* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download cell surface lipids and adhesion
Survey
Document related concepts
Biosynthesis wikipedia , lookup
Monoclonal antibody wikipedia , lookup
Vectors in gene therapy wikipedia , lookup
Polyclonal B cell response wikipedia , lookup
Butyric acid wikipedia , lookup
Lipid signaling wikipedia , lookup
15-Hydroxyeicosatetraenoic acid wikipedia , lookup
Evolution of metal ions in biological systems wikipedia , lookup
Biochemistry wikipedia , lookup
Specialized pro-resolving mediators wikipedia , lookup
Transcript
J. Cell Sci. 18, 357-373 (i97S) 357 Printed in Great Britain CELL SURFACE LIPIDS AND ADHESION II. THE TURNOVER OF LIPID COMPONENTS OF THE PLASMALEMMA IN RELATION TO CELL ADHESION A. S. G. CURTIS, F. M. SHAW AND V. M. C. SPIRES Department of Cell Biology, University of Glasgow, Glasgow Gi 1 6NU, Scotland SUMMARY The preceding paper showed that those conditions that ought to stimulate reacylation of lysolipids in cells can increase cell adhesion. Similarly we found that conditions that would be expected to lead to the accumulation of lysolipids in the cell surface diminish cell adhesion. This paper reports on the answers to the following questions. (1) Is reacylation of lysolipids in the cells stimulated by an external supply of CoA, ATP and a fatty acid ? (2) Does this reacylation lead to the incorporation of exogenous fatty acid in the plasmalemma ? (3) What range of fatty acids can be incorporated into the plasmalemma and into what compounds? (4) Does the plasmalemma contain the enzyme systems to effect this turnover, namely phospholipase A,, a CoA-ligase and an appropriate acyl transferase(s) ? (5) Do lysolipids accumulate in the plasmalemma under conditions which diminish cell adhesion ? We find that saturated fatty acids in the range C14-C18 and some unsaturated fatty acids are incorporated into the plasmalemmae of these neural retina cells. About 20 % of the plasmalemmal content of fatty acids can be turned over in 30'. Incorporation is mainly into phosphatidyl choline, serine and ethanolamine in both Rt and Ra positions. The plasmalemmae contain the enzymes to effect the turnover. Isolated plasmalemmae are active in this turnover. Incubation of the plasmalemmae with phospholipase A» leads to an accumulation of lysolipids. Very low levels of phospholipase stimulate turnover, possibly because endogenous phospholipase activity is the rate-limiting step in the system. These findings are discussed in relation to the possible mechanisms by which lipids might affect adhesion. INTRODUCTION The results reported in the previous paper (Curtis, Chandler & Picton, 1975 a) suggest that the state of the lipids in the cell surface might affect adhesion. Conditions which might be expected to lead to the accumulation of lysophosphatidyl compounds in the cell surface reduced cell adhesion. Conversely conditions which might lead to the reacylation of lysophosphatidyl compounds increased adhesion. However, it was not demonstrated in that paper that these changes do in fact take place in the plasmalemmal lipids or in other cell components when the appropriate conditions are applied. This paper is directed towards answering these questions. In the process incidental information on the nature of cell surface enzymes involved in lipid turnover has appeared. Warren & Glick (1968) found that the MC-labelling of the L cell surface with 358 A. S. G. Curtis, F. M. Shaw and V. M. C. Spires glucosamine or amino acids suggested that turnover did not occur unless growth was halted, when it had a half-life of about one day. Pasternak & Bergeron (1970) reported that phospholipid turnover as followed by the loss of a choline label in P815Y tumour cells was fairly fast even in rapidly growing cells. About 60% was lost in 35 h. It should be borne in mind that Pasternak & Bergeron studied turnover in a membrane fraction presumably derived from all membranous components of the cells. Phosphatidyl choline was the most rapidly turned over phospholipid. However they found that amino acid-labelled membrane components were turned over very slowly. Fischer et al. (1967) studying the turnover of phosphatidyl choline using a fatty acid label in erythrocytes, came to the conclusion that 50 % turnover could occur in 80 min and 20% in 20 min. Thus though there are differences between the cell types and the growth of the cells used by the 3 groups there is a strong suggestion that different labels reveal different turnover rates which indicates that various parts of the molecules and membrane may be replaced at differing rates. Thus there is some indication in the literature that fast turnover of the fatty acid parts of membranes can occur. The first question to be asked of the neural retinal cell system used in the previous paper is whether the medium used to stimulate reacylation leads to incorporation of fatty acid into either or both Rx or R2 positions of phosphatidyl compounds in the plasmalemma or other cell membranes. The second main question to be resolved in this paper is whether lyso-compounds accumulate in the cell surface when phospholipase A2 or lysolecithin is added to the incubation medium, or in the presence of Hanks' saline. MATERIALS AND METHODS Biological material Neural retinae from chick embryos (De Kalb strain) were dissected out on the 8th day of incubation (7-day-old eggs). They were dispersed into single cell suspensions by the method described by Curtis, Campbell & Shaw (1975 a). Incubations of cells in various media were carried out by the methods given in the previous paper (Curtis et al. 1975 a). Reagents Lysolecithin (mixed palmitoyl and stearoyl), Coenzyme A (free acid), oleic, palmitic, linoleic, stearic and linolenic acids were obtained from Sigma. Phospholipase A, bee venom (purified for sequence analysis) was a gift from Dr S. Doonan; its purity has been described in the previous paper. Dimyristoylphosphatidyl choline, distearoylphosphatidyl choline, Calbiochem., dilinoeylphosphatidyl choline, Lipid Products, sphingomyelin and other standards for TLC, Field Instruments; CoA acyl derivatives, Mann. 14C-labelled fatty acids, The Radiochemical Centre, Amersham. "C-labelled stearoyl-CoA, NEN Gmbh. Silica gel paper for TLC, Whatman; Polygram Sil N-HR sheets for TLC, Macherey-Nagel. Reagents for ATPase assay, Sigma. Ficoll, Pharmacia. Other reagents British Drug Houses (AR grade whenever possible). 14 C-labelled lysolecithin was prepared by incubation of 14C-labelled phosphatidyl choline (NEN Gmbh) with bee venom phospholipase At in 20 % n-propanol in 1 x io~2 M triethanolamine-isobutyrate buffer pH 82, followed by removal of free fatty acid and the enzyme on a Biodeminrolit column in 50 % w-propanol and 50 % of the above buffer. The purity of the lysolecithin produced was established by TLC (system C), as 97 %. Cell surface lipids and adhesion. II 359 Cell surface isolation Plasmalemmal fractions were isolated using a slight modification of Gahmberg's technique (Gahmberg & Simons, 1970). Cell suspensions, prepared by the method described above, were pelleted at 300 g. In the majority of experiments the yield of 30 retinae, ca. 25 x io7 cells, was used. The pellet was homogenized in a sucrose-MgSO4 medium (sucrose 0-25 M, MgSO4 2 x io~* M buffered with Tris-HCl 5 x io~3 M at pH 7-4). The homogenate was then made 1 x io~3 M with EDTA and centrifuged at 5000 g for 15 min to sediment nuclei, mitochondria, etc. The supernatant from this stage was recentrifuged at 3 x io4 g for 30 min. The pellet formed was resuspended in Tris buffer 1 x io~3 M, pH 8-6, to which had been added MgSO4 to make it 1 x io~3 M, using exposure to ultrasonic vibration to ease the resuspension. Care was taken to ensure that there was a minimal carry-over of the sucrose medium into the fraction containing the resuspended pellet. This fraction which contains endoplasmic reticulum and plasmalemmal components was then overlayed 3 ml of Ficoll solution made up in the pH 8-6 Tris buffer plus MgSO4 solution described above. The Ficoll solution had a density of 1-050 g cm"3 at 20 °C (i4-5 %, w/v). This density discontinuity system was centrifuged at 1-3 x io 6 g for 80 min. A thin opalescent band appeared at the density discontinuity and a small amount of material collected at the bottom of the tube. The opalescent band was recovered as the plasmalemmal fraction. Recognition of plasmalemmal components The following 3 techniques for identification of plasmalemmae were used: (1) Electron microscopy of the opalescent band was carried out for us by Dr E. Follett, using phosphotungstic acid and ammonium molybdate for negative staining. Results are shown in Fig. 3, which indicates that the fraction is composed of membranous fragments. (2) Fluorescent antibody staining of the band with a cell surface-directed antibody. This antibody was prepared by injection of ca. 2 x 1 o' cells (neural retinal) in 2 ml of Hanks' medium into rabbits, without the use of Freund's adjuvant, on 3 occasions separated by one week each. The rabbits were bled 10 days after the last injection and the serum prepared from the blood. An immunoglobulin fraction was then made by running the serum on a DEAEcellulose column, o-i M phosphate buffer pH 8-o. Fluorescent goat anti-rabbit immunoglobulin (Burroughs-Wellcome) was used to detect binding of the anti-neural retinal cell antibody. The double system gave ring staining of whole neural retinal cells when carried out at 4 °C on unfixed cells or at 20-30 °C on acetone-fixed cells. There was very little reaction of the antibody with other cellular components isolated by the series of centrifugations. Results of the staining of whole cells can be seen in Fig. 4. This antibody reacted extensively with the opalescent band but showed little reaction with other cell fractions. (3) Measurement of the sodium- and potassium-dependent ATPase activity of the fractions. The method of Slack, Anderton & Day (1973) was used. The results, which are given in Table 1, indicate that the opalescent layer contained a high specific activity of this enzyme unlike the other fractions. The results of applying these 3 criteria suggest that the opalescent layer contains a plasmalemmal fraction probably free from gross contamination with other organelles. Assay of enzyme activities in the plasmalemmal fraction Phospholipase A2 activity was measured by the conductimetric technique of Moores & Lawrence (1972) using the water-soluble substrate dihexanoyl phosphatidyl choline. Acyl transferase activity was assayed by following the incorporation of 14C-labelled oleate into phosphatidyl compounds, using the TLC techniques described below for detection of these compounds. CoA ligase activity was determined by following the incorporation of 14C-labelled stearic acid (sp. act. 50 Ci mol"1) into stearoyl-CoA, which was identified and separated by TLC methods (see later). 360 A.S.G. Curtis, F. M. Shaw and V. M. C. Spires Recovery of lipidfrom the plasmalemmalfraction The opalescent (plasmalemmal) layer from the density discontinuity centrifugation was extracted successively with n-propanol-chloroform (1 :z) and 2 further chloroform extractions. The pooled extracts were dried under a flow of oxygen-free nitrogen. Radioactive labelling of plasmalemmal lipids Two systems were used: (1) labelling of whole cells, and (2) labelling of plasmalemmal fractions. In both cases the incorporation medium contained ATP 1-25 x io~° M dm"3, CoA, free acid form, 5 x io~* M dm"3 and a chosen longchain fatty acid at 3-5 x io" 5 M dm"' in Hanks' saline. 14C-labelled fatty acids were added to the identical unlabelled fatty acids to give specific activities between 1-5 and i-8 /iCi//tmol. The fatty acids were dissolved in the medium with the aid of ultrasonication. In order to label cells 1 ml of the cell suspension was mixed with 9 ml of the incorporation medium, and the whole incubated at 37 °C for 20 min. Approximately 6 x io7 cells were present in each millilitre of thefinalincubation suspension. After incubation the cells were washed twice with cold Hanks' medium by centrifugation and replacement of the medium. Samples were taken to measure the crude total incorporation of label by the cells and the release of label during the washes. The cells were then homogenized for plasmalemmal isolation. Plasmalemmal fractions were labelled by mixing 1 ml of the incorporation medium with 1 ml of the Ficoll-Tris buffer discontinuity containing the plasmalemmal layer. The mixture was then incubated at 37 °C for 20 min. It proved to be impossible to effect a good recovery of the membranes after centrifugation and washing. Instead, unincorporated fatty acids were removed during the TLC runs. Separation and identification of plasmalemmal lipids TLC systems were used to separate and identify the lipids which had been extracted with chloroform:«-propanol. The following systems were used: (A) Whatman Silica gel paper. Eluant: CHC13 50, CH,OH 30, acetic acid 8, HjO 4 (v/v). This system retards lysolecithins relative to all other components, while phosphatidyl cholines run more slowly than phosphatidyl serine and ethanolamine. (B) Silica gel on plastic sheets (Sil N-HR). Eluant: CHC13 90, CH,OH 10, H,O 1 (v/v). This system retards phosphatidyl choline and phosphatidyl serine and sphingomyelin at the origin while allowing phosphatidyl ethanolamine to run with Rf = 020 and cerebrosides, ceramides and neutral lipids to run with Rf values increasing in that order. (C) Sil N-HR sheets. Eluant as for system A. Phosphatidyl serine and ethanolamine have Rf = 085 while phosphatidyl cholines run with Rt = o-6o. Use of this system in conjunction with B allows separate estimation of phosphatidyl ethanolamine and phosphatidyl serine. (D) Sil N-HR sheets. Two-dimensional. Eluant 1: CHC1, 95, CH3OH 5, 28% NHjOH, o-8. Eluant 2: CHC1, 32, acetone 6, CH,OH 1, acetic acid 1, H,O 1 (v/v). This retains glycolipids and phospholipids very close to the origin but ceramides have Rf values of Rfl = 050, R)t = 0-80-0-90, neutral lipids have Rfl = 095 and Rfl = 0-95 while free fatty acids have R/i = 0-25, Rft = 0-95. Use of this system allows separation of unincorporated fatty acids, ceramides and neutral lipids from each other and the phosphatidyl and glyco-lipids. (E) Sil N-HR sheets. Two-dimensional. Eluant 1: CHC1S 65, CH3OH 25, H,0 4. Eluant 2: Di-isobutyl ketone 80, acetic acid 50, HjO 10. This system separates phospholipids from each other close to the origin while glycolipids and neutral lipids run with larger Rf values in both dimensions. This system was mostly used for rough classification of classes of fatty acid incorporation. All TLC systems were run in the ascending mode. When radioactive materials were being run background counts on the plates did not exceed 40 dpm cm"1 ("C channel). Cell surface lipids and adhesion. II 361 Detection systems Appropriate standards were run in parallel with analytical runs and the Rf values of the standards provided some degree of identification of the spots obtained from the plasmalemmal extracts. In addition compounds were detected: (a) non-destructively by staining with iodine vapour, by making autoradiographs of the plates when labelled compounds were being used or by use of a Panax scanning radiochromatogram counter, and by use of the Malachite green spray (Ansell & Hawthorne, 1964) which is excluded by lysophosphatides after prolonged drying. (b) destructively by the formaldehyde-sulphuric acid charring reagent, or by use of the molybdenum blue reagent for phospholipids or with the orcinol-sulphuric acid spray for glycolipids. Spots were eluted for counting in a Beckman CPM 200 liquid scintillation counter, using the scintillation fluid to elute the lipid from the strips of gel paper or scraped off silica gel. The scintillation fluid used contained 0-4 % PPO in toluene. An additional method for the detection of phospholipids was used. Aliquots of the plasmalemmal lipid fractions were treated with 1 mg of phospholipase A, in 50 x io~9 m* of 4 0 % n-propanol in water for 10 min before spotting on to TLC System A. These runs were compared with the untreated plasmalemmal runs. Any spots missing from the enzyme-treated aliquots can be identified as phosphatidyl compounds. Two new spots should appear in the enzyme-treated material, free fatty acid and lysophosphatides. If this technique is applied to plasmalemmae which have incorporated labelled fatty acids the distribution of label in the free fatty acid and lysophosphatides respectively will reveal whether fatty acid is incorporated into the 1 or 2 position or into both. In general these methods permit the detection of 1-2 fig of lipid in a given spot by chemical methods and, of course, considerably less labelled lipid. RESULTS The main data on the plasmalemmal fractions obtained are summarized in Table 1. Recovery of the membrane fraction is about 15%. The ratio of phospholipid to protein is 100:55. Table 1. Properties of purified and partially purified plasmalemmal fractions Whole cell homogenate Phospholipid, /ig/107 cells Protein, fig/107 cells ATPase activity before ouabain after ouabain Phospholipase A activity CoA-ligase activity Reaction with cell surface antibody N.D. N.D. Membrane fraction Plasmalemmal fraction 210 189 20-O S-2 2-9 263 7-2 4-0 N.D. N.D. N.D. N.D. 0 0 4 Very weak + + ++ 37-6 2'5 i-35 Phospholipid measured as phosphate after HC1O4 digestion, calculated as phospholipid on basis of molecular weight of distearoyl lecithin and corrected for recovery. Protein measured by micro-Lowry method using bovine serum albumin as a standard, corrected for recovery. All enzyme activites expressed in units of fimol h" 1 per mg protein. N.D. = not determined. 362 A. S. G. Curtis, F. M. Shaw and V. M. C. Spires Incorporation offatty acids Into whole cells. Measurements of the incorporation of fatty acids into the cells were made by simple measurement of the loss of label from the incubation medium. They show that between 60-80% of the label is incorporated in individual experiments Table 2. Incorporation of [14C]oleic acid by neural retinal cells dpm per 1 x 10' cells A. Effect of components of incorporation medium 274 CoA plus fatty acid ATP plus fatty acid 436 CoA, ATP and fatty acid 1395 B. Classes of incorporation in presence of CoA and ATP Phosphatidyl choline 138 Phosphatidyl ethanolamine 250 Phosphatidyl serine 313 Phosphatidyl glycerol and inositol 55 Neutral lipids 277 Free fatty acid 325 Lysophospholipids 30 TLC methods used B, C, D, E. Incorporation methods, see Materials and methods. Specific activity of [14C]oleate I-I fiCi/fimol. Values are means of 5 runs. Table 3. Incorporation of [14C]o&ate into plasmalemmal lipids in intact cells Free fatty acid Phosphatidyl choline Phosphatidyl serine Phosphatidyl ethanolamine Phosphatidyl glycerol and inositol Lysophosphatides Sphingomyelin Cerebrosides Ceramides Neutral lipids 186 466 455 255 200 35 N.D. 94 81 552 Cells incubated in oleate-CoA-ATP incubation medium (see Materials and methods) for 20 min at 37 °C. Isolated plasmalemmal lipids run on TLC by methods B, C and D. Specific activity of [14C]oleate 1-5 ftCi/fimo\. Results expressed in dpm per 1 x 10' cells, conversion factor for dpm per mg protein = 735. Values are means of 3 runs, maximal standard deviations ± 20 %. N.D. Not detected. under the conditions used, corresponding to some 2 /tg of oleic acid per 1 x io7 cells. Use of TLC methods A and D on lipid extracts from the cell homogenates made immediately after exposure to the incorporation medium, shows that the incorporation is into phosphatidyl choline, ethanolamine, serine, glycerol and neutral lipids (not separately examined) and that some of the incorporation is of free fatty acid. Data on this whole cell incorporation are given in Table 2. Cell surface lipids and adhesion. II 363 When these incorporations are compared with the total lipid of the cells, it is clear that incorporation is fairly rapid. Subsidiary experiments showed that CoA-fatty acids can be incorporated as well as fatty acids. At this stage it is unclear as to whether this is de novo or turnover synthesis though the rapidity of labelling suggests that turnover must account for a large part of the labelling. I 2x10* - 50 25 Time (min) Fig. 1. Pulse-chase experiment. Isolated plasmalemmae unsonicated (to prevent loss through filter) labelled with [14C]stearic acid (3-0 /igstearate, specific activity 41-9 ftCi/ fimo\) for 20 min, washed on o-22-fim pore diameter filter with cold stearate medium (see Materials and methods). Then incubated, starting at time o, at 38 °C, with successive i-ml aliquots of cold medium for successive 5-min periods. • , cpm in plasmalemmae (of replicates). V, cpm released into cold medium (cumulative), as free fatty acid. Similar measurements were made on the incorporation of oleate by whole cells, into the plasmalemmal fraction, see Table 3. TLC systems A, B, C and D were used. The results show that about 20 % of the total plasmalemmal lipid can be substituted with oleic acid by this technique. This value is derived from calculation of the percentage recovery of plasmalemmal lipid. 1 x io7 cells (radius 4-2 /im) should have a 23 CEL 18 364 A. S. G. Curtis, F. M. Shaw and V. M. C. Spires plasmalemmal volume greater than 2 x io~B ml, corresponding to at least 25 fig of lipid and protein. In practice 2-6 fig phospholipid and 2-8 fig protein were obtained from this number of cells in the plasmalemmal fraction. Thus the total recovery is less than 20 %. The data in Table 3 show that about 1 nCi of oleic acid is incorporated nto 1 x io7 cells which corresponds to 0-15 fig oleic acid, or 0-45 /tg phospholipid. After correction for the percentage recovery this corresponds to about 18% of the total lipids. Similar results were obtained for other fatty acids. Data in Tables 2-7 are uncorrected for recovery. Table 4. Fatty acid incorporation by isolatedplasmalemmae Free fatty acid* 137 xio 8 Phosphatidyl choline 129230 Phosphatidyl serine 88461 Phosphatidyl ethanolamine and glycerol 230770 Phosphatidyl inositol 3477O Lysophosphatide 6538 0 Ceramides 64846 Neutral lipids • This figure is probably affected by our inability to remove all unincorporated fatty acid from the plasmalemmae after incubation. Incorporation of [14C]oleic acid by isolated neural retinal plasmalemmae: 20 min incorporation at 37 °C. Plasmalemmal lipids isolated by TLC methods B, C, D and E. Specific activity of ["CJoleate 1-5 /iCi//imol. 1 x 10' cells yield ca. 1-3 fig protein. Figures are the means of 5 determinations. Results expressed in dpm per mg protein. Thus incubation of these cells in the incorporation medium allows a very considerable replacement of the acyl side chains. Measurements of total lipid before and after incorporation and pulse-chase experiments show that this plasmalemmal incorporation is a turnover synthesis. Comparison of Tables 2 and 3 shows that about 10% of the incorporation is into the plasmalemma. Pulse-chase experiments (Fig. 1) on isolated plasmalemmae suggest that the incorporation is of a turnover type since incorporated labelled fatty acids can be displaced on incubation in excess cold fatty acid and cofactors. Into the plasmalemma. The first question to be answered is whether the plasmalemma contains the phospholipase A and the acyl-transferase as well as the CoA-ligase required to effect this turnover synthesis or whether the phosphatidyl compounds, etc., are acylated elsewhere and then exchanged for unlabelled lipid in the plasmalemma. The following techniques were used to answer this question. First, the ability of the isolated plasmalemmal fraction to incorporate labelled oleate was examined, using the same techniques as were used to investigate whole cell incorporation. The results are shown in Table 4. They show that the isolated plasmalemma incorporates label very readily into phosphatidyl compounds, cerebrosides and into neutral lipids. The labelling is relatively more extensive than into the plasmalemma of intact cells which might mean that any intermembrane exchange of labelled lipid is from, rather than to the plasmalemma. Owing to the method used for preparation of these incorporation Cell surface lipids and adhesion. II 365 samples it was impossible to remove unincorporated free fatty acid so that all that can be said is that most of the free fatty acid measured in each system must be in fact unincorporated material. The results imply that all the components required for incorporation must be present, namely a phospholipase to degrade diacyl phosphatidyl compounds to the 1 or 2 lyso-compounds, a CoA-ligase to link CoA and the external fatty acid and an acyl transferase. Second, we attempted to measure phospholipase A2 and CoA-ligase activities in the isolated plasmalemmae, see Materials and methods for details of techniques. Phospholipase A was measured by the conductimetric method on freshly isolated plasmalemmal fractions. CoA-ligase was measured by following the incorporation of 14C-labelled stearic acid into acyl CoA in isolated plasmalemmae. Acyl CoAs were detected by paper chromatography in n-butanol: acetic acid: H2O (5:2:3) on paper (Whatman No. 1). Ani? f value of 0-55 was obtained from stearoyl-CoA with this system. Results are shown in Table 1. They show that these enzyme activities are present in the membrane fraction. The total incorporation demonstrates that an acyl transferases are present in the membrane fraction. Table 5. Fatty acid incorporation into neural retinal plasmalemmae Caprate Myristate Palmitate Stearate Oleate Linoleate Linolenate NDB 282 1252 1290 989 820 971 Incorporation conditions, see text, dpm per 1 x io7 cells. Recovery ca. 10%. Specific activities I-2-I-6/tCi//imol. Values are means of 5 runs. NDB, not distinguishable from background. Range offatty acids incorporated into the plasmaUmma The incorporation of even-numbered saturated fatty acids in the range C8-C22 into the plasmalemmae of whole cells was measured, see Table 5. It is clear that the system only effects appreciable incorporation for acids in the range C^-C^. Ciy-unsaturated fatty acids in the C18 series were also studied in this way, see Table 5. Position offatty acid incorporation, control of total turnover rate and conditions for the accumulation of lysophosphatidyl compounds Slightly surprisingly these three questions have been elucidated by the same set of experiments. The measurements on plasmalemmal incorporation of fatty acids into phosphatidyl compounds, Table 3, showed that no incorporation took place into the single acyl side chain position for lysophospholipid. A variety of explanations could be put forward for this finding but most of these are excluded by the results of examining incorporation into membrane in the presence of fairly high levels of added phospholipase A2. These results are described in Table 6: they show that incorporation is much greater than for untreated membranes, see Table 4. In addition lysophosphatidyl compounds can be detected which carry appreciable label. These results 23-2 366 A.S.G. Curtis, F. M. Shaw and V. M. C. Spires suggest that the rate-controlling step for incorporation is the production of lysophosphatidyl compounds due to the action of intrinsic or added phospholipase A2. Presumably this lysis is fairly slow in the normal membrane so that re-acylation can keep the pool of lysophosphatidyl compounds at a very low level and hence almost no labelled lysolipid is found. Table 6. Incorporation of [14C]oleate into isolated plasmalemmae in the presence of added phospholipase A2 Control In presence of O'S fig ml"1 of enzyme Neutral lipid I , 537 I3s6 Free fatty acid j Phosphatidyl choline 3i» 53° 266 93i Other di-acyl phosphatides 28 Lysophosphatidyl choline 237 Not detected Other lysophosphatides 15° 7 Specific activity of oleate 1-3 /*Ci//ttnol. 1 x io cells per sample. 20 min incubation at 37 °C. 10-ml incubation volumes. Incorporations expressed as dpm per 1 x io7 cells. Increase in phospholipase A activity presumably results in enlargement of the lysophosphatide pool so that more material is available for reacylation. The appearance of labelled lysophosphatides, see Table 6, suggests that labelling can occur in both the 1 and the 2 positions, namely that lysophosphatide was labelled in one position and then reconverted to a lysophosphatide with the other position available for further labelling. This is, however, considered more fully in the Discussion. Clearer evidence for labelling in both positions comes from these experiments in which aliquots of the extracted plasmalemmal lipid (after labelling) were incubated with phospholipase A8 and then run on TLC (system A) beside untreated control aliquots, see Fig. 2. No lysophosphatidyl choline can be detected in untreated plasmalemmae but both labelled fatty acid and labelled lysolecithin can be found in the enzyme-digested sample. The snake venom phospholipase used is believed to form lysophosphatides with the fatty acid chain in the 1 position. If this is correct the appearance of label in both the free fatty acid and in the lysolecithin implies that labelling has taken place in both the 1 and 2 positions. In the first paper in this series it was reported that incubation of the cells in Hanks' at 37 °C produced effects on adhesion similar to those obtained by phospholipase A treatment. For this reason the question of the appearance of lysolecithin and other lysophosphatides in the plasmalemmae of cells that have been incubated in Hanks' was examined. Incubation was at 37 °C for 10 min. Results are shown in Table 7. It is clear that incubation of cells under these conditions allows accumulation of lysophosphatides in the plasmalemma. In order to discover whether lysolecithin added to the medium is in fact taken up in the plasmalemma, as would be expected from the results described in the first paper, 5 x io6 cells ml" 1 were incubated in Hanks' medium containing 0-5 fig ml" 1 of Cell surface lipids and adhesion. II 367 •4000- Lecithins r it LA 0 10 Fig. 2. TLC on silica gel paper (method A) of plasmalemmal lipids before (line over unshaded area) and after (line over shaded area) phospholipase A2 treatment of lipids. [14C]Myristic acid was incorporated into cells. Ordinate: cpm extracted from paper; abscissa: distance in cm from solvent front. Note the disappearance of the lecithin peak after enryme treatment and the appearance of a labelled free fatty acid peak at left-hand side and lysolecithin peak at the right-hand side. This suggests that substitution occurs in Rx and R, positions. The peaks at the left, near the solvent front, represent phosphatidyl serines, ethanolamines, etc., before enzyme treatment and their lyso derivatives afterwards. Table 7. Accumulation of lysophosphatides in the plasmalemmae of cells incubated in Hanks' medium at 37 °C Control Treated Free fatty acid 126500 151000 227000 212200 Neutral lipids 218000 164500 Phosphatidyl choline 369200 205300 Phosphatidyl serine and ethanolamine 36700 Lysophosphatidyl choline Not detected 71000 Not detected Other lysophosphatides 12 x 10' cells were incubated in 10 ml of incorporation medium containing [14C]oleate (see Materials and methods) for 20 min at 37 °C. Controls were then homogenized and their plasmalemmae isolated. 'Treated' cells were incubated in Hanks' medium at 37 °C for 10 min before preparation of plasmalemmae. Cold dimyristoyl lecithin was added to the membranes during isolation to stabilize them. Specific activity of the oleate 1 -4 /iCijfimoL Results expressed as dpm per mg protein. 368 A.S.G. Curtis, F. M. Shaw and V, M. C. Spires 14 C-labelled lysolecithin for 5 min. The cells were then collected and the distribution of label in the plasmalemmal and other fractions determined. Results are shown in Table 8. It can be seen that lysolecithin is incorporated into the plasmalemma, some of it being acylated into lecithin. Table 8. Uptake of lysolecithin into the plasmalemma of cells incubated in a lysolecithin solution 1 Total label added per I X I O cells Unincorporated after 5 min at 5 °C Plasmalemmal Isolation by Gahmberg procedure In 1st pellet In 2nd supernatant In Ficoll subnatant and supernatant to plasmalemmal layer In Plasmalemmal fraction, separation by TLC method C Lysolecithin Lecithin Free fatty acid Background subtracted Expt. 1 Expt. 2 29122 54000 2824 3894 857 38773 i°33 992 148 81 72 56 64 58 25 991 32 Sp. act. of the lysolecithin 08 /iCi//rtnol. Results expressed in dpm per 1 x io 8 cells. DISCUSSION The experimental results show that neural retinal cells will incorporate fatty acids from the medium when they are incubated in the presence of CoA and ATP. This incorporation, which is surprisingly rapid, is into the membrane phospholipids in the main. An appreciable fraction of the incorporation is into plasmalemmal phospholipid. Similar results have been obtained for liver cells by Wright & Green (1971) and for erythrocytes by Fischer et al. (1967). Is the plasmalemmal incorporation effected in situ or is the synthesis carried out in other cell organelles followed by intracellular exchange processes ? Is the incorporation de novo or turnover synthesis ? Both these questions are answered by the study of incorporation into isolated plasmalemmal preparations. The incorporation into isolated plasmalemmae shows that incorporation in life is probably directly into the plasmalemma. This incorporation is more extensive than that into the plasmalemmal fraction of whole cells, possibly because the isolated plasmalemma is accessible to the reactants from both sides or because exchange reactions with other cell organelles are of course impossible. It is of course most improbable that dc novo synthesis of phospholipid could take place in the isolated plasmalemmae. In any event pulse-chase experiments on isolated plasmalemma demonstrate that synthesis is of the turnover type. The plasmalemmal incorporation system must involve a CoA-ligase, an acyl transferase and phospholipases to produce the lysophosphatides required for reacylation. In these cells it appears to be able to accept fatty acids in the range C^-C^ from the Cell surface lipids and adhesion. II 369 saturated series and the m-unsaturated acids. Incorporation of fran^-unsaturated acids was not studied. The conductimetric assay used to establish the presence of phospholipase in the plasmalemmal fraction cannot distinguish between Ax and A2 enzymes nor resolve the question of whether some pure phospholipases may not possess both specificities, but the appearance of lysophosphatides with incorporated label argues strongly that both 1 and 2 positions are available for reacylation whether by hydrolysis in both positions or by intramolecular exchange. The absence of lysophosphatides in the plasmalemmae of cells freshly isolated with EDTA or in cells incubated in the presence of a fatty acid, CoA and ATP (leading to rapid fatty acid incorporation) suggests that the reacylation system is normally able to ensure that the cell surface is kept free from lyso-compounds. The function or functions of this repair synthesis are still uncertain, though Dawson (1973) has discussed possible explanations in a stimulating review. It is of particular interest in the context of the first paper in this series (Curtis et al. 1975 a) to note that incubation of the cells in Hanks' medium leads to the accumulation of lyso-compounds in the plasmalemmae. In that paper we reported that treatment of cells with phospholipase A2, or incubation of cells with lysolecithins led to a diminution in cell adhesion as did incubation in Hanks' medium. We suggested that the reason for this diminution might be the accumulation of lysophosphatides in the membrane. In the present paper we have shown that phospholipase treatment or treatment with lysolecithin or incubation in Hanks' leads to the accumulation of lysophosphatides in the plasmalemma. In the preceding paper (Curtis et al. 19750) we reported that conditions that would be expected to stimulate reacylation prevent the loss of, or restore the adhesive state of, the cells. We have shown in this paper that such treatments do indeed lead to the formation of phosphatidyl compounds from lysophosphatidyl compounds in the plasmalemma and other parts of the cells. This provides further evidence for the concept that the state of the lipids in the cell surface affects adhesion. In essence the following main categories of theory can be put forward to explain the role of plasmalemmal lipids in cell adhesion: (1) That rapid turnover is required to maintain an adhesive state. If such a theory is true one particular interpretation would be that new cell surface has to be synthesized to form and maybe to maintain an adhesion. Waddell, Robson & Edwards (1974) have put forward evidence for this particular theory. Alternatively we might regard evidence which appears to support such an interpretation as indicating that cells have to repair their surfaces after abnormal treatments such as trypsinization before becoming adhesive. (2) That the adhesive state depends upon the majority of the plasmalemmal lipids being in the di-acyl state and not in the lyso state. On such a theory turnover might be very active or very inactive but provided the balance was toward the di-acyl state the cell would be adhesive. Two particular sub-theories can be developed from this idea. First, that the state of membrane fluidity will affect adhesion, perhaps by determining the extent of aggregation of adhesive sites. Second that by altering the chain length and unsaturation of the plasmalemmal phosphatidyl lipids we might alter intermembrane van der Waals 370 A. S. G. Curtis, F. M. Shaw and V. M. C. Spires forces (Curtis, 1972) which may play a role in adhesion (Curtis, 1962). These matters are explored further in the next paper (Curtis et al. 19756). A subsidiary matter is considered next. It is of interest that phospholipase A2 treatment not only leads to the appearance of lysophosphatides in the plasmalemma but stimulates reacylation, presumably because the rate-limiting step in the turnover is the low activity of the membrane's own phospholipases. These findings also suggest that the balance between the accumulation of lysocompounds in the plasmalemma and the maintenance of the di-acylated state may be a delicate one dependent on quite small manipulations of the experimental conditions. For example many preparations of lysolecithin are of the stearoyl compound so that incorporation of this into the membrane may be followed by its rapid acylation to give perhaps a distearoyl lecithin in the plasmalemma which particularly aids adhesion (see Curtis et al. 19756). Our findings suggest that fatty acid can be incorporated into both 1 and 2 positions. Lands & Hart (1965) found that there are specific acyl transferases which at least show 1 and 2 position specificities. They found that incorporation into the 2 position, using a microsomal fraction acyl transferase, was fastest with unsaturated acids and then became slower with increasing chain length in the saturated series. Incorporation into the 1 position showed exactly the opposite effect, stearic acid being more rapidly incorporated than shorter or unsaturated acids. We have found that a variety of fatty acids varying in both chain length and saturation are approximately equally incorporated into plasmalemmal phospholipids. This result is not in contradiction with those obtained by Lands and Hart because it could be that stearate, for example, is more readily incorporated into the 1 position than other acids but oleate is equally well incorporated into the 2 position, while other acids are incorporated to about the same total level in both 1 and 2 positions. In most instances we did not distinguish between 1 and 2 position incorporation. We thank Science Research Council for a grant (B/SR49099). We should also like to thank Dr E. Follett of the Department of Virology, University of Glasgow, for preparation of the electron micrograph shown in Fig. 3, and Dr S. Doonan of the Department of Chemistry, University College, London, for the gift of phospholipase Aj. We should like to express our appreciation of the skilled technical assistance we have received from Miss Rose McKinney. REFERENCES G. B. & HAWTHORNE, J. N. (1964). Phospholipids: Chemistry, Metabolism and Function. Amsterdam: Elsevier. CURTIS, A. S. G. (1962). Cell contact and adhesion. Biol. Rev. 37, 82-129. CURTIS, A. S. G. (1972). Intra- and inter-membrane interactions of the cell surface. Subcell. Biochem. 1, 179—196. CURTIS, A. S. G., CAMPBELL, J. & SHAW, F. M. (1975a). Cell surface lipids and adhesion. I. The effects of lysophosphatidyl compounds, phospholipase Aj and aggregation-inhibiting protein. J. Cell Sci. 18, 347-356. CURTIS, A. S. G., CHANDLER, C. & PICTON, N. (19756). Cell surface lipids and adhesion. III. The effects on cell adhesion of changes in plasmalemmal lipids. J. Cell Sci. 18, 375-384. DAWSON, R. M. C. (1973). The exchange of phospholipids between cell membranes. Subcell. Biochem. a, 69-89. ANSELL, FISCHER, H., FERBER, E., HAUPT, I., KOHLSCHUTTER, A., MODELELL, M., MUNDER, P. G. & SONAK, R. (1967). Lysophosphatides and cell membranes. Protides biol. Fluids 15, 175-184. Cell surface lipids and adhesion. II 371 C. G. & SIMONS, K. (1970). Isolation of plasma membrane fragments from BHK21 cells. Ada path, microbiol. scand. 78, 176-182. LANDS, W. E. M. & HART, P. (1965). Metabolism of glycerolipids. VI. Specificities of acyl coenzyme A: phospholipid acyl transferases.,7. biol. Chem. 240, 1905-1911. MOORES, G. R. & LAWRENCE, A. J. (1972). Conductimetric assay of phospholipids and phospholipase A. FEBS Letters, Amsterdam 28, 201-204. PASTERNAK, C. A. & BERGERON, J. J. M. (1970). Turnover of mammalian phospholipids. Stable and unstable components in neoplastic mast cells. Biochem.J. 119, 473-480. SLACK, J. R., ANDERTON, B. H. & DAY, W. A. (1973). A new method for making phospholipid vesicles and the partial reconstitution of the (Na+, K+)-activated ATPase. Biochim. biophys. Acta 323, 547-559WADDELL, A., ROBSON, R. T. & EDWARDS, J. (1974). Colchicine and vinblastine inhibit fibroblast aggregation. Nature, Lond. 248, 239-241. WARREN, L. & GLICK, M. C. (1968). Membranes of animal cells. II. The metabolism and turnover of the surface membrane. J. Cell Biol. 37, 729-746. WRIGHT, J. D. & GREEN, C. (1971). The role of the plasma membrane in fatty acid uptake by rat liver parenchymal cells. Biochem.J. 133, 837-844. GAHMBERG, {Received 9 January 1975) 372 A. S. G. Curtis, F. M. Shaw and V. M. C. Spires Fig. 3. Negatively stained plasmalemmal fraction seen by electron microscopy. Stained 4 % ammonium molybdate. By kind permission of Dr E. Follett. Fig. 4. Fluorescence (blue light) view of an aggregate of neural retina cells showing ring staining. FITC-labelled sheep anti-rabbit immunoglobulin applied after rabbit antineural retina (whole cell) antibody. Cell surface Upids and adhesion. It 373