 264 is used to analyze ceramide species. The phytosphingosine backbone (t18:0) is abundant in fungal lipid extract, and is constituent of phytoceramides and IPC derivatives. A precursor ion m/z 282 is characteristic to t18:0 backbone and results from the loss of two water molecules. Similarly, a rather less abundant t20:0 (4R-Hydroxyeicosasphinganine) backbone is also detectable in phytoceramide and IPC structures using the positive ion precursor m/z 310. The presence of phosphorylinositol group/s in IPC, MIPC, and M(IP) 2 C structure often results in poor ionization efficiencies in positive ion mode. However, these lipids are readily ionized using negative ion mode and can be detected as [M-H] − . The precursor ion with m/z 241 which represents a dehydration product of the ion m/z 259 (the phosphorylinositol group) and can be used to analyze IPC and M(IP) 2 C species. Similarly, the precursor ion m/z 421 is used to analyze MIPC species (Table 1; Guan and Wenk, 2006;Ejsing et al., 2009;Sugawara et al., 2010).
The mass spectrometric analyses of purified lipids, especially cerebrosides, from different fungal sources have revealed several other complex backbone structures (Barreto-Bergter et al., 2011). Among these 9-methyl-4,8-sphingadiene (d19:2) is predominant in most fungal species like Cryptococcus, Aspergillus, Candida, and several others (Del Poeta et al., 2014). The d19:2 backbone has a characteristic fragment of m/z 276 that results from the loss of two water molecules. The corresponding hexosylceramide can be identified in positive ion mode using [M + H] + as the parent ion. The [M + H] + ions 756, 754, and 728 represent three most abundant hexosylceramides in fungi. The 2-hydroxy fatty acyl group in these structures is C 18:0 , C 18:1, and C 16:0 . It is important to mention here that often the ions for these lipid species present themselves as Na + or K + adducts in positive ion or Cl − adduct in the negative ion mode. Analysis of such samples requires optimization of the CE and accounting for the altered mass during the analysis. Interestingly, the differential fragmentation pattern of Cl − adduct of hexosylceramide in negative ESI-MS/MS mode can be used to identify the nature of the hexose moiety (glucose or galactose). This achievable by monitoring the peak intensity ratios of two characteristic product ions: m/z 179 and 89. The ion intensity patterns of 179 < 89 and 179 > 89 represent galactose and glucose moiety in the hexosylceramide structure, respectively (Han and Cheng, 2005).
Although, HPLC-ESI-MS/MS presents as a powerful tool for qualitative and quantitative analysis of sphingolipids in fungi as well as other biological systems. It is important to note that ESI-MS/MS only gives limited information about the structure, especially if we are looking at an uncharacterized structure. For example, it is not possible to assign positions of double bonds, hydroxyl groups, and methylation based on limited fragmentation information. Extensive MS n is necessary to provide some insights into the possible arrangement of the structure (Serb et al., 2009). However, such studies require more sophisticated instrumentation with high mass resolution, high sensitivity and sub part per-million accuracy are required. Confirmation of the exact structure is only possible using NMR spectroscopy of the purified lipid species (Sarkar et al., 2015). A routine sphingolipidome profiling of fungal lipids provides limited information regarding the rate at which these structures are synthesized or the activity of the enzymes that synthesize them. More detailed information of the metabolic flux of sphingolipid species may be achievable by using the labeling approaches like stable isotope (Ecker and Liebisch) or isotopomer analysis (Hellerstein and Neese, 1999). Regardless of its limitations HPLC-ESI-MS/MS remains a method of choice for quantitative sphingolipidome profiling.