Resolution and quantitation of diacylglycerol moieties ...

Can. J. Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by 199.201.121.12 on 06/05/13 For personal use only.

Resolution and quantitation of diacylglycerol moieties of natural glycerophospholipids by reversed-phase liquid chromatography with direct liquid inlet mass spectrometry S. PIND, a. KUKSIS,' J. J. MYHER,AND L. MARAI Depurtment of Biochemistry and Bunting and Best Department ofM~diculResearch, University of Toronto, 112 College Street, Toronto, Ont., Camda M5G 1L6 Received December 16, 1983 Pind, S., Kuksis, A., Myher, J. J. & Marai. L. (1984) Resolution and quantitation of diacylglycerol moieties of natural glycerophospholigids by reversed-phase liquid chromatography with direct liquid inlet mass spectrometry. Can. J. Biochern. Cell Biol. 62, 30 1-309 The sn-l,2-diacylglycerol moieties of natural phosphatidylcholines as the tertiary-butyldimethylsilyl ethers were resolved on the basis of their carbon number and degree of unsaturation by high pressure liquid chromatography on reversed-phase C18 colums. Using acetonitrile and propionitrile as eluting solvents and reagent gases, the yields of both quasi molecular and fragment ions were found to vary with the degree of unsaturation and positional distribution of the fatty acids in the diacylglycerol molecules, and appropriate calibration factors were necessary for accurate quantitation. In the absence of pure structural isomer and mixed acid standards, we have determined preliminary calibration factors for total and specific ion current responses, by comparing the peak area ratios obtained by liquid chromatography - mass spectrometry with the weight and mole proportions of molecular species known to be present in the samples from detailed analyses by capillary gas-liquid chromatography on polar liquid phases. It was found that the total chemical ionization current response agreed closely with the weight composition of the molecular species. The relative yields of the (MH - 132)' ions varied over one- to three-fold, while those of the (MH RCOBH)' ions varied over a three- to four-fold range of intensities. After suitable calibration of the relative ion response, it was possible to determine the identity and quantity of all common molecular species in the test diacylglycerophospholipids. Although the derived factors include both chromatographic and mass spectrometric effects and are obtained with a gradient of reagent gases, they appear to be generally applicable. Pind, S., Kuksis, A., Myher, J. J. & Marai, L. (1984) Resolution and quantitation of diacylglycerol moieties of natural glycerophospholipids by reversed-phase liquid chromatography with direct liquid inlet mass spectrometry. Can. J. Bischem. Cell B i d . 62, 301 -309 Aprks avoir converti les portions sn-1,2-diacylglycCrol des phosphatidylcholines naturelles en leurs Cthers tertiaires butyldimCthylsily1 correspondants, nous les avons sCparCes sur la base de leur nombre d'atomes de carbone et de leur degrC d'insaturation par chromatographie liquide ii haute pression sur des colonnes Cis ri phase inverse. Utilisant l'acktonitrile et le propionitrile c o m e solvants d9Clutionet gaz rCactifs, les rendements des ions quasi molCculaires et des fragments d'ions varient avec le degre d'insaturation et la distribution des positions des acides gras dans les molCcules de diacylglycCrolset des facteurs de calibration appropries sont necessaires pour une quantification prCcise. En absence d'isomkres stmcturaux purs et d'acides mixtes standard, nous avons dCterminC les facteurs de calibration prCliminaires pour les rCponses de courant ionique totales et spCcifiques en comparant les rapports des surfaces des pics obtenus par chromatographie liquide couplCe ii la spectromttrie de m s s e avec les proportions en poids et en moles des esp&cesmolCculaires dont la presence dans les Cchantillons est reconnue grice aux analyses dCtaillCes par chromatographie gaz-liquide sur colonne capillaire effectuke avec des phases liquides polaires. Nous avons trouvC que la rCponse du courant d9ionisationchimique totale s'accorde Ctroitement avec la composition en poids des espkces molCculaires. Les rendements relatifs des ions (MH - 132)' exckdent de un h trois fois les limites d'intensitb tandis que les rendements des ions (MW - RCOOH)' exckdent ces limites de trois B quatre fois. Aprks calibration appropriCe de la rCponse ionique relative, il est possible de dkterminer l'identitt et la quantitC de toutes les espkces molCculaires c o m u n e s dans l'analyse des diacylglyctrolphospholipides. Bien que les facteurs dCrivts incluent des effets chromatographiques et aussi de spectromttrie de masse et qu'ils soient obtenus avec un gradient de gaz rdactifs, ils semblent gCnCralement applicables. [Traduit par la revue]

Intr~ducti~n Previous work (1-3) has demonstrated that reversedphase HPLC with direct liquid inlet mass spectrometry AssaEvl~rroNs:HPLC, high Pressure liquid tography; t-BDMS, tertiary-butyldimethylsilyl; LC-MS , combined liquid chromatography and mass spectrometry; GLC, gas-liquid chromatography; TLC, thin-layer chromatography; TMS, trimethylsilyl; id, inside diameter; psi, pounds w r square inch; GC-MS, combined gas-liquid chromaiography and mass spectrometry; TI, total ion cirrent. -l~btGorto whom all c~rres~ondence should be addressed.

is well suited to separation, identification, and quantitation of molecular species of triacylglycerols in natural fats and oils. It has also been shown that the method is well suited to the separation and identification of free diacylglycerols (4), which are resolved according to both molecular weight and degree of unsaturation, as we" as the positional substitution of the glycerol In the present we have the free diacylglycerols to the corresponding t-BE)MS ethers, which are stable under the HPLC conditions and which yield excellent separations on the reversed-phase


302

CAN. J. BIOCHEM. CELL BIOL. VOL. 62, 1984

columns, as well as informative ion fragments upon chemical ionization mass spectrometry. In addition, we have obtained preliminary calibration factors for the t-BDMS ethers of various mixed acid diacylglycerols, by comparing the peak area proportions obtained by LC-MS with the proportions of the corresponding peaks resolved by capillary GLC on polar liquid phases and analyzed by hydrogen flame ionization detection, which yields correct weight response. The calibration factors have been applied to the quantitative analysis of several mixtures of diacylglycerols derived from natural glycerophospholipids and the results have been compared with the values obtained by capillary GLC resulting in an excellent agreement.

Materials and methods The various standard sn-1,2- and sn-l,3-diacylglycerols were available in the laboratory from previous studies (5). They had been stored in closed containers as either the free diacylglycerols or as their t-BDMS ethers. Prior to use they were purified by TLC as previously described (5). The cocoa butter oil was a gift from Dr. David Kritchevsky, Wistar Institute, Philadelphia, PA. The triacylglycerols were isolated from this oil by TLC with heptane - isopropyl ether - acetic acid (60:40:4) as the developing solvent. sn- 1,2(2,3)-Diacylglycerol and sn-1,3-diacylglycerol were prepared from these triacylglycerols by partial Grignard degradation and were purified by TLC as described (6). The sn- l,2-diacylglycerol moieties of phosphatidylcholine from egg yolk, soybean oil, and rat liver were obtained by hydrolysis with phospholipase C (Bacillus cereus, type V, Sigma Chemical Co., St. Louis, MO) as previously described (7). In all instances the parent phospholipids had been purified by TLC as described elsewhere (8). Acetonitrile (ealdon Laboratories Ltd., Georgetown, Ont .) and propionitrile (Fluka Chemical Corp., Hauppauge, NY) were HPLC grade. Trimethylchlorosilane and hexamethyldisilazane were from Pierce Chemical Co., Rockford, IL, while the t-BDMS-imidazole reagent was obtained from Applied Science Laboratories, State College, PA. All other chemicals and solvents were of reagent grade or better quality and were obtained from Fisher Chemical Co., Toronto, Ont. They were used without further purification. Preparation oJ TMS and t -BDMS ethers The TMS ethers of the diacylglycerols were prepared by reacting the sn-l,2-diacylglycerols with trimethylchlorosilane-hexmethyldisilazane-pyridine as described (5). The TMS ethers were dissolved in hexane and used for the GLC analyses. The t-BDMS ethers of the diacylglycerols were prepared by reacting the free sn-l,2-diacylglycerols with the t-BDMS-imidazole reagent as previously described (9). The t-BDMS ethers were recovered from the reaction mixture by extraction into petroleum ether (bp 30-60°C), followed by extensive washing with water. The petroleum ether extract was then passed through anhydrous sodium sulfate and concentrated to dryness under N2 at 40°C. For HPLC analyses the t-BDMS ethers were dissolved in propionitrile (5 mg/mL). GLC instrumentation and chromatographic conditions The GLC analyses were performed with a Hewlett-Packard

model 5880 capillary gas chromatograph equipped with a hydrogen flame ionization detector and a level 4 microprocessor. TMS ethers of the diacylglycerols were resolved on the basis of carbon number by means of a flexible quartz column (8 m x 0.30 mm id) coated with a permanently bonded nonpolar SE-54 liquid phase (Hewlett-Packard, Palo Alto, CA). The samples were injected on the column and the oven temperature was programmed from 40 to 150°C at 30°C /min, then to 230°C at 2O0C/min, to 280°C at 10°C/min, and to 340°C at S0C/min. The carrier gas was hydrogen at 5 psi (1 psi = 6.894757 kPa) head pressure. TMS ethers of the diacylglycerolswere resolved according to carbon number and degree of unsaturation by means of a capillary glass column (10 m X 0.25 mm id) coated with a polar SP-2330 liquid phase (Supelco). The column was operated isothermally at 250°C with hydrogen as the carrier gas (5 psi head pressure). The diacylglycerol peaks were identified by reference standards and by the relative retention times tabulated previously (6).

MINUTES

FIG. 1. LC-MS profiles of the sn-l,2-diacylglycerol moieties of the phosphatidylcholines of rat liver (A) and egg yolk (B). Peak identity is as given in Table 1. Column, Supelcosil-18 reversed phase (250 X 4.6 mm id); mobile phase, 30-90s linear gradient of propionitrile in acetonitrile; column temperature, 30°C; detection, total chemical ionization current. Ordinate, total ion current intensity with major peak as 100%. Abscissa, elution time in minutes. Other LC-MS conditions were as given in text. Sample: 500 pg of sn-1,2-diacylglycerol-3-t-BDMSethers in 250 FL of propionitrile.

PIND ET AL.

TABLE1. Major ions* in the chemical ionization spectra of t-BBMS ether of diacylglycerols as detected by LC-MS with propionitrile-acetonitrile gradient


HPEC peak?

(MH - RCOOH)' , m/z

(MH - 132)', m / ~

(MH)' , m/z

Molecular species$

*The identity of the ions is explained in the text. tPeak numbers refer to the numbering system used in Fig. 1. $The molecular species are written from left to right to indicate sn-1- and stl-2-positions. respectively.

HPLC instrumentation and chromatographic conditions The HPLC analyses were performed with a HewlettPackard model 1084B liquid chromatograph (Hewlett-Packard, Balo Alto, CA) equipped with a Supelcosil LC-18 column (250 x 4.6mm id) using a gradient of 30-9095 propionitrile in acetontrile. The columns were operated at a flow rate of 2.0 mL/min and 30°C oven temperature. About 1W-250 p L of the diacylglycerol solution was admitted to the column by means of an automatic sample injector. The diacylglycerol peaks were detected by chemical ionization mass spectrometry using a direct liquid inlet interface as described below. Mass spectrometric instrumentation and operating conditions The mass spectrometry was perfomed with a HewlettPackard model 5985B quadrupole mass spectrometer equipped with a Hewlett-Packud direct liquid inlet interface as previously described (I). About 1% of the HPLC column effluent was admitted to the mass spectrometer and full mass spectra (200-850 mass units) were recorded every 7 s over the entire elution profile. The data were analyzed by means of a Hewlett-Packard data system (model HP 1008E) and a graphics terminal (model WP 26488) as previously described (1). The total mass spectra were corrected for background by subtraction of a scan made with propionitrile-acetonitrile

alone. The LC-MS response to the various molecular species was determined by constructing mass chromatograms for the various ions or groups of ions of interest by means of the data system.

ResuZts and discussion Nature of PIPLC resolution Figure 1 shows the HPLC elution patterns obtained for the t-BBMS ethers of the sn-l,2-diacylglycerol moieties of phosphatidylcholines from rat liver and egg yolk using the propionitrile-acetonitrile gradient. The general order of peak elution is based on the carbon number and degree of unsaturation of the fatty chains of the diacylglycerols, with the shorter and more unsaturated species emerging first. Under identical elution conditions, the free diacylglycerols emerge about six partition numbers earlier than the corresponding tBBMS ethers. In both instances the sn- 1,3-diacylglycerols are eluted after the corresponding sn- 1,2(2,3)diacylglycerols , although a complete base-line resolution is not obtained (results not shown). Since the t-BDMS ethers are stable under the HPLC conditions, these derivatives are preferred over the free diacylgly-


CAN. 9. BIOCHEM. CELL BIOL. VOL. 62, I984

FIG. 2. Chemical ionization mass spectrum of sn-glycerol- 1-stearate 2-arachidonate-3-t-BDMS ether (peak 12, Fig. 1A). LC-MS conditions are as given in Fig. 1. Ordinate, relative ion current intensity with major ion as 100%. Abscissa, m j z . TABLE 2. Reproducibility of LC-MS total ion current profiles of t-BDMS ethers of sn- 1,2-diacylglycerd moieties of rat liver and egg yolk phosphatidylcholines

HPLC peak*

Rat liver, area % ( n = 39

HPLC peak

Egg yolk, area 5% ( n = 4)

5 6 7 9 BQ, 11 12 13 14, 15 16 17, 18 *Peak numbers refer to the numbering system used in Fig. 1. tResdts are expressed as mean 9 SD for M determinations.

cerols for the liquid chromatographic separation. The various peaks have been identified by the mass spectra and their identity was found to be consistent with the fatty acid compositisn of the original samples, as well as with the expected behavior of these molecules in the reversed-phase HPLC system. Table 1 gives the major ions detected in each of the various peaks and the corresponding identity of the t-BDMS ethers of the diacylglycerols. It can be seen that a large number of molecular species can be identified in this way, including a number of species present in very low amounts. Reproducibility s f H P L 6 pro$les Table 2 gives the results of repeat analyses of the t-BDMS ethers of the sn- 1,2-diacylglycerol moieties of

rat liver and egg yolk phosphatidylcholines by LC-MS. The peak area values represent the percentages of total chemical ionization current. The mns were recorded on different days, but under otherwise identical working conditions. It can be seen that the relative peak area ratios differ by less than 5% for peak areas making up more than 18% of the peak area and less than 18%for peak areas making up from 1 to 10% of total peak area. For peaks contributing less than 1% of the total peak area, the relative error could be much higher than 10%. Comparative relative errors were recorded for consecutive HPLC runs using direct liquid inlet LC-h4S as a detector. Nature of mass spectra Figure 2 shows the mass spectrum for sn-glycerol1-stearate 2-aachidonate-3- t-BDMS ether. The spectrum contains a significant protonated molecular-ion (m/z 7591, as well as a fragment representing the intact diacylglycerol (m/z 627). The major peaks are due to a loss-of Gachidonic acid (m/z 455) and of stearic acid ( m / z 475). However, both fatty acids are not lost at the same rate. Table 3 shows that the relative yields of the various ion fragments vary from species to species, depending on the degree of unsaturation and the positional distribution sf the fatty acids. Thus, the relative increase of the abundance of the (MH)' ion with unsaturation indicates that the presence of a polyunsaturated acyl chain tends to stabilize this ion. Furthermore, in all species the fatty acid in the sn-2-position is lost to a greater extent (three- to four-fold) than the acid in the sn-1-position. Thus, while the 18: 1 and B8:2 acids are released at comparable rates from the sn-2-position, the 18:2 acid is lost three to four times faster than the 18:1 acid, when the Batter is in the sn-1-position as in sn-glycerol-1-oleate-2-linoleate. The nature of the fatty

PHND ET AL.

TABLE3. Relative abundances sf the major ion fragments derived from the t-BDMS ethers of various sn- 1,2- and sn- l,3-diacylglycerols

Ion fragments, area % Molecular


species

(MH - R'COBH)'

(MH - W~COOH)' (MH - 132)'

(MH)'

*The sn-l,2-diacylglycemls were from phosphatidylcholines o f rat liver and egg yolk. The sn-1- and sn-2positions are writeen from left to right. ?Ratios o f a e a percentages relative t o 16:O 18:2 species o f sn-l,2-diacylglycerols. Ssn-1,3-Diacylglycerols were prepared from cocoa butter triacylglycerols by partial Grignard degradation and acid isornerization. N o distinction is being made between the sn-1- and sn-3-positions of these diacylglycerols. $Ratios o f area percentages relative to 16:O 18:l species o f the sn-l,3-diacyclglycerols.

acid is also important. Although an exact comparison is difficult to obtain, the results are consistent with the following relative order of release of the fatty acids from the same sn-position of the sn- l,2-diacylglycerols thus far examined: 2 2 5 > 20:4 > 18:2 r 18:1 > 18:O 14:O. Comparing the relative intensities of the ions resulting from the loss of either RCBOH or t-BDMS-OH from positions 1 or 3, we also see that RCOOH is the better leaving group. This is consistent with the apparent stabilization of the (MH)' ion when the poorer leaving group is in position 2 (sn- 1,3-diacylglycerols) . Thus, the (MH)' intensity is approximately 40% in the spectra of the t-BDMS ethers of the sn-l,3-diacylglycerols. Since the t-BDMS group is now in position 2, the (MH - 132)' ion is also more intense than in the spectra of the sn- l,2-diacylglycerols. In all other respects, the chemical ionization spectra of the t-BDMS ethers of the sn- 1,3- and sn- 1,2-diacylglycerols are similar. Table 3 also includes the relative response factors for the (MH - 132)' and the (MH - RCBOH)' ions for the major molecular species of the diacylglycerols. These factors relate the different ion responses to the total ion

current, which was already shown to be proportional to the mass of the molecular species. For exact quantitation via the ion response, these factors should be further adjusted for the slight differences in the yields of the total ion current among the different molecular species. The latter correction, however, was found to remain within the error of the overall analysis. Quantitatisn of resolved peaks Examination of the mass spectra of each HPLC peak in Fig. 1 indicates that most peaks contain single components and that only a few are made up of two or more molecular species in significant amounts. Since the quantitative composition of the molecular species of the diacylglycerol moieties of rat liver and egg yolk phosphatidylcholine is known (6), it is possible to derive a set of response factors for a diverse variety of molecular species. Within the precision of the method the response factors shown in Table 4 are almost the same for all the molecular species. Since these species include a wide range of unsaturation classes varying from saturates to hexaenes, it can be assumed that the total ion response is the same for all the individual

306

CAN. J. BIBCHERI. CELL BHOL. VOL. 62. 1984


TABLE4. Total ion current response factors for the t-BDMS ethers of selected sn- 12-diacylglycerol moieties of rat liver and egg yolk phosphatidylcholines" Molecular species

HPLC /GLC?

16:O 22:6 1830 20:4 18:O 226 B6:O 182 1810 20:4 16:O 18:l 18:O 182 16:O l6:O 18:O 18:l Average *Average values determined from analyses of the sn-l,2-diacylglycerol moieties of phosphatidylcholines from rat Iiver and egg yolk were used. The values are normalized to 1.O for 16:O 1 8 2 . TWEC values represent total ion sun1 from m / r 400 to YQO; the GLC values represent the hydrogen flame ionization response.

molecular species. The response factors for the various fragment ions, however, vary with the molecular species as indicated in Table 3.

Quantitatio~~ of unresolved peaks TOobtain an LC-MS estimate of the concentration of the unresolved components within an HBLC peak, it is necessary to quantitate the intensities of the characteristic ions. Generally, the intensity of one or both (MH RCOOH)' ions can be unambiguously assigned to a particular molecular species. Where only one of a pair can be measured the other can be calculated from the h s w n fragmentation pattern of reference compounds. Since it can be shown from the data in Table 3 that the abundances obtained for the sum of the two (MH RCOOH)' ions is a constant percentage of the total ($6.0 + 2. I), the quantities of the overlapping species are proportional to the respective sums. Alternately, it is possible to use the (MH - 132)' ions. In this case, however, the response is species dependent and must be appropriately corrected (see Table 3). Figure 3 shows the mass chromatograms recorded for the (MH - B 32)' ions for the major species of the diacylglycerol moieties of rat liver phosphatidylcholine. It is seen that the shoulders (peaks 6 and 8) sf peak 7 are made up of the species 16:0 2 2 5 + 18:1 20:4 and 18:O 205, respectively, which have identical masses of m / z 625. Since peaks 6 and 8 are completely resolved, their proportions can be obtained from the ratios of the peak

MINUTES

FIG.3. Mass chromatograms of the "diacylglycerol" (MH 132)' ions for the major species of rat liver phosphatidylcholine. LC-MS conditions are as given in Fig. 1. Peak identity is given in Table 1. The n z / z values are shown on the left and the ion scale is shown on the right of each mass chromatogram.

areas. The two components in peak 6 must be calculated from the relative abundances of the ions m / z 453 and 501, which represent the (MH - WCOOH)' ions retaining the 18:I and 2 2 5 acids, respectively. The species B6:0 20.4 emerges between peaks 6 and 8 and its peak area is traced out by the ion m / z 599 (MH 132)'. Since the (MM - WCOOM)' ions make up comparable proportions of the total ion current in the polyenoic diacylglycerols, the relative proportions of the sum of these four species can be calculated as a simple proportion of the sum of the peak areas recorded on the same ion scale. The overall relative concentration of each species is then computed as a proportion of the contributions of peaks 6 , 7 , and 8 to the total ion current, which is already known. The (MH - 132)' ions at m / z 577 and 603 in the mass chromatograms in Fig. 3 show that peak 14 is composed of 16:O l8:1 and B8:O 18:2 (and (or) 18:l 18:1) species, which overlap. The relative proportions ofthe B6:O B8:1 andofthe 18:O B8:2plusany 18:l B8:l species can be obtained from the relative intensities of

PIND ET AL.

TABLE5. Molecular species of sn-l,2-diacylglycerols of rat liver phosphatidyl-

cholines as determined by LC-hlS and by capillary GLC on polar liquid phases HPLC


peak*

Molecular speciest

- 132)'. area %

(MH

(MH - RCQQH)' area %

,

Capillary GLC, weight %

*Peak numbers refer to the numbering system used in Fig. 1. tThe molecular species are written from left to right to indicate sn-1- and sn-2-positions. respectively. $Not determined.

the (MW - 132)+ ions, since these species yield comparable proportions of the total ion current in the form of the (MW - 132)' ions. The relative content of the 18:1 18:1 species can be calculated from the relative proportions of the various (MH - RCOOH)' ions at m / z 453 (oleate retained), 45 1 (linoleate retained), 427 (palmitate retained), and 455 (stearate retained) in the (MH - RCOOH)' ions for the diacylglycerol species overlapping in peak 14. Quantitation of unknown mixtures The excellent agreement shown in Table 5 between the capillary GLC and LC-MS estimates following calibration indicates that the latter method of quantitation yields valid data, despite the use of a gradient system for peak elution and chemical ionization in the mass spectrometer. Obviously the exact composition of the acetonitrile-propionitrile mixture is not critical for the present work, as both of these reagents possess comparable potential for chemical ionization of the glycerolipid molecules. By means of calibration factors derived from the

analyses of the diacylglycerol moieties of rat liver phosphatidylcholines by LC-MS and by polar capillary GLC, we have determined the molecular species composition of the diacylglycerol moieties of other glycerophospholipids. Table 6 gives the composition of the diacylglycerol moieties of the phosphatidylcholines of egg yolk and soybean oil, along with the estimates from capillary GLC on polar columns. Overall, the agreement between the results is good. Certain discrepancies, however, remain and may be attributed to the use of secondary standards for the calibration. Alternatively, the differences may be real and may represent the incompleteness of the GLC analyses and (or) recoveries in the capillary columns. It is also possible that some of the difference is due to an incorrect assignment of the positional distribution of the fatty acids. Linearity of response In the absence of pure standards the linearity of the LC-MS response was assessed indirectly. It is obvious that both major and minor components are recovered in approximately the correct proportions from the diacyl-

308

CAN. S. BIOCHBM. CELL BIOL. V O L . 62, 1984

TABLE6. kfolecular species of sn-1,2diacylglycerol moieties of the phosphatidylcholines of egg yolk and soybean oil Area %


Species*

LC-MSt

GLC$

2.2 3.6 1.0 2.7 0.6 Soybean oil 2.2 29.8 3.2 0.7 5.5 10.7

39.0 8.2 0.7 *The molecular species are written from left to right to indicate sn-1- and sn-2-psitions, respectiveiy. t(MH - 132)*. $Capillary OLC values adopted from Ref. 6 . PNot determined.

glycerol moieties of egg yolk, rat liver, and soybean oil phosphatidylcholines, which have been analyzed in great detail previously by capillary GLC. The relative concentrations of these components vary from 0.1 to 25 % of the total diacylglycerol mixture, which represents a range of 1 to 250. Since both minor and major components differ in composition from one phosphatide to the other, it must be concluded that all the components give a linear response over the investigated range of concentrations. In most instances about 1 mg of total diacylglycerol mixture was injected and about 1% of the column effluent was admitted to the mass spectrometer. Therefore, the absolute concentration range over which a linear response was observed in the present experiments may be estimated to be 10 ng to 2.5 kg. These results are similar to those obtained previously for triacylglycerols under comparable working conditions (3).

Previous work Prior to the present studies, molecular species of natural diacylglycerols have been quantitated on the basis of the intensities of the molecular ions following direct probe or GLC method of sample introduction, and these reports have been reviewed along with our own work on the GC-MS of the t-BDMS ethers of diacylglycerols (5). Extensive qualitative and quantitative studies on both diacyl- and alkylacyl-glycerols from natural sources have been made using both TMS and t-BDMS derivatives in other laboratories (9- 13). The overall agreement between the GC-MIS and chemical data was reasonable, but major differences (up to 20%) were noted for some species (1 I). In the saturated diacylglycerol series, Myher et al. ( 5 ) were able to differentiate between the reverse isomers. because the fatty acid in the sn-2-position yielded less of the fragment (M - W2COO)+ than the acid in the sn-1-position (M - RICOO)+. On the basis of this observation it was suggested (14) that the reverse isomers of unsaturated diacylglycerols could be quantitated following deuteration, which would allow the identification of the original unsaturated acids in the sn-1- and sn-2-position on the basis of the deuterium content. Such analyses have been carried out by Dickens et al. (12). Comparable MPLC separations of natural sn-1,2diacylglycerols as the acetates have been recently reported by Nkagawa and Morrocks (15), who used GLC to identify and quantitate each peak on the basis of its fatty acid composition. The latter technique is less direct and much more laborious than HPLC with direct liquid inlet mass spectrometry. The present results are also consistent with those obtained previously by LC-MS for mixtures of natural triacylglycerols (2-4). The mass analyses obtained for the sn-1,2-diacylglycerols by LC-MS are similar to those derived by capillary SLC on polar liquid phases (6, 16), although the capillary GLC with hydrogen flame ionization detection is much more sensitive than the LC-MS combination. Theoretically, the polar capillary columns (6) could be connected to the mass spectrometer, but no practical combinations of this type have yet been devised for work with diacylglycerols. The main advantages of the LC-MS system described here are the ease of identification of the polyunsaturated species and the capability of a simultaneous measurement of the mass and stable isotope (deuterium or carbon-13) content of each molecular species of the diacylglycerols in the chemical ionization mode of operation of the LC-MIS system (17).

Acknowledgments These studies were supported by funds from the Medical Research Council of Canada, the Ontario Heart

PINB ET AL.


Foundation, and the Hospital for Sick Children Foundation, Toronto, Ont . 1. Kuksis, A., Mmi, L. & Myher. J. J. (1983) J. Chromatogr. 2 73, 43-66 2. Mxai, L., Myher, J. J. & Kuksis, A. (1983) Can. J. Biochem. Cekk Biok. 61, 840-849 3. Myher, J. J., Kuksis, A., Marai, L. & Manganaro, F. (1984) J. Chromatogr. 283, 289-301 4. Kuksis, A., Myher, J. J. & Marai, L. (1983) J. Am. Oil Chem. Soe. 60, 735 (abstr. 227) 5. Myher, J. J., Kuksis, A., Marai, L. & Yeung, S. K. F. (1978) Anal. Chem. 50, 557-561 6. Myher, J. J. & Kuksis, A. (1982) Can. J . Biochem. 60, 638-650 7. Myher, J. J. & Kuksis, A. (1979) Can. J. Biochem. 57, 117-124 8. Skipski, &I. P., Peterson, 8 . F. & Barclay, M. (1964) Biochern. J. 90. 374-378

309

9. Ogino, H.,Matsurnura, T o . Saeouchi, K. & Saito, K. (1979) Biochim. Biophys. Acta 574, 57-61 10. Satouchi, K. & Saito, K. (1979) Biomed. Mass Spectrom . 6, 396-402 11. Kins, M., Matsurnura, T., Garno, M. & Saieo, K. (1982) Biomed. Mass Speebrorn. 9 , 363-369 12. Dickens, B. F., Rernesha, C. S. & Thompson, S . A., Jr. (1982) Anal. Biochern. 127, 37-48 13. Dickens, B. F. & Thompson, G. A , , Jr. (1982) Biochemistry 21, 3604-361 1 14. Kuksis, A. & Myher, J. J. (1980) in ,kfemhrcane Fluidity (Kates, M. & Kuksis, A , , eds.), pp. 3-32, Hurnana Press, Clifton, NJ 15. Nakagawa, Y. & Hsrrocks. L. A. (1983) J. Lipid Res. 24, 1268- 1275 16. Myher, J. J. & Kuksis, A. (1984) Can. J. Biochern. Cell Biok. 62, in press 17. Kuksis, A. (1982) J. Am. Oil Chem. Soc. 59, 265A (abstr. 1)