phosphatidyldm5migu4zj3pb.cloudfront.net/manuscripts/104000/104679/jci62104679.pdf · cent methanol...

Journal of Clinical InvestigationVol. 41, No. 12, 1962

PLATELET PHOSPHATIDES. THEIR FATTY ACID ANDALDEHYDECOMPOSITIONANDACTIVITY IN DIFFERENT CLOTTING

SYSTEMS*

By AARONJ. MARCUS,HARRIS L. ULLMAN, LENOREB. SAFIER, ANDHAROLDS. BALLARD

(From the Hematology Section, N. Y. Veterans Administration Hospital, and the Departmentof Medicine, Cornell University Medical College, New York, N. Y.)

(Submitted for publication July 6, 1962; accepted August 22, 1962)

Previous studies from this laboratory revealedthat phosphatidyl serine (PS) isolated from plate-let phospholipid extracts by silicic-acid columnchromatography could replace blood platelets intests of thromboplastin generation and prothrom-bin consumption of platelet-poor "native" plasma(1, 2). When rigorous criteria of identificationwere employed, it was found that phosphatidylethanolamine (PE) fractions contained traces ofPS (2); coagulant properties were thereforedifficult to evaluate. Rouser, O'Brien, and Heller(3) recently accomplished the complete separa-tion of PE from PS by means of a double columntechnique using silicic acid and ammoniated silicicacid. Their studies were carried out on lipids ofbeef brain, but the methods could be applied tohuman brain and platelet extracts.

In 1961, an earlier study of the fatty acids ofplatelet phosphatides with gas-liquid chromatog-raphy was reported (4). PE and PS were then,of necessity, studied as a combined fraction, andsimilarly the choline phosphatides were not ade-quately separated. In addition, the significanceof oxidative changes consequent to prolongedseparations and exposure of phospholipids to airwas not appreciated, a factor which influenced theover-all results.

Zilversmit, Marcus, and Ullman (5) reportedthat 16 per cent (dry weight) of total platelet lipidextracts were in the plasmalogen form (phos-pholipids that liberate higher fatty aldehydes onhydrolysis). It could not, however, be determinedhow the plasmalogen was distributed among indi-vidual phosphoglycerides, although it was known(2) that the b)ulk of plasmalogen was associated

* Investigation supported by a grant from the NewYork Heart Association. Presented in part at the meet-ings of the Federation of American Societies for Experi-mental Biology, Atlantic City N. J., April, 1962.

with the combined PE-PS fractions. Whenmethyl esters of fatty acids of phosphoglyceridesare formed, the aldehydogenic chains of plasmalo-gens are simultaneously converted to dimethyl-acetals. These derivatives can be separated frommethyl esters on polar and nonpolar gas-liquidchromatographic columns (4, 6, 7), thus affordingan additional qualitative and quantitative meas-urement of plasmalogen to supplement other ana-lytical methods (5).

In the present study, human platelet PE and PSwere separated by techniques that minimized oxi-dative changes, which permitted an accurate evalu-ation of the activity of these lipids in differentblood clotting systems. In addition, it was pos-sible to carry out a detailed study of the fatty acidand fatty aldehyde composition of platelet phos-phatides by gas-liquid chromatography.

MATERIALS AND METHODS

In order to standardize and insure reproducibility ofthe extractions, separations, and coagulation tests, 8preliminary runs were performed with lipids of fresh,human brain tissue. The main data to be presented arethe findings of 7 similar runs with human platelet ma-terial. Although details of the brain separations willnot be presented, certain pertinent comparisons will bemade with platelets.

Solvents were freshly redistilled in glass, and 1 percent methanol was added to the chloroform after distil-lation. Procedures and transfers were carried out underhighly purified nitrogen, less than 10 parts oxygen permillion. Extracted lipids were not exposed to the at-mosphere. "Deoxygenation" of solvents was carriedout in a rotary evaporator connected to a water pumpvacuum. Solvents were subjected to reduced pressurefor 3 minutes after bubbling had ceased, and atmosphericpressure was restored with nitrogen. Absolute ethanolwas rendered aldehyde-free by the addition of 3 g per Lm-phenylenediamine dihydrochloride. After 24 hours,the ethanol was collected by distillation.

Preparation of platelet lipid extracts. Thirty-four units

2198

PLATELET PHOSPHATIDES

of blood from donors with normal or elevated plateletcounts were collected and processed (2). A large heter-ogeneous pool was used to minimize the possibility of al-terations in fatty acid composition due to dietary differ-ences. Extraction was performed on approximately 13 g(wet weight) of platelet material at a time and yieldedabout 400 mg of lipid which was the starting materialfor each separation. The platelets were weighed, trans-ferred to a Waring Blendor, and homogenized in 150 mlchloroform: methanol 2: 1 for 40 seconds in a nitrogenatmosphere. The contents of the Blendor were transferredto a new sintered glass filter of coarse grade. Gentlesuction was applied beneath the filter, which was cov-ered with a funnel through which nitrogen was passed.When filtration was complete, the material on the surfaceof the filter was dried, and 150 ml of the same solventwas added. After 5 minutes, gentle suction was resumedand the re-extraction repeated. Finally, the filtrate waspassed through fat-free sharkskin filter paper into a side-arm Erlenmeyer flask and was dried under reduced pres-sure and nitrogen in a 300 C water bath. During evapo-ration, deoxygenated absolute ethanol was added to theflask, thereby forming a water-ethanol azeotrope to aidin removal of last traces of water (8). The dry ma-terial was dissolved in chloroform: methanol 4: 1, asample removed for weight determination, and the re-mainder stored under nitrogen for a maximum of 12hours at -200 before column chromatography. Afterstorage, a precipitate was noted, which occasionally wasremoved by filtration without alteration of results. Inmost runs, the extract was placed on the columns intact.

Silicic acid chromatography. Mallinckrodt silicic acid,100 mesh, was used. Sixty g of silicic acid was washedthree times on a sintered glass filter with 180 ml ofmethanol, followed by 180 ml chloroform: methanol 1: 1,180 ml chloroform, and finally 60 ml methanol (3). Thewashing was facilitated by mild suction. The silicic acidwas then placed in a 1,000-ml three-necked flask and heated12 to 15 hours at 1200 C under negative pressure andnitrogen. After cooling, 200 ml of chloroform: methanol4: 1 was added and the slurry poured into the glass col-umn. The apparatus used was that of Hirsch and Ahrens(8) with one modification: column dimensions were al-tered to 2.5 X 40 cm so that 60 g of silicic acid extendedto a height of 20 cm after application of nitrogen pres-sure (3). About 400 mg of platelet lipid was appliedto the column in a small volume of chloroform: methanol4: 1, and the chromatography was done at 100 C. Astream of nitrogen was passed into each tube as itfilled. Fractions were collected by hand (9), which per-mitted immediate capping and storage at - 200 C. Sam-ples of the 10-ml fractions were removed and examinedby means of a rapid Ninhydrin 1 test (3), and approxi-mately every tenth tube was studied in greater detail bythin-layer and silicic-acid paper chromatography. Theflow rate was adjusted to 2 ml per minute by nitrogenpressure. The elution pattern was also followed by add-

ing 1 ml of eluate to 1 ml 5 N sulfuric acid in a 30-mlKjeldahl flask and heating at 3200 C for 5 to 7 minutes,after which no further charring occurred. The intensityof charring was an index of the amount of organic ma-terial present. Eluting solvents for the first silicic-acidcolumn separation (run A) were chloroform: methanol4: 1 until the fractions became negative to Ninhydrin.The run was completed with 990 ml chloroform: methanol5: 4 and 500 ml chloroform: methanol 1: 9 (10). One-mlsamples were removed for phosphorus determination, andbefore storage of fractions under nitrogen at - 200 C, 0.1ml concentrated ammonia was added to each tube in orderto retard plasmalogen breakdown (3). The final con-centration of ammonia in each tube was 0.15 M.

All Ninhydrin-positive fractions from run A wereverified as solely a mixture of PE and PS by thin-layerand silicic-acid paper chromatography. These werepooled, evaporated to 5 ml, stored at - 200 C, and ap-plied 12 hours later to a deoxygenated ammoniated silicicacid column (3). This procedure, run B, separated PEfrom PS. The PE was eluted with chloroform: methanol4: 1, and the retained acidic PS with methanol. Frac-tions were monitored as for run A.

Phosphorus was analyzed by the method of Bartlett(11). The procedure was carried out in 25-ml, gradu-ated, blood-sugar tubes (Nash, combination type) in aheating block. Standard curves were determined for theColeman Jr. spectrophotometer at 700 my and the Beck-man DU at 830 miz; for convenience, measurements weremade with the Coleman instrument.

Thin-layer chromatography. The identity and purityof lipids in individual fractions and pooled material wasdetermined by this useful technique. Silica gel plates 2

were prepared, the lipids applied, and the chromatogramsdeveloped in a solvent system of chloroform: methanol:water: concentrated ammonia, 75: 25: 4: 1 (3, 12). Foridentification, the air-dried plates were sprayed with 50per cent sulfuric acid and heated in a convection oven at2500 C for 30 minutes. Less than 25 jug of phospholipidcould be detected on the chromatograms. Although Rfvalues were not reproducible on thin-layer plates, theprincipal purpose of this chromatography was the detec-tion of oxidation products and minor components. Largeamounts, 0.5 to 1.5 mg, of some fractions were analyzedby thin-layer chromatography as a rigorous test of theirpurity: contamination with more than 0.01 mg of otherlipids could be excluded.

Silicic-acid paper chromatography. This was used tosupplement and confirm the findings with thin-layerchromatography. The methods were the same as previ-ously reported (1, 2).

Oxidation studies. The pooled PE and PS fractionswere examined for the presence of auto-oxidative changesas follows. Samples were dissolved in cyclohexane at aconcentration of 1 mg per ml and transferred to 1-cmcuvettes. The molar extinction coefficients were deter-mined at 235 and 275 my (3) in a Beckman DU spectro-

2 Brinkman Instruments, Inc.. Great Neck, N. Y.

2199

1 1,2,3-Indantrione hydrate.

MARCUS,ULLMAN, SAFIER, AND BALLARD

photometer. In addition, a continuous scan was madewith a' Beckman DB recording spectrophotometer.

Blood coagulation studies. Less than 24 hours after theseparated lipids were identified and oxidation studies- car-ried- out, a sample -was weighed. Amounts of 1 to 10 mgwere removed, evaporated under reduced pressure andnitrogen to a small volume, and transferred under ni-trogen to a Potter-Elvehjem 5A tissue grinder, :wherethey were dried. Imidazole buffered saline (2) wasadded, the lipid was rapidly emulsified, and transferredto a cork-stoppered test tube. The suspension producedwas considered the "undiluted specimen," from which asample was removed for phosphorus determination.Fractions of PS, as well- as of lecithin, sphingomyelin,and phosphatidyl inositol (PI), formed stable emulsions,but PE could only be partially suspended. Other pro-cedures were carried out in an attempt to emulsify PE.These included the following. a) Sonication with twotypes of apparatus (Raytheon model DS-101 and Bransonmodel S-75). Satisfactory emulsions were produced withthese instruments, but slight settling was noted after 40 to60 minutes. b) Admixture of PE with sodium desoxy-cholate.3 It was previously determined that sodium de-soxycholate at a concentration of 0.75 mg per ml did notinterfere with the thromboplastin generation test (TGT).Therefore, the undiluted specimen of PE was prepared inimidazole buffered saline containing 0.75 mg sodium de-soxycholate per ml, and a stable emulsion was formed.c) Suspension of PE with other materials: oleic acid,cholesterol, lecithin, and polysorbate 80 (Tween 80).All these emulsions showed some settling after 1 hour.The lipids to be tested were substituted for platelet re-agent in the TGT immediately after suspension in buf-fer (1, 2). The activity of the fractionated material wascompared to that obtained with total platelet lipid orcrude brain "cephalin." In addition, the effect of vari-ous phosphatides on the recalcified clotting time of plate-let-poor plasma was studied. Fresh citrated plasmafrom a fasting donor was used for each group of experi-ments. Blood was collected into silicone-treated test tubesthrough plastic tubing (Cutter "saftidonor") by gravityflow. During the test, 0.1 ml of freshly prepared lipidin imidazole buffered saline was added to 0.1 ml ofplasma in uncoated glass tubes, and after a 3-minuteincubation period at 370 C, 0.1 ml of 0.025 M calciumchloride was added (13).

Plasmalogen determinations. The plasmalogen contentof PE and the choline phosphoglycerides was determinedby iodine addition (5, 14). A 3-ml sample was removedfrom a total PE pool and the determinations were carriedout in quadruplicate. These samples contained 1.08Amoles phosphorus per ml. PS was not analyzed by thetechnique because of the small amounts recovered. Cal-culations of dimethylacetal peak areas by gas-liquid chro-matography --were correlated with the results of iodineaddition and the combined information was used as ameasure of plasmalogen.

8 Nutritional Biochemicals Corp.

Gas-liquid chromatography. A Barber-Colman model10 chromatographic apparatus equipped with two columnsand two strontium' detector cells was employed. U-shapedpyrex glass columns, 4 or 6 feet in over-all length withan internal diameter of 4 to 5 mm, were used. Columnperformance and linearity of detector response to masswere checked at least twice a week with standards ob-tained from the Metabolism Study Section of the Na-tional Institutes of Health. The range of error was 0to 3 per cent. Gas-liquid chromatographic analyses on in-dividual methyl ester samples were repeated many timeswith varying quantities before conclusions were drawnas to column performance. Initially, quantification of di-methylacetals gave results that were low. This wasfound to be a function of the pH of the stationary sup-port. Traces of acid resulted in distorted peaks of re-duced size-probably aldehydes-that separated poorlyfrom methyl esters.

Methyl esters (and dimethylacetals) were formedfrom column eluates as previously reported (4). In thelecithin and sphingomyelin fractions, methyl esters werealso produced by the method of Nelson and Freeman(15), as well as by sealed tube hydrolysis with 5 percent hydrogen chloride in superdry methanol (4).

Nonpolar columns. The stationary supports used wereChromosorb W, Gas-Chrom P, Gas-Chrom CL, Gas-Chrom Z (siliconized),4 Anakrom AB, Anakrom ABS(siliconized), and Celite 545.5 Initially, all of these 80-100 mesh supports were washed with acid and alkali. ThepH was 6.4, as determined by suspending a small amount ofpowder in distilled water. When the stationary supportswere directly coated with stationary phase, Apiezon L or M,the dimethylacetals did not appear. It was found that thestationary supports must not be brought back to neutralityafter the alkali wash (16, 17). They were treated there-after in the following manner. The material was washedwith 1 per cent methanolic potassium hydroxide, whichwas removed by filtration through a Buchner funnel.The support was dried in a convection oven for 12 hoursat 1200 C, and then at a concentration of 1 mg per mldistilled water, the pH of the powder was 7.4. The al-kaline stationary support was coated with 10 to 15 percent Apiezon L or M (18). The columns were condi-tioned for 12 to 24 hours at 2250 to 2500 C under argonpressure slightly above that used for normal operation.During the conditioning period, the columns were"soaked" with loads of 75 to 100 mg mixed methyl es-ters (19). The column efficiency was 2,700 theoreticalplates at methyl stearate. Detector cells were operatedat 750 to 1,250 v, depending on the amount giving optimallinearity, and the amplifier setting was 3 X 10' a. Thecolumn temperature was 1970 C, the detector was main-tained at 2200, and flash heater at 1970 C. Argon pres-sure was varied to achieve optimal flow rates. For com-parison, methyl esters of platelet PE and PS were si-multaneously analyzed on an Apiezon column at the

4Applied Science Laboratories, State College, Pa.5 Analabs, Hamden, Conn.

2200


PLATELET SEPARATION - RUN A

E-NINHYDRIN +

o 10 30 50 70 I10 130 150

TUBE NUMBER

FIG. 1. FRACTIONATION OF THE MAIN PHOSPHOLIPID CLASSES IN BLOODPLATELETS ON A SILICIC ACID COLUMN. Phosphatidyl ethanolamine (PE)and phosphatidyl serine (PS) were separated from the other platelet phos-phoglycerides. All the PE and most of the PS was in peak 1. Peak 2 wasmainly phosphatidyl inositol (PI), but contained small amounts of PS.Lecithin (peak 3) was well separated, but overlapped with sphingomyelin(peak 4).

Rockefeller Institute, New York, by Dr. John W. Farqu-har (18).

Polar columns. The stationary supports were the same

as those used for nonpolar columns. Pretreatment of thestationary supports with alkali was also helpful. The op-timal pH was 6.8. It was found advantageous to "soak"the columns with methyl ester mixtures during the con-

ditioning period. If apparent hydrolysis of dimethylacetalsoccurred, "resoaking" the column restored their analysisto the expected full response. The supports were coatedwith 10 to 15 per cent ethylene glycol adipate 6, 7 (EGA).Operating conditions were: cell voltage, 1,000 to 1,100;amplifier setting, 3 X 10-, a; column temperature, 1730 C;detector, 2200 C; and flash heater, 1730 C. Columnefficiency was approximately 2,900 theoretical plates atmethyl stearate. Ethylene glycol succinate was alsoused as stationary phase and gave results similar tothose for EGA (4).

6 Applied Science Laboratories, State College, Pa.7 Several batches of catalyst-free EGAof high molecular

weight (7) that provided columns of lower "bleed" rateswere furnished by Dr. J. W. Farquhar.

RESULTS

Column chromatographyWith accumulated experience, the qualitative

and quantitative aspect of the fractionations be-came more precise. When the starting materialconsisted of 10 to 16 g (wet weight) of platelets,the weight of the lipid obtained after extractionwas 400 to 420 mg, or approximately 32 mg perg platelets. At the conclusion of run A, about56 mg of PE and PS were obtained. After theirseparation in run B, the yield of PE was 25 to 35mg and of PS, 13 to 16 mg. The final amountsvaried with the number of tubes acceptable as

Ninhydrin-positive, since samples that gave traceNinhydrin reactions were arbitrarily rejected.Although all the PE finally recovered was confinedto one peak (Figure 3), there were variableamounts of PS in the inositol fractions (Figure

m

Z)I-

U)a_0n0

1:

2201


1); this probably represented the sodium salt ofPS (20), but was not investigated further. Theresults of run A are depicted in Figure 1. Allthe Ninhydrin-positive material was found in peak1. Peak 2 was mainly PI, but showed smallamounts of PS. Analysis of peak 3 showedlecithin only, and peak 4 contained mainly sphingo-myelin with small overlapping amounts of lecithin.No attempt was made to purify the sphingomyelinmaterial. Figure 2 is a photograph of a thin-layer plate showing different concentrations ofsamples from peak 1. The lipid was applied invarying amounts to insure that only PE and PSwere present. Figure 3 indicates the phosphorusanalyses on eluates from the ammoniated silicicacid column, run B. Although the ascending limbof the PE peak was always sharp, a certainamount of tailing was noted in the descending por-tion. It was important to allow at least 50 ml ofNinhydrin-negative eluate to appear before addingthe methanol. The PS elution consistentlyshowed a sharp configuration. Figure 4 is a pho-tograph of a thin-layer plate examined after thePE-PS fraction was separated, pooled, and con-

PLATELET SEPARATION- RUN B

20 30TUBE NUMBER

FIG. 3. SEPARATION OF PHOSPHATIDYLETHANOLAMINE(PE) FROM PHOSPHATIDYL SERINE (PS) ON AN AM-MONIATED SILICIC ACID COLUMN. The column was madealkaline by passage of 70 ml ammoniated chloroform:methanol 4:1 before application of PE-PS (3). Con-sequently PE was eluted first, but PS was retained andrecovered later with methanol.

- _ _ _ _ A

S_..

1l/ 2 3 #t d;

FIG. 2. THIN-LAYER CHROMATOGRAMON SILICA GELOF PEAK 1, RUN A. Before the combined phosphatidylethanolamine-phosphatidyl serine (PE-PS) fraction was

rechromatographed, it was evaporated to a small volumeand varying concentrations were applied to a thin-layerplate, to verify that other lipid material was not present.Lanes 1, 4, and 5 represent the PE-PS fraction, lanes 2and 3 are PE and PS markers, respectively.

centrated. Oxidation and breakdown products,which appear as extra spots (3), were absent.Identification was further confirmed by silicic acidpaper chromatography and the pattern shownby the fatty acids and aldehydes on gas-liquidchromatography.

Figure 5 represents the absorption spectra ofplatelet PE and PS, respectively, as well as of theoxidized forms of these phosphatides. The lipidsformed clear solutions in cyclohexane, which wasa prerequisite for spectrophotometry (3). Molarextinction coefficients at 235 and 275 mu averaged279 and 47 for PE, while the values for PS were618 and 179. These findings indicated that mini-mal oxidative changes took place (21). Figure 6depicts a thin-layer plate of the oxidized PE andPS samples shown in Figure 5.

Blood coagulation studies

A. Thromboplastin generation test. Table Ishows results obtained with the various phospho-lipids studied and with controls. The shortest

2202

0

fd.;:,


substrate clotting times obtained after 6 minutesof incubation in the TGT are presented, and theamount of lipid phosphorus introduced into thegenerating mixture is indicated. Separate TGTexperiments were carried out 6 times with similarresults.

PE had no activity in the TGT in any conetn-tration. The PE preparations in which emulsifi-cation was enhanced by ancillary means were alsoinert in this test. Clotting times obtained withPS were invariably within 2 to 3 seconds of con-trol values. The PI fractions (run A, peak 2)containing PS were also highly active. Clottingtimes with this combination of phospholipids werenot more than 1 second from control times atoptimal concentration. Lecithin and sphingomye-lin had no clotting activity. In general, optimalclotting times were noted with PS at a 1: 10 dilu-tion. When active materials were more concen-trated, a typical anticoagulant effect was observed

6000-5800-

5400

5000-

4600-

4200-

3600-

3400-

& 3000-

2600

2200-

1800-

1400-

4

10001

600-

200-0-

PS

Itai% ~~~~~~~~S220 240 260 280 300 320 340

mALFIG. 5. ABSORPTION SPECTRA OF PHOSPHATIDYL ETHA-

NOLAMINE (PE), PHOSPHATIDYL SERINE (PS), OXIDIZEDPE, AND OXIDIZED PS. The lower curves representspectra obtained with PE and PS after separation andmaintenance in a nitrogen atmosphere. For comparison,the upper curves illustrate spectra of the same lipids afterstorage and handling by conventional techniques.

PS PE

FIG. 4. THIN-LAYER SILICA GEL CHROMATOGRAPHYOF

PHOSPHATIDYLETHANOLAMINE(PE) AND PHOSPHATIDYL

SERINE (PS) AFTER RUN B. PS always shows a trailin this solvent system. Silicic acid paper chromatographyconfirmed these results.

(1, 2). Table I also shows that 2.5 ,ug lipid phos-phorus in the generating mixture gave optimalclotting times with PS, whereas 11.7 ptg wererequired for the total platelet lipid extract. Onthe other hand, when PS was compared with brain"cephalin," the amounts needed for maximal ac-tivity were similar.

After varying periods of storage in buffer at-20° C, the PS fractions showed considerableloss of activity in the TGT. After storage in cy-clohexane under nitrogen at - 200 C, however,

i"WPW

2203

I-VV


F",~~~~~~~~~~~~~~~~~ L

PE ~~~PSFIG. 6. THIN-LAYER SILICA GEL CHROMATOGRAMOF OXIDIZED PHOSPHATIDYL ETHA-

NOLAMINE (PE) AND PHOSPHATIDYL SERINE (PS). The absorption spectra of thesesamples are shown in the upper portion of Figure 5. Oxidized PE migrates in thesolvent front, shows streaking, and considerable material remains at the point of ap-plication. Oxidized PS develops marked streaking and shows material behind thiesolvent front and at the origin. Compare with Figure 4.

activity was preserved for at least 3 months. Thiscorrelated with spectrophotometric evidence ofoxidation.

B. Recalcified clotting time. Figure 7 depictsthe results obtained when serial dilutions of PE,

TABLE I

Separated platelet phosphatides in thethromboplastin generation test *

Optimalclotting

timeafter

6-minute P per lipidincuba- sample intion in generating

Platelet phosphatide test mixture

seconds yAgPE >90 All concentrationsPS 10 2.5

PI (+PS) 8 2.2

Lecithin >90 All concentrationsSphingomyelin >90 All concentrationsTotal platelet lipid 9 11.7

(chloroform: methanolextract)

Brain "cephalin" control 9 2.4(acetone-dried,chloroform extract)

* Abbreviations: PE=phosphatidyl ethanolamine, PS =phosphatidylserine, PI =phosphatidyl inositol. Results are compared with appro-priate controls. PS and PI( +PS) were active, while PE, lecithin, andsphingomyelin were inert.

PS, and total platelet lipid extracts were studiedin the recalcified clotting time of platelet-poorplasma. In marked contrast to results with theTGT, PE showed good activity in this system.PS was slightly more active than PE at all con-centrations. Coagulation times with total plateletlipid extracts were consistently shorter than thepurified phosphoglycerides. All concentrations oflecithin and sphingomyelin were inert. PI, whichcontained PS, gave the same values as PS. Thiscombination of lipids had good clotting activity.The series of recalcified clotting times were car-ried out on 5 separate occasions with the sameresults. All clotting studies carried out with brainphospholipids gave results similar to those ob-tained with platelet phosphatides.

Gas-liquid chromatography

Table II lists the molar distribution of the fattyacids and aldehydes in separated platelet phospha-tides on EGAcolumns. Table III lists the equiva-lent information provided by Apiezon columns, andalso lists the confirmatory findings with our PEand PS preparations analyzed independently byDr. John W. Farquhar. Figures 8 and 9 show

2204

.1 V <. lIi \ I I )

300 .m.y:....

250-

200

PE

I 0 0 A_ ~~~~~~~~~~~~~~Total Lipid

1 00

0 2 3 4 5 6 7 8 9

pg Lipid P Added To Incubation Mixture

FIG. 7. EFFECT OF PLATELET LIPIDS ON THE RECALCIFIED CLOTTING TIME OF PLATELET-POOR PLASMA. Both phos-phatidyl ethanolamine (PE) and phosphatidyl serine (PS) were capable of shortening the recalcified clotting time ofplatelet-poor plasma. Phosphatidyl inositol (PI)-PS gave the same results, but lecithin and sphingomyelin (notshown) had no activity. The total platelet lipid extract always showed the shortest clotting times.

TABLE 11

Fatty acids and aldehydes of platelet phosphatides on ethylene glycol adipate polyester at 173° C *

ShorthandFamiliar name designationt PE PS PI (+PS) Lecithin

moles % moles % moles % moles %Lauric 12:0 0.2 trace 0.1

14:0 DMA 0.2Myristic 14:0 trace 0.6 trace 0.1

15:0 0.1 0.5 0.316:0 DMA 8.8 0.9 0.4 1 0.2

Palmitic 16:0 3.1 0.9 1.9 35.0Palmitoleic 16:1 0.4 1.4 0.5 1.3

? 0.317:0 0.1 0.6 0.7

? 0.2 trace18:0 DMA 19.9 0.2 0.5

Stearic 18:0 18.2 46.9 46.1 13.7Oleic 18:1 7.2 20.1 11.7 31.2Linoleic 18:2 0.8 1.7 1.4 5.2Linolenic 18:3 1.3Arachidic 20:0 0.2 1.3 trace 0.8

20:1 1.3 1.120:3 1.5 0.5

Arachidonic 20:4 36.2 23.9 35.4 9.820:5 0.322: unt 1.922:5 0.222:5 0.122:6 0.8

* Abbreviations as in Table I, and DMA= dimethylacetals.t No. of carbon atoms:no. of double bonds.t un = unsaturated acid.


TABLE III

Flatly acids and aldehydcs of platelet phospualides on Apiezon L at 197' ( *

ShorthandFamiliar name designationt PE PE PS PS $ PI ( +PS) Lecithin

moles % moles % moles % moles % moles % moles %Lauric 12:0 trace 0.2Myristic 14:0 trace trace trace 0.3 trace

15:0 0.1 0.2 0.3 0.216:0 DMA 10.5 11.1 0.3 1.8 1.2 0.8Palmitic 16:0 3.3 3.5 0.7 1.8 2.7 33.4Palmitoleic 16:1 0.3 0.3 0.3 0.3 0.7 0.917:Obr§ DMA 0.317:0 0.3 0.2 0.4 trace18:0 DMA 19.0 21.1 0.2 0.8

Stearic 18:0 13.8 16.1 47.2 44.4 45.6 16.518:1 DMA 3.8 2.1

Oleic 18:1 8.4 6.9 22.6 22.4 11.7 31.3Linoleic and

linolenic 18:2 + 18:3 0.9 1.0 0.9 1.7 0.6 5.2Arachidic 20:0 1.3

20:1 0.620:3 1.1 0.9 0.9 1.4 1.1

Arachidonic 20:4(+20:5) 37.5 36.8 24.9 26.2 34.6 11.922:5 0.522:6 0.6

* Abbreviations as in Tables I and II.t No. of carbon atoms: no. of double bonds.t Results obtained by Dr. J. W. Farquhar (18).§ br = branched-chain acid.

tracings of chromatograms of PE on EGA andApiezon columns, and Figures 10 and 11 showthem of PS on similar columns.

Noteworthy features of each phosphatide are asfollows. a) PE. The main saturated fatty acid

18:0 DMA

PLATELET PE 16:0 DMAEGA on Gas Chrom-P 18:0173F3xlO-$ Amps.

16:0

0 10 20 30 40 50 60

20'4

8:2 20:0

70 80 90 150 160 170MINUTES

FIG. 8. TRACING OF GAS-LIQUID CHROMATOGRAMSHOW-ING FATTY ACIDS AND ALDEHYDESOF PLATELET PHOSPHA-TIDYL ETHANOLAMINE ON ETHYLENE GLYCOL ADIPATE(EGA) POLYESTER (POLAR). Note the large dimethyl-acetal (DMA) peaks preceding the 16:0 and 18:0 fattyacids.

was stearic, 15 moles per cent, and the main un-saturated acid was arachidonic, 36 moles per cent.The fatty aldehydes of platelet phospholipids wereconfined mainly to PE and almost all were of thesaturated straight-chain 16 :0 and 18:0 type. The18:0 aldehyde was twice the amount of 16:0 alde-hyde. Twenty- and 22-carbon aldehydes were notfound, but one type of branched 17-carbon satu-rated aldehyde (7) was noted on Apiezon. It issignificant that the molar ratio of saturated fattyacids plus fatty aldehydes to unsaturated fatty acidsapproached unity in the ethanolamine phospholipidclass (12) (Table IV). b) PS. The main satu-rated fatty acid was stearic, 46 moles per cent, andthe main unsaturated fatty acids were oleic, 21moles per cent, and arachidonic, 25 moles percent. The aldehydes found were principally of thesaturated straight-chain 16 :0 and 18 :0 type, butwere present in very small amounts. As in PE,the ratio of saturated fatty acids plus fatty alde-hydes to unsaturated fatty acids approached unity(Table IV). c) PI (with PS). The qualitativefatty acid and aldehyde composition of thesefractions was like PS in that stearate, oleate,and arachidonate predominated, but there wasmore arachidonate and less oleate. Although smallamounts of PS were present in these fractions, the

2206


PLATELET PEAPIEZON L on ANAKROMAB pH 7.4

1971C 3x10 Amps.

16:0

50 60 70 80

90 100 120 120 130 140MINUTES

FIG. 9. FATTY ACIDS AND ALDEHYDESOF PLATELET PHOSPHATIDYL ETHA-NOLAMINE (PE) ON APIEZON L (NONPOLAR). Arachidonic was the mainfatty acid of PE.

fatty acid pattern was felt to be representative ofPI. d) Lecithin. Large amounts of palmitic acid,34 moles per cent, were found in lecithin. Themajor unsaturated fatty acid was oleic, 31 molesper cent. Aldehydes of type 18:0 were not foundand there was less than 1 mole per cent 16:0 al-dehyde. Lecithin differed from PE in containingmore palmitic and oleic but less arachidonic acid.There were marked quantitative differences be-

tween PS and lecithin with regard to palmitate,stearate, and arachidonate. Results with samplesof PE and PS methyl esters chromatographed on

Apiezon columns at the Rockefeller Institute were

qualitatively and quantitatively similar to thosereported here (Table III).

The dimethylacetals in PE comprised about 33moles per cent of the total fatty chains. If thealdehydogenic group is linked to the a' (C-1)

PLATELET PSEGA on Gas Chrom-P173 3xIO10 Amps.

18:0

18:1

--

o 0 20 30 40 50 60

20:4

? 20:0 20:1

70 80 90 100 110 140 150 160 170MINUTES

FIG. 10. TRACING OF GAS-LIQUID CHROMATOGRAMOF FATTY ACIDS ANDALDEHYDESOF PLATELET PHOSPHATIDYL SERINE (PS) ON ETHYLENEGLYCOLADIPATE (EGA) POLYESTER. Very little plasmalogen (DMA) was presentin the PS fractions. PS contained large amounts of stearate.

0

8:1 DfMA

2207


PLATELET PS

APIEZON on ANAKROMAB

197' I x0-6 Amps.

16:1 16:0

18:1

18:2

I .. 20 30 4 5 6 0 0 90 I O 20 30 40 50 60 70 80 90

20:4

18:0 DMA 20:3

100 120 140

20:11 a

160MINUTES

FIG. 11. TRACING OF GAS-LIQUID CHROMATOGRAMOF PLATELET PHOSPHATIDYL SERINE (PS) ON APIEZON L.As in phosphatidyl ethanolamine, arachidonic was the main unsaturated fatty acid.

carbon atom of glycerol and an unsaturated fattyacid is linked to the (C-2) carbon as an acylester (see Discussion), then PE is 66 per cent inthe plasmalogen form. This agrees with resultsobtained by iodine addition, in which PE was

found to contain 65 per cent plasmalogen. Iodineaddition showed 5 per cent of the choline phospha-tides to be plasmalogen.

DISCUSSION

An important advance in lipid chemistry is thetechnique of quantitative isolation of PE sepa-

rately from PS in a high degree of purity (3, 9).It was possible to recover these lipids in relativelysubstantial amounts from platelets. Molecular ex-

tinction coefficients were determined at 235 mtu, theregion of conjugated diene hydroperoxide absorp-tion, and at 275 m/L, a region which shows other

products of auto-oxidation (21 ). The PE andPS preparations were essentially free of oxidationproducts, and findings were similar to those previ-ously reported by Rouser and associates (3). Inour experience, other techniques for the separa-

tion of PE from PS such as Folch fractionations or

single silicic-acid column runs invariably resultedin mixtures of lipid classes as well as oxidationproducts.

PI was eluted as a sharp peak, but thin-layer andpaper chromatography of concentrated samplesshowed the presence of traces of PS. Similaridentification procedures revealed lecithin fractionsto be pure. The sphingomyelin peak containedtraces of lecithin. Although lecithin and sphingo-myelin can now be separated (9), the procedurewas not used in this study. The over-all patternof run A was similar to that obtained with red cell

TABLE IV

Molar composition of fatty chains of platelet phosphatides*

A B CSaturated Fatty Unsaturated

fatty acids aldehydes fatty acids A/C A+B/CPlatelet

phosphatide EGA Apiezon EGA Apiezon EGA Apiezon EGA Apiezon EGA Apiezon

PE 21.7 t 17.2 28.7 t 33.6 49.2 49.3 0.44 0.35 1.02 1.03PS 50.4 49.5 1.3 0.5 48.4 50.2 1.04 0.99 1.07 1.00Lecithin 50.7 50.1 0.2 0.8 49.1 49.3 1.03 1.02 1.04 1.03

* Abbreviations as in Table I, and EGA=ethylene glycol adipate.t The figure for saturated fatty acids is probably too high and that for fatty aldehydes too low, since 18: 1 dimethylacetal has a retention time

similar to that of stearate on EGA.

20:0

ISO 200

2208

I


phospholipids and phosphatides of human plasmaby Hanahan, Watts, and Pappajohn (10). Thin-layer chromatography was valuable for rapid de-tection of contaminating lipids and oxidation prod-tucts. This procedure, in combination with silicic-acid paper chromatography and the fatty acid pat-terns on gas-liquid chromatography was sufficientfor evaluation of completeness and purity of theseparations.

Since adequate amounts of platelet PE separatedunder nitrogen were available to us for the firsttime, an intensive effort was made to study its prop-erties in the TGT, a system that has been foundsensitive and specific for evaluating clot-promotingproperties of phospholipids. This phosphatide isinert in the TGT, even if adequately suspended.Indeed, it is probable that activity in this test withPE preparations is an indication that they con-tain trace amounts of PS (2). This finding hasalso been reported by Slotta and Powers, whostudied brain "cephalin" fractions, separated bythe Folch technique (22, 23). In contrast to itsinert behavior in the TGT, PE was effective inshortening the recalcified clotting time and activein other test systems (13, 24).

Wallach, Maurice, Steele, and Surgenor (25)separated PE from egg yolk, and in the unoxidizedstate it showed good clot-promoting activity inthe thrombin generation test. These authorsstressed that coagulation properties of PE wereassociated with micelles of a limited size and sur-face configuration, dependent upon the presence ofa high percentage of polyunsaturated fatty acids.In this important study, the investigators havepointed out that information derived from colloidchemistry of soaps seems to be applicable to bloodcoagulation. Although these findings have addeda great deal to our knowledge, it remains for fu-ture research to determine whether the earlyphases of blood coagulation, involving phospho-lipids and plasma proteins, can be attributed mainlyto colloidal properties of the phosphatides. It isimportant to mention that human platelet PE isnot comparable to egg PE with special referenceto two points: a) platelet PE is 66 per cent in theplasmalogen form, whereas egg yolk contains noplasmalogen, and b) on the basis of the fatty acidcomposition given for egg yolk PE (25), a sub-stantial degree of diunsaturation must be postu-

lated, but from present evidence, the existence ofsuch molecules appears unlikely in platelet PE.

PS fractions were active in the TGT to the sameextent as previously noted (1, 2). Our observa-tions and those of other investigators were con-firmed (22, 26-28). It would have been of in-terest to study the clotting activity of PI, but con-tamination with PS was unavoidable. In TGTexperiments with PI isolated from peas,8 the ma-terial was inactive. The results with platelet andbrain PI, containing small amounts of PS, werestriking in that these showed excellent activity inboth the TGT and recalcified clotting times.

Lecithin was inactive in the TGT and recal-cified clotting time. Gas-liquid chromatographicstudies indicated that it contained approximately a1: 1 molar ratio of saturated and unsaturated fattyacids, and traces of aldehyde. This may indicatethat the presence of a 50 per cent molar content ofunsaturated fatty acids is not the only requisite forcoagulant activity of phospholipids.

PE and PS consistently and uniformly shortenedthe recalcified clotting time of platelet-poor plasma.The results with PE are in excellent agreementwith those of Rouser, White, and Schloredt (13)as well as with those of experiments performedwith these phosphatides in a collaborative studyby Ferguson, Marcus, and Robinson (24). Inour hands, the PS was more active than PE. Onthe basis of weight, however, the total plateletlipid extract and brain "cephalin" preparationswere even more active in the recalcified clottingtime.

In earlier experiments with the TGT (1, 2),PS fractions were many times more active thantotal lipid extracts on the basis of weight. It wassurprising that PS, present as about 4 per cent ofthe total platelet lipid, showed only a fivefold in-crease in activity over total platelet lipid controls.This result was less than expected if all the activityin platelets were due to PS. There are two pos-sible explanations: a) purification procedures mayexert some unknown effect to diminish the clottingactivity of PS, and b) the other platelet lipids, inthemselves inactive, may enhance the PS activity.Such findings may also indicate that under physio-logical conditions, all the platelet phospholipids areresponsible collectively for thromboplastic activity.

8 Provided by Dr. A. C. Wagenknecht.

2209


When the purified phospholipids were stored incyclohexane under nitrogen, their clotting activitywas retained for long periods. Even under thesecircumstances, however, they eventually showedspectrophotometric evidence of oxidation, and ac-tivity in the TGT was markedly diminished. Inthe Ferguson two-stage clotting test (24), theactivity of PS seemed to improve after oxidativechanges took place, and a similar phenomenon hasbeen observed in other coagulation systems (25).

When difficulty was encountered in gas-liquidchromatographic analysis of dimethylacetals onnonpolar columns, the initial impression was that"acid sites" in the inert supports were responsiblefor hydrolytic changes, despite the fact that theywere in a neutral pH range after acid and alkaliwashings. A systematic study of available gas-liquid chromatographic supports showed that un-coated packings that were made alkaline to a pHof 7.5 before application of the stationary phasegave good results. Pretreatment of stationarysupports with alkali was first suggested by James(19), and this procedure has been in common usefor chromatography of methyl esters and fattyaldehyde dimethylacetals on Apiezon columns (6).We also found that alkali washing of stationarysupports improved the results with EGAcolumns.It has become apparent from gas-liquid chromato-graphic studies on aliphatic diamnines (29) thattreatment with alkali is not merely neutralizationof "acid sites," but a more complicated phenonme-non, not yet understood. It has been emphasizedthat gas-liquid chromatographic analyses, espe-cially of plasmalogens, ought to be carried out ontwo types of columns, polar and nonpolar (7), andit seems desirable to confirm the accuracy of thismethod of plasmalogen calculation with other tech-niques such as iodine addition ( 14), or the p-nitro-phenylhydrazone reaction (5). It was advan-tageous to evaluate column performance and de-tector response by comparison of results with thoseobtained on the same sample with another gas-liquid chromatographic apparatus of establishedlinearity (18). Our quantification of platelet PEplasmalogen is 4 to 5 times higher than results re-ported in a recent publication by Blomstrand,Nakayama, and Nilsson (30). These authorsused the hydrolysis technique of Dawson (31),who had previously reported values for humanerythrocyte plasmalogens considerably lower than

those found by Farquhar (12). These differencesincrease the need for confirmatory methods ofdetermining plasmalogen.

Certain characteristics of the fatty acids andaldehydes that are distinctive for the various typesof platelet phosphatides are worthy of comment,but their relevance to the biological properties ofthese phospholipids is not yet known. The mainunsaturated fatty acid in PE and PS was arachi-donic, as previously reported (4) when the "cepha-lins" were studied as a group, but now the 20: 4distribution is better defined. Although stearicwas the main saturated fatty acid in PE, it ac-counted for less than 20 moles per cent of the fattychains, whereas about half of the fatty acid com-ponents of PS were stearate. The fatty acids oflecithin showed substantial quantitative differencesfrom PE and PS. Palmitic was the main satu-rated and oleic the main unsaturated fatty acid.The amount of oleate was increased over its levelin the PE and PS fractions. Although arachidonicacid was low in the choline phosphoglycerides,more was present than originally reported (4)when separations were carried out under lessstringent conditions with regard to oxidativechanges.

Recently, important studies on the structure orphosphoglycerides have been reported by Mari-netti, Erbland, and Stotz (32), Debuch (33), Tat-trie (34), and Hanahan, Brockerhoff, and Barron(35). They indicated that in the plasmalogenand diacyl forms of PE and lecithin, the a' (C-1)carbon of glycerol was linked to either a saturatedfatty acid or an aldehyde, and the , (C-2) carbon,to an unsaturated fatty acid. In platelet PE, themolar ratio of saturated fatty acids plus aldehydesto unsaturated fatty acids was unity (Table IV),as also for PS and lecithin. Farquhar has re-ported similar ratios in the fatty chains of erythro-cyte phospholipids (12). These findings furnishindirect evidence that the fatty chains of humanplatelet and erythrocyte phospholipids are arrangedas in other tissues (32-35). In general, there aremany similarities between red cells (12) andplatelets with regard to plasmalogen content andfatty acid composition. Farquhar found erythro-cyte PE to be 67 per cent plasmalogen and 66 percent of platelet PE is in this form. The main redcell fatty acids were similar to those of platelets inthe separated "cephalin" and choline classes.

2210

PLATELEST PHOSPHATIDES

Recently, the large percentage of plasmalogenin human erythrocytes (12) and human myelin(36) led to speculation that these lipids mightshare common functions in cell plasma membranes.Since platelet phospholipids are similar to red cellsand myelin, they may also be considered as largelyderived from the plasma membrane. It will be im-portant, however, to know the lipid compositionof platelet organelles such as mitochondria beforedefinite conclusions can be drawn. Further in-vestigation will also be required to determinewhether analogies of structure and function arepossible in tissues showing the same membranelipid constituents. Generalizations about the re-lationship of biological activity to lipid compositionmust be reconciled with differences in phospha-tide make-up between the same tissues of variousanimal species (37, 38). As Gray and Macfarlanehave demonstrated, investigations in comparativebiochemistry indicate the need to avoid prematurecorrelations of this nature (37, 38).

Although the function of plasmalogens is notknown, speculations have been made (12, 39).Participation in blood coagulation is possible, butearlier investigations (5. 13) have given evidenceagainst such activity.

SUMMARY

1. Two types of column chromatography, silicic-acid and ammoniated silicic-acid, were used to frac-tionate the phospholipids of human blood plateletsand obtain highly purified phosphatidyl ethanola-mine (PE) and phosphatidyl serine (PS). Allseparation procedures were carried out under ni-trogen, thus minimizing oxidative changes. Fattyacid and plasmalogen concentrations were meas-tured by gas-liquid chromatography on polar andnonpolar columns.

2. A total lipid extract of 400 mg contained 25 to35 mg PE and 13 to 16 mg PS.

3. There were distinct differences in the fattyacid and fatty aldehyde composition among theseparated platelet phosphatides. Stearate, oleate.and arachidonate predominated in the "cephalin"class, while the main fatty acids of lecithin werepalmitate and oleate. Plasmalogen was foundmainlv in the ethanolamine phospholipid group.Sixty-six per cent of platelet PE was in the plas-malogen form.

4. Both PE and PS shortened the clotting timeof recalcified platelet-poor plasma, but only PSwas active in the thromboplastin generation test.Phosphatidyl inositol fractions contained smallamounts of PS and showed marked thrombo-plastin-promoting properties in both tests.

5. Similar studies carried out on human brainphospholipids gave comparable results.

ACNOWLEDGMENTS

The authors are indebted to Dr. J. WV. Farquhar forhelpful discussions during the course of this work and toDr. T. H. Spaet for aid in preparation of the manuscript.

REFERENCES

1. Marcus, A. J., and Spaet, T. H. Platelet phospha-tides: their separation, identification, and clottingactivity. J. clin. Invest. 1958, 37, 1836.

2. Marcus, A. J., Ullman, H. L., and Wolfman, M.Studies on blood platelet phospholipids. J. LipidRes. 1960, 1, 179.

3. Rouser, G., O'Brien, J., and Heller, D. The separa-tion of phosphatidyl ethanolamine and phospha-tidyl serine by column chromatography. J. Amer.Oil chem. Soc. 1961, 38, 14.

4. Marcus, A. J., Ullman, H. L., and Ballard, H. S.Fatty acids of human platelet phosphatides. Proc.Soc. exp. Biol. (N. Y.) 1961, 107, 483.

5. Zilversmit, R. D., Marcus, A. J., and Ullman, H. L.Plasmalogen in human blood platelets. J. biol.Chem. 1961, 236, 47.

6. Gray, G;. M. The separation of the long chain fattyaldehydes by gas-liquid chromatography. J. Chro-mat. 1960, 4, 52.

7. Farquhar, J. W. Identification and gas-liquid chro-matographic behavior of plasmalogen aldehydesand their acetal, alcohol, and acetylated alcoholderivatives. J. Lipid Res. 1962, 3, 21.

8. Hirsch, J., and Ahrens, E. H., Jr. The separationof complex lipide mixtures by the use of silicicacid chromatography. J. biol. Chem. 1958, 233, 311.

9. Rouser, G., Bauman, A. J., Kritchevsky, G., Heller,D., and O'Brien, J. S. Quantitative chromato-graphic fractionation of complex lipid mixtures:brain lipids. J. Amer. Oil chem. Soc. 1961, 38,544.

10. Hanahan, D. J., Watts, R. M., and Pappajohn, D.Some chemical characteristics of the lipids ofhuman and bovine erythrocytes and plasma. J.Lipid Res. 1960, 1, 421.

11. Bartlett, G. R. Phosphorus assay in column chro-matography. J. biol. Chem. 1959, 234, 466.

12. Farquhar, J. W. Human erythrocyte phospho-glycerides. I. Quantification of plasmalogens.fatty acids, and fatty aldehydes. Biochim. biophys.Acta (Amst.) 1962, 60, 80.

221


13. Rouser, G., White, S. G., and Schloredt, D. PhOs-pholipid structure and thromboplastic activity. I.The phosphatide fraction active in recalcified nor-mal human plasma. Biochim. biophys. Acta(Amst.) 1958, 28, 71.

14. Rapport, M. M., and Alonzo, N. F. The structureof plasmalogens. V. Lipids of marine invertebrates.J. biol. Chem. 1960, 235, 1953.

15. Nelson, G. J., and Freeman, N. K. The phospholipidand phospholipid fatty acid composition of hu-man serum lipoprotein fractions. J. biol. Chem.1960, 235, 578.

16. James, A. T. Personal communication.17. Gray, G. M. Personal communication.18. Farquhar, J. W., Insull, W., Jr., Rosen, P., Stoffel,

W., and Ahrens, E. H., Jr. The analysis of fattyacid mixtures by gas-liquid chromatography: con-struction and operation of an ionization chamberinstrument. Nutr. Rev, 1959, 17, suppl., 1.

19. James, A. T. Qualitative and quantitative determina-tion of the fatty acids by gas-liquid chromatog-raphy in Methods of Biochemical Analysis, D.Glick, Ed. New York, Interscience, 1960, vol. 8,p. 1.

20. Marinetti, G. V., Erbland, J., and Stotz, E. Theseparation of ionic forms of phosphatidylserine bycolumn chromatography. Biochim. biophys. Acta(Amst.) 1958, 30, 41.

21. Lea, C. H. Some observations on the preparation andproperties of phosphatidylethanolamine in Bio-chemical Problems of Lipids, G. Popja'k and E.LeBreton, Eds. New York, Interscience, 1956, p.81.

22. Slotta, K. H. Thromboplastin. I. Phospholipidmoiety of thromboplastin. Proc. Soc. exp. Biol.(N. Y.) 1960, 103, 53.

23. Slotta, K. H., and Powers, J. K. Determination ofethanolamine and serine in phospholipids. Fed.Proc. 1962, 21, 293.

24. Ferguson, J. H., Marcus, A. J., and Robinson, A. J.Purified platelet phospholipids and blood coagula-tion. In preparation.

25. Wallach, D. F. H., Maurice, P. A., Steele, B. B.,and Surgenor, D. M. Studies on the relationshipbetween the colloidal state and clot-promoting ac-tivity of pure phosphatidylethanolamines. J. biol.Chem. 1959, 234, 2829.

26. Therriault, D., Nichols, T., and Jensen, H. Purifi-cation and identification of brain phospholipidesassociated with thromboplastic activity. J. biol.Chem. 1958, 233, 1061.

27. Troup, S. B., Reed, C. F., Marinetti, G. V., andSwisher, S. N. Thromboplastic factors in plate-lets and red blood cells: observations on theirchemical nature and function in in vitro coagula-tion. J. clin. Invest. 1960, 39, 342.

28. Barkhan, P., Silver, M. J., and OKeefe, L. Thelipids of human erythrocytes and platelets andtheir effects on thromboplastin formation in BloodPlatelets, S. A. Johnson, R. W. Monto, J. W. Re-buck, R. C. Horn, Jr., Eds. Boston, Little, Brown,1961, p. 303.

29. Smith, E. D., and Radford, R. D. Modification ofgas chromatographic substrates for the separationof aliphatic diamines. Anal. Chem. 1961, 33, 1160.

30. Blomstrand, R., Nakayama, F., and Nilsson, I. M.Identification of phospholipids in human thrombo-cytes and erythrocytes. J. Lab. clin. Med. 1962,59, 771.

31. Dawson, R. M. C. A hydrolytic procedure for theidentification and estimation of individual phos-pholipids in biological samples. Biochem. J. 1960,75, 45.

32. Marinetti, G. V., Erbland, J., and Stotz, E. Thestructure of beef heart plasmalogens. J. Amer.chem. Soc. 1959, 81, 861.

33. Debuch, H. Uber die Stellung des Aldehyds imColaminplasmalogen aus Gehirn. Hoppe-SeylersZ. physiol. Chem. 1959, 314, 49.

34. Tattrie, N. H. Positional distribution of saturatedand unsaturated fatty acids on egg lecithin. J.Lipid Res. 1959, 1, 60.

35. Hanahan, D. J., Brockerhoff, H., and Barron, E. J.The site of attack of phospholipase (lecithinase) Aon lecithin: a re-evaluation. Position of fatty acidson lecithins and triglycerides. J. biol. Chem. 1960,235, 1917.

36. Webster, G. R. Studies on the plasmalogens ofnervous tissue. Biochim. biophys. Acta (Amst.)1960, 44, 109.

37. Gray, G. M. The phospholipids of ox spleen withspecial reference to the fatty acid and fatty aldehydecompositions of the lecithin and kephalin frac-tions. Biochem. J. 1960, 77, 82.

38. Gray, G. M., and Macfarlane, M. G. Composition ofphospholipids of rabbit, pigeon and trout muscleand various pig tissues. Biochem. J. 1961, 81, 480.

39. Rapport, M. M. Plasmalogen structure and thechemical reactivity of a,fi-unsaturated ethers inThe Biology of Myelin, S. R. Korey, Ed. NewYork, Paul B. Hoeber, 1959, p. 282.

2212

phosphatidyldm5migu4zj3pb.cloudfront.net/manuscripts/104000/104679/jci62104679.pdf · cent methanol...

Documents