Membranes of Animal Cells

Surface membranes were isolated by the fluorescein mer:uric acetate and Tris procedures from mouse fibroblasts (L :ells) grown in suspension culture. Membranes isolated by :he fluorescein mercuric acetate method contained approxinately 4.7% of the protein and 13.8 % of the lipid of the whole :ell. Approximately 95 % of the plasma membrane lipid was accounted for. Comparison of lipid-protein ratios for whole :ells and surface membranes shows a d-fold enrichment of ipid in the surface membrane. Neutral lipids make up 40 % )f the total membrane lipid, with cholesterol and triglycerides accounting for 20 and 13%, respectively. The high choleserol content of the surface membrane is reflected in a :holesterol-phospholipid molar ratio of 0.69 while the cor,esponding ratio for total cell lipid is 0.26. Fifty-nine per :ent of the membrane phospholipids contain choline, and the ;urface membranes contain more sphingomyelin and less ecithin and phosphatidylethanolamine than the total cell ipid. Phosphatidylserine and sphingomyelin show a Zto L-fold enrichment of unsaturated and long chain fatty acids ,elative to other membrane phospholipids. The lipid comjosition of the surface membranes of L cells was compared o the composition reported in the literature for surface memcranes of animal cells. The comparison indicates that the ipid compositions of L cell and liver cell surface membranes Lre basically similar. Lipid composition may be useful in lassifying surface membrane preparations.

were isolated by the fluorescein mer-:uric acetate and Tris procedures from mouse fibroblasts (L :ells) grown in suspension culture.
Membranes isolated by :he fluorescein mercuric acetate method contained approxinately 4.7% of the protein and 13.8 % of the lipid of the whole :ell. Approximately 95 % of the plasma membrane lipid was accounted for.
Comparison of lipid-protein ratios for whole :ells and surface membranes shows a d-fold enrichment of ipid in the surface membrane.
Neutral lipids make up 40 % )f the total membrane lipid, with cholesterol and triglycerides accounting for 20 and 13%, respectively.
The high choleserol content of the surface membrane is reflected in a :holesterol-phospholipid molar ratio of 0.69 while the cor-,esponding ratio for total cell lipid is 0.26. Fifty-nine per :ent of the membrane phospholipids contain choline, and the ;urface membranes contain more sphingomyelin and less ecithin and phosphatidylethanolamine than the total cell ipid. Phosphatidylserine and sphingomyelin show a Z-to L-fold enrichment of unsaturated and long chain fatty acids ,elative to other membrane phospholipids.
The lipid comjosition of the surface membranes of L cells was compared o the composition reported in the literature for surface memcranes of animal cells.
The comparison indicates that the ipid compositions of L cell and liver cell surface membranes Lre basically similar.
Lipid composition may be useful in lassifying surface membrane preparations.
Iierent ljrogress in our understanding of the structure and unction of cell rncmbrallev has come from both physical and hemica data. Many of the physicochemical properties of rtificial binlolecular lipid membranes are similar to those of ~;~tural membranes (1). Studies on the interaction of lipids * This work TV~S supported by United States Public Health iervice Grants 1.ROllID-005201 and IIE-05285 and by American [~IKY~ Society Grant PNP-2S. and proteins have indicated that the organization of metnbrane structure may be dependent upon the lipid composition of metnbranes (2). Correlation of data from the study of model . . systems an vztro and of natural membranes can be achieved only when precise analytical data on membrane components are available.
The wide variation in the lipid composition of membranes (3) is l)robably important in determining the properties and functions of a particular membrane.
In this respect, the markedly different cholesterol contents of the surface membranes and internal mcmbrnncs of xnirnnl tissues (4)(5)(6) may indicate different affinities for the vnrious types of lipids involved in plasma membrane formation.
Until recently, studies on lipids of surface membranes lrere limited to t'he erpthrocyte membrane, which could be readily isolated and purified in large quantities.
Emmelot et al. (7) prepared a fraction from rat liver which consisted of lhs~n:~ mctnbrane fragrnent,s, but the lipids ivere only partially characterized.
Takeuchi and Terayarn;~ (8) and Skipski et crl. (0) attempted a complete characterization of the lipid components of similar preparations of mernbranes from rat liver. Both investigations were complicated by technical dificulties which do not alloy direct comparison with the more yuantitative studies of T)od and Gray (10) and Pfleger, Anderson, and Sn>-der (11).
Warren, Glick, and Nash (12) hare described several n&hods for the isolation of surface membranes from mouse fibroblasts (L cells) grown in suspension culture.
Preparations of intact surface nlelllbranes are obtained from as much as SOYi of the cells grown under controlled environmental conditions. This study describes the lipid composition of these surface metnbranes and makes a comparison n-ith the lipid composition of the whole cell and with analyses reported by others of the lipids of plasma, membranes of normal animal tissues. -1 preliminary report of this work has been presented (13).
Surface nlenlbrnnes were isolated Iron1 1, cells by the FhIA' and Tris methods of Warren et nl. (12). SLWOXY was removed fro111 the menlbrane preparations b\-dialysis on :I rocking stage for 24 hours at 4" against a flow of distilled water.
L cells were grown and harvested as previously described (12).
Pivc volunles of cold nlethanol were added to the cells or membranes in a glass centrifuge tube. The extraction proceeded for 15 min at 0" with occasional stirring; an equal volunle of cold chloroform was added to the tube, and stirring lvvas continued for 15 min. The lipid extract, recovered by centrifugation, n-as kept under an atmosphere of nitrogen at -20".
The cell or membrane pellet n-as re-estracted at O", then extracted twice at room temperature, and finally at 45" under a flow of nitrogen gas. Conipletenws of litlid extraction was determined by hydrolysis of the residrw n-ith 6 s II('1 at 100" for 16 hours followed b\-gas chrolllntogr:lphic estimation of the amount of fatty acids released.
Removal of Nonlipid Jf afcric~l&Two diff went procedures for removal of nonlipid material from L cell lipid extracts were used with comparable results.
In tht first procedure, the lipid extracts were washed with 0.2 volunle of water at, 4". The two phases were allowed t,o separate completely, and the chloroform layer was removed and rewashed with 0.1 volume of n-ater. Only trace amounts of phosph:ltidylseriIle n-erc found in the aqueous mcth:tnolic mashes. In the second procedure, the unwashed lipid extract xas dried under reduced pressure at 10" on a Buchi Rotavapor (Fisher), and the damp residue was dissolved in chloroforn-methanol (19: 1, v/v) saturated n-it,h Ivater and apl)lied to a column of Sephades G-25 (1 x 18 cm) at 4" for quantitative separation of lipids from nonlipids (14). Purification of the lipid extracts by Sephades chromatography resulted in 99.8y0 recovery of lipid and greater than 997; removal of added '4C-glucose and '4C-lcucine.
FMA rnembrane lipids were purified by this technique to eixurc complete remora1 of sucrose.
Completeness of separation 17as dcternlinctl by-thin la!-er chromatogr:aI)hy of the nclltral lipid fractions with the double development technique of Freeman and T\.est (19) Polar lipids lvere separated by two-dimensional thin l:Lyer chromatography according to the method of Abramson and I%lecher (20). Spots were visualized by brief exposure to iodine vapor or by spraying with alkaline bromthymol blue and esl)osing the 1)l:ltcs IO ammonia vapors (21). Purified lipids (hpplied Science Laboratories) and rat liver lipid which was freshly extracted were used as reference standards.
Total lipid was determined b>-the charring llirthod of 1Iarsh and J\-einsteill (23). After the initial analyses had determined the composition of n-hole cell and surfacc membrane lipids, caharring standards of identical composition were l)repared n ith mixtures of l'urified lipids.
These standards were used ill subsequent analyses for determination of total lipid recover)-.
Seutrnl lipids were determined after Florisil chromatography by the charring method, with a purified lipid of the same class as II stalldurd.
Cholesterol and cholesterol esters lvere also determillcd 1)~. tho method of Zlatkis,Zak,and Boyle (24); glyceritlcs 1,~. tlw method of Van Handel and Zilrersmit (25). l%ospholil)ids \~ert' determined by scraping the lipid spots off the Silica Gel (: plates and transferring the gel to digestion tubes; 0.15 1111 of concentrated sulfuric acid na+ added, and digestion was carried out in an aluminum he:Lting block at 220" for 90 min. l'hos~~horus content n-:rs measured by the method of Bartlett (26) and phospholipid content n-as calculated by multiplying the phosphorus value by 25. Plasn~nlogenbound aldehydes lrere detrrmined by the p~nitrophen!-lh!-crazine method of l'ries and Biittcher (2i). Sphingolipid base< were isolated bg methnnolysk of sphingomyelin according to the method of Rosenberg and Stern (28). The long chain bases were separated and nleasured by the gas chromatographic method of Carter and Gaver (29). Lipid fraction> were stored in deaerated hesane at -20" under a nitrogen atmosphere.
All analyses \T-ere performed in duplicate.
F&y _ I cid and _ I ldehyde k-nrrlysis-Fatty acid comlwsition was determined by two methods.
(n) The sample was hy drolyzed with constant-boiling HCl in sealed tubes unde, nitrogen for 12 hours at 100". Fatty acids were extracted n-it1 hesnne, washed n-ith water, and drird.
Methyl esters mere prepnrcd by treatment of the free fatty acids with freshly distiller by guest on January 13, 2018 http://www.jbc.org/ dinzomethane generated by alkaline treatment of IL--nitroso-N-FlZIh nlny be a reflection of small anlounls of free mercur\-which methyl urea (30). (b) Lipid s:m~ples were tran~estcrified with niay be fornird from the reagent during lipid storage. BFs-methanol by the procedure of Xlorriaon and Smith (31). Fatty aldeh!.dc dimethglacetals n-we prepnrcd lvith 13F3-RESULTS methanol and were freed of fatty acid methyl rsters by the Lipid E&tactions-The lipids of the L cell and surface mensaponification technique of Farquhar (32). lIethy esters and branc preparations were completely extracted with chloroformdimethylacetals were analyzed by gas-liquid chromatograph\-on methanol, 1: 1 (v/v). Phospholipid recovery from L cells was a l'!.(x argon chromatograph equipped with :I strontium ioniza-3 to 5(,', greater with this solvent than with chloroform-methanol, tion detector. A &foot column of 17.5(1 polyethylene glgcol 2: 1 (v/v), or alcohol-ether tnistures.
The initial extractions ?uccinate on 80 to 100 mesh (;:I>-('hrom I' (A\pplied Science were performed in the cold to minimize lipid osidation and Lnboratories) operated at 165" and 7.5 pounds of pressure of activation of any hydrolytic enzymes ljresent in the samples. argon TV:W uccd. Cnanturated fatty acids and long chain fatty Esantination of the h!-drolyzed cell and membrane residues acid> lvere also identified b) gas-liquid chrotnatograph\-on a revelled the presence of 1 to 2 pg of fatty acid per mg of protein.
-l-foot column of loi, Apiezon L on 100 to 120 mesh C'elite.
These trace amounts of pnlmitic and stearic acid are not solvent-Fatty acid nlcth\-1 ester identification and detector response estractable and may represent fatty acids col-alently linked to wrc dctermincd with S.T.H. Type Misturc KD and KF (Xl)-protein (33). The use of lipid mistures as standards for the plied Science Laboratories), total lipid determination was tested by preparing a grnrimetric Contakzafion of Lipid Extracts with Ileavg Jletccls and Lipid standard of L cell lipid and a lipid mixture of the same composi-;lu/oxidafior2-Atternl)ts to analyze quantitatively the lipids tion. The charring densities of equal quantities of the t1v-o of pln~ma membranes prel)ared by the %nClt l)rocedrtre of Warren standards agreed within +1.552. The total recovery of lipid et nl. (12) failed because of the presence of quantities of zinc in from I, cell and surface membrane lil'id extracts afler separation the lil)id esttwt sufficicllt to C:IUSC major changes in the chro-and measurement by the methods described above can be matographic propeltics of the l)olar lipids. The heavy tnetal readily estimated from the total lipid value calculated from the could not be removed by Scphadcs chloinat,ograp2i?-.
The use charring assay. Uy this nlcthod 95.7 =t 0.4Gj, of the lipids of of an orgaltometallic agent (FJI;\) in the isolation of surface the whole cell and 94.1 + 0.45, of the surface membrane lipids membranes requires special precautions to l)rcvellt lipid autosida-can be accounted for. tion due to the presence of heavy metals. Only small amounts Neufral Lipid Compositio~l~Table I su~nmarizes the data on of nlembrane-bound FI\IA could be solubilized by chloroform-the lipid composition of five samples of whole L cells, five samples methanol (1 :l) giving the lipid extract a yelloworange tint. of surface membranes isolated by the FMA procedure, and one Sephades chromatography of the extract did not remove the sample of surface membranes prepared by the Tris procedure. FJIA.
Subsequent silicic acid chromatography gave an FMA-L cells grown in medium containing 107; fetal bovine serum were free neutral lipid fraction, but the phospholipid fraction found to contain 0.225 mg of lipid per mg of protein.
The addition of 2, B-di-tert-per cent of the cell lipid is phospholipid; cholesterol and glycer-but\-l-p-cresol (5 pg per mg of lipid) to the membrane phos-ides comprise 11 and 8y0, respectively, of the remaining lipid. pholipid fraction completely protected the lipid; when it was Free cholesterol accounts for half of the neutral lipid of the cell, omitted, extensive autoridativc changes were observed similar triglycerides an additional 30%, and cholesterol esters only 2% of to those reported by Dodge and Phillips (17)  ww extracted, and the lipids were esomined and compared with Lecithin (43';;) and phosphntidylethanola~~li~~e (23 '%) make up untreated cell lipids. FRlh had no effect on the fatty acid, the bulk of the cell phospholipid and account for more than half aldehyde, neutral lipid, or phospholipid composition of the cells. of the cell lipid. The more acidic phospholipids (phosphatidyl-L cell lipids which remain ill the presence of FXIA and 2,6-di-serine, phosphutidylinositol, and phosphntidic acid) make up 16yo, tert-bllt!-lop-cresol for several days at -20" begin to show in-and cardiolipin less than 1 yc, of the cell phospholipid. The high lysophosphatide content (756) is a chromatogrsphic artifact creased lysolecithin and decreased plnsmalogcn content.
Since since lipid extracts not separated on silicic acid columns do not 0.005 M H&l:! can cleave l)las~~~:~logen-bouiid aldehydc (17)) the contain more than 1 :', lysophosphatide. decrease in plasmalogen content of L cell lipid in the presence of ('omparison of plasma membrane and cell phospholipids re- veals several significant differences. Although the content of the total cell lipid but only 8% of the total surface membrane choline-containing phospholipids is only slightly higher (59 as lipid. The acidic phospholipids account for 13% of both the against 54a/,) in surface membrane, total membrane lipid has cell and membrane lipid, with the membrane containing more almost twice as much sphingomyelin and significantly less phosphatidic acid but less phosphatidylserine than the whole lecithin + lysolecithin (21 as against 36%) compared to the cells. total cell lipid.
Ethanolamine phospholipids make up 22% of Cerebrosides and sulfatides were not detected in L cell lipid extracts and cardiolipin was not detected in plasma membrane preparations. The absence of cardiolipin (a major mitochon-d&l lipid (34)) in the membrane lipid extracts indicates that the plasma membrane preparations are not contaminated with mitochondria. This is supported by the inability to detect cytochromes a, a3, and c in the plasma membrane prcparations. 2 The snxdl amount of lipid phosphorus which does not migrate on thin layer chromntogral)hp plates contains small quantities of inositol (determined by gas-liquid chromatographic analysis) and probably represents traces of 1)olyphosphoinositides as well as SOI~F phospholipid degradation products. X0 major nonphosphorus-containing polar lipids were found by tw-dimensionnl chromatographic analysis. Glycolipids accounted for approximately O.TC/, of L cell lipid.3 Small amounts of ceramide diheroside were found in the polar lipid fraction, and several gangliosides were identified after separation from the lipid estract by Sephndex chromatography or the aqueous washing. Gangliosides were also found in the surface membranes.
Approximately 12 y0 of the L cell phospholipids and 5 71 of the membrane phospholipids occur in the form of plasmalogens. Two-thirds of the cell and rnernbrane plasmalogen is found in the lecithin fraction and the remainder in the phosphatidylethanolamine fraction.
Analysis of the plasmalogen-bound aldehydes of bovine fetal serum and L cells (Table III)  nearly all of the plasmnlogen nldehyde found in the cell. The distribution of aldehydes found in the surface membrane lipid is essentially identical.
Gas chromatographic analysis of the trirnethglsilyl et,her derivatives of the sphingolipid bases of the whole L cell revealed that C18-sphingosine accounted for 59.6$& of the bases. C18dihydrosphingosine (31.07;) and ('20.sphingosine (9.4%) were the only other bases found.
The total lipid and phospholipid composition of surface membrancs prepared by the Tris procedure shows excellent agreement with that of the surface nwnbrnnes prepared by the FN.1 procedure. XIenlbrnne preparations which had been stored frozen in dilute sucrose solutions showed significant loss of lipid to the aqueous medium upon defrosting.
Membranes stored for 2 IT-eeks to several months showed losses of 20 to 45% in cholesterol, lecithin, and I)hosphatid~lethnaol,zlllirle content. Therefore, all of the analytical work reported was performed on freshly prepared membranes.
Fatty Acid Paftcms~Table 1V presents a comparison of the  acids are unsaturated, l\-ith the polyunsaturated acids accounting for 127c of the total.
The fatty acid pattern of the plasma membranes shows several differences.
(a) Palmit'ic, stearic, and oleic acids make up slightly less than 5070 of the total fatt,\-acids, with long chain acids accounting for 147, of the total.
(b) The stearic and oleic acid contents of the membranes are usually half that of the whole cell. (c) Linolenic (18:3) and lignoceric (24:0) acids, xvhich make up only 56 of the cell fatty acids, account for over 357, of the membrane fatty acids. (d) Approximately half of the surface membrane fatty acids are unsaturated, with the polyunsaturated fatty acids accounting for 27 y0 of the total.
In Table V is seen a comparison of the fatty acid composition of the major phospholipids and the total neutral lipid fraction of surface membranes prepared by the PMA procedure.
The follolving features can be pointed out. (a) The neutral lipid fraction contains less long chain fatty acids (7%) than any of the major phospholipids (9 to 32%). (b) The neutral lipid fraction cont'ains less polyunsaturated fatty acids (9$&) than any of the major phospholipids (22 to 61 yO). Arachidonic acid (20:4) is a notable exception since it appears to be concentrated in the neutral lipid fraction.
(c) The major phospholipids can be separated into two groups.
DISCGSSIOS / L cells grown in suspension culture provide an esccilent source of surface membranes for the study of lipid composition. Bailey,Gey,and Gey (3'7) have shown that cultured mammalian cells derive most of the lipids required for growth from the lipids of the medium. Bailey (38) has shown that lipid synthesis de nova is inhibited up to 95'; m L cells grown in the presence of serum lipid.
The operation of metabolic controls of this type may be important in the mnilltenance of the remarkably constant lipid composition of the cells observed throughout the course of this and other investigations.
V:m Ileenen et al. (39) have sho\vn that despite large variations in diet the proportions between the major lipid classes of the erythrocyte membrane did not change, although the fatty acid patterns of the individual membrane lipids were subject to marked alterations. *Uthough 95y0 of the cell lipid can be derived from serum lipid, the composition of the cell 1il)id is markedly different from that of the serum lipid (36). The i;pccific utilization of serum lil)id by the L cell can be readily seen by comparing the fatty acid composition of the serum and the cells (Table IV).
This combination of specific utilization of lipids by the cell and the constant over-all lipid composition allows one to predict that the differences observed between whole cell lipids and the lipids of the isolated surface membrane \vill be significant and no>-represent specaific lipid requirements for surface membrane biosynthesis.
.\t, present there are several rel)orts in the literature \vhicl! supply data on the liljids of plasma nlembranes of rat, liver iso. lated by the variations of the method of Xerille (40) and or membranes isolated from other sources. Table VI presents B summary of the data taken directly or calculated from these reports. The low content 01 cholesterol and phospholipid and the very high lipid-protein ratic probably indicate gross contamination of the preparations witlsmooth endoplasmic reticulum (41). Skipski et al. (Line 4) reported that 40% of the dry weight of rat liver surface membrant was lipid-extractable, but they were unable to account for 26% of the lipid and suggested that this missing lipid was glycolipid although no glycolipid spots were found in the chromatographic analysis.
Dod and Gray (Line 5) have isolated surface mem branes from rat liver and reported a lipid composition which ir in good agreement with the lipid cornposit,ion of L cell surface membranes reported in this paper.
Pfleger et al. (Line 6) haw isolated rat liver l~lasma membranes by tonal centrifugation These preparations h:ave a much lower lipid to protein ratio ant lower neutral lipid and phospholipid content than previousl! Issue of August 10, 1969 Weinstein, .Marsh, Glick, and Warrejz lipid in rat liver membranes (9)(10)(11). L cell membrane neutral lipid contains 51 y0 cholesterol, 3Oa/, triglyceride, and 6% free fatty acid. Skipski et al. (9) found that 45y0 of the total neutral lipid recovered from rat liver membranes was cholesterol and that free fatty acids and triglycerides made up 20 and lo%, respectively, of the remaining neutral lipid. Dod and Gray (10)  the rat liver membrane lipid contained 47 to 59%, phospholipid (Y-11) and the L cell membrane, 59% phospholipid. Cholinecontaining phospholipids account for 48 to 74 70 of rat liver membrane phospholipid (7)(8)(9)(10)(11) and 59% of L cell membrane phospholipid.
Sphingomyelin, which normally makes up only 7% of L cell lipid and about 3C10 of rat liver lipid (9), makes up 14 and 20% (10) of the lipid of the surface membranes of these cells. Both plasma membranes show corresponding decreases in lecithin and phosphatidylethanolamine content compared to whole cell lipid.
Dod and Gray (10) did not find phosphatidic acid in their rat liver membranes; however, Skipski et al. (9) and Pfleger et al. (11) found "polyglycerophosphatide" fractions which account for 7 and 3%, respectively, of the membrane phospholipids. The high phosphatidic acid content of L cell surface membranes in our study does not appear to be a degradative artifact since the fatty acid pattern of this phospholipid is quite different from that of any of the other major phospholipids (Table V) 13osmann, Hagopian, and Eylar (41)  membranes from HeLa cells grown in suspension culture. These membranes have approximately the same lipid to protein ratio as rat liver and L cell surface membranes (Table VI, Line 8). HeLa cell surface membranes surprisingly have been reported to have a low phospholipid content (437; of the total lipid) and therefore a high cholesterol-phospholipid molar ratio of 1.05. The neutral lipid and phospholipid contents are quite different from that of L cell and rat liver surface membranes.
Since a high ratio of cholesterol to phospholipid appears to be a common feature of all surface membranes which have been studied at the present time, Bosmann et al. (41) have suggested that the very high ratio found for HeLa cell membranes is a good criterion of the purity of the preparations.
We suggest that the published figures for HeLa cell membrane lipids indicate that HeLa cell surface membranes may belong to a different class of surface membrane than do those of liver and L cell.
The IIeLa cell membrane appears to be more closely related to microvilli and erythrocyte membranes and myelin (Table VI, Lines 9 to II), all of which have cholesterol-phospholipid molar ratios of 0.9 or higher.
Purified fragments of plasma membranes from an avian source (influenza virus lipoprotein envelopes (Table VI, Line 12)) have also been shown to have a cholesterolphospholipid molar ratio greater than 1.
It can be seen from Table VI that surface membranes which have high cholesterol-phospholipid molar ratios may also have high glycolipid contents.
Dod and Gray (10) found at least fire alkali-stable components of rat liver plasma membranes which were chromatographically identified as glycolipids.
The glycolipid content of the plasma membranes was estimated to be 7y0 by gravimetric analysis. This number was calculated by subtraction of the weights of neutral lipid and phospholipid from the weight of the total lipid. So information is available concerning the presence of glycolipids in the HeLa cell membranes.
We have found by nongravimetric means that glycolipid accounts for less than 170 of the lipid of L cell surface membranes prepared by the FJIA procedure. 3 It may be possible to classify surface membranes into groups based upon their glycolipid tolltent since, as seen in ' Table VI, some membranes are rich in glycolipids and others arc quite poor.
The relatively constant composition of the FMA membrane lipids and the escellent agreement with the composition of L cell membranes prepared by treating cells with Tris-Cl buffer (pI1 7.4) are good indications that extensive degradation of lipids did not occur during the isolation of surface membranes. FAIL1 membranes are isolated as whole ghosts, and it can be estimated that these ghosts contain approsimately 4.7% of the protein of the whole cells from which they are prepared.
Similarly, it can be estimated that 13.8 To of the lipid of the L cell is in the surface membrane as isolated by the FMA method.
Coleman (6) and Skipski et al. (9) found that the rat liver plasma membranes contained less unsaturated fatty acids than the whole cells from which they were isolated.
L cells and theil surface membranes both contain 4SL% of unsaturated fatty acids. However, inspection of the fatty acid composition of L cell membrane lipids points out several interesting features.
(a) L cells grown in media containing linolenic acid (18 :3) appear to concentrate this fatty acid selectively in the surface membrane, accounting for 2570 of the total fatty acids of the membrane.
(b) There can probably be a certain amount of latitude in the degree of unsaturation of the surface membrane lipid without affecting its structural integrity and function.
L cells which are grown in the absence of endogenous lipid do not make polyunsaturated fatty acids (35) and under these conditions the total degree of unsaturation of the surface membrane would be much lower. Although differences in the chemical and quantitative distribut,ion of lipids may be related to differences in the surface properties, such as permeability (44), no definitive data concerning this type of relationship are currently available.
(c) Compared to the whole cell total lipid there is almost a 2-fold enrichment of sphingomyelin in the surface membrane of the L cell. Ninety per cent of the fatty acids of the membrane sphingomyelin are accounted for by linolenic acid (18 : 3) and lignoceric acid (24 : 0). This accounts for a large part of the concentration of these fatty acids in the surface membrane.
(d) All of the membrane phospholipids contain relatively large amounts of polyunsaturated and long chain fatty acids ( Table V). Analyses of whole cell phosphatidylserine and sphingomyelin fatty acids3 also indicate enrichment with polyunsaturated fatty acids, in contrast to most other mammalian cells which have been examined (47, 48). These fatty acid patterns indicate the operation of selection processes involving the utilization of fatty acids by L cells in tissue culture. Dodge and Phillips (47) have reported that the fatty acid distribut,ion of sphingomyelin from human erythrocytes is very different from that of the major " glvcerophospholipids of the erythrocyte.
Sphingomyelin contained less unsaturated fatty acids than the other phospholipids but contained most of the lignoceric (24:0) and nervonic (24: 1) acid found in the erythrocyte phospholipids.
Pfleger et nl. (11) found that the fraction from rat liver plasma membranes which contained phosphatidylserine and phosphatidylinositol had a higher arachidonic acid (20:4) content than any other lipid fraction.
They also found that sphingomyelin was the only membrane lipid that contained fatty acids longer than arachidonic acid. It would seem, then, that the fatty acid patterns of surface membrane sphingomyelin and phosphatidylserine are quite different from those of other lipids.
In summary, several characteristic features of animal cell membranes have emerged from analytical studies of membrane lipids.
(a) The lipid composition of the surface membrane of a cell is quite specific in that it differs from that of other membrane systems of the cell (4,41,46). (b) Except for variations in acyl chains due to environmental factors (39), the lipid compositions of the plasma membranes of animal cells show characteristic features such as greater cholesterol and neutral lipid content than other membranes from the same source. (c) Surface membranes from rat liver (10, ll), erythrocytes (44), and cultured mouse fibroblasts (Table II) show a greater sphingomyelin content and lower lecithin and phosphatidylethanolamine contents than internal cell membranes.
(d) Individual membrane lipids may contain very specific fatty acid patterns.
(e) Lipid parameters may be useful in the classification of surface membrane preparations.