ABSTRACT BIOCHEMISTRY AND METABOLISM OF MAMMALIAN BLOOD GLYCOSPHINGOLIPIDS BY Robert V. P. Tao The objective of this research was to chemically charac- terize the platelet glycosphingolipids and plasma gangliosides as well as to study the metabolic turnover of sphingolipids in a pig induced with reticulocytosis. Identifications of four neutral glycosphingolipids were made on the basis of sugar molar ratios, studies of permethyl- ation products, and the action of stereospecific glycosidases on these lipids. Lactosylceramide, the most abundant type, accounted for 64% of the total neutral glycolipid mixture. The platelets were rich in a ceramide fraction, representing 1.3% of the total platelet lipids. The neuraminic acid component of hematoside was N-acetylneuraminic acid. Treat- ment of platelets with trypsin, chymotrypsin or thrombin increased the yield of hematoside as compared with a control, while the level of ceramides was not changed. In contrast to human platelets, porcine platelets con- tained trihexosylceramide as the major neutral glycosphingo- lipid. Sulfatide, hematoside and ceramides were also detected. Robert V. P. Tao An investigation of the biosynthesis of glycosphingo- lipids in reticulocyte-rich blood was conducted by incubating [14C]glucose with the whole blood in_vit£9_at 37°C for 2 hr. The results suggested that reticulocyte-rich red cells can synthesize glucosylceramide and, to some extent, lactosyl- ceramide as well. Radioactivity was also detected in the hexose moiety of plasma glucosylceramide, suggesting a possible exchange between the plasma and erythrocyte glucosylceramide pools. A metabolic experiment was conducted to study the turn- over of sphingolipids in an anemic pig $2.23X2 by injecting [14C]glucose into the pig intravenously as a pulse label, and removing aliquots of blood samples for lipid analyses at frequent intervals throughout a period of 81 days. Analyses were made on plasma, mixed population of red cells, and red cells that had been fractionated into individual groups of cells according to age by density gradient ultracentrifuga- tion. The results suggested the radioactive glucose was rapidly incorporated into the membrane-bound globoside of the im- mature erythrocytes in the bone marrow. After being released from the marrow these cells lost a portion of their globo- side-containing membrane as they matured in the peripheral circulation. The remodeling of the cells within the circula- tion continued until it approached the size of a normal adult Robert V. P. Tao cell, after which time the turnover of globoside remained rather constant. The membrane-bound globoside remained with the cell until the time of red cell senescence, and then was released directly into the circulation as a whole unit before cell destruction. It is postulated that this is a major source of all four plasma neutral glycosphingolipids. Erythrocyte glucosylceramide and lactosylceramide did not follow the normal expected red cell survival; instead, they appeared to be in dynamic equilibrium with the plasma glyco- sphingolipids. A semilogarithmic plot of specific activity versus time emphasized the biphasic nature of the decay curves of all the sphingolipids studied, suggesting that there were at least two major pools in each of these lipid fractions. Similar half-times of 5.5 and 45.0 days were observed for both erythro- cyte GL—3a and GL-4. When globosides from the fractionated red cell bands were examined, half-times of 0.75, 1.0 and 4.0 days were obtained from the rapid turnover pools of Bands 1, 2, and 3, corre- sponding to turnover rates of 65.43, 47.04 and 11.50 umoles/ day; whereas 3.3, 5.5, and 9.3 days were obtained from the slow turnover pools of these respective bands which corre- sponded to turnover rates of 14.78, 11.90 and 6.89 umoles/day. Biphasic decay curves were also observed for both plasma GL-la (t 1/2=0.75, 7.5 dayS) and GM (t l/2=O.9, 1.9 days). 3 Robert V. P. Tao Approximately 93% and 9% of GL-la and 77% and 36% of GM3 were found to be metabolized each day. On the contrary, only approximately 21% was synthesized each day for plasma GL-2a, GL-3a and GL-4. BIOCHEMISTRY AND METABOLISM OF MAMMALIAN BLOOD GLYCOSPHINGOLIPIDS BY Robert V) PT Tao A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Biochemistry 1973 FOREWORD I would like to express my deepest and most sincere appreciation to my research director, Dr. Charles C. Sweeley, for his inestimable guidance, encouragement, and continual enthusiasm throughout the graduate experience. I would also like to thank Drs. Anthony J. Bowdler and Robert W. Bull of the Department of Medicine for the use of their laboratory in performing the hematological analyses throughout the in_yiyg experiment and the many inspiring and helpful dis- cussions concerning this study. I am extremely grateful to Dr. Elwyn R. Miller and his colleagues for providing and maintaining the experimental animal and for the blood samples which made it possible to perform this study. I am very much indebted to my close friend, Dr. Walter J. Esselman, for his valuable help in the density gradient experiment. I would also like to express my appreciation to Dr. Graham A. Jamieson of The American National Red Cross Blood Research Laboratory for providing the platelet lipid samples. Finally, I wish to express my sincere appreciation for the help I have received from Drs. Roger A. Laine, Ray K. Hammond and other associates from this laboratory and the Department of Biochemistry dur- ing this study. ii TABLE OF CONTENTS LIST OF TABLES O O O O O O O O O O O O O O O I O O 0 LIST OF FIGURES . O O O O C O O O O O O O O O C 0 LIST OF ABBREVIATIONS . . . . . . . . . . I. LITERATURE REVIEW. . . . . . . . . A. Sphingosine . . . . . . . . . B. Ceramide. . . . . . . . . C. Neutral Glycosphingolipids. . . l. Monohexosylceramide. . . . . . . . . . . a. Galactosylceramide (Galactocerebro- side, GL-lb). . . . . . . . . . . . b. Glucosylceramide (Glucocerebroside, G1-la). . . . . . . . . . . . . . . 2. Dihexosylceramides . . . . . . . . . a. Digalactosylceramide (Gl- -2b). . . . b. Lactosylceramide (GL- 2a). . . 3. Trihexosylceramides. . . . . . . . . a. Galactosyl- galactosyl-glucosylcer— amide (GL- 3a) . . . . . . . . b. Galactosyl- galactosyl- galactosylcer— amide (GL-3b) . . . . . . . . c.-N-Acetylgalactosaminy1- galactosyl— glucosylceramide. . . . . . . . . d. N- -Acetylglucosaminy1- galactosyl— glucosylceramide. . . . . . . . . . 4. Tetrahexosylceramides. . . . . . . . . a. N-Acetylgalactosaminyl- galactosyl- galactosyl- glucosylceramide (GL- 4). b. Cytolipin R . . . . . . . . . . . . c. Galactosyl-N-acetylgalactosaminyl- galactosyl-g1ucosylceramide . . . d. Galactosyl-N- acetylglucosaminyl- galactosyl- glucosylceramide . . . . 5. Pentahexosylceramides. . . . . . . . . a. Forssman hapten . . . . . . . . . iii Page m U1hk‘h4 ta OKDCIJQCD 11 11 ll 12 12 l3 l4 l4 l4 l4 TABLE OF CONTENTS—-continued II. b. N-acetylgalactosaminyl-N-acety1- galactosaminyl- galactosyl-galac- tosyl- glucosylceramide. . . . . . c. Galactosyl- galactosyl- N— acetylgalac- tosaminyl- galactosyl- glucosylcer- amide . . . . . . . . . . . . . . . d. Fucose-containing pentahexosyl- ceramides . . . . . . . . . . . . . (l) Lea hapten. . . . . . . . . . (2) X-hapten. . . . . . . . . . . (3) Blood group H substance . . . e. Polyhexosylceramides. . . . . . . . (l) Leb heptens . . . . . . . . . (2) Blood group A substance . . . (3) Blood group B substance . . . D. Acidic Glycosphingolipids . . . . . . . . . . 1. Sulfatides . . . . . . . . . . . . . . a. Ceramide monohexosyl sulphate . . . b. Ceramide dihexosyl sulphate . . . . 2. Gangliosides . . . . . . . . . . . . . . E. Biosynthesis of Neutral and Acidic Glyco- sphingolipids . . . . . . . . . . . . . . . . 1. Sphingosine. . . . . . . . . . . . . . . 2. Ceramide . . . . . . . . . . . . 3. Neutral glycosphingolipids . . . a. Galactosylceramide (GL-lb). . . . . -b. Glucosylceramide (GL-la). . c. Lactosylceramide (GL-2a). . . . . . d. Digalactosylceramide (GL-2b). . . . e. Trihexosylceramide (GL-3a). . . . . f. Globoside (GL-4). . . . . . . . . . g. Blood group substances. . . . . . . 4. Acidic glycosphingolipids. . . . . . . . a. Sulfatides. . . . . . . . . . . . . b. Gangliosides. . . . . . . . . . . . F. Turnover Studies. . . . . . . . . . . . . G. Catabolic Degradation of Glycosphingolipids . INTRODUCTION 0 O O O O O O O O O O O O O O O O O O I I I 0 EXPERIMENTAL . O O O I O O O O O O O I O O O O O O A. Materials . . . . . . . . . . . . . . . . . . 1. Non-chemicals. . . . . . . . . . . . . . 2. Chemicals. . . . . . . . . . . . . . . . B. Methods . . . . . . . . . . . . . . . . . . . iv Page 16 16 16 16 17 17 18 18 18 19 19 19 19 20 20 24 24 25 25 25 26 26 27 27 28 28 30 30 30 32 35 42 46 46 46 47 52 TABLE OF CONTENTS--continued U‘Iwai-J o o o 15. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. Human blood preparation . . . . . . . . . Human platelet preparation. . . . . . . . Porcine platelet preparation. . . . . . . Extraction of total lipids. . . . . Isolation of neutral glycosphingolipids (Silicic acid chromatography) . . . . . . Mild alkaline hydrolysis. . . . . . . . . Thin-layer chromatography . . . . . . . . Acid-catalyzed methanolysis . . . . . . . Gas-liquid chromatography . . . . . . . . Isolation and characterization of gang- liosides. . . . . . . . . . . . . . Identification of fatty acid methyl esters. . . . . . . . . . . . . . . . . . Identification of Sphingosine bases . . . Identification of N-acylneuraminic acid . Identification of ceramide. . . . . . . . Linkage studies . . . . . . . . . . . . . Anomerity study . . . . . . . . . . . . . Isolation and characterization of plate- let sphingomyelin . . . . . . . . . . . . Platelet phospholipids. . . . . . . . . . Platelet neutral lipids . . . . . . . . . Preparation of radioactive glucocerebro- side. . . . . . . . . . . . . . . a. Hydrolysis of G1- 1 by barium hydrox- ide. . . . . . . . . . . . . . . . . b. Coupling of glucosylsphingosine with [1- C] stearic acid . . . . . . . . Experiment with fetal pigs. . . . . . . . a. 45-day fetuses . . . . . . . . . . . b. 90-day fetuses . . . . . . . . . . . c. Analysis of blood samples. . . . . . Red Cell fractionation technique. . . . . a. Albumin solutions. . . . . . . . . . b. Preparation of gradients . . . . . . c. Ultracentrifugation. . . . . . . . Separation of young and mature red cells from a normal and anemic dog. . . . . . . Separation of young and mature erythro- cytes from a normal and an anemic pig . . Checking hemolysis of the erythrocytes in albumin solutions . . . . . . . . . . . . Red cell size distribution. . . . . . . . In vivo pig experiment. . . . . . . . . . a. Induction of anemia (reticulocytosis) b. Induction of anemia in pig 123—6 . . Page 52 52 53 55 56 56 57 58 59 60 61 62 63 63 64 65 66 67 68 69 69 7O 71 72 72 72 73 73 74 75 75 76 76 77 77 78 TABLE OF CONTENTS-~continued Page c. Radioisotope administration . . . . 79 d. Treatment of blood samples. . . . . 80 e. Red cell fractionation. . . . . . . 83 27. In vitro study . . . . . . . . . . . . . 85 IV. RESULTS . . . . . . . . . . . . . . . . . . . . . 86 A. Isolation, Purification and Characterization of Human Plasma GM3 Ganglioside . . . . . . . 86 1. Discovery of human plasma ganglioside. . 86 2. Search for a solvent system. . . . . . . 87 3. Separation of sialic acid-containing glycosphingolipid and globoside from human plasma . . . . . . . . . . . . . . 88 4. Identification of lipid B by gas—liquid chromatography . . . . . . . . . . . . . 88 5. Identification of sialic acid derived from lipid B . . . . . . . . . . . . . . 94 6. Linkage studies of human plasma hemato- side . . . . . . . . . . . . . . . . . 100 7. Hematoside from Folch upper phase. . . . 104 8. Fatty acid composition of human plasma hematoside . . . . . . . . . . . . . . 107 9. Quantitative estimation of the sugars by gas- liquid chromatography. . . . . . . . 107 10. Other human plasma gangliosides from upper phase. . . . . . . . . . . . . . . 109 B. Isolation, Purification and Characterization of Human Platelet Sphingolipids . . . . . . . 112 1. Human platelet concentrates. . . . . . . 112 2. Lipid composition of human platelets . . 112 3. Identification of platelet glycosphingo- lipids by thin-layer chromatography. . . 114 4. Identification of platelet glycosphingo- lipids by GLC. . . . . . . . . . . . . . 119 5. Identification of sialic acid from ' platelet lipid V . . . . . . . . 119 6. Linkage studies of platelet glycosph1ngo- lipids . . . . . . . . . . . . . . . . . 125 a. Lipid I (ceramide monohexoside) . . 125 b. Lipid II (ceramide dihexoside). . . 125 c. Lipid III (ceramide trihexoside). . 133 d. Lipid IV (ceramide tetrahexoside) . 133 e. Lipid V (sialo- dihexosylceramide) 134 7. Enzymatic hydrolysis of platelet neutral glycosphingolipids . . . . . . . . . . . 134 a. Platelet GL-4 . . . . . . . . . . . 137 vi TABLE OF CONTENTS--continued Page b. Platelet GL-3a. . . . . . . . . . . 137 c. Platelet GL- 2a. . . . . . . 137 8. Fatty acid composition of platelet glycosphingolipids . . . . . . . . . . . 138 9. Concentrations of platelet glycosphingo- lipids . . . . . . . . . . . . . . . . . 140 10. Platelet gangliosides. . . . . . . . . . 140 11. Platelet sphingolipids . . . . . . . . . 143 a. Ceramide. . . . . . . . . . . . . . 143 b. Sphingomyelin . . . . . . . 144 12. Sphingolipid content of trypsin- -treated and non- -treated platelets. . . . . . . . 144 13. Platelet phospholipids . . . . . . . 145 14. Fatty acid composition of platelet phospholipids. . . . . . . . . . . . . 146 15. Platelet neutral lipids. . . . . . . . 147 C. Isolation and Quantitative Determination of Porcine Platelet Glycosphingolipids . . . . . 147 1. Porcine platelet concentrates. . . . . . 147 2. Porcine platelet ceramides . . . . . . . 148 3. Glycosphingolipid content of porcine platelets. . . . . . . . . . . . . . . . 148 D. Globoside Concentration in Fetal Pigs . . . . 151 E. Radioactive Glucocerebroside. . . . . 152 1. Isolation of glucosylsphingosine? . . 152 2. Radioactive glucocerebroside ([ 4C] stearic acid). . . . . . . . . . . 155 3. Proof of the radioactive glucocerebro- side . . . . . . . . . . . . . . . . . . 155 4. In vivo experiment . . . . . . . . . . . 158 F. In Vitro Study. . . . . . . . . . . . . . . . 161 G. Red Cell Fractionation. . . . . . . 165 1. Separation by age of erythrocytes .from normal and anemic blood. . . . . . . . . 165 2. Separation of young and mature erythro- cytes from normal and anemic pig blood . 168 3. Electronic sizing of fractionated red cells. . . . . . . . . . . . . . . . . . 169 4. Checking hemolysis of erythrocytes in albumin solutions. . . . . . . . . . . . 169 H. In Vivo Studies . . . . . . . . . . . . . . . 172 —1. Induction of anemia in pig 123- 6 . . . . 172 2. Administration of radioisotope . . . . . 178 3. Concentration of porcine blood sphingo- lipids . . . . . . . . . . . . . . . . . 178 4. Metabolism of porcine blood sphingo- lipids . . . . . . . . . . . . . . . . . 191 vii TABLE OF CONTENTS--continued 1 l 1 1 l 5. l. 2. 3. 4. 5. Incorporation of labeled glucose into GL-4. . . . . . . . . . . . . . . . . . Incorporation of labeled glucose into GL-3a . . . . . . . . . . . . . . . . . Incorporation of labeled glucose into GL-Za . . . . . . . . . . . . . . . . . Incorporation of labeled glucose into GL-la . . . . . . . . . . . . . . . . . Incorporation of labeled glucose into Folch lower phase GM ganglioside . . . Incorporation of labeled glucose into Folch upper phase gangliosides. . . . . Incorporation of labeled glucose into ceramides . . . . . . . . . . . . . . . Fractionation of red cells according to age from pig 123- 6 during in vivo study Globoside (GL- 4) concentrat1on in por- cine red cells of different ages. . . . Globoside (GL—4) turnover in the frac- tionated red cells. . . . . . . . . . . Turnover values of plasma and erythro- cyte glycosphingolipids . . . . . . . . V. DISCUSSION. 0 O O O O O O C O O O O O I O O O O O A. Human Platelet Sphingolipids and Plasma Gangliosides . . . . . . . . . . . . . B. Porcine Platelet Glycosphingolipids. C. Fetal Erythrocyte GL- 4 . . . . . . . D. Metabolic Study in Pig 123- 6 . . . . . 'VI. SUMMARY REFERENCES. APPENDIX--LIST OF PUBLICATIONS. . . . . . . . . . . . viii Page 192 204 205 210 213 216 220 225 225 225 241 244 244 256 257 259 289 294 314 LIST OF TABLES TABLE 1. Gangliosides of Mammalian Brain. . . . . . . . . 2. Rations Pan-fed to Pig 123-6 . . . . . . . . . . 10. ll. 12. 13. 14. 15. Relative Retention Behavior of Trimethylsilyl Methyl Glycosides. . . . . . . . . . . . . . . . Concentration of GM3 Ganglioside and Globoside in Human Plasma. . . . . . . . . . . . . . . . . Retention Times of Permethylated Methyl Glyco- sides from Human Plasma Ganglioside. . . . . . . Retention Times of Partially Methylated Alditol Acetates from Human Plasma Ganglioside . . . . . Fatty Acid Composition of Plasma GM Ganglioside 3 Total Lipid Composition of Human Platelets . . . Concentration of Glycosphingolipids in Human PlateletS. O O O I O O O I O O I O I O O O O O O Retention Times of Permethylated Methyl Glyco- - sides from Platelet Glycosphingolipids . . . . . Retention Times of Partially Methylated Alditol Acetates from Platelet GlyCOSphingolipids. . . . Mass Spectrometric Identification of Methylated Alditol Acetates from Platelet Glycosphingo- lipids . O O O O O O O I I O O O O O O O O O O 0 Fatty Acid Composition of Human Plasma, Erythro— cyte and Platelet Glycosphingolipids . . . . . . Fatty Acid Composition of Human Platelet Ganglio- sides from Folch Upper Phase . . . . . . . . . . Concentration of Ceramide and Hematoside in Platelets Treated with Proteolytic Enzymes . . . ix Page 22 78 93 97 103 103 108 113 122 128 131 132 139 142 145 LIST OF TABLES--continued TABLE 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. Concentration of Glycosphingolipids in Porcine Platelets. O O O O O O O O O O O O O O O O O 0 Concentration of Globoside in Fetal Pig Erythro- cytes O C O O O O O O O O O O O 0 O O O O O O O O Incorporation of [14C]G1ucose into Neutral Glyco- sphingolipids of Pig Erythrocytes 20 Days post- Induction of Anemia. . . . . . . . . . . . . . . Incorporation of [14C]G1ucose into Neutral Glyco- sphingolipids of Pig Plasma 20 Days Post-Induc- tion of Anemia . . . . . . . . . . . . . . . . . Discontinuous Density Gradient Ultracentrifuga- tion of Canine Blood . . . . . . . . . . . . . . Hemoglobin Concentration of Individual Fractions Derived from Normal and Anemic Porcine Blood . . Mean Channel Number of Individual Bands Fraction- ated by Density Gradient Ultracentrifugation . . Hemoglobin Content of the Supernatant Solutions. Soret Band of the Supernatant Solutions. . . . . Concentrations of Plasma and Erythrocyte Sphingo- lipids from Pig 123-6 During the Metabolic Experiment . . . . . . . . . . . . . . . . . . . Specific Activities (cpm/umole) of Plasma and Erythrocyte Sphingolipids. . . . . . . . . . . . Specific Activities (cpm/ml) of Plasma and Erythrocyte Sphingolipids. . . . . . . . . . . . Random Errors for Various Steps in the Procedure Specific Activities of Hexose, Fatty Acid, and Sphingosine Moieties from Plasma GL-la and GL-4. Concentration and Specific Activities of Globo- side from the Fractionated Red Cell Bands During the Metabolic Experiment . . . . . . . . . . . . Page 151 152 162 163 168 170 171 171 171 186 189 190 191 193 228 LIST OF TABLES--continued TABLE Page 31. Turnover Values of Porcine Plasma and Erythro— cyte Glycosphingolipids. . . . . . . . . . . . . 242 32. Percentage of Blood Removed from Pig 123-6 Between Day 3 and Day 10 . . . . . . . . . . . . 264 33. Time of Reappearance of Label in Sphingolipids 284 During the In Vivo Study . . . . . . . . . . . . xi FIGURE 1. 2. 10. 11. LIST OF FIGURES Page Thin-layer chromatography of hematoside from human plasma. . . . . . . . . . . . . . . . . . . 90 Gas-liquid chromatography of trimethylsilyl methyl glycosides . . . . . . . . . . . . . . . . 92 Retention times of trimethylsilyl methyl glyco- sides from ganglioside of normal human plasma . . 96 Thin-layer chromatography of neuraminic acid derived from human plasma ganglioside . . . . . . 99 Gas-liquid chromatography of permethylated methyl glycosides from lactose, N-acetylneuraminyl- lactose, and human plasma ganglioside . . . . . . 102 Gas-liquid chromatography of partially methylated alditol acetates from N-acetylneuraminyllactose and human plasma ganglioside. . . . . . . . . . . 106 Thin-layer chromatography of Folch upper phase gangliosides from normal human plasma and platelets . . . . . . . . . . . . . . . . . . . . 111 Thin-layer chromatography of sphingolipids from normal human plasma . . . . . . . . . . . . . . . 116 Thin-layer chromatography of glycosphingolipid fraction from washed normal human platelets treated with trypsin. . . . . . . . . . . . . . . 118 Gas-liquid chromatography of trimethylsilyl methyl glycosides from major neutral glycosphingo- lipids of normal human platelets treated with trypsin . . . . . . . . . . . . . . . . . . . . . 121 Gas-liquid chromatogram of trimethylsilyl methyl glycosides from G ganglioside of normal human platelets treated with trypsin. . . . . . . . . . 124 xii LIST OF FIGURES--continued FIGURE 12. Gas-liquid chromatography of permethylated 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. methyl glycosides derived from lactose, Fabry trihexosylceramide (CTH), porcine erythrocyte globoside (GL—4), neuraminyllactose, and platelet glycosphingolipids. . . . . . . . . . . Gas-liquid chromatography of partially methyl- ated alditol acetates derived from lactose, neuraminyllactose, Fabry trihexosylceramide (CTH), porcine erythrocyte globoside (GL-4), and platelet glycosphingolipids. . . . . . . . . . Thin-layer chromatography of platelet glyco- sphingolipids and hydrolysis products by various glycosidases . . . . . . . . . . . . . . . Thin-layer chromatography of ceramides from normal porcine platelets . . . . . . . . . . . . Thin-layer chromatography of the hydrolysis product from Gaucher spleen glucosylceramide Gas-liquid chromatography of trimethylsilyl methyl glycosides and sphingosines from [ C] glucosylceramide . . . . . . . . . . . . . . . . Gas-liquid chromatography of fatty acid methyl ester from [14C]glucosylceramide . . . . . . . . Separation of young and mature erythrocytes from normal and anemic dogs . . . . . . . . . . . . . DevelOpment of anemia in pig 123-6 during the in- duction period and throughout the in_vivo study. Photomicrographs of blood cells from pig 123-6 during induction of anemia . . . . . . . . . . . Photomicrographs of blood cells from pig 123- 6 during the in vivo study . . . . . . . . . . . Thin-layer chromatography of erythrocyte cera- mides from pig 123-6 . . . . . . . . . . . . . . xiii Page 127 130 136 149 153 156 159 166 174 177 180 183 LIST OF FIGURES-~continued FIGURE 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. Thinwlayer chromatography of plasma ceramide from pig 123-6. . . . . . . . . . . . . . . . Thin-layer chromatography of Folch upper phase gangliosides from normal human erythrocytes, normal porcine plasma and erythrocytes. . . . . Turnover of plasma and erythrocyte globoside in pig 123-6 0 O I O I O O O O O O O O I O O O O O Turnover of red cell globoside in pig 123-6 over a lO-day period. . . . . . . . . . . . . . Turnover of erythrocyte globoside in pig 123-6. Turnover of plasma and erythrocyte trihexosyl- ceramide in pig 123-6 . . . . . . . . . . . . . Turnover of erythrocyte trihexosylceramide in pig 123-6 . . . . . . . . . . . . . . . . . . . Turnover of plasma and erythrocyte lactosyl- ceramide in pig 123-6 . . . . . . . . . . . . . Turnover of plasma and erythrocyte glucosyl- ceramide in pig 123- 6 . . . . . . . . . . . . Turnover of Folch lower phase GM3 ganglioside in plasma and erythrocyte of pig 123-6. . . . . Turnover of Folch upper phase gangliosides from plasma and erythrocyte of pig 123-6 . . . . . . Turnover of plasma and erythrocyte ceramides in pig 123-6 . . . . . . . . . . . . . . . . Turnover of erythrocyte ceramides in pig 123-6. Separation of young and mature erythrocytes from pig 123-6 during the in vivo study . . . . Distribution of radioactivity of red cell globoside from pig 123-6 at time interval after the initial pulse label . . . . . . . . . . . . xiv Page 185 188 195 199 201 203 207 209 212 215 218 222 224 227 232 LIST OF FIGURES--continued FIGURE Page 39. Bar graph presentation of the red cell globo- side specific activity from pig 123-6 at vari- ous time intervals after administration of the label. . . . . . . . . . . . . . . . . . . 235 40. Globoside turnover in individually fractionr ated groups of porcine erythrocytes. . . . . . 237 41. Globoside turnover in the tOp three bands of porcine erythrocytes fractionated by density gradient ultracentrifugation . . . . . . . . . 240 42. Loss of labeled GL—4 from individual bands as a function of time . . . . . . . . . . . . . . 273 XV LIST OF ABBREVIATIONS Acid Citrate Dextrose N-Acetylgalactosamine N-Acetylglucosamine N-Acetylneuraminic Acid Adenosine Diphosphate Adenosine Triphosphate Adenosine Triphosphatase Ceramide containing Hydroxy Fatty Acids Ceramide containing Non-hydroxy (normal) Fatty Acids Cytosine Monophosphate-NANA Fatty Acid Methyl Ester Fucose Galactose Gas-liquid Chromatography Gas-liquid Chromatography and Mass Spectrometry Glucose N-Glycolylneuraminic Acid Guanosine Diphosphate-Fucose Infrared Spectrosc0py Immunoglobulin E Mild Alkaline Hydrolysis Nuclear Magnetic Resonance 3'-Phospho-adenosine-5'-phosphosulfate Red Blood Cell Uridine Diphosphate-N-acetylgalatosamine Uridine Diphosphate-Galactose Uridine Diphosphate-Glucose xvi ACD GalNAc GlCNAc NANA ADP ATP ATPase Ceramide-HFA (Cer-HFA) Ceramide-NFA (Cer-NFA) CMP-NANA FAME Fuc Gal GLC GLC-MS Glc NGNA GDP-Fuc IR IgE MAH NMR PAPS RBC UDP-GalNAc UDP-Gal UDP-Glc I. LITERATURE REVIEW A. Sphingosine Much of our present knowledge in the field of glyco- sphingolipids is due to the pioneering work in brain chemistry by Ludwig Wilhelm Thudichum (1829-1901), who discovered spingosine (l) named after sphingein (Gr. "to bind tight"). His work was followed by many distinguished scientists such as Thierfelder, Levene, Rosenheim, and Klenk. Until 1940 the structure of Sphingosine was only partly known as an unsatu- rated‘dihydroxyamine with a long—chain carbon skeleton, and it was not until 1947 that the correct structure of sphingo- sine [I] was established by Carter and his collaborators (2). This achievement has paved the way for all the structural studies as well as studies on biosynthesis and metabolism of sphingolipids in the ensuing years. The term "Sphingolipid" was also introduced by Carter to designate the lipids derived from the parent base Sphingosine. CH3-(CH2)12-CH=CH-$H-$H-CH20H [I] OH NH 2 B. Ceramide Ceramide [II] is the trivial name for N-acylsphingosine; the name was derived from the waxy texture (cerum) of these substances (3). They are the only simple derivatives of sphingosines. The fatty acids are bound to the amino group of the long-chain base in amide linkage. Lignocerylsphingosine and cerebronylsphingosine are the outstanding representatives. Thudichum (4) was the first to isolate ceramide and psychosine from cerebroside [III] as degradative products after partial CH3-(CH2)lz-CH=CH—CH-CH20H NH 62:. 1'. [II] hydrolysis under different conditions. Alkaline hydrolysis gave fatty acids and psychosine (galactosylsphingosine)[IV], whereas mild acid treatment cleaved the glycosidic bond with the formation of ceramide. CH3’(CH2)12—CH=CH-gH—CH- OH" H NHZ CHZ—O-Gal + RCOOH[IV] CH3-(CH2)l2-CH=CH-CH-CH—CH2-O-Ga1¥\\\‘ 0H NH H CH —(CH.) —CH=CH-cH-cH- I 3 2 12 c=o OH H ) §=0 [III] R CHZOH + Gal This indicates that the fatty acid is in amide linkage and the hexose in glycosidic bond with Sphingosine. A number of investigators have isolated free ceramide from liver (5), spleen (6), and lung (7). Klenk (8) found lignocerylsphingo- sine in the red cells and Shimojo (9) also demonstrated the presence of ceramide in pig erythrocytes. Most recently, ceramides were found in human kidney (10), brain (11,12), plasma (l3), aorta (l4) and pig brain (15). Tannhauser (16) suggested that the presence of N—acylsphingosine in these organs or tissues could serve as a possible precursor in the biosynthesis of cerebroside and sphingomyelin. This View was supported by recent evidence from Kennedy and co—workers (17, 18). Ceramides have been shown to be part of the molecule in two types of sphingolipids. Sphingophosphatide is character- ized by linking ceramide to another polar group via a phos- phate diester, such as sphingomyelin [V]. It was discovered in human brain by Thudichum (1). Besides brain and nervous tissues, Sphingomyelin was also found to be present in human 0 CH3- (CH2) lz--CH=CH~.—lCH-fH-CHZ--O--£ll’--O--CH2--CH2-N+ (CH3) 3 [V] OH NH OH C=O R plasma and erythrocytes in rather large amounts. Ceramide phosphorylethanolamine [VI] (another variation of sphingo— myelin) and ceramide aminoethyl phosphonates [VII] (ceramide was linked via a phosphonate ester) were found to be present in marine invertebrates (19,20) as well as rumen protozoa (21) and the blowfly Calliphora erythrocephala (22). CH3-(CH2)12-CH=CH-CH-CH-CH2-O-P-OCH2-CH2NH2 [VI] OH NH H F“) R ‘3 - = - - — H -o—P-CH -CH NHCH CH3(CH2)12 CH CH CH CH C 2 l 2 2 3 [VII] OH NH OH I C=O I R The second class of sphingolipids is characterized by the presence of a glycosidic linkage between a carbohydrate unit (mono- or oligosaccharide) and the primary hydroxyl group of the ceramide, ceramide-O-carbohydrate. The carbo- hydrate chain may contain amino sugars and NANA or NGNA. Various neutral glycosphingolipids, sulfatides, gangliosides and blood group substances are representatives of this group. C. Neutral GlyCOSphingolipids The neutral glycosphingolipids contain one or more gly- cosyl moieties. They can be divided into monohexosylceramides, dihexosylceramides, trihexosylceramides, tetrahexosylceramides, pentahexosylceramides and polyhexosylceramides. Compounds containing amino-sugars are sometimes called aminoglycolipids. The individual compounds are often named as derivatives of ceramides. For example, a dihexosylceramide containing galac- tose and glucose in known sequence is called galactosyl- glucosylceramide. The lipid may be called lactosylceramide if the position and linkages are known, i.e., galactosyl- (81+4)-glucosyl-(Bl+l')-ceramide. l. Monohexosylceramide a. Galactosylceramide (Galactocerebroside, GL-lb) Thudichum (23) was the first to isolate this group of compounds and gave the name cerebrosides to a group of glyco- sidic substances found in the brain containing 1 molecule each of Sphingosine, hexose and fatty acid. The hexose component was identified as galactose by Thierfelder (24). The attach- ment of sugar to the C—1 position of the Sphingosine moiety was demonstrated by Carter and Greenwood (25). Evidence that the glycosidic linkage has the 8 configuration was presented on the basis of its susceptibility to hydrolysis by B-galac- tosidase (26) and IR measurements (27). Shapiro and co- workers further confirmed the structural analyses of brain cerebrosides by total synthesis of psychosine (28) and galac- tocerebroside (29). Brain cerebrosides contain four major fatty acids. The material first isolated by Thudichum was rather impure at that time; however, he was able to distinguish between the two major compounds of this class, cerasine [VIII] and phrenosine [IX], which were later found to contain ligno- ceric acid, by Levene (30), and a-hydroxylignoceric acid, by Thierfelder (24) and Klenk (31). Besides cerasine and phrenosine, a third cerebroside was obtained from beef brain CH3-(CHé);2-CH=CH—CH-CH-CH2-O-Gal OH NH _ ($H2)22 CH3 [VIII] — — = — — — —0— CH3 (CH2)12 CH CH $H 9H CH2 Gal 0H NH C=O I CH-OH I (OH2)21 CH3 [Ix] and identified by Klenk (32), who named it nervone [X]. The same author also postulated a fourth cerebroside, oxy- nervone [XI], containing d-hydroxynervonic acid. The sphingo- sine bases consisted predominantly of C -sphingosine (98%) 18 and a small quantity-ofClB-dihydrosphingosine (2%) (33). CH3-(CH2)12-CH=CH-CH-CH-CH -O-Ga1 I I 2 OH NH C=o I (CH ) in 13 [X] CH I (CH2)7-CH3 CH3“ (CH2) lZ-CH=CH-$H-ICH-CH2-O-Gal OH NH l C=0 l CIIH-OH (CH2)12 FH A (CH2)7-CH [XI] 3 Galactosylceramide was found exclusively in the human brain; this lipid accounted for 10-25% of the total lipids from the glial cells in the nervous system (34). It has also been found in extraneural tissues and fluids, such as human kidney (35), intestine (35) and blood. In blood, a small amount of galactosylceramide had been detected in serum (36) (about 10% of the total monohexosylceramides) and a trace amount in human plasma (37). Miras 33 31. (38) reported that galactosylceramide was the monohexosylceramide present in human leukocytes; however, this was at variance with the find- ings of Hildebrand gt El. (39) and Kampine 33 31. (40) who found glucosylceramide as the only cerebroside present in leukocytes. This could be explained by the fact that earlier authors based their conclusions on co-chromatography of the isolated material with a galactosylceramide standard in a con— ventional TLC system which was unsuitable for resolving glucosylceramide from galactosylceramide. Better resolution could be achieved if borate-impregnated silica gel G plates were used (41). b. Glucosylceramide (Glucocerebroside, Gl—la) Glucocerebroside was first isolated in 1940 from the spleen of a patient suffering from Gaucher's disease (26). Glucose instead of galactose was found to be present in this lipid (42). The structure was prOposed by Rosenberg and Chargaff (43) as l—O-(B-D-glucOpyranosyl)—N—docosanoyl—Qf erythro-sphingosine and confirmed by chemical synthesis (29, 44). Glycosylceramide had been found in a variety of living organisms. In human and pig plasma (37,45), Gl-la was the major neutral glyCOSphingolipid present, and accounted for about 48% of the total neutral glyCOSphingolipid contents. In erythrocytes (37,45), Gl-la represented 4% (human) and 5% (pig), respectively of the neutral glycosphingolipid frac- tion. In both human and pig leukocytes (38,45,46), Gl-la was also found to be a minor component. This lipid was shown to be elevated in both plasma and erythrocytes of patients with Gaucher's disease (47). 2. Dihexosylceramides a. Digalactosylceramide (Gl-2b) A digalactosylceramide was found in the kidney of a patient with Fabry's disease by Sweeley and Klionsky (48). The structure was shown to be Gal-(1+4)-Gal-(1+1')-ceramide (49), and recently the anomeric configurations of this lipid were assigned as Gal-(al+4)-Gal-(Bl+l')-ceramide with the aid of stereospecific glycosidases by Li e£_al, (50). In humans, GL-2b appears to be present only in kidney (51) and intes- tines (52), whereas in Fabry's disease abnormal amounts were found in kidney, urinary sediments and pancreas (53-55). GL-2b was never detected in plasma or erythrocytes of these patients. It was reported (38) to be present in human leuko- cytes; however, this finding could not be confirmed by several other investigators (39,40,52). Hence, the existence of GL-2b in leukocytes is questionable. There was also an early report on this lipid in the brain of a patient with Tay Sachs dis- ease (56). b. Lactosylceramide (GL—2a) In 1942, Klenk and Rennkamp (57) isolated a substance from bovine spleen and later from erythrocytes (58) , which was found to be composed of Sphingosine, fatty acid and two hexose residues (glucose and galactose). The structure of this lipid, isolated from erythrocytes and kidney by Yamakawa and co-workers (59), was shown to bear the following struc- ture Gal-(l+4)-Glc-(l+l)-ceramide. Besides human erythro? cytes, Gl-2a has been detected in human plasma (37), serum (60), bone marrow (61) and leukocytes (38-40) as well as porcine plasma (45), erythrocytes (45), and leukocytes (45,46). In human leukocytes (38), Gl-2a accounted for 16% by weight of the total lipids. 4-Sphingenine was the major long—chain base found in G1-2a of all the blood components, whereas the fatty acid compositions differed considerably. Recently, there 10 was a report on a patient with lactosylceramidosis, character- ized by high levels of Gl-2a accumulation in plasma, erythro— cytes, bone marrow cells and other visceral organs (52). 3. Trihexosylceramides a. Galactosyl:galactosyl-glucosylceramide (GL-3a) In 1953, Klenk and Lauenstein (58) first suggested the presence of a neutral trihexosylceramide, free of amino sugar, in human erythrocytes. Later, Svennerholm and Svennerholm (60) demonstrated that trihexosylceramide isolated from human serum, spleen and liver contained galactose and glucose in a molar ratio of 2:1. Trihexosylceramide had been identified as the pathologically accumulated glycosphingolipid in Fabry's disease by Sweeley and Klionsky (48). The same lipid was also elevated in the plasma of these patients but not in the red cells (47). Initially, the carbohydrate sequence of GL—3a was deduced as being Gal-Gal-Glc-ceramide upon partial hy- drolysis (48). Later, the same lipid was also isolated by Makita (62) and Martensson (51) from normal human kidneys and, on the basis of methylation studies, 1+4 linkages between the sugars were established by Makita and Yamakawa (63). Most recently, the structure of Gal-(1+4)-Gal-(l+4)-Glc-(l+l')- ceramide was confirmed by GLC of the methanolysis products before and after permethylation, periodate oxidation of GL-3a, and mild acid hydrolysis products (49). NMR studies coupled 11 with stereospecific enzymes (64-66) have established the anomeric configurations of the glycosidic bonds and gave evidence to support the following complete structure-- Gal-(al+4)-Gal-(81+4)-Glc-(81+l')—ceramide. Gl-3a was shown to occur in both the human and pig erythrocytes and also in the plasma fraction. In porcine leukocytes (46), G1—3a was the observed major neutral glycosphingolipid present. b. Galactosyl:galactosyl-galactosylceramide (GL-3b) This lipid has not yet been shown to occur in living organisms. It has been postulated (67), however, that a fur- ther addition of one molecule of galactose to the terminal galactose of digalactosylceramide might lead to the formation of this lipid, Gal-( ? )-Gal-(a1+4)-Gal-(Bl+1')-ceramide. It is possible that this material is present at a very low con- centration, which is not able to be detected by the analyti- cal methods employed at present. c. N-Acetylgalactosaminylegalactosyl—glucosylceramide This compound was first demonstrated in the brains of Tay-Sachs patients (56). However, recent studies have shown that the accumulation of this lipid was much more pronounced in Sandhoff's disease than in Tay-Sachs disease. d. N-Acetylglucosaminyl-galactosyl-glucosylceramide Recently, this lipid was shown to occur in human and cattle spleen (67) in rather large amounts. The structure was shown to be GlcNAc(Bl+3)-Gal-(81+4)-Glc—(l+l')—ceramide. 12_ 4. Tetrahexosylceramides a. N-Acetylgalactosaminyl-galactosyl—galactosyl- glucosylceramide (GL-4) In 1951, Klenk and Lauenstein (8) reported the isolation of a glycolipid from the human red blood cells which yielded galactosamine after acid hydrolysis. They found the compound to contain 3 hexose moieties (galactose and glucose in the proportion of 2:1) and l galactosamine (40%), fatty acid (29%) and Sphingosine (30%). This finding was later confirmed by Yamakawa and Suzuki (68) who showed the molar ratio of fatty acid: sphingosine:neutral hexose:N-acetylgalactosamine was l:l:3:l. Since the substance formed perfectly round globules (spherocrystals) under the microscope, it was called "Globoside", which is still the commonly used name today. Globoside is the most abundant neutral glycosphingolipid in human and pig red cell stroma (69). The structure of Gl-4 was first erroneously reported by Yamakawa g: 31. (70) as GalNAc-(1*6)-Gal-(1+4)-Gal-(164)-G1c-(lal)—ceramide. This error was mainly due to the misinterpretation of a gas chromatographic peak. However, after re-examination Yamakawa _E El. (71) proposed the correct structure of globoside, GalNAc-(1+3)-Gal-(1+4)-Gal-(l+4)-Glc-(1+l')—ceramide in 1965t Recently, on the basis of NMR, IR, use of stereospecific glycosidases and pmmmethylation studies, it was concluded that the exact structure of human red cell globoside is GalNAc- (81+3)-Gal-(al+4)-Gal-(8144)-Glc-(141)—ceramide (64). 13 In 1956, Matsumoto (72) reported that the chemical composition of the pig erythrocyte glycolipid consisted of galactose, glucose and galactosamine. This was further sug- gested by Shimojo et_al. (9) as being globoside. Later, Miyatake gt 31. (73) purified the glycolipid from pig red cells and discovered that the structure was actually identical to the human red cell globoside except for the fatty acid compo- sition. The same authors also demonstrated that the glyco- sidic bond linking N-acetylgalactosamine to galactose was in the 8 configuration, since the hexosamine was liberated by treatment of the lipid with a B-N—acetylhexosaminidase from pig epididymis tissue. The chemical structure of pig red cell globoside was postulated as being GalNAc-(Bl+3)-Gal- (81+4)-Gal-(81+4)-Glc-(Bl+l)-ceramide. Nevertheless, as a result of a recent study on the anomeric structure of human red cell globoside (64), it is believed that the correct structure for pig red cell globoside will probably be GalNAc- (Bl+3)-Gal-(ql+4)-Gal—(Bl+4)-Glc-(1+1)-ceramide. Recently, there were several reports (74-76) on Sandhoff's disease (variant of Tay-Sachs disease) which showed that these patients are characterized by extensive accumulation of Gl—4 in the visceral organs, with concomitant deficiency of B-N—acetyl- hexosaminidase activity. b. Cytolipin R In 1967, Rapport et 21' (77) isolated a glycosphingolipid from rat lymphosarcoma called cytolipin R. The lipid had the 14 same carbohydrate composition as G1-4, but differed from it in immunological activities and RF values on TLC. It was not until 1972 that the complete structure of cytolipin R, GalNAc-(81+3)-Gal-(al+3)-Gal—(81+4)-Glc-(Bl+l')-ceramide, was elucidated by Laine EE.E£‘ (78), based on the results obtained from the combination of linkage studies with GLC, GLC-MS and the uses of specific glycosidases. The difference in the immunological activities between cytolipin R and globoside were attributed to the difference of the internal linkage, -Gal-(dl+3)-Gal-, between the two lipids. c. Galactosyl-N-acetylgalactosaminyl-galactosyl— glucosylceramide This ceramide tetrahexoside was found in patients suffer- ing from G -gangliosidosis. This lipid is identical to M1 GMl-ganglioside in structure except that it lacks the neura- minic acid. d. Galactosyl-N-acetylglucosaminyl-galactosyl— glucosylceramide Unpublished results (67) have indicated the occurrence of Gal-(81+4)-G1CNAc-(81*3)-Ga1-(81+4)-Gal-(1+l)-ceramide in human and cattle spleen. 5. Pentahexosylceramides.: a. Forssman hapten Forssman was the first to report the immunologically active antigens in various organs of the guinea pig (79). 15 Later in 1939, Brunius (80) established the fact that Forssman hapten contained lipid and hexosamine. This was followed by the isolation of a glycosphingolipid containing hexose and hexosamine from sheep red cells which showed Forssman hapten activity by Papirmeister and Mallette (81). In 1966, Makita gg‘gl. (82) isolated Forssman hapten from horse kidney and spleen and showed that it possessed the same carbohydrate sequence (GalNAc-Gal-Gal-Glc-cer in l:1:1:l molar ratio) as GL-4 from human red cells and kidney, but differed in chrom- tographic behavior, Optical rotation and blood group activi- ties. The same authors further implicated that the determinant group of this hapten was probably the O-a-N-acetyl-galacto- saminyl-(le3)-galactosyl unit at the non-reducing end, differ- ing from GL-4 in the anomeric configuration of the disaccharide. This conclusion was later confirmed by Yamakawa and co-workers (83) from their studies involving NMR and stereOSpecific glycosidases, and the structure of GalNAc-(al+3)-Gal-(81+4)- Gal-(Bl+4)-Glc-(l+l)-ceramide was postulated. However, this was proven wrong recently by Siddiqui and Hakomori (84), who found that Forssman hapten is a ceramide pentahexoside contain- ing 2 moles of N-acetylgalactosamine, 2 moles of galactose, and 1 mole of glucose per ceramide. After hydrolyzing the terminal hexosamine with a-N-acetylgalactosaminidase from pig liver, they were able to obtain a compound identical to globoside in all respects, and which gave a positive precipitin reaction with antigloboside antiserum. With these results and the data 16 from methylation studies, a revised structure for Forssman hapten was proposed, GalNAc-(al+3)-Ga1NAc-(81+3)*Ga1—(a1+&)— Gal-(Bl+4)-Glc-(l+1)-ceramide. b. N-acetylgalactosaminyl-N-acetylgalactosaminyl- galactosyl-galactosyléglucosylceramide A very similar pentahexosylceramide was isolated from the dog intestine by Vance 32 a1. (85). Tentatively, the prOposed sequence of the carbohydrate chain was GalNAc- GalNAc-Gal-Gal—Glc-ceramide, which was further confirmed by McKibben (86). Details of the linkages between the mono- saccharide units and stereochemical configurations of the glycosidic bonds have not yet been established. 0. Galactosylegalactosyl-N-acetylgalactosaminyl--- galactosylsglucosylceramide. This ceramide pentahexoside was found to be the main component of rabbit erythrocytes and reticulocytes by Eto £5 31. (87). The structure proved to be Gal-(al+3)-Gal- (81+3)-GalNAc-(Bl+3)-Gal—(81+4)-Glc-(Bl+l')—ceramide. Since this lipid inhibited the agglutination of human B erythro- cytes with its corresponding antibody, the terminal galactose was concluded to be a-glycosidically bound (88). d. Fucose-containing:pentahexosylceramides, (l) Lea hapten A third ceramide pentahexoside had been isolated from human adenocarcinoma by Hakomori gt_§1. (89,90). The structure 17 of this lipid was proposed to beCkfl:(Bl+3)-GlcNAc-(1+3)- 1. 1 l Fuc Gal-(1+4)-Glc-(1+l)-ceramide (4) The carbohydrate backbone of this lipid is very similar to the tetrahexoside isolated from human and bovine spleen (see above) with the exception of a 1+3 linkage at the non-reducing end and, of course, the fucose residue which is attached to 'the N-acetylglucosamine residue in a 1+4 linkage. (2) X-hapten Another fucose-containing Sphingolipid with the following structure:Gal-(1+4)—GlcNAc-(1+3)-Gal-(l+4)-Glc-(l+1)-ceramide I) Fuc was isolated from erythrocytes and chemically characterized by Yang and Hakomori (91). This lipid contained a novel type of ceramide, which composed of 4—hydroxysphinganine and long- chain 2 hydroxy fatty acids, The carbohydrate moieties were similar to those of the Lea hapten in composition but there were differences at the terminal galactose (1+4 linked instead of 1+3) and fucose (1+3 linked instead of 1+4). This lipid exhibited no blood group A, B, H or Lewis specificities. (3) Blood group H substance The H substance, isolated from type 0 blood cells (92), was shown to contain glucose, galactose, fucose and l8 N-acetylglucosamine. The prOposed structure was Fuc-(1+2)- Gal-(l+4)-GlcNAc-(1+3)-Gal-(l+4)-Glc-ceramide. This lipid was active in the inhibition of both H and Leb hemagglutina- tion. e. Polyhexosylceramides (l) Leb'Haptens Two Leb haptens are known (88). One is a ceramide hexo- side with a linear arrangement of oligosaccharides as follows: Fuc-(1+2)-Gal-(l+3)-GlcNAc-(l+3)-Gal-(1+4)-Glc-ceramide 4 + 1 FL. Another Leb active glycosphingolipid has been proposed to be a ceramide octahexoside, Fuc-(1+2)-Ga1-(1+3)-GlcNAc-(l+3)-Gal-GlcNAc-(1+3)-Gal-(1+4)- Glc-ceramide 4 1 FL. (2) Blood group A substance Hakomori and Strycharz (92) obtained three different fractions from type A erythrocytes, all of which are immuno- logically active. All three glycolipids contain the basic carbohydrate contents of galactose, glucose, fucose, N-acetyl— galactosamine and N-acetylglucosamine, except one which contains additional sialic acid(s). The postulated sequence 19 for the A substance is GalNAc-Gal—G1cNAc—Gal—Glc-ceramide. Fuc (3) Blood group B substance Glycosphingolipid with blood group B activity was iso- 1ated from type B blood cells. It differs from the A sub- stance by possessing a galactose instead of a N-acetylgalac- tosamine at the non-reducing end of the carbohydrate chain, Gal-Gal-GlcNAc-Gal-Glc-ceramide. It has been mentioned that Fuc the serological specificities of the determinants from blood group-active glycolipids are very similar to the blood group-active glycoproteins; hence, one could speculate that the linkages between the sugar units would probably be 1+3, 1+3 or 4, 1+3 and 1+4. This certainly needs to be verified. D. Acidic Glycosphingolipids 1. Sulfatides a. Ceramide monohexosyl sulphate Thudichum (l) was again the first to isolate a sulfur- containing lipid from human brain and he named it sulfatide. In 1933, Blix (93) gave more detail descriptions of the sulfatide from brain by showing a composition of Sphingosine, amide-bound fatty acid and galactose esterified with sulfuric acid. Initially, the linkage of sulfate to galactose was erroneously reported to be at the C-6 position (94,95); 20 however, later studies (59,63) clearly demonstrated the sul- fate group was located at position 3 of the galactose. This was also confirmed later by chemical synthesis (96,97). The structure was shown to be SO4-Gal-(l+l)-ceramide. b. Ceramide dihexosyl sulphate A sulfatide with both galactose and glucose (lactosyl) has been isolated from human kidney by Martensson (98,99). Periodate oxidation and degradation studies indicated that the lactosylceramide was esterified with sulfuric acid at C-3 of the galactose. Methylation studies further confirmed the result and showed that the linkage of galactose to glucose was 1+4, hence the structure is SO4-Gal-(1+4)-Glc-ceramide. 2. Gangliosides Gangliosides were first isolated and described by Klenk (100) in 1942. Since these lipids appeared to be character- istic of the ganglion cells, Klenk invented the name "gangliosides". These lipids were found to contain Sphingo- sine, fatty acid, hexose, and a substance which gave a strong purple color with Bial's reagent and was called neuraminic acid. This substance was later shown to be the same compound as sialic acid isolated by Blix (101) from submandibularis mucin. Gangliosides are now defined as acidic glycosphingo- lipids that contain sialic acid. In 1956, Svennerholm (102) launched a detailed study of the brain ganglioside fractions and more than a dozen different o.- , v“ is. '1 O c... .4 21 gangliosides have been recognized since then. The isolation, differentiation and structural determination of all these compounds were the result of contributions mainly from the laboratories of Klenk, Kuhn and Svennerholm. These works have been discussed in several excellent reviews and will not be discussed in detail here. The major brain gangliosides all contain Gal-GalNAc-Gal-Glc-ceramide as the basic carbohydrate backbone. Some of the representatives are listed in Table l. Gangliosides have been detected in many extraneural tissues and fluids, such as spleen (103), liver (104,105), lung (106), kidney (107-110), intestine (111), placenta (112), lens (113,114), adrenal medulla (115), adrenal cortex (116), erythrocytes (117,118) and cerebrospinal fluid (119). In 1951, Yamakawa and Suzuki (120) obtained a glyco- sphingolipid, which they called hematoside, from equine erythrocyte stroma. This glycolipid was later confirmed by Klenk and Wolter (121) and re-examined by Klenk and Lauen- stein (8) . The equine hematoside was shown to contain N-glycolylneuraminic acid attached to the third position of the galactose molecule of the lactosylceramide. Two kinds of hematoside were demonstrated in dog erythrocytes (122) , N-acetyl (73%) and N-glycolyl (27%) , both of which were linked to the lactosylceramide. Lignoceric acid was the main fatty acid in equine hematoside, while the dog hematoside contained more stearic and nervonic acids. In 1965, the ganglioside 22 Table 1.--Gangliosides of Mammalian Brain Symbol Chemical structure GM3 NANA- (2+3) -Gal- (Bl+4)-Glc— (1+1) -Cer GD3 NANA— (2+8) -NANA- (2+3) —Gal-— (81+4) -Glc— (1+1) -Cer GMZ GalNAc- (81+4)-Ga%- (81+4)-Glc- (1+1)—Cer II I 2 NANA GD2 GalNAc- (81+4)-Gal- (Bl+4)-Glc- (1+1) -Cer I 2 NANA-(8+2)-NANA GMl Gal- (Bl-+3) —GalNAc- (81+4) +Ggl- (81+4) -Glc- (1+1) -Cer I I I 2 . NANA GDla Gal- (81+3) —GalNAc- (81+4 ) -Gz§11— (81+4 ) -Glc- (1+1) -Cer I I I 2 NANA- (8+2 ) -NANA (:le ngl- (Bl-+3) -GalNAc-— (81+4)-Ga1- (81+4) -Glc- (1+1) -Cer (I I ,2 i NANA N NA G Gal— (Bl-+3) -GalNAc- (81+4 ) -Gal- (Bl->4 ) -G1c- (1+1)-Cer T1 3 I3 I I) I) NAN N NA-(8+2)-NANA 601 G 1- (Bl-+3) -GalNAc- (81+4I-G l- (81+4) -G1c- (1+1) -Cer III III N A- (8+2) -NANA NA A- (8+2 ) -NANA 23 from cat erythrocytes was shown to have one more sialic acid than either the dog or horse (123). It was a disialolactoside with the structure of N—glycolylneuraminyl—(2+8)-N-glycolyl— neuraminyl-(2+3)-galactosy1-(1+4)-glucosyl—ceramide. LignOF ceric and nervonic acids were the major fatty acids. Bovine erythrocytes (122) showed a different variation by having Neacetylglucosamine instead of N-acetylgalactosamine in the carbohydrate chain. The prOposed structure was Gal-(1+3)- GlcNAc-(l+4)-Gal—(l+4)-Glc-(1+1)-ceramide. Hematoside was the major erythrocyte glycosphingolipid in sea lamprey and it was also shown to be present in calf serum (124). This lipid has also been detected in human erythrocytes (117). It is interesting to note that variations exist among the major erythrocyte glycosphingolipids of various mammals. The cat, dog and horse contain mainly hematosides, whereas glObosides are the major components in the red cells of man, Pig, sheep, goat and guinea pig (124). Bovine stroma have both components, whereas the chicken has neither. The pre- 0188 physiological role of such differences is not understood at Present; nevertheless, such differences are important in the Selection of a suitable animal for glycosphingolipid metabolic studies. 24 E. Biosynthesis of Neutral and Acidic Glycosphingolipids Studies of glycosphingolipid biosynthesis have gained nmch impetus in recent years; numerous reports have appeared involving in_viyg and in XEEEQ techniques which led to the elucidation of several biosynthetic pathways in the formation of glycosphingolipids. The biosynthesis of complex glyco- sphingolipids is believed to proceed through the stepwise addition of monosaccharides from sugar nucleotide intermediates (UDP-Gal, UDP-Glc, UDP-GalNAc, GDP-Fuc, etc.) to an appro- priate receptor molecule by a multitransferase system (88). It is also believed-that different enzymes exist for each step and they are different in substrate and metal requirements, pH Optimum, lability to heat and different inhibitors (88). These anabolic enzymes are bound to microsomes and have a pH Optimum around 7.0. 1- §phingosine In_yi!g, Sphingosine is formed by the condensation of gtserine and palmityl-CoA, with the formation of 3-dehydro- sPhinganine as the intermediate (125,126). This compound has been synthesized by Mendershausen and Sweeley (127) and by Gaver and Sweeley (128). The reaction requires pyridoxal phosphate. The next step is the catalytic formation of SPhinganine by a S-3-dehydrosphinganine:NADPH oxidoreductase (129~131). The final introduction of a double bond gives sphing'osine. 25 (3-dehydrosphing- E COOH anine) ~ PLP synthetase s - ' x -— — (ll-5H31 fi’ CoA + CH CHZOH ~§% I C15H31 C H CHZOH C2 OH NH ~ 2 2 (palmityl-CoA) (grserine) NADPH reductase H H . .- = —.- — O 1 — — — C15H29 CH CH CH2 H .i: C15H31 CH CH CHZOH H H2 OHNH2 (Sphingosine) (sphinganine) 2 - Ceramide An enzyme which catalyzes the biosynthesis of ceramides frcun Sphingosine and acyl-CoA derivatives was found in Chixzken liver, rat and guinea-pig brain by Sribney (132). It \was found that either threo- or erythro—sphingosine could Ibel.incorporated into ceramide. The specificity of the acyla- ti4311 process of long-chain bases by acyl-CoA esters has been Stnlxiied extensively by Morell and Radin (133), and was es- ta~131ished that ceramide containing nonhydroxy-fatty acids were tile! precursors of gangliosides and Sphingomyelin. The hydroxy- fEitrty acid containing ceramides were mainly incorporated into galactosylceramide of the cerebron and oxynervon [type (133- 136). 3 ' IAlgutral glycosphingolipids a. Galactpsylceramide (GL—lb) Two pathways have been postulated for the biosynthesis Of’ . . . . . <-‘—erebr051des on the baSIS of $.12 Vitro studies, 26 Psychosine Pathway: psychosine + fatty acyl—CoA+cerebroside Ceramide Pathway: ceramide + UDP-Gal+cerebroside Emycosine has been synthesized enzymatically from sphingosine and UDP-Gal by a number of investigators (18,137,138) and the fbrmation of cerebroside by acylation of psychosine with acyl-CoA was suggested by Brady (139) in 1962. However, other investigators have proposed that the formation of cerebroside might proceed via the acylation of sphingosine followed by the addition of hexose (136,137,140,l41). These two pathways also .have supportive evidence from in vivo experiments. Recently, (H1 the basis of mass spectrometric analyses Hammarstrom pre- sented evidence that the formation of galactosylceramide CCnataining non-hydroxy fatty acid occurred via both pathways. IHE! then found that psychosine can be acylated with acyl-CoA nohenzymatically (142) . Whether such dual pathways exist it! cerebroside containing hydroxy fatty acids still remains to be investigated . b. Glucosylceramide (GL-la) In 1968, Basu (143) demonstrated the presence of a gly- <3<>syl transferase in a particulate fraction from chicken brain which catalyzed the formation of GLI-la from UDP[14C] glucose and ceramide. This finding was recently confirmed by Haunmarstréim using rat brain microsomes. c. Lactosylceramide (GL—Za) A rat Spleen homogenate was shown by Hauser (140) to 1'ncorporate UDP[l-3H]galactose into lactosylceramide and 27 similar results were obtained by Basu (143) with a particu- late fraction from embryonic chicken brain. Kampine 33,31. (40) obtained radioactive monohexosylceramide and dihexosyl- ceramide when human leukocytes were incubated with [14C]glu- cose and [14C]ga1actose. Since human leukocytes contained mainly lactosylceramide, most of the labeling was located in the terminal galactose of the carbohydrate chain. Similar results were obtained irrespective of which sugar was being This finding seemed to favor a step— used as the precursor. When pig bone [wise synthesis of lactosylceramide via GL-la. Imxrrow cells were incubated with [U-14C]glucose, radio- éurtivity was detected in all three of the hexose units of GIr-Ba (major glyCOSphingolipid), which seemed to indicate E novo biosynthesis of the hexose moieties by the marrow cells. However, the majority of the label was located in GIv-la and GL-3a with low incorporation detected in GL-2a. This contradicted the theory of stepwise addition of carbo- hYdrate units to a receptor molecule. d. Digalactosylceramide (GL-2b) Synthesis of digalactosylceramide was studied in Gray's ‘léilboratory (144) using a kidney homogenate. The galactosyla- ‘tdican of galactosylceramide by UDP-Gal was demonstrated. e. Trihexosylceramide (GL-3a) Hildebrand and Hauser (145) demonstrated that the trans- ifear- of a galactosyl moiety from UDP-Gal to lactosylceramide ‘ 28 was catalyzed by an enzyme present in rat spleen homogenate. It was reported that two reactions, catalyzed by different enzymes, occurred in the spleen tissue glucosylceramide + UDP—Gal -* lactosylceramide + UDP lactosylceramide + UDP-Gal -* trihexosylceramide + UDP Although these two enzymes had the same optimal pH (6.0), they differed from each other in their lability to heat, degree of activation by Mg++, and the extent of inhibition by various sphingolipids. f. Globoside (CL-4) The biosynthesis of GL-4 in zit£2_has not been fully investigated. It would probably proceed via the N-acetyl- galactosylation of GL-3a by UDP-GalNAc, however. Brady (146) was able to obtain incorporation in the glucose, galactose, and N-acetylgalactosamine moieties of globoside after incu- bating beef bone marrow cells with labeled glucosamine. In a similar type of experiment, Dukes (147) was able to in- corporate glucosamine into glycolipids of bone marrow cells ifl.§i£23 the products were tentatively identified as hema— tosides, gangliosides and globosides. 9. Blood group substances The blood group-active glycosphingolipids all share the common tetrahexosylceramide backbone, Gal-GlcNAc-Gal—Glc- ceramide. It was shown by Basu et 31. (148,149) that the 29 committed step in the biosynthesis of these compounds is the N-acetylglucosylation of lactosylceramide. The transfer of galactose from UDP-Gal to receptor GlcNAc-Gal-Glc—ceramide was demonstrated by the same authors. A GalNAc or Gal—GlcNAc-Gal-Glc-ceramide B Gal/I I Fuc Fuc H Lea “"W—J Leb It is well established that blood group A and B specificities differ only by a single substituent on the terminal galactose group., The corresponding glycosyl transferases responsible for the formation of these determinants were isolated from the appropriate tissues of blood group A and B individuals respec- tively (150). Before these transferases can act, the receptor molecule has to be in the right conformation, which is the addition of a fucosyl group at C—2 of the terminal galactosyl moiety. The transferases confer A and B specificity only when this fucosyl group is present. Marcus and Cass (151) found that Lea- and Leb- active glycosphingolipids were associated with the high- and low- density lipoproteins of plasma and were transfered or inte- grated into the erythrocyte surface membranes. The site of their biosynthesis is not known. 30 4. Acidicyglycosphingolipids a. Sulfatides In 1960, Goldberg (152) reported the incorporation of [35$]sulfate into rat kidney and liver sulfatides from labeled 3'-phospho-adenosine—5'-phOSphosu1fate (PAPS). This was later confirmed by McKhann and associates (153) as well as Balasu- bramanian and Bachawat (154), who demonstrated an enzyme system which could catalyze the transfer of sulfate from PAPS to galactosylceramide acceptor in yitrg. The sulfate transferase activity was very high in the microsomal fraction of the rat kidney. The major sulfatide of human kidney, sulfo-lactosyl— ceramide, was shown by McKhann and Ho (155) to be the product when lactosylceramide was used as the substrate for the sul- fate transferase. Recently, Cumar gt 31, (156) provided addi- tional evidence that exogenous cerebrosides would accept [35$]sulfate from PAPS. In addition, Stoffyn, Stoffyn and Hauser (157) obtained radioactive sulfatide from [l4clgalac- tose-labeled phrenosine and PAPS in_zitgg, using the bio- synthetic system described by McKhann and Ho (155). They further showed that the structure of this biosynthetic sulfa- tide was sulfated at C-3 of the galactose. b. Gangliosides The biosynthesis of ganglioside is believed to Operate in the same manner as the neutral glycosphingolipids, involv— ing a series of reactions whereby specific sugars are 31 transferred to the growing glycolipid acceptor by a series of specific glycosyl transferases. The glycosyl transferases were ++ glycolipid-acceptor + nucleotide sugar Mn ’ transferase; glycolipid-product + nucleotide proposed to exist as a membrane-bound multiglycosyltransferase system (158). Different transferases catalyze the transfer of a glycosyl unit to the acceptor, each transferase is speci- fic for its acceptor, and the product of each step is the preferred substrate for the next reaction. In addition to the normal sugar nucleotides mentioned earlier, CMP-NANA is involved in the biosynthesis of gangliosides. The sialyl- transferases are a family of enzymes which catalyze the transfer of NANA or NGNA from the corresponding CMP-NANA or CMP-NGNA to the glycolipid receptor. It is believed that a number of different sialyltransferases are involved in ganglio- side biosynthesis. Transfer of sialic acid from CMP-NANA to glycosphingolipid was first demonstrated by Kanfer gt 31. (159) ifl.!i££2§ By using a rat kidney homogenate as the enzyme source, and asialoganglioside and tetrahexosylceramide as the acceptors, they were able to obtain a ganglioside fraction identified as GM The role of sialyltransferases l. in the biosynthesis of gangliosides have been studied exten- sively and systematically by Roseman and associates (158). These workers have described a number of transferases from embryonic chicken brain which catalyze the stepwise synthesis 32 of gangliosides from ceramide and the appropriate sugar donors (reactions 1—6). The biosynthetic pathway is shown in Scheme I. Experiments from other laboratories suggested that this pathway also occurred in adult frog brain (160,161) as well as young rat brain (162-165). Recent evidence (166,167) also pointed to the existence of a second pathway (reactions 7, 9 and 10) and a third possible pathway (reactions 7, 8 and 5) where the trihexosylceramide and tetrahexosylceramide back- bones of gangliosides could be assembled before the addition of sialic acid(s). The galactosyl transferases catalyzing reactions 2 and 5 have been shown to be different by Hilde- brand st 31. (163), whereas the same galactosyl transferases were involved in the catalysis of reactions 5 and 9 (168). Synthesis of complex gangliosides might also occur via reac- tion 11 (169); experimental evidences for reactions 7, 8 and 14 had been presented by Cumar et 31. (170) and Arce et_al. (165). F. Turnover Studies Relatively few reports are available about the turnover of glycosphingolipids outside the nervous system. In 1965, Kanfer (171), injected labeled GL-la into lO—day old rats and recovered most of the activity in the extracted GL-la and ceramide. Radioactively labeled sulfatides (35804) were employed by Davison and Gregson (172) and Pritchard (173) to 33 Scheme I. Biosynthesis of Gangliosides Ceramide CD Glc-cer Gal-Glc-cer @ (D Q) Gal-Glc-cer GalNAc-Gal-Glc-cer I . /I;ANA @ l© Gal-Glc-cer Gal—GalNAc-Gal-Glc-cer I GalNAc-Gal-Glc-cer NANA I NA @ NANA G M1 GalNAc-Gal-Glc-cer Gal-GalNAc—Gal-Glc-cer NANA NANA DineuraminylganglioSIdes I . .. NANA C) Polyneuraminylgangliosides Gal-GalNAc-Gal-Glc-cer Gal-GalNAc-Gal-Glc-cer Gal-GalNAc-Gal-Glc-cer NANA NANA NANA NANA I I NANA NANA L (9 Gal-GalNAc-Gal-Glc-cer - NA - a - l - Gal Gal c G'l G c cer NANA NANA NANA NANA I I NANA NANA Grl-GalNAc-Gal—Glc-cer NANA NANA NANA NANA s. ”‘0. Ap‘ vy' '- I 34 study the synthesis and turnover of myelin; most rapid syn- thesis of sulfatide was found to occur between 20 and 25 days. Dawson and Sweeley (45) studied the turnover of neutral glycosphingolipids of porcine plasma and erythrocytes in 3129 over a 2-month period using [14C1glucose as a pulse label. Their results suggested that red cell GL-2a, GL-3a and GL-4 were synthesized in the bone marrow and released into the plasma during the time of red cell catabolism, whereas GL-la was not synthesized in the bone marrow and exchanged freely between the plasma and erythrocytes. Glucose and galactose (174-176), glucosamine (177) and mannosamine (178) are the precursors commonly employed in ganglioside metabolic studies. The radioactivity is in- corporated into all portions of the carbohydrate chain of gangliosides and hexosamines are the most efficient precursor of NANA and GalNAc. Radin (174) was the first to demonstrate the incorporation of [1-14C]galactose into cerebrosides and gangliosides of rat brain. This was later confirmed by Burton 32 31. (177) who showed that all the sugar moieties of gangliosides had the same specific activity. The half-life of the total ganglioside fraction in rat brain was estimated to be about 8-10 days for a glucose or galactose label, and about 24 days for amino sugars (179). This was somewhat dif- ferent from the data of 20 days obtained by Suzuki (180) using [U-14C]glucose as the precursor. Recently, Suzuki (181) and 35 Maccioni gt 31. (182) independently used [1—14C1glucosamine and [6-3H]glucose to study the formation and turnover of rat myelin and brain gangliosides. Suzuki found that the forma- tion and turnover of GMl ganglioside in myelin was different from that of whole brain lipid or whole brain gangliosides. Neither investigator could detect any precursor-product relationship in vivo. G. Catabolic Degradation of Glycosphingolipids GlyCOSphingolipids are degraded in a stepwise removal of sugar units by a family of glycosylceramide hydrolases. These enzymes are assured to be located in lysosomes (183) of various organs throughout the body (184). Elucidation of the catabolic pathways of both neutral and acidic glyco— sphingolipid metabolism gained much success in recent years; this was partly due to the deveIOpment of enzyme assays which made possible the detection of enzyme deficiencies that occurred in various lipid storage diseases. In 1964, Sandhoff (185) obtained an enzyme preparation from pig kidney which was able to catabolyze Gal—GalNAc-Gal- Glc-ceramide to ceramide with the formation of trihexosyl- ceramide, GL-Za and GL-la. Later, Statter and Shapiro (186) demonstrated the recovery of radioactivity in GL-2a and GL-la from rat after an initial injection of labeled GL-4. An enzyme (hexosaminidase) which could hydrolyze the terminal but - IA‘ '1 - 4 A "5 36 N-acetylgalactosamine moiety of globoside was isolated from calf brain by Frohwein and Gatt (187). The same enzyme was equally active towards GalNAc-Gal-Glc-ceramide or GlcNAc- Gal-Glc-ceramide. Recently, two forms of hexosaminidase (A and B form) were identified by Sandhoff (188,189) and by Okada and O'Brien (190-191). In Tay-Sachs disease, the A form was missing while the B form remained normal or ele- vated (188,190). This correlated well with the finding that in Tay-Sachs disease the main reason for the accumulation of Tay-Sachs ganglioside was due to the absence of an enzyme to hydrolyze it. However, in a variant form of Tay-Sachs disease, characterized by the accumulation of globoside in visceral organs, both A and B forms were missing (188). It was inferred that hexosaminidase A played a physiological role in ganglioside metabolism, whereas hexosaminidase B was involved in globoside degradation (192). Once the terminal N-acetylgalactosamine was removed from globoside, the remaining trihexosylceramide could be further catabolyzed by an a-galactosidase to lactosylceramide, as shown in Scheme II. In patients with Fabry's disease, the accumulation of an excessive amount of GL-3a was due to the absence of a GL—3a-c1eaving enzyme in the plasma (193) and other organs (194) of these patients. On the contrary, enzymatic activity was detected in normal spleen, small intestine, kidney, brain and liver (195). 37 Scheme II. Biodegradation of Neutral Glycosphingolipids GalNAc-Gal-Gal-GchCer l -GalNAc Gal-Gal—Glc-Cer GlcNAc-Gal-Glc-Cer -Gal —GlcNAc Gal-Glc-Cer 1 -Gal Glc-Cer Gal-Cer -Glc -Ga1 Ceramide -FA Sphingosine - .nuo y...’ "In" _ I ".1. o .'0 he. Q-c' ‘I~ . :w .5. . ‘F u ‘5‘ PA‘ - "o. I fit- in: "v I. I M, P. If! 38 Digalactosylceramide was shown to possess the same terminal disaccharide as that of GL-3a in linkage and anomeric configuration; however, whether or not its accumulation in Fabry's disease was due to the absence of the same cleaving enzyme is not uncertain. Recent evidence by Mapes and Sweeley (196) indicates that GL-2b is hydrolyzed by a differ- ent enzyme than the d-galactosidase involved in GL-3a metabol- ism. Lactosylceramide is degraded to GL-la by a B-galactosi- dase from rat brain (197,198). The enzyme has been partially purified. An enzyme with similar activity was also demon- strated in a particulate fraction from rat kidney (185). Recently, a new glycosphingolipid storage disease, lactosyl- ceramidosis, was discovered (52). This unusual lipidosis is characterized by a deficiency of lactosylceramide: galactosyl hydrolase activity coupled with excessive accumulation of GL-2a in plasma, erythrocytes, bone marrow, urine sediment and other neural and non-neural tissues. Further degradation of GL-la to ceramide has been studied by Brady 22 21. (199) and by Gatt and Rapport (200). The GL-la-cleaving enzyme was purified 82-fold from a 100,000 g supernatant fraction of human spleen (201). The enzyme was present in normal but absent in Gaucher spleen (202). GL-la— cleaving enzyme has also been detected in white blood cells (203). Evidence seems to imply that the major source of accumulated GL-la in Gaucher's disease might be turnover of 39 glycosphingolipids from leukocytes (204) or erythrocytes (205). An enzyme that cleaves galactosylceramide to galactose and ceramide was purified by Hajra et 31. (206) from pig brain. A ceramidase which catalyzes the hydrolysis of ceramides to long-chain base and fatty acids has been purified ZOO-fold from rat brain by Gatt (207) and Yavin and Gatt (208). Information about sphingosine degradation has been gen- erated mainly by Stoffel's laboratory, based on in yiyg_(209- 211) and in yi££9_studies (212-214). Long-chain bases were phosphorylated, then cleaved between C—2 and C-3 to give phosphorylethanolamine and palmitaldehyde, hexadecenal and 2-hydroxypalmita1dehyde from Sphinganine, Sphingosine and 4-thydroxy3phinganine, respectively. ATP SPHINGOSINE R-fH-CH-CHZOH KINASE r OH NH2 SPHINGANINE or R-CH-TH-CHz-O- -OH ——€> 4-2fHYDROXYSPHINGANINE OH NH2 OH IO R-C-H + CHZ-CHz-O-P-OH \\ I I O NH2 OH The reaction is initiated by an ATP-dependent kinase (212- 215), which has been detected in human erythrocytes (215). Then a pyridoxal-dependent microsomal 1yase (214) splits the 40 phosphorylated base into the C -fragment, phosphorylethanol- 2 amine, and, in the case of Sphingosine, palmitaldehyde. The biodegradation of gangliosides proceeds via the stepwise removal of monosaccharide units of the carbohydrate backbone by specific hydrolases and sialidases. The scheme has been proposed (216), for the hydrolysis of GT to GM 1 3 and verified by Gatt and associates (217), who isolated and purified several of the glycolipid hydrolases. The degrada- tive pathway is outlined in Scheme III. The degradation of GT1 proceeds mainly via GD as shown by human brain siali- 1b dase (218) as well as those from bacteria origin (219). This finding suggests that the terminal galactose—bound sialic acid is the first to be hydrolyzed, whereas the sialic acid bound to the internal galactose remains intact. It is proposed that the internal sialic acid is more resistant because of steric hindrance caused by the N-acetylgalacto- samine bound to C-4 of the internal galactose (220). After N-acetylgalactosamine is liberated by hexosaminidase, the internally bound sialic acid is more easily removed by sialidase (221). 41 Scheme III. Biodegradation of Brain Gangliosides Gal-GalNAc-Gal-Glc-Cer (GQl) NATA NANA NANA NANA -N A -NANA AN -NANA Gal-GalNAc-Gal-Glc-Cer £—-Gal-GalNAc-Gal-Glc-Cer-——+>Gal-GalNAc-G?l~Glc-Cer NANA NANA NANA NANA (GT1) NANA (GDla) NANA NANA (Gle) -NANA -NANA Gal-GalNAc-Gal—Glc-Cer NANA (GMl) -Gal -NANA GalNAc-Gal-Glc-Cer ;-GalNAc-Gal-Glc-Cer NANA (G ) NANA (G ) I D2 M2 NANA l -GalNAc -NANA Gal-Glc-Cer 7’ Gal-Glc-Cer NANA I (G ) | (G ) NMM 'M3 D3 NANA -NANA Gal-Glc-Cer (CL-2a) 1 -Gal Glc-Cer (GL-la) l -Glc Cer I FA + Sphingosine II . INTRODUCTION Glycosphingolipids are important components of mammalian cell membranes. Extensive studies have been made of the composition and metabolism of neutral glyOOSphingolipids in human erythrocytes (37,222). However, no detailed informa- tion is available about the Sphingolipid composition and metabolism of human platelets, although several brief reports (223-225) on the existence of sphingolipids in platelets have appeared. Nothing is known about the biochemical structures of the platelet glycosphingolipids but some recent studies have been made of the phospholipid component (226) and glyco- proteins (227,228). The lack of information is due, in part, to difficulties in obtaining a sufficient amount of material for adequate biochemical studies. To complement previous studies on the glycosphingolipids of normal human plasma (37), erythrocytes (37), and leuko— cytes (38,39) as well as studies of these lipids in patients with storage diseases (47,52,229) and leukemia (39), part of this dissertation was devoted to the determination and charac- terization of the major sphingolipid constituents of human Platelets. It was also interesting to examine the lipids of Pig platelets so that a better assessment of the feasibility 42 u. ‘n ‘D 43 of studying platelet glycosphingolipid metabolism 12.XEXE could be made. As part of a general investigation of the blood glyco- sphingolipids, human plasma gangliosides as well as the major erythrocyte glycosphingolipids in fetal pigs were also studied. Another area of investigation of this thesis centered on the turnover of various blood sphingolipids in an anemic pig. Previous studies (222) with a normal human and a patient with Fabry's disease demonstrated that the early synthesis (such as liver or other organs) of plasma glycoshpingolipids could only account for 10-20% of the total plasma glycolipids when a stable isotope labeled compound, [6,6-2H2]glucose, was used as the precursor, which indicated that the remaining plasma pool had to be derived from some other source, such as the red cells. However, since the incorporation of the label into the red cell glycosphingolipids was not suffi- ciently high enough to be detected, this hypothesis was not verified until later when Dawson and Sweeley (45) conducted an £2,21X9 experiment in a pig, involving a large pulse label l4C]glucose. of [U- The pig was found to be an excellent experimental animal nmdel because it was rather similar to humans in the plasma and erythrocyte glycosphingolipid profiles. Furthermore, the size of the animal permitted withdrawal of 30-40 m1 of blood from the animal at various time intervals without affecting 44 its metabolic status in any way. Dawson and Sweeley (45) concluded from the experiment that GL-la exchanged freely between plasma and erythrocyte pools, and was not synthesized in the bone marrow, whereas the red cell GL-2a, GL-3a and GL-4 were synthesized in the marrow and were subsequently released into the plasma during the time of erythrocyte catabolism, suggesting that the erythrocyte glycosphingolipids were a major source of the plasma glycolipids. Nevertheless, several questions remained unanswered. One was that the maximum Specific activities of erythrocyte GL-2a, GL-3a and GL-4 were reached around 5-7 days after the label was given, and 40-60% of the label was lost from GL-Za and GL-3a, but not from GL-4 during the next few days. This phenomenon might be explained in several ways. One is the possibility of contamination of red cell preparations by other cell compon- ents such as leukocytes and platelets. This could result from unsuccessful washing or removal of these components dur- ing the centrifugation of the whole blood. The second and more probable explanation may be the presence of some highly metabolically active cells (such as reticulocytes) in the isolated red cell fraction. It is possible that the early loss of label was associated with membrane changes as the reticulocytes mature into normal erythrocytes. If this were true, one would expect to see the specific activity of red cell glycolipids (especially the major erythrocyte glyco- sphingolipid, globoside) decrease gradually during the entire 45 life span of the erythrocytes; this was not observed, however. This hypothesis about globoside metabolism in senescent erythrocytes may be verified by conducting a turnover experi- ment in a pig that has been previously induced with reticulo- cytosis. Another puzzling fact was the marked differences observed in the fatty acid composition of the different pig erythro- cyte glycosphingolipids. The relative amounts of various fatty acids were similar in GL-3a and GL-4, but distinctively dif- ferent from GL—la and GL-2a. For this matter, it is difficult to envisage that GL-Za and GL-3a are the precursors of GL-4. It was the intention of this study in the anemic pig to in- vestigate the turnover of both plasma and red cell ceramides, gangliosides and the neutral glycosphingolipids so that a better perspective could be obtained of the interrelationship between the various plasma and erythrocyte sphingolipid pools. In addition, the experiment was planned to include more time points throughout an eighty day period, since the time points chosen for the earlier study were shown to be inadequate. As part of this in 2122 study, an attempt was made to separate the reticulocytes from the normal cells using gradient ultracentrifugation. If such separation could be achieved, then one could study the turnover of globoside in the young cells versus the old; so that the metabolism of globoside in the red cell aging process could further be assessed. I I I . EXPERIMENTAL A. Materials 1. Non-chemicals Lipid sources Human blood Human platelets Porcine blood Equine blood In vivo study Female Yorkshire pigs Pregnant gilts Miscellaneous Dialysis tubing (size 20) Diazo projection paper (l919-A, size-8%"xll", cat. no. 30012) Lansing Regional Blood Center, American Red Cross, Lansing, Mich. American Red Cross Blood Research Center, Bethesda, Md. Meat Laboratory, Dept. of Animal Husbandry, Michigan State Univ., East Lansing, Mich. Veterinary Clinic, Michigan State Univ., East Lansing, Mich. Porcine Research Center, Michigan State Univ., East Lansing, Mich. Maternity Ward, Swine Research Center, Michigan State Univ., East Lansing, Mich. Sargent-Welch Scientific Co., Chicago, Ill. B. K. Elliot Co., Pittsburgh, Pa. 46 Instruments Gas Chromatograph Model 402 Sorvall Refrigerated Centrifuge Model RC2-B Beckman Model L3-50 Ultracentrifuge Beckman LS-150 Liquid Scintillation Counter I Mass Spectrometer, LKB 9000 Advanced Osmometer, Model 3L Gilford Spectropho- tometer 2400 2. Chemicals Solvents Chloroform, methanol, acetone, etc. 47 Hewlett-Packard Analytical Instruments, Avondale, Pa. Ivan Sorvall Inc., Newton, Mass. Spinco Division, Beckman Instru- ments, Inc., Palo Alto, Calif. Beckman Instruments, Inc., Palo Alto, Calif. Sweden LKB, Stockholm, Advanced Instrument, Inc., Needham Heights, Mass. Gilford Instrument Laboratories, Inc., Oberlin, Ohio J. T. Baker Chemical Co., Philipsburg, N: J. Mallinckrodt Chemical Works, St. Louis, Mo. Silicic acid chromatography Unisil (200-325 mesh) Clarkson Chemical Co., Williamsport, Pa. Thin-layer chromatography Precoated silica gel G plates (250 u) Precoated silica gel G plates (250 and 500 u) Glycosphingolipid standards Quantum Industries, Fairfield, N. J. Analtech, Inc., Newark, Del. Prepared from human, porcine and equine red blood cells. 48 Thin-layer chromatography (cont'd) Ceramide (hydroxy fatty Sigma Chemical Co., acids) St. Louis, Mo. Ceramide (normal fatty A gift from Dr. Karin Samuelsson; acids) also from Sigma Chemical Co., St. Louis, Mo. Ganglioside standards Supelco, Inc., Bellefonte, Pa. Acid-catalyzed methanolysis Hydrogen chloride (lec- Matheson Gas Products, ture bottle) East Rutherford, N. J. Silver carbonate Mallinckrodt Chemical Works, St. Louis, Mo. Acetic annydride J. T. Baker Chemical Co., Philipsburg, N. J. Hexane (nanograde) Mallinckrodt Chemical Works, St. Louis, Mo. Dyes Bromothymol blue Matheson Coleman and Bell, Norwood, Ohio Rhodamine 6 G Allied Chemicals, Morristown, N. J. Methyl orange Fisher Scientific Co., Fair Lawn, N. J. a-Naphthol Sigma Chemical Co., St. Louis, Mo. Resorcinol Mallinckrodt Chemical Works, St. Louis, Mo. Gas-liquid chromatography Hexamethyldisilazane I Applied Science Laboratories, and Trimethylchloro- State College, Pa. silane Anspec Co., Ann Arbor, Mich. Fatty acid methyl ester Applied Science Laboratories, standards State College, Pa. 49 Gas-liquid chromatography_(cont'd) Polysaturated fatty Supelco, Inc., Bellefonte, Pa. acid methyl esters (PUFA No. 1) Normal and hydroxy Supelco, Inc., Bellefonte, Pa. fatty acid methyl esters Mannitol Nutritional Biochemical Corp. Cleveland, Ohio 3% EGSS-X on Gas- Applied Science Laboratories, Chrom Q (100-120 mesh) State College, Pa. 15% ethylene glycol Supelco, Inc., Bellefonte, Pa. adipate on Chromosorb WHP (80-100 mesh) 16% ethylene glycol suc- Supelco, Inc., Bellefonte, Pa. cinate on Gas-Chrom P (80-100 mesh) 3% GC-grade SE-30 on Supelco, Inc., Bellefonte, Pa. Supelcoport (100-120 mesh) 3% ECNSS-M on Supelco- Supelco, Inc., Bellefonte, Pa. port HD (100-120 mesh) N-Acetylneuraminyl- Sigma Chemical Co., St. Louis, lactose Mo. QL-Sphingosine and Miles Laboratories, Inc., DIhydrosphingosine Elkhart, Ind. Assay for sphingosine (Lauter and Trams) QL-Sphingosine, di- see GLC fiYdrospingosine Methyl orange see Dyes Assay for ester groups (Rapport and Alonzo) Hydroxyamine hydro- Mallinckrodt Chemical Works, chloride St. Louis, Mo. 50 Assay for ester groups (Rapport and Alonzo) (cont'd) Ferric perchlorate (non-yellow) Tripalmitin Permethylation study Sodium hydride (57% oil dispersion) Sodium borohydride Anomerity study a-Galactosidase (fig), B-Galactosidase (jack bean), B-Hexosaminidase (jack bean) a-Hexosaminidase Sodium taurochelate (ox bile) G. Frederick Smith Chemical Co., Columbus, Ohio Supelco, Inc., Bellefonte, Pa. Alfa Inorganics, Ventron Corp., Beverly, Mass. Sigma Chemical Co., St. Louis, Mo. Gifts from Dr. Y. T. Li, Tulane Univ., New Orleans, La. Prepared from pig liver by Joseph Sung of this laboratory according to the method of Weissman and Hinrichsen (Biochemistry 8:2034, 1969). Sigma Chemical Co., St. Louis, Mo. Assay for organic_phosphorus (Bartlett) Ammonium molybdate' Sodium bisulfite l-Amino-2-naphthol-4- sulfonic acid Fisher Scientific Co., Fair Lawn, N. J. Fisher Scientific Co., Fair Lawn, N. J. Mallinckrodt Chemical Works St. Louis, Mo. 51 Gradient Ultracentrifugation Albumin, bovine (powder) (Cohn Fraction V; JACS 68:459, 1946) Cellulose nitrate tubes (size 1%”x3é"; l"x3%") Sugars 2-(+)-Galactosamine Hydrochloride D-(+)-G1ucosamine hydro- Chloride Fucose Galactose Glucose N-Acetylneuraminic acid and N-Glycolyl- neuraminic acid Radioactive isotopes [U-14C]glucose (S.A. 192 mCi/mM; in ethanol- water 9:1; lot no. 580-031) [U-14C1glucose (S.A. 180 mCi/mM; in 20% ethanol; lot no. 19- 119107) [1-14C]stearic acid (S.A. 46.08 mCi/mM; in benzene; lot no. 308-211) Sigma Chemical Co., St. Louis, Mo. Beckman Instruments Inc., Palo Alto, Calif. Calbiochem, Los Angeles, Calif. Nutritional Biochemical Corp., Cleveland, Ohio Nutritional Biochemical Corp., Cleveland, Ohio Fisher Scientific Co., Fair Lawn, N. J. Mallinckrodt Chemical Works, St. Louis, Mo. Sigma Chemical Co., St. Louis, Mo. New England Nuclear, Boston, Mass. International Chemical and Nuclear Corp., Irvine, Calif. New England Nuclear, Boston, Mass. 52 Miscellaneous l-Ethyl-3-(3'-dimethyl- Ott Chemical Co., Muskegon, aminoprOpy1)carbodi- Mich. imide hydrochloride Armidexan Bardley Products, Chicago, Ill. Heparin Nutritional Biochemical Corp., Cleveland, Ohio Experimental drug #744 Parke, Davis and Co., Ann Arbor, Mich. Aquasol New England Nuclear, Boston, Mass. B. Methods 1. Human blood preparation Human blood (non-outdated or fresh) was obtained from the Regional Blood Center of the American Red Cross in Lansing, Michigan. The separation of plasma from erythrocytes was done according to the method described by Vance and Sweeley (37). 2. Human platelet preparation Human platelet concentrates were obtained by Dr. Graham A. Jamieson in his laboratory at the American Red Cross Research Center, Bethesda, Maryland. The platelet concen- trates were prepared according to a previously established procedure (230) which involved repeated differential centrifu- gation to remove erythrocytes. The pellet, obtained after centrifuging at 10,000 rpm (18,000 g) at 0°C for 5 min, was 53 used for total lipid extraction. Preliminary studies were done on the total lipid extract from the residues of 73 platelet units after the isolation of platelet membrane glycopeptides by brief trypsin treatment (228). In addition, lipid extracts from non-treated platelets as well as those treated with chymotrypsin and thrombin were prepared for comparison. 3. Porcine platelet preparation Ten liters of blood were collected from normal healthy pigs at the Meat Laboratory of the Department of Animal Husbandry at Michigan State University, using ACD solution (NIH formula A) as anticoagulant. The pigs were killed by electric shock, then the jugular and carotid vessels were severed. Blood was collected in large round chromatography jars containing 1290 m1 of ACD solution (67.5 ml ACD/450 ml blood) with constant agitation. Preparations of platelet concentrates were made according to the method of Cohen and Derksen (231) involving differential centrifugation. After the initial centrifugation, the platelet rich suspensions were transferred to a fine-tip conical centrifuge tube and centri- fuged at 1500 g in a desk tOp clinical centrifuge in order to pack the residual erythrocytes. After each centrifugation, the contents were transferred to another tube without dis- turbing the erythrocyte pellet. This process was repeated until a faint pinkish ring was observed at the tip of the tube. 54 The contents were then mixed well and an aliquot was removed for leukocyte, erythrocyte and platelet counts. A smear was also made with the suspension and stained with Wright Stain for microscOpic examination. An aliquot was removed and fixed for electron microsc0pic examination. Finally, the contents were centrifuged once more and the White platelet preparation was freeze—dried in the centrifuge tube leaving the pinkish ring intact. The tube was placed in a 1yophili- zation jar (standing upright) and lyophilized overnight. The dry platelet fraction was removed from the tube in one piece (cone shape), and the tip with the pinkish ring was dissected with a scalpel in order to further minimize red cell contamination. For electron microscopy, and aliquot of the washed platelet concentrate was fixed with 4% glutaraldehyde in Sorensen's phosphate buffer (pH 7.4) and post-fixed with 1% osmium tetroxide in buffer for 1 hour. The specimens were dehydrated rapidly in a graded series of cold alcohols rang- ing from 70-100% and embedded in Epon. Thick and thin sec- tions were cut on the LKB Ultratome III. The thick sections were stained with toluidine blue and the thin sections were stained with uranyl acetate and lead citrate, and the speci- mens were examined on the Phillips 300 electron microscope at 500 A°. 55 4. Extraction of total lipids Extraction of total lipids from plasma and erythrocytes was performed according to a previously established pro- cedure (37). After the Folch-type partition, total lipids were recovered from the lower phase, while the gangliosides remained in the upper phase. In the case of human platelets, the trypsinized plate- lets were extracted with chloroform-methanol 2:1 (v/v) according to the method of Folch, Lees, and Sloane-Stanley (232), using a Waring blender to prepare the chloroform- methanol 2:1 homogenate. The mixture was filtered through a sintered-glass filter, and the extraction was repeated three times with chloroform-methanol 2:1, after which 0.2 volume of 0.75% NaCl was mixed with the combined filtrates (233). The biphasic mixture was allowed to stand overnight. The upper phase was then removed by aspiration, and the lower phase was washed with Folch's "pure solvents upper phase" (chloro- form-methanol-water 3:48:47 (v/v)) (234) three times. The combined lower phases were evaporated to dryness in vacuo to yield a fraction of crude total lipids. LyOphilized porcine platelets were stirred with 400 ml of methanol and 800 ml of chloroform for 30 min at room temperature. The solution was filtered and the residue was washed with 50 ml of chloroform-methanol 2:1 (v/v). The residue was further extracted at a gentle reflux temperature with 250 m1 of chloroform-methanol 2:1 (v/v) for 2 hr. 56 After filtration, the combined extracts were mixed thoroughly with 300 ml of distilled water and allowed to stand in the cold room overnight. The lower phase was washed two more times with Folch's pure solvent upper phase. 5. Isolation of neutra1_glycosphingolipids.(Silicic acid chromatography) Crude total lipids (lower phase) were fractionated into neutral lipids, glycosphingolipids and phospholipids by silicic acid chromatography as described previously (37). After detecting the presence of free ceramides together with the glycosyl ceramides in the acetone-methanol 9:1 (v/v) fraction from human platelets, a modification was made to elute the ceramides separately with chloroform-methanol 95:5 (v/v) or 98:2 (v/v) before the elution of glycosphingolipids from the Unisil column. The following elution scheme was adopted for all the platelet studies as well as the in vivo pig experiment: 1) freshly distilled pure chloroform without methanol added as preservative 2) freshly distilled chloroform-methanol 95:5 or 98:2 (V/V) 3) acetone-methanol 9:1 (v/v) 4) methanol. 6. Mild alkaline hydrolysis The crude glycosphingolipid fractions (acetone-methanol 9:1 (v/V) fraction) were treated with base according to the 57 procedure of Vance and Sweeley (37) and Dawson (52). This procedure destroys phospholipid contaminants which contain labile ester linkages. The hydrolysate is then,dialyzed for 48 hours in a large chromatography jar against several changes of distilled water in the cold to remove free gly- cerol, sugars, salts and other contaminants. The biphasic dialysate was reduced to dryness in vacuo. 7. Thin-layer chromatography Commercially available pre-coated silica gel G plates were used throughout this study. Both 250 u and 500 u plates were employed, depending upon the amount of lipids available for analysis. Plates were usually developed in a paper-lined tank saturated with chloroform-methano1-water 100:42:6 (v/v) or 70:30:5 (v/v) (single development). Samples were suspended in approximately 200 ul of chloroform-methanol 2:1 (v/v) and -applied to the plate in one single streak with a Radin-Pelick TLC streaker. This process was repeated two more times with 100 pl of chloroform—methanol 2:1 (v/v). Standard mixtures of glycosphingolipids were always spotted alongside as markers for identification. The TLC plate was air-dried at room temperature after solvent development, and the glycosphingo- lipid bands were visualized by nonspecific staining with iodine vapor. The individual bands were encircled with a fine hypodermic needle and a record of the TLC plates was made by exposure to diazo projection paper and subsequent develop- ment of the paper in a tank saturated with ammonia vapor. 58 Initial studies, using Quantum plates, indicated that GL-4 and GM3 migrated in close proximity and the separation be- tween them was unsatisfactory in the neutral solvent system. Hence, the bands were recovered from the plate in one frac- tion and the lipids were eluted from the gel as before. The mixture was then separated by TLC using another solvent sys- tem: chloroform-methanol-7% NH OH 55:40:10 (v/v). This 4 problem was not encountered when Analtech plates were used. 8. Acid-catalyzed methanolysis After complete sublimation of the iodine vapor, indi- vidual bands were scraped with a razor blade, and the neutral glyCOSphingolipids were eluted from the gel (0.2-0.4 g) with 40-50 ml of chloroform-methanol-water lOO:50:10 (v/v) (37) at room temperature. Hematoside was eluted under the same con- ditions with 40-50 ml of methanol-chloroform-water-pyridine 56:40:12:2 (v/v) (235) and ceramides were eluted with ethyl acetate. The recovered glyCOSphingolipids were subjected to acid-catalyzed methanolysis. A stock solution of 0.75 N hydrogen chloride in anhydrous methanol was prepared by bubbling dry gaseous HCl into dry methanol at room temperature. The solution could be stored at room temperature for periods up to 2 weeks. In practice, a fresh solution was usually pre— pared on the day of analysis. In order to make a reliable quantitative estimation of the glycosphingolipids present, an internal standard of mannitol was introduced. The stock solu- tion of mannitol contained 36.4 mg of mannitol in 100 ml of 59 methanol containing 1% water. A mixture of the glycosphingo- lipid sample (up to 1 mg), 100-300 ml of mannitol stock solu- tion (0.2-0.6 umoles) and 3 m1 of 0.75 N methanolic hydrogen chloride was heated for 20-24 hr at 75-80°C in a small culture tube with a Teflon-lined screw-cap. The solution was then cooled to m25°C and about 200 mg of silver carbonate was added to neutralize the hydrogen chloride. For the amino sugars, 0.2 ml of acetic anhydride was added and the mixture was allowed to stand at m25°C for 18 hours in order to convert methyl glycosides of liberated galactosamine and neuraminate to N-acetyl derivatives. The mixture was then centrifuged (1500 g) in a small clinical centrifuge for several minutes, and the supernatant was transferred to another small screw- cap culture tube. The residual silver carbonate was washed 2 more times with a small aliquot of methanol and centrifuged as before. Fatty acid methyl esters were recovered by 3-5 extractions of the combined supernate with equal volumes of hexane, and the methanol solution was evaporated to dryness under a gentle flow of nitrogen. 9. Gas-liquid chromatography Methyl glycosides were converted to O-trimethylsilyl derivatives by dissolving the residue in 35 ul of a freshly prepared 5:2:1 (v/v) mixture of dry pyridine, hexamethyldi- silazane and trimethylchlorosilane (37). After 15 minutes at m25°C an aliquot of the cloudy mixture was injected into the gas chromatographic column. 60 If no hexosamine or sialic acid were present, gas chromatography was normally carried out isothermally at 160° or 170°C on a glass column (6'xt"i.d.) packed with 3% SE-30 (or 3% OV-l) on 100-200 mesh, acid-washed, silanized di- atomaceous earth. Alternatively, the separation could be made by linear temperature programmed analysis on the same column, with an initial temperature of 160°C and a programming rate of 2°C/min to an upper temperature of 230°C. Identifica- tions of fucose, galactose, glucose, galactosamine, glucosamine, sialic acid, and inositol were made by comparison of the observed retention times relative to that of mannitol. The yield of each component was calculated from the total area produced by the various anomeric forms of given sugar, using the area produced by the known amount of mannitol for compari- son a 10. Isolation and characterization of gangliosides The Folch upper phase was dialyzed in the cold for 72 hr against 4 changes of distilled water. The dialysate was 1yOphilized and subjected to mild alkaline hydrolysis by the method described previously. After dialysis, the hydrolysate was reduced to dryness in 23222 and dissolved in a small amount of solvent for preparative TLC. Alternatively, the Folch upper phase was concentrated to dryness in_yagug and the residue was subjected to mild alkaline hydrolysis as described above. Several solvent systems were utilized in the separation Of gangliosides: 61 TLC plate Solvent system Quantum (250 u) ascending TLC in a tank saturated with chloroform-methanol-2.5 N NH OH (60:40:9) without paper lineés; the plate was developed 2 times with ade- quate drying between each develOpment. Uniplate (250 u) heat activation of plate for 30 minutes at 120°C, one-dimensional ascending TLC in a two-solvent sequential system described by Klibansky et 31. (236). Appropriate bands were detected by iodine vapor and recovered by eluting the gel with methanol-chloroform-water- pyridine 56:40:12:2 (v/v). After acid-catalyzed methanolysis (0.5 N), methyl glycosides and fatty acid methyl esters were recovered and subjected to GLC analysis. 11. Identification of fatty_acid methyl esters Methyl esters of fatty acids were recovered from the acidic methanolysate by hexane extraction. The methyl esters were then purified by preparative TLC on silica gel G with hexane-diethyl ether 85:15 (v/v) as the developing solvent (237), using methyl esters of palmitic and a-hydroxy palmitic acid as markers. Bands were made visible with bromothymol blue, and subsequently scraped from the plate within an hour to avoid extensive losses of short-chain esters by evaporation. A suspension of the silica gel in diethyl ether was packed into a small glass column and the esters were eluted with diethyl ether (40 ml per gm of silica gel) (238). Purified methyl esters were analyzed by GLC at 190°C on a glass column (6 ft by 1/8 in. i.d.) packed with 15% ethylene glycol adipate. 62 The hydroxy acids were analyzed as their trimethylsilyl methyl ester derivatives (239). The methyl esters were identified by comparing their retention times with those of standards and by co-injection of the unknown with an appropriate standard. Plots of relative retention time versus carbon number were employed for the identification of fatty acids not represented in the standard. Areas were calculated from peak heights and widths at half height and the compositions were expressed as percentages of uncorrected total area. 12. Identification of Sphingosine bases Purified glycosphingolipids were subjected to methanolysis by the method of Gaver and Sweeley (240) using 1 N aqueous methanolic HCl. After methanolysis, the reaction mixture was extracted with hexane 3 to 5 times to remove methyl esters, and the lower phase was neutralized with silver carbonate. The mixture was centrifuged and the supernatant fraction was evaporated to dryness under nitrogen. The residue was dis- solved in chloroform and applied to a column containing about one gram of Unisil in chloroform. The column was eluted with 10-15 ml of chloroform (discarded) and the long-chain bases were then recovered with 10-15 ml of methanol. The methanol eluate was evaporated to dryness and bases were N-acetylated with 50 ul of methanol-acetic anhydride 4:1 (v/v) at room temperature overnight. To facilitate the removal of excess acetic anhydride, butanol was added and the mixture co- evaporated under a stream of nitrogen. The acetylated bases 63 were then converted to O-trimethylsilyl derivatives for GLC at 230°C on 3% SE-30. Reference N-acetyl sphingosines were used as standards. 13. Identification of N-agylneuraminic acid N-Acylneuraminic acid was liberated from the acidic glyco- sphingolipid under mild conditions with 0.03 N aqueous hydro- chloric acid. After partitioning the hydrolysate with chloro- form, the aqueous phase was evaporated to dryness and the product was further purified and identified by column and thin-layer chromatography according to the method of Puro (241). Authentic N-acetyl- and N-glycolylneuraminic acids were used as standards. The N-acylneuraminic acid was also characterized by GLC and GLC-MS as the trimethylsilylated derivative at 220°C on 3% SE-30. 14. Identification of Ceramide The ceramide fraction, isolated by TLC in the neutral solvent system, was further purified on a 500 u silica gel G Uniplate (Analtech, Inc.) in chloroform-methanol-glacial acetic acid 192:5:8 (v/v) (10). In other instances, ceramides obtained from silicic acid chromatography (98:2 fraction) were further purified on a 250 u silica gel G Quantum plate with chloroform-methanol 95:5 (v/v) as the develOping solvent (9). Authentic ceramides containing normal and hydroxy fatty acids were used as standards. After exposure to iodine, the band which corresponded to the standard was eluted from the gel 64 with chloroform-methanol 2:1 (v/v) and ethyl acetate after the iodine had sublimed. The fraction was evaporated to dry- ness in_yaggg and the residue was methanolized (37). Fatty acid methyl esters were recovered by hexane extraction and the extract was divided into two equal portions. Ester groups were estimated quantitatively by the method of Rapport and Alonzo (242) on one aliquot and the distribution of fatty acids was analyzed by GLC as described above. The Sphingo- sine content was determined by a modification (243) of the method of Lauter and Trams (244) and the long-chain bases were identified by GLC and GLC-MS analyses of the N-acetyl deriva- tives as 1,3-di-O-trimethylsilyl ethers. 15. Linkage studies Neutral glchSphingolipids or gangliosides (1.0 mg) were dissolved in 0.5 ml of dry dimethyl sulfoxide and methylated according to the method of Hakomori (245), using 0.5 ml of the carbanion solution. The contents were sealed in small vials with Teflon caps and sonicated for 30 min, and the re- action was then allowed to proceed for six hours, after which methyl iodide was added (245). After washing the contents with water and chloroform, the chloroform layer was reduced to dryness under a gentle nitrogen flow. The residue was dis- solved in 1 m1 of 2 N methanolic HCl and methanolysis was carried out at 120°C for 5 hr. Fatty acid methyl esters were extracted into petroleum ether. A small fraction of the 65 permethylated methyl glycosides in the lower phase was ana- lyzed by GLC on 3% ECNSS-M at 160°C. The remainder of the permethylated methyl glycoside fraction was further hydrolyzed to free sugars with 1 m1 of 2 N aqueous HCl at 100°C for 3 hr. The hydrolysate was neutralized, the methylated sugars were reduced with sodium borohydride and alditol acetates were prepared with acetic anhydride-pyridine 1:1 (v/v) accord- ing to the method of Pepper and Jamieson (227). The alditol acetates were identified by GLC and GLC-MS (246-248) on 3% ECNSS-M at 175°C. Reference compounds of lactose, N-acetyl- neuraminyllactose, Fabry kidney trihexosylceramide and globoside (porcine red cells) were employed as standards. 16. Anomerity study Purified platelet GL-2a, GL-3a, and GL-4 were incubated with a-galactosidase, B-galactosidase, a-N-acetyl-hexosamini— dase and B-N-acetyl-hexosaminidase according to the methods described by Hakomori gt it, (64) and Laine gt gt. (78). Approximately 100-200 ug of the purified lipid (except 800 pg of GL-4 was used in the incubation with B-N-acetyl-hexo- saminidase) was dissolved in 150-300 ul of 0.05 M sodium citrate buffer (pH 4.3) containing 150-300 ug of sodium tauro- cholate in a conical centrifuge tube, and the reaction mixture was incubated at 37°C for 20 hr. The reaction mixture was reduced to dryness under a fine stream of nitrogen. One ml of chloroform was added and the contents were sonicated for 3 min 66 after which it was left at room temperature overnight. After brief centrifugation, the supernatant was removed and the residues were further extracted with chloroform two more times as before. The combined chloroform extracts were evaporated to dryness under a gentle flow of nitrogen and the residues were redissolved in small aliquots of chloroform. Approxi- mately 25% of the lipid was applied to a silica gel G plate together with the authentic glycosphingolipid standards and the purified platelet GL-2a, GL-3a and GL-4 that had not been incubated with the enzymes. The plate was developed in chloroform-methanol-water 100:42:6 (v/v) system. Spots were made visible with a-naphthol and sulfuric acid sprays. 17. Isolation and characterization of platelet Sphingomyelin A crude mixture of phOSpholipids (140 mg), eluted from the silicic acid column with methanol, was subjected to mild alkali-catalyzed methanolysis, using 1 ml of 0.6 N methanolic NaOH for each 10 mg of lipid. The solution of lipids was allowed to stand overnight at room temperature. Sphingomyelin was separated from alkali-stable lipids by TLC in chloroform- methanol-water 100:42:6 (v/v) (249). Long-chain bases were liberated from the sphingomyelin (22 mg) by acid-catalyzed methanolysis in 10 ml of the modified aqueous methanolic HCl (240). Selective N-acetylation of the free bases and con- version of the N-acetylated derivatives into 1,3-di—0-tri- methylsilyl ethers was accomplished as outlined above, except 67 that trimethylsilylation was carried out as described by Carter and Gaver (250). The derivatized bases were analyzed by GLC at 230°C on 3% SE-30 and 3% OV-l7 columns. Reference samples of N-acetylated sphingosines were employed as standards. Plots of relative retention times versus carbon numbers were used for identifications of bases not represented in the standard. The structures were further confirmed by GLC-MS of the trimethylsilyl derivatives on both polar and non-polar columns, and the aldehydes derived from periodate oxidations of the free long-chain bases (251) were analyzed by GLC and GLC-MS on a 3% EGSS-X column at 130°C. 18. Platelet Phospholipids Phospholipids, eluted from the silicic acid column (methanol fraction), were concentrated to dryness on a flash evaporator and the residue was dissolved in a known amount of methanol. An aliquot was applied to a 250 u Quantum plate and developed (ascending) with a two-dimensional technique reported by Klibansky gt gt. (236). In addition, a single dimensional system was also utilized to separate the indi- vidual phospholipids on a preparative basis. Uniplate (250 u) was activated at 110°C for 90 mins. Sample was spotted on 10-15 tracks and the plate was developed in the appropriate solvent system. Human plasma phOSpholipids were adopted as standards. Lipids were made visible with iodine vapor. After complete sublimation of the iodine vapor, each individual 68 group of spots (including the origin) VNHS eluted with methanol from the gel. The methanol eluate was evaporated to dryness under nitrogen flow and the residue was weighed. The residue was resuspended in a known amount of solvent and an aliquot was removed for phosphorus determination by the technique of Bartlett (252). The fatty acid components of the individual phospholipids were also investigated. This was accomplished by separating each group of lipids according to the single-dimensional TLC. The plate was sprayed with 0.005% aqueous solution of Rhodamine 6 G. Each group of individual phospholipid was eluted as before and methanolized with 0.75 H HCl in anhydrous methanol. Fatty acid methyl esters were extracted and identified by GLC at 170°C iso- thermally on a glass column (6 ft by 1/8 in. i.d.) packed with either 3% EGSS-X or 16% EGS. Peaks of fatty acid methyl esters were resolved by comparing their retention times with the polyunsaturated fatty acid methyl esters derived from cod liver oil and the commercially available standard PUFA No. l. Co-injection of the unknown with the standard was also used in the process of identification. In addition, fatty acids were also identified by comparing their equivalent chain length values with the established results (253). 19. Platelet neutral lipids Neutral lipids were fractionated by column chromatography on silicic acid, as described above. Column eluate was con- centrated to dryness and resuspended in chloroform. Aliquots ‘F G. 69 were spotted on multiple tracked TLC plate (Uniplate, 250 u). The plate was develOped with hexane:diethyl ether:glacial acetic acid 90:10:l.5. Individual lipids were located on the developed plate by spraying with a 0.005% solution of Rhodamine 6 G. The lipids were eluted with diethyl ether from the scraped gel. An additional plate was prepared for the identification of cholesterol and cholesterol esters. The plate was Sprayed with a mixture of conc. sulfuric acid and acetic acid 1:1 and heated at 90°C for 15 minutes (254). Both cholesterol and cholesterol esters gave red spots against a white background. Fatty acids of triglycerides and chol- esterol esters together with the free fatty acids were con- verted to methyl esters by HCl-methanolysis (2 N). The re— sulting methyl esters were recovered and checked by GLC as indicated above. 20. Preparation of radioactive glucocerebroside a. Hydrolysis of GL-l by barium hydroxide The method of hydrolysis employed here was a modifica- tion of the procedure described by Carter and Fujino (255), adapted for small samples. To 20 mg of GL-l (Gaucher Spleen) 1 m1 of dioxane was added and warmed slightly to dissolve all material. One ml of 10% Ba(OH)2 was then added drop-wise and the mixture was refluxed for 12 hr at 90°C in a pear-shaped flask fitted with a small distillation unit. The reaction mixture was transferred to a conical centrifuge tube containing 70 3.25 ml of water. The reaction flask was further rinsed with dioxane-Ba(OH)2 1:1 (V/v) and water to ensure removal of all the contents. The combined reaction mixture was allowed to stand at room temperature overnight. The contents were centrifuged and the residue was extracted twice with hot ethanol and once with hot chloroform-methanol 1:1 (v/v). The combined extracts were concentrated i2 ygggg_to give crude glucosylsphingosine which was purified by silicic acid chromatography in order to remove any unhydrolyzed Gl-l. The crude sample was applied to a column packed with 0.9 gm of Unisil in chloroform. The materials were eluted subsequently with 30 ml each of chloroform-methanol 8.5:l.5 and methanol. Both fractions were further purified by TLC using two differ- ent solvent systems, chloroform-methanol-2.5 N NH OH 60:40:9 4 and chloroform-methanol-water-conc. NH OH 48:14:l:l. Standard 4 glycosphingolipid and psychosine were used as markers. GlucosylSphingosine migrated slightly ahead of galactosyl- sphingosine (psychosine) in both systems, and it was eluted from the gel with methanol. b. Coupling of glucosylSphingosine with [1-14C] stearic acid The coupling reaction was done according to a previous report (256) on the synthesis of radioactive ceramides with a slight modification. Five mg of [1-14C]stearic acid and 5.1 mg of glucosylsphingosine were dissolved in 3 m1 of CH2C12-methanol 1:1 in a large heavy-walled conical centrifuge 1. l‘ .- H‘U 1 PI- ~\V 71 tube containing 10 mg of carbodiimide. The tube was stoppered and left standing in a closed incubator at 40°C overnight. To the centrifuge tube 25 m1 of chloroform and 25 m1 of water were added, mixed and centrifuged. The lower layer was washed twice with 1% NaHCOB, twice with 0.1 N HCl and then twice with water. After removal of water from the last wash- ing, some ethanol was added to the lower phase and the con- tents were concentrated to dryness 12.2gggg. The material was purified by silicic acid chromatography and TLC as described above. The band that corresphnded to the GL—la standard was scraped and eluted from the gel with chloroform- methanol-water 100:50:10 (v/v). The recovered material was dissolved in a small amount of solvent and an aliquot was removed for radioactivity counting. Acid-catalyzed methan- olysis was also carried out on a small aliquot of the sample in order to analyze fatty acids, Sphingosine and sugar derived from this lipid. 21. Experiment with fetal pigs Two Hampshire sows of l and 1 1/2 year old were selected from the experimental station herd. Both were bred to Hampshire boar; one was in her second gestation while the other was her first. Using the absence of further estrus as an in- dication of conception, the pregnant sows_were chosen at specific periods during gestation for Caesarean section. The Operations were performed by Dr. David J. Ellis of the College 72 of Veterinary Medicine, Michigan State University, at inter- vals of 45 and 90 days after the first breeding. The gilts were anesthetized by I. M. injection of an experimental drug #744 in a dosage of 1 mg/lb of body weight. a. 45-day fetuses The fetuses were removed from the uterus and a hemostat clamped on the umbilical cord to prevent loss of blood. The litter contained 11 fetuses in total. Each fetus was weighed and blood was removed by direct heart puncture. Difficulties were encountered in sampling the blood; attempts to remove blood from the umbilical veins and arteries were also unsuccessful. Out of the 11 fetuses, only 1 m1 of blood was collected in a heparinized tube and the majority of this sample was contaminated by the mother's blood. b. 90-day fetuses A total of 11 fetuses were also obtained from this litter. Blood samples were obtained from the anterior vena cava as described by Carle and Dewhirst (257). An average of 3 m1 of blood was obtained from each fetus. c. Analysis of blood samples Red cells were separated from the samples as described previously and followed with lipid extractions. Globoside was isolated andquantitated. In addition, a sample of the blood from the gilt was also analyzed which served as a control. 73 22. Red cell fractionation technique Separation of young and mature erythrocytes was by ultra- centrifugation over a discontinuous density gradient of isotonic albumin solution according to the method described by Winterbourn and Batt (258), which was a modification of the procedure originally described by Piomelli gt gt, (259), adapted for use on a larger scale. a. Albumin solutions Albumin was dissolved in water to give a 40% solution. This was done in the cold room with stirring at low speed to avoid foaming. The osmolality, density and albumin concentra- tion of the solution were measured, and the osmolality adjusted to 290 milliosmoles by adding solid NaCl. Osmolality was recorded on an Advanced Osmometer Model 3L in Dr. Anthony Bowdler's laboratory. Albumin concentration was measured by reading the Optical density of a series of diluted solutions of unknown concentration at 280 nm in a Gilford Model 2400 recording spectrophotometer. Densrty was measured with a hydrometer ranging 1.000-1.2000 in scales or calculated di- rectly from the albumin concentrations (259). The stock albumin solution was diluted with 0.92% saline solution to give six solutions with albumin concentration ranging between 30-40% (densities ranging between 1.075 and 1.100). b. Preparation of gradients Gradients were prepared at 4°C in cellulose nitrate centrifuge rubes. The size of tubes varied with the amount vb AIV' ynl MAI I. .4 v.- '73 74 of albumin solutions and red cell suspensions employed in the experiment. Albumin solutions of decreasing specific gravity were layered on each other with the aid of a peristaltic pump. The pump was fitted with 2 ft of Tygon tubing attached with a fine capillary tip at both ends. One end was placed in the tube containing the albumin solution, and the other end was placed inside the cellulose nitrate tube with the capil— lary tip leaning against the inside wall. By operating the pump at low to moderate speed, the albumin solutions were layered without creating any air bubbles. Interfaces between layers were clearly visible. c. Ultracentrifugation Red cell suspensions (hematocrit approximately 75%) were carefully layered over the albumin with a Pasteur pipette. Tubes were balanced using a small amount of isotonic saline. The tubes were centrifuged at 4°C for 30 min at 25,000 rpm in a Beckman Model LS-50 Ultracentrifuge with a swinging bucket, 25.1, 25.2 or 41 depending on the size of the centrifuge tubes used. After centrifugation, the different red cell bands were collected in another centrifuge tube using the peristaltic pump in a similar manner as described previously, only by switching the two ends of the Tygon tubing around so that the materials were removed instead of being delivered. The red cells were separated from albumin by mixing with an equal volume of isotonic saline and centrifuged for 10 min 75 at 4000 rpm in a Sorvall refrigerated centrifuge. After washing 3-5 times, the cells were ready for further study. 23. Separation ofgyoung and mature red cells from a normal and anemic dog To check the fractionation technique, normal and anemic canine blood was used since samples were readily available. Three tubes of gradients were prepared, as described above, in cellulose nitrate tubes (1"x 3 1/2"0 and 1.2 ml each of red cell suspensions from old normal human blood, normal dog and anemic dog (hemolytic anemia, etiology unknown; reticulo- cyte count 15-20%) were layered over the albumin solutions and ultracentrifugation was done in a SW 25.1 rotor. After washing the cells thoroughly with isotonic saline, the amount of red cells was estimated visually in a calibrated centrifuge tube. The cells were resuspended in saline and blood smears were prepared, air-dried and stained for microscOpic exami- nation. 24. Sgparation ofgyoung.and mature erythrocytes from a normal andan anemictpig. Layers of 1.9 ml each of albumin solutions of decreasing specific gravity were carefully layered on each other as be- fore in four cellulose nitrate tubes of 9/16" in diameter by 3 1/2" in length. Interfaces were clearly visible. Aliquots of 0.6 ml of red cell suspensions from normal and anemic dog, normal and anemic pig (20% reticulocytes) were layered on the top of the gradient. The tubes were centrifuged in a SW 41 76 rotor and the hemoglobin contents of the recovered cell layers were measured and expressed as the percentage of the total. 25. Checking hemolysis of the erythrocytes in albumIH solutions To Six conical centrifuge tubes containing 1 ml of 30, 32, 34, 36, 38, and 40% albumin solutions, respectively, were added 0.1 ml of packed pig erythrocytes; the contents were mixed well on a Vortex mixer, and incubated at 4°C for 30 min. Then 1 m1 of isotonic saline was added to each tube, mixed, and centrifuged at 4000 rpm for 10 min. The supernatant solu- tion was examined for the Soret band and hemoglobin content at 416 nm and 540 nm, respectively, in a Gilford spectropho- tometer. 26. Red cell size distribution This was obtained in a Celloscope Model 112 red cell counter with a setting of 75 u for the opening. Cells were diluted in phosphate buffered saline (pH 7.4). The CelloSOOpe was linked to a multi-channel analyzer and by a teletype print-out. The relative number of cells in successive channels was recorded.. The mean channel number was calculated by the following equation, Xi.Ci mean channel number = 2 C1 where i is equal to the channel number and Ci is the number of cells in a given channel i. 77 27. IE vivo pig experiment a. Induction of anemia (reticulocytosis) Generation of reticulocytosis in animals can be accom- plished with chemicals (such as phenylhydrazine) or bleeding. Since the effects of chemicals on blood glyCOSphingolipid metabolism in the pig are not known, it was decided to induce anemia in the pig by bleeding. Bleeding was performed by Dr. Elwyn Miller with daily removal of blood from the anterior vena cava according to the method of Carle and Dewhirst (257). Standard hematological techniques were employed in monitoring the degree of anemia. Microhematocrits were obtained in duplicate with capillary tubes and centrifuged for 5 min in an International Model MB centrifuge. Hemoglobin content was determined at 540 nm using the cyanomethehemoglobin method (260). Air-dried blood smears in sextuplicate were stained with new methylene blue for the supravital staining of reticulocytes. Routinely, 1000 red cells were examined on 3-4 blood smears and the average of number of cells containing reticulum was expressed as percentage of the total number of erythrocytes. Preliminary bleeding of several pigs, ranging in size between 10-13 kg, demonstrated that removal of 50 ml blood daily gave a rather slow anemic response judging by the hema- tological criteria. Hence, it was decided to bleed the pig on a 100 ml basis in the following i2 vivo experiment. 78 b. Induction of anemia in Pig l23-6 A female Yorkshire pig (123-6), weighing 10 kg, was put on an iron-deficient ration described in Table 2. Table 2.--Rations Pan-fed to Pig 123-6 Ingredients Percent Ground yellow corn (shelled) 79.00 Soybean meal, 49% C.P. 18.00 Dicalcium phOSphate 1.00 Ground limestone 1.00 Iodized salt 0.50 Vitamin B mix* 0.20 Vitamin D2 premix (9000 I.U./g) 0.02 Vitamin A palmitate (30,000 I.U./g) 0.04 Vitamin E premix (125,000 I.U./1b) 0.25 Vitamin B12 premix (60 mg/lb) 0.03 Zinc sulfate (36% Zn) 0.02 *Contained riboflavin, 2 g/lb; pantothenic acid, 4g/lb; niacin, 9 g/lb; choline chloride, 90 g/lb. The Special ration was prepared by Dr. Elwyn Miller, and the calculated analyses of the ration indicated the composition of 16% protein, 0.65% calcium, 0.54% phosphorus and 30 ppm of iron (which is one-half the standard requirement for this size pig). The pig was pan-fed with 400 gm of this ration per day (two times/day). 79 During this time daily removal of 100 ml blood had al- ready begun and anemia was expected to develop. On day 9 (and subsequently on day 15 and 24), 200-300 mg of Armidexan (elementary iron as ferric hydroxide in complex with a low molecular weight dextran form) was given to the pig I.M. in order to ensure the availability of an adequate iron pool for red cell production. At the same time, the ration was changed to the normal growth diet containing 60 ppm of iron; however, the ration was still pan-fed in order to limit the growth rate of the pig. Between day 15 and the actual day of radioisetopeAadministration, 100-150 ml of blood was taken daily in order to maintain a steady level of anemia in the pig. Achievement of a steady level (state) was judged on the basis of leveling-off of both the packed cell volume and hemoglobin content, and also the oscillation of the reticulo- cyte counts within a fairly narrow range. c. RadioisotOpe administration . Once the anemic condition was maintained at a steady level, 10 mCi of [U-14C]glucose (192 mCi/mM, in ethanol-water 9:1 solution) was given to the pig I.V. This was done by restraining the pig in a tray developed by Dr. Elwyn Miller (261). Two syringes (one filled with the isotOpe while the other was empty) were connected to a 3-way valve which was in turn attached to 10 cm rubber tubing fitted with an 18 gauge needle. This set-up prevented any possibility of needle displacement from the site of injection in case the pig moved. 80 All parts in contact with blood were pre-rinsed with‘heparin- ized saline. Once the needle was in the vein, approximately 5-10 m1 of blood was drawn into the empty syringe; the valve was turned to the second syringe containing the isotOpe, which was subsequently administered. Immediately following that, the valve was turned back to the first syringe whereby the previously drawn blood was re-injected back into the pig to flush out any remaining isotope trapped in the assembly line. The pig was individually confined. d. Treatment of blood samples Blood samples were withdrawn from the pig according to the following schedule. Approximately 100-130 ml of blood was removed daily between Day 0 and Day 10, while 40 m1 of 6 hr Day 7 Day 35 Day 63 12 hr 8 41 66 Day 1 9 46 70 2 10 50 75 3 14 53 81 4 20 55 5 25 57 6 30 60 blood was removed on the other days. The weight of the pig was maintained constant for the first ten days by restricting the food intake as before. After the initial 10 day period, the pig was allowed to grow normally by feeding from a self- feeder g9, 112- Erythrocytes were separated from plasma by centrifuga- tion and washed extensively with saline as before. Lipids were extracted from all of the plasma and erythrocyte samples. 81 In addition, on days 2, 3, 4, 6, 7, 8, 10, 14, 20, 30, 41, 50, 60, 70 and 81, aliquots of the packed red cells were sub- jected to gradient fractionation by the ultracentrifugation technique. Extraction of total lipids and isolation of glycosPhingolipids were carried out according to the pro- cedure described above in the Methods, and presented schematically below. Scheme IV. Analysis of Blood Glycosphingolipids. Plasma or RBC extraction lower phase +— Folch {ash ———)- upper phase silicic acid column~a . . reduced to dryness neutral‘VI' phosphalipids lipids ceramides glycosphingolipids mild alkaline hydrolysis (MAH) acid- counting MAH dialysis catalyzed { ‘methanolysis dia ysis reduced to dryness FAME base TLC C:M (1:1) cOLnt f// \\\ I ‘ assay counting ’individual bands aliquot removed for ._ counting aliquot for acid-catalyzed methanolysisf counting ‘ g . bjfe * FAfE methyl glycosides counting counting cOfi;:ing GLC The only apparent difference in the procedure cited above was the counting of radioactivity in the glchlipid and/or its components derived from these molecules. Radioactivity was monitored in two ways--the purified Sphingolipid was 82 dissolved in an appropriate solvent and a known amount of the lipid was transferred to a scintillation vial. The contents were concentrated to dryness under a stream of nitrogen and 10 ml of Aquasol was added for counting. Specific activity was expressed as counts per minute per m1 of plasma or erythrocyte. Alternatively, in the case of neutral glyco- sphingolipids, individual components of fatty acid hexose and sphingosine were separated according to their different solu- bility properties after acid-catalyzed methanolysis (45). The hexose moieties were divided into two equal halves and one was used for the quantitation of methyl glycosides by GLC, while the other was used for radioactivity determination. Specific activity was expressed as counts per minute per umole. Counting was done in a Beckman Model LS-150 liquid scintalla— tion counter. Plasma and erythrocyte ceramides were purified by TLC and recovered by solvent extraction. An aliquot of the lipid was counted in Aquasol. In addition, erythrocyte ceramides containing normal and hydroxy fatty acids were further sub- jected to acid-catalyzed methanolysis after which the fatty acid methyl esters were recovered by hexane extraction. The hexane extracts were reduced to dryness in a scintillation vial and 10 m1 of scintillation fluid was added. Long-chain bases were obtained by reducing methanol fractions to dryness under a flow of nitrogen and redissolving the residues in a known amount of ethyl acetate. Aliquots were removed in 83 duplicate for sphingosine assay and the remainder was counted for radioactivity. Upper phases from both the plasma and erythrocyte samples were concentrated to dryness 12 ygggg. The residues were sub- jected to mild alkaline hydrolysis at room temperature for 1 hr, after which the hydrolysate was extensively dialyzed. The dialysate was again concentrated to dryness ig.ygggg and the residues were dissolved in a small amount of chloroform- methanol 1:1 and an aliquot was transferred toIa scintillation vial for radioactivity determination. Since specific activi- ties were not similarly expressed in all of the sphingolipids studied, the following table summarizes how results were presented. Specific Activity (cpm/ Specific Activity (cpm/ml umole) RBC or plasma) plasma Gl-la, GL-2a, GL- erythrocyte GL-la, GL—2a, 3a, GL-4 and GM3 GL-3a, GL-4 and GM3 erythrocyte ceramides erythrocyte gangliosides erthrocyte GL—4 (8 time erythrocyte ceramides points from first 10 days) plasma ceramides plasma gangliosides e. Red cell fractionation Cell fractionations were done on red cell samples ob- tained at 2, 3, 4, 6, 7, 8, 10, 14, 20, 30, 41, 50, 60, 70 and 81 days. Layers of 9 ml each of albumin solutions of decreaSP ing Specific gravity were carefully layered on each other in 84 three cellulose nitrate tubes of 1 1/2 inches in diameter by 3 1/2 inches in length. Aliquots of 2.5 ml of red cell suspensions (hematocrit approx. 75%) were layered on top of the albumin solution. The tubes were centrifuged at 4°C for 30 min at 25,000 rpm with a SW 25.2 swinging bucket. At several time points (beginning and end) throughout the 80 day period photographs were taken of the gradient tubes after centrifugation so that a permanent record could be kept. Individual red cell bands were recovered as before, pooled, and the hemoglobin contents were measured. The cells were then washed with isotonic saline extensively to remove all traces of albumin. The volume of the packed cells was recorded and the lipids were extracted from the cells as be- fore. Glycosphingolipids were isolated by silicic acid chromatography and further purified by TLC. Globoside was eluted from the gel and subsequently degraded by acid- catalyzed methanolysis. Fatty acid methyl esters were removed by hexane extraction and the hexane layers reduced to dryness in a scintillation vial and counted. The methyl glycosides and sphingosine bases were partitioned with Chloroform- methanol 2:1 and water. Half of the methyl glycoside fraction was quantitated by GLC while the other half and the sphingo- sine fraction were concentrated to dryness in scintillation vials and radioactivity determined after adding 10 ml of Aquasol. Specific activity was expressed as counts per minute per umole. 85 27. In_yitrg study During the generation of reticulocytosis in pig 123-6 (before administration of the radioisotope), 100 ml of blood (reticulocyte count 40%) was removed from the pig and incu- bated with 30 uCi of [U-14C]glucose in an Erlenmeyer flask at 37°C (Dubnoff Shaker) for 2 hr. The flask was saturated with 100% O2 and stOppered. The pH of the blood was 7.25. After incubation, the plasma was removed by centrifugation at 600 g and the cells were washed three times with 0.92% of isotonic saline. Three tubes containing 9 ml each of the albumin solutions were prepared as before and 6 m1 of the red cell suspension (hematocrit approx. 75%) was layered on albumin in each tube. The tubes were centrifuged at 4°C for 30 min at 25,000 rpm in a SW 25.2 rotor. After ultracentrifugation the tubes seemed overloaded with cells; hence, the top layer which contained most of the cells was removed from each tube and pooled. The cells were washed extensively with isotonic saline and glycosphingolipids were isolated. The lipids were subjected to acid-catalyzed methanolysis and radioactivity of the indi- vidual components was determined. The methyl glycosides derived from each glycosphingolipid were also quantitated by GLC. Specific activity was expressed as counts per minute per umole. IV . RESULTS A. Isolation, Purification and Characterization of Human Plasma GM3 Ganglioside 1. Discovery of Human Plasma GM3 Gapglioside The presence in plasma of a ganglioside was discovered as a result of changing from one brand of TLC plates to another. In this laboratory commercially available pre- coated silica gel G plates have been employed routinely in the isolation of glycosphingolipids. In the past, Brinkman plates were always employed and the separation of various glycosphingolipids was very satisfactory in the neutral solvent system; globoside usually gave a single band. However, delivery of these plates from the manufacturer was slow and resulted in work delays. Hence, I tested another brand of pre-coated plates (Quantum Industries) that had just appeared on the market. The separation and yield of glyco- sphingolipids from these plates were equally satisfactory as the Brinkman plates. Furthermore, scraping gel from these plates was more easily accomplished because they contained less binder. When using Quantum plates in the purification of human plasma glyCOSphingolipids, it was observed that the GL-4 86 87 fraction was resolved into two closely migrating bands, at times apparently depending on the daily fluctuation of room temperature and humidity. It was originally believed that the second band was a fraction of GL-4 with a different fatty acid composition, but further investigation ruled out this possibility. Gas-liquid chromatography of the methyl glycosides derived from the material after acid-catalyzed methanolysis gave a peak which corresponded in retention time to the methyl ester methyl ketal of trimethylsilylated sialic acid on a programmed analysis. The presence of a sialic acid-containing glycosphingolipid was therefore suspected. 2. Search for A Solvent System Initially, an attempt was made to develop the TLC plates in the neutral solvent system (ascending) two times in the same direction. This was a useful modification since the two bands separated a little better than with one develOpment but was not sufficient for complete purification. A search was then made for a different solvent system to resolve the mixture. Efforts were mainly concentrated on published gang- lioside solvent systems using various combinations or prOpor- tions of chloroform, methanol, water and ammonia. Standard GL-4 (isolated from erythrocytes by Mr. Paul Snyder of this laboratory) and gangliosides (isolated from horse erythrocytes by the author) were employed as markers. It was discovered 88 that chloroform-methanOl-7% NH4OH (55:40:10, v/v/v) gave the best separation between the two lipids. 3. Separation of Sialic Acid-Containing Glycosphingolipid and Globoside from Human Plasma Glycosphingolipids derived from 50 ml of human plasma were first purified by TLC (Quantum plate) in the chloroform- methanol-water 100:42:6 (v/v) system. The two bands (in the region of the Standard GL-4) were removed from the plate in one fraction and eluted from the gel with chloroform-methanol- water mixture. The lipids were then separated by a second TLC (Quantum plate) using chloroform-methanol-7% NH4OH 55:40:10 (v/v) as solvent. As shown in Figure 1, the unknown lipid (Band B, RF = 0.51-0.54) and globoside (Band C, RF = 0.38-0.42) were well separated from each other and from another compound (Band A) that has not been identified. 4. Identification of Lipid B by Gas-liguid Chromatogtgphy Gas-liquid chromatography enables one to obtain detailed information about the nature and relative amounts of the sugar constituents in a given glycosphingolipid. An identification of fucose, galactose, glucose, galactosamine, glucosamine and sialic acid can be made by comparison of the observed retention times relative to that of mannitol with those of standards as shown in Figure 2 and Table 3. Figure l. 89 Thin-layer chromatography of hematoside from human plasma. Thin layer chromatographic separation of an unknown lipid (Band A), hematoside (Band B) and globoside (Band C) from human plasma. The plate was develOped in chloroform-methanol- 7% NH4OH (55:40:10, v/v/v). 90 Figure l 91 . 3.: 30m OflcflEmnsmcamumOMIz can “Awa.ma.aav mewammoosHmawpmomIz “ANH.OHV mcfiemmouomammamumomlz “Amy Houflccme «Am.mv mmoosam “Am.m.vv mmouomHmm “Am.m.av mmoosm "mummsm mcfl3OHHOm may Op Ocomm Iwuuoo mxmmm .Asmmouuflsv GHE\HE ca mo mums 30am mam MOHHHMO a runs cae\oom em ooomm on .ema seem emssmnmone .omumm wm co metamoomam stuma mo mm>fipm>flmmp HmaflmamnumEfiHpIo mo mammamgm .mmpflmoomam stumfi Hmaamamsuwfifluu mo hammumoumEOHSO pflsqflalmmw .N musmflm 92 Om mm ON N musmdm mmSEE m. GSUOdSGJ .IoIoaIap 93 Table 3. Relative Retention Behavior of Trimethylsilyl Methyl Glycosides Compound Retention Timea Correction Factorb Fucose 0.17, 0.19, 0.21 1.56 Galactose 0.47, 0.53, 0.62 1.25 Glucose 0.69, 0.77 1.25 Mannitol 1.00 1.00 N-Acetylgalactosamine 1.15, 1.43 1.36 N-Acetylglucosamine 1.34, 1.66, 1.87 1.36 Methyl N-acetylneuraminate not determined 0.98 aThe retention times are relative to the time for hexa-O-tri- methylsilylmannitol, which was 15:1 min on 3% SE-30 at 160°C (isothermal) with 48 ml per min nitrogen flow rate. bThese factors are used to correct observed areas in a gas- liquid chromatOgraph with flame ionization detector; they are based on molecular weight ratios. The lipid recovered from Band B was subjected to acid- catalyzed methanolysis and the mixture was reacted with acetic anhydride to convert the neuraminate to an N-acetyl deriva- tive. Methyl glycosides were identified by gas-liquid chroma- tography as O-trimethylsilyl derivatives on a 6 ft. U-shaped glass column packed with 3% SE-30. After injection of the sample the column temperature was increased linearly from 160°C to 230°C at a rate of 2°/min. Peaks of galactose, glu- cose and neuraminic acid derivatives were identified by their 94 retention behavior and comparison of their mass spectra with those of authentic compounds. A typical gas-liquid chroma- togram is shown in Figure 3. Based on the areas of peaks on the gas-liquid chromatograms, the calculated average molar ratios of galactose and neuraminate to glucose were 0.99 and 0.63 for the plasma ganglioside, as shown in Table 4. 5. Identification of Sialic Acid Derived from Lipid B The lipid was reacted with aqueous acid under mild con- ditions to liberate intact N-acetylneuraminic acid (115). After extraction with chloroform the aqueous phase was evap- orated to dryness and the product was purified by column chromatography and TLC according to the method of Puro (241). Thin-layer chromatography (Figure 4) showed that the RF value (0.41) of the sialic acid from the lipid was identical with that of authentic N-acetylneuraminic acid, whereas N-glycolyl- neuraminic acid, liberated from hematoside of equine erythro- cytes, had an RF value of 0.28. The lipid recovered after mild acid hydrolysis was sub- jected to acid-catalyzed methanolysis and the trimethylsilyl methyl glycosides were analyzed by GLC. The only components found were galactose and glucose in a molar ratio of 1:1 as shown in Figure 3. It was therefore concluded that human plasma lipid B was an N-acetylhematoside (GM3 ganglioside). 95 .mflmmaoupwc pfiom OHHE umumm Opamoflamcmm mEmmHm .mcaomuu Eonuom .Amv Odom OflsflfimusmcH>umomIz Ocm “Am.vv wmoosHm “Am.m.av wmouomHmm "mHMOSm mcflsoaaom ecu ou Ocommmnuoo mxmwd .mammaoupmn paom OHAE OHOMOQ wpflmoflamcmm mEmMHm .mcflomup doe .cHE\om mo mums m um Uoomm Op uoowa Eoum maummcfia Ommwmuocfl mus musumummEmu SESHOO why madfimm mnu mo cowpomflnfl ca umum< .cmmumoumeounu mew mew pumxomqupmHBOm m Ga omlmm mm nuHB Omxomm CESHOO mmmHm OmmmnmID .um m n so mm>fium>fluwp H>Hflmawcumfiauulo mm pmfimwucopfl mums mwOHmOO>Hm Hague: .mEmmHm cmES: HmEuo: mo OOHmOHHmcmm Eoum mopflm00>am Hmnumfi Hwaflmamnumfifiuu Mo mOEflu coflucmumm .m Ouflvflh 96 mm m musmHm 65.32:): 52: 202.29er ON d v mp OF — J J) m N EISNOdSEItI 80133130 97 .mmmmnm umBoa can some: cfl mpflmoflamcm 0 m2 .anm OHCHEmusmsawumomIz a U prm>oomn Hmuou now one mOSHm> when e.~ om.o mm.H m.m me.e mm.o cams a.m ma.e mm.H m.m No.0 em.o m e.m om.o mm.H m.m mm.e mm.o N m.m Ha.o as.a H.m mo.o He.H H Ha\mmaose UHO\oam mpsuo MOM ma Gama» one . 0 pam mUHEmumo mopsaoch we m mm com: mo owa m mm mm vs mum m cm vvm m Hm «N mm mum A m8 m5 0E me has NO m usmfimz has mo w unmflmz has mo w unmflmz fiance madamm mpflmflaonmwonm ompflmfiammcflnmmoowaw mpflmmq Hmuudmz mumamumHm cmfism mo coflpflmomaoo pflmflq Hopoe .m manna 114 fraction, whereas the shpingolipids and phOSpholipids accounted for 8% and 65%, respectively. 3. Identification of Platelet Glycosphingo- lipids by Thin-layer Chromatography Commercially available pre-coated plates of silica gel G were used for the separation of glycosyl ceramides. Previous experience with Quantum plates had demonstrated that the separation of globoside and hematoside could be achieved by means of two solvent systems. The use of two plates for this procedure is not economical when there are many samples to be analyzed, such as in a turnover study. Hence, an attempt was made to resolve these components with TLC in a single develop- ment. This was achieved with Uniplates from Analtech. An example of the TLC of glycosPhingolipids from 50 ml of plasma is shown in Figure 8. Plasma globoside (Band B) and hemato- side (Band F) were separated very well. In addition, ceramide (Band A) could also be identified, but with lipids derived from erythrocytes and platelets further purification of ceramides was required. Uniplates also gave good separation of platelet globoside (RF = 0.16) and hematoside (RF = 0.07), as shown in Figure 9. ‘After development in chloroform-methanol-water 70:30:5 (v/v), iEour neutral glycosphingolipids (Figure 9--I, II, III, and IV) ‘Here located in the same areas as glucosylceramide, lactosyl- ceramide, Fabry trihexosylceramide and porcine globoside, Figure 8. 115 Thin-layer chromatography of sphingolipids from normal human plasma. Thin-layer chromatography (on 500u pre-coated silica gel G plates from Analtech, Inc.) of sphingolipids from normal human plasma. In lane 1 reference standard ceramide. In lane 2, reference standards from top to bottom are glucosylceramide, lactosylceramide, Fabry tri- hexosylceramide and porcine globoside. In lane 3, ceramide (A), major neutral glycosphingolipids (B-E) and hematoside (E) are separated from crude total glycosphingolipid fraction of plasma. Unlabeled zone at top contains primarily methyl esters that were released from contaminating phospholipid by mild alkali-catalyzed methanolysis. The solvent system used was chloroform-methanol- water (70:30:5, v/v/v). 116 l. ... nos .1... . . .7, ‘0‘ Figure 8 Figure 9. 117 Thin-layer chromatography of glycosphingolipid fraction from washed normal human platelets treated with trypsin. Thin-layer chromatography (500p pre-coated silica gel G plates from Analtech, Inc.) of glyCOSphingo- lipid fraction from washed normal human platelets treated with trypsin. In lane 1, reference standards from top to bottom are glucosylceramide, lactosylceramide, Fabry trihexosylceramide and porcine globoside. In lane 2, major neutral glycosphingolipids (I—IV) and acidic glycosphingo- lipid (V) are separated from crude total glyco- sphingolipid fraction of platelets. Unlabeled zone at top contains primarily ceramides along with methyl esters that were released from con- taminating phOSpholipid by mild alkali-catalyzed methanolysis. The solvent system used was chloroform-methanol-water (70:30:5, v/v/v). 118 Figure 9 119 respectively, and an additional band (V) which migrated between the origin and globoside in this solvent system was shown to contain sialic acid by resorcinol spray. 4. Identification of Platelet Glyco— sphingolipids by GLC Representative gas-liquid chromatograms of the O-tri- methylsilylated methyl glycosides of the neutral glycosphingo- lipids are shown in Figure 10. The calculated average molar ratios of galactose to glucose were 1.1, 2.2, and 2.1 for lipids II, III, and IV, reSpectively (Table 9). Glucose was the only sugar component in lipid I. N-Acetylgalactosamine was also present in IV, and the molar ratio of N-acetyl- galactosamine to glucose was 0.8. The sialic acid-containing lipid (V) had a galactose to glucose ratio of 1.2 and a molar ratio of N-acylneuraminic acid to glucose of 0.54, as shown in Figure 11. These results agree with the assignment of lipids I-V as monohexoside, dihexoside, trihexoside, tetra- hexoside and sialo-lactoside. 5. Identification of Sialic Acid from Platelet Lipid V After splitting the neuraminic acid by mild acid hydroly- sis, thin-layer chromatography revealed the presence of N-acetylneuraminic acid (RF = 0.40) when compared with the authentic N-acetylneuraminic acid standard. This was further confirmed by GLC of the liberated neuraminate as the trimethyl- silylated derivative. Figure 10. Gas-liquid chromatography of trimethylsilyl methyl glycosides from major neutral glyco- sphingolipids of normal human platelets treated with trypsin. Analysis of O-trimethylsilyl derivatives of methyl glycosides with a Hewlett—Packard 402 gas chromatograph equipped with a 6 foot 3% SE-30 column maintained at 165°C. Peaks correspond to the following sugars: galactose (1,2,3); glucose (4,5); mannitol (6); and N-acetylgalac- tosamine (7,8). From tOp to bottom: GL-4 (IV), GalNAc-(l+3)-Gal-(l+4)-Gal-(1+4)-Glc-(l+l)- ceramide containing Gal:Glc:GalNAc (2.1:1.0:0.8); GL-3 (III), Gal-(l+4)-Gal-(l+4)-Glc-(1+1)- ceramide containing Gal:Glc (2.2:l.0); GL-2 (II), Gal-(1+4)-Glc-(1+1)-ceramide containing Gal:Glc (l.l:l.0); and GL-l (I), Glc-(1+1)- ceramide containing only glucose. DETECTOR RESPONSE 121 I I * MIL Wk 1 UL \l/U LI 1 5 IO I5 RETENTION TIME (MINUTES) Figure 10 122 .Uqw an pmcflfiuwump mm mummsm mo moaumu HmHozn .cwmmmuu nuflz pmummuulmum yo: mumz noHSS mumammem Goes: cosmos Eouw uomuuxm UHQHH mcu ma m meEmm .mmspflmmu umHmumam pmummHuICwmmmnu Eonm uomnwxm UHQHH may ucmmmummu away «Auummm mauQOE o3u usonmv mmEHu ucmummmwp um pmnwamsm mums can munmfim3 ucmHmMMHp mum N can H mmamammm X mm.o mm.o o.a mv.a om.o m.H on.a m.a mm.v o.H mm.o m mm.H vm.o N.H vm.H mm.o H.N mm.a ~.N mm.¢ H.H mm.o com: ov.a mm.o m.H Hm.a om.o H.~ mm.H m.m mo.m H.H Hm.o m mm.a mm.o H.H bH.H mm.o H.m Hm.H H.m Hm.v H.H N~.o H as m cam use as m .oaw cam as m oao .qa 6 mac .ne m mmH061 ¢zHum>fiump HmaflmahnquHHulo mo Emumoumeouao Uflsvflalmmw .cflmmmuu nufi3 pmummup mumawumHm amass HmEHoc mo mpflmoflamcmm mzw Eoum mmpflmoomam Hmnumfi Hwaflmamsumafluu mo Smumoumeouno pwsqflalmmo .HH musmflm 124 mu Ha wusmfim 8332.2. us: 26:55.... 8 K\o_ _ \\ _ ' ESNOdSBH 80103130 125 6. Linkage Studies of Platelet Glyco- sphingolipids a. Lipid I (ceramide monohexoside) It is well established that the hexose is linked to ceramide through a glycosidic linkage; limited amounts of material prevented permethylation studies. Analysis of the trimethylsilyl derivatives showed that only methyl glucoside was present in this lipid. b. Lipid II (ceramide dihexoside) The methylated products of the dihexoside contained a and B methyl-2,3,4,6-tetra-O-methyl-galactoside and a and B methyl-2,3,6-tri-O-methyl-glucoside (Figure 12 and Table 10). Gas-liquid chromatography of the partially methylated alditol acetates revealed two peaks for lipid II with the same retention times (Figure 13 and Table 11) as 1,5-di-O- acetyl-2,3,4,6-tetra-O-methy1-galactitol and 1,4,5-tri-O- acetyl-2,3,6—tri-O-methy1-glucitol. The structures of the methylated monosaccharides were further confirmed by GLC-MS; their fragmentation patterns were identical to those reported for l,5-di-O-acetyl-2,3,4,6-tetra-O-methyl-hexitol and 1,4,5- tri-O-acetyl-Z,3,6-tri-O-methyl-hexitol (241,247,248) and agreed with those obtained with reference standards (Table 12). The results were consistent with the structure of the ceramide dihexoside from human platelets being Gal-(l+4)-Glc-(l+l)- ceramide. Figure 12. 126 Gas-liquid chromatography of permethylated methyl glycosides derived from lactose, Fabry trihexosylceramide (CTH), porcine erythrocyte globoside (GL-4), neuraminyllactose, and platelet glycosphingolipids. Peak 1 corresponds to methyl-2,3,4,6-tetra-O- methyl-a and B-galactoside; peak 2 to methyl- 2,3,6-tri-O-methyl-B-galactoside; peak 3 to methyl-2,3,6-tri-O-methyl-B-glucoside; peak 4 to methyl-2,4,6-tri-O-methy1-B-galactoside; and peak 5 to a anomers of 2,3, and 4. Peaks 2 and 3 are partially resolved, not quantitative with respect to peak areas. Analysis was made on a 6 ft. glass column packed with 3% ECNSS-M on 100-120 mesh Supelcoport maintained isothermally at 160°C using helium as carrier gas. Peak identifications were made by comparing with pub- lished results (262). RESPONSE DETECTOR LACTOSE 903cm: cm 23 ”HO IV THME———” Figure 12 128 .mmmum xmmm ou uowmmmu suflz m>aumufiucmsq uo: .qu wn pm>a0mmu waamfiuummn .v paw m .N uo mumEocm a .m umpflmouomammlmIamnumEIOIHHu-m.«.muamSme .v “mpflwoosHmlmlamsumE -o-Huu-o.m.m assume .m imoflmouomamm-m-H>sums-o-Huu-m.m.N-Hmnpms .N “mcflmonomamm-m cam a IawsquIOImuumulm.v.m.mlamnuwe .H mm pmuocmo mum mmpflmoomam Hmnume pmumamnmeHmm mo mxmmmm s.n m.o q.m -- -- > cfiaflq w.b m.m v.m IIII III- mmouomaHmcHEmusmzamumom-z v.n m.o v.m -- -- >H mamas v.n m.m m.m N.m lull vlqw mpmoouzumum mcwouom m.~ -- v.m m.m m.m HHH oflmflq v.a -- m.m m.m m.m m-qo smccflx munmm ... ---- ..m ---- m.~ HH uuaflq m.n IIII m.m IIII m.m mmouomq m v am am H loomao umawumam Eoum mmpflmoohaw mmoflmflaomcflnam ashamz omumHmsumsumm mo mmsae coflucmumm .oa magma Figure 13. 129 Gas-liquid chromatography of partially methylated alditol acetates derived from lactose, neuraminyl- lactose, Fabry trihexosylceramide (CTH), porcine erythrocyte globoside (GL-4), and platelet glyco- sphingolipids. Peak (a) corresponds to l,5-di—O-acetyl-2,3,4,6- tetra-O-methylgalactitol; peak (b) to 1,3,5-tri-O- acetyl-2,4,6-tri-O-methyl galactitol; and peaks (c) and (d) to 1,4,5-tri-O-acety1-2,3,6-tri—O- methyl galactitol and -glucitol, respectively. Analysis was made on a 6 ft. glass column of 3% ECNSS-M on 100-120 mesh Supelcoport maintained iso- thermally at 175°C using helium as carrier gas. RESPONSE DETECTOR (a) (d) J LAC TOSE Mater UPID u AEURAMINYLLACTOSE (b) (d) Wt umo v M FABRY CTH PLATELET lIPID III (b) (c) Id) PORCINE 61-4 PLATEIET LIPID IV TIME—fl Figure 13 131 Table 11. Retention Times of Partially Methylated Alditol Acetates from Platelet Glycosphingolipids GLC Peaks (a) (b) (C) (d) Lactose 15.9 --- --- 32.9 Lipid II 16.3 --- --- 32.7 Fabry kidney trihexosylceramide 16.3 --- 31.7 33.2 Lipid III 16.3 --- 31.7 32.9 Porcine globoside --- 29.9 31.5 33.0 Lipid IV --- 30.1 31.8 33.8 N-acetylneuraminyllactose --- 30.1 --- 33.3 Lipid V --- 29.7 --- 32.9 aPeaks of partially methylated alditol acetates are repre- sented as (a), 1,5-di-(3-acetyl-2,3,4,6-tetra-O-methyl- galactitol; (b), l,3,5-tri-O-acetyl-2,4,6-tri-O-methyl- galactitol; (c), 1,4,Svtri-O-acetyl-Z,3,6-tri-O-methyl- galactitol; and (d), 1,4,5-tri-O-acetyl-2,3,6-tri-O-methyl- glucitol. 132 HouHosHO-HmnuOs-o-Huu-O.O.O-HOumom-o-HuO-O.O.H .HOO-xmmm HouHuoOHOO-HOOuOs-o-HuO-O.O.O-Hmumom-o-Huu-O.O.H .xov xmmm HouHuomHOO-HmsuOs-o-Hun-O.O.O-Hmumom-o-Huu-O.O.H .an xmmm HouHuomHOO-HOOOOE-o-Ouumu-O.O.O.O-Haumom-o-HO-O.H .HOO xmmm OH O OH O OH O O OH OH O OH O - OH O - HH - HH - OOO - - - - - - - - - - - - O - - O - O - OH OOO O OH O OH O O OH O O OH O O OH O O OH O OH O OH HOH - - - - - - - - - - - - OH - - OH - OH - Om OOH O HO O OO O O OO _O O OO O O OO O O OO O OO O OO OOH - O - O - - O - - O - - - - - - - - - - OOH NO NO HO OO OO OO OO HO OO HO OO Om Om Om OO Om HO Om HO Om OHH OH - OO - Om HO - ON ON - OH OO O OH OO O HO O OO O OHH OH OH OH OH OH OH OH OH OH OH OH .OH NO OH OH Om OH OO OH NO HOH OH O OH O OH O O OH OH O OH OH - HH OH - O - OH - OO OH O OH O OH OH O OH OH OH OH OH OH OH OH OH OH OH OH OH OO O - O - O O O O O O O O O O O O O O O O OO O O O O O O O O O O O O O O O O O O O HH HO OH Om OH Om OH OH Om OH ON ON OH OH OO OH OH OO OH NO OH Om OO OOH OOH OOH OOH OOH OOH OOH OOH OOH OOH OOH OOH OOH OOH OOH OOH OOH OOH OOH OOH OO H3 H3 :3 H8 5 H8 H8 H2 H3 H8 HOV H8 :3 HOV H8 33 as :3 H3 :3 Hmmmgm um mmouoOH O-Ho O-Ho OO-Ho OO-HO OO-qu axe -ono Ozo -HOOHaOusoc OOHOOOHO 0mm OcHouom OOHOOOHO Ounmm OOHOOOHO mmouomq OuHmamucH OOHOOOHO -Hmumo<-z .AOOO-OOOO .HO um Hmpcuoflm ou mcwpuooom MO umnaac mmms ou huwmsmusw xmwm mo mmmucmonom mm pmmmmumxm mum mumnanz .mpflmwaomaflfimmoowaw umeumam Eonm mmumumod Houflpfim vmumHhsumz mo COOHOOHwHucmpH owuumfioupommm mmmz .NH magma 133 c. Lipid III (ceramide trihexoside) Analysis of the methylated derivatives of the platelet trihexoside disclosed three peaks with retention times identical with a and B methyl-2,3,4,6-tetra-O—methyl-galacto- side, a and B methyl-2,3,6-tri-O-methyl-galactoside and a and B methyl-2,3,6-tri-O-methyl-glucoside derived from the corre- sponding standards (Figure 12 and Table 10). Partially methylated alditol acetates gave three peaks with retention times identical with those of l,5-di-O-acetyl-2,3,4,6-tetra-O- methyl-galactitol, 1,4,S-tri-O-acetyl-Z,3,6-tri—O-methyl- galactitol and 1,4,5-tri-O-acetyl-2,3,6-tri-O-methyl-glucitol by GLC (Figure 13 and Table 11). With the reasonable assump- tion, based on previous reports (51,106,263), that the glucose is linked 1+1 to ceramide the gas chromatographic evidence is thus consistent with the characterization of the trihexoside as Gal-(1+4)-Ga1-(l+4)-Glc-(1+1)-ceramide. d. Lipid IV (ceramide tetrahexoside) After methylation and acid hydrolysis of the neutral glyCOSphingolipid IV, GLC of partially methylated hexoses revealed three peaks (Figure 12 and Table 10) corresponding to the retention times of a and B methyl-2,3,6- and -2,4,6- tri-O-methyl-galactosides and a and B methyl-2,3,6-tri-O— methyl-glucoside, respectively. After reduction and acetyla- tion, three peaks having the same retention times as 1,3,5- tri-O-acetyl—2,4,6—tri—O—methyl—galactito1, 1,4,5—tri—O—acetyl— 2,3,6-tri-O-methyl-galactitol and 1,4,5-tri—O-acetyl-2,3,6-tri- O-methyl-glucitol were obtained from the ceramide tetrahexoside 134 (Figure 13 and Table 11). These results indicate that the two galactose residues are linked to the glucose unit in a similar fashion as that of the platelet GL—3a, with the exception of an additional N-acetylgalactosamine which is attached to C-3 of the external galactose moiety. Hence, the structure of this tetrahexoside is GalNAc-(l+3)-Gal- (l+4)-Gal-(l+4)-Glc-(1+1)-ceramide. e. Lipid V (sialo-dihexosylceramide) GLC of partially methylated monosaccharides after methyl- ation and hydrolysis of intact platelet acidic glyCOSphingo- lipid V disclosed the presence of a and B methyl-2,3,6-tri-O- methyl-glucoside and a and B methyl-2,4,6—tri-O-methyl- galactoside (Figure 12 and Table 10). GLC of the partially methylated alditol acetates gave two peaks which corresponded to the retention times of l,3,5-tri-O-acetyl-2,4,6-tri-O- methyl-galactitol and 1,4,5—tri-O-acetyl-2,3,6-tri-O—methyl- glucitol (Figure 13 and Table 11). These experiments revealed that the structure of the carbohydrate chain of this ganglio- side was identical with the N—acetylneuraminyllactose standard. Thus, the structure of this ganglioside is NANA-12+3)-Gal- (1+4)-Glc-(1+1)-ceramide. 7. Enzymatic Hydrolysis of Platelet Neutral GlyEOSphingolipids A TLC separation of products resulting from incubation of platelet neutral glycosphingolipids with specific glycosi- dases is shown in Figure 14. 135 .HOHNOHOOHV amumslaoamaume IEHOOOHOHLU mo Emummm uam>Hom m auH3 Umma mmB H.0aH .aommamaav mumaa w Hmm moaaam pmumomlmam 10mm < .H mama aH mm mEmm mpamaaomaflammommHm pampamum moamummma .ma mama .uamEummHu mahnam aaoauaz Olao umamumam HmaHoHHO .HH mama .uamEumme mshwam uaoauaz mmlam umamumam amaamauo .oa mama .uamEummau mfihmam uaoauHB mmlaw umamumam Hmaamaao .m mama .uamaHEmuaoo mmmpamouomammaa m Uam mmnao mo umapoam m maHBOam mmmpHaHEmmoxmaawpmom121m auHB Olaw umamumam mo mummwaoupma .m mama .Onaw amaHmHHo mau pmam>oo Ima cam pmaaamoo aOHuommH oa .mmmpHaHEmmoxmaHmummmIZIO auak Olaw umampmam mo mummmaoapma .O mama .MHao malao maaaamuaom .mmmpamouomHmmlm amma xomn maam mmmpamouomammla mam aua3 mmlaw umamumam mo mummmaoupma .O mama .mmlaw HmaHmHHo map pmam>omma mam aoauomma oa maa3oam rmmmpamouomammlm amma xmmh auaz mmlaw umamumam m0 mummwaoapha .m mama .uoapoam pmumaoapma map mm mmlao maazoam .mmmpamouomammlo mam auH3 mmlao umamumam mo mummmaoupha .O mama .mmlaw pmmmaoapmaaa mo paDOEm HHmEm m cam mataw maHBOam .mmmp Iamouomamm-m amma xomm aua3 mmlaw umamumaa mo mummhaoupma .m mama .mm-aw Hmaamaao pmam>oomu cam aoauomma oa maakoam .mmmpamouomammla mam auHB mmlaw umamumam mo mummhaoapma .N mama .HOlawv mpamoaon maamaom Uam Ammuaov mpHEmHmmammoxmaHHu magma mammlawv mpaamamoammouoma .Hmalauv mpHEmamoahm Iooaam mam Eouuoa ou mou 80am mpamaaomaaammommam pumpamum moamammmu .H mama .mmmmpamoowam maoHHm> %Q muoapoum mamhaoapma cam mpamaaomaaammoowam umamumam mo wammnmoumfioaao Hmwmalaaae .OH OOOOHO 136 OH muamHm m_ __ O _ m w .2 ’OII‘ O o Hm a 137 a. Platelet GL-4 When pig liver a-N-acetylhexosaminidase was incubated with GL-4, no reaction occurred. Jack bean B-N-acetylhexos- aminidase, however, hydrolyzed the lipid to a product with TLC behavior similar to Fabry GL-3a and platelet GL-3a. b. Platelet GL-3a Degradation of platelet GL—3a to ceramide dihexoside took place with fig d-galactosidase but not with jack bean B-galac- tosidase. Combined reaction of fig a-galactosidase and jack bean B-galactosidase with GL-3a produced one component with a TLC migration rate similar to that of GL-la. c. Platelet GL-2a When fig o-galactosidase was incubated with platelet GL-2a, no reaction occurred; jack bean B—galactosidase cleaved the lipid to a product with TLC properties similar to those of GL-la. The yield of the reaction was limited because of lower activity of the enzyme and the excessive amount of sub- strate present in the reaction mixture. These results showed that the stereochemical configura- tion between the galactose and glucose was 8 in GL-Za, and the anomeric linkage between the two galactose residues in GL-3a had an a—configuration while the internal galactose was B-glycosidically linked to the glucose. In GL-4, the anomeric linkage between the N-acetylgalactosamine and galac- tose had a B-configuration. Due to the limited supply of the galactosidases the hydrolysis product of GL-4 was not 138 degraded further by these enzymes. Nevertheless, in View of the similar TLC behavior of the hydrolyzed product with that of the authentic platelet GL-3a standard, it is likely that the anomeric configurations in the product are identical with those of platelet GL-3a. Therefore, the complete structures of platelet glycosphingolipids are tentatively assigned as Glc-(1+l)-ceramide (GL-la), Gal-(Bl+4)-Glc-(1+1)-ceramide (GL-2a), Gal-(al+4)-Gal-(Bl+4)-Glc-(1+1)—ceramide (GL-3a) and GalNAc-(Bl+3)-Gal—(dl+4)-Ga1-(Bl+4)-Glc-(1+1)-ceramide (GL-4), respectively. The stereochemical configurations of the glycosidic linkages in GL-la and G ganglioside were not M3 determined. 8. Fatty Acid Composition of Platelet Glycosphingolipids The fatty acid compositions of platelet glycosphingo- lipids are listed in Table 13. All of the lipids contained primarily 20:0, 22:0, 24:0 and 24:1, although 16:0 and 18:0 were also present. 2-Hydroxy acids were not found. Long— chain fatty acids (3>20:0) constituted 73-77% of the total normal fatty acids in GL-2a, GL-3a, GL-4 and G , whereas M3 9-14% of normal acids were dominated by 16:0 and 18:0. GL-la showed a high percentage of 18:0 (64.0%) and 18:1 (21.9%) with less of 16:0 (3.0%) and the long-chain acids. .mEmamoumEOHSU mmm ao mmam Hmuou mo mmmuamo Iumm m mm am>Hm mam Uaaow muaDOEmimaB .mpaon mHador mo HmaEaauaumamH aHmaU mm Umuoamp mum mpHum muummm 139 m.m o.mH N.m v.0 0.0 o.m 0.0 m.m o.H o.oH o.m O.o mumauo o.m H.m m.0H m.mH 0.0 m.mm o.mH O.mm m.HH m.mH I .Hu HHON m.mH H.mm m.Hm H.mv 0.0m H.wv m.0m m.ov H.0N o.hm I v.mm ouvm m.m m.m O.m o.O m.m I o.m I m.m 0.0 I o.O oumm I m.m H.v I .Hu 0.m N.v m.m m.v 0.0H .au .Hu HHNN o.mm H.Hm m.mm 0.0m 0.0m 0.0H O.mm m.MH 0.0m m.O .au m.mm oumm m.m m.v 0.0 m.H m.0 I H.n m.N o.O 0.5 m.m m.NH OHON I I O.H I O.H I H.H I O.o I O.m I mumH m.m .Hu H.H o.H N.H .uu H.H .Hu m.H H.O m.Hm O.H HumH m.OH 0.v m.m 0.H v.m m.m m.v m.m m.m 0.0 o.v0 O.VH oan O.H I O.H I m.H I m.H I O.o I m.H O.o ouOH O.oH m.m 0.m m.m O.m O.m 0.0 m.m O.m m.O o.m O.MH ou0H m.o I .au I H.H I m.o I 0.0 I o.H .Hu oumH O.o I .uu I o.H I m.o I m.H I 0.0 m.o OHOH mpaom >uumm Hmuou 00 w OEOOHO omm umHmumHm omm umHouOHO omm OOHOOOHO 0mm umHOpOHa omm umHmuOHa umeuOHa newcoaeoo HOOOOa nmsoHv Ozo O-Ho O-Hu OIHO H-ao OOHEOHOU mOUHaHHomaHammoowHw umHmumHa 0am muhooaauwam .memHm amaam mo aoHuHmomEou pHoa huumm .MH mHamB 140 9. Concentrations of Platelet Glycosphingo- lipids The composition of the platelet glycosphingolipid frac- tion is shown in Table 9, in which the average from duplicate analyses are reported. It is evident that GL-2a was the major neutral glyCOSphingolipid in the platelet lipids, and ac- counted for 64% of the total neutral glycosphingolipid frac- tion. There was an appreciable amount of G GL—3a and M3' GL-4 in platelets too, but GL-la was present in rather small amounts as compared with other tissues. 10. Platelet Gangliosides Total lipids, extracted from 73 units of trypsinized platelets, were partitioned at the National Red Cross Blood Research Laboratory according to the method of Folch §E_ai, (232) using 0.75% of NaCl solution. In so doing, the upper phase was accidentally discarded by Dr. Jamieson. Partition of the lipid with a salt solution suppresses the loss of polar lipids into the upper phase, but it leads to considerable losses of the less polar gangliosides (such as 6M3) into the chloroform phase (264). Since deionized water produces maximal partition of extraneural gangliosides (such as kidney) into the upper phase (264), the lower phase was partitioned a second time, using water according to the method of Folch §p_gi. Thus it was hoped that some of the platelet ganglio- sides could be isolated for structural analysis. 141 On TLC (Figure 7, lane 2) the water—soluble platelet lipids from the upper phase revealed 3 bands (A, B, and C; in the order of their distance from the solvent front). Band A was the major component and the fastest moving band. Gas—liquid chromatograms of the trimethylsilylated methyl glycosides derived from these bands showed a ratio of galactose to glucose of 1.04, 1.03 and 0.86 for Band A, B and C, reSpectively. The molar ratio of N-acylneuraminic acid to glucose was 0.83, 0.53 and 0.40, respectively. In addition, N-acetylgalactosamine was shown to be present in Band C and exhibited a ratio of N-acetylgalactosamine to glucose of 0.40. Table 14 shows the normal fatty acid distribution of platelet gangliosides from the upper phase. Of the normal C C and C were predominant in Band 20:0' 22:0' 24:0 24:1 A; fatty acids with a chain length greater than C20-0 consti- acids, C tuted 80% of the total normal acids. Palmitate and stearate were the major acids in Bands B and C and accounted for 59 and 65% of the total. There was substantially less C22:0' C24:0 and C24:l than in the mixture for the A band. N-acetyl- neuraminic acid was detected in Band A. Sphingosine was the major long-chain base in Bands A and B, along with a trace of sphinganine. Because of the unusually high molar ratio of N—acetyl- neuraminic acid to glucose from Band A, the possibility was not ruled out that Band A was a disialohematoside (GD3 142 Table 14. Fatty Acid Composition of Human Platelet Ganglio— sides from Folch Upper Phasea Component Band A Band B Band c % of total fatty acids 14:0 1.4 3.5 3.9 15:0 0.7 1.7 1.9 ]6:0 13.5 27.7 30.8 1730 tr. 1.4 1.5 18:0 6.9 13.8 15.3 18:1 5.0 7.7 8.5 18:2 --— --- —-- 20:0 10.0 3.4 3.8 22:0 39.5 12.5 13.9 22:1 tr. --- tr. 23:0 2.8 10.1 tr. 24:0 13.3 11.2 12.5 24:1 5.2 2.1 2.4 others 1.8 4.9 4.4 L. aFatty acids are denoted as chain length:number of double bonds. The amounts are given as a percentage of total area on gas chromatograms. 143 ganglioside). Ganglioside A was subjected to mass spectro- metric analysis in the direct probe inlet of an LKB 9000 single-focusing mass spectrometer according to a previously published method (265). Ions from N-acetylneuraminic acid were observed at m/e 173, 186 and 205. No detectable peaks were observed at m/e 261 and 274 which rules out the possi- bility of N-glycolyl forms (a shift of ion location by 88 mass units). The mass spectra also disclosed the presence of sphingosine (m/e 311) as the major long-chain base with sphinganine (m/e 313) as the minor component. These results suggest that ganglioside A is an N-acetyl- hematoside, and ganglioside B is probably also a hematoside with a different fatty acid composition. Ganglioside C is assumed to be similar to G but this needs to be verified M2’ by more detailed analyses. No other slow-migrating ganglio- sides (GD G and TTl) were detected, but it is likely la' le that these gangliosides were lost in the first aqueous salt partition. 11. Platelet Sphingolipids a. Ceramide Ceramide from trypsinized platelets was further purified by thin-layer chromatography (see Method). A single band was obtained, suggesting that there were no ceramides containing a-hydroxy fatty acids. After acid-catalyzed methanolysis, the sphingosine content and ester groups were determined 144 colorimetrically. The molar ratios of bases to esters were 0.77/1.00 and 0.80/l.00 in duplicate analyses. 4-Sphingenine (83.2%) and sphinganine (9.8%) were the long-chain bases present. The fatty acids were mainly 20:0, 22:0 and 24:0. b. Sphingomyelin Sphingomyelin was obtained from the crude phospholipid fraction isolated by silicic acid chromatography. 4-Sphin- genine (75%) was the major long-chain base, sphinganine (15%) was also present, and minor peaks were identified as hexa- decasphingenine (5%), heptadecasphingenine (2%) and octa- decasphingadienine (3%). Platelet sphingomyelin contained a high prOportion of the long-chain saturated fatty acids 22:0 (20.3%), 24:0 (8.2%) and 24:1 (24.0%). 12. Sphingolipid Content of Trypsin-treated and Non-treated Platelets An attempt was made to determine whether there was a difference between the glycosphingolipid content of trypsin- treated and non-treated platelets. As shown in Table 9, the concentration of hematoside was much lower in platelets that were not treated with trypsin; otherwise, the composition was not changed by trypsin treatment. A more detailed investiga- tion of the effect of proteolytic enzymes on the GM3 and ceramide levels in platelets is summarized in Table 15. The concentration of GM3 was increased in all of the treated platelets. Ceramide levels were unaffected by pre-treatment of the platelets with the proteolytic enzymes. 145 Table_15. Concentration of Ceramide and Hematoside in Plate— ‘ lets Treated with Proteolytic Enzymesa Lipid Component Control Trypsin Chymotrypsin Thrombin umoles/g TL GM3 0.26 0.46 0.73 1.25 Ceramideb 9.85 9.44 8.59 9.08 aThese analyses were carried out on separate batches of ap- proximately 3 units of platelet concentrate (l platelet unit is from 450 m1 of whole blood), incubated with trypsin (l mg/unit), thrombin (100 NIH units/unit), and chymotrypsin (l mg/unit), reSpectively, at 37°C for 30 min (except chymo- trypsin which was incubated for 60 min) in 0.85% saline at pH 6.7 before extraction of the lipids. The control was incubated in the same way without added proteolytic enzyme. bAverage of two determinations, each in duplicate. The concentration of ceramide (washed platelets) was surpris- ingly high and accounted for 55% of the observed total Sphingo- lipids including GM3' l3. Platelet PhOSpholipids Platelet phospholipids were studied by TLC using two dif- ferent techniques. DevelOpment in two dimensions was utilized for identification only, while the one dimensional system was used for preparative isolation of individual components. Phosphatidylcholine (230.4 umoles P/g total lipid), phOSpha- tidylethanolamine (128.4 pmoles P/g total lipid) and sphingo- myelin (131.6 Umoles P/gm total lipid) were the major 146 phOSpholipids in this mixture. Phosphatidylinositol could not be found in the methanol fraction from the silicic acid column. This is probably due to the fact that phOSphatidyl- inositol can be eluted from the silicic acid column in acetone-methanol 9:1 (v/v) fraction (105), and is destroyed by the mild alkaline hydrolysis step. Furthermore, initial studies indicated the presence of inositol in a ganglioside fraction from upper phase of Folch wash. The inositol was not free, however, since direct trimethylsilylation of the intact lipid gave no inositol peaks on GLC. Nor was free inositol liberated from lyophilized lipid of upper phase by mild alkali-catalyzed methanolysis. It has been concluded that the upper phase lipid fraction contained a phoSpho- inositide. l4. Fatty Acid Composition of Platelet Phospholipids The major fatty acids in phosphatidylcholine and phos- phatidylethanolamine were 16:0 (26.6 and 4.0%), 18:0 (18.9 and 20.4%), 18:1 (30.6 and 7.9%) and 20:4 (15.1 and 34.9%) whereas phosphatidylserine contained mostly 18:0 (47.9%), 18:1 (15.9%) and 20:4 (15.0%). The phosphatidylethanolamine fraction also contained dimethylacetals of 16:0 (5.8%) and 18:0 (10.0%) derived from the plasmalogen phosphatidylethanol- amine. 1J0 ,- Or .3. RIM 147 15. Platelet Neutral Lipids Beside ceramide, the platelet neutral lipid fraction consisted of triglycerides, cholesterol, cholesterol esters and free fatty acids that were recovered from the chloroform fraction after silicic acid chromatography. Cholesterol, cholesterol ester, triglyceride and free fatty acids were identified by TLC. Palmitate (33.4%), stearate (29.2%), oleate (23.6%) and linoleate (4.2%) were the major fatty acids found in the free fatty acid fraction, while cholesterol esters and triglycerides contained 16:0 (15.7 and 32.4%), 16:1 (30.2 and 10.3%), 18:0 (7.1 and 15.9%) and 18:1 (32.9 and 20.0%) as the major constituents. Triglycerides contained some linoleate (6.2%) as well, but this component was nearly absent in the cholesterol ester fraction. C. Isolation and Quantitative Determination of Porcine Platelet Glycosphingolipids l. Porcine Platelet Concentrates Porcine nlatelet concentrates were obtained from ten liters of whole blood by differential centrifugation. The final suspension contained 2.78 x 107 platelets, 3230 erythro- cytes and 1480 leukocytes per mm3. Homogeneity of the plate- let concentrates was observed by microscopic examination of the air-dried smear prepared from the suspension. Electron microsc0pic examination of the suspension showed a detailed rn 148 structure of the platelets similar to those reported previ- ously (266,267). 2. Porcine Platelet Ceramides Analysis of the ceramide fraction (chloroform-methanol 98:2) revealed three separate bands, designated ceramide A, B, and C. A typical preparative TLC of the ceramide frac- tion is shown in Figure 15. Band A was most abundant, fol- lowed by B and C (very faint bands). Band A migrated in the same area as the ceramide standard containing normal fatty acids. Band B was assumed to be the same general type of ceramide with different chain lengths of acyl groups. Band C was the slowest moving band, and was assumed to be due to the presence of a-hydroxy fatty acids as the acyl moieties, since it had the same R values as the ceramide standard con- F taining hydroxy fatty acids. 3. Glycosphingolipid Content of Porcine Platelets Glycosphingolipids were isolated from crude total lipid (300 mg) from 1.89 g of platelets by a combination of silicic acid and thin-layer chromatography. Thin-layer chromatog- raphy disclosed the presence of four neutral glyCOSphingo- lipids and hematoside. In addition, an extra band was observed between GL-2a and GL-3a. Previous experience indicated that sulfatide exhibited similar TLC behavior in the neutral sol- vent system. Figure 15. 149 Thin-layer chromatography of ceramides from normal porcine platelets. The ceramide fraction, isolated from the total lipids by silicic acid chromatography (chloro- form-methanol 98:2 fraction), were further purified on a 2500 silica gel G Quantum plate with chloroform-methanol 95:5 (v/v) as the developing solvent. Lane 1, reference standard ceramide containing normal fatty acids. In lane 2, reference standard ceramide containing 2-hydroxy fatty acids. In lane 3, ceramides from porcine platelets. 150 Figure 15 151 Gas-liquid chromatography of the methyl glycosides derived from these lipids gave molar ratios indicating these fractions were GL-la, GL-2a, GL-3a, GL-4 and GM3' The molar ratio and concentration of each of the glycosphingolipids are given in Table 16. It is evident that GL-3a was the major neutral glycosphingolipid in the porcine platelet lipids, accounting for 52% of the total neutral glycosphingo- lipid fraction. Table 16. Concentration of Glycosphingolipids in Porcine Platelets GL-la GL-2a I. . GL-3a GL-4 GM3 v—f Umoles Gal Umoles Gal Umoles Gal GalNAc Umoles Gal NANA Umoles 9 TL Glc 9 TL Glc 9 TL Glc Glc g TL Glc Glc 9 TL 0.38 1.2 0.42 1.9 1.05 1.7 0.70 0.18 0.87 0.60 0.24 D. Globoside Concentration in Fetal Pigs Total lipids were extracted from the erythrocytes of the gilt (5.2 ml), 90-day fetus (3.0 ml) and 45-day fetus (1.0 ml), respectively. Glycosphingolipids were isolated and purified by TLC as before. The concentration of GL-4 was determined quantitatively and is summarized below in Table 17. It is obvious that the red cell GL-4 concentration in the 45—day fetus is low when compared to the normal adult which is usually 152 Table 17. Concentration of Globoside in Fetal Pig Erythrocytes Gal GalNAc nmoles Glc Glc ml RBC Gilt 2.00 0.72 234.6 90-day fetus 2.07 0.75 166.7 45-day fetus 2.02 0.81 130.0 around 20 nmoles/100 ml red blood cells, whereas the concen- tration of this glycolipid was at the lower normal range in red cells from a 90-day fetus. If one takes into account the contamination of the fetal blood (45-day) by the gilt's blood during sampling, the actual GL-4 concentration is probably lower than that reported here, but this possibility must be confirmed. E. Radioactive Glucocerebroside 1. Isolation of Glucosylsphingpsine The purity of the hydrolyzed product from Gaucher spleen GL—la was determined by TLC (chloroform-methanol—2.5 NNH4OH 60:40:9) after silicic acid chromatography, using a glyco- sphingolipid mixture (GL-la, GL-2a, GL-3a and GL—4) and galactosylsphingosine (psychosine) as standards. As shown in Figure 16, glucosylsphingosine migrated slightly ahead of the galactosylsphingosine. The chloroform-methanol (8.5:l.5) Figure 16. 153 Thin-layer chromatography of the hydrolysis product from Gaucher spleen Glucosylceramide. In lane 1, methanol eluate from silicic acid chromatography containing exclusively gluco- sylsphingosine. In lane 2, the chloroform- methanol (8.5:l.5) fraction from silicic acid chromatography containing some glucosylceramide and a small amount of unhydrolyzed GL—la. In land 3, reference standard glycosphingolipids from top to bottom are GL-la, GL-2a, GL-3a and GL-4. In lane 4, standard galactosylsphingosine (psychosine). The plate was developed in chloroform-methanol-2.5 N NH4OH (60:40:9). 154 d A FFFF Figure 16 155 fraction contained some glucosylsphingosine and a small amount of unhydrolyzed GL-la, while the methanol eluate contained exclusively glucosylsphingosine. After preparative TLC, the combined weight of the recovered glucosylsphingosine was 7.9 mg (40% yield). 2. Radioactive Glucocerebroside ([14C] stearic acid) The product of the coupling reaction was isolated as described under Methods by combined silicic acid chromatog- raphy and TLC on silica gel G. After development of the plate, the labeled GL-la was eluted from the gel and weighed. The material (4 mg) was dissolved in 4 m1 of chloroform— methanol 2:1 and 10 ul was removed for radioactivity counting. There was 456,567 cpm in 10 ul of material and hence 183 x 106 cpm in 4 ml of total material. With a counting efficiency of 96%, 79.7 uCi of radioactivity was recovered in the 4 mg sample, indicating a specific activity of 0.02 uCi/ug. Assuming a molecular weight of 727 (stearoylpsychosine), the specific activity of the radioactive GL—la was 14.5 uCi/umole. 3. Proof of the Radioactive Glucocerebroside To prove that the synthesized radioactive GL-la was composed of glucose, sphingosine, and fatty acids in a 1:1:1 molar ratio, the product was subjected to acid-catalyzed methanolysis and the individual components were separated and examined by GLC. Figure 17 shows the presence of long-chain bases (tracing at upper right corner) and methyl glucosides Figure 17. 156 Gas—liquid chromatography of trimethylsilyl methyl glycosides and sphingosines from [ 4C] glucosylceramide. The methyl glycosides and long-chain bases of [14C]Gl-la were recovered after acid-catalyzed methanolysis by solvent partition and trimethyl- silylated with pyridine-hexamethyldisilazane- trimethylchlorosilane (10:4:2). The trimethyl- silyl derivatives were analyzed with a Hewlett- Packard 402 Gas Chromatograph equipped with a 6 ft. 3% SE-30 column. The trimethylsilyl methyl glycosides were analyzed at 165°C iso- thermally. The O-trimethylsilyl derivatives of long-chain bases were analyzed by linear temperature programming on the same column, with an initial temperature of 160°C and a program- ming rate of 2°C/min to an upper temperature of 260°C. Top tracing, long-chain bases from [14C] GL-la. Bottom tracing, methyl glucoside from [14C]GL-la, mannitol was added as a marker. 157 1.1, I LUCOSE I ( (Jaw LONG - C HAIN BASES MANN ITOL —v Figure 17 158 (tracing at lower left corner). Figure 18 shows the fatty acid methyl ester derived from the synthesized radioactive compound. Stearic acid was the only fatty acid present. In addition, the fatty acid methyl ester was monitored for radioactivity. An aliquot of the isolated C fatty 18:0 acid was injected into a 3% SE-30 column in an F and M Model 400 Gas-liquid Chromatograph equipped with a Barber- Coleman Model 5190 Radioactivity Monitor. Radioactivity was associated exclusively with methyl stearate from the GLC column. The amount of radioactivity was determined by count- ing 0.5 ul of a 0.5 m1 sample in the scintillation counter. The registered counts were 215,200. The amount of radio- activity was calculated to be 9.4 uCi in the 0.5 mg sample. The recovery of l4C-label in the hexane extracts after acid- catalyzed methanolysis was estimated to be approximately 95%. These results indicate that'the synthetic product of the coupling reaction is a cerebroside containing glucose, sphingosine and radioactive stearic acid. 4. In Vivo Experiment After establishing the identity of the synthetic GL-la, an experiment was planned to administer the labeled GL-la intravenously into a pig so that the turnover of plasma and erythrocyte GL-la could be studied by sampling blood accord- ing to a prescribed time course. A 10 kg Yorkshire pig was chosen from the Swine Research Center. The labeled GL-la was Figure 18. 159 Gas-liquid chromatography of fatty acid methyl ester from [l4Cngucosylceramide. The fatty acid methyl ester of [14C1GL-la was recovered from the acidic methanolysate by hexane estraction. The methyl ester was then analyzed with a Hewlett-Packard 402 Gas Chroma- tograph equipped with a 6 ft. 15% ethylene glycol adipate column maintained at 190°C. Top tracing, standard fatty acid methyl esters: peak 1, C 6-0; peak 2, C18- ; and peak 3, C .0. Bottom tracing, stearic acid methyl ester derived from [14C]GL—la after acid-catalyzed methanolysis. 160 L Figure 18 F ("J n; 161 dissolved in 1 ml of ethanol and injected into the restrained pig, but unfortunately, one of pig's feet got loose during the injection and subsequently displaced the needle at the site of injection. The experiment was discontinued. F. ip_Vitro Study Incubation of intact red cells in plasma containing [14C] glucose led to the incorporation of labeled glucose mainly into GL-la of the neutral glycosphingolipid fraction of the red cells and, to a lesser extent, into GL-2a, as shown in Table 18. GL-3a contained very little radioactivity and no significant amount of activity was detected in GL-4. The radioactivity in GL-la was localized mainly in the hexose moiety, whereas the amount of label recovered in the long- chain base and fatty acid fractions was progressively greater with increasing size of the oligosaccharide group. In GL-3a about 50% of the radioactivity was in the fatty acid frac- tion, Analyses of the neutral glycosphingolipid fractions from PLasma revealed significant incorporation of [14C] glucose jINKD GL-la. The label was found exclusively in the hexose mCiiety, as shown in Table 19. The amount of label present in the long-chain bases and fatty acids did not show any inclrease with increasing chain length of the carbohydrate uni ts . 162 Hm OH HO O O m H.MOH OIaw mm mm mO mm Om OH 0.00 mMIau OH mm >0 mm Om OO 0.0H mNIaU 0 HH mm OO ONH OOO 0.0 mHIaw pHoa muumm mmmm mmoxmm pmma-mupmm- mmmm mmoxmm HE\mmHoEa aOHuomam HOV OOHOOOHHOOHO HOH>HO0OOHOOO HOH>Hno< oHOHoadO .uamaomeoo pHmHH mo mHoEJ\Emo mm 0mmmmhmxm mH >0H>Huom UHMHommm .an ha pmumuHuamav 0am oae >9 pmamHHampmm mmB maoHuomam UHQHHomaHammoomHm Hmauama map 00 muHuamUH mas .aEaHoo pHmm UHUHHHm o O m a0 UmumaoHuomHm mHmB HHS MHV mHHmo pma mo mpHmHH maB .aOHumm IDMHHuammmHuHa >3 mum on maHpaooom pmumaOHuomam Hmaguam mm3 aonammmam HHmm 0mm maB .UoOm um Ha N now No wOOH Hmpaa maop mm3 aOHquaoaH ammooaHm HUOHH 00 HO: ON 0am pooHa mHoa3 mo HE OOH UmaHmuaoo xmmHm aoHumoaaaH mae mHEmam mo aoHuoapaHIumoa mmmd ON mmumooaaumum mHm mo mpHaHHomaHammoomaw Hmnuamz ouaH mmooanHoOHH mo aOHumHomHooaH .OH mHamB 163 mm mm mm OHN mom mm N.N OIaw m m0 Hm mm OON OO O.H mmlao mm mm ON ON Om OH 0.0 mmlao H H mm mH om mOOm O.m mHIaU pHoa wuumm mmmm , mmoxmm pHoa muumm mmmm mmoxmm HE\mmHoEa maoHuomam HOV doHOOOHsOmHO OOH>HOOOOHOOO -HOH>HO0O OHOHOOOO .uamaomfioo pHaHH mo mHoE:\an mm pmmmmamxm mH muH>Huom UHMHommm .aEaHoo pHom OHUHHHO m O m a0 UmumaoHummum 0mHm mm3 HHE omv mammHm mo mpHmHH mas .mHHmo 0mm map op HmoHuampH mm3 aoHquaoo aoHumaaoaH mas mHEma< mo aoHuoapaHIumom mama om mEOmHm mHm mo OUHQHHomaHammoome Hmauamz ouaH mmooaHuHUOHu mo aOHumHomHooaH .mH mHamB 164 These results suggest that the reticulocyte-rich cells are capable of incorporating labeled glucose into the hexose moiety of GL-la biosynthetically. The synthetic mechanism would probably involve the enzyme glucosyl transferase and ceramide as the substrate, since porcine erythrocytes were shown to contain an appreciable amount of ceramides (9). The data may account for the early source of GL—la observed previously in the ip_yiyg_experiment (45). Recently, similar results had been observed with normal rat red cells i2 viE£2’(268). Sloviter (269) has suggested that reticulo- cytes can synthesize glycosphingolipids, and it is uncertain whether the reticulocytes or erythrocytes were responsible for the biosynthetic activity in this study. The separation of the red cells according to age by ultracentrifugation technique was not complete due to overloading of the red cell suspension. It is likely that the erythrocyte fraction used in the incubation contained most of the reticulocyte population and perhaps some other cells. The recovery of a significant amount of label in erythro- cyte GL-2a suggests that this lipid can also be synthesized ip_yippg, presumably by stepwise reaction of ceramide with glucosyl and galactosyl transferases. The lack of signifi- cant radioactivity in GL-3a and GL-4 indicates that further synthesis does not occur during the 2 hr incubation time. 165 G. Red Cell Fractionation 1. Separation by Age of Erythrocytes from Normal and Anemic Blood Separation of a red cell population into groups having different specific gravities was obtained by ultracentrifuga- tion in discontinuous density gradient of bovine serum albumin, devised by Piomelli ep.ai. (259). In this technique, the youngest cells (reticulocyte-rich), which are dense and larger in size, remain on top of the gradient, while the older cells (reticulocyte-poor) are more dense and remain in the lower part of the gradient. A preliminary experiment with normal and anemic canine blood disclosed six bands after ultracentrifugation, as shown in Figure 19. In the normal dog (tube A), two very faint bands were observed at the top; most of the cells were congre- gated at Band 3 and 4 with some in Band 5 and 6 and a few cells were at the bottom of the tube (Band 7). In the anemic dog (tube B), the reverse was true. Most of the cells were in Band 1, 2 and 3, some were in 4, and a faint ring of cells were observed in Bands 5 and 6. Each band was recovered and washed extensively with saline, and the volume of packed cells ‘was visually estimated in a calibrated centrifuge tube. Band 7 was not estimated. No attempts were made to recover cells in between these discrete bands. Cells derived from the individual bands of the anemic dog were suspended in saline and air-dried smears Figure 19. 166 Separation of young and mature erythrocytes from normal and anemic dogs. Approximately 1.2 ml each of red cell suspensions from normal and anemic dogs (reticulocyte count, 15-20%) were centrifuged on 6 layers of albumin solution ranging in specific gravity of 1.075- 1.100 in cellulose nitrate tubes (1" x 3 1/2"). The tubes were centrifuged at 4°C for 30 min at 253000 rpm in a Beckman Model LS-SO Ultracentri- fuge with a 25.1 swinging bucket. The red cells were fractionated into 7 bands. Band 7 at the bottom of the tube is partly hidden from the picture. A small percentage of cells is dispersed in the albumin layers. A, normal dog. B, anemic dog. 167 Band Band Band Band Band Band Band H Figure 19 II>(.II.)I\)I--l 0‘ II 7 168 Table 20. Discontinuous Density Gradient Ultracentrifugation of Canine Blood Bands Volume of Packed Cells (ml) Normal Anemic 1 0.05 0.25 2 0.05 0.50 3 0.30 0.10 4 0.45 0.03 5 0.06 0.05 6 0.05 0.05 were prepared and stained with new methylene blue for micro- sc0pic examination. Reticulocytes were prevalent in Band 1 and a few were observed in Band 2, but no reticulocytes were seen in Band 3. The cells from Band 1 were large, and the cell size seemed to decrease from Band 1 to Band 6. Photo- micrographs were made of these slides; unfortunately, they did not turn out. 2. Separation of Young and Mature Erythrocytes from Normal and Anemic Pig Blood A similar experiment was carried out with normal and anemic porcine erythrocytes to determine whether pig red cells would behave in the same way as dog erythrocytes in the albumin gradients. The profile of red cell separation by gradient ultracentrifugation was exactly identical to that of 169 canine red cells. The hemoglobin content of each individual red cell band is listed in Table 21. 3. Electronic Sizing of Fractionated Red Cells The red cell size distribution of individual bands was obtained by a Celloscope on a blood sample 30 days after giving the glucose label. Mean channel number of each layer was calculated as described in Methods, and is liSted below (see Table 22). There was a significant difference in the size of Bands 2 and 5. The small differences observed between, the middle bands are probably due to experimental errors. 4. Checking Hemolysis of Erythrocytes in Albumin Solutions After incubating the pig erythrocytes with the respective albumin solutions at 4°C for 30 min, the cells were washed‘ with saline and centrifuged. The supernatent solutions seemed to be clear. The hemoglobin content of each supernatant solu- tion was recorded and summarized in Table 23. In addition the Soret band of the albumin and supernatant solutions was examined at 416 nm against saline as blank. The results indicate that there was no significant amount of hemolysis of the red cells in the albumin solutions (see Table 24). 170 aHaonoEmmm O.m 0.0 0.0 0.0 m.O 0.0 O N.O O.m m.O O.HH O.NH m.H m N.ON H.ON H.N O.HO m.NO 0.0 O N.Om m.mm O.N 0.0H 0.0H O.H m N.OH 0.0H m.H O.H O.H N.O N 0.0H N.OH m.H O.H m.H N.O H mHmEmm HmaHmHao Hmuou Aw Emvmam mHmEmm HmaHmHHo Hmuou Aw Emv am pamm aH aOHuomHm HO O aH aOHpomHm «o w m aomm mo mam w aomm mo mam w mHa UHEmaa mHm Hmanz .poon maHoaoa OHEmaa 0am Hmanz Eoum pm>HHmo maOHuomHm HmapH>HpaH mo aOHumHuamUaoo aHaonoEmm .HN mHamB Table 22. Mean Channel Number of Individual Bands Fraction- ated by Density Gradient Ultracentrifugation Bands Mean Channel Number 2 52 3 49 4 50 5 48 6 45 Table 23. Hemoglobin Content of the Supernatant Solutions Albumin Concentration (%) Optical Density of Supernate 30 32 34 36 38 40 0.000 0.005 0.002 0.000 0.002 Table 24. Soret Band of the Supernatant Solutions Optical Density Albumin Concentration (%) Albumin Supernate 30 0.506 0.541 32 0.564 0.545 34 0.626 0.580 36 0.607 --- 38 0.760 0.730 40 1.53 1.849 4 172 H. 13 Vivo Study 1. Induction of Anemia in Pig 123-6 In pig 123-6 rapid blood regeneration was induced by phlebotomy. The reticulocyte count is the simplest and most effective means of measuring bone marrow activity, since the reticulocyte numbers in the peripheral blood increase to a degree which is related to the severity of the anemia pro- duced (270,271). The hematological profile of this anemic pig is shown in Figure 20. When 10% or more of the blood volume was removed daily from the pig, considerable reticulo- cytosis developed. The close correlation between reticulo- cytes, hemoglobin and hematocrit was apparent. Reticulocyte count was expressed as absolute percentage (271), and the error of reticulocyte measurement was~estimated to be : 5-10%. No attempt was made to classify reticulocytes on the basis of relative maturity. To assess bone marrow production, correction of the absolute reticulocyte count was made according to the calcu- lation described by Hillman and Finch (271) for humans, using an average packed cell volume and reticulocyte count (absolute %) of 25 and 24, respectively, derived from the results ob- tained on the first ten days after administration of the l4C-labeled glucose. A corrected reticulocyte count of 12 was obtained, which directly reflects the rate of red cell production. In this connection, the pig was making blood at 173 .UOHHmm OmUIHO may mo 0am maa HHuaa HH mmp anw HHHB um pamHmB aHmm 00 0m3oHHm mm3 mHm map 0am .0 amp ao mHmaoam>mHuaH am>Hm mm3 mmooaHmHUOHIDO H05 OH mo mm00 mHmaHm a .UOHHmm mHau maHaac maOHumH map 00 mxmpaH map maHHOHHummH ma uamumaoo UmaHmpaHmE mmB mHa may mo pamHmB mae .mmEaHo> pooHa pmumEHumm may 00 wOH meumEonammm pmHmavm >HHm0 aBmHUauH3 UOOHQ mo maaHo> mae .OH mmp amaoaau maHpmmHa mo pamum map SOHO mHHmp meuowamm mm? >anoamHam .mvaum o>H> mm map uaoamaoaap 0am poHHmm aoHuoapaH map maHaap 0ImNH mHm aH mHEmam mo HamEmon>mo .OO OHOOHO 174 (+ ' % mm .LIHOOLVWBH $8 8 S (+‘%WSELLKJOTIO T 40 30 TIME (DAYS) O N _ LL33X+.%ub) NIBO'IOOWBH Figure 20 175 a rate approximately twelve times higher than normal. Nevertheless, one should approach this value with caution, since it is based on the assumption that the means of cor— recting reticulocyte count in humans can be similarly applied to the pig. At any rate, pig 123-6 did not appear to be defective in erythropoiesis and the bone marrow seemed to be quite active in response to the bleedings. Figure 21 is a microscopic view of the dried blood smears prepared on Day 0 (A, before the bleeding starts) and at 8 (B), 12 (C), and 25 (D) days post-induction of anemia. A relatively high concentration of reticulocytes were present on the 8th, 12th, and 25th days; some of the reticulocytes, which were rather large in size, were probably macroreticulocytes (imma- ture reticulocytes). Various degrees of anisocytosis and poikilocytosis were also observed. Starting on Day 18, the reticulocytes began to oscillate within a fairly narrow range (20-27%) while the hemoglobin and hematocrit rebounded to 6-8 g % and 20-25 volume %, reSpectively (Figure 20). To maintain a steady anemic state in the pig, daily removal of blood samples was necessary. After intra— venous injection of the radioisotope on Day 25 (which is Day 0 for the ip_ziyg experiment), the weight and the anemic condi— tion of the pig was maintained constant for an additional 10 days by daily bleedings and restriction of ration. Beginning on Day 11 through Day 81, the weight of the pig was allowed to gain by lifting the diet regimen. Figure 21. 176 Photomicrographs of blood cells from Pig 123-6 during induction of anemia. Air-dried smears of freshly drawn heparinized blood stained with supravital stain new methylene blue. The smears were examined under oil—immersion. A, day 0 (before bleeding starts); B, 8 days post- phlebotomy; C, 12 days post-phlebotomy; D, 25 days post-phlebotomy. The large cells containing heavily staining, aggregated clumps of reticulum represent young reticulocytes. The cells contain- ing smaller amounts of reticulum in "punctate" foci are reticulocytes in an advanced stage of maturation. 177 Figure 21 178 Blood was sampled at approximately 5 day intervals, withdrawing 40 ml at each prescribed time point. Soon after the termination of the daily phlebotomy, the reticulocyte count, hemoglobin and hematocrit all returned to the normal range and plateaued. This is also evident from Figure 22, which shows the blood smears obtained on Days 10 (A), 30 (B), 50 (C), and 70 (D) after the radioactive label was given. It can be seen that a significant number of reticulocytes were still present on Day 10; however, subsequent to the termination of bleeding on daily basis, the blood picture returned to normal. 2. Administration of Radioisotope Because of the previous experience in the ip_yiyp_experi- ment with the radioactive GL-la, routine practices of saline injections were made on other available pigs at the swine barn and a standard procedure was developed to prevent any more of the previously encountered problems in injection. The injec- tion of [14C1glucose was given on Day O after the initial start of bleedings, as indicated in Figure 20. The injection went smoothly and the pig was in excellent condition after— wards, except for a few wobbling minutes on her feet. 3. Concentration of Porcine Blood Sphingolipids The results of TLC and GLC analyses indicated that the porcine plasma and erythrocytes contained the same neutral glycosphingolipids as previously described (45). In addition, Figure 22. 179 Photomicrographs of blood cells from Pig 123-6 during the in vivo study. Air-dried smears of freshly drawn heparinized blood stained with supravital stain new methylene stain. A, 10 days after injection of [U-14C] glucose as a pulse label. B, 30 days post- injection of the label; C, 50 days post-injection of the label; D, 70 days post-injection of the label. 180 Figure 22 181 the erythrocytes contained an appreciable amount of ceramides consisting of normal fatty acids and a small proportion of d-hydroxy fatty acids, as shown in Figure 23. The plasma fraction contained a rather small amount of normal acyl-type ceramide; no d-hydroxy fatty acid—containing ceramide was detected by TLC (Figure 24). GM3 ganglioside from the Folch lower phase was detected in both plasma and red cell fractions of the blood. The con— centration of plasma G was in the same range as the plasma M3 GL-4 (Table 25). GM2 ganglioside (Figure 25, lane 4, Band B) was partially identified as the major ganglioside from the Folch upper phase of plasma by TLC, and G was also present M3 as expected (Band A). In porcine erythrocytes (Figure 25, lane:3) GM2 ganglioside was the major component from the upper phase, along with some G (Band A) and GMl (Band C). M3 This is different from human red cell gangliosides (upper phase) wherein the major component (Band X) was reported (117) to migrate between GM2 and GMl' as shown in lane 2 of Figure 25. Concentrations of porcine plasma and erythrocyte sphingo- lipids over a 81-day period are summarized in Table 25. The concentrations generally did not fluctuate to any great extent. Specific activities of porcine plasma and erythrocyte sphingolipids over a period of 81 days are tabulated in Tables 26 and 27. Figure 23. 182 Thin—layer chromatography of erythrocyte cera- mides from Pig 123-6. The ceramide fraction, isolated from the total lipids by silicic acid chromatography (chloro- form—methanol 98:2 fraction), were further purified on-a 250u silica gel G Quantum plate with chloroform-methanol (95:5) as the develop- ing solvent. In lane 1, reference ceramide standard containing normal fatty acids. In lane 2, reference ceramide standard containing 2- hydroxy fatty acids. In lanes 3, 4, and 5, erythrocyte ceramides from pig 123-6 obtained on days 81, 70, and 66 respectively. E CEIE' ema 'Mm 183 Figure 23 Figure 24. 184 Thin-layer chromatography of plasma ceramide from Pig 123-6. The ceramide fraction, isolated from the total lipids by silicic acid chromatography (chloro- form—methanol 98:2 fraction), were further puri— fied on a 250p silica gel G Quantum plate with chloroform—methanol (95:5) as the developing solvent. In lane 1, reference ceramide standards containing normal fatty acids (fast—moving com- ponent) and 2—hydroxy fatty acids (slow-moving component). In lanes 2, 3, 4, and 5, plasma ceramide from pig 123-6 obtained on days 14, 10, 9, and 8 respectively. 185 3 4 5 ”In a: 2 Figure 24 186 Concentrations of Plasma and Erythrocyte Sphingolipids from Pig 123-6 During the Metabolic Experiment Table 25. RBC (nmoles/ml) Plasma (nmoles/ml) GL-la GL-2a GL—4 Cer-HFA CervNFA M3 GL-3a GL—4 G Time 20.6 1.4 1.6 250 2.1 1.6 2.0 1.4 2.1 2.3 6 hr. 12 hr. Day-l 19.9 .3 .2 2.2 1.9 2 8.7 8.7 22.2 -23.0 258 192 232 1.5 2.4 1.3 6.3 2.5 2.8 1.7 4.1 .0 1.8 1.9 205 1.4 210 7.9 23.1 1.8 1.8 2.5 1.6 2.2 1.6 1.8 2.3 21.7 23.1 2.1 206 1.8 4. 7.4 1.9 2.2 2.3 2.0 .8 3.0 1.5 2.6 1.3 1.1 223 .7 6.8 13.8 2.1 1.7 2.8 1.8 2.8 1.9 3.3 2.0 2.6 3 5.8 10 14 20 25 30 35 41 18.2 .6 .5 3.4 5.3 17.8 1.1 6.9 3.6 21.6 1.6 7.5 19.7 2.4 2.0 2.7 1.5 2.0 1.6 2.1 1.9 1.8 19.2 .3 2.4 19.9 46 50 53 23.6 0.9 . 3.5 2.7 13.4 1.8 1.4 2.9 1.4 1.8 1.3 3.7 11.3 .2 2.5 . 3.6 5.2 55 57 21.0 1.2 3.8 24.1 2.0 1.5 .1 2.3 60 63 19.5 1.9 1.7 0.9 2.4 1.4 2.3 3.0 2.7 9.1 66 70 75 81 1.5 2.0 1.7 1.3 1.2 10.8 16.0 3.1 3.0 . 19.0 222 1.6 2.3 7.2 Mean Figure 25. 187 Thin—layer chromatography of Folch upper phase gangliosides from normal human erythrocytes, normal porcine plasma and erythrocytes. The Folch upper phases from human erythrocytes (17 ml), porcine erythrocytes (15 ml) and porcine plasma (15 ml) were dialyzed, hydrolyzed by mild alkali and dialyzed again in the cold for 72 hr. The crude ganglioside mixtures were purified by TLC, using 250u Uniplate (heat activiated at 120°C for 30 min) and developed in two—solvent sequential system. Solvent systems were chloroformdmethanol— 1.25 N NH4OH (60:40:9) and n—prOpanol-water (70:30). In lane 1, reference standard ganglioside mixture containing mono-, dir, and trisialoganglioside. In lane 2, gangliosides from human erythrocytes which contained hematoside (A), and the major ganglioside (X), structure unknown. In lane 3, porcine erythrocyte gangliosides partially and tentatively identified as hematoside (A), G (B), and GM (C). In lane 4, porcine plasma gangiio- sides partially and tentatively identified as hematoside (A), and GM2 (B). 188 Figure 25 189 Table 26. Specific Activities (cpm/pmole) of Plasma and Erythrocyte Sphingolipids GL-4 Plasma (corrected RBC for anemic Cer- Cer— Time GL-la GL-2a GL—3a GL-4 volume) GM3 GL-4 NFA HFA 6 hr. 1132 351 1165 418 418 3053 23 1597 4984 12 hr. 1118 305 843 350 350 1174 2086 9535 Day 1 709 260 2291 1320 1320 1786 485 1173 5050 2 502 252 1587 901 901 674 1247 1020 3297 3 85 171 1390 720 720 328 1268 1618 6335 4 1100 145 932 910 910 199 1076 1333 7301 5 82 64 761 855 847 199 828 1340 8590 6 73 52 356 329 329 131 1482 6236 7 75 73 670 395 348 95 608 1079 5793 8 68 94 315 180 182 --— 1070 4713 9 28 39 430 298 301 16 420 2112 6364 10 50 63 469 208 208 56 1791 8291 14 25 39 145 79 73 20 1738 5457 20 38 118 164 110 81 --- --- --- 25 --- 20 99 31 23 --- 1426 8444 30 141 —-- -—- 153 116 --- 1667 10328 35 67 29 156 78 60 166 1326 7139 41 --- 139 84 77 57 69 1311 2590 46 38 --- 154 106 77 958 759 7119 50 515 81 171 156 115 --- 800 7365 53 --- --- 54 173 129 --- 1349 23005 55 153 --- 464 65 48 90 1682 6735 57 --- 243 122 130 96 315 1236 9370 60 336 --- 390 1262 928 --- 1645 44541 63 226 89 424 797 588 --- 2847 40427 66 --- 282 665 207 154 --- 1404 --- 70 144 --- 147 .274 206 --- 3497 --- 75 --- --- 200 199 148 2472 1474 8502 81 37 --- 175 200 149 --— 3464 3741 190 Om OH III Om O.H NO OO N mO OOH Hm ON ON III 0H 0.0 Om ON O O OH mO Om Om OH Om N.H mO O0 ON Om mm OO OO O O OH m.H OO O0 0 III mH 00 Nm NN m mH O.H mmH OHH mH HO mmH m0 OO O III OH O.N HOH 0mH OH mO HO O0 OO O m 0H N.H HOH ONH OH ON OO Om mm O O 0H O.H O0H 0mH mH 0 ON mm NO HH ON mH m.H ONH OOH MH w mm mm mm O III HN O.H OON HOH ON Om Nm Om MH OH O OH O.H NmH OmH NH Hm Nm 0O N0 0 HH mN O.H OOH OOH OH OH 0N HO OH O m ON O.H mHN NmH 0H m 0H mm OH HH OH mm O.H OMN mON OH O NO Om HO OH O OH O.H OmN mON HN 0 O mN III OH III OO O.H mOm OHm OH O III ON Om OH OH 0O O.H NOO mOm OO O HO OH III ON 0H OO O.m HON HON Om NN mN OH 0O OH OH O0 O.N Omm Omm mm NH mm O OOH OH mH mO O.H 0mm Hmm Om HH mN m O0 0H mH Om O.H OOO mmO mO HH OH O N0 0H OH mO O.H OOO OOO Om HN ON 0 OOH OH OH OO O.N NmO NmO 0O OH 0N O OO ON 0H OO m.N OOOH OOOH OOH OH Nm O OO Nm 0N HNH O.N OONH OONH HNH HN N0 m mm O0 O H0 0.0 OHOH OHOH 0m HH ON N OO mm OH OO O.m m0O m0O Om mH Om H Omo 0N HO Nm NmH 0.H mOH mOH ON HN 0O Ma NH mN 0m OH mO III ON ON OH OH 00 an 0 ..mm0HmoHHOamw mOHmoHHOamo ammIHmu «mZIHmU aszamU mEaHo> OIao mmlau mNIao mHIao mEHB . mmmam mmmaa 0mm 0mm mEOmHm UHEmam HOW mammHm mEOmHm mEOmHm mammHm HmmmO Hmmmo Omuomuaoov 0mm mEmmHm OIaO 0mm mvaaHomaHamm mHOUOHCOOHm 0am mEmmHm mo AHE\EQUV OmHuH>Huo¢ UHMHommm .ON mHamB 191 4. Metabolism of Porcine Blood Sphingolipids As stated in the Methods, specific activities were ex- pressed as cpm/ml of plasma or red cells and/or cpm/umole hexose; in the latter case, the specific activity of an indi- vidual hexose moiety was obtained by dividing by the number of monosaccharide residues present in the particular glyco- sphingolipid, since it had been shown previously in several studies that the sugar moieties are equally labeled (45,179). Turnover curves for each of the sphingolipids were constructed from linear and/or semilogarithmic plots of specific activi- ties versus time. The vertical bars at each data point represent the accumulated random errors for various steps in the procedure. Table 28. Random Errors for Various Steps in the Procedure Procedure % Error Measurement of plasma and RBC volume 2.0 Lipid extraction and silicic acid column chromatog- raphy 0.5 Mild alkaline hydrolysis and dialysis 0.5 TLC and elution of individual fractions 1.0 Acid-catalyzed methanolysis and GLC 2.0 Pipetting 2.0 Sphingosine assay 2.0 Radioactivity counting (Beckman LS-150 Manual) Counts 100 20.0 200 15.0 400 10.0 800 7 0 1600 5.0 4000 3 0 192 The estimation of the error of a computed result R from the errors of the component terms A, B, and C is summarized below: If R is calculated as product or quotient where R = AB/C, then the random error for R is equal to'{(SA/A)2 + (SB/B)2 + (SC/C)2}% where R is the calculated result; A, B, and C are the measured quantities from which R is derived. SA’ B SC are the absolute standard errors 0 A, B, and C. To maintain an anemic pig in a state of hematopoietic equilibrium for an extended period of time is difficult; hence, the pig was allowed to gain weight after the initial lO-day period and corrections were made in the calculations for the large increase in blood volume that occurred after the initial 10 days. 5. Incorporation of Labeled Glucose into GL—4 The labeled glucose was rapidly incorporated into hexose, sphingosine and fatty acid moieties of plasma GL-4 (Table 29), which reached a maximum specific activity at 24 hr after the initial pulse label (Figure 26) and then declined rapidly for the next 10 days. The specific activity of plasma GL-4 remained rather constant over the next 35 days until around Day 50, when significant relabeling occurred throughout the whole molecule, reaching maximum activity on Day 60. Semi-logarithmic plots of plasma GL-4 specific activity versus time for the first 10 days and at Day 60 subsequent 193 Table 29. Specific Activities of Hexose, Fatty Acid, and Sphingosine Moieties from Plasma GL-la and GL-4 Specific Activities (cpm/umole) Time GL-la _ GL-4 Sugar Base FAME Sugar Base FAME 6 hr 1132 70 113 418 331 428 12 hr 1118 101 101 350 220 580 Day 1 709 166 207 1320 1199 562 2 502 380 354 901 561 714 3 85 165 249 720 793 470 4 1100 169 174 910 1005 1580 5 82 211 116 855 714 740 6 73 182 116 329 603 256 7 75 149 124 395 1289 960 8 68 95 34 180 1969 911 9 28 118 111 298 267 350 10 50 124 160 208 254 392 14 77 116 99 79 198 158 20 38 56 44 110 897 189 25 --- --- --- 31 401 123 30 141 141 130 153 427 183 35 67 150 80 78 312 150 41 --- 4 2 77 1497 413 46 38 171 57 106 1830 704 50 515 676 162 156 389 233 53 --- --- -—- 173 108 260 55 153 688 420 65 392 196 57 --- ——— --- 130 695 695 60 336 773 202 1262 2825 689 63 226 339 225 797 1328 1593 66 --- —-- --- 207 826 275 70 144 168 125 274 1205 657 75 --- --- --- 199 1789 1391 81 37 45 22 200 1900 300 194 .OOHamm OmOIHO am am>o mEHu mamam> HHE\EQUV mumo Ioaapwam 0am Ammoxma mHOEJ\anv mEOmHa anm .mOHEmHmoImHOIHO+HVIHmo -HO+HV-HOO-HO+HO-oOzHOO .O-HO Oo mmHuH>Huom OHOHOOOO Oo OoHa “OOOHH .0ImNH OHm aH mOHmonoHO mphmoaaumam 0am mEmmHm mo Hm>oaaae .ON maaOHm ON muaaHa s g § § é. ("D“) BSOXBH 310W "IMHO-0,, § 35.50 NEE. 00 CO Om ON 0. H H H w _ _ O \\||.|d |||||||||| II", .\- . a a // I I I I/ 0 \ 42041.n— 0 0mm 0 8.5-040.440 .48- 934-30 § § 8 0 (+‘Iw/wd0-o,.) OIdI‘IOOA‘IO :JO All/\llOVOIOVB § E. _I--* --‘ § 196 to maximum incorporation revealed a decline in radioactivity at a rate suggesting half times of approximately 3.6 and 5.0 days, respectively. A linear plot of specific activity (cpm/m1) versus time revealed that the daily increment of incorporation of label into the erythrocyte GL-4 (Figure 26) was slower and the GL-4 did not reach maximum specific activity until 3 days after the injection of [U-14C]glucose. This corresponds well with the finding of Bush gp gi. (273) that the time of maximum incorporation of 59Fe into the erythrocyte was 3 days and less for the normal and anemic pigs, respec- tively. After maximum incorporation was attained the specific activity fell steadily; approximately 65% of the label was lost during the next 10 days, after which there was a slow and gradual decrease until around Day 60 when a much more abrupt decline occurred. Corrections for the increase in blood volume were based on the normal blood volume figures reported by Bush 32 ii. (272), since attempts to correct for the volume increase on the basis of anemic pig blood volume (273) did not reveal any significant changes (Tables 26 and 27) in the profile of the turnover curves. There was no change in the specific activities between 6 hr and Day 14 for both plasma and erythrocyte GL-4. Between Days 20 and 81, the specific activ- ities of erythrocyte GL-4 were raised slightly while the specific activities of plasma GL-4 declined by about 20%. However, the basic profiles of the two turnover curves re- mained the same. 197 Figure 27 shows the linear plot of Specific activity (cpm/pmole hexose) versus time of erythrocyte GL—4 from seven data points obtained during the first 10 days subsequent to the initial pulse label. It can be seen that following the initial incorporation phase, there was an exponential decrease in specific activity of the glycolipid, while the hemoglobin and packed cell volume of the red cells remained rather con— stant throughout this period. If one compares the peak of radioactivity of erythrocyte GL-4 with that reported previously for the normal pig (45), there is an approximately 6—fold increase in the amount of 14C incorporation in the red cells of the anemic pig. The semilogarithmic plots of specific activities of samples obtained from the pig 123-6, plotted against time, are depicted in Figure 28. It can be seen that the part of the curve between Day 25 and Day 57 is not a flat plateau; instead, it follows a gradual exponential decline in specific activity which is terminated by a more precipitous drop (this is also evident from the linear plot in Figure 26). This suggests that the turnover of erythrocyte GL-4 represents more than one process. The exponential phase may be ascribed to random destruction, whereas another process, which is age-dependent, is evidenced by the precipitous terminal decline in specific activity. Figure 28 also suggests that there are at least three GL-4 fractions (pools) in the red cell pOpulation which undergo 198 Figure 27. Turnover of red cell globoside in pig 123-6 over a lO-day period. Linear plot showing the decay of the 14C- labeled hexose moiety of red cell GL-4 in a period of 10 days post-injection of [U-14C] glucose. 199 ON mHaOHm TIQIEXQEQQOEOEXKSZHI RBC GNbeflmbGNd3££EVWIE I200)- 800% 600+ T01 OOOXO: 52328-0... IO 5 TMHEUNWS) 200 .m>aao Omomm mau mo maauma OHmmaaHa map OmNHmmaaEm mEHu mo aoHuoaam m mm OIaw mo HHE\EmoV OHH>Humm UHMHomam 00 uon oHeauHamOoHHEmm .0ImNH OHm aH mpHmOQOHm muwmoaaumum mo am>oaaae .ON mHaOHm 201 ON muaOHm $2922; 8 2. 8 OO 9. OO 8 O. H H H H H H H H . :09 .H. as OO .H. :OOOO.H. Hugs—418-040 H. 4o --_40- 04.2440 0mm -0. On 1 OOO.“ H188. § (IN/W900...) Oldl‘IOOA'IO :IO AllAllOVOIOVB ..9.n 202 .OOHHmm OmUIHO am um>o mEHp mo aoauoaam m mm HHE\EmoH mummouaumnm 0am Hmmoxma mHoEJ\EmmO mEmmHm anw .mpHEmHmOIonIHOIHH -HOw-HO+HO-HOO .OO-HO mo mmHnH>Hnom OHOHomam Oo uoHa “OOOHH .0ImNH OHm aH mOHEmamOmeoxmaHau muwmoaaumam 0am mammHm Mo am>oaHaB .ON masOHa 203 ON mHaOHm Hm>49 ME:- 4H2m4-E 0 0mm 0 wo_$_4mmo-o-.o.-_4o.-_4o (Hg—‘IuI/uIda-c)M ) OldI‘IOOA'IO :IO MIAIiOVOICIVU filé § 1 g. § 1% 4% («a») asoxz-IH 310WTI/Wd3-0.. 204 turnover at different rates. Within experimental inaccuracies, the rate of decrease of specific activity from Day 3 to Day 9 can be fitted to a single exponential component having a half- time of 5-6 days. If one assumes that GL-4 is an integral part of the red cell membrane and remains with the cell through- out its normal life span, the data can be correlated with a cellular mean life span of 8 days. It is likely that this represents turnover of reticulocytes. The value was estimated from the fact that the mean life time is 1.44 times the half- time when a process is represented by a single exponential decay (274,275). The rate of radioactivity decay from Day 25 to 57 can also be fitted to a single exponential component having a half-time of 45 days, which would indicate a mean cellular life span of 65 days for these cells. Finally, a mean life Span of 9 days (t 1/2 = 6.2 days) can be obtained for the rate of decrease in specific activity of GL-4 from Day 60 to Day 75. It is interesting to note that the half- time of erythrocyte GL-4 between Days 60 and 75 is very similar to the half-time of reticulocytes subsequent to maximum label- ing on Day 3. 6. Incorporation of Labeled Glucose into GL-3a Labeled glucose was rapidly incorporated into plasma GL-3a (Figure 29), which reached a maximum Specific activity at 24 hr. The specific activity then declined rapidly before reaching a plateau around Day 15. Significant relabeling did 205 not occur until Day 50, and GL—3a reached a new peak of specific activity on Day 65. A semilogarithmic plot of plasma GL-3a specific activity over the first lOvday period revealed a half-time of 3.5 days. Erythrocyte GL-3a showed a similar pattern of incorpora- tion (Figure 29) as that of red cell GL-4; except for differ- ences in specific activities, the two curves were almost superimposable. Maximum incorporation of the radioactive label occurred on Day 3, followed by a decline with a loss of approximately 65% of the label during the next 10 days. After that the level of specific activity decreased slowly with time until around Day 60 when a precipitous loss occurred. A semilogarithmic plot of Specific activity versus time (Figure 30) disclosed the presence of at least three GL—3a pools. The rate of decrease of specific activity from Day 3 to Day 9, Days 20-57 and Days 60-81 could be fitted to exponential components having half-times of 5.5, 45.0 and 6.3 days, respectively. This would indicate cellular mean life spans of 8, 65 and 9 days. 7. Incorporation of Labeled Glucose into GL-2a Plasma GL-2a showed a rapid incorporation of the label (Figure 31), reaching a maximum specific activity after 6 hr, which was much earlier than observed with either GL-3a or GL-4, then the radioactivity was lost during the next 9 days in a similar manner as the other two plasma glycosphingolipids. 206 .uamamaam mmB m>aao Omomc may no maauma OHmmamHm .mEHp mo aoHumaam m mm mHHmo 0mm maHoaom aH mmIaw UmHQOHIOOH mo Hm>oaaau man OaHzoam uon 0HeauHHmOoHHEmm .0ImNH OHm aH mOHEmHmonmoxmaHHu muhmoaaumam mo Hm>oaaae .om aaaaaa 207 om mHaOHm Hm>4OH 22:. 00 0O 00 00 CO on ON 0. fl _ H _ H _ H H H OOOOOOOIH. mo .24mm0 I040-440 -440 um m .0. oo— (Iw/de-ofl) Oldl‘IOOA'IO :IO All/\llOVOIOVH é 208 .OoHamm OmUIHO am Hm>o mEHu no aoHumaam m mm OHHmmaHH OmupoHa mums HHE\EQUH mumooaauwam 0am Hmmoxma mHoE:\EmoH mammHm maHoaom Eoum .mUHEmamOImHOIHOIHHIHmw .mNIa0 mo mmHuH>Huom OHMHmmmm mae .0ImNH OHm aH mOHEmammHmmouomH mumooaauwam 0am mEmmHm mo Hm>oaaae .Hm maaOHm 209 HO mHaOHm Hm>4e 92:. 42044.”. 0 00¢ o 09—24mw0-040-440 . q. "o’—" ' co 0 HI. 8 (+‘Iuvwd0-o,.I OIdI‘IOOA'IO :IO All/\llOVOlClV-H 4 6 § 2 i8? 3. (--u--) BSOXEIH 31OWTI/ was-3,. 210 The half-time for turnover of GL-2a during this lO-day period was 3.0 days. The lipid was relabeled beginning on Day 50, peaking at Days 57 and 66, after which the Specific activity declined nearly to zero after 70 days. When red cell GL-2a was examined (Figure 31), two small peaks of maximum specific activity were observed at 12 hr and on Day 3, respectively. Approximately 50% of the label was lost during the next 20-day period, after which signifi- cant relabeling occurred with a complex pattern that showed maximum of radioactivity on Days 50, 63, 70, and 81. The turnover of GL-2a in red cells was different from that of GL-3a and GL—4 in several respects. First, there was an ad- ditional early peak of incorporation at 12 hr. Second, the radioactive peaks between Days 50 and 81 were more complex. 8. Incorporation of Labeled Glucose into GL-la The patterns of incorporation of label into plasma and erythrocyte GL-la (Figure 32) were quite similar to those of GL-2a. Rapid incorporation of the label was observed imme- diately after administration of [U-14C]glucose, with a peak of specific activity at 6 hr. Loss of this label occurred within the next 10 days and haletimes of 0.75 and 7.5 days could be obtained from the semilogarithmic decay curve subse- quent to maximal incorporation. Erythrocyte GL-la also showed rapid incorporation at 6 hr and on Day 3, and,half-times of 0.8 day and 4.2 days could 211 .UOHHmm mmOIHO am Hm>o mEHu mamam> OHHmmaHH OmHHOHm mam3 HHE\EmoH mgmoouaumam 0am Hmmoxma mHOEJ\EmoO mEmmHa maHmaom 80am .mOHEmamOIHHIHHIOHo .mHIa0 mo mmHuH>Humm OHMHommw mas .0ImNH OHm aH mOHEmHmoHOmomaHO mumooaaumam 0am mEmmmm mo Hm>oaaae .Nm mHaOHm 212 GLC-CERAM I DE 0 RBC °I I I [,4 / / "C(\ .. 9 / L\‘§~ . ~——o-—a.:"""’ - “\‘ v-—O—-' ~\\ 0+ \ \ .c \ \ 4 al‘ . 2 ‘- 2 I—.—. n H .1 CL -0- D 1 DIP I- .——.——- q (+‘lm/wd3-Ofl) Oldl'IOOA'IO :IO All/\llOVOIOVH l l.500~ § s (--n—-) BSOXBH 310w Tf/ Wd0-0,. 50 60 TIME (DAYS) 30 4O 20 IO Figure 32 213 be obtained from the loss of radioactivity of these two early peaks on semilogarithmic plots of specific activity versus time. The specific activity of GL-la hexose was 35% higher at 6 hr than it was at Day 3, indicating a much more active and rapid incorporation at the earlier time. The amount of radioactivity incorporated was higher in the plasma fraction than the corresponding red cells. The early peaks of incorp- oration from both plasma and red cell GL-la were higher than those observed for the plasma and erythrocyte pools of GL-2a. Although this seems to be consistent with a precursor-product relationship, other criteria cannot be established and hence one cannot relate this to a classical precursor-product rela- tionship. Between Day 50 and Day 70, significant relabeling occurred in the plasma pool of GL-la, reaching maximum specific activi- ties on Days 50, 60, and 70. These peaks paralleled exactly the relabeling peaks observed in erythrocyte GL-la, suggesting that the plasma and red cell membrane pools of GL-la are in rapid exchange, which confirms the observation made in the ip.yip£g experiment mentioned earlier. 9. Incorporation of Labeled Glucose into Folch Lower Phase G“ Ganglioside Maximum incorporation of label into plasma GM3 ganglio- side occurred at 6 hr (Figure 33) and 90% of the label was lost during the next 10 days. A semilogarithmic plot of specific 214 .OOHamm mmOIHm am am>o mEHu mo aOHuoaam m mm OHHmmaHH Omuuon mums HHE\EQUH mumooaaumam 0am Hmmoxma m$081\8mov mEmmHm maHmaom anm Hmmmam Hm3oH aoHomH mOHOOHHOamO 0 mo mmHuH>Huom OHMHommm mae z .0ImNH OHm mo mumooaaumam 0am mammHm aH mOHOOHHOamO m 0 mmmaa amBoH amHom mo Hm>oaaae .mm maaOHm mm maaOHm Hm>49 E22. 00 04 00 ON 0. 215 . VIA. . l \.\\II1 Juli-Alli . l I/ \Q. / D \ I II.I H W ,m x + x + . 1.. . m ,4 xi .. W H x O l x . u H \ . .IAIH \ ”_ .A 00. 000._ 0 u 9 _ m m H.H H _ 0 OOO M188 _ A... _ d u m w 4.2m44d n .I 00m 0 + 09.241000404404242 000 (4000.0 (--O--) BSOXEIH E'IOWTT/WdO-O. 216 activity versus time suggested a decay curve which was bi- phasic in nature, indicating that there were two pools of hematoside with half-times of 0.9 day and 1.9 days, reSpec- tively. There was no detectable label in plasma GM3 between Day 10 and Day 35, and significant relabeling did not occur until Days 46, S7, and 75. The Specific activities of plasma GM3 during this period were higher than those of erythrocyte G . M3 The incorporation of radioactive label into red cell GM3 ganglioside (Figure 33) was rather slow and did not reach a maximum specific activity until 3 days after administration of the labeled glucose, which was similar to the results ob- served with GL-3a and GL-4 of erythrocytes. At this time, the plasma G fraction had reached its peak incorporation and M3 the specific activities were rapidly decreasing. An additional peak of radioactivity was observed on Day 6 which was absent in the other glycosphingolipids examined so far. The appearance of radioactivity between Day 40 and Day 81 was very similar to the patterns noticed previously for GL-2a and GL-la; maximum relabeling was recorded on Days 50, 60, 70, and 81. 10. Incorporation of Labeled Glucose into Folch Upper Phase Gangliosides The patterns of radioactivity incorporation into the upper phase gangliosides of both plasma and red cells bear some simi— larities to that observed in the G ganglioside (Figure 34). M3 217 .GOHHoQ mmplam cm Hm>o wEHu mo cofluocsm m mm omuuoam mums wumoonsumum cam mammam mcflouom Scum mmowm Ioflamcmm woman woman noaom mo AHE\EQOV mmwufl>wuom oamaomam one .mumma and mo musoouausum 0cm mammHm Eonw mooflmOHHmomm mmmnm nomad noaom mo Hm>ocuse .vm musmflm 218 I-a-Iw/de-ow) CIIdI'IODA'IS) :IO All/\ILOVOIOVH 8 8 8 8 s I I I I I *9” _ m ,1 LOU / a I Q ,/ —-I I‘ —1 ‘3 x 8 “\ ° \ \ I \\ Lu 2 "t? - E 45- GP 6 < . o. S. P a d. _ 0- 02 I I 0:0. : (g on :0_ “- I u—.—-o udl-a I +0} - I l H“ "I I {r a I I 45. 1 [I ._c}—.: I - ‘ 0 'i 3 £34} M l | - 0 0 0 0 0 9 co 0 <- N (+‘Iw/wd0-3,.) Oldnoom :JO All/\llOVOICIVH 70 IO TIME (DAYS) Figure 34 219 The labeled glucose was incorporated rapidly, reaching maximum specific activity at 6 hr, and approximately 70% of this label was lost during the next 10 days. A semilogarith- mic plot of specific activity versus time during this period clearly illustrates the biphasic nature of the decay curve, suggesting that there were two pools of glycosphingolipids with half-times of 0.9 day and 1.8 days, respectively. After the initial loss of the label from the plasma gangliosides, the specific activity remained rather constant for 50 days, then on Day 60 relabeling of this lipid occurred, reaching maximum specific activity at around Day 70. A semilogarithmic plot of the loss of the label versus time subsequent to Day 70 disclosed a half-time of 7.5 days. When the upper phase gangliosides from erythrocytes were examined, maximum specific activities were detected on Day 3 and Day 8, followed by loss of approximately 70% of this label. Relabeling of the erythrocyte gangliosides are shown to occur around Day 35, and peaks of specific activities were detected on Days 41, 50, 60, 66, and 81. From the peaks of specific activities, the amount of radioactivity seemed to be decreasing from Day 41 to Day 81 with time during this period. Semilogarithms of the loss of label subsequent to Day 50 and Day 66 were plotted against time and revealed half- times of 6.3 and 11.2 days, respectively. 220 ll. Incorporation of Labeled Glucose into Ceramides Linear plots of specific activity (cpm/ml RBC) versus time for ceramide containing normal and d-hydroxy fatty acids (Figure 35) demonstrated rapid incorporation of the label at the early times similar to the results observed in GL-la and GL-2a, namely, reaching maximum incorporation at 12 hr and on Day 3 and loss of approximately 50% of their label during the next 7 days. Relabeling of these two lipids occurred between Day 50 and Day 81; ceramide containing normal fatty acids exhibited peaks of maximum specific activities at Days 70 and 81, while d-hydroxy fatty acid-containing ceramide gave peaks of maximum relabeling at Days 53 and 70. The l4C-labeling of the ceramide fraction of plasma was negligible compared with the other glycosphingolipids (Figure 35). When specific activity (cpm/umole sphingosine bases) was plotted against time for the red cell ceramides (Figure 36} the profiles of radioactivity incorporation at the earlier times were shown to be essentially the same as mentioned immediately above. However, ceramide containing normal fatty acids was shown to be relabeled on Day 53 and Day 60 in addi- tion to Days 70 and 81; while the fraction of ceramides con- taining a-hydroxy fatty acids exhibited peaks of relabeling on Days 53, 60 and 75. There were similarities in the patterns of relabeling of ceramide containing normal fatty acids, Gl-la, GL—2a, and also GM3. 221 .Uoflhmm mmolam cm Hm>o mEHu momum> maummcwa owuuon mHoB muhoounuhum can mammHm mafionom Eoum mmoflfimumo mo AHE\Emov mmflufl>fiuom oamwommm one .wlmma mam CH mmoHEmnmo muwoougumum can memHm mo H¢>OCHDB .mm onsmwm 222 mm munmflm $549 wsE. m0.o4 E4“. émozifiwg 4 00.04 >._.._.4m >x0m0>Idmm o m0_04 >._I_.4m I_<_2moz.0mm o mugs—$.30 0m 0v Om O.N 0. a + 0+ 8 * 0.? —.-—t A <3 (0 O (D 8 (Iw/de-ofl) OIdI'lOOXlS) :JO AilAllOVOIGV/B 0N. 223 .moflom wuumm mxoupmnlm can Hmauoc mcwcwmucoo mmowamumo Hawo own mo momma camao Imcoa omaonmalova on» no Emwaonmumo on» mGHBOnm .mEHu msmum> Ammmma mawmomoflnmm mHoEn\Em00 muw>flpom owuaommm mo uon Hmmcwa 4 .mlmma mHm aw mmowfioumo mumooucuhum mo um>ocude .mm musmwm 224 mm wHDmfim Am>49 m: _.r 8 8 8 on 8 o. _ 4 VII! . _ a _ _ I .. , 4 .. MW .4 p \ I/ _ x + 4‘ — \ / + I. // N “ w x ’* fl 4 // _ . . I i _. . ,i ........ ” .fl .4 1% + J... 1080. .._ , .._ , _ .. . u n _. .. A «m o r : m a .- _. / m _ t. r .n n n s _oooIm.8Q8W — . 0 . 3 an. n _ M w N . . .2 W a i m 609.388» m : m m j m m. + H r 8.8 the >xomefn ( ..I if u. 8.8. Es. .2252. .89 .82 $9258 0mm 225 12. Fractionation of Red Cells According to Age from Pig l23-6 During ;n_Vin Study Fractionation of red cells into individual bands by the density gradient technique was performed according to the procedure outlined in Methods. Figure 37 shows the separation of red cells as bands after ultracentrifugation of the red cell samples obtained on Days 8 (A), 40 (B), and 70 (C), in a discontinuous density gradient of bovine albumin solution. On Day 8 (A) Bands 1 and 2 were clearly visible at the top of the gradient, representing reticulocytes and young cell frac- tions of the red cell population. On Day 40 and Day 70, these bands were hardly noticeable because the pig was no longer as anemic and no significant amount of reticulocytes were present at that time (Figure 20). This was also evident by the appearance of heavier bands at the lower half of the tube which is quite characteristic of the red cell separation patterns observed previously with normal pig and dog. 13. Globoside (GL-4) Concentration in Porcine Red Cells of Different Ages The concentrations of erythrocyte GL-4 from 53 samples obtained between Days 2 and 81 are tabulated in Table 30. No changes with age in the relative proportions of the major erythrocyte glycosphingolipid component have been observed. 14. Globoside (CL-4) Turnover in the Fractionated Red Cells Since GL-4 is the major porcine erythrocyte membrane glycosphingolipid, efforts were made to concentrate on the Figure 37. 226 Separation of young and mature erythrocytes from pig 123-6 during the in vivo study. Two and one-half milliliters of porcine red cells were centrifuged on 6 layers of albumin solution ranging in specific gravity of 1.075- l.lOO in cellulose nitrate tubes (1 1/2“ x 3 1/2"). The tubes were centrifuged at 4°C for 30 min at 25,000 rpm in a Beckman Model LS-50 Ultracentrifuge with a 25.2 swinging bucket. A, 8 days post-injection of the radioactive label. B, 40 days post—injection of the radioactive label. C, 70 days post-injec— tion of the radioactive label. 227 Figure 37 228 450 4.0 005 0H0 40H 045 00H 00.0 00.0 m 000 0.0 050 050 000 0H5 000 05.0 04.0 0 mm 0.0 04 00H HOH 00 0H0 44.0 00.0 H OH 000 0.0 0H0 50H 00H 000 000 Hm.0 04.H m 000 4.0 000 004 004 0m0H 50H 00.0 00.4 0 00 0.0 54 05 00 00 000 00.0 00.H H 0 000 4.0 H00 000 H50 0000 II 000 00.0 00.4 m 05 0.H 00H 00H MHH 40H 000 00.0 00.H 0 00 0.0 00 54 H4 00 40H 4H.0 05.0 H 5 000 0.0 H05 05 00 4M0 0H0 00.0 04.0 4 HO0H 5.H HHOH 000 000 000H 05H 04.0 04.0 m 00H 0.H 04H 00H 40H 00H H40 Hm.0 00.H 0 00 0.H 00 HOH 05 40 400 00.0 0m.H H 0 000 0.0 000 HOH 40 000 00H 0H.0 00.0 4 040! 0.0 HMOH 000 00H 000 00H 40.0 00.H m 00 0.0 000 00H 00 00H 400 4H.0 05.0 0 0m m.H 00 00 5mI 04 040 0H.0 00.0 H 0 045 0.0 5HOH H0 00 440 000 00.0 00.0 4 HOHH 0.0 HO0H 0H0 04m 05Hm 00H H5.0 05.0 m 000H 4.H 000H 0H4 40m 405H 00H 0m.0 0H.0 0 000 0.H 000 040 000 000 000 Hm.0 00.H H 4 H00 0.0 005 00H 00H 4HOH 00H 04.0 00.0 m mmmH 0.H HHOH 4Hm 400 0400 00H 00.0 0H.0 0 000 0.0 000H 044 000 000H 0H0 00.0 00.H H m H4H 0.0 00m 04 m4 004 000 00.0 04.H m 000 0.H 000H 00H 00H 00mH 0H0 0H.0 00.0 0 MHOH 0.0 4000 0H4 5Hm 0000 000 00.0 00.H H 0 onEmm mHmEMm mHmEmm omm HE mHmEMm cows mwcmm mwmo «4.4.0 Q0 «.4.0 “02400 Ammmmmv ammmoxmmv emu .mmmmmm mmmmmm omm Ego Emu .HE ucoEHuomxm OHHonmumz on» 0CHuda mwcmm HHoo com wouMGOHuomum wnu Scum oGHmOQOHo mo moHuH>Hpom UHMHoomm 0cm coHumuucoocou .00 oHnma 229 ham 0 \Emov wuH>Huom UHMHommm n .4.0** AmHoan\Emov wuH>wuom OHMHommm u .4.m« 000 cum: 0 0.0 44H mm 0H 0H 05H 00.0 00.0 5 0H H.4 04 00 0m 4m 000 40.0 00.4 0 H0 m.H HMH 00 mm 0H 00H 0H.0 05.0 0 H0 50 m.H 040 00 00 00 00H 00.0 4H.0 5 0 0.0 0H mm 40 4 00H 04.0 00.0 0 00 0.0 00 00 40 0H 000 04.0 04.0 0 05 0H 0.4 000 40 0H OH 040 40.0 0H.0 5 00 4.4 04 04 44 40 000 00.0 00.0 0 0H 0.0 0H 00 m0 0 000 04.0 00.0 0 00 00 0.0 540 0H 0 0H 000 H0.0 00.0 5 00 0.4 00 40 00 om 4H0 40.0 00.0 0 0H 0.0 50 mm 00 4H H00 04.0 0H.0 0 00 0H 0.0 0H0 00 40 0H 040 40.0 0H.0 5 05 5.0 00H 40 00 00 00H 00.0 00.0 0 50 4.0 05 00 54 04 400 04.0 00.0 0 0H0 0.0 004 5H 0H 40 500 40.0 0H.0 4 H4 0 o.H 0m 0 4H 0 000 40.0 0H.0 5 00 0.0 00H 00 04 05 00H 00.0 00.H 0 00 0.4 00 55 H5 00 00H 00.0 0H.0 0 00H 4.0 00H 50 0H 00 040 50.0 50.0 4 om 00H 0.0 000 0MH 00H 050 . 00H H4.0 0H.0 m 05 0.4 H0 00H 40H 45H 0H0 40.0 00.4 0 0MH 4.0 HOH H0 00 00 H50 HH.0 04.0 H 00 050 H.H 404 50H 00 . 000... HHO H0.0 00.H m 00H 0.4 00H 004 0H0 000 0H0 4H.H 04.0 0 04 4.H 00 55 00 04 000 00.0 00.H H 4H 230 isolation of this lipid and study its turnover in the frac- tionated groups of red cells. Specific activities were expressed as cpm/umole hexose and cpm of total hexoses/g hemoglobin. In both cases, specific activities were corrected for the increase in blood volume of the pig during the in_vivo experiment. It will be noted that not all the bands were analyzed for radioactivity. Samples were randomly chosen at the beginning, middle and end of the experiment, with emphasis on analyses of the top and bottom layers of cells. Table 30 summarizes the specific activities obtained from GL-4 of each individual group of red cells fractionated during the ig_vizg study. Labeled glucose was incorporated into the hexose, sphingosine and fatty acid moieties of the erythrocyte GL-4 molecule, although in the majority of cases, most of the label was found in the hexose residues. The data in A of Figure 38 show that the radioactivity of GL-4 was concentrated in the top layers at the beginning of the experiment, then moved gradually downward to reach the bottom of the tube at 50 days. At Day 60, only 25% of radio- activity (from Day 50) was found to remain in Band 7. However, at Day 70, radioactivity reappeared in Band 7, reaching an almost identical specific activity as that of Day 50. This label was subsequently lost to approximately the same level as that of Day 60 by Day 81. During this period, relabeling seemed to occur in Band 5 also. Essentially similar results were obtained when specific activity was expressed as cpm of 231 .mflmmHmom uHEHmm ou HHmEm oou mmB mmEHu >Humw map um 5I4 mccmm CH mHHmo 0mm Umum>oomu mo ucoosm one .mwsum o>H> mm map 0CHHDU mHm>uoucH 08H» moowum> um ucmHUmnm wuHmcmo co cowummsmfluucmomuuHo ma cmumCOHuomHm mmumoonnuwum mQHouom,Eoum 4Iqo UmHmanIU H 00 wchMHmoEwa 0\mmmoxmc Hmuou Mo 8000 wuH I>fiuom OHMHommm .m .DomEHummxo o>0> :H on» 0GHMDU mHm>umucH oEHu msowum> um ucmemum muHmomc co coHuwmsmeucmomuuHo 0Q wouMGOHuomHm mmuhoounpmum mcwouom Eoum 4IA0 UmHQOHIU4H mo Ammoxm: mHOE:\Emov mpH>Huom UHMHommw .4 .HonH mesm HMHuHsH map Hmpmw Hm>uwucH oEHu um oImNH 8H0 scum mwflmonon Hamo can «0 auH>HDOMOH4mu no aofiusnfluumflo .00 musmwm 232 mm munmflm 2988.2 5.0. x 28.0... $86: .52 “.0 E>53 D5><0 00>3 850 .450 098 850 950 03 0 Box! 30.243.02.286... 0.000 00.000 00.000.00an 00. o E [ FL PIIL .L E _ _ _ i. l... .28 958 8.3 8.3 .18 8.3 8.3 9.3 4 0 >40 233 total hexoses/gm hemoglobin, as shown in Figure 38, B. The distribution of radioactive GL—4 from individual bands at various time intervals was quite similar to that observed in A of Figure 38, except that the radioactive label was not concentrated in the bottom layer until Day 70. There was no appreciable radioactivity detected at Day 81. The specific activity of Band 5 seemed to increase with time between Days 60 and 81. Figure 39 is another variation of presenting the data shown in Figure 38, B. Essentially similar comments can be made. Figure 40 is a linear plot of the specific activity (cpm/umole hexose) versus time of the data from Table 30. In A of Figure 40, specific activities of Bands 1, 2, and 3 were plotted from Day 2 to Day 20 whereas specific activities of Bands 5, 6, and 7 were plotted from Day 30 to Day 81 as demonstrated in Figure 40, B. The [U—14C]glucose was rapidly incorporated into GL-4 of the top layer, reaching maximum specific activity 48 hr after the initial pulse label. The label then disappeared rapidly until, at Day 6, almost none was detectable. Approximately 98% of this label was lost during this time period, suggesting that the turnover rate of GL—4 of this reticulocyte—rich fraction was very rapid. The remaining 2% of the radioactivity was maintained throughout the rest of the 20-day period. Incorporation of the radio- active label into GL-4 of Band 2 was quite rapid, reaching a 234 .wmoosHm m>wuom000~u on» no cofluowncqumom m>m0 H0 wow 00 somzhmn 5 can .0 .0 mvcmm Eoum 4IA0 mo wufl>Huom UHMHommm may mucommummu .hnmflm .4H >00 smsonnu 0 >60 um 0GHccwmon 0 0cm .0 .H wwcmm Eoum 4IA0 mo 0uH>Huom UHNHommm msu mpcmmonmmn .ummq .ucmHumum muHmcmw so cowummsmHuucmomuuHs ma wmumcoHuomum mumz mwumoounwmuo ocHoHom .HmQMH may 00 coaumuuchHsvm “ovum mHm>uopaw 08H» mDOHHm> um 0I00H 0H0 Eoum >DH>Huom UHMHommm oUHmonon HHmo won on» no COHumucmmmum £0000 H00 .00 wh90Hm 235 8'5 RBC GALNAc-GAL-GAL-aC-CERAMIDE ®_ K) \ 60 w 50 TIME (DAYS) W—JI’I' "I'. 40 $.00 Figure 39 I4 IO 8 TIME (DAYS) (W W0) sasoxaII “mos: .an All/\ILOSVOIOVU 2;. 3: 3 g E I: g I E: g E E I 5’ E E r I E F“ F I‘“ F F 1:: F V §§ §§ é (W W3.) sasoxaH ‘IVlOl :IO All/\IOVOIOVB 34567 2 500— v -p ——~_—__- 236 .unmmm mmmw 0H um H0 0mm 0:0 00 000 cmmBuwn 5 can .0 .0 mUcmm mo mHHmo Eoum 4Iqw mo Hm>ocusu may mucmm Iwnamu .0 .00 0cm 0 0mm :mwsuwn 0 now .0 .H mvcmm mo mHHmo Eoum 4Iqw mo um>ocuou ms» mucommummu .4 .usmwvmum muHmcmw co coHummsMHnucmo ImuuHs 0Q mHHmo 00 came woumc00uomum mHHmDUH>0©cw Eoum Uo>HumU 4Iqu How 08H» 00 COHuocsm m mm >00>Hpom OHMHommm map 00 uon HamCHH 4 .mmumoonsuwum mcHouom mo mmooum 0wHMGOHuomum MHHMDUH>HUCH :H Hm>OCH5u memonoHo u .IHMWH .04 musmwm 237 04 musmflm 4 6.29 ms: 8.25. m2: 9. 8 8 0.4 cm 8 4. o. m m . III. n o u + O O o d m n m 3 .. 0254 m n 024m. 0 ozuoo mooon onu mo ououoC OHmonm IHn one .0 0Com CH 4Iqo mo no>0CuCu .uanm .0 0Com CH 4Iqo mo Co>o ICHCu .oHUUHz .H 0Com CH 4Iq0 mo uo>OCHCu .umoq .uCoHnoum muHmCoU Co COHuomsmHHuCoomuuHC 0n nonoCOHuooum mHHoo Eoum 0o>Huo0 4Iqw you oEHu mo CoHuoCCm o mo muH>Huoo OHMHoomm 00 uon UHEanHomoHHEom .COHuomsMHHuCoooHuHC uConoum muHmCoU an oouoCoHuooCm mohmoounpmuo oCHoHom mo mUCon oounu mow on» CH uo>0CHCu oUHmonon .H4.oH50Hm 240 H4 ousmHm .0>40. 0.2.... .940. 0.2.... .940. 0.2.... 00 0. _ 0 00 0. o 00 o. -0. _ . _ _ H 4 £8 00".. £80.70. . . :8. m mbgumhfifiwp / 9.80.0". I .08.. 8.550.348.3338 8.. 0.0 0-0 70 000.0. BSOXBH BWOWTI/WdO-O . 241 much slower turnover time; the half-times were 4.0 and 9.3 days for the two corresponding pools. 15. Turnover Values of Plasma and Erythrocyte Glycosphingolipids Turnover values were determined from semilogarithmic plots of specific activity (cpm/umole hexose) versus time over the period subsequent to maximum labeling. The semi- logarithmic plot provides the half-time (t 1/2) (275), from which turnover times can be calculated (tt = 1.44 t 1/2). Turnover rates (p) can then be calculated from the total circulating pool sizes of the glyc05phingolipid (r) and the turnover time (tt) from the simple relationship described by Zilversmit (275), as shown below. The calculated turnover values for the plasma and erythro— cyte glycosphingolipids are summarized in Table 31. Because of the biconcave nature of the turnover curves, two turnover values (pool A and B) were obtained for both plasma GL-la and GM3 ganglioside. In GL-la, pool A turns over ten times more rapidly than pool B; in GM3 pool A has turnover twice as high. It is also apparent that GL—Za turns over more rapidly than GL-3a and GL-4. The turnover times of GL-2a, GL—3a, and GL-4 are approximately the same, whereas the turnover times of GL-la and GM in pool A are much shorter than in pool B. 3 242 onu How .050. amm.mm nmom 0n nouuomou oECHo> ouxoounuwno 0C0 osmon Eouw wouoHooHoo mH oNHm Hoom Houoe .0Hm UHEoCo n .oEHu mamCo> .omoxon oHOEJ\EQo. wuH>Huom UHMHoomm 00 uon UHEnuHummoHHEom Eoum 0o>Hno0 moB 0\H no H000 H0009 MO UCDOEd n m s mm.o mm.~m 04.ma m.m 4H m.CH m4.oo ok.m 0.4 mIocmm ma m.aH 4m.4m No.5 m.m am 40.54 4n.ko 44.H o.H mswcmm Hm me.4a Nm.05 m5.4 m.m mm m4.mo oo.oH mo.4 m5.o Huocmm 4-46 omm mm 00.0 mm.H 4H.m o.H m2 Hp om.o NH.H om.s 0.0 o «swung om 4m.o 4b.. ma.m o.m 4-46 mammHm om mm.o mo.n 4o.m m.m mmIAo mammsm mm mn.o mn.m mm.4 o.m 44-46 mammHm m 4m.o m5.m om.oH om.s mm 4m.m mm.m mo.H mn.o 44-46 mammHm .4. .mmw\mmnoea. .mwnosn. .m044. .m046. oflmHHoozno xoc Com UoNHmonuC>0 opom Co>OCCCe oNHm Hoom Houoe oEHe Co>OCH5e oEHuIMHom onHmHHomCHnmmoo>H0 ouxoounuxum 0C0 mamon oCHouom mo mooHo> Mo>OCCCe .H0 oHnme 243 Approximately 20% of GL-Za, GL-3a, and GL-4 is metabolized per day, a finding that agrees quite well with previous stud- ies in humans (222) and pig (45). This cannot be said for pool A of GL-la and G however, since both pools are very M3' active metabolically and approximately 90 and 80% of the total glycolipid in this pool are synthesized per day. Table 31 also presents the turnover values of GL-4 derived from the red cell groups during the ultracentrifuga- tion study. It is evident that two pools of GL-4 (designated as pool A and B) were present in each of the fractionated bands. The turnover times of both the A (early) and B (late) pools became longer as the cells aged, while the turnover rates were rapid in Band 1 but became less rapid with aging. This seems to indicate that the young cells were more meta- bolically active and thus turning over much faster than the older cells. This was also obvious from the percentage con- tribution to the total GL-4 pool each day listed in the last column of Table 31. V. DISCUSSION Since the initial discovery of globoside and hematoside in erythrocytes (8,68,120), much has been learned about the occurrence of various neutral and acidic glycosphingolipids in plasma, erythrocytes and to some extent, leukocytes. However, as an important member of the formed elements in blood, platelets have not been examined for glycosphingo- lipids in great detail. The first part of this thesis re- search was concerned with the isolation and structural characterization of platelet glycosphingolipids. In addition, studies were made of the composition of human plasma ganglio- sides. A. Human Platelet Sphingolipids and Plasma Gangliosides The results presented in this dissertation were obtained with a pooled sample of trypsinized platelets from 73 donors. Simultaneous analyses of the glyc0protein composition of these platelets necessitated the trypsin treatment; however, a com- parison was made onaismaller scale of untreated platelets and platelets treated with different proteolytic enzymes. 244 245 The solvent-soluble Sphingolipid fraction of platelets consisted of sphingomyelin, ceramide, a family of neutral glycosphingolipids resembling those of plasma and erythro- cytes (37), and GM3 ganglioside. Separation of the neutral glycosphingolipids and hematoside was easily accomplished on a single commercially available TLC plate. The molar [—1 ratios of sugars and permethylation studies coupled with use of specific glycosidases, confirmed that the major neutral 3-- glycosphingolipids were galactosyl(Bl+4)-glucosylceramide (GL-Za), galactosyl(dl+4)-galactosyl(Bl+4)glucosylceramide (GL—3a), and N—acetylgalactosaminyl(81+3)-galactosyl(dl+4) galactosyl(Bl+4)glucosylceramide (GL-4). The GM3 ganglioside consisted exclusively of N-acetylneuraminyl(2+3)—galactosyl- (1+4)glucosylceramide. The stereochemical configurations of the glycosidic linkages in GL-l and GM3 ganglioside have not been determined. The composition of the neutral glycosphingolipid frac- tion of human platelets was recently reported by Snyder, Desnick, and Krivit (224), who found that GL-2a (38-43%) and GL-4 (23-31%) were the major components. In my study, GL-Za was relatively much more dominant (64%) and GL-3a were present in about the same proportions (ca. 16%). The rather large amount of GL—2a present in platelets makes them quite unique among the formed elements of blood, and it is tempting to speculate that GL-2a may play a special role in the main- tenance of platelet structural integrity or may be involved 246 in immunologic reactions, since the GL-Za from human epiderm- oid carcinoma explants in rats has been shown to have strong haptenic properties (276). GL-Za has also been shown to be the major neutral glycosphingolipid of human leukocytes (38, 39). Platelets and leukocytes differ substantially in this respect from erythrocytes, in which GL—4 is the major neutral - glycosphingolipid (37). In porcine leukocytes, the concen- T—“T tration of GL-3a exceeds that of GL-Za (46). Leukocytes are able to carry out the synthesis of GL-la and GL-2a (40) as are cultured cells of bone marrow (147). No comparable experiments have been done with platelets to assess their enzymatic activity for glycosphingolipid biosyn- thesis. The similarity of fatty acid composition in platelet GL-Za, GL-3a, GL-4 and GM3 seemed to support the concept of metabolic interconversion commonly believed to exist in glycosphingolipid metabolism, GL—2a has been indicated as an /GM3 GL-2a \GL-3a 3:;- GL-4 intermediate in the biosynthesis of ganglioside and globoside (192). However, the marked differences in fatty acid compo- sition observed in GL-la and ceramides were indeed puzzling. In previous studies of porcine blood glycosphingolipids, the fatty acid composition of GL-la was found to be significantly different from GL-4 (45,277); the latter glycolipid contained 247 appreciable proportions of 22:0 and 24:0 whereas more 16:0 and 18:0 were found to be present in GL-la. Furthermore, in both porcine platelets and erythrocytes (45,278) the concen- tration of 16:0, 18:0, and 18:1 decrease with an increasing number of hexose units in the glycosphingolipids, whereas the concentrations of 22:0 and 24:0 increase. This does not seem to be the case with human platelet glycosphingolipids, which points to variations in species specificity among these cellular components. It was previously reported that platelets contain GM3 ganglioside (223,224) but structural studies were not included in these earlier investigations. On the basis of molar ratios of sugars and permethylation studies, the structure of platelet GM3 ganglioside was found to be NANA—(2+3)-Gal-(l+4)- Glc-ceramide, which is exactly identical to the human plasma hematoside in structure. The ratios of N-acetylneuraminic acid to glucose were somewhat lower than the theoretical value in both human plasma and platelet GM3' Low recoveries of neuraminic acid are not uncommon (105), however, and may be the result of incomplete N-acetylation, partial acid destruction during methanolysis, overexposure to pyridine during the preparation of the tri- methylsilyl derivatives (279), and loss during silicic acid chromatography (280) or gas-liquid chromatography (281). It is possible to obtain better relative values for neuraminic acid by using 0.5 N methanolic HCl during methanolysis and 248 exposing to pyridine in the trimethylsilylation reaction for a short period of time, since the neuraminic acid values decrease slowly with time (279). The concentrations of the water-soluble gangliosides were not determined, and hence one cannot comment on the relative importance of GM compared with other platelet gangliosides. 3 On the basis of a GM3 partition ratio of 0.4 between the upper and lower phases of a partition system (282), it was calculated that the total platelet concentration of GM3 was about 1.9 nmoles/g total lipid, a value that agrees reasonably well with the relative yields of G given by Snyder gt 31. M3 (224). The finding that G is the major ganglioside along M3 with some hexosamine-containing gangliosides in both human platelets and plasma has been confirmed recently by Marcus gt El' (283) and Yip and Ledeen (284). However, individual ganglioside structures have not yet been determined, and the presence of glucosamine indicates that some of the structures must have different oligosaccharide moieties than those char- acteristic of brain. Previous studies had demonstrated that the relatively non-polar neuraminic acid-containing glycolipids were the major representatives of the extraneural tissues (264). Hence, it is generally believed that the pattern of gangliosides in extraneural tissues is appreciably simpler than in the brain. However, recent studies by Puro (264) demonstrated the presence of several slow-moving gangliosides in a variety of 249 tissues. In pig platelets (278), 8 bands were detected in the ganglioside fraction by TLC. This finding, coupled with the detection of slow-moving minor gangliosides in both human plasma and platelets, further extend the observation of Puro from extraneural tissues to body fluids and its component. Thus, the ganglioside patterns in the extraneural tissues and fluid are at least as complex as those in the brain. The fatty acid distributions of both human plasma and platelet GM are different from the results reported for 3 brain gangliosides in man (285-288) and vertebrates (288) wherein stearic acid comprised 80-90% of the total fatty acids. In the plasma and platelet GM3 ganglioside from lower 20-C24 represented 68-79% of the total mixture, which is in accordance with phase the longer-chain fatty acids from C some of the reported studies on extraneural gangliosides (113,289,290). In the upper phase, platelet gangliosides were shown to contain principally 16:0, 18:0, and 18:1, with the exception of ganglioside A which had more of 22:0. The physiological significance of gangliosides is largely unknown. Inherited diseases with a storage of gangliosides in different tissues, especially in brain, are termed ganglio- sidoses (67). There have been only two cases found with an accumulation of ganglioside G (52,291). There are many M3 facts, however, supporting the view that gangliosides play a dynamic role in the function of the cell membrane and thus are 250 not merely structural lipids (292). As for G it has been M3’ suggested that this ganglioside may have an immunologic role in transformed cells (293); it may also be involved as a receptor site on platelet membranes for serotonin (294,295), and in platelet aggregation (224). As it was stated earlier, a comparison was made on a smaller scale of untreated platelets and platelets treated with several proteolytic enzymes to determine whether the composition of the platelet sphingolipids was changed by pre- incubation with these enzymes. The yield of hematoside was substantially higher when the cells were trypsinized and was higher still after incubation with thrombin, but no other measurable changes were observed. It has been reported that thrombin (and also trypsin) produces striking stimulation of phosphatidylserine formation in platelets (296). When trypsin was preincubated with soybean trypsin inhibitor, the effect of trypsin was abolished, suggesting that the proteo— lytic action of trypsin may be required to produce the effect of phosphatidylserine formation, since trypsin inhibitor alone exerted no effect on phosphatidylserine formation (296). The mechanism(s) of the thrombin effect on the stimulation of both phosphatidylserine and G ganglioside and its physio- M3 logical significance are not clear at the present time. Thrombin is a proteolytic enzyme which not only clots fibrinogen but also causes platelets to aggregate and undergo "viscous metamorphosis" (296). Trypsin has also been shown 251 to aggregate washed platelets (297) in a manner similar to that observed with thrombin. Both trypsin and chymotrypsin were obtained from Calbiochem and were classified as Grade A purity. Thrombin was from Parke, Davis and Co., and was presumably also of good quality. Nevertheless, the possibil- ity of having some other substance(s) present in the thrombin preparation has not been completely ruled out and further studies are required to clarify this point. Thrombin affects platelet metabolism in a number of ways. It causes an increase in ATPase activity and lactate production (298), and a decrease in the intracellular ADP and ATP levels (298-300). It is possible that the increased G formation may be an indirect effect of the changes of M3 intracellular sugar nucleotide levels by the action of thrombin on the platelet membrane. Indirectly, the proteolytic enzymes may act on the plate- let surface, splitting a target protein and producing an active peptide that might act as a messenger to stimulate the release reaction (301). During the release reaction, acid hydrolases are believed to be released in soluble form (302). If this is true, the possibility exists for the stepwise removal of sugar units from more complex gangliosides (such as GMZ' 6M1) by these hydrolases. However, examination of the gangliosides derived from the Folch upper phase by TLC did not reveal any discernable changes between control plate- lets and those treated with proteolytic enZymes. 252 Another possible explanation is that thrombin causes some conformational changes on the platelet membrane surface possibly by proteolytic action which might in turn alter the spatial configuration of enzymes, such as sialyltransferase, that will actively transfer NANA from some other macromole- cules (such as the glycoproteins) to the receptor GL-2a, the major glycosphingolipid in platelets. It has been reported (303) that washed human platelets incubated with [14C]CMP-NANA in the presence of homogenized rat liver as the source of sialyltransferase showed an increase in the amount of sialic acid bound to the platelet membrane. Furthermore, an increase in sialic acid bound to the Platelet + [14C]CPM-NANA Slallltransr‘erase: [PlateletW sialic acid + CMP surface favors the aggregation of serotonin on platelets. Sialic acid has been implicated as a component of the receptor complex for serotonin on platelet membrane as well as on smooth muscle of the guinea-pig intestine (304). It has been suggested that GM3 ganglioside may be involved at the receptor site for serotonin on the platelet membrane (294, 295). It is also possible that the burst of G synthesis is M3 involved in blood coagulation, since thrombin is an active participant in the clotting mechanism. It is of interest to note that phosphatidylserine thus far seems to be the best 253 replacement for platelet membranes in coagulation (305), although it is not as effective as the platelet membranes themselves (306). Nevertheless, it is uncertain whether or not GM3 ganglioside can replace the platelet membranes in the coagulation process and further study is required to evaluate this point. An interesting fact, although not related to the throm- bin effect, is that injection of bacterial neuraminidase intravenously causes thrombocytopenia in the rats (266). It was concluded that the neuraminidase cleaved the sialic acid from the platelet membrane, altering and deforming the sur- face characteristics and thus provoking removal of the damaged platelet from the circulation by the reticuloendothelial sys- tem. This seems to imply that the sialic acid-containing component plays an important role in the structural integrity of the platelet membrane. Another interesting aspect about neuraminidase is that it inhibits the release of histamine by rat mast cells in the IgE antibody-mediated reaction. In this reaction, the IgE antibody binds to the receptor site on the surface of the mast cell which in turn activates the release of histamine (307). Preliminary evidence (307) sug- gests that neuraminidase acts on the receptor site for IgE rather than on substrates involved in the reaction sequence (307). If this is true, then the receptor site would seem to contain sialic acid residue(s), probably a ganglioside as part of this receptor complex on the membranes of both 254 platelets and mast cells and other cells for that matter. In this regard, there seems to be a functional role for the presence of this ganglioside in a variety of extraneural tissues. Platelet is the only formed element of blood which exhibits the characteristics of accumulating serotonin (308) and dopamine (309). The mechanism of the release and trans- port of these amines to target organs or tissues in the blood stream is not clear at present. Platelets may be unique among mammalian cells in the high proportion of total sphingolipids accounted for by free cera- mide, which confirms recent reports by Krivit and Hammarstrdm (225) for human platelets and by Heckers and Stoffel (278) for pig platelets. It is clear from a variety of studies that ceramides are common constituents of animal tissues, including erythrocytes (9), plasma (l3), aorta (14), liver (5), spleen (6), lung (7), kidney (10), and brain (15). In none of these cases is the concentration of ceramide as high as it is in platelets, however. It is possible that the ceramide fraction represents accumulation from glycosphingolipid turn- over especially since significant activities of various glyco- syl hydrolases have been reported in platelets (224). The fatty acid composition of the various sphingolipid fractions does not support this theory, however. The ceramides con- tained significantly more 16:0 and 18:0 than any of the glyco- sphingolipids, while the level of 24:1 was much lower. The ceramide fraction also differed considerably in fatty acid 255 composition from that of ceramides (13) and other sphingo- lipids of human plasma, and was confirmed by Krivit and Hammarstrém (225,310). These results and the high concentra- tion of the ceramide fraction suggest that its role in plate- lets may not be limited to that of an intermediate in the biosynthesis and degradation of more complex Sphingolipids. The yields of total phospholipids (65%) and neutral lipids (25%) in platelets and their composition were not remarkably different from the results reported by Marcus, Ullman and Safier (226). Phosphatidylcholine was the major phospholipid and free cholesterol accounted for most of the neutral lipid fraction. The proportion of sphingomyelin was approximately equal to that reported for erythrocytes and was considerably higher than that of normal lymphocytes and polymorphonuclear leukocytes (311). The fatty acid composi- tion of the various phospholipids, triglycerides and choles- terol esters were in close agreement with the results obtained by Marcus g£_al. (226) including the interesting fact that the cholesterol ester fraction is practically devoid of 18:2. It was also noted that the phosphatidylethanolamine fraction contained about 20% of plasmalogen form, consisting primarily of molecular species containing 16:0 and 18:0 alkenyl groups, whereas the other phOSpholipids contained very much lower levels of plasmalogens. 256 B. Porcine Platelet GlyCOSphingolipids Analyses of porcine platelets revealed that GL-3a was the major neutral glycosphingolipid present, which confirmed recent studies of Heckers and Stoffel (278). In addition, GL-la, GL-Za, GL-4, and GM3 were also found. The presence of GL-4 in my samples may have been due to a small amount of erythrocytes in the platelet preparation, since this lipid was not found by Heckers and Stoffel. An interesting observa- tion was that an extra band was detected by TLC, which migrated between GL-2a and GL-3a and had a RF value similar to that of sulfatide. Heckers and Stoffel reported that sulfatide is one of the acidic glycoshpingolipids in porcine platelets. They also found that the major porcine platelet ganglioside consisted of a double band on TLC; one band contained a lacto- sylceramide residue and the other a diglucosylceramide residue which had never been found before. N—Glycolylneuraminic ac1d was found to be present in these lipids. GD3 was not detected in this study, because it is present exclusively in the Folch upper phase and was not analyzed in this case. The chromatographic behavior and identification of the ceramides confirmed the findings by Heckers and Stoffel for porcine platelets. The fact that human and porcine platelets are both rich in ceramides further substantiate the importance of this class of lipid in platelet physiology. It is interesting to note that GL-Za is the major mem- brane-bound glycosphingolipid in both human leukocytes and 257 platelets, whereas GL-3a is the major glycolipid of porcine leukocytes and platelets. Thus, there appears to a degree of species specificity in the composition of neutral glyco- sphingolipids of these cells which is not observed in the erythrocytes. C. Fetal Erythrocyte GL-4 The fetal erythrocyte has been shown to possess a number of unique structural and metabolic properties (312). Some of these are related to cell age whereas others are unique to the fetal cell. Decreased survival of the fetal erythrocyte, to approximately 80 days, has been clearly demonstrated by many techniques (313). These findings may have important implications in certain neonatal diseases. Many possibilities have been proposed for the decreased survival of fetal erythrocytes (312), and membrane lipid al- teration is one of them, since lipids play an important role in the erythrocyte membrane function. As part of the inter- est in blood glycosphingolipid metabolism, an attempt was made to study the major glycosphingolipid composition of fetal red cells during development. In particular, the question was posed whether globoside is part of the erythrocyte membrane structural component during the early embroyonic stages or whether it is incorporated into the membrane structure only at certain stages of the developmental process such as near or at term. 258 It is not practical to obtain blood samples from human fetuses. Although the pig has a gestation period of W 114 days (about 1/2 of humans), as an animal model it seems to provide all the necessary requirements from which a study of this nature can be performed. First of all, the pig also possesses GL-4 as the major erythrocyte glycosphingolipid. Secondly, the pig fetus is large enough in size so that ade- quate amounts of blood can be removed for analysis. And finally, in contrast to many other animals, the placenta of the pig is a rather primitive one (314). The chorion and the urterine mucosa lie in close contact with each other but are not fused, so they can easily be peeled apart. The maternal and fetal circulation are always separated. No definite conclusions can be drawn from the analyses of erythrocyte GL-4 of the 45- and 90-day fetuses. However, the low GL-4 concentration found in the 45-day fetuses is of interest and may be significant when one considers the contami- nation of mother's blood in the sample. Hence, in actuality, the values may be lower than what was observed or GL—4 may actually be totally absent. If blood-sampling had been de- layed until perhaps Day 52 or 55 (of gestation) it may have been possible to withdraw blood from the vessels (as that of the 90-day fetuses) without any difficulties'and contamination. 259 D. Metabolic Study in Pig 123-6 The second half of this thesis is concerned with metab- olism of neutral and acidic glycosphingolipids in an anemic pig. The primary objective was to study the metabolism of globoside as a function of erythrocyte senescence, and to determine the nature of interrelationships between plasma and red cell pools of glycolipids. One approach to study the turnover of globoside in red cell fractions involved separa- tion of red cells by density gradient ultracentrifugation. Induction of anemia in pig 123-6 was accomplished by daily bleeding. The data from the hematological profiles indicated that the pig responded well to the stress of bleed- ing with an active marrow production. The reticulocyte. percentage was used to define the response of the anemic pig, and the hemolytic state was identified on the basis of a high circulating reticulocyte count. No reports were available for comparison on the induction of reticulocytosis in a pig with such a degree and for such a long period of time. Separation of reticulocytes from the mature erythrocytes was achieved by density gradient ultracentrifugation. A dis- continuous gradient offers the advantage over a continuous one is that larger quantities of red cells can be centrifuged. HowevFr, even with the degree of resolution obtained by dens- ity gradient ultracentrifugation, pure fractions containing 100% of cells of a given age cannot be obtained from normal 260 animal blood (259). In the discontinuous system used in the present study, it was estimated (259) that 95% 0f the reticu- locytes were present in the tOp-most layer; the remaining 5% were found in the second layer, and only occasional reticulo- cytes were seen in any other layer. Hence, it is to be noted that the peripheral blood used in this turnover study was not a homogenous mixture, whereas in the density gradient experi- ment, individually fractionated groups of cells were analyzed. Two types of red cell glycosphingolipid turnover curves were obtained from this in_yiyg study, one of which seemed to follow the normal red cell survival (GL-3a and GL-4), while the others did not (GL-2a, GL—la, GM3’ upper phase ganglio- sides, and ceramides). The results from the incorporation of the labeled glucose into both plasma GL-3a and GL-4 (Figures 26 and 29) confirm the previous observation (45) that there is an early synthesis of these two lipids, probably in the liver, which exhibits peaks of incorporation at 24 hr. This agrees well with stud- ies of 32P incorporation into canine plasma lecithin (315). The loss of specific activity subsequent to maximum incorpora- tion was rapid in both lipids, presumably reflecting a loss of the labeled lipid from plasma to other tissues, as well as their catabolism and elimination from the animal. The fact that the specific activities of both glycosphino- lipids remained rather constant between Days 20 and 40 rules out any possibility of exchange between these lipid pools. 261 The reappearance of label in GL-3a and GL-4 began around Day 50, at a time when the specific activities of erythrocyte GL-3a and GL-4 were gradually declining. The plasma glyco- sphingolipids reached maximum relabeling around 60 days, at which time a precipitous decline of specific activities occurred in both red cell GL-3a and GL-4 fractions, indicat- ing the possibility of a loss of red cell glycosphingolipid into the plasma without exchange. The plasma glycosphingo- lipids were relabeled around 60 days, which is the average reported life span of porcine erythrocytes (273). The 14C-labeling of the erythrocyte GL-3a and GL-4 shows a consistent pattern (Figures 26 and 27). Because of the similarities between these two lipids during metabolic turn- over, they will be discussed together. Emphasis, however, will be placed on GL-4, and it can be said that the metabolic fate of GL-3a is very similar to GL-4. Both GL-3a and GL-4 decay curves are essentially composed of three identifiable pools of each lipid. The Specific activity of each red cell glycosphingolipid reaches a maximum peak at Day 3, which coincides with the approximate time of entrance of erythrocytes from the bone marrow into the peripheral circulation when pulse labels of radioactive iron and glycine are utilized (272,273). The data from the in zitrg study did not reveal any evidence of GL-3a and GL-4 biosynthesis by the reticulocyte-rich cells; if any, it was insignificant. This was further confirmed by the 262 absence of any early peak of incorporation in the erythro— cyte glycosphingolipid fractions before Day 3 from the in 2322' study. It seems evident, therefore, that most if not all of the erythrocyte GL-3a and GL—4 is synthesized in the bone marrow, and that synthesis in the reticulated circulating red cells is qualitatively of minor significance. It was previously observed (45) that approximately 60% of the label in red cell GL-3a (subsequent to maximum incorpora— tion at Day 5) was lost by Day 9, while there was no such loss in the GL-4 fraction. In the present experiment with the anemic pig, such losses were demonstrated to occur in both glycosphingolipids. If one compares the turnover curves of GL-4 between the normal and anemic pig, it can be seen that the part of the curves between Days 30 to 81 are very similar indeed in both animals. The only difference which has defi— nitely been established is the presence of the initial peak in the anemic pig, and such initial incorporation merely indicates the rate of uptake by new red cells of the radioactive glucose utilized in the biosynthesis of this membrane glycolipid. It has been demonstrated that incorporation of 59Fe and [lsN] glycine into erythrocytes was faster in the anemic pig ( ) and rabbit ( ) than the normals. It is quite clear, then, that the difference observed in the slope and extent of GL—4 and GL-3a incorporation between the normal and anemic pig presumably reflects the uptake of label by young cell frac— tion(s) present in the marrow, most likely the precursors of reticulocytes. 263 The precipitous decrease of Specific activities between Day 60 and Day 81 in red cell GL-3a and GL-4 is unique but not uncommon. Previous attempts using radioisotopes of [2-14C]glycine, 59Fe and [15N]glycine in red cell survival studies of normal and anemic pig (272,316), human (317), dog (318), cat (319), and rabbit (320) has demonstrated such phenomena in the survival curves, and it was implicated to be the result of an age-dependent process, loss of label due to cell destruction. Had this in_yiyg_experiment been con- tinued beyond 81 days, it may have been possible to see a plateau after this abrupt decline, so that a sigmoidal (or S) shaped curve may be observed for this late period. Several possibilities may be advanced to account for the loss of label observed subsequent to Day 3. Since the daily removal of blood from the pig was still enforced during the ‘ first 10 days, it is possible that the loss of specific activ- ity may be associated with the amount of radioactivity removed by sampling. Table 32 shows that approximately 10-13% of the circulating blood was removed daily. From the packed cell volume of the blood samples obtained each day, an average of 11% can be calculated from the removal of red cells from the total pool. The percentage change in Specific activities of both mixed population (cpm/ml) and fractionated red cells (cpm total hexoses/ml) varied considerably. Day to day changes were not constant. Hence, it seems that the decline of label is attributed to some other process or processes rather than .mUCmn wouoCOHuomum no 000 HE Mom 4IH0 CH momoxon Houou mo ouCCHE Mom muCCoqumbHuHom UHMHoomm H.4.0n .HmHHoo 0oxHEV omm HE mom 4IH0 mo ouCCHE mom quCooI0UH>Huom UHMHoomm H.4.0o 264 H00 HH 00H 05HH H00 000 0H 00I 000 0H 00H 05HH H00 000 0 VI. 04H H00 0H 00H 00HH 000 040 0 40I 00I 000 HOH 005 0H 00H 00HH 000 040 5 00I 5HI 00I 040 000 505 HH 00H 000H H00 HHO 0 04I 041 00I H00 5H4 000 HH 40H 000H H00 HHO 0 H0I 00I 0HI 000H 0H0 050H HH 00H 000H 050 005 4 0 I 40I 5HI 044H 405H 000H 0H 00H H00 405 0 40I 00I 000H 4H00 0H 0HH 000 000 045 0 .4. .0. 0 0Com H 0Com .00 mHHoU .0. .HE. .HEV .HE. AHEV omConO 0 0Com omConu H 0Com omcmnu notz noboaom ©o>oEom oECHo> oECHo> oECHo> mama .4.m n.<.m .<.m n.4.m .m.m 6.4.m voon voon nooflm Hmuoe omm. mammHm 0H moo 0C4 0 moo Coo3uom 0I00H 0H0 Eoum ©o>oEom noon mo omouCoouom .00 oHnoe 265 the removal of a constant fraction of blood cells present. However, the possibility exists that the observed decrease in specific activity can be a combination of two different processes. Since approximately 12% of the blood was removed daily (Table 32) during the early period, it seems that the amount of erythrocyte GL-4 radioactivity removed by sampling may be significant with respect to the total circulating red cell GL-4 radioactivity and this loss might affect the calculated t l/2 values seriously. An attempt was made to correct for this. The correction used was that suggested by Brown and Eadie (318). If a is the total circulating erythrocyte GL-4 l radioactivity on Day 3 and if b is the GL-4 radioactivity l removed in a sample on that day, then the total circulating GL-4 radioactivity at time of sample on Day 4, a2, must be divided by Similarly the total circulating red cell GL—4 radioactivity shown on Day 5 must be divided by (al-bl)(a2—b2) alaz etc. After correction, a half-time value of 5.0 days was obtained from the semilogarithmic plot of specific activity versus time. 266 An attempt was also made to take into consideration the replacement rate of red cell mass from the bone marrow. Normally, the replacing rate of red cells is 1/60 to its total red cell mass, since the life span of the porcine erth— rocytes is approximately 60 days, whereas the plasma volume remains rather constant. In acute bleeding, the rate of re- placement is believed to be around 8—10 times higher than normal in humans. No information is available about the replacement rate in pigs; however, if one assumes that an estimate of 10 times normal is also applicable to the pig, then the maximum rate of replacement in the pig is 1/6 of the red cell mass per day. On this basis, after correction, the semilogarithmic plot of specific activity versus time revealed a half-time value of 7.0 days for the decline of radioactivity between Days 3—10. The slope of the decay curve is less steep, but not to a great extent. Another explanation is dilution of the existing radio- active red cell GL-4 by red cells containing GL-4 of low isotOpe content. Since the pig is still in the steady state of hematopoiesis, the hemoglobin, hematocrit and blood volume 267 was rather constant as expected (Figure 27). If the anemic condition were moderate (moderate dilution) one would expect to see a curve showing a gradual decay whereas, if the anemic condition were severe with lots of incoming cells from the marrow, then the decay would be more steep. In either case, the decay curves should not contain any breaks or plateau in the specific activity versus time plot. This was not observed, however. A third, and more probable, explanation is loss of globosidemcontaining membrane fragments from the reticulocytes as they mature in the circulation. Red cells are known to lose membrane lipids as they age in yizg (321,322). This change occurs primarily during the first part of the cell's life span (321,258), perhaps due to regiculocyte maturation (323). Although reticulocytes lack a nucleus and contain little, if any, rough or smooth endoplasmic reticulum, they possess mitochondria and free ribosomes (324). Glycolipids are not 268 found in mitochondria or membranes of the endoplasmic reticu— lum and they have not been found in the nucleus (325,326). Rather, this important class of lipids seems to be unique to the plasma membrane. The fact that globoside is confined to the erythrocyte plasma membrane rests on the demonstration that the lipid can be removed in its entirety in ghosts pre- pared from erythrocytes (8,68). It has been established that under an anemic condition, premature delivery of marrow reticulocytes to the blood stream occurs in rat (327). These immature reticulocytes are termed macroreticulocytes; their presence was detected in pig 123-6 as expected. Recent studies (327,328) have investi— gated the maturation of these cells from animals during acute anemia. It was found that macroreticulocytes undergo normal i2_yizg maturation without an appreciable reduction in life span. The cell undergoes remolding within the circulation, involving loss of membrane and cell contents until it approaches approximately the size of the normal adult red cell. One-third of the plasma membrane lipids (cholesterol and phos- pholipids) was estimated to be lost during maturation due to factors operative in yiyg (327). Spleen is one of these factors (327). The lipids were lost largely during the first two days of reticulocyte maturation and the process was com- pleted by Day 5. Although red cells at this point were "mature" as defined by the reticulocyte stain, loss of membrane lipids and cell contents continued over the remaining 6 days. 269 It is unclear how long this process normally continued; however, from the extrapolation of the data, it was suggested that normal values would have been achieved by Days 20 to 25 (327). In this respect, if one assumes that the life span of macroreticulocytes or reticulocytes is normal or near normal, as demonstrated by Ganzoni et al. (328) and Shattil and Cooper (327), then one can envisage the early decline of Specific activity in red cell GL-4 subsequent to maximum in- corporation as a result of loss of membrane components due to maturation. In a given situation, with a group of young cells containing GL-4 which are maximally labeled at Day 3, assume that it takes 11 days for these cells to mature during which time the globoside-containing membrane fragments and hence the radioactive label are being lost. Other cellular contents, such as mitochondria, etc. are being lost also but, since they do not contain any glycolipids, will not be con— sidered here. Therefore, by Day 14, the majority of the young cells would have reached maturation and the specific activity should begin to be constant. If one also assumes that the maturation process will continue on until Day 25 as suggested by Shattil and Cooper (327), then all the cells will be normal adults and thus resuming a normal age-dependent process. Then, one would expect the specific activity of the red cell GL-4 to remain rather constant after Day 25, until such time when cell destruction occurs and the specific 270 activity of GL-4 will take a rather sharp drop. If this were the case, one would expect to see a turnover curve of red cell GL-4 quite similar to that shown in Figure 28 and 30. The slight decrease in Specific activity observed between Days 25 and 60 is probably due to random destruction of red cells, since a perfectly normal youthful cell may meet with a fatal accident while speeding around the circulation. Random destruc- tion of pig red cells has been implicated previously in sur- vival study (273). Half-times of 0.75 and 1.0 day were obtained from the GL-4 decay curves of Band 1 and Band 2 (Figure 41) during the period subsequent to maximum labeling. These data suggest that the reticulocyte GL-4 pools have a very short half-time and are turning over very rapidly. One may relate this loss of labeled GL-4 as evidence of a short-lived population. However, it should be realized that the detection of labeled GL-4 loss cannot be equated with cell death, since it may be due to loss of a portion of red cell component without actual cell destruction. If there were cell destruction one should observe reappearance of the GL-4 label in the plasma fraction at a time subsequent to the rapid loss of specific activity from red cell GL-4. No significant peak of relabeling was observed for either GL-3a pr GL—4. Furthermore, Ganzoni gt_§$, '(328) could not demonstrate any significant shortening of rat blood cell life span by repeated red cell mass measurements or radioautography techniques. Hence, some other mechanism must 271 be involved to account for the loss of cell membrane and reduc- tion in size without causing a loss of individual red cell. Another interesting point is that Figure 40 shows quite clearly that as the specific activity of GL-4 is being lost from the reticulocyte fraction, there is a concomitant in- crease in the label of GL-4 of more mature red cells. This is quite evident between Band 2 and Band 3. In other words, the labeled GL—4 is progressing downward through the gradient with red cell aging, and hence the cells are not dying, but instead, maturing. Figure 42 shows plots of nmoles hexose/g hemoglobin for GL-4 versus time on three of the individually fractionated red cell bands during the first eight days post-injection of the pulse label. As the specific activity of GL-4 declines (Table 30) the amount of GL-4 per gram hemoglobin also de- creases, while the hemoglobin content and packed cell volume remain constant throughout this period. These data most strongly indicate that the loss of radioactivity from GL-4 of the young cells is most likely associated with plasma membrane loss due to maturation. However, it is to be noted that the percentage loss of GL-4 in early days is greater than that pre- dicted for the reticulogyte maturation. Such loss perhaps represents a combination of processes such as cell maturation as well as random cell destruction and removal of cells by sampling. There were three distinct pools of erythrocyte GL-3a and GL-4 as shown by the semilogarithmic plots of specific activity Figure 42. 272 Loss of labeled GL—4 from individual bands as a function of time. Plots of amount of globoside per g hemoglobin as a function of time during the first 8-day period post-injection of the pulse label. pMOLE HEXOSE/S Hb (N— N (N— A A l y. 1 ' I I 2 4 6 8 TIME(DAY-S) Figure 42 274 versus time (Figures 30 and 28). It is possible even more than three pools may exist, if one considers the slepes around the breaks of the curves. The presence of more than one pool of a given molecule, whether simple or complex, in a bio- logical system is not uncommon. The turnover of human plasma cholesterol has been proposed to involve a two-pool system (329). From their studies of the turnover of myelin phospho- lipids in the adult and developing rat brain, Jungalwala and Dawson (330) presented supportive evidence that two pools also exist in myelin. Turnover studies of the glycoproteins in adult and developing rat brain (331) have also revealed the presence of at least two glycopeptide components with differing half-times. And recently, in studies on the char- acterization and metabolism of glycosphingolipids in normal human skin fibroblasts (332,222) and embryonic mouse fibro- blasts (334) evidence was presented for two pools of glyco- sphingolipids. The results in this study suggest that red cell GL-3a and GL-4 each contain two major components which differ in turnover rate. The pool with a half-time of 5—6 days accounts for the more rapid decline which occurs between Day 3 and Day 10. It is to be noted that the apparent half- times obtained are probably maximum values, since the specific activities during this early period were derived from blood samples of heterogenous populations. The true half-time for the younger cell fractions is more likely to be less, as shown in Figure 41. The pool with a half-time of 45 days accounts for the decline in label between Day 25 and Day 57. 275 This slower half-time is an additional indication that GL-4 and GL-3a are components of red cell plasma membrane or components of structural units as classified by Burton (179). In addition, a third pool was evident in both lipids, with a half-time of approximately 6 days, which accounts for the decline in Specific activities between Days 60 and 81. The Slope of the decay for GL-4 is very similar to that ob- served for the early period. This suggests that the pools of red cell GL-3a and GL-4 which turn over at the time of red cell senescence have a very similar t 1/2 to that observed earlier. Or one may speculate that the glycosphingolipids metabolized at this late period may have been derived from the corresponding red cell glycolipid pools from the early period which have gone through the maturation process and finally are being released from the membranes into the plasma before phagocytosis of the cells begins. The finding of several GL-4 pools in the porcine erythro- cytes may be correlated with the reports of Hanahan gt_al. (335) and Coles and Foote (277) that there are two forms of globoside. Both have the same basic structure of GalNAc-Gal- Gal-Glc-ceramide; the only difference is that one contains 2-hydroxy fatty acids while the other contains normal straight- chain fatty acids (335). 2-Hydroxy fatty acids were detected in other porcine erythrocyte glycosphingolipids too (277), but they are most abundant in red cell GL-4. The similarity in the 2-hydroxy fatty acid compositions of porcine erythro- cyte GL-3a and GL-4 further suggests a close metabolic 276 relationship between these two lipids. In the present study, no attempts were made to separate the two forms of globoside or GL-3a on the basis of their fatty acid differences. However, these components may account for the different turn— over rates observed in the in yiyg study. It is possible that these two forms of globoside are synthesized from two differ— ent fatty acid pools at different ”sites“ or fpockets' of the erythrocyte membrane with different rates; one may reside at the surface membrane, while the other is imbedded in the membrane. It is also possible that these two forms of globo— side are synthesized from a common pool of fatty acids, some of which are nonselectively hydroxylated. Whether one is a structural component and the other a functional component is uncertain. The turnover curves of plasma GL—la and GL—2a were simi— lar but different from GL-3a and GL-4 in that maximum incorporation was detected much earlier, at 6 hr. Similar to GL—3a and GL-4 these two lipids were also relabeled between Day 50 and Day 70. However, one noticeable difference was the presence of several peaks during this period instead of one. One objective of this in_gizg_study was to study as many samples as possible throughout the prescribed time course, especially during the latter part of the time sequence, so that a better perspective could be obtained with regard to the turnover curves. The curves turned out to be more complex than anticipated. If one were to use the time points chosen 277 in a previous pig experiment (45) the decay curves of plasma GL-la and GL—2a for the anemic pig would look quite similar to those observed in the normal pig, with a single broad peak of relabeling around Day 60. The initial incorporation peak of radioactive precursor into plasma GL-la and GL—2a at 6 hr indicates an early syn- thesis of these lipids, possibly in the liver. Earlier studies had indicated that triglyceride and cholesterol esters were labeled with radioactive linoleate at around 6 hr (336). Nevertheless, another source of these two plasma glycolipids should not be ignored, that is, through the exchange with its corresponding erythrocyte GL-la and GL-2a pools, since the results from the in 31352 study indicated that reticulocyte- rich red cells could synthesize GL—la and GL—2a. Erythrocyte GL-la and GL-2a turned over quite similarly to the red cell GL-3a and GL-4 in several ways. First, both lipids exhibited a peak of incorporation at Day 3, indicating that bone marrow is one of the possible contributors to these two glycosphingolipid pools. This contradicts the previous finding (45) that bone marrow contributes significantly to the red cell GL-2a, GL-3a and GL-4, but not to GL—la. The observation was based on the fact that no increase in specific activity was detected in GL-la at the time when the erythro— cytes were released from the marrow. Another similarity is that the label is rapidly lost and the specific activity was maintained rather constant between Days 10 and 40. 278 Several differences were also noticed. One was that in addition to the maximum specific activity detected at Day 3, both red cell GL—la and GL—2a showed an additional rapid incorporation of radioactive label, reaching maximum Specific activity at 12 hr. In GL-la, the peak at 12 hr was 33% higher than the peak at Day 3; whereas in GL-2a the two peaks were of equal height. The presence of an additional early peak in erythrocyte GL-2a was not detected previously (45), whereas the early synthesis of GL-la was noted and liver was implicated to be the possible contributing source. Erythro- cyte GL—la was also believed to derive from the plasma through rapid exchange (45). However, in View of the findings from the in zitrg study, an additional source of erythrocyte GL-la and GL-Za Should be considered; that is, the synthesis of these lipids by red cells; and thus, indicating the presence of ceramide-B-glucosyltransferase and GL-la-B-galactosyltrans- ferase. These enzymes are probably membrane-bound. There was no evidence that red cell GL-2a acted as a substrate for the biosynthesis of erythrocyte GL-3a. Recently, similar results were obtained when rat erythrocytes were incubated with radioactive glucose (268). Since the plasma GL—la had a higher specific activity than the corresponding erythrocyte GL-la, it is possible that the red cell GL-la labeled under the i§_yi££g_condition originated from the plasma, thus exhibiting a precursor-product relationship. Nevertheless, the in_vivo studies conducted in the normal as well as the 279 anemic pig could not support such a theory. Furthermore, quantitative analyses revealed a significantly higher concen— tration of ceramides in porcine red cells during the study (1.9 nmoles/100 m1). It is difficult to envisage, contrary to time data from the in 21352 study, that red cell GL—la is derived from plasma and then galactosylated to form GL—2a, without taking advantage of the available ceramide pool present in the cells. Another difference observed in the turnover curves of both erythrocyte GL-la and GL-2a from that of GL—3a and GL—4 is the reappearance of label between 50 and 70 days. Relabel— ing in erythrocyte GL-2a was not noted previously. The data suggest that GL-2a as well as GL—la do not follow the expected normal survival as was previously believed (45). Instead, the red cell GL-la and GL-2a are probably in rapid equilibrium with the plasma counterparts. Figure 32 established clearly the nature of exchange between the plasma and erythrocyte GL-la pool Simply by the fact that these two turnover curves were parallel with each other, except for one point at Day 4. This point was omitted from the curve on the grounds that there were no other points on the slopes up to or after this point to make a peak. However, if there were a peak at Day 4 it would coincide rather closely with the corresponding erythro— cyte GL-la peak at Day 3-thus accounting for all peaks. The turnover curve of plasma GL—2a (Figure 31), on the contrary, does not Show an exchange equilibrium as ideal as 280 that of GL—la. Notably, there is no exchange between the red cell and plasma GL-2a at Day 70. Since the results from the ig_yit£g study did not give any clue in this respect, it is rather difficult to ascertain whether or not any exchange occurs. However, on the basis of the similarities in the Specific activity profiles between GL—la and GL—2a and their difference from GL-3a and GL-4, it is tempting to Speculate that exchange between the GL-2a pools may occur. An uncon- firmed report by Krivit and Kern (337) indicated that a Similar type of experiment was performed on a rabbit using a pulse label of [14C]g1ucose. It was found that rabbit erythrocyte GL—2a and GL-la were apparently in rapid equi— librium with the plasma glycosphingolipids whereas a penta- hexosylceramide, the major erythrocyte glycolipid of rabbit (Gal-Gal-GalNAc-Gal-G1c—ceramide), followed the expected red cell survival in the same manner as the porcine red cell GL-4. The accumulating GL-la and GL—3a in Gaucher's and Fabry's disease have been proposed to derive from erythrocyte glyco- sphingolipid metabolism (124). In Gaucher's disease, GL—la was found to be elevated in both plasma and red cells. Since GL-la is exchangeable between these two pools, it is be- lieved that the excessive GL-la present in erythrocyte is derived from plasma where this lipid accumulates due to the absence of glucosylceramide hydrolase to metabolize this catabolic product of complex glycosphingolipids. With respect 281 to the results obtained from the in yit£g_study, one might speculate that red cell may contribute additional GL-la to the existing red cell pool simply by utilizing the adequate amount of ceramides that red cell provides for synthesis. It is interesting to note that red cell ceramide contains a significant proportion of 2-hydroxy fatty acids, whereas plasma ceramide does not. And yet, 2-hydroxy fatty acids occurred in significant quantities in GL-la of the plasma low density lipoproteins. Hence, the possibility exists that the plasma GL-la can be derived from liver, which had been shown to contain GL-la with 2-hydroxy fatty acids. Alterna- tively, plasma GL-la may be derived from GL-2a which origi- nates either from red cell GL—2a through process of exchange or from GL-3a (and GL-4) released into the plasma when the red cells are catabolized. These red cell glycolipids contain high concentrations of 2-hydroxy fatty acids. In Fabry's disease, GL-3a accumulates in plasma but not in the red cells. Hence, it might be concluded from the in yitrg as well as the in_gizg_experiment that red cells do not synthesize GL-3a from GL-2a nor engage in any exchange reaction with the plasma pool of GL-3a. According to the model, red cell GL-3a remains with the cell throughout its life span; from the time of GL-3a being incorporated into the membrane structure of erythrocyte until the time of cell destruction, the level of this lipid Should remain the same throughout any accumulation. But once it is released into the 282 plasma, this lipid cannot be metabolized further due to the absence of GL—3a-cleaving enzyme and hence it accumulates. The red cell pools of GM and upper phase gangliosides 3 were quite similar in parts of their turnover curves. Both lipids were maximally labeled at around Day 3 to Day 5, indicating that bone marrow is at least partially responsible for their synthesis. The gangliosides also showed an addi- tional peak of incorporation at around 6-8 days. This could mean some newly labeled cells are entering the circulation from the marrow, Since the pig is still in an active hemato- poietic state. It could also represent reutilization of the label derived from the catabolism of other glycosphingolipids during the early turnover period (Day 3 or before), which had found its way back to the bone marrow via plasma. Significant relabeling occurred during the late period for the gangliosides. There does not appear to be any ex- change process involved, although several plasma peaks were observed in close parallel to those of red cells. Definite evidence for a classical exchange equilibrium between the plasma and erythrocyte pools of gangliosides cannot be estab- lished. It is interesting to note that maximum specific activity of relabeling was reached on Day 41, and the amount of radioactivity present in the other peaks during this late period seemed to decrease in size and height with time, which could mean a slow gradual decrease of specific activity. If so, this would indicate that the red cell upper phase 283 gangliosides were lost to the plasma without exchange, a situation rather similar to that observed in red cell GL-3a and GL-4. It is possible that such turnover represents a combination of processes involving several pools, one in exchange with plasma and one not, such probability is only a Speculation and cannot be ascertained at this time. The turnover curves of red cell ceramides (Figures 35 and 36) also showed that they were labeled within a short time after injection of the initial pulse label, and reached maximum specific activity in about 12 hr. This seems to indi- cate again that liver and perhaps the erythrocyte itself can account for the early synthesis. The fact that an additional peak of uptake was also observed at Day 3 further confirmed that bone marrow is one of the major sources for all of the sphingolipids studied thus far. The extent of synthesis, as judged from the peak size, varied between individual lipids from GL-4 to ceramide. The decay curves of red cell ceramides did not demonstrate that they followed the normal expected survival. There was a trough at around 40 days, suggesting loss of label, and subsequent increase in the Specific activity betweeb Days 50 and 81. It is rather difficult to rationalize such a phenomenon since there is a substantial body of evidence, summarized by Van Deenen and de Gier (338), that mature erythrocytes are incapable of carrying out anabolic lipid processes. No conclusion can be made with regard to the possibility of having an exchange process, since the specific 284 activities of the plasma ceramides were too low to permit any meaningful interpretation. A summary of the results, showing maximum relabeling periods for each lipid, is given below. Table 33. Time of Reappearance of Label in Sphingolipids During the In Vivo Study Red Cell Sphingolipids Days Upper phase gangliosides 40, 50, 60, 66, 81 Lower phase GM3 50, 60, 70, 81 GL-2a 50, 63, 70, 81 GL-la 50, 63, 70, 81 Ceramide-NFA 55, 63, 70, 81 Ceramide-HFA 53, 60, 75 As can be seen, Similarities exist in the pattern of relabel- ing among all of these lipids. The fact that ceramide became relabeled might be an indication of the catabolism of some of the complex glycosphingolipids, such as gangliosides and globoside. gangliosides (upper phase) Gl-4 GM3(lower phase) GL-3a GL-2a GL-la ceramide 285 Lipid loss has been reported during the course of in 3139 aging (321,233) as well as during in £3239 incubations of abnormal cells (339). In these instances, the loss of lipid has been found to involve all of the neutral and phos- pholipid components. The concomitant formation of small membrane fragments of varying Sizes was also observed in some of these studies (340). Perhaps it is possible that these membrane fragments are adsorbed by circulating erythro- cytes, or perhaps they can be incorporated into the membrane. It is more likely that the cells which take up the l4C-labeled tracer slowly will outlive, on the average, cells formed earlier, and will be present to a larger extent in the blood after the end of the major senescence phase, that is, the period of the sharp decrease in specific activity. The results from the previous metabolic study in the normal pig (45) did not seem to support some of the discrep- encies observed from the fatty acid analyses of the plasma and erythrocyte glycosphingolipids. It was postulated that the fatty acid composition of plasma and red cell GL-la should be the same, because they exchange freely, and this was found to be the case and confirmed by the present study. The data from the in yit£g_study indicated that synthesis of GL-2a by the reticulocyte-rich cells can occur in addition to its forma- tion in the bone marrow. Hence, some similarity in the dis- tribution of fatty acids should exist between the red cell GL-la and GL-Za. Previous reports on the analyses of normal 286 fatty acids clearly indicated that these lipids contained primarily 16 and 18 carbon acids, whereas red cell GL-3a and GL-4 contained mostly fatty acids of 20 carbons and higher. The percentage distribution between red cell GL-la and GL-2a is not exactly identical, which is not unexpected for the reason given above and because GL-2a is also involved as an intermediate in the biosynthesis and catabolism of ganglio- sides and it should contain a rather heterogenous population of fatty acids. An interesting observation from the in_yit£9_ study is that the amount of label recovered in the fatty acid fraction seemed to increase with the size of the oligosac- charide chain (Table 18). It is possible that this represents the elongation of the fatty acid chain with the radioactive acetate derived from the matabolism of the labeled glucose, since the reticulocytes are known to possess the normal comple- ments of TCA cycle (341) and the capability of incorporating fatty acids into neutral lipids (342). Similarities in the distribution of 2-hydroxy fatty acids in red cell GL-2a, GL-3a and GL-4 have been demonstrated (277). However, although the biosynthesis of GL-3a and GL—4 have not “shown in_yit£g_by reticulocytes in whole blood, it‘does not rule out the possibility that GL-2a is an anabolic intermediate in the biosynthesis of GL—4. The process of step-wise addition of monosaccharides to GL-2a can very well be the mechanism of synthesis operational in the bone marrow. 287 Since quantitative determinations of the plasma and erythrocyte glycosphingolipids were made, the pool sizes of each lipid as well as the turnover rate and the fractional turnover rate could be calculated (275). Although the turn— over rate for GL-2a differed somewhat from those of GL—3a and GL-4, there was close correlation with the total pool sizes. The proportion of GL—Za, GL-3a, and GL—4 synthesized each day was quite the same; approximately 21% of each plasma pool was newly synthesized each day. This finding was in good agreement with the data obtained from human (222) and pig plasma (45). Approximately 90% of plasma GL—la was syn— thesized each day, which is considerably different from the value of 20% observed previously in the normal pig (45). This rapid turnover of plasma GL-la could be the reflection of the rapid synthesis of GL-la by the reticulocytes in cir- culation. Approximately 75% of plasma GM3 was shown to be metabolized each day, Whether this represents the normal rate or not is uncertain. Since the turnover values of erythrocyte GL-4 from the normal pig are not available for direct com- parison, it is evident from this study that the daily produc- tion of GL-4 is pronounced (90%) in the reticulocyte fraction (Table 30, Band 1) which is in keeping with the hemolytic state of the anemic pig. The amount metabolized each day for the more mature cells (Band 3) was much less. Globoside is the major glyCOSphingolipid in human (222) and porcine (45) erythrocytes, and accounts for about 70% of 288 the total red cell glycolipids. The fact that globoside forms an integral part of the red cell membrane and follows the normal red cell survival is interesting indeed. Integrity of the red cell membrane can be disrupted by fragmentation, indirect enzyme hydrolysis or hemoglobin malfunction. If the cellular degradative hydrolases were involved, GL-4 would not be released into the plasma as a whole molecule at or around senescence. Hence, the possibility of direct intra— vascular disruption must be considered. It is tempting to speculate that the release of GL-4 from the erythrocyte mem- brane may cause an alteration in the protein—lipid configura- tion of the membrane complex in such a manner that the deformed erythrocyte will be recognized and hence removed from the circulation by the reticuloendothelial system, such as the spleen. However, it is evident that one must still understand more about the architecture and physiology of the red cell membrane before one can go directly from lipid analyses to an understanding of the erythrocyte aging process. For example, we still know very little about the functions and involvement of the red cell membrane proteins. Neverthe— less, it is hoped that the results of metabolic experiment presented in this thesis may provide some insight into the function and structural organization of the red cell membrane. Firm conclusions are not yet justified, but as the body of data grows the image brightens. VI . SUMMARY Six fractions of sphingolipids were isolated from washed trypsinized human platelets. The concentrations and chemical structures were studied in detail. On the basis of sugar molar ratios, studies of permethylation products, and the action of stereOSpecific glycosidases three major neutral glyCOSphingolipids were identified, of which GL-2a was the most abundant type. In addition, GL-la and GM3 ganglioside were also found. The structures of human plasma and platelet G were found to be identical. Treatment of platelets with M3 trypsin, chymotrypsin or thrombin increased the yield of G M3 ganglioside as compared with a control, while the level of ceramides was not changed. In contrast to human platelets, porcine platelets con— tained GL-3a as the major neutral glycosphingolipid. Sulfatide and G ganglioside were the acidic glycosphingolipids present M3 in pig platelets. Both human and porcine platelets were found to be rich in ceramides, suggesting that this class of lipid may be unique to platelet physiology. No definite conclusions can be drawn about the level of globoside in fetal red cells of pigs since the blood samples were contaminated. However, the level of this glycosphingo— lipid in the erythrocytes of 45-day old fetuses was low when 289 290 compared to values for new-born and adult pigs. The results from the i2_yit£9_study indicated that reticulocyte-rich red cell fractions can synthesize GL—la and, to some extent, GL-2a as well. Radioactive label was detected in the hexose moiety of plasma GL-la also, sug- gesting either that plasma could synthesize GL-la or that plasma GL-la was derived from the red cell GL-la by active exchange of the two pools. Porcine plasma and erythrocytes contained significant quantities of ceramides and G ganglio- M3 side in addition to GL-la, GL-2a, GL-3a and GL-4. A metabolic experiment to study the turnover of these sphingolipids in an anemic pig was conducted by injecting [14C]glucose into the pig intravenously as a pulse label, and removing aliquots of blood samples for lipid analyses at frequent intervals throughout a period of 81 days. Turnover studies were per- formed on the plasma, a mixed population of red cells, and celIS'vfifixflr had been fractionated into individual groups according to age by density gradient ultracentrifugation. All of the plasma glycosphingolipids became labeled within a relatively Short time after the initial pulse label, and reached maximum Specific activities at 6 or 12 hr. The label was then rapidly lost and remained rather constant be— tween Days 15 and 40. From Day 50 to Day 81, all of the glycosphingolipids in plasma were relabeled. Maximum speci- fic activities for GL-3a and GL-4 were reached at around Day 60, which was the average life Span of porcine erythrocytes. 291 The other plasma glycolipids showed several additional peaks of relabeling besides the one observed on Day 60 during the latent period. The presence of such complex peak is not clear at the moment. Plasma and red cell GL-la behaved essentially the same throughout the duration of the experiment; the two turnover curves paralleled each other and hence it was concluded that plasma and erythrocyte GL-la pools were in rapid equilibrium. This finding was further supported by the data obtained from the in yitrg study. The decay curves of plasma and erythro- cyte GL-2a were not as superimposable as those observed for GL-la, there was a peak of relabeling which could not be accounted for; however, on the basis of the similarities in turnover between GL-la and GL-2a, it was tentatively con— cluded that plasma GL-2a may also be involved in an exchange reaction. This needs to be verified. The rate of incorporation of isotope into all fractions of erythrocyte sphingolipids was faster in the anemic than the normal pig. A peak of incorporation was observed at Day 3 for all of the lipids studied, indicating that bone marrow is one of the prime sources of these lipids. In addition, a contribution by early synthesis was also evident for erythro' cyte GL-la and GL-2a. Possible sources for these syntheses could be the liver or red cell itself. Two types of red cell glycosphingolipid turnover curves were observed; one seemed to follow the normal red cell 292 survival (GL-3a and GL—4) while the others did not (cera- mides, GL-la, GL-2a, lower phase GM3 ganglioside and upper phase gangliosides). With regard to erythrocyte GL-4 (and GL—3a), the results obtained from the turnover studies of mixed cell population, and the fractionated individual groups of cells by density gradient ultracentrifugation, were consistent with the hypothe- sis that the radioactive glucose was rapidly incorporated into the membrane-bound GL-4 of the immature erythrocytes in the bone marrow. After being released from the marrow these cells lost a portion of their globoside-containing membrane as they matured in the peripheral circulation. The remodeling of the cells within the circulation continued until it ap- proached the size of a normal adult cell, after which time the turnover of GL—4 remained rather constant. The membrane- bound GL-4 remained with the cell until the time of red cell senescence, and then was released directly into the circula— tion as a whole unit when cell destruction began. Thus, it became the major source of all four plasma neutral glycosphingo- lipids. Erythrocyte GL-la and GL-2a appeared to be in dynamic equilibrium with plasma glycolipids and did not follow the expected survival. No definite conclusion could be drawn with respect to the other Sphingolipids, Since the results seemed to imply that more than one process was present, one involves exchange and the other without exchange. However, from the 293 similarities in the turnover curves of these lipids, some metabolic interrelationships were evident. This study also revealed that glycosphingolipids are a family of complex and heterogeneous lipids. At least two pools of each glycosphingolipid were present in the plasma and erythrocytes of porcine blood, with rapid and slow turn- over rates. The semilogarithmic plots of specific activity versus time reveals half-times of 5.5, 45.0 and 6.3 days for red cell GL-3a; whereas erythrocyte GL-4 gave half-times of 5.5, 45.0 and 6.2 days, respectively. When the globosides from fractionated red cell bands were examined, biphasic decay curves were also demonstrated. Half-times of 0.75, 1.0 and 4.0 days were obtained from the rapid turnover pools of Bands 1, 2 and 3; whereas 3.3, 5.5 and 9.3 days were obtained from the slow turnover pool of these cell bands. Calculated turnover rates were 65.4, 47.0 and 11.5 nmoles/day for the rapid turnover pools and 14.8, 1.9 and 6.9 nmoles/day for the slower pools of these respec- tive bands. Biphasic decay curves were also observed for the plasma GL-la (t l/2=0.75, 7.5 days) and G (t l/2=O.9, 1.9 days), M3 again indicating the presence of rapid and slow turnover pools. Approximately 93% and 9% of GL-la and 77% and 36% of GM3 were found to metabolize each day. This is in contrast to an average value of 21% found for the plasma GL—2a, GL-3a and GL-4. 10. ll. 12. l3. 14. REFERENCES Thudichum, J. L. W. A Treatise on the Chemical Constitu- tion of the Brain. Bailtiere, Tindall and Cox, London (1884). Carter, H. E., F. J. Gldck, W. P. Norris, and G. R. Phillips. J. Biol. Chem. 170, 285 (1947). Frankel, E., and F. Bielshowsky. Z. Physiol. Chem. 213, 58 (1932). Thudichum, J. L. W. Report of the Medical Officer of. Privy Council and Local Government Board. New Sefies, NO. 7, 134 (1877). Thannhauser, S. J., and E. Frankel. Z. Physiol. Chem. 203, 183 (1931). Tropp, C., and V. Wiedersheim. Z. Physiol. Chem. 222, 39 (1933). Tropp, C. Z. Physiol. Chem. 237, 178 (1935). Klenk, E., and K. Lauenstein. Z. Physiol. Chem. 288, 220 (1951). Shimojo, T., A. Kataura, H. Yamaguchi, and K. Ohno. Sapporo Med. J. 28, 85 (1965). Moser, H. W., A. L. Prensky, H. J. Wolfe, and P. N. Rose- man. Am. J. Med. 41, 869 (1969). Klenk, E., and R. T. C. Huang. Z. Physiol. Chem. 349, 451 (1968). Klenk, E., and R. T. C. Huang. Z. Physiol. Chem. 350, 373 (1969). Samuelsson, K. Scand. J. Clin. Lab. Invest. 21, 371 (1971). Royer, M., and J. L. Foote. Chem. Phys. Lipids Z, 266 (1971). 294 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 295 Akino, T. Tohoku J. Exptl. Med. 98, 87 (1969). Thannhauser, S. J., and G. Schmidt. Physiol. Rev. 26, 275 (1946). “ Sribney, M., and E. P. Kennedy. J. Biol. Chem. 233, 1315 (1958). Cleland, W. W., and E. P. Kennedy. J. Biol. Chem. 235, 45 (1960). Hori, T., O. Itasaka, and H. Inoue. J. Biochem. (Tokyo) 59, 570 (1966). Simon, G., and G. Rouser. Lipids 2, 55 (1967). Dawson, R. M. C., and P. Kemp. Biochem. J. 105, 837 (1967). Dawson, R. M. C., and P. Kemp. Biochem. J. 106, 319 (1968). Thudichum, J. L. W. Report othhe Medical Officer of the Privy Council and Local Government Board. No. 5, 113 (1874). Thierfelder, H. Z. Physiol. Chem. 14, 209 (1890). Carter, H. E., and F. L. Greenwood. J. Biol. Chem. 199, 283 (1952). Halliday, N., H. J. Deuel, L. J. Tragerman, and W. E. Ward. J. Biol. Chem. 132, 171 (1940). Stoffyn, P. J. Am. Oil Chem. Soc. 43, 69 (1966). Shapiro, D., E. S. Rachamann, and T. Sheradsky. J. Am. Oil Chem. Soc. 86, 4472 (1964). Shapiro, D., and H. M. Flowers. J. Amt Chem. Soc. 83, 3327 (1961). Levene, P. A. J. Biol. Chem. 15, 359 (1913). Klenk, E. Z. Physiol. Chem. 145, 244 (1925). Klenk, E. 32. Physiol. Chem. 74, 214 (1928). Karlsson, K. A. Chem. Phys. Lipids 5, 6 (1970). 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 296 Fewster, M. E., and J. F. Mead. J. Neurochem. 15, 1041 (1968) . '— Suzuki, C., A. Makita, and Z. Yoshizawa. Arch. Biochem. BiOphys. 127, 140 (1968). Svennerholm, E., and L. Svennerholm. Nature 198, 688 (1963). Vance, D. E., and C. C. Sweeley. J. Lipid Res. 8, 621 (1967). Miras, C. J., J. D. Mantzos, anui G. M. Levis. Biochem. 8. 88, 782 (1966). Hildebrand, J., P. Stryckmans, and P. Stoffyn. J. Lipid Res. 88, 361 (1971). Kampine, J. P., E. Martensson, R. A. Yankee, and J. N. Kanfer. Lipids 8, 151 (1967). Kean, E. L. J. Lipid Res. 1, 449 (1966). Klenk, E. Z. Physiol. Chem. 267, 128 (1941). Rosenberg, A., and E. Chargaff. J. Biol. Chem. 233, 1323 (1958). Brady, R. 0., J. Kanfer, and D. Shapiro. J. Biol. Chem. 240, 39 (1965). Dawson, G., and C. C. Sweeley. J. Biol. Chem. 245, 410 (1970). Levis, G. M. Lipids 4, 556 (1969). Vance, D. E., W. Kirvit, and C. C. Sweeley. J. Lipid Res. 88, 188 (1969). Sweeley, C. C., and B. Klionsky. J. Biol. Chem. PC 3148 (1963). Sweeley, C. C., P. D. Snyder, and C. E. Griffin. Chem. Phys. Lipids 8, 393 (1970). Li, Y. -T., S. -C. Li, and G. Dawson. Biochem. Biophys. Acta 260, 88 (1972). Martensson, E. Biochim. Biophys. Acta 116, 296 (1966). 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 297 Dawson, G. J. Lipid Res. 88, 207 (1972). Sweeley, C. C., B. Klionsky, W. Krivit, and R. J. Desnick. "Glycolipid Lipidosis: Fabry's Disease.“ In The Metabolic Basis of Inherited Disease. (J. B. Stanbury, J. B. Wyngaarden and D. S. Fredrickson, eds.), McGraw- Hill, New York, 663 (1966). Miyatake, T. Jap. J. Exptl. Med. 88, 35 (1969). Schibanoff, J. M., S. Kamoshita, and J. S. O'Brien. J. Lipid Res. 10, 515 (1969). Gatt, S., and E. R. Berman. J. Neurochem. 88, 43 (1963). Klenk, E., and F. Rennkamp. Z. Physiol. Chem. 273, 253 (1942). Klenk, E., and K. Lauenstein. Z. Physipl. Chem. 295, 164 (1953). Yamakawa, T., N. Kiso, S. Handa, A. Makita, and S. Yokoyama. J. Biochem. (Tokyo) 88, 226 (1962). Svennerholm, E., and L. Svennerholm. Biochim. Biophys. Acta 18, 432 (1963). Wajda, M. Biochem. J. 88, 252 (1965). Makita, A. J. Biochem. (Tokyo) 88, 269 (1964). Makita, A., and T. Yamakawa. J. Biochem. (Tokyo) 88, 365 (1964). Hakomori, S. -I., B. Siddiqui, Y. -T. Li, S.-C. Li, and C. G. Hellerqvist. J. Biol. Chem. 246, 2271 (1971). Clarke, J. T. R., L. S. Wolfe, and A. S. Perlin. J. Biol. Chem. 246, 5563 (1971). Li, Y. -T., and S. -C. Li. J. Biol. Chem. 246, 3769 (1971). Wiegandt, H. "Glycosphingolipids." In Advances in Lipid Research. (R. Paoletti and D. KritchevSky, eds.), AcademiE Press, New York, 249 (1971). Yamakawa, T., and S. Suzuki. J. Biochem. (Tokyo) 88, 393 (1952). 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 298 Yamakawa, T., S. Yokoyama, and N. Kiso. J. Biochem. (Tokyo) 88, 228 (1962). Yamakawa, T., S. Yokoyama, and N. Handa. J. Biochem. (Tokyo) g3, 28 (1963). Yamakawa, T., S. Nishimura, and M. Kamimura. Jap. J. Exptl. Med. 88, 201 (1965). Matsumoto, M. J. Biochem. (Tokyo) 88, 53 (1956). Miyatake, T., S. Handa, and T. Yamakawa. Jap. J. Exptl. Med. 88, 135 (1968). Sandhoff, K., W. Andreae, and H. Jatzkewitz. Life Sci. 1, 283 (1968). Sandhoff, K., W. Andreae, and H. Jatzkewitz. Path. Europ. 8, 278 (1968). Krivit, W., R. J. Desnick, J. Lee, J. Moller, F. Wright, G. C. Sweeley, P. D. Snyder, and H. L. Sharp. Am. J. Med. 88, 763 (1972). Rapport, M. M., H. Schneider, anui L. Graf. Biochim. Biophys. Acta 137, 409 (1967). Laine, R., C. C. Sweeley, Y. -T. Li, A. Kisic, and M. M. Rapport. J. Lipid Res. 13, 519 (1972). Forssman, J. Biochem. Z. 7, 78 (1911). Brunius, F. E. Aktiebolaget Fahlcrantz Bocktrypkeri (Stockholm) (1936). Papirmeister, B., and M. F. Mallette. Arch. Biochem. 88, 94 (1955). Makita, A., C. Suzuki, and Z. Yosizawa. J. Biochem. (Tokyo) 88, 502 (1966). Ando, S., and T. Yamakawa. Chem. ngs. Lipids 8, 91 (1970). Siddiqui, B., and S. -I. Hakomori. J. Biol. Chem. 246, 5766 (1971). Vance, W. R., C. P. Shook, and J. M. McKibbin. Biochemistry 8, 435 (1966). McKibbin, J. M. Biochemistry 8, 679 (1969). 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 299 Eto, T., Y. Ichikawa, K. Nishimura, S. Ando, and T. Yamakawa. J. Biochem. (Tokyo) 88, 205 (1968). Stoffel, W. "Sphingolipids." In Annual Review of Biochem- istry. (E. E. Snell, P. D. Boyer, A. MeiSter, and R. L. Sinsheimer, eds.), vol. 40. Annual Reviews Inc. Palo Alto, 57 (1971). Hakomori, S. -I., and R. W. Jeanloz. J. Biol. Chem. Pc 3607 (1964). Hakomori, S. -I., J. Koscielak, K. J. Block, and R. W. Jeanloz. J. Immunol. 88, 31 (1967). Yang, H. -J., and S. -I. Hakomori. J. Biol. Chem. 24 , 1192 (1971). Hakomori, S. -I., and G. D. Strycharz. Biochemistry (Wash.) 1, 1279 (1968). ‘ Blix, G. Z. Physiol. Chem. 219, 82 (1933). Lloyd, A. D., and K. S. Dodgson. Biochim. BioPhys. Acta 88, 116 (1961). Thannhauser, S. J., J. Felling, and G. Schmidt. J. Biol. Chem. 215, 211 (1955). Stoffyn, P., and A. Stoffyn. Biochim. BiOphys. Acta 28, 218 (1964). Malone, M., and P. Stoffyn. Biochim. Biophys. Acta 88, 218 (1965). Martensson, E. Acta Chem. Scand. 81, 1174 (1963). Martensson, E. Biochim. Biophys. Acta 116, 521 (1966). Klenk, E. Z. Physiol. Chem. 273, 76 (1942). Blix, G. Scand. Arch. Physiol. 89, 46 (1938). Svennerholm, L. Nature 177, 524 (1956). Wiegandt, H., H. W. Bficking. Europ. J. Biochem. 15, 287 (1970). “' Eeg-Olofsson, 0., K. Kristensson, P. Sourander, and L. Svennerholm. Acta Paediat. 88, 546 (1966). 105. 106. 107. 108. 109. 110. 111. 112. 113. 114. 115. 116. 117. 118. 119. 120. 121. 122. 123. 300 Kwiterovich, Jr., P. 0., H. R. Sloan, and D. S. Fredrick- son. J. Lipid Res. 88, 322 (1970). Gallai-Hatchard, J. J., and G. M. Gray. Biochim. Biophys. Acta 116, 532 (1966). Martensson, E., A. Percy, and L. Svennerholm. Acta Paed. Scand. 88, l (1966). Puro, K. Biochim. Bipphys. Acta 187, 401 (1969). Jap, J. Exptl. Med. 191 (1969). Kawanami, J. 88, Kawanami, J. J. Biochem. (Tokyo) 88, 625 (1968). Puro, Biocfl' . Biophys. Acta 187, K., P. Maury, and J. K. Huttunen. 230 (1969). Svennerholm, L. Acta Chem. Scand. 88, 1506 (1965). Feldman, G. L., L. S. Feldman, and G. Rouser. 21 (1966). Lipids‘8, Windler, A. S., and G. L. Feldman. Lipids 8, 167 (1969). Ledeen, R., K. Salsman, and M. Cabrera. Biochemistry. l, 2287 (1968). ' McCluer, R. H. Chem. Phys. Lipids 8, 220 (1970). Wherrett, and B. L. Brown. (1969). J. R., Neurology 88, 489 Kuhn, R., and H. Wiegandt. (1964). Z. Naturforsch. 88, 80 Bernheimer, H., Klin. Wochschr. 8, 258 (1968). Yamakawa, T., and S. Suzuki. J. Biochem. 199 (1951). (T015170) _3_§. Klenk, E., and H. Wolter. (1952). Z. Physiol. Chem. 291, 259 Yamakawa, T., and S. Handa. Jap. J. Exptl. Med. 293 (1964). .3_4. Handa, N., and S. Handa. 331 (1965). Jap. J. Exptl. Med. 88, 124. 125. 126. 127. 128. 129. 130. 131. 132. 133. 134. 135. 136. 137. 138. 139. 140. 141. 301 Sweeley, C. C., and G. Dawson. "Lipids of the Eruthro- cytes." In Red Cell Membrane. (G. A. Jamieson, and T. J. Greenwalt, eds.), J. E. Lippincott, Philadelphia, 172 (1969). Stoffel, W., D. LeKim, and G. Sticht. 348, 1570 (1967). Z. Physiol. Chem. Braun, P. E., and E. E. Snell. J. Biol. Chem. 243, 3775 (1968). Mendershausen, P. B., and C. C. Sweeley. Biochemistry 8, 2633 (1969). Gaver, R. C., and C. C. Sweeley. J. Am. Chem. Soc. 3643 (1966). g. Stoffel, W., D. LeKim, and G. Sticht. 349, 637 (1968). Z. Physiol. Chem. Braun, P. E., P. Morell, and N. S. Radin. J. Biol. Chem. 245, 335 (1970). Morell, P., and N. S. Radin. 342 (1970). J. Biol. Chem. 245, Sribney, M. Biochim. Biophys. Acta 125, 542 (1966). Kopaczyk, K. C., and N. S. Radin. J. Lipid Res. 140 (1965). g. Fujino, Y., and S. Ito. Biochim. Biophys. Acta 152, 627 (1968). Basu, S., A. Schultz, and M. Basu. Fed. Proc. 540 (1969). 29.. Morell, P., and N. S. Radin. Biochemistry 8, 506 (1969). Kanfer, J. N. Lipids 8, 163 (1969). Neskovic, N. M., J. L. Nussbaum, and P. Mandel. Letters 8, 199 (1969). FEBS Brady, R. O. J. Biol. Chem. 237, PC 2416 (1962). Hauser, G. Biochem. Bi0phys. Res. Comm. 88, 502 (1967). Morell, P., and N. s. Radin.’ Fed. Proc. 83, 409 (1970). 142. 143. 144. 145. 146. 147. 148. 149. 150. 151. 152. 153. 154. 155. 156. 157. 158. 159. 160. 302 Hammarstrdm, S. Biochem. Biophys. Res. Comm. 88, 459 (1971). Basu, S. Fed. Proc. 81, 396, 1968. Coles, L., and G. M. Gray. Biochem. Biophys. Res. Comm. 88, 520 (1970). Hildebrand, J., and G. Hauser. J. Biol. Chem. 244, 5170 (1969). Brady, R. O. Metab. Physiol. 8ignificance Lipids. Proc.; Cambridge, England 1963. 95 (1964). Dukes, P. Biochem. Biophys. Res. Comm. 88, 345 (1968) Basu, S., M. Basu, H. Den, and S. Roseman. Fed. Proc. 88, 410 (1970). Basu, M., and S. Basu. Fed. Proc. 88, 1133 (1971). Pardoe, G. I., and G. Uhlenbruck. Med. Lab. Tech. 88, 1 (1971). Marcus, D. M., and E. Cass. Science 164, 553 (1969). Goldberg, I. H. J. Lipid Res. 8, 103 (1961). McKhann, G. M., R. Levy, and W. Ho. Biochem. BiOphyS. Res. Comm. 88, 109 (1965). Balasubramanian, A. S., and B. K. Bachawat. Ind. J. Biochem. 8, 212 (1965). McKhann, G. M., and W. Ho. J. Neurochem. 88, 717 (1967). Cumar, F. A., H. S. Barra, H. J. Maccioni, and R. Caputto. J. Biol. Chem. 243, 3807 (1968). Stoffyn, P., A. Stoffyn, and G. Hauser. J. Lipid Res. 88, 318 (1971). Roseman, S. Chem. Phys. Lipids 8, 270 (1970). Kanfer, J. N., R. S. Backlow, L. Warren, and R. 0. Brady. Biochem. Biophys. Res. Comm. 88, 287 (1963). Dain, J. A., M. Mark, M. C. M. Yip, J. Yiamouyiannis, and Y. Cha. Abstr. 154th Meet. Am. Chem. Soc. (Chicago) p. 69 c (1967). 161. 162. 163. 164. 165. 166. 167. 168. 169. 170. 171. 172. 173. ;174' 175. 176. 177. 178. 303 Yip, M. C. M., and J. A. Dain. Biochem. J. 118, 247 (1970). ' Yip, M. C. M., and J. A. Dain. 206, 252 (1970). Biochim. Biophys. Acta Hildebrand, J., P. Stoffyn, and G. Hauser. J. Neurochem. 81, 403 (1970). DiCesare, J. L., and J. A. Dain. Biochim. Biophys. Acta 231, 385 (1971). Arce, A., H. J. Maccioni, and R. Caputto. Biochem. J. 121, 483 (1971). Yip, M. C. M., and J. A. Dain. Lipids 8, 270 (1969). Handa, S., and R. M. Burton. Lipids 8, 589 (1969). DiCesare, J. L., and J. A. Dain. J. Neurochem. 88, 403 (1972). ' Kaufman, B., S. Basu, and S. Roseman. J. Biol. Chem. 243, 5804 (1968). Cumar, F. A., P. H. Fishman, and R. 0. Brady. J. Biol. Chem. 246, 5075 (1971). Kanfer, J. N. J. Biol. Chem. 240, 609 (1965). Davison, A. N., and N. A. Gregson. (1966). Biochem. J. 38, 915 Pritchard, E. T. J. Neurochem. 88, 13 (1966). Radin, N. S., F. B. Martin, and J. R. Brown. Chem. 224, 499 (1957). J. Biol. Moser, H. W., and M. L. Karnovsky. 8 (1959). J. Biol. Chem. 234, Suzuki, K., and S. R. Korey. (1964). J. Neurochem. 88, 647 Burton, R. M., L. Garcia—Bunuel, M. Golden, and Y. M. Balfour. Biochemistry 8, 580 (1963). Kolodny, J., R. 0. Brady, and B. W. Volk. Biochem. BiOphys. Res. Comm. 81, 526 (1969). 179. 180. 181. 1 812.0 183. 184. 185. 186. 187. 188. 189. 190. 191. 192. 193. 194. 195. 196. 304 Burton, R. M. "The Turnover of Lipids." In Handbook of Neurochemistry. (A. Lajtha, ed.), Vol. 5, Plenum Press, New York, 199 (1971). Suzuki, K., S. E. Poduslo, and W. T. Norton. Biochem. Biophys. Acta 144, 375 (1967). Suzuki, K. J. Neurochem. 81, 209 (1970). Maccioni, H. J., A. Arce, and R. Caputto. Biochem. J. 125, 1131 (1971). Gatt, S. Biochim. BiOphys. Acta 137, 192 (1967). Weinreb, N. J., R. 0. Brady, and A. L. Tappel. Biochim. Biophys. Acta 159, 141 (1968). Sandhoff, K., H. Pilz, and H. Jatzkewitz. Z. Physiol. Chem. 338, 281 (1964). Statter, M., and B. Shapiro. Israel J. Med. Sci. 8, 514 (1965). Frohwein, Y. Z., and S. Gatt. Biochemistry 8, 2775 (1967). Sandhoff, K. FEBS Letters 8, 351 (1969). Sandhoff, K., H. Jatzkewitz, and G. Peters. Naturwissen- schaften 88, 356 (1969). Okada, S., and J. S. O'Brien. Science 165, 698 (1969). O'Brien, J. S., S. Okada, A. Chen, and D. L. Fillerup. New Eng. J. Med. 283, 15 (1970). Morell, P., and P. Braun. J. Lipid Res. 88, 293 (1972). Mapes, C. A., R. L. Anderson, and C. C. Sweeley. FEBS Letters 1, 180 (1970). Brady, R. 0., A. E. Gal, R. M. Bradley, E. Martensson, A. L. Warshaw, and L. Laster. New Eng. J. Med. 276 1163 (1967). Brady, R. 0., A. E. Gal, R. M. Bradley, and E. Martens- son. J. Biol. Chem. 242, 1021 (1967). Mapes, C. A., and C. C. Sweeley. J. Biol. Chem. In review. 197. 198. 199. 200. 201. 202. 203. 204. 205. 206. 207. 208. 209. 210. 211. 212. 213. 305 Gatt, S., and M. M. Rapport. Biochem. J. 101, 680 (1966). Gatt, S., and M. M. Rapport. Biochim. Biophys. Acta 113, 567 (1966). Brady, R. 0., J. N. Kanfer, and R. M. Bradley. J. Biol. Chem. 240, 3766 (1965). Gatt, S., and M. M. Rapport. Israel J. Med. Sci. 8, 624 (1965). Brady, R. O., J. N. Kanfer, and D. Shapiro. J. Biol. Chem. 240, 3943 (1965). Brady, R. 0., J. N. Kanfer, and D. Shapiro. Biochem. Biophys. Res. Comm. 8, 221 (1965). Kampine, J. P., R. 0. Brady, and J. N. Kanfer. Science 155, 86 (1967). Kattlove, H. E., J. C. Williams, E. Gaynor, M. Spivack, R. M. Bradley, and R. 0. Brady. Blood 88, 379 (1969). D. S. "Cerebroside Lipidosis: Gaucher's In The Metabolic Basis of Inherited (J. B. Stanbury, J. B. Wyngaarden, and D. S. 565 (1966). Fredrickson, Disease." Disease. Fredrickson, eds.), McGraw-Hill, New York, Hajra, A. K., D. M. Bowen, Y. Kishimoto, and N. S. Radin. J. Lipid Res. 1, 379 (1966). Gatt, S. J. Biol. Chem. 2 8, PC 3131 (1963). Biochemistry 8, 1692 (1969). Yavin, E., and S. Gatt. Stoffel, W., and G. Sticht. Z. Physiol. Chem. 348, 941 (1967). Stoffel, W. and R. Henning. Z. Physiol. Chem. 349, 1400 (1968). Stoffel, W., D. LeKim, and G. Sticht. 350, 1233 (1969). Z. Physiol. Chem. Stoffel, W., G. Sticht, and D. LeKim. 349, 1745 (1969). Z. Physiol. Chem. Keenan, R. W., and A. Maxam. Biochim. Biophys. Acta 176, 348 (1969). 214. 215. 216. 217. 218. 219. 220. 221. 222. 223. 224. 225. 226. 227. 228. 229. 306 Stoffel, W., and G. Assmann. Z. Physiol. Chem. 351, 1041 (1970). Stoffel, W., G. Assmann, and E. Binczek. Chem. 351, 635 (1970). Z. Physiol. Svennerholm, L. "The Metabolism of Gangliosides in Cerebral Lipidosis'" In Inborn Disorders of Sphingo- lipid Metabolism. (S. M. Aronson, and B. W. Volk, eds.), Pergamon Press, Oxford, 169 (1967). Leibovitz, Z., and S. Gatt. Acta 152, 136 (1968). Biochim. Biophys. Ohman, R. Metabolism of Gangliosides. Ph. D. Thesis, University of G6teborg (1970). L. Svennerholm, J. Neurochem. 88, 613 (1963). Kuhn, R., and H. Wiegandt. Chem._per. 88, 865 (1963). L. "Gangliosides." In Handbook of Neuro- (A. Lajtha, ed.), Vol. 5. Plenum PreSS, 425 (1971). Svennerholm, chemistry. New York, Studies on the Chemistry and Metabolism Ph. D. Thesis, Vance, D. E. of Human Blood Glycosyl Ceramides. University of Pittsburg (1968). J., H. L. Ullman, L. B. Safier, and H. S. Fed. Proc. 88, 315 (1970). Marcus, A. Ballard. Snyder, P. D., R. J. Desnick, and W. Krivit. Biophys. Res. Comm. 88, 1857 (1972). Biochem. Krivit, W., and S. Hammarstrdm. (1972). J. Lipid Res. 88, 525 Marcus, A. J., H. L. Ullman, and L. B. Safier. J. Lipid Res. 88, 108 (1969). Pepper, D. S., and G. A. Jamieson. 3362 (1969). Biochemistry 8, Pepper, D. S., and G. A. Jamieson. 3706 (1970). Biochemistry 8, Snyder, R. A., and R. 0. Brady. 331 (1969). Clin. Chim. Acta 88, 230. 231. 232. 233. 234. 235. 236. 237. 238. 239. 240. 241. 242. 243. 244. 245. 246. 247. 307 Pepper, D. S., and G. A. Jamieson. Nature 219, 1252 (1968). and H. van den Bosch. 128 (1970). P., A. Derksen, 9, Cohen, Invest. J. Clin. Folch, J., M. Lees, and G. H. Sloane-Stanley. Chem. 226, 497 (1957). J. Biol. Abramson, D., and M. Blender. Biochim. Biophys. Acta 88, 117 (1965) . Kanfer, J. N. Methods in Enzymolqu. XIV, Academic Press, New York, "Preparation of Gangliosides." In (J. M. Lowenstein, ed.), Vol. 660 (1969). R. J. Am. Oil Chem. SOC. 621 (1966). Ledeen, 88, Klibansky, C., A. Saifer, N. I. Feldman, L. Schneck, and B. W. Volk. J. Neurochem. ii, 339 (1970). J. Am. Oil Chem. 39, Vioque, E., and R. T. Holman. Soc. 63 (1962). Foote, J. L., and E. Coles. J. Lipid Res. 8, 482 (1968). Wood, R. D., P. K. Raju, and R. Reiser. J. Am. Oil Chem. Soc. 3, 81 (1965). Gaver, R. C., and C. C. Sweeley. J. Am. Oil Chem. Soc. 88, 294 (1965). Puro, K. Biochim. Biophys. Acta 189, 401 (1969). Rapport, M. M., and N. Alonzo. J. Biol. Chem. 217, 193 (1955). Sweeley, C. C. J. Lipid Res. 8, 402 (1963). Lauter, C. J., and E. G. Trams. J. Lipid Res. 8, 136 (1962). Hakomori, S. -I. J. Biochem. (Tokyo) 88, 205 (1964). Bj5rndal, H., B. Lindberg, and S. Svensson. Acta Chem. Scand. 88, 1801 (1967). Bjdrndal, H., B. Lindberg, and S. Svensson. Carbohydrate Res. _5_, 433 (1967). 248. 249. 250. 251. 252. 253. 254. 255. 256. 257. 258. 259. 260. 261. 2620 263. 264. 308 Bj5rndal, H., C. G. Hellerqvist, B. Lindberg, and S. Svensson. Angew. Chem. Internat. Edit. 8, 610 (1970). Polito, A. J., T. Akita, and C. C. Sweeley. Biochem— istry l, 2609 (1968). Carter, H. E., and R. C. Gaver. J. Lipid Res. 8, 391 (1967). Sweeley, C. C., and E. A. Moscatelli. J. Lipid Res. 8, 40 (1959). Bartlett, G. R. J. Biol. Chem. 234, 466 (1959). Ackman, R. G. "Gas-Liquid Chromatography of Fatty Acids and Esters." In Methods of Enzymology. (J. M. Lowenstein, ed.), Vol. XIV, Academic Press, New York, 329 (1969). Jatzkewitz, H., and E. Mehl. Z. Physiol. Chem. 320, 251 (1960). Carter, H. E., and Y. Fujino. J. Biol. Chem. 221, 879 (1956). Hammarstrdm, S. J. Lipid Res. 88, 760 (1971). Carle, B. N., and W. H. Dewhirst. J. Am. Vet. Med. Assn. 101, 495 (1942). Winterbourn, C. C., and R. D. Batt. Biochim. Biophys. Acta 202, 1 (1970). Piomelli, S., G. Lurinsky, and L. R. Wasserman. J. Lab. Clin Med. 88, 659 (1967). Wintrobe, M. M. Clinical Hematology. Kimpton, London, 5th ed., 95 (1961). Ramirez, C. G., E. R. Miller, D. E. Ullrey, and J. A. Hoefer. J. Anim. Sci. 2;, 1068 (1963). Naiki, M., and T. Taketomi. Jap. J. Exptl. Med. 88, 549 (1969). Makita, A., L. Iwanaga, and T. Yamakawa. J. Biochem. (Tokyo) 2.5.: 202 (1964). Puro, K. Studies on Extraneural Gangliosides. Ph.D. Thesis, University of Helsinki (1969). 309 265. Dawson, G., and C. C. Sweeley. J. Lipid Res. 88, 56 (1971). 266. Choi, S.-I., J. V. Simone, and L. J. Journey. Brit. J. Haemat. 88, 93 (1972). 267. Rossi, E. C. Med. Clin. North Amer. 88, 25 (1972). 268. Petit, M. Glycolipides du Plasma et Des Globules Rouges de Rat. Ph.D. Thesis, Universite Paris (1972). 269. Sloviter, H. A. International Symposium on Glyco- saminoglycans, Glycoproteins and Glycosipids, Tokyo (1967). 270. Bremner, K. C. Aust. J. Exptl. Biol. Med. Sci. 88, 251 (1966). 271. Hillman, R. S., and C. A. Finch. Brit. J. Haemat. 81, 313 (1969). 272. Bush, J. A., W. N. Jensen, H. Ashenbrucker, G. E. Cartwright, and M. M. Wintrobe. J. Exptl. Med. 103, 161 (1955). 273. Bush, J. A., N. I. Berlin, W. N. Jensen, A. B. Brill, G. E. Cartwright, and M. M. Wintrobe. J. Exptl. Med. 101, 451 (1955). 274. Berlin, N. I., and C. Lotz. Proc. Soc. Exptl. Biol. Med. 78, 788 (1951). 275. Zilversmit, D. B. Amer. J. Med. 88, 832 (1960). 276. Rapport, M. M., V. P. Skipsky, and N. F. Alonzo. Nature 181, 1803 (1958). 277. Coles, E., and J. L. Foote. J. Lipid Res. ii, 433 (1970). 278. Heckers, H., and W. Stoffel. Z. Physiol. Chem. 353, 407 (1972). 279. Sweeley, C. C., and B. Walker. Anal. Chem. 88, 1461 (1964). 280. Wolfe, L. S., and J. A. Lowden. Can. J. Biochem. 88, 1041 (1964). 281. Chambers, R. E., and J. R. Clamp. Bdochem. J. 125, 1009 (1971). 282. 283. 284. 285. 286. 287. 288, 289. 290. 291. 292. 2930 294. 295. 296. 297. 298. 299, 310 Tao, R. V. P., and C. C. Sweeley. Biochim. Biophys. Acta 218, 372 (1970). Marcus, A. J., H. L. Ullman, and L. B. Safier. J. Clin. Invest. 88, 2602 (1972). Yu, R. K., and R. W. Ledeen. J. Lipid Res. 88, 680 (1972). Svennerholm, L., and A. Raal. Biochim. BiOphys. Acta 88, 422 (1961). Kuhn, R., H. Wiegandt, and H. Egge. Angew. Chem. 28, 580 (1961). Klenk, E., and W. Gielen. Z. Physiol. Chem. 326, 144 (1961). Trams, E. G., L. E. Giuffrida, and A. Karmen. Nature 193, 680 (1962). Klenk, E., and G. Padberg. Z. Physiol. Chem. 327, 249 (1962). Philippart, M., and J. Menkes. Biochem. Biophys. Res. Comm. i8, 551 (1964). Pilz, H., K. Sandhoff, and H. Jatzkewitz. J. Neurochem. 88, 1273 (1966). Lehninger, A. L. Proc. Nat. Acad. Sci. 88, 1069 (1968). Hakomori, S.-I., and W. T. Murakami. Proc. Nat. Acad. Sci. 88, 254 (1968). Wild, G., D. W. Wooley, and B. W. Gommi. Biochemistryf 8, 1671 (1967). Gielen, W. Z. Naturforsch. 21(b), 1007 (1966). Lewis, N., and P. W. Majerus. J. Clin. Invest. 88, 2114 (1969). Haslam, R. J. Nature 202, 765 (1964). Bettex—Galland, M., and E. F. Lfischer. Thromn. Diath. Haemorrh. 8, 178 (1960). Karpatkin, S., and R. M. Langer. J. Clin. Invest. 81, 2158 (1968). 300. 301. 302. 303. 304. 305. 306. 307. 308. 309. 310. 311. 312. 313. 314. 315. 311 Born, G. V. R. Biochem. J. 88, 695 (1958). Holmsen, H., H. J. Day, and H. Stormorken. Scand. J. Haemat. (Suppl. 8), 3 (1969). Day, H. J., G. A. T. Ang, and H. Holmsen. Proc. Soc. Exptl. Biol. Med. 139, 717 (1972). Mester, L., L. Szabados, G. V. R. Born, and F. Michal. Nature New Biol. 236, 213 (1972). Gaddum, J. H., and Z. P. Picarelli. Brit. J. Pharmacol. Chemother. 88, 323 (1957). Marcus, A. J. "The Role of Lipids in Blood Coagulation." In Advances in Lipid Research. (R. Paoletti and D. Kritchevsky, eds.), Vol. 4, Academic Press, New York, 1 (1966). Marcus, A. J., D. Zucker-Franklin, L. B. Safier, and H. L. Ullman. J. Clin. Invest. 5, 14 (1966). Bach, M. K., J. R. Brashler, K. J. Bloch, and K. F. Austen. "Studies on the Receptor Site for IgE Antibody on the Peritoneal Mast Cell of the Rat." In Biochemistry of the Acute Allergic Reactions. 2nd Internat. Symp. (K. F. Austen and E. L. Becker, eds.), Blackwell Scientific Publications, Oxford, 65 (1971). Michal, F. Nature 221, 1253 (1969). Solomon, H. M., N. M. Spirt, and W. B. Abrams. Clin. Pharmacol. Therap. ii, 838 (1910). Beutler, E., and W. Kuhl. J. Lab. Clin. Med. 18, 747 (1970). . Gottfried, E. L. J. Lipid Res. 88, 531 (1971). Gross, G. P., and W. E. Hathaway. Pediat. Res. 8, 593 (1972). Pearson, H. A. J. Pediat. 18, 166 (1967). Patten, B. M. "Embryology of the Pig." 3rd Edition, McGraw-Hill, New York, 103 (1948). Zilversmit, D. B., and Van Handel. Arch. Biochem. Biophys. 73, 224 (1958). 312 316. Jensen, W. N., J. A. Buch, H. Ashenbrucker, G. E. Cartwright, and M. M. Wintrobe. J. Exptl. Med. 103, 145 (1955). 317. Shemin, D., and D. Rittenberg. J. Biol. Chem. 166, 627 (1946). 318. Brown, I. W., and G. S. Eadie. J. Gen. Physiol. 88, 327 (1953). 319. Valentine, W. N., M. L. Pearce, R. F. Riley, E. Richter, and J. S. Lawrence. Proc. Soc. Exptl. Biol. Med. 22, 244 (1951). 320. Neuberger, A., and J. S. F. Niven. J. Physiol. 112, 292 (1951). 321. Van Gastel, C., D. van der Berg, J. de Gier, and L. L. M. van Deenen. Brit. J. Haemat. ii, 193 (1965). 322. Westerman, M. P., L. E. Pierce, and W. N. Jensen. J. Lab. Clin. Med. 88, 394 (1963). 323. Shattil, S. J., and R. A. Cooper. Clin. Res. i8, 431 (1971). 324. Hallinan, T., E. Eden, and R. North. Blood 88, 547 (1962). 325. Rouser, G., G. J. Nelson, S. Fleischer, and G. Simon. "Lipid Composition of Animal Cell Membranes, Organelles, and Organs." In Biological Membranes, Physical Fact and Function. (D. Chapman, ed.), Academic Press, New York, 5 (1968). 326. Fleischer, S., G. Rouser, B. Fleischer, A. Casu, and G. Kritchevsky. J. Lipid Res. 8, 170 (1967). 327. Shattil, S. J., and R. A. COOper. J. Lab. Clin. Med. 18, 215 (1972). 328. Ganzoni, R., S. Hillman, and C. A. Finch. Brit. J. Haemat. i8, 119 (1969). 329. Nestel, P. J. "Cholesterol Turnover in Man." In Advances in Lipid Research. (R. Paoletti and D. Kritchevsky, eds.), Vol. 8, Academic Press, New York, 1 (1970). 330. 331. 332. 333. 334. 335. 336. 337. 338. 339. 340. 341. 342. 313 Jungalwala, F. B., and R. M. C. Dawson. Biochem. J. 123, 683 (1971). 1. Holian, 0., D. Dill, and E. G. Brunngraber. Arch. Biochem. Biophys. 142, 111 (1971). Dawson, G., R. Matalon, and A. Dorfman. J. Biol. Chem. 247, 5944 (1972). Dawson, G., R. Matalon, and A. Dorfman. J. Biol. Chem. 247, 5951 (1972). Chatterjee, S. Biochemical Studies on Cultured Mouse Embryo Cells. Ph.D. Thesis, University of Toronto (1973). Hanahan, D. J., J. E. Elkhom, and B. Benson. Biochim. Biophys. Acta 231, 343 (1971). Nichaman, M. Z., R. E. Olson, and C. C. Sweeley. Am. J. Clin. Nutr. 88, 1070 (1967). Krivit, W., and L. Kern. Blood 88, 858 (1969). van Deenen, L. L. M., and J. de Gier. “Chemical Composition and Metabolism of Lipids in Red Cells of Various Animal Species." In The Red Blood Cells. (C. Bishop and D. M. Surgenor, eds.), Academic Press, New York, 243 (1964). Reed, C. F., and S. N. Swisher. J. Clin. Invest. 88, 777 (1966). Weed, R. I., and A. J. Bowdler. J. Clin. Invest. 88, 1137 (1966). Beutler, E. Exptl. Eye Res. ii, 261 (1971). Marks, P. A., A. Gellhorn, and C. Kidson. J. Biol. Chem. 235, 2579 (1960). APPENDIX Tao, LIST OF PUBLICATIONS R. V. P., and Sweeley, C. C. Occurrence of Hematoside in Human Plasma. Biochim. Biophy§. Acta, 218, 372 (1970). Sweeley, C. C., and Tao, R. V. P. Gas Chromatographic Esti- Tao , mation of Carbohydrates in Glycosphingolipids. £2: Methods in Carbohydrate Chemistry. R. L. Whistler and J. N. BeMiller (Editors), Vol. VI, Academic Press, New York, 8 (1972). R. V. P., Sweeley, C. C., and Jamieson, G. A. Sphingo- lipid Composition of Human Platelets. J. Lipid Res., 14, 16 (1973). 314