185,36-43
ANALYTICALBIOCHEMISTRY
(1990)
A Sensitive, Radiometric Assay for Lysophosphatidylcholine’ David J. Dobmeyer, Peter B. Corr, and Michael H. Creer Cardiovascular Division, Department of Internal Medicine and Department of Pharmacology, Washington University School of Medicine, St. Louis, Missouri 63110
Received
September
11,1989
tion, prevention of the accumulation of LPC during myocardial ischemia in viuo is associated with a pro1found antiarrhythmic effect (6). We have developed a model system utilizing isolated adult canine cardiac myocytes exposed to hypoxia and acidosis in vitro to further investigate the biochemical mechanisms responsible for the accumulation of LPC during myocardial ischemia. To accurately quantify the small amounts of LPC in the isolated cell preparations, an improved method for measurement of LPC which would be capable of detecting subnanomolar quantities in extracts of isolated cells was required. The analysis of LPC in isolated cardiac cells or in myocardial tissue is complicated by several factors including (i) LPC accounts for only 2 to 3 mol% of the total phospholipid composition of the myocardium (7) and therefore is present in very small quantities. (ii) Conventional methods for isolation of LPC by one-dimensional TLC do not completely resolve LPC from all other phospholipid classespresent in tissue (8,9). (iii) TLC separation in two dimensions, although providing the resoluPress. Inc. tion necessary to isolate LPC, requires at least 10 to 20 nmol of LPC to identify the LPC region on the TLC plate. (iv) HPLC methods for isolation of LPC freAlterations in the phospholipid composition of the quently employ the use of mobile phases which are comIn this mobile myocardial cell membrane, or sarcolemma, have been prised of acetonitrile/methanol/water. phase, cardiolipin (diphosphatidylglycerol), which acimplicated in the pathogenesis of ventricular arrhythcounts for 10 to 15% of the total phospholipid composimias during myocardial ischemia (1). We, and others, tion of the myocardium (lo), is not completely soluble have demonstrated that lysophosphatidylcholine (LPC), and elutes as a broad band resulting in 30 to 40 mol% an amphiphilic phospholipid derived from phosphatidylcontamination of the LPC fraction (7). (v) Several curcholine, accumulates in ischemic myocardium and has rently available HPLC methods do not completely seppotent arrhythmogenic properties in vitro (2-5). In addiarate sphingomyelin and LPC. (vi) In several animal species, plasmalogens (1-0-alk-Y-enyl-2-acyl phospho1 Research from the authors’ laboratory was supported in part by lipids) account for 30 to 40% of the total content of diraNational Institutes of Health Grant HL 17646, SCOR in Ischemic dyl choline phospholipids in myocardium (10,ll). ChoHeart Disease and NIH Grants HL 28995 and HL 36773. line plasmalogens are readily hydrolyzed in acidified 2 Abbreviations used: LPC, lysophosphatidylcholine; FAME, fatty extraction media leading to artifactual production of 2acid methyl ester; DMAP, dimethylaminopyridine; [“CILPC, l-(1acyl-lysophosphatidylcholine (12). This may also ac[%]palmitoyl)-2-lysophosphatidylcholine.
To facilitate investigation of the metabolism of lysophosphatidylcholine and choline lysoplasmalogen in small quantities of tissue, a method for the quantiflcation of these phospholipid species that is capable of accurate and reproducible analysis in samples which contain less than 1 nmol of total choline lysophospholipid was developed. The procedure employs chloroform and methanol extraction of phospholipids from isolated tissue with subsequent separation of the choline lysophospholipid fraction by high-performance liquid chromatography. The choline lysophospholipids are then acetylated with [3H]acetic anhydride and the [‘HIacetyl-lysophosphatidylcholine product is isolated by thinlayer chromatography and quantified by liquid scintillation counting. The choline lysophospholipid content in the sample is determined from a standard curve constructed from samples containing a known amount of synthetic lysophosphatidylcholine with correction for recovery based on the inclusion of [14C]lysophosphatidylcholine as an internal standard. o 1990 Academic
36 All
Copyright 0 1990 rights of reproduction
0003-2697/90 $3.00 by Academic Press, Inc. in any form reserved.
SENSITIVE,
RADIOMETRIC
ASSAY
count, in part, for the apparent increase in recovery of LPC previously reported with methods requiring the addition of acid to the extraction media (13). (vii) Due to the high content of choline plasmalogens in myocardial tissue, the LPC fraction isolated from extracts of myocardial tissue or from isolated cardiac myocytes may also contain a significant amount of choline lysoplasmalogen. The quantification of LPC by GC analysis of the fatty acid methyl esters (FAME) produced after basecatalyzed hydrolysis of LPC in methanol (14) does not permit detection of choline lysoplasmalogens since they are not hydrolyzed under these conditions (15). In our experience, even when acidic conditions are employed for hydrolysis of LPC fractions isolated by HPLC, analysis of the FAME derivatives (produced from monoacyl LPC) and dimethylacetal derivatives (produced from choline lysoplasmalogen) by GC is limited by two factors. First, GC methods require at least 2 to 5 nmol of each individual LPC species for accurate quantitation. Second, trace amounts of fatty acid contamination often cause significant interference. This interference is not easily eliminated despite intense efforts to remove the contaminating fatty acids by acid washing, silanizing all glassware, and using solvents from several different manufacturers. To overcome these limitations of conventional approaches to the quantification of LPC, we developed a method based on the acetylation of the free hydroxyl group of LPC with radiolabeled acetic anhydride to permit the accurate quantification of LPC in samples containing less than 1 nmol of total LPC. This method is not subject to interference by fatty acids, sphingomyelin, or cardiolipin which frequently contaminate the LPC fraction isolated by HPLC procedures. The method is also applicable to the measurement of choline lysoplasmalogens. MATERIALS
AND
METHODS
Soybean LPC, synthetic l-heptadecanoyl-sn-glycero3-phosphocholine (17:0 LPC), 1-0-alkyl-2-acetyl-snglycero-3-phosphocholine (platelet activating factor), cardiolipin, dimethylaminopyridine, sphingomyelin, KHzPOl, and acetic anhydride were obtained from Sigma Chemical Co. (St. Louis, MO). Phospholipids obtained from Sigma were chromatographically pure and were used without further purification except for the 17: 0 LPC which was purified by HPLC prior to use as described below. Choline lysoplasmalogen was prepared by base-catalyzed hydrolysis of bovine heart phosphatidylcholine (Avanti Polar Lipids, Birmingham, AL) as previously described (16), and 1-(1-[14C]palmitoyl)-2-lysophosphatidylcholine ( [14C]LPC) was obtained from Amersham Corp. (Arlington Heights, IL). [3H]Acetic anhydride was purchased from New England Nuclear (Boston, MA). Liquid scintillation cocktail was pur-
FOR
37
LYSOPHOSPHATIDYLCHOLINE
chased from Research Products International (Mt. Prospect, IL). Thin-layer chromatography plates, glacial acetic acid, and chloroform were obtained from Fisher Scientific (St. Louis, MO). Bovine serum albumin was purchased from Bio-Rad (Richmond, CA). Acetonitrile, methanol, water, and toluene were purchased from Taylor Chemical Co. (St. Louis, MO). Partisil-SCX HPLC columns (Whatman) were purchased from PJ Corbert Associates, Inc. (St. Louis, MO).
Protein
Assay
Cell samples were analyzed for total protein by a modification of the Lowry method as described by Markwell et al. (17). Lyophilized bovine serum albumin served as the protein standard.
Microsphosphate
Assay
Total lipid phosphorous was determined in selected samples and standards by microphosphate assay after ashing of the samples in 1.2 M Mg(N03)2 in ethanol (7). The phosphate content of the ashed material was determined by the microphosphate assay method of Chen et al. (18). Standards containing 2 to 50 nmol of phosphate were prepared from anhydrous reagent grade KH,PO,.
Extraction
of Myocyte Phospholipids
Calcium tolerant isolated adult canine myocytes were harvested from canine epicardium by collagenase dispersion as previously described (19). Isolated cardiac myocytes (approximately 2-5 X lo5 total cells, equivalent to 2-5 mg total cellular protein) were suspended in 2 ml Hepes buffer (115 mM NaCl, 5 mM KCl, 35 mM sucrose, 10 mM glucose, 10 mM Hepes, and 4 mM taurine, pH 7.4). A 250-~1 aliquot was removed for protein analysis. The remaining cell suspension was then subjected to chloroform and methanol phospholipid extraction according to the method of Bligh and Dyer (20). [ 14C]LPC (17,000 dpm) was added to the cells before extraction as an internal standard to correct for loss of LPC during the course of analysis. Following extraction, the lower organic layer was taken to dryness under N, at 37°C. The residue was resuspended in 3 ml of chloroform/methanol (l/l, v/v), filtered through 0.2-pm filters (Acrodisc), and stored at -4°C in vials sealed with Teflon-lined silicon rubber septum caps.
High-Performance
Liquid
Chromatography
The phospholipid extracts were taken to dryness under N2 and resuspended in 100 ~1 of l/l chloroform/ methanol (v/v). The entire volume was then injected onto a silica-based cation-exchange HPLC column (10 pm Partisil-SCX). Phospholipids were eluted with a mobile phase comprised of acetonitrile, methanol, and wa-
38
DOBMEYER,
CORR, AND CREER
ter (400/100/32, v/v) as previously described (21). Collection of myocyte LPC fractions was based on the retention time of unsaturated molecular species of LPC determined following a separate injection of soybean LPC standard by monitoring absorbance in the ultraviolet range at 203 nm. The LPC fraction was collected in 18 X 150-mm test tubes, taken to dryness under Nz at 37”C, resuspended in 3 ml l/l chloroform/methanol, and stored at -4°C in 3-ml conical vials sealed with Teflon-lined silicon rubber septum caps. 3H Acetylution Reaction Following separation of phospholipid fractions by HPLC, the purified LPC fraction was taken to dryness under N2 at 37°C. The residue was resuspended in 25 ~1 of 0.33 M dimethylaminopyridine in acetonitrile and 25 ~1 of [3H]acetic anhydride in toluene. The [3H]acetic anhydride reagent was prepared by diluting [3H]acetic anhydride with unlabeled acetic anhydride to a specific activity of 50-70 dpm/pmol and adjusting the final concentration to 100 nmol/pl in toluene. The reaction was allowed to continue for 12 h at 37°C. At the end of this period, the acetonitrile-toluene solution was evaporated under N2 at 37°C. One hundred microliters of water was then added to each vial and incubated at 37°C for 15 min, after which the water was evaporated under Nz at 37°C. The remaining residue was then resuspended in 25 ~1 l/l chloroform/methanol in preparation for thin-layer chromatography. A standard curve was prepared by adding HPLC-purified 17:0 LPC in amounts ranging from 0.5 to 15.0 nmol to separate vials followed by addition of dimethylaminopyridine and [3H]acetic anhydride as described above for individual samples. [ 14C]LPC (17,000 dpm) was also added to each standard curve vial before acetylation to correct for incomplete recovery of the acetyl-LPC product following TLC separation. Reagent blanks were prepared by addition of all reagents except LPC to separate vials. Three blanks were prepared for each assay (one for each TLC plate). In two separate vials, 0.5 pmol of soybean LPC was added to each vial and the solvent removed by evaporation under N2 . To each of these vials, 10 ~1 of unlabeled acetic anhydride (neat), 30 ~1 of toluene, and 25 ~1 of 0.33 M dimethylaminopyridine were added, the vials were incubated for 12 h at 37”C, and the reaction was terminated as described above. The residue in each vial was then resuspended in 50 ~1 of l/l chloroform/methanol. Five microliters of this solution (containing 50 nmol of unlabeled acetylated soybean LPC) was then added to each sample (17:0 LPC standards and LPC samples derived from cell extracts) and to the reagent blanks immediately prior to the application to the TLC plates. The unlabeled acetylated soybean LPC added to the samples served as an internal marker for identification of the region of the TLC plate containing the 3H acetylated LPC product in these samples following Iz staining as described below.
Thin-Layer Chromatography Thin-layer chromatography was performed on silica gel G plates, 20 X 20 cm, 250 pm thick using a solvent system composed of chloroform/acetone/methanol/glacial acetic acid/water (6/8/2/2/l, v/v) (system I). Samples were applied as 2-cm-wide “streaks” in adjacent lanes of the TLC plate (eight samples per plate). On each plate, 10 ~1 of soybean LPC (approx 100 nmole) was applied to lane 1, a reagent blank was applied to lane 2, and samples or standards were applied to lanes 3 through 8. TLC separation was accomplished by two successive developments in solvent (system I). At the conclusion of the TLC separation, the plates were allowed to air-dry in a fume hood for 1 h. The region corresponding to acetyl-LPC in lanes 2 through 8 was then identified by passing air through a Pasteur pipet containing Iz crystals that was directed over the TLC plate. The acetyl-LPC region was outlined with a pencil and the TLC plate was allowed to stand for 1 h in a fume hood before scraping the plate for liquid scintillation counting. This was done to allow the Iz to evaporate from the surface of the TLC plate to avoid any potential quenching effects of 12. Liquid Scintillation Counting The TLC plate regions from the reagent blanks and from samples containing the [3H]acetyl-LPC product were scraped into liquid scintillation vials. Five hundred microliters of water and 10 ml of liquid scintillation cocktail were added to each vial. The vials were vortexed thoroughly and remained in the dark overnight to minimize autofluorescence. Liquid scintillation counting was performed on a Beckman Model LS-3801 liquid scintillation counter utilizing a 14C, 3H dual-label program. The scintillation counter provides automatic conversion of sample counts per minute to disintegrations per minute utilizing the sample channels ratio technique of quench correction (22). Additional quenching in samples obtained from the TLC plates was assessed by counting samples before and after addition of a 14C- or 3H-labeled internal standard as described by Fox (23). This approach in combination with the quench correction provided by the instrument resulted in greater than 95% overall counting efficiency for both radioisotopes in samples containing silica that were obtained from the TLC plate and treated as described above. Calculation of Assay Results After determining the 14C and 3H disintegrations per minute in the samples and standards, the observed 3H disintegrations per minute were adjusted according to the equation adjusted 3H dpm = (observed 3H dpm - 3H dpm in corresponding reagent blank) (observed 14C dpm/14C dpm added to each sample) *
SENSITIVE,
RADIOMETRIC
ASSAY
A standard curve was constructed by least-squares linear regression analysis of the adjusted 3H disintegration per minute results for the standards (y data) and the known amount of 17:O LPC added to each standard (X data). The LPC content of the samples was then determined from the equation derived by the linear regression analysis using the adjusted 3H disintegration per minute results for the samples. RESULTS
Optimization Separation
of Reaction Conditions of Reaction Products
ana’ TLC
In initial experiments, [‘4C]LPC was utilized to determine appropriate reaction conditions for quantitative acetylation of LPC. To prevent hydrolysis of acetic anhydride, aprotic solvents were used. On the basis of previous studies of the acetylation of 1-0-alkyl-2-lysophosphatidylcholine (lyso-platelet activating factor), pyridine was initially used as a catalyst (24). In these experiments, samples were incubated in 25 ~1 of pyridine in the presence of lo- to loo-fold molar excess of unlabeled acetic anhydride (100 nmol//*l in acetonitrile) for selected intervals of time up to 18 h at room temperature and at 37°C. The samples were applied to the TLC plate with LPC and platelet activating factor used as markers to identify regions corresponding to LPC and acetylLPC, respectively. Regions corresponding to LPC, acetyl-LPC, and fatty acid (solvent front) were scraped and counted to determine the extent of the reaction and also to determine whether hydrolysis of LPC occurred during the course of the incubation period. The results of these initial studies demonstrated that the yield of acetyl-LPC product increased when a large (greater than 50-fold) molar excess of acetic anhydride was used and when the incubation time was extended to 18 h. The extent of product formation was also increased for those incubations at 37°C compared to incubations performed at room temperature. Incubation at 37°C for 18 h resulted in a near quantitative (i.e., greater than 95%) decrease in the radioactivity in the LPC region. However, only 60% of the radioactivity in the sample was recovered in the region corresponding to the acetyl-LPC product. The remaining radioactivity migrated to the solvent front (corresponding to fatty acid) indicating that under these conditions approximately 35% of the LPC was hydrolyzed to fatty acid and glycerophosphorylcholine. Incubation of LPC in acetonitrile with pyridine in the absence of acetic anhydride resulted in less than 2% hydrolysis of LPC. Similarly, incubation of LPC with acetic anhydride in the absence of pyridine resulted in less than 5% conversion of LPC to acetyl-LPC. Thus, an acylation catalyst (e.g., pyridine) is needed for the reaction to proceed to completion; however, in the presence of pyridine and acetic anhydride approximately 35% of
FOR
LYSOPHOSPHATIDYLCHOLINE
39
the LPC is hydrolyzed. One possible explanation for these observations is that in the presence of pyridine and acetic anhydride, the acetylation of LPC results in the production of a stoichiometrically equivalent amount of acetic acid and the accumulation of free acetic acid promotes further hydrolysis of LPC. Accordingly, dimethylaminopyridine (DMAP) was selected as an alternative catalyst for the acetylation reaction. Since DMAP contains a basic, tertiary amine group, this catalyst may also serve to “buffer” the production of acid produced during the course of the acetylation reaction and thereby prevent hydrolysis of LPC. Using DMAP as the acylation catalyst, there was no evidence of hydrolysis of LPC. Although the extent of the reaction increased with time with incubations performed at room temperature, the reaction did not reach completion within 24 h. In contrast, samples incubated for 12 to 15 h at 37°C in the presence of a 200-fold molar excess of acetic anhydride resulted in a greater than 95% conversion of LPC to acetyl-LPC. Therefore, for all subsequent analyses of LPC, DMAP was employed as the acylation catalyst, a 200-fold or greater molar excess of acetic anhydride was used, and all incubations were performed at 37°C for at least 12 h. After the appropriate conditions for avoiding hydrolysis while achieving quantitative acetylation of LPC using [14C]LPC were established, another series of experiments were performed using [3H]acetic anhydride and unlabeled LPC to permit analysis of LPC in extracts of isolated ventricular cells or whole tissue. Since the [3H] acetic anhydride from commercial sources is supplied in toluene, this solvent was used to prepare both labeled and unlabeled acetic anhydride solutions. The substitution of toluene for acetonitrile had no effect on the rate or extent of completion of the acetylation reaction. The same protocol was used to quantify the formation of [3H]acetyl-LPC from unlabeled LPC. After TLC separation and liquid scintillation counting of the [3H]acetylLPC product, the observed disintegrations per minute far exceeded those expected for a complete reaction in all samples. Examination of all regions of the TLC plate in lanes which contained only [3H]acetic anhydride showed 3H radioactivity in all regions. Most of the radioactivity was found at the solvent front but significant radioactivity was also found in the area corresponding to the acetyl-LPC region. Thus, in this solvent system C3H]acetic anhydride and/or [3H]acetic acid does not migrate as a discrete zone. Activation of the TLC plates by heating had no appreciable effect on the distribution of 3H radioactivity. Several alternative solvent systems for TLC were examined but this also did not improve the resolution of [3H]acetic acid. Selective extraction of [3H]acetyl-LPC from [3H]acetic anhydride and [3H]acetic acid using a variety of different binary and ternary solvent systems was also unsuccessful. To overcome this problem, the TLC plates were exposed to two successive
40
DOBMEYER,
CORR,
AND
CREER
developments in the same solvent system (system I). The TLC plates were removed from the tank and allowed to air-dry for 20 to 30 min before proceeding with the next solvent development. With this approach, it was possible to reduce the “background” 3H radioactivity in the [3H]acetyl-LPC region to consistently less than 0.1% of the total radioactivity applied to the plate. Under these conditions, with samples containing only 1 nmol of LPC and 100 nmol of [3H]acetic anhydride, the [3H]acetyl-LPC product would account for greater than 90% of the total 3H radioactivity in the acetyl-LPC region of the TLC plate. For all other samples, the background 3H radioactivity should be less than 10%. Thus, for all assays, a separate reagent blank was prepared for each TLC plate to allow correction for background 3H radioactivity and the reaction products formed following incubation of LPC with C3H]acetic anhydride were separated by two successive developments of the TLC plate in solvent system I. A representative example of the TLC separation obtained using this approach is illustrated in Fig. 1.
Evaluation
of the Assay Using [3H]Acetic
--)
3
solvent front
A
Anhydride
To determine the linearity of the assay, duplicate samples containing 0.5 to 15.0 nmol of LPC were prepared and analyzed. The results, shown in Fig. 2, demonstrate the linearity and reproducibility of the assay. The sensitivity of the assay was sufficient to permit accurate quantification of LPC in samples containing only 500 pmol of LPC. The specific activity of the [3H]acetylLPC product was approximately 60% of the value predicted based on the calculated specific activity of the [3H]acetic anhydride. This discrepancy may be due to hydrolysis of [3H]acetic anhydride or it may reflect the fact that the specific activity of the r3H]acetic anhydride was less than the value specified by the manufacturer. The specific activity of the [3H]acetyl-LPC product did not change appreciably when assays were performed 3 to 4 weeks after the [3H]acetic anhydride reagent was prepared and stored at -4°C. The results of quantification of choline lysoplasmalogen using this radiometric assay procedure are shown in Fig. 3. These results correlate closely with those observed for measurements of monoacyl-LPC and demonstrate that the assay is able to accurately quantify both monoacyl and choline lysoplasmalogen molecular species. Choline lysoplasmalogen and monoacyl LPC both contain one free hydroxyl group per molecule that is acetylated to produce labeled products of the same specific activity. Since choline lysoplasmalogen coelutes with monoacyl LPC during the course of HPLC separation and the acetylated products of choline lysoplasmalogen corn&ate with acetyl-LPC during TLC separation, the radiometric assay method does not distinguish these two choline lysophospholipid subclasses. Accord-
+ 1
2
origin
3
FIG. 1. TLC separation on silica gel G plates of 50 nmol each of soybean LPC (lane l), acetylated soybean LPC (lane 2), and I-O-alkyl-2-acetyl-sn-glycero-3-phosphocholine (platelet activating factor) (lane 3), following two successive developments in a solvent system comprising chloroform/acetone/methanol/acetic acid/water (6/g/2/ 2/l, v/v). Following TLC separation, the plate was stained by exposure to I2 vapor. The faint staining region shown by the arrow represents dimethylaminopyridine present in the reaction mixture used to prepare the acetylated soybean LPC product as described under Materials and Methods. The region indicated by the letter A to the right of the figure corresponds to the region that is scraped for liquid scintillation counting of 3H and “C dpm in samples following reaction with 13H]acetic anhydride.
ingly, in samples containing both subclasses, the result obtained by the radiometric assay procedure represents the sum of the monoacyl LPC and choline lysoplasmalogen content of the sample. A methods comparison study was performed to compare the results obtained by our method using [3H]acetic anhydride with those results obtained based on measurements of lipid phosphorus. Duplicate samples containing 6 to 30 nmol of LPC were prepared from the same LPC stock solutions. In one sample, the LPC content was determined by assay of lipid phosphorus and, in the other sample, the LPC content was determined by radiometric assay using [3H]acetic anhydride. For this study, relatively large amounts of LPC were required for each sample to permit accurate measurements of lipid
SENSITIVE,
0.0
3.2
RADIOMETRIC
12.8
6.4 9.6 nmol LPC
ASSAY
FOR
41
LYSOPHOSPHATIDYLCHOLINE
0.01 ’ ’ ’ ’ ’ ’ ’ ’ ’ 1 0.0 6.4 12.8 19.2 25.6 32.0 nmol LPC (radiometric)
16.0
FIG. 2. Representative radiometric assay results for the analysis of monoacyl LPC. Samples containing 0.5 to 15.0 nmol of l-heptadecanoyl-sn-glycero-3-phosphocholine (17:O LPC) were prepared and incubated with [3H]acetic anhydride, [‘%]LPC, and dimethylaminopyridine for 12 h at 37°C. The reaction was terminated, the reaction products were separated by TLC, and adjusted 3H dpm were determined as described under Materials and Methods. The correlation coefficient (r) shown at the bottom right portion of the figure was determined by least-squares linear regression analysis (n = 10 samples).
FIG. 4.
phosphorus. As shown in Fig. 4, there is excellent agreement between the values obtained for LPC by the two independent methods. The results of measurements of lipid phosphorus were slightly lower than those ob-
tained by the radiometric assay. However, the radiometric assay results agreed more closely with the values expected based on the known concentration of LPC in the stock solution.
Influence of Cardiolipin, on LPC Quantification
4.03.22.4 I .6 0.8 0
I 0.0
FIG. 3.
Methods comparison study. Paired samples containing 6.0 to 30.0 nmol of 1-heptadecanoyl-sn-glycero-3-phosphocholine (17:O LPC) were prepared. For each pair of samples, the LPC content was determined for one sample by microphosphate assay and by radiometric assay for the other sample as described under Materials and Methods. Shown at the right portion of the figure are the results of least-squares linear regression analysis of the data obtained from the sample pairs (n = 10 paired samples).
I
I
I
I
I
IIll
1.2 2.4 3.6 4.8 nmol LPC PLASMALOGEN
6.0
Representative radiometric assay results for the analysis of choline lysoplasmalogen (LPC plasmalogen). Samples containing 0.8 to 6.0 nmol of LPC plasmalogen were prepared and incubated with [3H]acetic anhydride, [“CILPC, and dimethylaminopyridine for 12 h at 37°C. The reaction was terminated, the reaction products were separated by TLC, and adjusted 3H dpm were determined as described under Materials and Methods. The correlation coefficient (r) shown at the bottom right portion of the figure was determined by least-squares linear regression analysis (n = 5 samples). The specific radioactivity of the [3H]acetic anhydride used for measurement of LPC plasmalogen was approximately fourfold greater than that of [3H]acetic anhydride used for the assay of monoacyl LPC shown in Fig. 2.
Sphingomyelin, by Radiometric
and Fatty Acid Assay
Since cardiolipin, sphingomyelin, and fatty acid frequently contaminate the LPC fraction isolated by conventional HPLC methods (vide supra), additional experiments were performed to determine the effects of these lipids on the measurement of LPC by radiometric assay. Samples containing 1.0 to 5.0 nmol of LPC were prepared and 25 nmol of cardiolipin, spingomyelin, or fatty acid was added to each sample. The quantity of cardiolipin, sphingomyelin, or fatty acid added was far in excess of that expected in the LPC fraction following HPLC isolation of LPC in extracts from whole tissue or isolated cardiac myocytes. Comparison of the results of LPC analysis obtained in the absence and presence of cardiolipin, sphingomyelin, or fatty acid demonstrated that the presence of these lipids did not significantly affect the results of the measurement of LPC using the radiometric assay procedure. Cardiolipin and sphingomyelin both contain a free hydroxyl group that could be acetylated under the conditions employed. However, the lack of interference of these lipids on the measurement of LPC indicates that the acetylated products formed from sphingomyelin or cardiolipin do not comigrate with acetyl-LPC under the conditions used for TLC separation of reaction products. If present in very large
42
DOBMEYER,
CORR,
amounts (i.e., greater than 50 to 100 nmol/sample), both cardiolipin and sphingomyelin could interfere with the quantification of LPC as a result of a reduction in the large molar excess of [3H]acetic anhydride available for acetylation of LPC. However, this is unlikely to occur in extracts of biological samples that are subjected to HPLC for isolation of the LPC fraction.
LPC Measurements
in Isolated Cardiac Myocytes
Because the TLC separation step could potentially allow the isolation of acetyl-LPC from other acetylated products in crude phospholipid mixtures obtained following extraction of cells or tissue, experiments were performed to determine whether isolation of the LPC fraction by HPLC was necessary since elimination of this step would considerably reduce the time required to perform LPC measurements. The results obtained in crude extracts were 2- to lo-fold greater than the values obtained for LPC measurement after isolation of the LPC fraction by HPLC. When trace amounts of [‘“ClLPC were added to the extracts before measurement of the LPC using [3H]acetic anhydride and both 14C and 3H radioactivities determined in the acetyl-LPC region of the TLC plate at the conclusion of the assay, the extent of conversion of [14C]LPC to [3H,‘4C]acetyl-LPC varied from 35 to 85%. Thus, although far less than quantitative conversion of LPC to acetyl-LPC was achieved in the extracts of tissue or cells analyzed directly, the apparent amount of LPC was 2- to lo-fold greater than expected. These results indicate that the overwhelming majority of 3H radioactivity detected in the acetyl-LPC region following TLC separation is due to the presence of acetylated products derived from other chloroformmethanol extractable substances in tissue or isolated cells which corn&ate with acetyl-LPC under the chromatographic conditions employed. These substances could be phospholipids, neutral lipids, or other chloroform-methanol extractable compounds containing a nucleophilic functional group capable of reacting with acetic anhydride. Since the presence of these substances resulted in a marked, positive interference in the direct measurement of LPC in extracts from tissue or isolated cells, it is apparent that isolation of the LPC fraction by HPLC is required prior to the reaction with [3H]acetic anhydride. Because of the potential loss of LPC which can occur during homogenization of the cells, extraction with chloroform-methanol, HPLC separation, and TLC separation of the reaction products produced after reaction with [3H]acetic anhydride, additional recovery studies were performed. For these experiments, carrier-free [14C]LPC was added to each sample of cells prior to homogenization. The recovery of 14C radioactivity was then determined at the conclusion of each step during the course of analysis. As shown in Table 1, excellent
AND
CREER TABLE
1
Recovery of LPC during Analysis by the Radiometric Assay Procedure Analytical
step
Bligh and Dyer extraction HPLC separation Acetylation and TLC separation
Recovery
(X)
96 f 1.5 95 f 2.3 86 f 9.0
Note. For this study, 1-[“C]pa1mitoy1-sn-g1ycero-3-phosphocho1ine ([“C]LPC) was added to isolated canine cardiac myocytes (equivalent to approximately 5 mg total cellular protein) and extracted, the LPC fraction was isolated by HPLC, and the acetylation reaction and TLC separation were performed as described under Materials and Methods. Aliquots were taken after extraction and HPLC separation to determine recovery of “C dpm at the conclusion of each of these analytical steps. Recovery following acetylation and TLC separation was determined as the percentage of total 14C dpm in the LPC fraction after HPLC separation that was recovered in the acetyl-LPC region of the TLC plate after acetylation and TLC separation. The recovery value shown therefore reflects both the extent of completion of the acetylation reaction and the recovery of “C dpm from the TLC plate. Values shown represent the means f SE of three experiments.
recovery of LPC during each step of the analytical procedure was obtained. However, recovery was not quantitative and even a small loss of LPC at each step would have a cumulative effect on the final result. In addition, there was significant variability in recovery after acetylation and TLC separation of the acetyl-LPC product. Accordingly, carrier-free [14C]LPC (17,000 dpm, 200 pmol) was added to the cell suspension before extraction and to the LPC standards before performing the acetylation reaction. The 14C radioactivity measured in the acetyl-LPC region of the TLC plate at the conclusion of the analytical procedure was then used to correct for the small but variable loss of LPC during the course of analysis. Since each sample of cells contains the equivalent of 4-10 mg of total cellular protein or approximately 2-5 nmol of total LPC, the mass of carrier-free [14C]LPC (200 pmol) added for determination of recovery would not significantly affect the total mass of LPC present in the sample. For analysis of samples containing less than 1 nmol of total LPC, it would, of course, be necessary to reduce the amount of carrier-free [14C]LPC added for correction of recovery. The results of analysis of LPC in isolated, adult canine cardiac myocytes using the present radiometric method were compared to results obtained by GC quantification of LPC. The GC method for measurement of LPC is based on the quantification of the fatty acid methyl esters and dimethylacetals produced by acid-catalyzed methanolysis of LPC using synthetic l-heptadecanoyl-an-glycero-3-phosphocholine (17:0 LPC) as an internal standard as previously described (6). The GC measurements of LPC required relatively large amounts of isolated cells (equivalent to 15-25 mg of cellular pro-
SENSITIVE,
RADIOMETRIC
ASSAY
tein) compared to that required for analysis of LPC by the radiometric assay using [3H]acetic anhydride (2-5 mg of cellular protein). The results obtained by the radiametric assay procedure (0.38 f 0.04 nmol/mg protein (mean f SE), n = 16 samples) compared very favorably to those results obtained by GC analysis of LPC (0.39 f 0.08 nmol/mg protein (mean f SE), n = 15 samples). Thus, in isolated adult canine cardiac myocytes, the results of the present method are in close agreement with those results obtained by other methods employed for the analysis of LPC. DISCUSSION
The radiometric method developed for the quantification of LPC in tissue samples or in isolated cell preparations offers several advantages compared to other conventional assays. For example, the analysis of LPC based on measurements of total lipid phosphorus or by GC quantification of the fatty acid methyl esters or dimethylacetals produced by acid-catalyzed methanolysis of LPC requires at least 5-10 nmol of LPC per sample. The present method is much more sensitive permitting analysis of LPC in samples containing less than 1 nmol of LPC. Further enhancement of sensitivity could be achieved by increasing the specific activity of the [3H]acetic anhydride reagent. The radiometric assay procedure is also a cost-effective approach for the analysis of LPC in biological samples. At the present time, commercial sources of [3H]acetic anhydride provide the reagent at a relatively low cost of less than 50 cents per assay sample. Our approach to the analysis of LPC is also not subject to interference by fatty acids, sphingomyelin, or cardiolipin which frequently contaminates the LPC fraction isolated by HPLC methods which employ acetonitrile, methanol, and water in the mobile phase (vide supra). Thus, removal of fatty acids and cardiolipin by anionexchange chromatography of the LPC fraction obtained following HPLC separation (6) is not required prior to the analysis of LPC. The present method is able to accurately quantify choline lysoplasmalogen which coelutes with monoacyl LPC during the course of HPLC isolation by those methods which separate phospholipids into classes based on differences in polar headgroup composition. Other derivatization procedures based on CC separation and quantitation of the fatty acid methyl ester derivatives which employ basic conditions for methanolysis preclude detection of lysoplasmalogens (14,15). Thus, the ability of the present method to detect choline lysoplasmalogen molecular species is a distinct advantage for the analysis of choline lysophospholipids in myocardial tissue samples or in isolated cardiac myocyte preparations where plasmalogens account for a significant fraction of the to-
FOR
43
LYSOPHOSPHATIDYLCHOLINE
tal choline phospholipid pool. The radiometric assay approach for the quantification of LPC developed in this study should be readily adaptable for analytical studies of LPC metabolism in other tissues and also in isolated cell preparations derived from a variety of different sources. REFERENCES 1. Corr, P. B., Gross, 135-154.
R. W., and Sobel,
B. E. (1984)
Res. 66,
Circ.
2. Corr, P. B., Snyder, D. W., Cain, M. E., Crafford, W. A., Jr., Gross, R. W., and Sobel, B. E. (1981) Circ. Res. 49,354-363. 3. Snyder, D. W., Crafford, W. A., Jr., Glashow, Sobel, B. E., and Corr, P. B. (1981) Amer. J. H707. 4. Arnsdorf, 5. Clarkson, 556.
M. F., and Sawicki, C. W., and Ten
G. J. (1981) Eick,
J. L., Rankin,
Physid. 241,
Circ.
R. E. (1983)
D., H700-
Res. 49,16-30. Res. 52, 543-
Circ.
6. Corr, P. B., Creer, M. H., Yamada, K. A., Saffitz, J. E., and Sobel, B. E. (1989) J. Clin. Invest. 83,927-936. 7. Creer, M. H., Pastor, C., Corr, P. B., Gross, R. W., and Sobel, B. E. (1985) Anal. B&hem. 144,65-74. W. W. (1973) 8. Christie, ford, NY.
Lipid
Analysis,
p. 192, Pergamon,
Elms-
9. Kates, M. (1975) Techniques of Lipidology, p. 552, North Holland/Elsevier, Amsterdam/New York. 10. White, D. A. (1973) in Form and Function of Phospholipids (Ansell, G. B., Hawthrone, J. N., and Dawson, R. M. C., Eds.), p. 448, Elsevier, Amsterdam/New York. 11. Horrocks, L. A., and Sharma, M. (1982) in Phospholipids (Ansell, G. B., and Hawthorne, J. N., Eds.), p. 51, Elsevier, Amsterdam/ New York. 12. Mogelson,
S., Wilson,
G. E., and Sobel,
Biochim. Bio-
B. E. (1980)
phys. Acta 619,680-688. 13. Bjerve,
K. S., Daae,
L. N. W., and Bremer,
J. (1974)
Anal. Bio-
&em. 58,238-245. 14. Shaikh,
N. A., and Downar,
E. (1981)
15. Kates, M. (1975) Techniques land/Elsevier, Amsterdam/New
Circ. Res. 49,316-325.
of Lipidology, York.
p. 361, North
Hol-
16. Creer, M. H., and Gross, R. W. (1985) J. Chromatogr. 338,61-69. 17. Markwell, M. A. K., Haas, S. H., Tolbert, N. E., and Bieber, L. L. (1981) in Methods in Enzymology (Lowenstein, J. M., Ed.), Vol. 72, p. 296, Academic Press, New York. 18. Chen, P. S., Toribara, 28,1756-1758.
T. Y., and Warner,
19. Heathers, G. P., Yamada, K. A., Kanter, (1987) Circ. Res. 61,735-746. 20. Bligh,
E. G., andDyer,
W. J. (1959)
H. (1956) E. M.,
And. Chem.
and Corr,
P. B.
J. Biochem. Physiol. 37,911-
917. 21. Gross,
R. W., and Sobel,
B. E. (1980)
J. Chromatogr. 197,79-85.
22. Fox, B. W. (1976) Techniques Scintillation Counting, pp. Amsterdam/New York.
of Sample Preparation for Liquid 212-220, North Holland/Elsevier,
23. Fox, B. W. (1976) Techniques Scintillation Counting, pp. Amsterdam/New York.
of Sample Preparation for Liquid 207-211, North Holland/Elsevier,
24. Polonsky, Benveniste,
J., Tence, J. (1980)
M.,
Varenne,
P., Das,
B. C., Lunel,
J., and
Proc. Natl. Acad. Sci. USA 77,7019-7023.
ANALYTICALBIOCHEMISTRY
(1990)
A Sensitive, Radiometric Assay for Lysophosphatidylcholine’ David J. Dobmeyer, Peter B. Corr, and Michael H. Creer Cardiovascular Division, Department of Internal Medicine and Department of Pharmacology, Washington University School of Medicine, St. Louis, Missouri 63110
Received
September
11,1989
tion, prevention of the accumulation of LPC during myocardial ischemia in viuo is associated with a pro1found antiarrhythmic effect (6). We have developed a model system utilizing isolated adult canine cardiac myocytes exposed to hypoxia and acidosis in vitro to further investigate the biochemical mechanisms responsible for the accumulation of LPC during myocardial ischemia. To accurately quantify the small amounts of LPC in the isolated cell preparations, an improved method for measurement of LPC which would be capable of detecting subnanomolar quantities in extracts of isolated cells was required. The analysis of LPC in isolated cardiac cells or in myocardial tissue is complicated by several factors including (i) LPC accounts for only 2 to 3 mol% of the total phospholipid composition of the myocardium (7) and therefore is present in very small quantities. (ii) Conventional methods for isolation of LPC by one-dimensional TLC do not completely resolve LPC from all other phospholipid classespresent in tissue (8,9). (iii) TLC separation in two dimensions, although providing the resoluPress. Inc. tion necessary to isolate LPC, requires at least 10 to 20 nmol of LPC to identify the LPC region on the TLC plate. (iv) HPLC methods for isolation of LPC freAlterations in the phospholipid composition of the quently employ the use of mobile phases which are comIn this mobile myocardial cell membrane, or sarcolemma, have been prised of acetonitrile/methanol/water. phase, cardiolipin (diphosphatidylglycerol), which acimplicated in the pathogenesis of ventricular arrhythcounts for 10 to 15% of the total phospholipid composimias during myocardial ischemia (1). We, and others, tion of the myocardium (lo), is not completely soluble have demonstrated that lysophosphatidylcholine (LPC), and elutes as a broad band resulting in 30 to 40 mol% an amphiphilic phospholipid derived from phosphatidylcontamination of the LPC fraction (7). (v) Several curcholine, accumulates in ischemic myocardium and has rently available HPLC methods do not completely seppotent arrhythmogenic properties in vitro (2-5). In addiarate sphingomyelin and LPC. (vi) In several animal species, plasmalogens (1-0-alk-Y-enyl-2-acyl phospho1 Research from the authors’ laboratory was supported in part by lipids) account for 30 to 40% of the total content of diraNational Institutes of Health Grant HL 17646, SCOR in Ischemic dyl choline phospholipids in myocardium (10,ll). ChoHeart Disease and NIH Grants HL 28995 and HL 36773. line plasmalogens are readily hydrolyzed in acidified 2 Abbreviations used: LPC, lysophosphatidylcholine; FAME, fatty extraction media leading to artifactual production of 2acid methyl ester; DMAP, dimethylaminopyridine; [“CILPC, l-(1acyl-lysophosphatidylcholine (12). This may also ac[%]palmitoyl)-2-lysophosphatidylcholine.
To facilitate investigation of the metabolism of lysophosphatidylcholine and choline lysoplasmalogen in small quantities of tissue, a method for the quantiflcation of these phospholipid species that is capable of accurate and reproducible analysis in samples which contain less than 1 nmol of total choline lysophospholipid was developed. The procedure employs chloroform and methanol extraction of phospholipids from isolated tissue with subsequent separation of the choline lysophospholipid fraction by high-performance liquid chromatography. The choline lysophospholipids are then acetylated with [3H]acetic anhydride and the [‘HIacetyl-lysophosphatidylcholine product is isolated by thinlayer chromatography and quantified by liquid scintillation counting. The choline lysophospholipid content in the sample is determined from a standard curve constructed from samples containing a known amount of synthetic lysophosphatidylcholine with correction for recovery based on the inclusion of [14C]lysophosphatidylcholine as an internal standard. o 1990 Academic
36 All
Copyright 0 1990 rights of reproduction
0003-2697/90 $3.00 by Academic Press, Inc. in any form reserved.
SENSITIVE,
RADIOMETRIC
ASSAY
count, in part, for the apparent increase in recovery of LPC previously reported with methods requiring the addition of acid to the extraction media (13). (vii) Due to the high content of choline plasmalogens in myocardial tissue, the LPC fraction isolated from extracts of myocardial tissue or from isolated cardiac myocytes may also contain a significant amount of choline lysoplasmalogen. The quantification of LPC by GC analysis of the fatty acid methyl esters (FAME) produced after basecatalyzed hydrolysis of LPC in methanol (14) does not permit detection of choline lysoplasmalogens since they are not hydrolyzed under these conditions (15). In our experience, even when acidic conditions are employed for hydrolysis of LPC fractions isolated by HPLC, analysis of the FAME derivatives (produced from monoacyl LPC) and dimethylacetal derivatives (produced from choline lysoplasmalogen) by GC is limited by two factors. First, GC methods require at least 2 to 5 nmol of each individual LPC species for accurate quantitation. Second, trace amounts of fatty acid contamination often cause significant interference. This interference is not easily eliminated despite intense efforts to remove the contaminating fatty acids by acid washing, silanizing all glassware, and using solvents from several different manufacturers. To overcome these limitations of conventional approaches to the quantification of LPC, we developed a method based on the acetylation of the free hydroxyl group of LPC with radiolabeled acetic anhydride to permit the accurate quantification of LPC in samples containing less than 1 nmol of total LPC. This method is not subject to interference by fatty acids, sphingomyelin, or cardiolipin which frequently contaminate the LPC fraction isolated by HPLC procedures. The method is also applicable to the measurement of choline lysoplasmalogens. MATERIALS
AND
METHODS
Soybean LPC, synthetic l-heptadecanoyl-sn-glycero3-phosphocholine (17:0 LPC), 1-0-alkyl-2-acetyl-snglycero-3-phosphocholine (platelet activating factor), cardiolipin, dimethylaminopyridine, sphingomyelin, KHzPOl, and acetic anhydride were obtained from Sigma Chemical Co. (St. Louis, MO). Phospholipids obtained from Sigma were chromatographically pure and were used without further purification except for the 17: 0 LPC which was purified by HPLC prior to use as described below. Choline lysoplasmalogen was prepared by base-catalyzed hydrolysis of bovine heart phosphatidylcholine (Avanti Polar Lipids, Birmingham, AL) as previously described (16), and 1-(1-[14C]palmitoyl)-2-lysophosphatidylcholine ( [14C]LPC) was obtained from Amersham Corp. (Arlington Heights, IL). [3H]Acetic anhydride was purchased from New England Nuclear (Boston, MA). Liquid scintillation cocktail was pur-
FOR
37
LYSOPHOSPHATIDYLCHOLINE
chased from Research Products International (Mt. Prospect, IL). Thin-layer chromatography plates, glacial acetic acid, and chloroform were obtained from Fisher Scientific (St. Louis, MO). Bovine serum albumin was purchased from Bio-Rad (Richmond, CA). Acetonitrile, methanol, water, and toluene were purchased from Taylor Chemical Co. (St. Louis, MO). Partisil-SCX HPLC columns (Whatman) were purchased from PJ Corbert Associates, Inc. (St. Louis, MO).
Protein
Assay
Cell samples were analyzed for total protein by a modification of the Lowry method as described by Markwell et al. (17). Lyophilized bovine serum albumin served as the protein standard.
Microsphosphate
Assay
Total lipid phosphorous was determined in selected samples and standards by microphosphate assay after ashing of the samples in 1.2 M Mg(N03)2 in ethanol (7). The phosphate content of the ashed material was determined by the microphosphate assay method of Chen et al. (18). Standards containing 2 to 50 nmol of phosphate were prepared from anhydrous reagent grade KH,PO,.
Extraction
of Myocyte Phospholipids
Calcium tolerant isolated adult canine myocytes were harvested from canine epicardium by collagenase dispersion as previously described (19). Isolated cardiac myocytes (approximately 2-5 X lo5 total cells, equivalent to 2-5 mg total cellular protein) were suspended in 2 ml Hepes buffer (115 mM NaCl, 5 mM KCl, 35 mM sucrose, 10 mM glucose, 10 mM Hepes, and 4 mM taurine, pH 7.4). A 250-~1 aliquot was removed for protein analysis. The remaining cell suspension was then subjected to chloroform and methanol phospholipid extraction according to the method of Bligh and Dyer (20). [ 14C]LPC (17,000 dpm) was added to the cells before extraction as an internal standard to correct for loss of LPC during the course of analysis. Following extraction, the lower organic layer was taken to dryness under N, at 37°C. The residue was resuspended in 3 ml of chloroform/methanol (l/l, v/v), filtered through 0.2-pm filters (Acrodisc), and stored at -4°C in vials sealed with Teflon-lined silicon rubber septum caps.
High-Performance
Liquid
Chromatography
The phospholipid extracts were taken to dryness under N2 and resuspended in 100 ~1 of l/l chloroform/ methanol (v/v). The entire volume was then injected onto a silica-based cation-exchange HPLC column (10 pm Partisil-SCX). Phospholipids were eluted with a mobile phase comprised of acetonitrile, methanol, and wa-
38
DOBMEYER,
CORR, AND CREER
ter (400/100/32, v/v) as previously described (21). Collection of myocyte LPC fractions was based on the retention time of unsaturated molecular species of LPC determined following a separate injection of soybean LPC standard by monitoring absorbance in the ultraviolet range at 203 nm. The LPC fraction was collected in 18 X 150-mm test tubes, taken to dryness under Nz at 37”C, resuspended in 3 ml l/l chloroform/methanol, and stored at -4°C in 3-ml conical vials sealed with Teflon-lined silicon rubber septum caps. 3H Acetylution Reaction Following separation of phospholipid fractions by HPLC, the purified LPC fraction was taken to dryness under N2 at 37°C. The residue was resuspended in 25 ~1 of 0.33 M dimethylaminopyridine in acetonitrile and 25 ~1 of [3H]acetic anhydride in toluene. The [3H]acetic anhydride reagent was prepared by diluting [3H]acetic anhydride with unlabeled acetic anhydride to a specific activity of 50-70 dpm/pmol and adjusting the final concentration to 100 nmol/pl in toluene. The reaction was allowed to continue for 12 h at 37°C. At the end of this period, the acetonitrile-toluene solution was evaporated under N2 at 37°C. One hundred microliters of water was then added to each vial and incubated at 37°C for 15 min, after which the water was evaporated under Nz at 37°C. The remaining residue was then resuspended in 25 ~1 l/l chloroform/methanol in preparation for thin-layer chromatography. A standard curve was prepared by adding HPLC-purified 17:0 LPC in amounts ranging from 0.5 to 15.0 nmol to separate vials followed by addition of dimethylaminopyridine and [3H]acetic anhydride as described above for individual samples. [ 14C]LPC (17,000 dpm) was also added to each standard curve vial before acetylation to correct for incomplete recovery of the acetyl-LPC product following TLC separation. Reagent blanks were prepared by addition of all reagents except LPC to separate vials. Three blanks were prepared for each assay (one for each TLC plate). In two separate vials, 0.5 pmol of soybean LPC was added to each vial and the solvent removed by evaporation under N2 . To each of these vials, 10 ~1 of unlabeled acetic anhydride (neat), 30 ~1 of toluene, and 25 ~1 of 0.33 M dimethylaminopyridine were added, the vials were incubated for 12 h at 37”C, and the reaction was terminated as described above. The residue in each vial was then resuspended in 50 ~1 of l/l chloroform/methanol. Five microliters of this solution (containing 50 nmol of unlabeled acetylated soybean LPC) was then added to each sample (17:0 LPC standards and LPC samples derived from cell extracts) and to the reagent blanks immediately prior to the application to the TLC plates. The unlabeled acetylated soybean LPC added to the samples served as an internal marker for identification of the region of the TLC plate containing the 3H acetylated LPC product in these samples following Iz staining as described below.
Thin-Layer Chromatography Thin-layer chromatography was performed on silica gel G plates, 20 X 20 cm, 250 pm thick using a solvent system composed of chloroform/acetone/methanol/glacial acetic acid/water (6/8/2/2/l, v/v) (system I). Samples were applied as 2-cm-wide “streaks” in adjacent lanes of the TLC plate (eight samples per plate). On each plate, 10 ~1 of soybean LPC (approx 100 nmole) was applied to lane 1, a reagent blank was applied to lane 2, and samples or standards were applied to lanes 3 through 8. TLC separation was accomplished by two successive developments in solvent (system I). At the conclusion of the TLC separation, the plates were allowed to air-dry in a fume hood for 1 h. The region corresponding to acetyl-LPC in lanes 2 through 8 was then identified by passing air through a Pasteur pipet containing Iz crystals that was directed over the TLC plate. The acetyl-LPC region was outlined with a pencil and the TLC plate was allowed to stand for 1 h in a fume hood before scraping the plate for liquid scintillation counting. This was done to allow the Iz to evaporate from the surface of the TLC plate to avoid any potential quenching effects of 12. Liquid Scintillation Counting The TLC plate regions from the reagent blanks and from samples containing the [3H]acetyl-LPC product were scraped into liquid scintillation vials. Five hundred microliters of water and 10 ml of liquid scintillation cocktail were added to each vial. The vials were vortexed thoroughly and remained in the dark overnight to minimize autofluorescence. Liquid scintillation counting was performed on a Beckman Model LS-3801 liquid scintillation counter utilizing a 14C, 3H dual-label program. The scintillation counter provides automatic conversion of sample counts per minute to disintegrations per minute utilizing the sample channels ratio technique of quench correction (22). Additional quenching in samples obtained from the TLC plates was assessed by counting samples before and after addition of a 14C- or 3H-labeled internal standard as described by Fox (23). This approach in combination with the quench correction provided by the instrument resulted in greater than 95% overall counting efficiency for both radioisotopes in samples containing silica that were obtained from the TLC plate and treated as described above. Calculation of Assay Results After determining the 14C and 3H disintegrations per minute in the samples and standards, the observed 3H disintegrations per minute were adjusted according to the equation adjusted 3H dpm = (observed 3H dpm - 3H dpm in corresponding reagent blank) (observed 14C dpm/14C dpm added to each sample) *
SENSITIVE,
RADIOMETRIC
ASSAY
A standard curve was constructed by least-squares linear regression analysis of the adjusted 3H disintegration per minute results for the standards (y data) and the known amount of 17:O LPC added to each standard (X data). The LPC content of the samples was then determined from the equation derived by the linear regression analysis using the adjusted 3H disintegration per minute results for the samples. RESULTS
Optimization Separation
of Reaction Conditions of Reaction Products
ana’ TLC
In initial experiments, [‘4C]LPC was utilized to determine appropriate reaction conditions for quantitative acetylation of LPC. To prevent hydrolysis of acetic anhydride, aprotic solvents were used. On the basis of previous studies of the acetylation of 1-0-alkyl-2-lysophosphatidylcholine (lyso-platelet activating factor), pyridine was initially used as a catalyst (24). In these experiments, samples were incubated in 25 ~1 of pyridine in the presence of lo- to loo-fold molar excess of unlabeled acetic anhydride (100 nmol//*l in acetonitrile) for selected intervals of time up to 18 h at room temperature and at 37°C. The samples were applied to the TLC plate with LPC and platelet activating factor used as markers to identify regions corresponding to LPC and acetylLPC, respectively. Regions corresponding to LPC, acetyl-LPC, and fatty acid (solvent front) were scraped and counted to determine the extent of the reaction and also to determine whether hydrolysis of LPC occurred during the course of the incubation period. The results of these initial studies demonstrated that the yield of acetyl-LPC product increased when a large (greater than 50-fold) molar excess of acetic anhydride was used and when the incubation time was extended to 18 h. The extent of product formation was also increased for those incubations at 37°C compared to incubations performed at room temperature. Incubation at 37°C for 18 h resulted in a near quantitative (i.e., greater than 95%) decrease in the radioactivity in the LPC region. However, only 60% of the radioactivity in the sample was recovered in the region corresponding to the acetyl-LPC product. The remaining radioactivity migrated to the solvent front (corresponding to fatty acid) indicating that under these conditions approximately 35% of the LPC was hydrolyzed to fatty acid and glycerophosphorylcholine. Incubation of LPC in acetonitrile with pyridine in the absence of acetic anhydride resulted in less than 2% hydrolysis of LPC. Similarly, incubation of LPC with acetic anhydride in the absence of pyridine resulted in less than 5% conversion of LPC to acetyl-LPC. Thus, an acylation catalyst (e.g., pyridine) is needed for the reaction to proceed to completion; however, in the presence of pyridine and acetic anhydride approximately 35% of
FOR
LYSOPHOSPHATIDYLCHOLINE
39
the LPC is hydrolyzed. One possible explanation for these observations is that in the presence of pyridine and acetic anhydride, the acetylation of LPC results in the production of a stoichiometrically equivalent amount of acetic acid and the accumulation of free acetic acid promotes further hydrolysis of LPC. Accordingly, dimethylaminopyridine (DMAP) was selected as an alternative catalyst for the acetylation reaction. Since DMAP contains a basic, tertiary amine group, this catalyst may also serve to “buffer” the production of acid produced during the course of the acetylation reaction and thereby prevent hydrolysis of LPC. Using DMAP as the acylation catalyst, there was no evidence of hydrolysis of LPC. Although the extent of the reaction increased with time with incubations performed at room temperature, the reaction did not reach completion within 24 h. In contrast, samples incubated for 12 to 15 h at 37°C in the presence of a 200-fold molar excess of acetic anhydride resulted in a greater than 95% conversion of LPC to acetyl-LPC. Therefore, for all subsequent analyses of LPC, DMAP was employed as the acylation catalyst, a 200-fold or greater molar excess of acetic anhydride was used, and all incubations were performed at 37°C for at least 12 h. After the appropriate conditions for avoiding hydrolysis while achieving quantitative acetylation of LPC using [14C]LPC were established, another series of experiments were performed using [3H]acetic anhydride and unlabeled LPC to permit analysis of LPC in extracts of isolated ventricular cells or whole tissue. Since the [3H] acetic anhydride from commercial sources is supplied in toluene, this solvent was used to prepare both labeled and unlabeled acetic anhydride solutions. The substitution of toluene for acetonitrile had no effect on the rate or extent of completion of the acetylation reaction. The same protocol was used to quantify the formation of [3H]acetyl-LPC from unlabeled LPC. After TLC separation and liquid scintillation counting of the [3H]acetylLPC product, the observed disintegrations per minute far exceeded those expected for a complete reaction in all samples. Examination of all regions of the TLC plate in lanes which contained only [3H]acetic anhydride showed 3H radioactivity in all regions. Most of the radioactivity was found at the solvent front but significant radioactivity was also found in the area corresponding to the acetyl-LPC region. Thus, in this solvent system C3H]acetic anhydride and/or [3H]acetic acid does not migrate as a discrete zone. Activation of the TLC plates by heating had no appreciable effect on the distribution of 3H radioactivity. Several alternative solvent systems for TLC were examined but this also did not improve the resolution of [3H]acetic acid. Selective extraction of [3H]acetyl-LPC from [3H]acetic anhydride and [3H]acetic acid using a variety of different binary and ternary solvent systems was also unsuccessful. To overcome this problem, the TLC plates were exposed to two successive
40
DOBMEYER,
CORR,
AND
CREER
developments in the same solvent system (system I). The TLC plates were removed from the tank and allowed to air-dry for 20 to 30 min before proceeding with the next solvent development. With this approach, it was possible to reduce the “background” 3H radioactivity in the [3H]acetyl-LPC region to consistently less than 0.1% of the total radioactivity applied to the plate. Under these conditions, with samples containing only 1 nmol of LPC and 100 nmol of [3H]acetic anhydride, the [3H]acetyl-LPC product would account for greater than 90% of the total 3H radioactivity in the acetyl-LPC region of the TLC plate. For all other samples, the background 3H radioactivity should be less than 10%. Thus, for all assays, a separate reagent blank was prepared for each TLC plate to allow correction for background 3H radioactivity and the reaction products formed following incubation of LPC with C3H]acetic anhydride were separated by two successive developments of the TLC plate in solvent system I. A representative example of the TLC separation obtained using this approach is illustrated in Fig. 1.
Evaluation
of the Assay Using [3H]Acetic
--)
3
solvent front
A
Anhydride
To determine the linearity of the assay, duplicate samples containing 0.5 to 15.0 nmol of LPC were prepared and analyzed. The results, shown in Fig. 2, demonstrate the linearity and reproducibility of the assay. The sensitivity of the assay was sufficient to permit accurate quantification of LPC in samples containing only 500 pmol of LPC. The specific activity of the [3H]acetylLPC product was approximately 60% of the value predicted based on the calculated specific activity of the [3H]acetic anhydride. This discrepancy may be due to hydrolysis of [3H]acetic anhydride or it may reflect the fact that the specific activity of the r3H]acetic anhydride was less than the value specified by the manufacturer. The specific activity of the [3H]acetyl-LPC product did not change appreciably when assays were performed 3 to 4 weeks after the [3H]acetic anhydride reagent was prepared and stored at -4°C. The results of quantification of choline lysoplasmalogen using this radiometric assay procedure are shown in Fig. 3. These results correlate closely with those observed for measurements of monoacyl-LPC and demonstrate that the assay is able to accurately quantify both monoacyl and choline lysoplasmalogen molecular species. Choline lysoplasmalogen and monoacyl LPC both contain one free hydroxyl group per molecule that is acetylated to produce labeled products of the same specific activity. Since choline lysoplasmalogen coelutes with monoacyl LPC during the course of HPLC separation and the acetylated products of choline lysoplasmalogen corn&ate with acetyl-LPC during TLC separation, the radiometric assay method does not distinguish these two choline lysophospholipid subclasses. Accord-
+ 1
2
origin
3
FIG. 1. TLC separation on silica gel G plates of 50 nmol each of soybean LPC (lane l), acetylated soybean LPC (lane 2), and I-O-alkyl-2-acetyl-sn-glycero-3-phosphocholine (platelet activating factor) (lane 3), following two successive developments in a solvent system comprising chloroform/acetone/methanol/acetic acid/water (6/g/2/ 2/l, v/v). Following TLC separation, the plate was stained by exposure to I2 vapor. The faint staining region shown by the arrow represents dimethylaminopyridine present in the reaction mixture used to prepare the acetylated soybean LPC product as described under Materials and Methods. The region indicated by the letter A to the right of the figure corresponds to the region that is scraped for liquid scintillation counting of 3H and “C dpm in samples following reaction with 13H]acetic anhydride.
ingly, in samples containing both subclasses, the result obtained by the radiometric assay procedure represents the sum of the monoacyl LPC and choline lysoplasmalogen content of the sample. A methods comparison study was performed to compare the results obtained by our method using [3H]acetic anhydride with those results obtained based on measurements of lipid phosphorus. Duplicate samples containing 6 to 30 nmol of LPC were prepared from the same LPC stock solutions. In one sample, the LPC content was determined by assay of lipid phosphorus and, in the other sample, the LPC content was determined by radiometric assay using [3H]acetic anhydride. For this study, relatively large amounts of LPC were required for each sample to permit accurate measurements of lipid
SENSITIVE,
0.0
3.2
RADIOMETRIC
12.8
6.4 9.6 nmol LPC
ASSAY
FOR
41
LYSOPHOSPHATIDYLCHOLINE
0.01 ’ ’ ’ ’ ’ ’ ’ ’ ’ 1 0.0 6.4 12.8 19.2 25.6 32.0 nmol LPC (radiometric)
16.0
FIG. 2. Representative radiometric assay results for the analysis of monoacyl LPC. Samples containing 0.5 to 15.0 nmol of l-heptadecanoyl-sn-glycero-3-phosphocholine (17:O LPC) were prepared and incubated with [3H]acetic anhydride, [‘%]LPC, and dimethylaminopyridine for 12 h at 37°C. The reaction was terminated, the reaction products were separated by TLC, and adjusted 3H dpm were determined as described under Materials and Methods. The correlation coefficient (r) shown at the bottom right portion of the figure was determined by least-squares linear regression analysis (n = 10 samples).
FIG. 4.
phosphorus. As shown in Fig. 4, there is excellent agreement between the values obtained for LPC by the two independent methods. The results of measurements of lipid phosphorus were slightly lower than those ob-
tained by the radiometric assay. However, the radiometric assay results agreed more closely with the values expected based on the known concentration of LPC in the stock solution.
Influence of Cardiolipin, on LPC Quantification
4.03.22.4 I .6 0.8 0
I 0.0
FIG. 3.
Methods comparison study. Paired samples containing 6.0 to 30.0 nmol of 1-heptadecanoyl-sn-glycero-3-phosphocholine (17:O LPC) were prepared. For each pair of samples, the LPC content was determined for one sample by microphosphate assay and by radiometric assay for the other sample as described under Materials and Methods. Shown at the right portion of the figure are the results of least-squares linear regression analysis of the data obtained from the sample pairs (n = 10 paired samples).
I
I
I
I
I
IIll
1.2 2.4 3.6 4.8 nmol LPC PLASMALOGEN
6.0
Representative radiometric assay results for the analysis of choline lysoplasmalogen (LPC plasmalogen). Samples containing 0.8 to 6.0 nmol of LPC plasmalogen were prepared and incubated with [3H]acetic anhydride, [“CILPC, and dimethylaminopyridine for 12 h at 37°C. The reaction was terminated, the reaction products were separated by TLC, and adjusted 3H dpm were determined as described under Materials and Methods. The correlation coefficient (r) shown at the bottom right portion of the figure was determined by least-squares linear regression analysis (n = 5 samples). The specific radioactivity of the [3H]acetic anhydride used for measurement of LPC plasmalogen was approximately fourfold greater than that of [3H]acetic anhydride used for the assay of monoacyl LPC shown in Fig. 2.
Sphingomyelin, by Radiometric
and Fatty Acid Assay
Since cardiolipin, sphingomyelin, and fatty acid frequently contaminate the LPC fraction isolated by conventional HPLC methods (vide supra), additional experiments were performed to determine the effects of these lipids on the measurement of LPC by radiometric assay. Samples containing 1.0 to 5.0 nmol of LPC were prepared and 25 nmol of cardiolipin, spingomyelin, or fatty acid was added to each sample. The quantity of cardiolipin, sphingomyelin, or fatty acid added was far in excess of that expected in the LPC fraction following HPLC isolation of LPC in extracts from whole tissue or isolated cardiac myocytes. Comparison of the results of LPC analysis obtained in the absence and presence of cardiolipin, sphingomyelin, or fatty acid demonstrated that the presence of these lipids did not significantly affect the results of the measurement of LPC using the radiometric assay procedure. Cardiolipin and sphingomyelin both contain a free hydroxyl group that could be acetylated under the conditions employed. However, the lack of interference of these lipids on the measurement of LPC indicates that the acetylated products formed from sphingomyelin or cardiolipin do not comigrate with acetyl-LPC under the conditions used for TLC separation of reaction products. If present in very large
42
DOBMEYER,
CORR,
amounts (i.e., greater than 50 to 100 nmol/sample), both cardiolipin and sphingomyelin could interfere with the quantification of LPC as a result of a reduction in the large molar excess of [3H]acetic anhydride available for acetylation of LPC. However, this is unlikely to occur in extracts of biological samples that are subjected to HPLC for isolation of the LPC fraction.
LPC Measurements
in Isolated Cardiac Myocytes
Because the TLC separation step could potentially allow the isolation of acetyl-LPC from other acetylated products in crude phospholipid mixtures obtained following extraction of cells or tissue, experiments were performed to determine whether isolation of the LPC fraction by HPLC was necessary since elimination of this step would considerably reduce the time required to perform LPC measurements. The results obtained in crude extracts were 2- to lo-fold greater than the values obtained for LPC measurement after isolation of the LPC fraction by HPLC. When trace amounts of [‘“ClLPC were added to the extracts before measurement of the LPC using [3H]acetic anhydride and both 14C and 3H radioactivities determined in the acetyl-LPC region of the TLC plate at the conclusion of the assay, the extent of conversion of [14C]LPC to [3H,‘4C]acetyl-LPC varied from 35 to 85%. Thus, although far less than quantitative conversion of LPC to acetyl-LPC was achieved in the extracts of tissue or cells analyzed directly, the apparent amount of LPC was 2- to lo-fold greater than expected. These results indicate that the overwhelming majority of 3H radioactivity detected in the acetyl-LPC region following TLC separation is due to the presence of acetylated products derived from other chloroformmethanol extractable substances in tissue or isolated cells which corn&ate with acetyl-LPC under the chromatographic conditions employed. These substances could be phospholipids, neutral lipids, or other chloroform-methanol extractable compounds containing a nucleophilic functional group capable of reacting with acetic anhydride. Since the presence of these substances resulted in a marked, positive interference in the direct measurement of LPC in extracts from tissue or isolated cells, it is apparent that isolation of the LPC fraction by HPLC is required prior to the reaction with [3H]acetic anhydride. Because of the potential loss of LPC which can occur during homogenization of the cells, extraction with chloroform-methanol, HPLC separation, and TLC separation of the reaction products produced after reaction with [3H]acetic anhydride, additional recovery studies were performed. For these experiments, carrier-free [14C]LPC was added to each sample of cells prior to homogenization. The recovery of 14C radioactivity was then determined at the conclusion of each step during the course of analysis. As shown in Table 1, excellent
AND
CREER TABLE
1
Recovery of LPC during Analysis by the Radiometric Assay Procedure Analytical
step
Bligh and Dyer extraction HPLC separation Acetylation and TLC separation
Recovery
(X)
96 f 1.5 95 f 2.3 86 f 9.0
Note. For this study, 1-[“C]pa1mitoy1-sn-g1ycero-3-phosphocho1ine ([“C]LPC) was added to isolated canine cardiac myocytes (equivalent to approximately 5 mg total cellular protein) and extracted, the LPC fraction was isolated by HPLC, and the acetylation reaction and TLC separation were performed as described under Materials and Methods. Aliquots were taken after extraction and HPLC separation to determine recovery of “C dpm at the conclusion of each of these analytical steps. Recovery following acetylation and TLC separation was determined as the percentage of total 14C dpm in the LPC fraction after HPLC separation that was recovered in the acetyl-LPC region of the TLC plate after acetylation and TLC separation. The recovery value shown therefore reflects both the extent of completion of the acetylation reaction and the recovery of “C dpm from the TLC plate. Values shown represent the means f SE of three experiments.
recovery of LPC during each step of the analytical procedure was obtained. However, recovery was not quantitative and even a small loss of LPC at each step would have a cumulative effect on the final result. In addition, there was significant variability in recovery after acetylation and TLC separation of the acetyl-LPC product. Accordingly, carrier-free [14C]LPC (17,000 dpm, 200 pmol) was added to the cell suspension before extraction and to the LPC standards before performing the acetylation reaction. The 14C radioactivity measured in the acetyl-LPC region of the TLC plate at the conclusion of the analytical procedure was then used to correct for the small but variable loss of LPC during the course of analysis. Since each sample of cells contains the equivalent of 4-10 mg of total cellular protein or approximately 2-5 nmol of total LPC, the mass of carrier-free [14C]LPC (200 pmol) added for determination of recovery would not significantly affect the total mass of LPC present in the sample. For analysis of samples containing less than 1 nmol of total LPC, it would, of course, be necessary to reduce the amount of carrier-free [14C]LPC added for correction of recovery. The results of analysis of LPC in isolated, adult canine cardiac myocytes using the present radiometric method were compared to results obtained by GC quantification of LPC. The GC method for measurement of LPC is based on the quantification of the fatty acid methyl esters and dimethylacetals produced by acid-catalyzed methanolysis of LPC using synthetic l-heptadecanoyl-an-glycero-3-phosphocholine (17:0 LPC) as an internal standard as previously described (6). The GC measurements of LPC required relatively large amounts of isolated cells (equivalent to 15-25 mg of cellular pro-
SENSITIVE,
RADIOMETRIC
ASSAY
tein) compared to that required for analysis of LPC by the radiometric assay using [3H]acetic anhydride (2-5 mg of cellular protein). The results obtained by the radiametric assay procedure (0.38 f 0.04 nmol/mg protein (mean f SE), n = 16 samples) compared very favorably to those results obtained by GC analysis of LPC (0.39 f 0.08 nmol/mg protein (mean f SE), n = 15 samples). Thus, in isolated adult canine cardiac myocytes, the results of the present method are in close agreement with those results obtained by other methods employed for the analysis of LPC. DISCUSSION
The radiometric method developed for the quantification of LPC in tissue samples or in isolated cell preparations offers several advantages compared to other conventional assays. For example, the analysis of LPC based on measurements of total lipid phosphorus or by GC quantification of the fatty acid methyl esters or dimethylacetals produced by acid-catalyzed methanolysis of LPC requires at least 5-10 nmol of LPC per sample. The present method is much more sensitive permitting analysis of LPC in samples containing less than 1 nmol of LPC. Further enhancement of sensitivity could be achieved by increasing the specific activity of the [3H]acetic anhydride reagent. The radiometric assay procedure is also a cost-effective approach for the analysis of LPC in biological samples. At the present time, commercial sources of [3H]acetic anhydride provide the reagent at a relatively low cost of less than 50 cents per assay sample. Our approach to the analysis of LPC is also not subject to interference by fatty acids, sphingomyelin, or cardiolipin which frequently contaminates the LPC fraction isolated by HPLC methods which employ acetonitrile, methanol, and water in the mobile phase (vide supra). Thus, removal of fatty acids and cardiolipin by anionexchange chromatography of the LPC fraction obtained following HPLC separation (6) is not required prior to the analysis of LPC. The present method is able to accurately quantify choline lysoplasmalogen which coelutes with monoacyl LPC during the course of HPLC isolation by those methods which separate phospholipids into classes based on differences in polar headgroup composition. Other derivatization procedures based on CC separation and quantitation of the fatty acid methyl ester derivatives which employ basic conditions for methanolysis preclude detection of lysoplasmalogens (14,15). Thus, the ability of the present method to detect choline lysoplasmalogen molecular species is a distinct advantage for the analysis of choline lysophospholipids in myocardial tissue samples or in isolated cardiac myocyte preparations where plasmalogens account for a significant fraction of the to-
FOR
43
LYSOPHOSPHATIDYLCHOLINE
tal choline phospholipid pool. The radiometric assay approach for the quantification of LPC developed in this study should be readily adaptable for analytical studies of LPC metabolism in other tissues and also in isolated cell preparations derived from a variety of different sources. REFERENCES 1. Corr, P. B., Gross, 135-154.
R. W., and Sobel,
B. E. (1984)
Res. 66,
Circ.
2. Corr, P. B., Snyder, D. W., Cain, M. E., Crafford, W. A., Jr., Gross, R. W., and Sobel, B. E. (1981) Circ. Res. 49,354-363. 3. Snyder, D. W., Crafford, W. A., Jr., Glashow, Sobel, B. E., and Corr, P. B. (1981) Amer. J. H707. 4. Arnsdorf, 5. Clarkson, 556.
M. F., and Sawicki, C. W., and Ten
G. J. (1981) Eick,
J. L., Rankin,
Physid. 241,
Circ.
R. E. (1983)
D., H700-
Res. 49,16-30. Res. 52, 543-
Circ.
6. Corr, P. B., Creer, M. H., Yamada, K. A., Saffitz, J. E., and Sobel, B. E. (1989) J. Clin. Invest. 83,927-936. 7. Creer, M. H., Pastor, C., Corr, P. B., Gross, R. W., and Sobel, B. E. (1985) Anal. B&hem. 144,65-74. W. W. (1973) 8. Christie, ford, NY.
Lipid
Analysis,
p. 192, Pergamon,
Elms-
9. Kates, M. (1975) Techniques of Lipidology, p. 552, North Holland/Elsevier, Amsterdam/New York. 10. White, D. A. (1973) in Form and Function of Phospholipids (Ansell, G. B., Hawthrone, J. N., and Dawson, R. M. C., Eds.), p. 448, Elsevier, Amsterdam/New York. 11. Horrocks, L. A., and Sharma, M. (1982) in Phospholipids (Ansell, G. B., and Hawthorne, J. N., Eds.), p. 51, Elsevier, Amsterdam/ New York. 12. Mogelson,
S., Wilson,
G. E., and Sobel,
Biochim. Bio-
B. E. (1980)
phys. Acta 619,680-688. 13. Bjerve,
K. S., Daae,
L. N. W., and Bremer,
J. (1974)
Anal. Bio-
&em. 58,238-245. 14. Shaikh,
N. A., and Downar,
E. (1981)
15. Kates, M. (1975) Techniques land/Elsevier, Amsterdam/New
Circ. Res. 49,316-325.
of Lipidology, York.
p. 361, North
Hol-
16. Creer, M. H., and Gross, R. W. (1985) J. Chromatogr. 338,61-69. 17. Markwell, M. A. K., Haas, S. H., Tolbert, N. E., and Bieber, L. L. (1981) in Methods in Enzymology (Lowenstein, J. M., Ed.), Vol. 72, p. 296, Academic Press, New York. 18. Chen, P. S., Toribara, 28,1756-1758.
T. Y., and Warner,
19. Heathers, G. P., Yamada, K. A., Kanter, (1987) Circ. Res. 61,735-746. 20. Bligh,
E. G., andDyer,
W. J. (1959)
H. (1956) E. M.,
And. Chem.
and Corr,
P. B.
J. Biochem. Physiol. 37,911-
917. 21. Gross,
R. W., and Sobel,
B. E. (1980)
J. Chromatogr. 197,79-85.
22. Fox, B. W. (1976) Techniques Scintillation Counting, pp. Amsterdam/New York.
of Sample Preparation for Liquid 212-220, North Holland/Elsevier,
23. Fox, B. W. (1976) Techniques Scintillation Counting, pp. Amsterdam/New York.
of Sample Preparation for Liquid 207-211, North Holland/Elsevier,
24. Polonsky, Benveniste,
J., Tence, J. (1980)
M.,
Varenne,
P., Das,
B. C., Lunel,
J., and
Proc. Natl. Acad. Sci. USA 77,7019-7023.
