| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Methods |
Institute for Clinical Chemistry and Laboratory Medicine, University of Regensburg, D-93042 Regensburg, Germany
Published, JLR Papers in Press, August 1, 2003. DOI 10.1194/jlr.D300025-JLR200
1 B. Lieser and G. Liebisch contributed equally to this work. 
2 To whom correspondence should be addressed. e-mail: gerd.schmitz{at}klinik.uni-regensburg.de
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The developed HPLC-tandem mass spectrometry methodology, with an analysis time of 3.5 min, simple sample preparation, and automated data analysis, allows high-throughput quantification of sphingoid bases from crude lipid extracts and is a valuable tool for studies of cellular sphingolipid metabolism and signaling.
Abbreviations: CER, ceramide; C17-SPH, 17-carbon chain sphingosine; CV, coefficient of variation; ESI, electrospray ionization; FB1, fumonisin B1; IS, internal standard; MS/MS, tandem mass spectrometry; OPA, ortho-phthaldialdehyde; SPA, sphinganine; SPH, sphingosine; SPP, sphingosine-1-phosphate
Supplementary key words high-performance liquid chromatography • sphingolipids • metabolism • high throughput
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The existing detection methods for sphingoid bases involve different derivatization steps in order to enable quantification of the low cellular concentrations. A commonly used method established by Merrill et al. (6, 7) utilizes HPLC with fluorescence detection after ortho-phthaldialdehyde (OPA) derivatization of SPH and SPA. Other procedures are based on the conversion of SPH to radioactive derivates. Thus, SPH can be acetylated by [3H]acetic-anhydride to N-[3H]acetyl-SPH, which is separated by TLC, scraped, and analyzed by liquid scintillation counting (8). Other enzymatic methods include SPH kinase-mediated conversion of SPH to [32P]SPP, followed by TLC separation and autoradiography (9, 10) or the conversion of SPH to C6-CER, with subsequent generation of C6-CER-1-[32P]phosphate by diacylglycerol-kinase (11). All of these methodologies based on SPH-derivatization are multistep procedures, which are, in part, complicated or time consuming and therefore of limited value for high-throughput analysis.
As an alternative approach, mass spectrometry, as a highly sensitive and selective methodology, has been applied for SPH quantification. Two methods have been described that involve an HPLC separation prior to tandem mass spectrometric detection (LC-MS/MS) (12, 13). One protocol, published by Mano et al. (12), used N-acetyl-SPH (C2-CER) as internal standard (IS) and allowed the simultaneous analysis of several sphingolipids within a total run time of 15 min. Sullards et al. (13) described the separation of SPH, SPA, and their corresponding 1-phosphates within a run time of 13 min. Recently, Seefelder et al. (14) presented a methodology for quantification of SPH and SPA without lipid extraction prior to LC-mass spectrometry (LC-MS) analysis. Phytosphingosine was used as IS, and the methodology showed a total run time of 14 min.
The aim of the present study was to develop a simple and fast methodology based on the high sensitivity and selectivity of electrospray ionization-tandem mass spectrometric (ESI-MS/MS) for the quantification of SPH and SPA that is applicable for high-throughput analysis.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Cell culture
Fibroblasts were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with L-glutamine, nonessential amino acids, and 10% fetal calf serum (FCS) in a humidified 5% CO2 atmosphere at 37°C, and confluent cells were used for the experiments. Other cell types were cultured according to American Type Culture Collection recommendations. Adherent cell types were cultured to confluence, and suspension cells were grown to the recommended cell density before analysis.
Sample preparation
Cells were rinsed two times with ice-cold phosphate-buffered saline (PBS) and scraped in PBS. After centrifugation at 240 g for 7 min, the resulting pellet was homogenized in distilled water by sonication. An aliquot was taken for protein determination. C17-SPH was added as IS (175 ng per mg of cellular protein). Lipids were extracted according to the method of Bligh and Dyer (15). The dried lipid extracts were resolved in the mobile phase. Calibration lines were generated by adding different amounts of naturally occurring SPH and SPA to cell homogenates.
LC-MS/MS
Samples were quantified by LC-MS/MS. A Waters Alliance 2790 (Milford, MA) was used with a constant flow of 300 µl/min and a mobile phase of methanol-chloroform (3:1; v/v) containing 0.1% (v) formic acid. The HPLC column was a Thermo Hypersil-Keystone Beta Basic CYANO, 3 µm, 50 x 2 mm (Kleinostheim, Germany). The injection volume was 5 µl, and the column temperature was set to 30°C. The triple quadrupole mass spectrometer (Quattro Ultima, Micromass, Manchester, UK) was operated with the following settings: capillary voltage, 3 kV; cone voltage, 35 V; collision energy, 14 V; and a collision gas pressure of 1.75e-3 Torr argon.
The mass resolution for quadrupole 1 and quadrupole 3 was above unit resolution. The mass spectrometer was operated in multiple reaction monitoring modus monitoring the following transitions: m/z 300 > 282, 300 > 264, and 300 > 252 for SPH; 302 > 284, 302 > 266, and 302 > 254 for SPA; and 286 > 268, 286 > 250, and 286 > 238 for C17-SPH. Data analysis was performed with MassLynx software (Micromass) coupled to self-programmed Excel macros. The SPA peak area was corrected for the overlap resulting from the M+H+2 isotope peak of SPH. The percentage overlap was determined within each run by a triplicate injection of an SPH standard. The theoretical overlap of the 300 > 282 transition with the 302 > 284 transition was calculated using the MassLynx Mass Calculator.
Protein determination
Protein concentrations were measured according to the method of Smith et al. (16).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
50%, but had only a minor effect on the retention time. A further doubling of formic acid concentration to 0.1% reduced the retention time significantly, to 2.65 min (Fig. 2)
, but did not induce a further increase of signal intensity (Table 1). Using 0.1% formic acid as ionization additive, other phospholipids, such as phosphatidylcholine, sphingomyelin, phospatidylethanolamine, and phosphatidylserine, were clearly separated from the sphingoid bases (Fig. 2). Furthermore, comparison of IS signal intensity injected without lipid matrix and spiked to crude fibroblast lipid extract revealed no signal suppression (data not shown). Therefore, a mobile phase of methanol-chloroform (3:1; v/v) containing 0.1% formic acid was selected for further experiments. The use of crude lipid extracts had no negative effect on the stability of the HPLC column, with a lifetime of more than 1,000 injections and without any cleaning step (data not shown).
|
|
12% of the -H2O transition for SPH), and the loss of two water molecules (-2H2O; 10% of the -H2O transition for SPH). The intensities of all three transitions were sufficient for the analysis of fibroblast SPH levels. However, at equal concentrations, SPA exhibited only one-third the signal intensity of SPH. In addition, cellular SPA concentrations are typically lower than SPH concentrations. Therefore, only the intensities of the -H2O transition were suitable for the quantification of cellular SPA levels.
Quantification
In order to compensate for variations in sample preparation and ionization efficiency, C17-SPH, a non-naturally occurring species, was added as IS prior to lipid extraction. The ratio between analyte and IS peak area was used for quantification.
Due to the coelution of SPH and SPA, an overlap existed between the SPH+2 isotope peak (m/z 302) and the SPA transition 302 > 284. To address this issue, the percentage of overlap was experimentally determined by injection of pure SPH standard within each run. Typically, values of 2.5% were found, which is close to the theoretically calculated value of 2.29%. The experimentally measured value was used to correct peak intensity of the 302 > 284 transition. This approach was validated by the addition of SPH to a fibroblast homogenate, which did not result in an increase of SPA concentration after overlap correction (Table 2).
|
|
|
Recovery and limit of detection
SPH recovery was tested by the addition of 5–50 ng SPH prior to lipid extraction. The mean recovery was 92.0 ± 6.4%, and no correlation to the spiked amount was observed (data not shown).
For evaluating the limit of detection, an analyte-free matrix would be optimal, because particularly for ESI applications, the influence of matrix components has to be considered. However, because we did not detect any influence of sample matrix on sphingoid base signals after HPLC separation, it may be acceptable to use pure standard for determining the detection limit. Therefore we used authentic standards to assess the limit of detection, defined as a peak with a signal-to-noise ratio above three. The observed detection limits for SPH (300 > 282) and SPA were 9 fmol and 21 fmol, respectively. If SPH quantification was based on the transitions 300 > 264 and 300 > 252, the detection limit was 25 fmol.
Interday and intraday precision
The precision of the developed assay was assessed with fibroblast samples. The intraday CV was below 6% for all transitions monitored (Table 3), with the best precision for the 300 > 282 SPH transition, showing a CV of 2.8%. For SPA, an intraday CV of 4.4% was observed. Similarly, the best interday CV of 7.6% was observed for the 300 > 282 transition. For SPA, the total CV was 13%.
Sample stability
In order to investigate sample stability, sphingoid base concentrations of a fibroblast cell homogenate were measured immediately, after 4 h, and after 24 h at room temperature (Table 4). After 4 h and 24 h at room temperature, SPH and SPA concentrations were increased by 20% and 80%, compared with the freshly analyzed sample, respectively. Immediately frozen samples did not differ in sphingoid base levels compared with freshly analyzed cell homogenates (data not shown).
|
50% in medium containing BSA (Table 5). Confluent primary human keratinocytes showed sphingoid base levels comparable to those of fibroblasts, with 180 pmol/mg for SPH and 60 pmol/mg for SPA. In contrast, the leukocyte cell lines HL60 and U937 showed significantly lower sphingoid base levels. The SPH concentrations in HL60 and U937 cells were 11 pmol/mg and 14 pmol/mg, respectively (Table 5).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
20 min and the required OPA derivatization. In contrast to these conventional methodologies, mass spectrometry offers advantages because of its high sensitivity and selectivity, and may serve as platform for sphingolipid profiling. Several methodologies for mass spectrometric SPH analysis have already been described, and all of these methods use C18 reversed-phase HPLC columns for the separation of the sphingolipid analytes (12–14). Because these columns usually separate the analyte from the IS, an accurate correction of possible interferences with the ESI process or of machine instabilities is not ensured. The method of Mano et al. (12) allows the quantification of several different sphingolipids; however, C2-CER was used as IS for SPH quantification. This could be a potential source of error, because CER clearly differs in its chemical structure from SPH and shows a different ionization and fragmentation behavior. A second study by Sullards (13) applies as IS C20-SPH, which occurs naturally (18) and which may disturb the analysis. In addition, the paper does not include data concerning precision and detection limits, and leaves unclear whether the IS coelutes with the analytes. Therefore, a detailed comparison with the method described in this manuscript is not possible. A recently published methodology by Seefelder et al. (14) uses LC-MS for SPH and SPA quantification without prior lipid extraction. This methodology has a total run time of 14 min, and SPH shows a partial chromatographic overlap with the IS phytoSPH, but both are clearly separated from SPA. Although their study was performed with a triple quadrupole mass spectrometer, fragment analysis was not used to improve specificity. This may possibly be due to signal-suppressing matrix effects from the injected crude sample, resulting in a low sensitivity, which does not permit the analysis of molecule fragments.
The aim of the current study was to establish a fast and simple quantitative detection method for SPH and SPA from crude lipid extracts using ESI-MS/MS. Because quantification of sphingoid bases by direct flow injection analysis was prevented by high signal suppression, a fast HPLC methodology with a clear separation between sphingoid bases and other major phospholipids was developed. The signal suppression for sphingoid bases was completely eliminated, and correction for possible interferences with ionization and fragmentation as well as for machine instabilities was ensured by the coelution of the IS with SPH and SPA. C17-SPH should be superior, compared with the IS used in other mass spectrometric methods, because it is a non-naturally occurring analog and differs only by one methylen group from the main mammalian SPH species, C18-SPH. In contrast to already-existing MS/MS methods, which use only the collisional-induced loss of water, two additional transitions were monitored for SPH quantification. Linear calibration lines for SPH and SPA were generated by the addition of standards, and these calibration lines were independent of the cellular matrix, allowing the use of one single calibration line for samples from different cell types. Cellular SPH concentrations calculated from standard curves based on the three different transitions were similar (Table 3), supporting the high specificity of the method. For SPA, only the collision-induced loss of water delivers a signal intensity sufficient to monitor cellular concentrations; this is due primarily to the reduced ionization efficiency of SPA compared with SPH. This phenomenon may be related to a reduced charge stabilization because of the missing double bond in SPA. Moreover, cellular SPA concentrations are commonly lower than those of SPH (Table 5).
Because the SPH+2 isotope interferes with SPA quantification, SPA measurements were corrected by an experimentally determined overlap factor, which is similar to the theoretically calculated value (Table 2). This factor should be determined within each run, because it may be influenced by mass calibration and the resolution of the mass spectrometer.
The applied simple, one-step lipid extraction showed almost complete recovery of SPH, and removed protein and hydrophilic contaminants from the sample. Thus, after more than 1,000 injections, no significant changes of analyte peak shape or signal intensity were observed (data not shown). The methodology showed an excellent in-run precision, with a CV below 6% for all monitored transitions and a total CV below 13%. For SPH, the highest precision was obtained for the 300 > 282 transition, which may be due to the higher calibration line slope and signal intensity of this transition, compared with those of 300 > 264 and 300 > 252 (Fig. 3). As a consequence, the 300 > 282 transition shows a significantly lower detection limit, compared with the other SPH transitions. Therefore, this transition should be used for SPH quantification, whereas the 300 > 264 and 300 > 252 transitions could serve as an internal quality check. The achieved detection limit was significantly lower, compared with that of already-published methodologies for SPH quantification. Thus, Ohta et al. (8) described a working range down to 10 pmol for a radioactive acetylation assay. Twenty-five picomoles as the lower limit of quantitation was reported for an assay applying a combination of chemical and enzymatic conversion by Van Veldhoven et al. (11), as well as for an SPH kinase-mediated generation of radioactive SPP (10).
Sphingoid base levels were analyzed in several different cell types, and significant differences were observed. Compared with fibroblasts and keratinocytes, the leukocyte cell lines HL60 and U937 were characterized by lower sphingoid base concentrations. In accordance with our findings, Ohta et al. (8) reported 4-fold higher SPH levels in a gastric cancer cell line, compared with HL60 and U937 leukemia cell lines. This study used a derivatization-based methodology, with conversion of SPH into radioactive C2-CER, and reported SPH concentrations of 6.3 pmol/106 cells for HL60 and 6.8 pmol/106 cells for U937. We found SPH concentrations of 11.1 and 14.3 pmol/mg cell protein in HL60 and U937 cells, respectively. Related to cell numbers, the concentrations were 1.4 pmol/106 HL60 cells and 1.9 pmol/106 U937 cells (data not shown). The higher concentrations found by Ohta et al. may, in part, arise from methodological differences, because their method did not differentiate between SPH and SPA. Alternatively, such differences may also result from different cell culture conditions, as has been suggested by previous publications (19, 20). Indeed, we observed a great difference in sphingoid base levels for NIH 3T3 fibroblasts cultured in medium containing either 10% FCS or 0.2% BSA. However, another study using MS/MS (12) reported SPH concentrations of 2.0 pmol/106 HL60 cells, which is in good accordance with our value. In HACAT cells, Min et al. (7) showed an SPH concentration of 230 pmol/mg and an SPA concentration of 110 pmol/mg using OPA derivatization and HPLC analysis. These data are similar to the sphingoid base levels in primary human keratinocytes as measured in the present study (Table 5).
Further evidence for the validity of the developed methodology derives from the observed increase of sphingoid base levels after treatment of NIH 3T3 fibroblasts with the CER synthase inhibitor FB1 (Table 5) (17). The more-pronounced increase of cellular SPA compared with SPH concentrations maybe due to a more active sphingolipid biosynthesis compared with SPH reacylation or a higher inhibitory effect of FB1 on sphingolipid biosynthesis.
In summary, all of these data support the validity of the developed ESI-MS/MS method. In contrast to already existing protocols, the assay offers advantages with respect to simple sample preparation, a short analysis time of 3.5 min per sample, as well as the ruggedness and sensitivity of the measurement. These features, coupled with highly automated data processing software, make this methodology a valuable tool for investigating the metabolism and signaling of sphingolipids.
|
ACKNOWLEDGMENTS |
|---|
Manuscript received July 9, 2003 and in revised form July 25, 2003.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  S. C.D. van IJzendoorn, J. M. van der Wouden, G. Liebisch, G. Schmitz, and D. Hoekstra Polarized Membrane Traffic and Cell Polarity Development Is Dependent on Dihydroceramide Synthase-Regulated Sphinganine Turnover Mol. Biol. Cell, September 1, 2004; 15(9): 4115 - 4124. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| All ASBMB Journals | Journal of Biological Chemistry |
| Molecular and Cellular Proteomics | ASBMB Today |
