STUDES on. eevcosmmeoumw THE ANOMERIC conn-eumm 0F mew TRTHEXOSYL cmmms AND THE REmeme or AN EXCEPT'IONAL CASE OF TAY-SACHS DISEASE wzm VlSCERAL INVOLVEMENT Thesis for the Degree of M. S. MIGHE‘GAN STATE UNIVERSITY PAUL DANl£L SNYDER. JR. 1 9 7 O TTTT TTTTTT TTTTTTTT TTTTTTTTTTTTTTTTT TTT TTTTTT 3 1293 19751 15 . LIBRARY Michigan State University ABSTRACT STUDIES ON GLYCOSPHINGOLIPIDS: THE ANOMERIC CONFIGURATION OF FABRY TRIHEXOSYL CERAMIDE AND THE RECOGNITION OF AN EXCEPTIONAL CASE OF TAY-SACHS DISEASE WITH VISCERAL INVOLVEMENT By Paul Daniel Snyder, Jr. Two separate research topics are reported in this thesis: (l) the determination of the anomeric configura— tion of the glycosidic linkages of the carbohydrate moiety Of Fabry kidney trihexosyl ceramide and (2) the recogni- tion of an exceptional case of Tay-Sachs disease with visceral involvement and the partial identification of the three accumulating glycosphingolipids. To establish the anomeric configuration of Fabry trihexosyl ceramide, NMR analyses were carried out on the trimethylsilyl derivatives of the intact lipid, the in- tact trisaccharide moiety cleaved from this lipid, and glucosyl ceramide derived from Fabry trihexosyl ceramide via mild acid hydrolysis. Comparison of these findings with NMR data from trimethylsilyl derivatives of Gaucher Spleen glucosyl ceramide, reference mono- and disacchar- ides, and other model compounds established an all 8 Paul Daniel Snyder, Jr. configuration for Fabry trihexosyl ceramide, 115., galactosyl-(Bl+4)-galactosyl-(Bl+4)-glucosyl~(Bl+l)- ceramide. Todd Thomey was an 18-month old child who died from a lipid storage disorder diagnosed initially as clas- sical Tay-Sachs disease. Hepatosplenomegaly, however, indicated an unusual visceral involvement and fresh- frozen sections of brain, kidney, spleen, and liver from this patient were analyzed for their glycosphingolipid content. Brain, kidney, spleen, and liver from normal juvenile controls were also investigated and the data compared. Three glycosphingolipids had accumulated in Thomey's organs: (1) normal kidney globoside (galNAc— gal-gal-glc-cer) in the visceral organs, (2) asialo ganglioside GMZ (galNAc-gal-glc-cer) in the brain and liver, and (3) ganglioside GMZ (galNAc-gal-glc-cer) ex- NANA clusively in the brain. Thus, Thomey was not a classical case of Tay-Sachs disease, but an exceptional case with gross visceral involvement. Since all three accumulating glycosphingolipids possess a terminal B-N-acetylgalactosamine moiety, it is reasonable to postulate a B-N-acetylhexosaminidase defi- ciency as the cause of this disease. This possibility Paul Daniel Snyder, Jr. is discussed in relation to the only other reported case of exceptional Tay-Sachs disease in which a deficient B-N-acetylhexosaminidase was demonstrated. STUDIES ON GLYCOSPHINGOLIPIDS: THE ANOMERIC CONFIGURATION 0F FABRY TRIHEXOSYL CERAMIDE AND THE RECOGNITION OF AN EXCEPTIONAL CASE OF TAY-SACHS DISEASE WITH VISCERAL INVOLVEMENT By Paul Daniel Snyder, Jr. A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE 'Department of Biochemistry I970 (galls/3 5403/0 ACKNOWLEDGMENTS I would like to express my most sincere thanks to Dr. Charles C. Sweeley for his timely advice and continu- ous interest during the progress of these studies. Deep thanks are also due Dr. William Krivit and Dr. Robert Desnick for their gifts of organ sections from Todd Thomey and the normal controls analyzed in this research. I would also like to thank Mr. Jack Harten for his assist- ance with mass spectrometry and Dr. Glyn Dawson for his helpful advice in glycosphingolipid isolation and identi- fication techniques. Special thanks are due my parents, whose interest and understanding during this work were a steady source of strength and inspiration. ii TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES THE ANOMERIC CONFIGURATION OF FABRY TRIHEXOSYL CERAMIDE INTRODUCTION AND LITERATURE REVIEW . EXPERIMENTAL A. Materials B. Methods RESULTS A. NMR Analyses . B. Optical Rotation of Fabry and Red Cell GL- 3 . . . . . . . . . . DISCUSSION SUMMARY Page vi A CASE OF TAY-SACHS DISEASE WITH VISCERAL INVOLVEMENT: EVIDENCE FOR THE ACCUMULATION OF THREE GLYCOSPHINGOLIPIDS IN THE BRAIN AND VISCERAL ORGANS INTRODUCTION AND LITERATURE REVIEW . EXPERIMENTAL A. Materials B. Methods RESULTS A. TLC of Neutral Glycosphingolipids and Gangliosides . . . 32 38 38 39 42 42 Page B. GLC of TMS Neutral Glycosphingolipids, Gangliosides, and FAME . . . . . . 52 C. Mass Spectrometry of TMS Glycosphingolipids. 64 DISCUSSION . . . . . . . . . . . . . . . . . . . . 70 SUMMARY . . . . . . . . . . . . . . . . . . . . . 77 BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . 78 iv Table 2A. 28. 10. ll. 12. LIST OF TABLES Programmed GLC of TMS Reference Sugars and TMS Fabry Trisaccharide . Neutral Glycosphingolipids of .Thomey Kidney . . . . . . . . . Neutral Glycosphingolipids of Thomey Kidney . . . . . . . . Neutral Glycosphingolipids of Thomey Spleen . . . . . . Neutral Glycosphingolipids of Thomey Liver . . . . . . . . . . . . . . Neutral Glycosphingolipids of Normal Human Juvenile Kidney . . . . . Neutral Glycosphingolipids of Normal Human Juvenile Spleen . . . . Neutral Glycosphingolipids of Normal Human Juvenile Liver . . . . Neutral Glycosphingolipids of Thomey and Normal Juvenile Cerebral Grey and White Matter Thomey and Normal Juvenile Cerebral Grey Matter Gangliosides Thomey and Normal Juvenile Cerebral White Matter Gangliosides FAME from Thomey Abnormal and Corres- ponding Normal Juvenile Visceral Glyc05phingolipids FAME from Thomey and Tay- Sachs Cerebral Glycosphingolipids Page 20 53 54 55 56 57 58 59 6O 61 62 65 66 Figure 10. ll. 12. 13. LIST OF FIGURES Thin- layer chromatography of Fabry kidney glycosphingolipids . Density gradient elution curve of Fabry kidney glycosphingolipids . . NMR spectrum of Fabry trihexosyl ceramide (GL-3) dissolved in CDCl3 Structure of Fabry GL-3, alactosyl- (Bl+4)-galactosyl-(Bl+4T-glucosyl- (Bl+l)-ceramide . . . . . . . . TLC of Thomey kidney neutral glyc05phingolipids . TLC of normal human juvenile kidney neutral glycosphingolipids . . . TLC of normal human juvenile spleen neutral glycosphingolipids TLC of Thomey liver neutral glycosphingolipids TLC of normal human juvenile liver neutral glycosphingolipids TLC of Thomey cerebral white matter neutral glycosphingolipids TLC of normal human juvenile cerebral white matter neutral glycosphingolipids TLC of Thomey cerebral white matter gangliosides TLC of normal juvenile cerebral white matter gangliosides vi Page 14 16 25 28 43 44 45 46 47 48 49 50 51 Figure . Page l4. Mass spectra of TMS glyCOSphingolipids . . . 67 15. Mass spectrum of Thomey liver TMS tri- hexosyl ceramide, galNAc-gal-glc-cer . . . 68 vii THE ANOMERIC CONFIGURATION OF FABRY TRIHEXOSYL CERAMIDE INTRODUCTION AND LITERATURE REVIEW Interest in the neutral glycosphingolipids has increased greatly over the past decade due to their impli- cation as haptenic compounds in antigen—antibody reactions and their involvement in several genetic diseases classi- fied as inborn errors of lipid metabolism. By nature, neutral glycosphingolipids are compounds composed of (l) the long-chain base sphingosine and its homologs; (2) the sugars glucose, galactose, N-acetylgalactosamine, N-acetylglucosamine, and, in the case of the blood group substances, fucose; (3) long-chain fatty acids of varying chain length. Though present in most mammalian tissues and organs, as well as in blood, evidence as to the structure and overall function of these compounds has remained scant until recent years. With the development of more sophisticated and efficient techniques of detec- tion and isolation, however, the chemical structures of many glyc05phingolipids have now been elucidated. In one of the earliest structural studies, Nakayama determined via methylation in l950 that the glycosidic linkage of galactocerebroside was between the anomeric carbon of galactose and the primary hydroxyl group of sphingosine (l). Fujino and Negishi in a follow-up investigation used enzymatic hydrolysis with B-galactosidase to show that this linkage was of the B-configuration (2). The glycosidic linkage of gluco- cerebroside isolated from Gaucher spleen was also shown to be of the B-configuration, based on infrared spectro- scopy of the glucosyl sphingosine sulfate prepared from this lipid via the addition of methanolic H2504 to the acyl-free lipid dissolved in anhydrous methanol (3). By careful degradation of kerasin (galactocerebroside) and Gaucher cerebroside and by comparison of the products, Carter, gt_gl. in l96l confirmed the results of Rosenberg and Chargaff and concluded that kerasin and Gaucher cerebroside were identical compounds except in their sugar content (4). At about the same time, Rapport, gt_al. used hap- ten inhibition in an immunochemical study to show that the disaccharide moiety of cytolipin H (lactosyl ceramide) has the B-configuration (5). Galactosyl-(l+4)-glucosyl- ceramide isolated from human erythrocytes was found by Yamakawa, gt_gl. to also be of the B-configuration (6). Using permethylation and infrared spectrosc0py, these in- vestigators determined that the linkage between galactose and glucose in the lactosyl moiety was of the Bl+4 type. Continuing their work on human red blood cell glycosphingo- lipids, Yamakawa, gt_gl. in 1962 carried out a series of studies on globoside (tetrahexosyl ceramide) extracted from human red cell stroma (7). Combining periodate con- sumption data with that obtained from permethylation, gas-liquid chromatography, optical rotation, and infrared spectroscopy, these workers proposed the globoside struc- ture to be N- acetylgalactosaminoyl-(Bl+6)—galactosyl-(l+4)- galactosyl-(l+4)-glucosyl ceramide. Makita, gt_gl. in- vestigated the structure of human kidney globoside, however, using the same methods and postulated N-acetylgalactosaminoyl- (Bl+3)-galactosyl-(l+4)-galactosyl-(l+4)-glucosyl ceramide as the structure for this compound (8). In the same study, human red cell globoside was reexamined and evidence for a terminal 1+3 linkage was found rather than the l+6 linkage previously reported (7). In l965, Yamakawa, et al. settled the problem of human red cell globoside structure (9). Using improved gas-liquid chromatographic techniques, they discovered that the peak previously identified as 2,3,4- trimethylgalactoside was a product of inadequate solid support and stationary phase column conditions and not actually due to the presence of a terminal 1+6 linkage in the globoside carbohydrate moiety. Rather, they found a substantial peak for 2,4,6-trimethylgalactoside and, to- gether with periodate consumption data, these new data indicated a terminal 1+3 linkage. Thus, human red cell globoside was assigned the structure N-acetylgalactos- aminoyl-(l+3)-galactosyl-(l+4)-galactosyl-(l+4)-glucosyl ceramide and labeled globoside I to distinguish it from the blood-group active globoside II and III, also isolated from human red cell stroma and found to contain fucose, glucosamine, and sialic acid, in addition to glucose, galactose, and galactosamine. Globoside I was found by Rapport and Graf (10) to be immunochemically identical to cytolipin K, a tetrahexosyl ceramide previously obtained from human kidney containing fatty acid, sphingosine, glucose, galactose, and galactosamine in the molar ratio l:l:1:2:l and shown to have haptenic properties (11). Another glycosphingolipid of interest, possessing haptenic activity and having the general globoside structure, is the Forssman hapten isolated from equine kidney and spleen. This compound has the same composition as globoside I and cytolipin K, but was shown by Makita, gt_al. to differ from these lipids in its mobility on thin-layer chromato- graphy, its optical rotation, and its immunological reac- tivity (12). Chemical investigation of this glycosphingo- lipid led to the assignment of the structure N-acetylgalac- tosaminoyl-(al+3)-galactosy1-(Bl+4)-galactosyl-(1+4)-glucosyl ceramide, thus differing from globoside I and cytolipin K only in the anomeric configuration of the terminal galac- tosidic linkage. This finding dramatizes how such a seemingly unimportant structural difference can profoundly affect the physicochemical and serological behavior of glycosphingolipids of identical gross chemical composition. Glycosphingolipids in various disease states have also been studied structurally. In 1963, Sweeley and Klionsky found that trihexosyl ceramide is the lipid that accumulates in the kidney and other organs and tissues of persons afflicted with Fabry's disease, an X-linked lipidosis (13). Further investigations of this glyco- sphingolipid by these workers revealed the structure to be galactosyl-(1+4)—galactosyl-(1+4)-glucosyl-(l+1)- ceramide (14). The determination of the anomeric con- figuration of the glycosidic linkages of the carbohydrate moiety remained unsolved until the present and will be the topic of part one of this thesis. A study by Makita and Yamakawa of glycosphingolipids isolated from normal human kidney revealed that trihexosyl ceramide from this tissue was structurally identical to that isolated from Fabry kidney (15). Numerous investigations on glycosphingolipids isolated from mammalian sources other than human have been conducted in recent years. Gray in 1964 isolated and characterized glycosphingolipids from BP8/C3H mouse ascites-sarcoma cells and found lactose and galactosyl- (Bl+6)-galactose among the hydrolysis products of dihexo- syl ceramide and trihexosyl ceramide, respectively (16). In a later study, Adams and Gray elucidated the oligosac- charide moieties of pig lung glycosphingolipids and found structures similar to glycosphingolipids identified in human kidney and red blood cells (17). Combining a two-stage methylation-methanolysis with gas-liquid chromatographic and infrared spectrosc0pic data, they identified the following glycosphingolipids: glucosyl- (Bl+l)-ceramide, galactosyl-(Bl+4)-glucosy1-(Bl+l)- ceramide, a small amount of galactosyl-(Bl+4)-galactosyl- (Bl+l)-ceramide, galactosyl-(l+4)-galactosyl-(l+4)- glucosyl-(l+l)-ceramide, and N-acetylgalactosaminoyl- (l+3)-galactosyl-(l+4)-galactosyl-(l+4)-glucosyl-(l+l)- ceramide. More recently, Miyatake, gt_gl. determined the structure of the main glycosphingolipid of hog erythrocytes to be N-acetylgalactosaminoyl-(Bl+3)-galactosyl-(1+4)- galactosyl-(l+4)-glucosy1-(l+l)-ceramide (18). The termi- nal B-linkage was uncovered via the liberation of the terminal N-acetylgalactosamine with B-N-acetylhexosamini- dase obtained from hog epididymis tissue. Structures 0f the carbohydrate moieties of glycosphingolipids from the kidneys of five different strains of mice were elaborated in 1968 by Adams and Gray (19). In all five strains studied, the major mono-, tri-, and tetrahexosides found were glucosyl-(l+l)-ceramide, galactosyl-(l+4)-galactosyl- (l+4)-glucosyl-(l+l)-ceramide, and N-acetylgalactosaminoyl- (l+3)-galactosy1-(l+4)-galactosy1-(l+4)-glucosyl-(l+l)- ceramide, respectively. In one strain, however, galactosyl- (31+4)-glucosy1-(Bl+l)-ceramide was the major dihexoside and in the other strains the less common galactosyl—(Bl+4)- galactosyl-(Bl+l)-ceramide was dominant. A similar study of rat kidney glycosphingolipids by Kawanami confirmed the structure of rat kidney globoside to be N-acetyl- galactosaminoyl-(l+3)-galactosyl-(l+4)-galactosyl-(l+4)- glucosyl ceramide, with a small amount of N-acetylgalac- tosaminoyl trigalactosyl ceramide also being present (20). One of the most complex glycosphingolipids to be struc- turally identified is the pentahexosyl ceramide of rabbit erythrocytes and reticulocytes. Eto, gt_al. in a chemical and immunochemical study proposed the structure of this lipid to be galactosyl-(l+3)-galactosyl-(l+3)-N-acetyl- glucosaminoyl-(l+3)-galactosyl-(l+4)-glucosyl ceramide (21). It was also shown that this glycosphingolipid in- hibits the agglutination of human B-red blood cells with its corresponding antibody, suggesting a terminal a-galactosidic linkage as is found in blood group B carbohydrate. Another approach to the structural study of glycosphingolipids has been enzymatic hydrolysis. As cited earlier, Fujino and Negishi used a- and B-galacto- sidase to determine the B-configuration of galactocerebro- side in one of the earliest enzymatic analyses (2). Lactosyl ceramide from ox spleen was hydrolyzed by rat brain B-galactosidase in a study carried out by Gatt and Rapport, thus confirming the structure of this glycosphingo- lipid to be ga1actosy1-(81+4)-glucosyl-(l+lT-Ceramide (22)- In a subsequent study, Gatt confirmed the B-configuration of Gaucher glucosyl ceramide using enzymatic hydrolysis, with ox brain B-glucosidase (23). Frohwein and Gatt em- ployed calf brain B-N-acetylhexosaminidase in the hydroly- sis of synthetic N-acetylgalactosaminoyl-(Bl+4)-galactosyl- (Bl+4)-glucosyl-(Bl+l)-ceramide, human red blood cell globoside, and lay-Sachs ganglioside (24). This work provided further proof for the terminal B-galactosidic linkage in globoside from human red cells. As indicated earlier, Miyatake, gt_gl. in a recent investigation used B-N-acetylhexosaminidase from hog epididymis tissue to establish the terminal B-galactosidic linkage of globo- side isolated from hog erythrocytes (18). Thus, enzymatic hydrolysis with the appropriate glycosidases has proven valuable in determining or confirming the anomeric con- figuration of glycosidic linkages of the carbohydrate moieties of various glycosphingolipids. Nuclear magnetic resonance spectrosc0py (NMR) has been used infrequently to determine anomeric configura- tions in the glycosphingolipids. In the only published study to date, Kawanami used NMR to establish the struc- ture of trihexosyl ceramide isolated from Nakahara- Fukuoka sarcoma tissue as galactosyl-(al+4)-galactosyl- (Bl+4)-glucosyl-(l+l)-ceramide (25). This terminal a-galactosidic linkage is, however, inconsistent with results obtained for the terminal glycosidic linkage in 10 other glycosphingolipids, with the exception of the Forss- man hapten and Gray's BP8/C3H mouse ascites-sarcoma trihexosyl ceramide. In the following study, the complete structure of the carbohydrate moiety of trihexosyl cera- mide isolated from the kidneys of two Fabry patients will be discussed. Convincing NMR spectrosc0pic evidence for exclusive 8 configurations, 115;, galactosyl-(Bl+4)- galactosyl-(Bl+4)-glucosyl-(Bl+l)-ceramide, is presented, thus completing the structural work on this glycosphingo- lipid begun by Sweeley and Klionsky in 1963 (13). EXPERIMENTAL A. Materials Fabry trihexosyl ceramide (GL-3) was extracted from two portions of fresh-frozen kidney obtained from two deceased male Fabry patients. Glucosyl ceramide (GL-l) was isolated from the spleen of a patient with Gaucher's disease. Glycosphingolipids used as reference standards for thin-layer chromatography (TLC) were iso- lated and purified from packed human red blood cells by the method of Vance and Sweeley (26). All solvents and chemicals used were analytical or reagent grade unless otherwise indicated. Heat-activated Unisil (200/325 mesh, Clarkson Chemical Co., Williamsport, Pa.) was used in all silicic acid column chromatography. For all TLC, glass plates were spread-coated to a thickness of 250p with Silica Gel G (Brinkmann Instruments, Inc., Westbury, New York) using a Desaga TLC spreader. Gas-liquid chromatography (GLC) was carried out on a F & M Model 402 gas chromato- graph (Hewlett Packard Co., Avondale, Pa.) equipped with a glass, six-foot 3% SE-30 or OV-l column (packing from Applied Science Co., State College, Pa.). Hexamethyl- disilazane and trimethylchlorosilane used in the silylating ll 12 (TMS) reagent were purchased from Applied Science Co., State College, Pa. B. Methods Lipid Extraction Total lipids weighing 3.414 gm were extracted according to the method of Folch, gt_gl. (27) from 32.8 gm (wet weight) of Fabry kidney from two deceased male patients. Prior to column chromatography, the total lipids in chloroform were shaken vigorously with several grams of Celite filter aid and suction-filtered to obtain a clear filtrate. The residue was then washed with chloroform-methanol (CM) 1:1 (v/v) to obtain a second clear filtrate. This filtration removed all non-lipid or CM-insoluble impurities which would interfere with efficient column chromatography. Column Chromatography and Analytical TLC A llO-gm Unisil column was prepared and pre-washed with 10 column volumes of CM 9:1 (v/v) and enough CHCl3 to restore column translucency. The CHCl3 filtrate was applied and eluted with CM 9:1 (v/v) until the Lands lipid spot test (28) showed an absence of lipid in the eluate. To remove glycosphingolipids and some polar lipids, CM 1:1 (v/v) was next added, followed by pure methanol to remove remaining lipids. The Lands test was used after these elutions as before. 13 Analytical TLC was done on 100 pl each of the CM 9:1 (2.399 gm eluted) and 1:1 (452 mg eluted) column eluates and the CM 1:1 (378 mg) Celite filtrate to detect the presence of glycosphingolipids. Upon iodine vapor visualization, spots corresponding to lactosyl ceramide (GL-2) and GL-3 were found in the CM 1:1 column eluate and Celite filtrate, as is shown by Figure 1. Celite 545 in CM 2:1 (v/v) was added to a 1.5-cm i.d. column to give a bed height of 19 cm. The CM 1:1 eluate, redis- solved-in CM 2:1, was applied and eluted with CM 2:1 to remove colored impurities appearing on TLC. This eluate was then dialyzed against distilled water in the cold overnight with two changes of water. The original Celite filtrate was subjected to mild aklali-catalyzed methanoly- sis (26) to remove phospholipid impurities and then eluted from a 15-gm Unisil column with CM 19:1, 2:1, and 1:4 (v/v), respectively. Glycosphingolipids were detected via TLC in the CM 2:1 eluate and this fraction was like- wise dialyzed against distilled water in the cold. Lipids from the combined dialysates, totaling 528 mg, were then ready for density gradient column chromatography as a final purification step. Densitnyradient Column Chromatography In this procedure, a density gradient is estab- lished between two solvents of unequal density. The level 14 I II III IV V T TT .9; o 0 _2_ Figure 1. Thin-layer chromatography of Fabry kidney glyc05phingolipids. 1. Human red blood cell glycosphingolipid reference standards: (1) lactosyl ceramide (GL-2), (2) tri- hexos l ceramide (GL-3), and (3) globoside (GL-4 . 11. CM 9:1 fraction. III. CM 1:1 fraction. IV. Celite filtrate. V. Human plasma phospholipids. 15 of the less dense solvent must be higher than that of the denser solvent to prevent backward flow due to the force of gravity. The required height difference is given by the following relationship: -- 2:: 3:23:11: '12:: 3:22: 31:22:. For this study, the mixing bottle contained 1700 m1 of CM 94:6 (v/v, sp. gravity = 1.447) and to another bottle was added enough CM 65:35 (v/v, sp. gravity = 1.225) to give a solvent height 1.16 greater than that in the mix- ing bottle. A 150-gm Unisil column with 2.5-cm i.d. was prepared for this elution. After fitting a Gilson auto- matic fraction collector with 225 30-ml test tubes, the elution was started and allowed to run overnight. The Lands spot test showed lipid present in tubes 70 through 190. The contents of every three tubes Were pooled and, beginning with tubes 70-72 (fraction 24), were taken to dryness and weighed. Fractions 28-30, 32-35, and 41-54 contained the bulk of the weight as is shown by the graph in Figure 2. Analytical TLC of these fractions showed GL-2 to be present in 28-30 and 32-35 and GL-3 in 41-54. Phospholipid impurities persisted, however, and the fractions were subjected to mild alkali-catalyzed methanoly- sis. After repeated hexane extraction of the methanolysates, analytical TLC showed the fractions to contain pure GL-2 and GL-3. Final weights of pure GL-2 and GL-3 were 63 mg and 272 mg, respectively. 16 4o - ‘ 20 - III I l I l 0 50 60 Figure 2. Density gradient elution curve of Fabry kidney gly- cosphingolipids. Ordinate: weight of lipid eluted in m ; abcissa: fraction number (3 tubes = 1 frac— tionT. I & II = dihexosyl ceramide; III = trihexo- syl ceramide. . 17 Degradation of GL-3 To obtain the intact oligosaccharide moiety for structural analysis, 100 mg of pure GL-3 was subjected to the method of Hakomori (29). This procedure involves an osmium tetroxide-catalyzed periodate oxidation of the sphingosine moiety of the glycosphingolipid. The follow- ing is a stepwise account of the degradation: (1) One hundred mg of Fabry GL-3 was dissolved in 25 ml of pyridine-acetic anhydride 3:2 (v/v) and a1- 1owed to stand at room temperature in a stoppered flask overnight. (2) To remove the acetylating agent, 25 ml of water was added to the flask and the contents lyophilized. (3) The yellowish residue was dissolved in a small volume of CM 1:1, transferred to a screw-cap test tube, and evaporated to dryness under N2. To insure complete removal of water, first absolute ethanol and then acetone were successively added, evaporating to dryness under N2 each time. (4) The residue was next dissolved in 25 ml of dry dioxane and 3.75 ml of freshly-prepared 0.2M NaIO4 plus 0.50 ml of 1% 0504 in dry ethyl ether added with thorough mixing. The mixture was allowed to stand in the cold overnight. (5) A few drops of ethylene glycol were then added to react with excess NaIO4. When no additional 18 precipitation was observed, the mixture was divided be- tween two 40-ml centrifuge tubes, twice the volume of CHCl3 was added to each with mixing, and both tubes were spun in a table centrifuge for about five minutes. The CHCl3 lower phases were removed, divided among four 40-ml centrifuge tubes, equal volumes of distilled water added, and the tubes centrifuged again for five minutes. This water wash was repeated ten times, the aqueous upper phases being discarded each time. The four CHCl3 extracts were then combined and evaporated to dryness under N2. (6) Vacuum desiccation over KOH pellets for 24 hours was employed to remove all harmful OsO4 vapors. (7) The dry residue was dissolved in 10 ml of methanol, 2.50 ml of freshly-prepared 0.5% NaOMe in methanol added with mixing,and the solution allowed to stand at room temperature for one hour. Next was added enough aqueous 0.5% HOAc to bring the solution to neu- trality. After the subsequent addition of twice the volume of distilled water, the solution was centrifuged for ten minutes. The clear, yellowish supernatant was removed, taken to dryness on a flash evaporator, and the residue dissolved in 7.0 m1 of methanol. GLC of Oligosaccharide Two umoles of a mannitol standard solution (2.0 umole/ml) were added to 0.14 ml (1/50 of the original 19 material) of the Fabry oligosaccharide in MeOH, the solu- tion evaporated to dryness, and 3.0 m1 of 0.75 N HCL in dry MeOH added for overnight methanolysis at 82°C. Sub- sequent isothermal GLC on a 3% OV-l column at 170°C of the TMS derivative of the dry methanolysate indicated a trisaccharide with a galactose:glucose ratio of 2.06:1.00. The total amount of GL-3 oligosaccharide obtained was cal- culated to be 20.5 umoles. To prove the intactness of the trisaccharide, another 1/50 of the original material in solution was evaporated to dryness and TMS reagent added directly and allowed to stand overnight. Likewise, an aqueous solu- tion of 5 mg each of D-glucose, D-lactose, and raffinose was allowed to equilibrate for one hour and then evaporated to dryness under N2. One ml of TMS reagent was next added to this mixture and allowed to stand overnight. A pro- grammed GLC run on a 3% OV-l column, beginning at 140°C and rising at 2°/min, of the TMS sugar mixture gave the a- and B-anomers of the three sugars at the retention times listed in Table l. The TMS Fabry trisaccharide programmed under identical conditions gave a- and B-anomers only in the region of the trisaccharide raffinose. NMR of TMS Oligosaccharide The TMS derivative of the oligosaccharide was prepared and purified as a pale yellow oil upon two 20 Table l.--Programmed GLC of TMS Reference Sugars and TMS Fabry Trisaccharide. Retention Time (Minutes) Sugar Anomer Sugar Mixture Fabry Trisaccharide Glucose a 20.5 - B 25.5 - Lactose a 53.5 - B 58.5 - Raffinose a 100.5 87.7 B 105.5 111.7 21 washings with dry ethyl acetate. The oil was then taken up in a small volume of dry ethyl acetate and sent to Dr. C. E. Griffin at the University of Pittsburgh for NMR spectrometry. A Varian Associates HA-100 spectrom- eter was used in this analysis. External CHCl3 was used as a lock signal with the TMS oligosaccharide being dis- solved in CDC13. Sweep widths of 50-200 Hz were used in scanning the anomeric proton region. All signals were accumulated by a Northern Scientific computer and 100-200 sweeps were taken to obtain chemical shift and coupling constant values. The data collected in this analysis were coupled with NMR data previously obtained from intact TMS Fabry GL-3 and compared with spectra of TMS lactose and other established NMR values for anomeric configurations of glycosidic linkages cited in the literature. To establish the anomeric configuration of the remaining glycosidic bond in the glucosyl-(1+l)-ceramide portion of Fabry GL-3, mild acid hydrolysis (0.3N HCl in CM 2:1, v/v, at 60°C for three hrs.) was used on 50 mg of GL-3 to cleave off the two galactose units, leaving the GL-l moiety intact. Purity of the derived GL-l (yield, 5 mg) was established by TLC and the pure TMS derivative was prepared for NMR as before. The TMS deriva- tive of known glucosyl-(81+1)-ceramide from Gaucher Spleen (3,4) was likewise prepared and, together with the TMS 22 Fabry GL-l, was sent to Dr. Griffin for NMR analysis. The data were compared with the NMR spectra of TMS a- and B-D-methyl glucoside and galactoside and of intact TMS Fabry GL-3. Optical Rotation of GL-3 Optical rotation data were obtained using a Zeiss polarimeter fitted with a 1.0-cm cell. Pure Fabry GL-3 and pure GL-3 isolated from human red blood cells were dissolved in dry pyridine and their rotations versus the mercury 578 mu and 546 mu wavelengths compared. RESULTS A. NMR Analyses The NMR spectrum of TMS GL-l derived from Fabry GL-3 via acid hydrolysis exhibited an anomeric proton 7.0 Hz. These data com- resonance at T = 6.03 ppm, J pare very favorably with those obtained for TMS Gaucher spleen GL-l, glucosyl-(Bl+1)-ceramide (3,4): 1 = 6.02 ppm, J = 7.0 Hz; TMS lactose: Hg, T = 5.86 ppm, J = 7.6 Hz; H“, T = 5.11 ppm, J = 3.0 Hz; TMS methyl B-galactoside: 1 6.06 ppm, J 7.0 Hz; and methyl B-glucoside: (30). Values for T 5.95 ppm, J = 7.0 Hz, in DMSO-d 6 the corresponding o-anomers are quite different, indi- cating a B-linkage for Fabry GL-l. For TMS maltose, O-a-D-glucopyranosyl-(1+4)-B-D-glucopyranose, these parameters are: Hg, T = 4.69 ppm, J = 3.2 H2; H8, T = 5.34 ppm, J 7.1 Hz, in acetone-d6 (31); TMS methyl a-galactoside: 1 5.53 ppm, J = 3.0 Hz; and methyl a-glucoside: T = 5.45 ppm, J = 3.0 Hz, in DMSO-d (30). 6 Two anomeric proton resonances were identified for TMS Fabry GL-3 trisaccharide: a doublet at T = 5.98 ppm, J = 7.3-7.5 Hz, with relative intensity of 2.5, and a doublet at 1 = 5.00 ppm, J = 3.0 Hz, with relative inten- sity of 0.5. Broadened upfield resonances indicated 23 24 protons with slightly differing chemical shifts or coup- ling constants. The Hg of the two galactosidic linkages plus the anomeric proton of B-glucose adequately account for the 2.5 proton doublet at T = 5.98 ppm, while the anomeric proton of a-glucose is assumed to contribute the 0.5 proton doublet at T = 5.00 ppm. In addition, the spec- trum of intact TMS Fabry GL-3 exhibited broadened doublets at T = 6.01 ppm, J = 7.0 Hz, and T = 5.81 ppm, J = 7.0 Hz, in the H8 chemical shift region and showed no doublets in the H3 chemical shift region (T = 4.5-5.5 ppm), as is illustrated by Figure 3. The two smaller resonances present in this region will be discussed in the following section. These data, therefore, coupled with NMR param- eters cited in the preceding paragraph strongly suggest an all-B anomeric configuration for Fabry GL-3. B. Optical Rotation of Fabry and Red Cell GL-3 For Fabry GL-3, it was found that [“1526 = +300 and [a]§;8 = +260 (c = 2.01). For human red blood cell GL-3, it was found that [“1526 = +180 and [“1538 = +150 (c = 2.00). These data cannot readily be compared to those determined by Kawanami for Nakahara-Fukuoka sarcoma GL-3 (25): [a1ggg = +23.9° (c = 1.02), and by Makita for normal human kidney GL-3 (32): [91:29 = +26.1o (c = 2.13), since the standard sodium D-line (589 mu) was not used. An earlier rotation, however, taken by Sweeley and Klionsky (14) gave [a]§gg = +34.2° (c = 2.02) for Fabry GL-3. 25 ‘7’ Figure 3. NMR spectrum of Fabry trihexosyl ceramide (GL-3) dissolved in CDC13. DISCUSSION The method of Hakomori (29) used in this work to obtain the intact trisaccharide unit from Fabry GL-3 pro- vides a convenient means of obtaining the oligosaccharide moieties from both simple and complex glycosphingolipids. Free from the long-chain base (LCB) and acyl portions of the original molecule, the anomeric configuration of the oligosaccharide moiety can then be examined via NMR spectrometry. Thus, possible interference from protons in the LCB and acyl moieties is eliminated and anomeric configurations can be unambiguously assigned. In order to assign anomeric configurations to the glycosidic linkages of Fabry GL-3, it was necessary to compare the NMR data obtained with values for reference sugars analyzed under identical conditions and with known NMR parameters for various sugars cited in the literature (30,31,33-35). Typical chemical shift values for glyco- sidic linkages are those established by van der Veen (in 020): Hz, T = 4.88 s 0.33 ppm; Hg, T (34). This difference is further emphasized when the = 5.56 i 0.06 ppm C-1, C-2 proton-proton spin-spin coupling constants are considered. The diaxial relationship of the protons in B-linkages yields a coupling constant (J1 = 7.2 i 0.2 Hz) 2 26 27 more than twice that found for the axial-equatorial arrangement in a-linkages (J12 = 3.2 i 0.6 Hz) (34). This is a result of the fact that C-l protons of B-glyco- sidic linkages are more highly shielded than those of a-linkages. From NMR data previously cited for intact TMS Fabry GL-3 and TMS Fabry GL-l derived from GL-3, unambiguous anomeric assignment of the three glycosidic linkages of Fabry GL-3 as all-B has been established. The resonance at T = 5.38 ppm was due to the presence of some CHCl3 as an impurity in the CDC13. The slight resonance at T = 5.18 ppm remains unassigned, since it is downfield from the HS region in TMS methyl a-galacto- side (T = 5.53 ppm) and is displaced upfield from the expected anomeric region (T = 4.6-4.9 ppm) for a galacto- syl-(al+4)-galactosyl linkage. It can thus be assumed that this resonance is not due to an anomeric proton. Figure 4 illustrates the structure of the intact Fabry GL-3 molecule. The findings contrast with those of Kawanami (25) and Gray (16) in their studies of glycosphingolipids of cancer tissues. It may be possible, however, that the GL-3 configurations found by these investigators are peculiar to glycosphingolipids of cancerous tissues and, therefore, cannot be accurately compared with the data presented here. 28 OH CHzOH . HO OH 011,014 H0 cw.ow O OH HO 0 O 011 \€le 0 ijm "0H Figure 4. Structure of Fabry GL-3, galactosyl-(Bl+4)—galac- tosyl-(Bl+4)-glucosyl-(Bl+l)-ceramide. 29 A very recent paper by Kint (36) shows the com- plete absence in leukocytes of the enzyme a-galactosidase from males with Fabry's disease and a reduced level of this enzyme in female carriers. Kint uses this evidence, without structural study of Fabry GL-3 itself, to suggest that the accumulating GL-3 possesses a terminal a-galacto- sidic linkage. Since the NMR analyses of highly purified Fabry trisaccharide and intact GL-3 have unequivocally shown the glycosidic linkages of this compound to be all 8, Kint's hypothesis seems untenable. In addition, his enzyme assays were carried out using artificial substrates; nowhere does he show that his nonspecific a-galactosidase preparation cleaves the terminal galactose moiety of Fabry GL-3. Until this is demonstrated, coupled with a careful structural analysis of purified GL-3 from his patients, his proposal of a GL-3 with a terminal a-galac- tose moiety is indeed questionable. SUMMARY From the data presented, it has been concluded that the glycosidic linkages in Fabry GL-3 are unmistak- ably in the B anomeric configuration. Thus it now seems certain that Fabry's disease is strictly an enzyme defi- ciency disease and not a combination of a deficient enzyme and defective substrate. NMR analysis of the TMS derivative of the intact glycosphingolipid proved invaluable in this study, indi- cating that the determination of anomeric configurations of other glycosphingolipids via this method is quite feasible. Despite the high proton content of the intact compound, the NMR absorptions of interest occur in a region of the spectrum relatively free from conflicting absorptions. An assignment of anomeric configuration could therefore be made without ambiguity. 3O A CASE OF TAY-SACHS DISEASE WITH VISCERAL INVOLVEMENT: EVIDENCE FOR THE ACCUMULATION OF THREE GLYCOSPHINGOLIPIDS IN THE BRAIN AND VISCERAL ORGANS 31 INTRODUCTION AND LITERATURE REVIEW Classical Tay-Sachs disease is a biochemical malady affecting the central nervous system (CNS), re- sulting in death before the age of four. Onset generally begins by six months of age and involves progressive de- velopmental retardation, dementia, paralysis, and blind- ness, characterized by a cherry-red spot in the retina. These clinical symptoms are a direct result of the steady accumulation of ganglioside GM2’ N-acetylgalactosaminoyl- (31+4)-galactosyl-(1+4)-glucosyl-(l+1)-ceramide (37,24), (3 2. N-acetylneuraminyl and, to a lesser extent, its asialo derivative, N-acetyl- galactosaminoyl-(Bl+4)-galactosyl-(1+4)-glucosy1-(1+l)- ceramide (37,24), in the ganglion and glial cells of the CNS. An excellent review (38) covers the medical and biochemical history of classical Tay-Sachs disease and, therefore, those details will not be discussed here. In- formation is not so abundant, however, concerning aberra- tions of Tay-Sachs disease involving the visceral organs. Only a few such instances have been recorded and some of them have not been clearly differentiated from other gang- lioside storage disorders. 32 33 Perhaps the earliest recorded definitive case of Tay-Sachs disease with visceral involvement was described by Turban in 1944 (39). A non-Jewish child died at age two from a progressive neurological illness. Histologi- cally, the brain was typical of Tay-Sachs disease and the cerebral cortex exhibited a fivefold increase in gangliosides. Vacuolation of the liver and renal tubules was noted and foam cells were found in the pulmonary alveoli. Similarly, Norman, §t_gl,(40) in 1959 reported a case of a non-Jewish child who died at age 17 months from Tay-Sachs disease and who exhibited gross visceral in- volvement. These workers found that, in addition to a large excess of gangliosides in the brain, the liver, Spleen, lymph glands, thymus, bone marrow, adrenals, lung, and intestine exhibited lipid storage. A large excess of hexosamine was discovered upon chemical analysis of the liver and spleen. This case had originally been diagnosed as a Niemann-Pick, but absence of sphingomyelin accumula- tion in the brain and viscera plus the excessive ganglio- side accumulation confirmed that this was a Tay-Sachs patient. No attempt was made in this study to characterize the accumulating ganglioside or hexosamine material. In still another case reported in 1964, Norman gt_al. (41) investigated the brain and viscera of a non- Jewish child who died at age 16 months from Tay-Sachs 34 disease with visceral involvement. Gangliosides, as estimated from the neuraminic acid content, were elevated 3-1/2 times in cerebral cortex and 10 times in white matter. Foam cells were found in the alveolar septa and vacuolation of liver and renal tubular cells was evident. The ganglioside accumulation in this case also resembled that of gargoylism, especially the levels noted in the cortex. But a final diagnosis of Tay-Sachs disease was made on the basis of the tenfold increase in white matter gangliosides. Confusion existed at this time because ganglioside separation techniques were still not entirely reliable and the nature of the stored ganglioside in Tay- Sachs disease remained uncertain. Svennerholm and Raal (42) had shown in 1961 that more than 90% of the total lay-Sachs gangliosides were of an “abnormal" type with higher Rf-values on TLC than normal brain gangliosides. But the final structure of Tay-Sachs ganglioside remained somewhat ambiguous until Ledeen's work in 1965 (37) finally established its structure to be N-acetylgalactosaminoyl- (l+4)-galactosyl-(l+4)-glucosyl-(l+1)-ceramide, verifying Ti) 2 N-acetylneuraminyl the structure proposed by Makita and Yamakawa in 1963 (43). Eeg-Olofsson, gt_al, in 1966 (44) and Suzuki, gt al. in 1969 (45) reported cases of lay-Sachs disease in which ganglioside GM2 was elevated in the livers and 35 spleens of their patients. These were the first instances in which Tay-Sachs ganglioside was shown to accumulate in organs other than the CNS. Eeg-Olofsson, g£_al. found that the liver G of their patient contained equal M2 amounts of C18 and 022 + 024 fatty acids. On this basis, they concluded that Tay-Sachs disease could be considered El a generalized disorder of ganglioside metabolism, since ETW the fatty acid pattern indicated both a neural and vis- T ceral synthesis of ganglioside 6M2“ TTT Sandhoff, ££_1l- (46,47) recorded still another :9 variation of Tay-Sachs disease in 1968 when they reported a case of a non-Jewish male who died at age 2-1/2 in which normal kidney globoside, N-acetylgalactosaminoyl- (81+3)-galactosyl-(1+4)-ga1actosyl-(l+4)-glucosy1-(l+l)- ceramide (8,9,24), accumulated in the visceral organs in addition to the excessive presence of ganglioside GM2 and its asialo derivative in the brain. Three cases of classical Tay-Sachs disease examined by these investiga- tors showed no increased globoside levels in the visceral organs. Enzyme assays of brain and visceral tissue homo- genates from seven normals, three classical lay-Sachs cases, and the exceptional case of Tay-Sachs, showed nor- mal to somewhat above normal activities for neuraminidase and nonspecific B-N-acetylhexosaminidase in the controls and classical Tay-Sachs patients. The exceptional case of Tay-Sachs, however, exhibited an almost complete 36 absence of hexosaminidase in all tissues assayed; neura- minidase showed normal activity. These investigators therefore postulated the missing hexosaminidase as the cause of glycosphingolipid accumulation, since all three excessive compounds possess a terminal B-N-acetylgalacto- samine moiety. This same case was also reported by Pilz, gt_gl. (48). In 1969 Taketomi and Kawamura (49) reported a thorough investigation of cerebral and visceral glyco- sphingolipids in a case of lay-Sachs disease. They found the usual excessive accumulation of ganglioside GM2 in the brain, but failed to detect any ganglioside storage in the kidney, spleen, or liver. In addition, they found only a trace of glucosyl ceramide and a small amount of lactosyl ceramide in the brain. 0n the other hand, lactosyl ceramide was the major glycosphingolipid of Tay-Sachs spleen and trihexosyl ceramide and hematoside (ganglioside 6M3) dominated the liver glycosphingolipids. Asialo GM2 was limited to the Tay-Sachs brain. Kidney globoside was present in the kidney, spleen, and to a lesser extent in the liver of this patient. Amounts of this glycosphingolipid in these organs were not quanti- tated, but from the discussion presented, they did not seem exceptional. In a similar case of Tay-Sachs disease with visceral involvement, Grégoire, gt_al. (50) detected the storage of asialo GM2 in both cerebral grey and white 37 matter and the liver, in addition to the usual ganglio- side GM2 accumulation in the brain. This study presents the case of Todd Thomey, a child who died at age 18 months from a lipid storage dis- order diagnosed as classical Tay-Sachs disease. The cherry-red spot was present in the retina and hepato- splenomegaly was observed, strongly suggesting visceral involvement. Neutral glycosphingolipids and gangliosides were extracted from Thomey's cerebral grey and white mat- ter, kidney, spleen, and liver and compared with those from normal controls identically extracted. The data obtained indicate a type of Tay-Sachs disease with vis- ceral involvement very similar to the case of Sandhoff, et al. (46,47). EXPERIMENTAL A. Materials Portions of Thomey brain, kidney, liver, and Spleen (fresh-frozen at autopsy) were shipped to this laboratory over dry ice via air express, courtesy of Dr. William Krivit, the University of Minnesota Medical School, Minneapolis, Minnesota. Identical organ sections from three different normal controls were also obtained from Dr. Krivit. Glycosphingolipids used as reference standards for TLC were extracted and purified from packed human red blood cells by the method of Vance and Sweeley (26). All chemicals and solvents used in this work were analytical or reagent grade unless otherwise noted. Heat-activated Unisil (200/325 mesh, Clarkson Chemical Co., Inc., Williamsport, Pa.) was used in all silicic acid column chromatography. With the exception of the first Thomey kidney, the Thomey spleen, and the second Thomey liver glycolipid separations, all analyti- cal and preparative TLC was carried out on Quantum pre- coated TLC plates (Quantum Industries, Chicago, Ill.). The other TLC separations were done on glass plates coated to a thickness of 250p with Silica Gel G (Brink- mann Instruments, Inc., Westbury, New York) using a 38 39 Desaga TLC spreader. A F & M Model 402 gas chromatograph (Hewlett Packard Co., Avondale, Pa.) fitted with a glass, Six-foot 3% SE-30 or 0V-l column (packing from Applied Science Co., State College, Pa.) was used in all GLC analyses. Hexamethyldisilazane and trimethylchlorosilane used in the silylating (TMS) reagent and bis-trimethyl- silyltrifluoroacetamide (BSTFA) used in direct probe mass spectrometry (MS) analyses were purchased from Applied Science Co., State College, Pa. B. Methods Lipid Extraction and Purification Total lipids were extracted from all organs ac- cording to the method of Folch, gt_gl. (27). Glycosphingo- lipids were prepared from the total lipid extract by the method of Vance and Sweeley (26). In the majority of analyses, a modification of the Folch extraction was used. This procedure involved taking up the total lipids in 30-40 ml of CHC13, pouring the solution into a suitable length of dialysis tubing, and dialyzing against a large volume of distilled water for two days with several changes of water. Ideally, this method minimizes reten- tion of low molecular weight gangliosides in the organic phase, which often occurs in the conventional Folch pro- cedure (51). 4O TLC of Glycosphingolipids Two TLC systems were utilized in this work. For separation of neutral glycosphingolipids, the CHC13:Me0H:H20 100:42:6 (v/v) system was used with single development (26). Ganglioside separation was accomplished using the CHC13zMe0Hz2.5N NH3(aq) 60:40:9 (v/v) system with double development (52). Pre-coated TLC plates were used in the majority of separations after sharper bands and minimal overlapping were observed with their use. Preparation of TMS Glycosphingolipids The silylating agent consisted of dry pyridine, hexamethyldisilazane, and trimethylchlorosilane mixed in the ratio 5:2:1 (v/v) in that order. Generally, 50-100 pl of TMS reagent was added to 100-300 pg of glycosphingo- lipid and allowed to stand about 10 minutes before GLC injection. GLC of TMS Glycosphingolipids and Fatty AEid Methyl Esters Thomey and normal liver, kidney, and spleen TMS glycosphingolipids and fatty acid methyl esters (FAME) were run on a six-foot 3% OV-l column. A six-foot 3% SE-30 column was used in Thomey and normal brain analy- ses. All neutral TMS glycosphingolipids were isothermally eluted at 170°C; all TMS gangliosides were programmed from 41 150°C to 220°C at 20/min. All FAME were programmed from 180°C toz4o°C at 2°/m1n. Direct Probe Mass Spectrometry of TMS GlycosfihThgolipids The BSTFA derivatives of whole Thomey and normal glycosphingolipids were prepared and analyzed according to the method of Sweeley and Dawson (53). Only those glycosphingolipids which accumulated in Thomey's viscera and their normal counterparts were examined. An LKB 9000 mass spectrometer (LKB Produktor, Stockholm, Sweden) con- sisting of a gas chromatograph and a single-focusing 60O magnetic sector mass spectrometer directly coupled to Becker-Ryhage type molecular separators was employed for these analyses. The GC column used was a six-foot, coiled glass column packed with 3% 0V-l on 100-200 mesh, acid- washed, silanized Gas Chrom S. For direct probe analysis, the instrument was operated at 3500 volts and 70 eV elec- tron energy with a 60 uamp electron current and ion source temperature of 290°C. RESULTS A. TLC of Neutral Glycosphingolipids and Gangliosides Figures 5 through 13 illustrate the preparative TLC separations of neutral glycosphingolipids and gang- liosides from each type of tissue extracted. From 6 to 12 mg of material was applied to each plate. The shaded bands in the chromatograms of both Thomey and normal lipids are the bands of interest to be compared. For all neutral glycosphingolipid separations, the solvent system was CHCl :MeOH:H 0 100:42:6 (v/v); for all ganglioside 3 2 separations, the CHC13zMeOH:2.5N NH3(aq) 60:40:9 (v/v) system was used. Human red blood cell glycosphingolipids were used as reference standards for all neutral glyco- sphingolipid TLC separations. The identification of the major gangliosides on TLC, however, rested solely on the literature citations for the solvent system used, as suitable ganglioside reference standards were not avail- able. Verification of these assignments was accomplished via GLC analyses after acid-catalyzed methanolysis. 42 43 U‘l C \D C ‘ \9 C" 2‘ 3 '— Figure 5. TLC of Thomey kidney neutral glycosphingolipids. A. (1) Normal red blood cell gal-gal-glc-cer. B. (l) Glc-cer and gal-cer, (2) gal-gal-cer and gal-glc-cer, (3) $0?gal-glc-cer, (4) gal-gal-glc-cer,and (5) galNAc-gal- ga ~glc-cer. In FiguresESthrough 11, the solvent system was CHC13zMe0HzH20 100:42:6 (v/v). 44 4! / ’I/I Figure 6. TLC of normal human juvenile kidney neutral glyco- Sphingolipids: (l) glc-cer and gal-cer, (2) gal- al- cer and gal-glc-cer, (3) gal-gal-glc-cer, and (4? galNAc-gal-gal-glc-cer. 45 3C A .@g//fl///_/Z//_fl//7z///////////_/_//f/B 2 C — c: _A A 1 Figure 7. TLC of normal human juvenile spleen neutral glycosphingo- lipids. A. (l) Gaucher spleen glc-cer and (2) normal red blood cell gal-gal-glc-cer. B. (1) Glc-cer and gal-cer, (2) gal-glc-cer, (3 gal-gal-glc-cer, and (4) galNAc-gal- gal-glc-cer. 46 1 f A 2C: I x ’b O \ \ \ \ ‘ q 0 \ \ \\\‘\\\ I~Os‘1\~ \ \.\|\‘§ \\ \ ‘ \\‘\:\\ \ ‘ . ‘ 34‘ ‘ . ‘g ’ "“K ‘1”.‘ 1‘ \\ \“.‘ \\\\ ‘\§ \\\\‘ \’1\.‘.l,..“,“ ¢,\‘ C‘ ' \\ ‘ \ \\ \ \ \t\ \ \1\ 3 \ ‘\\\‘ \\ \ ‘5 ‘ \\ \|\ ‘\\\ ‘|Q \\\. “ bee‘ 9‘ \‘\Sn"’ ‘\\\\\ \ ‘\\10\"on\'\.‘1‘ 4Q/fl/1f/H //7/ Hg, 41/ ”gym Figure 8. TLC of Thomey liver neutral glycosphingolipids. A. (1) Normal red blood cell gal-gal- 1c-cer. B. (l) Glc-cer and al-cer, (2) gal-glc-cer, T3) galNAc-gal-glc-cer, and T4) galNAc-gal-gal-glc-cer. 47 1.\,°\\11M1\\H|I\Hb 3 ‘\\\ ‘\ o..)’.\‘..‘ D \\\\ H“ \‘\.. {.DH.".*‘I‘V‘\\ \I‘.\’\‘c‘..‘.\.0\33:‘ \ a.‘ 0 .I II \ H \ \ \\ \ \, \ | \\ l ‘ H. q. . .\ 0“ \ \.\‘ I‘\\ \‘0 ‘ f. .I \D 'IIPLL\1 5* 5 sa.-=====================:__ 4 W/I/lf/Zl/flfllim ,__:= Figure 9. TLC of normal human juvenile liver neutral 1 cosphingo- lipids. A. (l) Gaucher spleen glc- -cer and T2 normal red blood cell gal- gal- glc- cer. B. (1) Glc- -cer and 981- cer, (2) gal- glc- cer, (3 gal- 961- 910- cer, (4) galNAc- -gal- -ga1- glc- cer, and (5)G M3’ NANA- gal- glc- -cer. 48 4 Wf/i // 41/17 /f///7/ /j///b Figure 10. TLC of Thomey cerebral white matter neutral glycosphingo- lipids. A. (l) Gaucher spleen glc-cer, (2) normal red blood cell gal-glc-cer, and (3) normal red blood cell gal-gal-glc-cer. B. (l) Glc-cer and al-cer, (2) sulfa- tide, SO3gal-cer, and gal-glc-cer, (3T galNAc-gal-glc-cer, and (4) galNAc-gal-gal-glc-cer. 49 ‘7 ' J G 3— Figure 11. TLC of normal human juvenile cerebral white matter neutral (1) Gaucher spleen glc-cer and 1 cosphingolipids. TZT normal red blood cell gal-gal-glc-cer. A. (l) Gal-cer, (2) sulfatide, $03ga1-cer, and gal-glc-cer, and (3) gal- gal-glc-cer (tentative). 50 A B 1CD 16" i 4': 202: _,_____ 3D WW / 4 TT' *2:- 5C‘ -:3 of : A :1 7 S====ll=======llll==¥—e 44— f‘:2 8:: ~ 9 # Figure 12. TLC of Thomey cerebral white matter gangliosides. A. (l) T Gaucher spleen 1c-cer, (2) normal red blood cell gal-gal- ‘ glc-cer, and (3T normal red blood cell galNAc- al-gal-glc- cer. B.(1) Unclassified, (2) 0M3, (3) 5M2: (4T unclassified, (5) GM], (6) 301a, (7) 601b, and (8) GT]. Solvent system: CHC13:MeOH:2.5N NH3 (aq) 60:40:9 (v/v). 51 A B C 1% Z‘C:> 1 (_¥7 “":::::3 1c::) 3 b 2 V/J/7/LTIL/IITIII/IIIW 3T<===::====l==::, 77’ 44:; 4L :7 5L A 41— =7; 6 2:: CL_ 7 <:f ,#;:D Figure 13. A. (l) Gaucher spleen glc-cer, (2) gal-gal-glc-cer, and (3) normal red blood cell gal—glc-cer. B. (l) Unclassified, (2) G 4) GM], (5) 001a, (6) 601b, and (7) GT1- Co GMz. Solvent system: TLC of normal juvenile cerebral white matter gangliosides. normal red blood cell galNAc-gal- M2’ (3) unclassified, (1) Thomey CHCl3zMeOH:2.5N NH3 (aq) 60:40:9 (v/v). 52 B. GLC of TMS Neutral Glycosphingolipids, Gangliosides, and FAME Tables 2A through 10 Show the levels of neutral glycosphingolipids and gangliosides in the tissues examined expressed as umole hexose/gm wet weight of tis- sue. Prior to overnight acid-catalyzed methanolysis at 82°C, 0.1 or 0.2 umole of a mannitol standard solution (2.0 umole/ml, in MeOH) was added to each fraction to insure quantitation of material. Areas of GLC peaks were calculated via planimetry and the amounts of glucose (glc), galactose (gal), N-acetylgalactosamine (galNAc), and N-acetylneuraminic acid (NANA) were calculated from the following formulas: x 1.25 x umoles mannitol added = umoles (1) area glc or gal area mannitol glc or gal (2) area galNAc . 4 x 1.36 x umoles mann1tol added = umoles area mann1th galNAc (3) area NANA . . x 0.98 x umoles mann1tol added = pmoles area mann1tol NANA To carry out the re-N-acetylation of galNAc and NANA, about 0.2 m1 of acetic anhydride was added to each FAME-extracted, neutralized methanolysate suspected of containing galNAc or NANA and allowed to stand in a closed culture tube a minimum of six hours at room temperature. Silver carbonate used in the neutralization was allowed to remain in the methanolysate during re-N-acetylation, 53 .mzmmwp mo agave: um: Em\mw_oea pamFm>P=cm cm>wm mm; cowmcm>coo ”Amm.mmv mm.o x mmzmmwp Amcuwx Fasco: we “swam; Ace Em\mmposn mew mm=Pm> omen»; .umcwsgmpmu pozm .Acmu-o_mupem-_mmrowpmpcmp .Acmo-u_m-pmmmomv muwume_:m _zmoxmgwu n JADmu .Acmoupmm-_wm new emoropm-_mmv muesmcmo Famoxmgwu n ozoo .Aemo-_mm use emu-u_mv mu_Eequ _xmoxmzocos n 0123 .msmmwu mo ucmwmz pm; Em\Apmm co upmv wmoxm; m—oE: mm cm>wm mew mmzrm> .mwpamp cmgpo PPM use wasp :Hm m¢.o Nm.o mo.o op.o N_.o scammeoeeaz m8.o wo.~ __.oneo.o mm._ Fo.onm_.o om.F _o.oemo.o F¢.m No.oem_.o mm.o eo.oe~_.o momeeo>< 88.0 mm.P 65.8 mm._ m_.o e_.F mo.o om.o mp.o mm.o op.o HH No.0 mm._ mm.o _m._ mp.o 8m._ mo.o mN.N N_.o mm.o ~_.o H e we.o no.N mm.o oo.~ __.o mm._ mo.o “a.“ ~_.o mm.o mo.o >H - - m- - a- - 6. mm.@ m_.o Ne.o mo.o HHH a- 8N.N me.o mm., __.o - - NN.o_ .F.o _m.o 8N.o HH a- ON.N we.o mm._ N_.o - - mm.m o~.o om.o op.o H _ o_o o_w o.o o_o ope o_u . . \omcuw¥ mesogp Lo muwawpoucwcqmooxpw Pmcuzmzrr.og .oeo .QIQ .0126 mm.o ¢o.m mm.~ ©¢.F w~.¢ mv.o mm.omm©.m ¢N.owoo.o mo.OHm—.o mo.OHm¢.o o—.ow¢m.o mmmmsm>< ow.o mo.N om.m OF.N Nm.o vm.~ NF.o mm.m mm.o ©¢.o NN.O HH mm.o mo.N mm.m mm.— mm.o 1 1 mN.m m¢.o _o.o ¢¢.o H m ¢@.o mm.F mm.m cm.P mm.o mm.p m_.o mm.m mm.o om.o mw.o HH mm.o om.~ mo.¢ om.P No.0 1 1 mm.N ~©.o mm.o ¢¢.o H N 0P0 ope upw upu upw UFG . . .xmcnwx zmeoch mo muwawpomcwsamooapo _mgpaw211.mm mpnmp 55 .mpnmu ecu cw umgcmmmca mm=Fm> ecu m>wm op mko on umucm use «#30 umuumgp unzm .umumpsupmu we: uxmp cw mOJw mo pcaosm asp .uth cw ucmmmea umc .upm .ozo .uzzm mm.o oo.N P¢.P mm.o 1 mm.omvo.m No.omwp.o Fo.omm—.o No.omom.o mmmmgm>< Pm.o eo.m mN.N uwm.F v—.o oo.~ _P.o n1 mm.o N mm.o mm._ Nw.F umv._ P_.o om.o vp.o n1 mm.o F upw\uFw —mcp:mz-r.m «Fame .msmmep mo pemwmz pm: Em\o_m «Foe: mp.p we: Hewempes we e_m_> Heeme meo mm emeemome moeu wee uzmh new mpepa UHF mueEmEoe e eo meow we: eowpeeeamm mweho . mo p He .AHeo emurupm-_emtu< «5.0 Hm.H Ne.o He.o eo.H Nm.H em.o oH.o No.0 OH.o HH Hw.o mm.H mm.o Hm.o mo.H 8H.H HH.H NH.o mm.o mH.o H e Hm.o Ho.m e~.o em.o eo.H mm.H em.o mo.o me.o mH.o H m - - o- - - o- oH.H Ho.o om.o oH.o H N am.o 8H.H NN.o mm.o eo.H me.o eo.H eo.o - e- H H oHe oHe oHe oHe oHe oHe . . \oezHee \Hee mOHe \oezHee \Hee agree \Hae gee \Hee 0:: oz Hoe .ew>we xmsoeh Lo meanFome_eamouxFu Feeuemz--.¢ mpnep .< NH.0 00.0 0H.0 00.H HN.0 00.0 00.0 00.0 00.0 0 0H.0 0H.H 00.0 ~0.H 00.0 0e.0 HN.0 H0.0 00.0 m 0H.0 00.H 00.0 00.H 00.0 00.0 00.0 00.0 0H.0 N e- 0H.N 00.0 H0.~ 00.0 0e.~ 00.0 00.0 00.0 H oH0\oezHe0 oH0\He0 00H0 oH0\He0 0:00 oH0\H00 0:0 oH0\He0 0:: Hopmno0 .zmeewx m—wem>=0 e050: Feeeoz mo mewawpomeweamouxpo Heepemz-1.m 0000» 58 .pemmmeq 003 0H0 x—eoe 00.0 NN.H 00.H 00.H 00.0 N0.0000.0 00.0000.0 00.0n0H.0 N0.0eNH.0 000eeo>< 00.0 00.H HH.0 00.H 0H.0 0H.H 0H.0 00.0 00.0 m 0H.0 00.H 00.0 00.H 00.0 00.H 00.0 NH.0 00.0 N N0.0 N0.H 00.0 H0.H 00.0 00.H 0N.0 0- NN.0 H oH0\oza ewes: Heeeoz Ho mewawHomeweamouan Heepemz--.o mHemh 59 .Hemmmee 003 0H0 xHeon .xHeo emuruHmrH0mrp0m n 0:000 H0.0 00.0 00.0 N0.H 00.H HH.H 00.0 H0.0veH0.0 H0.0000.0 H0.0000.0 N0.0e0H.0 H0.0h00g0000eeo>< 00.0 0N.H H0.0v H0.0 N0.H N0.0 0H.N 00.0 0H.H 00.0 0H.0 00.0 0 00.0 H0.0 H0.0v 00.0 00.N 00.0 00.H 00.0 00.H NH.0 0- No.0 N 00.0 0N.H H0.0 00.0 00.H N0.0 00.H 00.0 NH.H HH.0 0- 00.0 H oH0 0H0 02 0H0 oH0 0H0 oH0 0H0 .02 \0202 \H00 0 \oezH00 \H00 0000 \H00 00:0H \H00 000 \H00 0:: HoeHeo0 » .em>04 wHwem>00 e050: H0Eeoz mo mewawHooeHeemoowa H0epemzuu.N 0H00H 60 .000 :o m_00>Ho0m::: mem3 00:00 oz: 0Poeueou H0500: 000 .Aeow0e .emurF0mmomV m00000H00 0:0 Aeoews .00000000 000 003 00000 .0m0ompm0 go: 003 u:0 005002 0:0 .empu0z mpwez xmsoe» mo 00000Homeweq0ouz_w _0euemz--.m mH00H 61 .uem0000 003 upm >Feon .Amev Em00x0.0em>_o0 mE00 men :0 0m00000Fme00 F000memu 00 000 00 000000 0me00_0=0 0003 :000e0eeou 00 00000 mgu :o m00E mem3 0m0meueme00 e0 memnsze Hm0ez 0:0000emw0m0 .ewmweo 00m: on :300 mew0em0 1xm 0:0 peoem 0em>Ho0 00m: meweewmmn .000 :0 00:00 mo moemzcm0 pem0meame 00000020 A—huv mo.— 00.0 mm._ H—.o 0 2:000 2:000 _0.F 00.0 mm.F om.o m om._ 00.0 00.0 00.0 m m 0 H H000 H H000 m0.F 00.0 00.N 00.0 m m_.~ 00.0 00.0 mm.o 0 :03 0:00: 00.0 mm.o mo.~ mm.o n 00.0 00.0 mm.~ om.o m 0N000 0N000 00.0 m00ep 0~._ 00.0 0 00.0 00.0 op._ m0.m m m2 m2 0 3 0 e H0.0 mu0ep 00.0 —o.o m mm.o 1 Fn.o 00.0 _ mm.o m00ep 00.0 no.0 0 1 1 01 Po 0 m 1 1 01 Po 0 N 1 1 mm.o Po.o F 0~0\00 _0Eeoz 0:0 xmsoeh11.m mpn0» 62 .0 mH00L :0 00 mE00 me“ 000 0:0000em_0m0 0:0 0:0:m05020 0:00 0:00 00._ 00.0 Nm.F 00.0 N 00.0 00.0 00.0 No.0 0 “00000 “00000 00.0 m0.o m0.— No.0 0 mo._ Nw.o 00.0 00.0 N m w A Fowv A 0000 _0.— 00.0 mo.N m—.o m 00.0 00.0 00.0 00.0 0 :03 1000 _n.0 NN.0 00.N 00.0 0 00.0 No.0 Nn.— 00.0 m 00.0 mm.o NF.— 00.0 0 ANzwv H0.0 mu0ep N_._ _o.o m 00.0 00.0 mo.F No.0 m ANzwv Amzwv mm.o m0.o 00.0 —0.0 N 00.0 m0000 NN.0 No.0 N 00000 1 ~0.F _o.ov _ mu0eu 1 00.0 00.0 F 0F0\a0 P0Eeoz 0:0 0050:011.o_ mpn0p 63 since it has been observed to catalytically promote re-N- acetylation by an unknown mechanism (54). Sugar ratios given in the tables are all relative to glc and were found by taking ratios of the umolar amounts of the moieties present. In the neutral glycosphingolipid deter- minations, bands appearing on TLC closer to the origin than globoside sometimes contained NANA, in addition to glc and/or gal. Only in the normal liver did NANA con- sistently appear and, considering the gal/glc ratios, this substance was probably hematoside. No attempt was made to characterize other bands appearing lower than globoside in separations where the compound could not be identified as hematoside from the calculated ratios. Where available, values from the literature for visceral and cerebral glycosphingolipids are included in the tables. The Thomey kidney data deserve special mention. First, the discrepancy between extractions 1 and 4, and 2 and 3, is most likely due to an error in weighing the wet tissue, Since the glycosphingolipid amounts in both sets of data are proportional. These extractions were all done in the same manner and a weighing error there- fore appears to be the best explanation for this discrep- ancy. Second, the gal/glc values for trihexosyl ceramide indicated the presence of two glyCOSphingolipids. A separation was achieved after the total glycosphingolipids 64 had been standing in CM 2:1 for several days or more and then applied to a TLC plate. When the glycosphingolipids were run immediately after mild alkali-catalyzed metha- nolysis, however, only one band was observed for trihexosyl ceramide. The gal/glc ratio for this late-separating gly- cosphingolipid indicates that it is probably the dihexosyl sulfatide isolated from human kidney by Martensson (55). However, due to some overlap of trihexosyl ceramide, the gal/glc ratio does not Show equimolar amounts of gal and glc. Similar separation problems were not encountered with other visceral and cerebral glycosphingolipids. Tables 11 and 12 illustrate the kinds of FAME present in the visceral and cerebral glycosphingolipids isolated. Values are given as percent of total FAME present (total area of all FAME GLC peaks). Where avail- able, values from the literature are presented for com- parison. Identifications of FAME were made on the basis of retention times of a standard mixture of 16:0, 18:0, 20:0, 22:0, and 24:0 FAME and by "carbon-number" (56). C. Mass Spectrometry of TMS Glycosphingolipids The mass spectra shown in Figures 14 and 15 veri- fy that Thomey kidney globoside is the same compound as that of normal kidney, due to the presence of a terminal galNAc moiety. The presence of galNAc is indicated by 65 .Aemo1uHm1H0m1H0mv m0HE0emu H>0oxmeH00 .Aem01uHm1H0a1o0oxmeH00 .H000H0 00N .0HH .0000 .mHQE00 em>Hm 0 :0 Hem0mea mz<0 .0xeq0H0 .EHeuon ..0 .eo00emHe0z .011L 0 0=0Ho 0 H0000 we Hemoema Hem0meam: 0m0H0>0 0.N0 N.0 H.HH 0.0H 0.00 0.N0 0.NN N.0N HN0N N.0N N.0N 0.N N.00 0.0N 0.0N 0.H0 0.0N 0H0N H.0 0.0H 0.H 0.0 N.0 0.N H.0 N.0 000N N.0 - - - H.H 0.H 0.0 - HHNN 0.HH H.0N 0.0 0.HN 0.0H 0.0N N.NN 0.0N 0HNN H.0 N.0 N.N H.0 H.0 0.0 N.0 0.0 0u0N 0.N N.N N.0 0.H 0.H 0.0 H.H 0.0 HH0H 0.0 H.0 0.00 N.N 0.0 N.0 0.0 0.0 0H0H 0.0 - - - - 1 - - HU0H 0.N 0.NH N.0 0.0 0.0 0.0 H.H 0.0 0H0H H.000 N0 H.000 00 00H0000H0 0000+ 0000+ 00H0000_0 00H0000H0 000H0000H0 00H000oH0 00H0000H0 0:00 emsH0 z 00>H0 z 00>H0 H 000H00 2 000H00 H 0000H: 2 0000H0 z 0000H0 H 000H0HHomeHee0oo>H0 H0em00H> mHHem>00 H0Eeoz meH0eoe0meeo0 0:0 H0seoen< xmsoeH 0000 mz<011.HH mH00H 66 .Amompv mo— .oo ..Em:uowm .6 ..z .mgssmzmx v20 ..H .wEOmemho .Agmu-uHm-H0muuHm 0 :0 mz0 0.0 - - - - - - H000 0.0 - - - - - - 0000 - - - - - - - 0000 0.H - - - - - - H000 0.0 - - - 0.H - - 0000 0.0 0.0 0.0 H.0 0.0 0.0 0.0 0H00 0.0H 0.0 0.H H.H 0.0 0.0 0.0 HH0H 0.00 0.H0 0.00 0.00 H.00 0.H0 0.H0 0H0H 0.H - 0.0 H.H - - - HH0H 0.HH 0.0 0.0 0.0 0.0 0.0 0.0 0H0H 0000000H0 0:0 0:0 020 00:0H m000H M000H 0HMHW0¢H 000000-00H 0HMHW00H 0MWW00£ 000000-00H 00MHH00H 00ww0wfl 0200 0.0uHQHHomchqmooxHo ngnmgmu mcummuzmH 0:0 xoeosH EoLH mz<0--.~H mHamH 67 .LmuuupmupmmupomuuHH zmsogH Ho Ezguumam 0002 Orb omw 0mm _h__hhhhpb — _ 0mm on _ _PPPkPPH me _ 00m om¢ om¢ rPHPLHLF~F0FLP oru . h H H ohm _ L om H m 0mm 0mm _ _ __ _; 33:: A:XII\ _ H _ 0 FF—_____ =«_J:041- .0. 000000 ofiN OFH 00H om Om 0 :ON 00¢ ow 0 :0m mJQAHF 69 an ion at m/e 173 and a terminal galNAc exhibits a strong ion at m/e 420 (53). Both of these ions are present in the spectra of Thomey and normal globoside. The mass spectrum of Thomey liver trihexosyl ceramide confirms this lipid to be asialo ganglioside GMZ’ since the ions at m/e 173 and 420 indicate a terminal galNAc. DISCUSSION Quantitative determinations of visceral and cere- bral glycosphingolipids in either disease or normal states are scarce in the literature and therefore few comparisons can be made with the data obtained in this study. Martens- son's investigation of normal kidney glycosphingolipids (57), however, was very thorough and the data for Thomey kidney and the control kidneys are in close agreement with his values. The glycosphingolipid levels in Martensson's study were originally given as umole/gm dry weight of tis- sue. Assuming visceral tissues to be composed of 75% water on the average (45), multiplication of Martensson's values by 0.25 transforms the data to umole/gm wet weight of tissue. The Thomey data showed an elevation in globo- side and a depression in trihexosyl ceramide (gal-gal-glc- car) as compared with normal levels. This finding is consistent with a metabolic block at the site of globo- side catabolism to gal-gal-glc-cer, assuming that gal- gal-glc-cer is derived from globoside via enzymatic cleavage of the terminal B-N-acetylgalactosamine moiety. A similar result was obtained with spleen from Thomey. The depression in the amount of gal-gal-glc-cer was not observed here, but there was an enormous 7O 7l twenty-fivefold increase in the amount of globoside pres- ent. Svennerholm and Svennerholm (58) reported that normal spleen consisted chiefly of lactosyl ceramide (gal-glc-cer, 50-60%); globoside was present only to the extent of l2-18% of the total glycosphingolipids. The three normal controls corresponded closely to the Svenner- holms' criteria, but the Thomey spleen contained an over- whelming proportion of globoside. Results with liver were somewhat different. In the same study, Svennerholm and Svennerholm (58) reported that the dominant glycosphingolipid of normal liver was gal-glc-cer (65-75%), while gal-gal-glc-cer and globoside were present in far lesser amounts, 7-lO% and 8-l2%, respectively. Levels for the normal controls corroborated these findings, but Thomey liver exhibited a thirtyfold increase in galNAc-gal-glc-cer and an elevenfold increase in globoside. Normal gal-gal-glc-cer was not found in Thomey's liver. More significant, however, is the fact that Thomey galNAc-gal-glc-cer was found via GLC and MS analysis to be asialo ganglioside GMZ' This finding is in accord with a diagnosiscH’Tay-Sachs disease with vis- ceral involvement. Gregoire, gt_gl. (50) had previously reported the storage of asialo GMZ in the liver in a case of Tay-Sachs disease with visceral involvement. In all normal and Thomey visceral tissues studied, no gangliosides were detected in the aqueous phases of the 72 Folch washes or the dialysis-modified Folch procedure. In the Thomey kidney and spleen there was a trace of globoside observed on TLC of the ly0philized material from the Folch upper phase. Due to the excess of globo- side in these tissues, traces of this compound partitioned into the aqueous methanolic Folch upper phase. Eeg-OlofS- 5; son, gt_al. (44) and Suzuki, e£_al. (45), however, reported Lw1 ganglioside GM2 accumulation in the livers and spleens of their Tay-Sachs patients. Taketomi and Kawamura (49), on [‘1 the other hand, did not detect ganglioside G storage in M2 the viscera of their Tay-Sachs patient. These differences might be attributed to variations in methodology, since it seems inconsistent that asialo GMZ would accumulate to such an extent in the liver, for example, without at least some storage of ganglioside GMZ occurring as well. More careful ganglioside extraction and assay procedures would perhaps reveal the presence of ganglioside GM2 in the vis- ceral tissues. Alternatively, these storage differences might be due to genetic variations that are not yet understood. Analysis of Thomey cerebral grey and white matter revealed the classical storage pattern for Tay-Sachs ganglioside. Asialo GMZ dominated the mixture of neutral glycosphingolipids in grey matter and was also present to a significant extent in the glycosphingolipids of white matter. The ganglioside GMZ was present in ten- to 73 twentyfold excess over the other grey and white matter gangliosides found in the aqueous phase of the modified Folch extracts. In addition, glucocerebroside (glc-cer) was detected to a significant extent in Thomey grey mat- ter and the levels of gal-cer and sulfatide (SO3gal-cer) were severely depressed in Thomey white matter. Gal-glc- cer was present to a minor extent in both Thomey and nor- mal grey and white matter. These observations are con- sistent with previous studies on neutral glycosphingolipids of Tay-Sachs brain (44,49,59) and of normal brain (45,49). The galNAc/glc and NANA/glc ratios found in these studies deserve special comment. All of these ratios were consistently low and were especially so in the case of the brain polysialo gangliosides. The consistency of these low ratios probably indicates an inadequate re-N- acetylation procedure. It has been the practice in our laboratory to add some (93, 0.21nl)acetic anhydride to the neutralized methanolysate in the presence of AgZCO3 and let the re-N-acetylation mixture stand for a minimum of 6 hours up to a maximum of 24 hours. From the vari- ability of the data, it is obvious that this procedure will have to be made more rigorous if reliable results are to be obtained. Although the mechanism of AgZCO3 catalysis is not known, optimal acetic anhydride concen- tration and reaction time could be achieved via re-N- acetylation experiments with known micromolar amounts of 74 N-acetylneuraminic acid. The standardization of this procedure would greatly facilitate the accurate GLC analysis and identification of hexosamine- and NANA- containing glycosphingolipids. The investigations of the glycosphingolipid and ganglioside content of Thomey brain and visceral organs, coupled with identical studies on normal controls, have led to an accurate identification of Thomey's lipid storage disorder. An examination of the data accumulated conclusively establishes Todd Thomey as a case of Tay- Sachs disease with gross visceral involvement. Three glycosphingolipids were found to accumulate in excessive amounts: (1) normal kidney globoside (galNAc-gal-gal- glc-cer) in the visceral organs, (2) asialo ganglioside GMZ (galNAc-gal-glc-cer) in the brain and liver, and (3) ganglioside GM2 (galNAc-gal-glc-cer) in the brain. These NANA three compounds share a common structural feature: all possess a terminal B-N-acetylgalactosamine moiety (8,24). (The presence of a terminal 8 linkage was not proven, however, for the compounds isolated in this study.) Thus, it appears that Todd Thomey is a case of Tay-Sachs disease with visceral involvement very similar, if not identical, to the case described by Sandhoff, gt_gl. (46,47). In that study, an absence of the enzyme B-N-acetylhexosamini- dase was postulated as the cause of the excessive kidney 75 globoside accumulation in the visceral organs and of the accumulation of ganglioside GMZ and its asialo derivative in the CNS. Classical Tay-Sachs patients exhibited nor- mal B-N-acetylhexosaminidase activity; normal neuramini- dase activity was observed in both the exceptional and classical cases. In a later publication, Sandhoff (60) investigated this enzyme deficiency more closely and found that B-N-acetylhexosaminidase consisted of two main fractions with isoelectric points at pH 7.3 and 5.0. In classical cases of Tay-Sachs disease, patients ex- hibited a lack of the fraction with isoelectric point at pH 5.0 in brain and liver, but the activity of the pH 7.3 fraction was elevated three- to fourfold in the brains of these patients. The exceptional case of Tay- Sachs disease, however, exhibited a complete absence of both enzyme components. These preliminary data would suggest a correlation of kidney globoside storagewith the absence of the pH 7.3 fraction of B-N-acetylhexosami- nidase, but sufficient evidence has not yet been obtained to warrant such a conclusion. At about the same time, Okada and O'Brien (6l) in a study of B-N-acetylhexosamini- dase activity in Tay-Sachs patients and normal controls found this same dual-component activity in the controls and labeled these components A and B. These investigators also found an absence of component A in all Tay-Sachs patients studied, thus paralleling Sandhoff's work (60). 76 These findings suggest that B-N-acetylhexosaminidase assays of Thomey cerebral and visceral tissues would have been of immense value in clarifying the distinction between classical lay-Sachs disease and Tay-Sachs disease with visceral storage. The data in this study, however, unequivocally identify Thomey as a case of Tay-Sachs disease with visceral involvement and strongly suggest an identity with the exceptional case of Sandhoff, g£_al. (46,47). Full clarification, however, must await careful enzymatic analysis and characterization of the remaining tissues. SUMMARY Glycosphingolipids were isolated from cerebral grey and white matter, kidney, spleen, and liver of Todd Thomey, an lB-month old child who died from a lipid storage disorder diagnosed as classical Tay-Sachs disease. Three different glycosphingolipids accumulated in exces- sive amounts: (1) normal kidney globoside (galNAc-gal- gal-glc-cer) in the visceral organs, (2) asialo ganglio- side GM2 (galNAc-gal-glc-cer) in the brain and liver, and (3) ganglioside GMZ (galNAc-gal-glc-cer) in the brain. NANA These findings classified Thomey as an exceptional case of Tay-Sachs disease with visceral involvement. A strik- ing similarity was noted between this case and a case reported by Sandhoff, et al. (46,47) in which a deficiency of B-N-acetylhexosaminidase was cited as the cause for glycosphingolipid storage in the brain and visceral organs. 77 10. ll. 12. l3. l4. BIBLIOGRAPHY Nakayama, T., J. Biochem., 31, 309 (1950). Fujino, Y., and Negishi, T., Bull. Agr. Chem. Soc. Japan, 4, 183 (1956). Rosenberg, A., and Chargaff, E., J. Biol. Chem., 233, 1323 (1958). Carter, H. E., Rothfus Lipid Res., 2, 228 J. A., and Gigg, R., J. 961). (i L., and Yariv, J., Arch. 92, 438 (1961). Rapport, M. M., Graf, Biochem. Biophys., Yamakawa, T., Kiso, N. Shizuo, H., Makita, , and A. Yokoyama, S., J. Biochem., 52, 226 (1962). J. M Yamakawa, T., Yokoyama, S., and Handa, N., Bio- chem., 53, 28 (1963). Makita, A., Iwanaga, M., and Yamakawa, T., J. Bio- chem., _5, 202 (1964). Yamakawa, T., Nishimura, S., and Kamimura, M., Japan. J. Exptl. Med., 35, 201 (1965). Rapport, M. M., and Graf. L., Fed. Proc., 24, 176 (1965). Rapport, M. M., Graf, L., and Schneider, H., Arch. Biochem. Biophys., 105, 431 (1964). Makita, A., Suzuki, C., and Yosizawa, Z., J. Biochem., 60, 502 (1966). Sweeley, C. C., and Klionsky, B., J. Biol. Chem., 238, PC3148 (1963). Sweeley, C. C., and Klionsky, B., Abstracts Sixth International Congress of Biochemistry, New York (1964). 78 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 79 Makita, A., and Yamakawa, T., J. Biochem., £5, 365 (1964). Gray, G. M., Biochem. J., 94, 91 (1965). Adams, E. P., and Gray, G. M., Chem. Phys. Lipids, l, 368 (1967). Miyatake, T., Handa, S., and Yamakawa, T., Japan J. Exptl. Med., 48, 135 (1968). Adams, E. P., and Gray, G. M., Chem. Phys. Lipids, 2, 147 (1968). Kawanami, J., J. Biochem., g4, 625 (1968). Eto, T., Ichikawa, Y., Nishimura, K., Ando, 5., and Yamakawa, T., J. Biochem., 64, 205 (1968). Gatt, S . and Rapport, M. M., Biochem. J., 101, 680 (1966). Gatt, 5., Biochem. J., 101, 687 (1966). Frohwein Y. Z., and Gatt, 5., Biochem., g, 2783 (1967). _—'_' Kawanami, J., J. Biochem., 2, 105 (1967). Vance, D. E., and Sweeley, C. C., J. Lipid Res., 8, 621 (1967). Folch, J., Lees, M., and Sloan Stanley, G. H., J; Biol. Chem., 226, 497 (1957). Lands, N. E. M., and Dean, C. S., J. Lipid Res., 3, 129 (1962). Hakomori, S., J. Lipid Res., 1, 789 (1966). Casu, B., Reggiani, M., Gallo, G. G., and Vivegani, A., Tetrahedron, 24, 803 (1968). Kamerling, J. P., Rosenberg, D., and Vliegenhart, J. F. G., Biochem. Biophys. Res. Comm., 38, 794 (1970). Makita, A., J. Biochem., pg, 269 (1964). 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 80 Emsley, J. N., Feeney, J., and Sutcliffe, L. H., High Resolution Nuclear Magnetic Resonance Spec- troscopy, Vol. 2, Pergamon Press, Oxford, 700 van der Veen, J. M., J. Organic Chem., 28, 564 (1963). Horton, D., and Turner, N. N., J. Organic Chem., 88, 3387 (1965). Kint, J. A., Science, 1 7, 1268 (1970). Ledeen, R., and Salsman, K., Biochem., 4, 2225 (1965). Fredrickson, D. S., and Trams, E. G., in The Meta- bolic Basis of Inherited Disease, J. B. Stanbury, J. BI Wyngaarden, and 0. SI Fredrickson, Eds., McGraw-Hill, New York, 523 (1966). Turban, H., Thesis, Freiburg, i/Br (1944). Norman, R. M., Urich, H., Tingey, A. H., and Good- body, R. A., J. Path. Bact., 28, 409 (1959). Norman, R. M., Tingey, A. H., Newman, C. G. H., and Hard, S. P., Arch. Dis. Chil., 88, 634 (196 ). Svennerholm, L., and Raal, A., Biochim. Biophys. Acta, 88, 422 (1961). Makita, A., and Yamakawa, T., Japan. J. Exptl. Med., 88, 361 (1963). Eeg-Olofsson, 0., Kristensson, K., Sourander, and Svennerholm, L., Acta Paediat..Scand., 546 (1966). P., 55, Suzuki, K., Suzuki, K., and Kamoshita, S., 82 Neuropath. Exptl. Neurol., 28, 25 (1969). Sandhoff, K., Andreae, U., and Jatzkewitz, H., Path. Europ., 8, 278 (1968). fe —lo Sandhoff, K., Andreae, U., and Jatzkewitz, H., L Sci., 2, 283 (1968). Pilz, H., MUller, D., Sandhoff, K., and ter Meulen, V., Deutsche Med. Noch., 88, 1833 (1968). Taketomi, T., and Kawamura, N., J. Biochem., 88, 165 (1969). 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 81 Gregoire, P. and Cape , Jonniaux, G. , Loeb, H., Voet, N., Fran. D'Etude Clin. Biol., E. e, R., Rev. 9 11 14, 568 (1 69). Ledeen, R., J. Am. Oil Chem. Soc., 48, 57 (1965). Wherrett, J. R., and Cummings, J. N., Biochem. J., 88, 378 (1963). Sweeley, C. C., and Dawson, G. Res. Comm., 81, 6 (1969). Dawson, G., unpublished work. Martensson, E., Biochim. , Biochem. Biophys. Biophys. Acta, 116, 521 (1966). Noodford, F. P., and van Gent, C. M., J. Lipid Res. 1. 188 (1960). Martensson, E., Biochim. Biophys. Acta, 116, 296 (1966). Svennerholm, E., and Svennerholm, L., Nature, 198, 688 (1963). Suzuki, K., and Chen, G. (1967). C J. Lipid Res., 8, 165 Sandhoff, K., FEBS Letters, 4, 351 (1969). Okada, 5., and O'Brien, J. (1969). S. , Science, 165, 698 HICHIGRN STRTE UNIV. LIBRRRIES |||||| lllll ||||||||| Ill llll lllllll II III 9 1 1 312 3 075 5664