Quantitative determination of phospholipid compositions by ESI-MS: effects of acyl chain length, unsaturation, and lipid concentration on instrument response

Electrospray ionization-mass spectrometry (ESI-MS) is a very promising tool for the analysis of phospholipid compositions, but is hampered by the fact that not all molecular species are detected with equal efficiency. We studied this and other issues that need to be taken into account to obtain truly quantitative compositional data. The key findings were as follows: First, the instrument response for both saturated and unsaturated phospholipid species decreased with increasing acyl chain length. This effect became increasingly prominent with increasing overall lipid concentration. Second, the degree of acyl chain unsaturation also had a significant effect on instrument response. At the highest concentration studied (10 pmol/μl), polyunsaturated species gave 40% higher intensity than the fully saturated ones. The effect of unsaturation diminished and nearly disappeared with progressive dilution. Third, the instrument response for the different head group classes varied markedly depending on the infusion solvent used. Notably, inclusion of ammonia in the infusion solvent eliminated sodium adduct formation in the positive ion mode, thus greatly simplifying the interpretation of the spectra. The fact that instrument response is dependent on many structural features, overall lipid concentration, solvent composition, and instrument settings makes it necessary to include several internal standards for each phospholipid class to obtain accurate data. Preferably, both unsaturated and saturated standards should be used. Finally, we quantified the major phospholipid classes of BHK cells using ESI-MS. The data agreed closely with those obtained with thin-layer chromatography and phosphorus analysis. This study indicates that quantitative compositional data can be obtained with ESI-MS, provided that proper attention is paid to experimental details, particularly the choice of internal standards. —Koivusalo M., P. Haimi, L. Heikinheimo, R. Kostiainen, and P. Somerharju. Quantitative determination of phospholipid compositions by ESI-MS: effects of acyl chain length, unsaturation, and lipid concentration on instrument response. J. Lipid Res. 2001. 42: 663–672. Electrospray ionization-mass spectrometry (ESI-MS) is a very promising tool for the analysis of phospholipid compositions, but is hampered by the fact that not all molecular species are detected with equal efficiency. We studied this and other issues that need to be taken into account to obtain truly quantitative compositional data. The key findings were as follows: First, the instrument response for both saturated and unsaturated phospholipid species decreased with increasing acyl chain length. This effect became increasingly prominent with increasing overall lipid concentration. Second, the degree of acyl chain unsaturation also had a significant effect on instrument response. At the highest concentration studied (10 pmol/μl), polyunsaturated species gave 40% higher intensity than the fully saturated ones. The effect of unsaturation diminished and nearly disappeared with progressive dilution. Third, the instrument response for the different head group classes varied markedly depending on the infusion solvent used. Notably, inclusion of ammonia in the infusion solvent eliminated sodium adduct formation in the positive ion mode, thus greatly simplifying the interpretation of the spectra. The fact that instrument response is dependent on many structural features, overall lipid concentration, solvent composition, and instrument settings makes it necessary to include several internal standards for each phospholipid class to obtain accurate data. Preferably, both unsaturated and saturated standards should be used. Finally, we quantified the major phospholipid classes of BHK cells using ESI-MS. The data agreed closely with those obtained with thin-layer chromatography and phosphorus analysis. This study indicates that quantitative compositional data can be obtained with ESI-MS, provided that proper attention is paid to experimental details, particularly the choice of internal standards. —Koivusalo M., P. Haimi, L. Heikinheimo, R. Kostiainen, and P. Somerharju. Quantitative determination of phospholipid compositions by ESI-MS: effects of acyl chain length, unsaturation, and lipid concentration on instrument response. J. Lipid Res. 2001. 42: 663–672. Compositional analysis of phospholipids is of considerable interest because they are involved in many crucial cellular functions such as signal transduction, apoptosis, and protein sorting (1Martin T.F. Phosphoinositide lipids as signaling molecules: common themes for signal transduction, cytoskeletal regulation, and membrane trafficking.Ann. Rev. Cell Dev. Bio. 1998; 14: 231-264Google Scholar, 2Bevers E.M. Comfurius P. Dekkers D.W. Zwaal R.F. Lipid translocation across the plasma membrane of mammalian cells.Biochim. Biophys. Acta. 1999; 1439: 317-330Google Scholar). At present, the analysis is complicated by the low sensitivity and complexity of the commonly available methods. Regarding the former, hundreds of nanomoles of total phospholipid are needed to establish the molecular species profiles of the different classes (3Patton G.M. Fasulo J.M. Robins S.J. Separation of phospholipids and individual molecular species of phospholipids by high-performance liquid chromatography.J. Lipid Res. 1982; 23: 190-196Google Scholar). This precludes detailed analyses when subcellular fractions obtained from cultured cells, for instance, are studied. The complexity, on the other hand, derives from the fact that several chromatographic steps and/or enzymatic reactions are required to establish the molecular species patterns. In view of these complications, it is hardly surprising that, so far, a complete molecular species profile (i.e., including all major classes) has been quantitatively determined for very few, if any, mammalian cells or membranes. Mass spectrometry (MS) offers an attractive alternative for the analysis of phospholipid compositions because of its high sensitivity, specificity, and (apparent) simplicity. However, MS has rarely been used for this purpose until very recently. The main reason for this is that the ionization methods previously available (fast-atom bombardment, etc.) cause extensive fragmentation of the lipid molecules; this, along with the extreme complexity (hundreds of different molecular species) of most biological samples, has precluded compositional analysis. The recent introduction of "soft" ionization methods has opened completely new vistas in this field. In particular, electrospray ionization-mass spectrometry (ESI-MS) has been shown to be a very promising technique (4Kerwin J.L. Tuininga A.R. Ericsson L.H. Identification of molecular species of glycerophospholipids and sphingomyelin using electrospray mass spectrometry.J. Lipid Res. 1994; 35: 1102-1114Google Scholar, 5Kim H.Y. Wang T.C. Ma Y.C. Liquid chromatography/mass spectrometry of phospholipids using electrospray ionization.Anal. Chem. 1994; 66: 3977-3982Google Scholar, 6Han X. Gross R.W. Electrospray ionization mass spectroscopic analysis of human erythrocyte plasma membrane phospholipids.Proc. Natl. Acad. Sci. USA. 1994; 91: 10635-10639Google Scholar, 7Myher J.J. Kuksis A. Electrospray MS for lipid identification.INFORM. 1995; 6: 1068-1072Google Scholar, 8Han X. Gubitosi-Klug R.A. Collins B.J. Gross R.W. Alterations in individual molecular species of human platelet phospholipids during thrombin stimulation: electrospray ionization mass spectrometry-facilitated identification of the boundary conditions for the magnitude and selectivity of thrombin-induced platelet phospholipid hydrolysis.Biochemistry. 1996; 35: 5822-5832Google Scholar, 9Brügger B. Erben G. Sandhoff R. Wieland F.T. Lehmann W.D. Quantitative analysis of biological membrane lipids at the low picomole level by nano-electrospray ionization tandem mass spectrometry.Proc. Natl. Acad. Sci. USA. 1997; 94 ([published erratum appears in Proc. Natl. Acad. Sci. USA. 1999. 96:1>0943]): 2339-2344Google Scholar, 10Ramanadham S. Hsu F.F. Bohrer A. Nowatzke W. Ma Z. Turk J. Electrospray ionization mass spectrometric analyses of phospholipids from rat and human pancreatic islets and subcellular membranes: comparison to other tissues and implications for membrane fusion in insulin exocytosis.Biochemistry. 1998; 37: 4553-4567Google Scholar, 11Schneiter R. Brugger B. Sandhoff R. Zellnig G. Leber A. Lampl M. Athenstaedt K. Hrastnik C. Eder S. Daum G. Paltauf F. Wieland F.T. Kohlwein S.D. Electrospray ionization tandem mass spectrometry (ESI-MS/MS) analysis of the lipid molecular species composition of yeast subcellular membranes reveals acyl chain-based sorting/remodeling of distinct molecular species en route to the plasma membrane.J. Cell. Biol. 1999; 146: 741-754Google Scholar, 12Hsu F.F. Ma Z. Wohltmann M. Bohrer A. Nowatzke W. Ramanadham S. Turk J. Electrospray ionization/mass spectrometric analyses of human promonocytic U937 cell glycerophospholipids and evidence that differentiation is associated with membrane lipid composition that facilitate phospholipase A2 activation.J. Biol. Chem. 2000; 275: 16579-16589Google Scholar). However, there are still many issues to be resolved before ESI-MS can become a routine tool for quantitative determination of lipid compositions. A key problem is that different molecular species are not detected with equal efficiency (6Han X. Gross R.W. Electrospray ionization mass spectroscopic analysis of human erythrocyte plasma membrane phospholipids.Proc. Natl. Acad. Sci. USA. 1994; 91: 10635-10639Google Scholar, 9Brügger B. Erben G. Sandhoff R. Wieland F.T. Lehmann W.D. Quantitative analysis of biological membrane lipids at the low picomole level by nano-electrospray ionization tandem mass spectrometry.Proc. Natl. Acad. Sci. USA. 1997; 94 ([published erratum appears in Proc. Natl. Acad. Sci. USA. 1999. 96:1>0943]): 2339-2344Google Scholar, 13Fridriksson E.K. Shipkova P.A. Sheets E.D. Holowka D. Baird B. McLafferty F.W. Quantitative analysis of phospholipids in functionally important membrane domains from RBL-2H3 mast cells using tandem high-resolution mass spectrometry.Biochemistry. 1999; 38: 8056-8063Google Scholar). For instance, it has been shown that the detection sensitivity (instrument response) can depend markedly on the acyl chain length (9Brügger B. Erben G. Sandhoff R. Wieland F.T. Lehmann W.D. Quantitative analysis of biological membrane lipids at the low picomole level by nano-electrospray ionization tandem mass spectrometry.Proc. Natl. Acad. Sci. USA. 1997; 94 ([published erratum appears in Proc. Natl. Acad. Sci. USA. 1999. 96:1>0943]): 2339-2344Google Scholar). However, other important factors remain to be studied. First of all, it is not clear whether and how the instrument response depends on the degree of unsaturation of phospholipid acyl chains. Second, it is unknown whether or not the relative response factors are affected by the overall lipid concentration. The data on the effect of the phospholipid head group structure on instrument response are somewhat contradictory. In this study, we have investigated these and other issues that need to be considered when using ESI-MS for quantitative determination of (phospho)lipid compositions. Synthetic PC, PE, and PS standards were obtained from Avanti Polar Lipids (Birmingham, AL), fatty acids from Larodan AB (Malmö, Sweden), and sphingosylphosphorylcholine from Matreya, Inc. (Pleasant Gap, PA). Phospholipase D from Streptomyces species (type VII) was obtained from Sigma (St. Louis, MO), and silica gel 60 high performance thin-layer chromatography (HPTLC) plates from Merck (Darmstadt, Germany). Sphingomyelins (SMs) with a 14:0, 17:0, 23:0, or 25:0 fatty acid residue were synthesized from sphingosylphosphorylcholine and the respective fatty acid as described previously (14Cohen R. Barenholz Y. Gatt S. Dagan A. Preparation and characterization of well defined D-erythro sphingomyelins.Chem. Phys. Lipids. 1984; 35: 371-384Google Scholar, 15Ramstedt B. Slotte J.P. Interaction of cholesterol with sphingomyelins and acyl-chain-matched phosphatidylcholines: a comparative study of the effect of the chain length.Biophys. J. 1999; 76: 908-915Google Scholar). The SM product was purified using a NH2-bonded silica cartridge (16Kaluzny M.A. Duncan L.A. Merritt M.V. Epps D.E. Rapid separation of lipid classes in high yield and purity using bonded phase columns.J. Lipid Res. 1985; 26: 135-140Google Scholar), followed by reverse-phase HPLC on an Ultrasphere ODS column (Beckmann, 5-μm particle size, 250 × 4.6 mm) eluted with 5% chloroform in methanol at 1 ml/min. Di-16:0 and di-18:0-PI species were isolated from hydrogenated yeast PI by reverse-phase HPLC. Alternatively, Di-16:0-PI was obtained from Larodan AB. Di-21:0-PE and -PS were synthesized from the corresponding PC by using phospholipase D-mediated transesterification (17Kasurinen J. Somerharju P. Metabolism and distribution of intramolecular excimer-forming dipyrenebutanoyl glycerophospholipids in human fibroblasts. Marked resistance to metabolic degradation.Biochemistry. 1995; 34: 2049-2057Google Scholar). The products were purified on a Lichrosphere 100 DIOL column (5-μm particle size, 250 × 4.6 mm; Alltech, Deerfield, IL) as described previously (18Silversand C. Haux C. Improved high-performance liquid chromatographic method for the separation and quantification of lipid classes: application to fish lipids.J. Chromatogr. Biomed. Sci. Appl. 1997; 703: 7-14Google Scholar). All solvents were of high performance liquid chromatography (HPLC) or analytical grade and were purchased from Merck or Rathburn Chemicals Ltd. (Walkerburn, Scotland). All standard mixtures were prepared in silane-treated autosampler vials (Alltech) in chloroform/methanol (C/M) 1:2 to avoid loss due to adsorption to glass. Several different equimolar PC mixtures were prepared. The compositions of these mixtures are indicated in the respective figure legends. PE, PS, and PA standard mixtures were prepared from these PC mixtures by phospholipase D-mediated transesterification or hydrolysis (see above). For the quantification of the cellular phospholipid species, a mixture consisting of the following internal standards was used: 14:0/14:0-PC, 19:0/19:0-PC, 21:0/21:0-PC, 24:0/24:0-PC, 14:0-SM, 17:0-SM, 23:0-SM, 25:0-SM, 14:0/14:0-PE, 16:0/16:0-PE, 21:0/21:0-PE, 14:0/14:0-PS, 16:0/16:0-PS, 21:0/21:0-PS, 16:0/16:0-PI, and 18:0/18:0-PI. BHK-21 cells were cultured on plastic dishes (Nunc, Roskilde, Denmark) in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 2 mM glutamine, penicillin (200 U/ml), and streptomycin (200 μg/ml) under 5% CO2 at 37°C. Cellular lipids were extracted (19Folch J.M. Lees M. Sloane-Stanley G.H. A simple method for the isolation and purification of total lipids from animal tissue.J. Biol. Chem. 1957; 226: 497-509Google Scholar) in silane-treated screw-cap tubes, and the extract was divided into three aliquots. One aliquot was used to determine the total phospholipid content of the extract (20Bartlett E.M. Lewis D.H. Spectrophotometric determination of phosphate esters in the presence and absence of orthophosphate.Anal. Biochem. 1970; 36: 159-167Google Scholar), the second aliquot was used for lipid class distribution analysis by TLC (21Stone S.J. Vance J.E. Cloning and expression of murine liver phosphatidylserine synthase (PSS)-2: differential regulation of phospholipid metabolism by PSS1 and PSS2.Biochem. J. 1999; 342: 57-64Google Scholar) and phosphorus analysis (20Bartlett E.M. Lewis D.H. Spectrophotometric determination of phosphate esters in the presence and absence of orthophosphate.Anal. Biochem. 1970; 36: 159-167Google Scholar), and the third aliquot was spiked with PC, SM, PE, PS, and PI internal standards and used for ESI-MS analysis. Most experiments were carried out both with an ion trap instrument (Esquire-LC, Bruker-Franzen Analytik, Bremen, Germany) and a triple quadrupole instrument (Perkin Elmer Sciex API 300). The lipids were dissolved in C/M 1:2 with or without NH4OH (0.25–1%) or 0.1 mM NaCl, and were infused to the electrospray source via a 50-μm id fused silica capillary using a syringe pump at the flow rate of 5 μl/min. With the ion trap, nitrogen was used as the nebulizing (at 5–6 psi) and the drying gas (5–7 l/min at 200°C). The potentials of the spray needle, capillary exit, and skimmer 1 were set to ± 4,000, 90–150, and 25–50 V, respectively. For each spectrum 100–500 scans were averaged. With the triple quadrupole instrument, the spray capillary voltage was 4,000 V and orifice voltage 25 V in positive and negative ion ESI. Synthetic air was used as the nebulizing gas, and nitrogen was used as the curtain and collision gas. The m/z range scanned in the MS measurements was from 500 to 1,000 (5 s/scan), and in MS/MS from m/z 50 to 10 mass units above the m/z value of precursor ion (15–300 s/scan). Collision-activated dissociation was applied in the second quadrupole at a collision energy of 25–40 eV. The phospholipid species were identified on the basis of i) their characteristic m/z value, ii) fragmentation analysis, and iii) precursor ion or neutral loss scans. In the positive ion mode, PE and PS give the characteristic neutral losses of 141 or 185, respectively, whereas PC and sphingomyelin can be identified by scanning for precursors of m/z 184 (9Brügger B. Erben G. Sandhoff R. Wieland F.T. Lehmann W.D. Quantitative analysis of biological membrane lipids at the low picomole level by nano-electrospray ionization tandem mass spectrometry.Proc. Natl. Acad. Sci. USA. 1997; 94 ([published erratum appears in Proc. Natl. Acad. Sci. USA. 1999. 96:1>0943]): 2339-2344Google Scholar, 22Lehmann W.D. Koester M. Erben G. Keppler D. Characterization and quantification of rat bile phosphatidylcholine by electrospray-tandem mass spectrometry.Anal. Biochem. 1997; 246: 102-110Google Scholar, 23Duffin K.L. Henion J.D. Shieh J.J. Electrospray and tandem mass spectrometric characterization of acylglycerol mixtures that are dissolved in nonpolar solvents.Anal. Chem. 1991; 63: 1781-1788Google Scholar), and PI species can be selectively detected by scanning for precursors of m/z 241 in the negative ion mode (9Brügger B. Erben G. Sandhoff R. Wieland F.T. Lehmann W.D. Quantitative analysis of biological membrane lipids at the low picomole level by nano-electrospray ionization tandem mass spectrometry.Proc. Natl. Acad. Sci. USA. 1997; 94 ([published erratum appears in Proc. Natl. Acad. Sci. USA. 1999. 96:1>0943]): 2339-2344Google Scholar). The instrument response can vary depending on the phospholipid acyl chain length (9Brügger B. Erben G. Sandhoff R. Wieland F.T. Lehmann W.D. Quantitative analysis of biological membrane lipids at the low picomole level by nano-electrospray ionization tandem mass spectrometry.Proc. Natl. Acad. Sci. USA. 1997; 94 ([published erratum appears in Proc. Natl. Acad. Sci. USA. 1999. 96:1>0943]): 2339-2344Google Scholar). To account for this and for the isotope effect, several internal standards at equimolar concentrations were included. A correction function fstd(m) can then be defined by plotting the measured intensity of the internal standards versus m/z. The concentration of an analyte species can be obtained from the equation: C(I,m) = I * Cstd/fstd(m), where Cstd is the concentration of the internal standards and m and I are the m/z value and peak intensity of the analyte species, respectively. At low total lipid concentrations, the instrument response is a linear function of m/z, thus correction function is of the form fstd(m) = k * m + c. At high total lipid concentrations the exponential function fstd(m) = a * e bm better approximates the chain length dependency. The degree of lipid acyl chain unsaturation can also affect the instrument response. In principle, it is possible to correct for this by including unsaturated internal standards as well. However, because such standards were not available for all lipid classes (e.g., PI), the samples were measured at low total lipid concentrations, at which the effect of unsaturation is eliminated or at least greatly diminished (see Results). Phospholipid classes were separated on HPTLC silica gel 60 plates using chloroform–methanol–acetic acid–formic acid–water 70:30:12:4:2, (v/v/v) as the solvent (21Stone S.J. Vance J.E. Cloning and expression of murine liver phosphatidylserine synthase (PSS)-2: differential regulation of phospholipid metabolism by PSS1 and PSS2.Biochem. J. 1999; 342: 57-64Google Scholar). The lipid bands were scraped off the plates and their phosphorus content was determined (20Bartlett E.M. Lewis D.H. Spectrophotometric determination of phosphate esters in the presence and absence of orthophosphate.Anal. Biochem. 1970; 36: 159-167Google Scholar). A previous study indicated that the acyl chain length of PC can have a significant effect on the instrument response (9Brügger B. Erben G. Sandhoff R. Wieland F.T. Lehmann W.D. Quantitative analysis of biological membrane lipids at the low picomole level by nano-electrospray ionization tandem mass spectrometry.Proc. Natl. Acad. Sci. USA. 1997; 94 ([published erratum appears in Proc. Natl. Acad. Sci. USA. 1999. 96:1>0943]): 2339-2344Google Scholar). To study this issue in more detail, an equimolar mixture of PC standards containing 11 saturated and 5 diunsaturated species (each at 5 pmol/μl) was prepared and analyzed with the ion trap instrument. As apparent from Fig. 1, the response decreased markedly and in a nearly linear manner with increasing acyl chain length for both the saturated and diunsaturated species. Notably, however, the instrument response for the unsaturated species was somewhat higher than for the corresponding saturated species. This effect of acyl chain unsaturation will be studied further below. Parallel results were obtained with the triple quadrupole instrument, and also for PE, PS, and PA standard mixtures (data not shown). To study whether the total phospholipid concentration influences the chain length dependency of instrument response, an equimolar mixture of saturated PC species was diluted to 10, 5, 1, 0.5, 0.2, or 0.1 pmol/μl per species (120–1.2 pmol total lipid/μl). At the highest concentration (10 pmol/μl), the relative response decreased strongly and in a seemingly exponential manner with the acyl chain length (Fig. 2), which is in agreement with the previous study (9Brügger B. Erben G. Sandhoff R. Wieland F.T. Lehmann W.D. Quantitative analysis of biological membrane lipids at the low picomole level by nano-electrospray ionization tandem mass spectrometry.Proc. Natl. Acad. Sci. USA. 1997; 94 ([published erratum appears in Proc. Natl. Acad. Sci. USA. 1999. 96:1>0943]): 2339-2344Google Scholar). However, when the concentration was decreased to 5 pmol/μl (60 pmol/μl total lipid) and below, the response was an essentially linear function of the chain length. At the same time, the differences betweenthe short and long chain species diminished. For example, the response for 24:0-PC was only 2.5-fold higher than for 48:0-PC at 0.1 pmol/μl, whereas it was 5-fold higher at 10 pmol/μl (Fig. 2). We also studied the chain length dependency of instrument response with (di)unsaturated PC species (Fig. 3). Analogous to the saturated species, i) the response decreased with acyl chain length, and ii) the decrease became more prominent with increasing lipid concentration. However, the response remained a linear function of the acyl chain length, even at the highest concentration studied (10 pmol/μl). Parallel data were obtained with PE, PS, and PA standards, both with the ion trap and the triple quadrupole instrument (data not shown). To study the effect of lipid unsaturation and concentration on the instrument response in more detail, an equimolar mixture of PC standards, including five 36-carbon PC species with 0, 1, 2, 4, or 6 double bonds, was analyzed at various concentrations. As can be seen in Fig. 4, the response was markedly dependent on lipid unsaturation, particularly at the higher concentrations. For instance, at 10 pmol/μl per species (140 pmol/μl total lipid), the response for the polyunsaturated species was 40% higher than that for the fully saturated ones. However, when the mixture was progressively diluted, the effect of unsaturation gradually diminished, and virtually disappeared at 0.1 pmol/μl per species (1.4 pmol/μl of total lipid). The influence of the phospholipid head group structure on instrument response in ESI-MS has been studied previously, but the results are somewhat contradictory (see Discussion). In addition, the response for PI relative to other phospholipids has not been studied using acyl chain matched species. Therefore, an equimolar mixture of six dipalmitoyl species, i.e., 16:0/16:0-PC, -PE, -PS, -PG, -PA, and -PI was prepared, diluted to a varying degree in C/M 1:2 with or without 1% NH4OH, and spectra were then obtained with the ion trap instrument operated inthe negative or positive ion mode. In the negative ion mode with C/M 1:2 as the solvent, there were major differences in instrument response between classes, as shown in Fig. 5A. The highest response was observed for PG, whereas PI, PA, and PS gave intermediate responses, and PE and the chloride adduct of PC gave low responses. Inclusion of ammonia (1%) in the infusion solvent markedly increased the response for PE and PA relative to other species (Fig. 5C). Notably, the relative responses did not vary significantly with lipid concentration, independent of whether ammonia was included or not (Fig. 5B and C). In the positive ion mode, with C/M 1:2 as the solvent, the highest peak was observed for protonated PC, followed by the Na+-adducts of PC, PE, and other lipids (Fig. 5D). It is notable that in the case of PE and most other lipids, the Na+-adduct was much more prominent than the protonated form, whereas the opposite was true for PC (Fig. 5E). When ammonia was included, the relative intensitieschanged dramatically (Fig. 5F). At this point the intensity of protonated PC far exceeds that of the others, probably because under these basic conditions PC is a zwitterion, and the other lipids exist as anions. The relative responses varied only slightly with the total lipid concentration (Fig. 5E and F). Note that, in agreement with previous studies (24Hvattum E. Hagelin G. Larsen A. Study of mechanisms involved in the collision-induced dissociation of carboxylate anions from glycerophospholipids using negative ion electrospray tandem quadrupole mass spectrometry.Rapid Commun. Mass Spectrom. 1998; 12: 1405-1409Google Scholar), the adjustment of certain instrument parameters can markedly change the relative (and absolute) responses from those shown in Fig. 5 (data not shown). The linearity of instrument response is obviously an important concern in quantitative analysis. Previous studies have shown that a linear response can be obtained with a quadrupole instrument over at least two orders of magnitude (6Han X. Gross R.W. Electrospray ionization mass spectroscopic analysis of human erythrocyte plasma membrane phospholipids.Proc. Natl. Acad. Sci. USA. 1994; 91: 10635-10639Google Scholar, 22Lehmann W.D. Koester M. Erben G. Keppler D. Characterization and quantification of rat bile phosphatidylcholine by electrospray-tandem mass spectrometry.Anal. Biochem. 1997; 246: 102-110Google Scholar). To determine whether this is the case with an ion trap, an equimolar mixture of seven PC species varying in chain length and unsaturation was prepared and analyzed at various dilutions (0.1–10 pmol/μl per species), using the ion trap instrument. As shown in Fig. 6A, the intensities were linearly dependent on the lipid concentration up to about 1 pmol/μl, but tended to level off at higher concentrations. Such signal saturation is a typical phenomenon in ESI-MS and results from the saturation of the surface of the spray droplets by the analyte molecules (see Discussion). Importantly, however, the relativeresponses were linearly dependent on the concentration up to at least 5 pmol/μl, as shown for some species in Fig 6B. Parallel results were obtained for the PE, PS, and PA standard mixtures (data not shown). To investigate whether such linear behavior is also observed under more biologically relevant circumstances, an equimolar mixture of five PC species (30:0, 32:2, 34:4, 38:0, and 44:2) was prepared and diluted to concentrations varying from 0.1 to 10 pmol/μl per species. Each diluted mixture was then combined with BHK cell total lipid extract containing 140 pmol/μl of total phospholipid and analyzed with the ion trap instrument. The relative response was a linear function over the whole 100-fold concentration range studied for each PC species (data not shown). Thus a linear response over a considerable range of concentrations can be obtained for phospholipids, even in the presence of other cell-derived lipids and other compounds present in the crude lipid extract. After having determined the effect of structural factors and lipid concentration on instrument response, we used this information to quantitatively determine the molecular species profiles of the major phospholipids of BHK-21 cells. To this end, a total lipid extract was prepared and spiked with a mixture of internal standards for each major phospholipid class, and then analyzed with the triple quadrupole instrument using class-specific detection (9Brügger B. Erben G. Sandhoff R. Wieland F.T. Lehmann W.D. Quantitative analysis of biological membrane lipids at the low picomole level by nano-electrospray ionization tandem mass spectrometry.Proc. Natl. Acad. Sci. USA. 1997; 94 ([published erratum appears in Proc. Natl. Acad. Sci. USA. 1999. 96:1>0943]): 2339-2344Google Scholar). Representative spectra are shown in Fig. 7. Because unsaturated standards were not available for all ph