Eur. J. Biochem. 64, 373-380 (1976)
Immunochemical Detection of the Thomsen-Friedenreich Antigen (T-antigen) on the Pig Lymphocyte Plasma Membrane Roland A. NEWMAN, Wolfgang M. GLOCKNER, and Gerhard G. UHLENBRUCK Department of Immunobiology, Medical University Clinic, Cologne (Received October 20, 1975/January 26, 1976)
A method for the isolation of lymphocytes from pig peripheral blood and preparation of plasma membranes was developed. The method resulted in a ten-fold increase in 5’-nucleotidase activity and neuraminic acid, relative to protein. A cholesterol/phospholipid ratio of 0.93 was obtained. Alkaline borohydride treatment of isolated plasma membrane, after desialylation, released the disaccharide ~-D-galactosyl-(l+3)-N-acetyl-~-galactosaminitol (the immunodominant group of the T-antigen), which was identified by gas chromatography using a crystalline standard. The disaccharide was not found without prior desialylation, indicating that all the disaccharide units were substituted by neuraminic acid. Serological evidence using the agglutinins from Arachis hypogoea and Helix pomatia confirmed the presence of this disaccharide (T-antigen) in the lymphocyte membrane and indicated that it was the major alkali-labile oligosaccharide of the pig lymphocyte membrane.
It is generally accepted that the cell surface structure determines many functional properties within the cell, as well as reactions between itself and its environment or neighbour. Earlier investigations by our group have already demonstrated that human T and B lymphocytes can be differentiated by the use of the agglutinin from Helix pomatia, which reacts specifically with neuraminidase-treated human T cells [l]. This finding was also observed by Hammarstrom [2]. During the course of this work we have continued a line of research to investigate the presence of neuraminic acid-covered cryptic antigens which, if present, can be used as secondary markers for the characterisation of different types of cells. Different authors, using serological methods, have suggested the presence of receptors for anti-T (Thomsen-Friedenreich phenomenon) antibodies on the surface of lymphocytes from human peripheral blood after treatment with neuraminidase [3,4]. As the immunodominant group of the T-antigen is a disaccharide of defined structure [ 5 ] ,we decided to develop a method to demonstrate the presence of this disacEnzymes. 5’-Ribonucleotide phosphohydrolase (EC 3.1.3.5); glucose-6-phosphatase (EC 3.1.3.9); succinic oxidoreductase (EC 1.3.99.1); orthophosphate monoester phosphohydrolase (EC 3.1.3.2).
charide, by chemical means, on the surface of lymphocyte membranes. In order to obtain sufficient amounts of lymphocyte plasma membranes we isolated lymphocytes from the peripheral blood of pigs. Other authors also have published methods for the isolation of lymphocyte plasma membranes [6,7] but from pig mesenteric lymph nodes, which are known to contain a different subpopulation of lymphocytes to that of peripheral blood [S]. In future, chemical characterisation methods will be applied to different subpopulations of normal and leukaemic human lymphocyte plasma membranes.
MATERIALS AND METHODS
Materials Blood was obtained from freshly slaughtered pigs and prevented from clotting by the addition of sodium citrate solution (1 vol. 3 . 8 % sodium citrate solution to 9 vol. of blood). Dextran T-250 and Ficoll were obtained from Pharmackd Ltd (Uppsala, Sweden). Uromiro-300, a 65% aqueous solution of the methyl glucamine salt of acetylamino-methylacetylitmino triiodobenzoic acid, was obtained from D r Fwnz Kohler-Chemie (Alsbach, West Germany). Enzyme substrates, glucose-6-
374
phosphate, adenosine monophosphate sodium salt, glycerol 3-phosphate and dichlorophenol-indophenol, as well as trypan blue for viability staining, were obtained from Serva Ltd (Heidelberg). Crystalline P-D-galactosyl-(l+ 3)-N-acetyl-~-galactosamine isolated from human brain gangliosides [9] was a gift from Prof. W. Gielen (Pharmakologisches Tnstitut der Universitat Koln). All other general reagents were obtained from Merck Ltd (Darmstadt). Arachis hypogoea lectin was prepared as previously described by Dahr et al. [lo]. Enzyme Determinations
5'-Nucleotidase was used as a marker for plasma membrane and glucose-6-phosphatase for endoplasmic reticulum. Both enzymes were assayed as described by Aronson and Touster [I 11. Succinic oxidoreductase was assayed as described by Earl and Korner [12] as a mitochondria1 enzyme and acid phosphatase as a marker for lysosomes [13].
Analyticd Procedures Neuraminic Acid. Total neuraminic acid, released by hydrolysis in 0.1 M H,SO, for 1 h at 80 "C, was assayed by the method of Aminoff [14] and corrections made for any interfering nucleic acids as described by Warren [15]. Lipids. Cholesterol and phospholipids were determined after chloroform/methanol extraction as described byzlatkisetal, [16] and Lowry [I71 respectively. Phosphate. Free phosphate after enzymatic release was determined by the method of Chen et al. [18]. Protein. Protein was measured using the FolinCiocalteau reagent as described by Lowry [19]. Hexoses and Hexosamines. Individual hexoses and hexosamines in plasma membrane fractions were determined by hydrolysis followed by gas chromatography. Samples (containing approx. 400 pg of membrane protein) were hydrolysed in 3 M HCI for 4 h at 100 "C, after which erythritol (7.5 pg) was added as an internal standard. Neutralisation with Ag,CO, , reacetylation and preparation for gas chromatography was carried out as previously described [20, 211 and chromatographed on columns of SE-30 or OV-17 on Gas Chrom Q (Serva Ltd). A temperature programme from 125-230 "C (4 "C per min) with a nitrogen flow rate of 45 ml per minute was used. Disaccharides. Alkali-labile oligosaccharides were released from the membrane by dissolving plasma membrane samples (approx. 1 mg membrane protein) in 0.05 M NaOH containing 1.0 M NaBH, (2 ml) as described previously [20, 211. The released material was then desialylated (0.1 M H,SO, for 1 h at 80 "C) and prepared for gas chromatography. Free disac-
Thomsen-Friedenreich Antigen o n Pig Lymphocyte Membrane
charide was assayed by isothermal chromatography at 250 "C using trehalose as an internal standard [20,21]. Isolation ? f Lymphocytes,from Peripheral Blood
The isolation procedure was based on that described by Boyum [22,231. 10 I of citrate-containing blood was mixed with 6 % dextran solution (1100 ml) in Tris/saline buffer (0.01 M Tris, pH 7.4 containing 0.15 M NaCI) and allowed to stand at room temperature for 45 min. The resulting leukocyte-rich supernatant was removed from the sedimented erythrocytes, centrifuged (500 x g , 20 min) and then resuspended in phosphate-buffered saline. The resuspended cells were again centrifuged (500 x g, 10 min) and resuspended once again in phosphate-buffered saline. This process was repeated until no thrombocytes, as judged by phase-contrast microscopy, remained in the supernatant. By differential counting, using May-Griinwald and Giemsa solution to stain smears taken from the suspension, the method was found to remove 100% of the thrombocytes and reduce the erythrocyte count to approx. 0.5 %, of the original blood suspension. The thrombocyte-free cell pellet was then resuspended in Tris/saline (0.01 M Tris, pH 7.4 containing 0.15 M NaCl), in one tenth of the original blood volume, and aliquots (30 ml) layered onto 12 ml of a Uromiro/Ficoll mixture (Uromiro-300 mixed with 8 % Ficoll to give a solution with a specific weight of 1.077) in 50-ml centrifuge tubes. After centrifugation (800 x g, 20 min, MSE 6L centrifuge) a layer of white cells was obtained at the interface of the two layers, the white cell count of which contained 99.7% lymphocytes, the other cells being granulocytes. No thrombocytes or macrophages were present in this layer when measured by differential staining, although erythrocytes were still present. The lymphocyte layer was washed by resuspension in Tris/saline buffer followed by centrifugation (500 x g, 10 min). Cell viability as judged by the trypan blue exclusion test was always found to be greater than 90%. Residual erythrocytes were removed from the suspension by resuspending the cells in a solution of 0.155 M ammonium chloride containing 0.01 M NaHCO, and 0.01 mM EDTA at pH 7.4. After incubation for 10 min in an ice-bath, the suspension was centrifuged (500 x g , 10 min) and the lymphocyte pellet washed by resuspension in Tris/saline buffer. Throughout the isolation procedure cells were kept at 4 "C or below. Membrane Preppara t ion
The purified lymphocyte preparation, containing a total of 2 x lo1* cells was suspended in Tris/saline buffer (110 ml) and placed in an ice-cooled nitrogen cavitation bomb. (Artisan Industries Inc., Massachu-
R. A. Newman. W. M. Glockner, and G. G. Uhlenbruck
315
Lymphocyte Suspension 1. N
2 2 . 500
cavitation bomb 30 atmos., 1 5 min.
xg, 1 5 min.
I
I
sediment (nuclei)
supernatant 1 7 500
I
sediment
x g , 20 min.
supernatant
(crude mitochondria1 fraction)
I
I
sediment
supernatant
(microsomal fraction)
(cytosol fraction)
I
hypotonic shock 0.01 M Tris 135 000 x g, 60 min.
I
sediment
supernatant
(microsomes)
(soluble protein)
I
20-40%
discontinuous sucrose gradient 135 000 xg, 17h.
membrane fraction
Scheme 1
setts, U.S.A.). A pressure of 30 atmospheres for 15 min was found to be optimal for preparation of plasma membranes with least breakage of nuclei. The suspension of broken cells released from the bomb was centrifuged (500 x g, 15 min) to remove the nuclei and the resulting supernatant further centrifuged, (17 500 x g , 20 min) to obtain a mitochondria1 fraction. After pelleting of the mitochondria, the supernatant from this fraction was centrifuged (135000 x g, 60 min) using a Beckman L5/65 ultracentrifuge equipped with a swing-out rotor (SW 27).
Ultracentrifugation gave a pellet containing the microsomes and left soluble proteins in solution (Scheme 1). In order to remove any trapped proteins from vesiculated membrane fragments, the microsomes were resuspended with 3-4 strokes of a Dounce homogeniser, in hypotonic buffer (0.01 M Tris, pH 7.4). This suspension was then recentrifuged as above to repellet the microsomal membranes. Sucrose gradient centrifugation of the microsomal fraction was carried out by resuspending the pellet in 1 vol, of Tris buffer (0.01 M, pH 7.4) and mixing with
316
Thomsen-Friedenreich Antigen on Pig Lymphocyte Membrane
Table 1. Enzymatic content of the suhceNular,fractions Values are averages of 3 determinations. Variation was < 15 Specific activity is expressed as nmol ofproduct liberated . mg protein- I . min-' at 37 'C. Figures in parentheses are enrichment factors. Microsomal fractions are (1) before and (2) after hypo-osmotic shock
x.
~
~~~
Subcellular fraction
Protein
5'-Nucleotidase
Acid phosphatase
Glucose-6-phosphatase
Succinic oxidoreductase
specific activity
total activity
specific activity
specific activity
specific activity
mg
nmol x min-' x mg-'
nmol/min nmol x min-' x mg-'
nmol/min nmol x min-l x mg-'
nmol/min nmol x min-' xmg-'
nmol/min
366.0 61 .O 54.0 19.0 10.3 211.1 3.5
23.4 20.4 43.5 115.7 204.3 8.5 236.1 (10.1)
8 564 1238 2349 2 198 2104 1794 826
17677 2946 7339 1453 699 3 968 277
3 587 287 464 217 62 1836 57
9 809 1452 6 674 216 45 0 0
total activity
total activity ~
Homogenate Nuclear fraction Mitochondria1 fraction Microsomes 1 Microsomes 2 Cytosol fraction Plasma membrane
48.3 48.3 135.9 76.5 67.3 18.8 79.3 (1.6)
1 vol. of Tris buffer that contained 40% sucrose. This suspension was then layered onto a discontinuous sucrose gradient from 20 - 40 % and centrifuged (135000xg, 17 h at 4 "C in a Beckman L5/65 ultracentrifuge with SW 27 swing-out rotor). A band was obtained at the interface of the two sucrose layers containing the plasma membrane fragments. In addition, a small amount of material was found above the 20% sucrose layer, as well as a pellet. Before performing the various chemical assays, the plasma membrane fraction was washed twice with distilled water.
9.8 4.1 8.6 11.4 6.0 8.7 16.2 (1.7)
total activity
~
26.8 23.8 123.6 11.4 4.4 0 0
activity by a factor of 10 between homogenate and the purified plasma membrane fraction. Succinic oxidoreductase was confined mainly to the mitochondrial fraction with only small amounts present in the microsomal fraction. After sucrose density centrifugation however, no detectable activity was found in the plasma membrane fraction, suggesting that mitochondrial fragments are removed by this step. Acid phosphatase and glucose-6-phosphatase were distributed throughout all the fractions although the most was removed in the mitochondria1 fraction.
Chemical Analysis of Membrane Fraction Immunological Assays Arachis hypogoea lectin was used to detect the presence of the immunodominant disaccharide of the T-antigen in lymphocyte plasma membrane fragments, as described previously [2]. The agglutinin from Helix pornatia was used to detect the presence of terminal N-acetyl-galactosamine residues [24, 251. Desialylated erythrocytes were prepared by incubation of a 2 % suspension in physiological saline (20 ml) with 200 p1 of neuraminidase solution for 1 h at 37 "C, following which the cells were then washed three times with saline. 2% suspensions were used for immunological assays.
RESULTS
Isolation of Plasma Membranes The enzymic content of the various subcellular fractions outlined by scheme 1 is shown in Table 1. 5'-Nucleotidase, which was used as a marker for plasma membrane, was found to increase in specific
Chemical analysis carried out on the various subcellular fractions showed that, as well as a ten-fold increase in nucleotidase activity, there was a corresponding increase in the relative amounts of neuraminic acid, cholesterol and phospholipid, expressed per milligram of protein (Table 2). It is generally agreed that the sialic acid content of a cell is located almost exclusively at the cell surface and the increase in sialic acid found in the membrane fraction over that of the homogenate, when expressed per milligram of protein, provides a convenient measure of the percentage recovery (Table 2). Cholesterol and phospholipid, relative to protein, are also found in increased amounts in the plasma membrane fraction. The cholesterol/phospholipid ratio was found to be 0.93 in the plasma membrane fraction. The enrichment in 5'-nucleotidase activity was paralleled by a similar increase in sialic acid content between homogenate and plasma membrane fractions. Hexoses and hexosamines determined individually by gas chromatography are shown in Table 3. For comparison we have included the results obtained by Kornfeld et al. [26] for calf thymocytes and results
311
R. A. Newman, W. M. Glockner, and G . G. Uhlenbruck
Table 2. Chemical composition ojssubcellulur ,fractions Values arc averages of 3 determinations with variation i 37; for neuraminic acid determination and < 10% for cholesterol and phospholipid determination. Figures in parentheses arc enrichment factors. Microsomal fractions are (1) before and (2) after hypo-osmotic shock
Neuraminic acid
Homogenate Nuclear fraction Mitochondria1 fraction Microsomal 1 Microsomal 2 Cytosol fraction Plasma membrane
Cholesterol
specific
total
specific
total
specific
total
nmol/ mg protein
pmol
nmol/ mg protein
pmol
nmol/ mg protein
kmol
2.9 0 7.5 16.2 20.7 1.6 49.3 (17)
1.06 0 0.41 0.31 0.21 0.34 0.17
63.1 59.4 143.7 372.4 324.8 25.7 663.8 (10.5)
23.1 3.6 7.8 7.1 3.3 5.4 2.3
93.8 94.3 206 450 430 16.2 717 (7.7)
34.3 5.8 11.1 8.6 4.4 3.4 2.5
Table 3. CurboJ7ydrate composition of plasma membrane fruction Values arc averages of 3 batches obtained by gas chromatography of their trimethylsilyl derivatives after acid hydrolysis. Variation of mono and disaccharide determinations were < 10% and for neuraminic acid < 3%. Glucosamine was standardised as 3.0 for comparison of molar relationships. P-Gal(1 - 3)GalN, P-o-galactosyl( 1-3)N-acetyl-o-galactosaminitol ~~~
Sugar
Pig lymphocyte membrane molar ratio
Fucose Mannose Galactose Glucose N-Acetyl-Dgalactosamine N-Acetyl-Dglucosamine Neuraminic acid p-Gal(l 3)GalN
Phospholipid
~
~~
Pig erythrocyte Calf thymocyte membrane membrane molar ratio molar ratio (Kornfeld et al. W1)
ms/ 100 mg protein 0.5 1.4 3.6 3.6
0.6 1.7 4.3 4.3
0.7 1.2 5.7 1.3
0.34 1.8 3.7 1.6
0.6
0.6
2.0
0.7
3.1
3.0
3.0
3.0
1.5
1.1
1.6
0.7
0.42
0.24
0.34
lymphocyte membranes and erythrocyte membrane glycoprotein, both showed approximately the same absolute values when expressed per milligram of protein or glycoprotein respectively. Following desialylation of the plasma membrane fraction and subsequent alkaline borohydride treatment, a peak was obtained on gas-chromatography which corresponded to that of a standard of fi-D-galactosyl-( 1-+ 3)-N-acetyl-~-galactosaminitol.A single peak was found on columns of either OV-17 or SE-30 and the disaccharide only observed if desialylation was first carried out. The galactosaminitol content of the disaccharide fi-D-galactosyl-( 1+3)-N-acetyl-11galactosaminitol represented 40 % of the total galactosamine content of the membranes. The above results are consistent with the presence, in the pig lymphocyte membrane, of the disaccharide /h-galactosyl-( 1 3)-N-acetyl-~-galactosamine linked to protein via an alkali-labile linkage (serine or threonine) and substituted by sialic acid. -+
Serological Assays
~
obtained by us for pig erythrocytes (unpublished results). It can be seen that the molar ratio of monosaccharides in pig lymphocyte membranes show striking similarities to those obtained by Kornfeld er al. [26] for calf thymocytes, the major difference in the two determinations being the glucose content, which could not be removed by extensive dialysis. The molar ratio of monosaccharides from pig lymphocyte membrane also resembles that of pig erythrocyte membrane glycoproteins except that in the lymphocyte, sialic acid and N-acetylgalactosamine are lower. As well as similarities in the molar ratios of sugars between pig
Desialylated lymphocyte plasma memb -anes showed inhibitory activity towards the lectin from Arachis hypogoeu but not towards the agglutinin from Helix pomutiu. The inhibition of the lectin from A . hypogoea, reported to be specific for the disaccharide /3-Dgalactosyl-(1 -+ 3)-N-acetyl-~-galactosamine[27], confirms the chemical evidence for the presence of this structure in the lymphocyte plasma membrane. (Table 4). The Failure of the membrane suspension, native or desialylated, to inhibit the H . pomutiu agglutinin suggests that no monosaccharide units of galactosamine, either free or subslituted by neuraminic acid, are linked to protein. The results of serological tests, together with chemical evidence, suggests that the
378
Thomsen-Friedenreich Antigen on Pig Lymphocyte Membrane
Table 4. Inhihition of‘u~glutinarionhy A. hypogoea and H. pocnatia Inhibition titres are expressed as their reciprocal against 4 agglutination doses of agglutinin. Membrane preparations of 5 mg protein/ml were used against desialylated erythrocytes of blood group 0. Disaccharide was used at a concentration of 3 mg/ml ~~
Fraction
Inhibition of agglutination by Aruchis
by Hcliw pornutiu
hypogoea
Native membrane (4 Desialylated membrane z4 p-Galactosyl-(l - 3)-Nacetyl-galactosamine 2’
9 0
9
major alkaline-labile oligosaccharide of the lymphocyte membrane is confined to structures containing the disaccharide b-D-galactosyl-( 1+3)-N-acetyl-~-galactosamine and is further substituted by neuraminic acid.
DISCUSSION By using the nitrogen cavitation method of plasma membrane preparation, originally developed by Wallach and Kamat [29], we were able to obtain a 10-fold increase in the specific activity of 5’-nucleotidase. The protein content of the plasma membrane fraction represented 1 of the original homogenate and is in good agreement with other authors, notably Allan and Crumpton [6] and Ferber et ul. [7],who found recoveries of 1.2% and 1.7% respectively. The enrichment of 5’-nucleotidase in plasma membrane fractions is usually reported to be within the range 8 - 1Sfold [6, 30, 311 although there are reports of recoveries as high as 25-fold [7] from pig lymph node lymphocytes and as low as 1.7-fold from mouse leukaemic cells [321. Preliminary investigations by us showed that homogenisatioii with a Dounce homogeniser produced indiscriminate breakage of cells with a higher percentage of ruptured nuclei than by the nitrogen cavitation method. Lymphocytes have a large nucleus/cell volume ratio and the shear forces produced by the Dounce homogeniser are too variable to allow reproducible results. Cell fractions obtained using this method led to variable amounts of marker enzymes being present, as well as greater overlapping of fraction markers. For preparation of purified plasma membrane fractions by sucrose density gradient centrifugation, we used a step from 20”/;’,-40%. Preliminary investigations using smaller differences or a multi-step gradient gave the same results as those obtained with a 20?-40‘;/, gradient.
.Ox
The observed 10-fold increase in 5’-nucleotidase activity between homogenate and the plasma membrane fraction together with a absence of succinate oxidoreductase and only marginal increases in acid phosphatase and glucose-6-phosphatase, indicates our plasma membrane fraction to be relatively pure. Van Blitterswijk et al. [33] and Ferber et at. [7] also found no pronounced difference in glucose-6-phosphatase levels between the different fractions and the views of these authors, that in the case of lymphocytes this enzyme is a poor marker for endoplasmic reticulum as it may be adsorbed to plasma membranes, were confirmed by us. The cholesterol/phospholipid of cell membranes is a convenient measure of the purity of plasma membranes. It is usually agreed that plasma membranes have a value of about 1.0 although this can vary between different types of cell. A value of 0.93 in our preparation compares favourably with those reported by other authors, although van Blitterswijk et a / . [33] found the ratio to be as low as 0.61 for calf thymocytes and Schmidt-Ullrich et al. [34] as high as 1.26 for the same cells. The relatively high cholesterol/phospholipid ratios in our nuclear and mitochondria1 fractions are probably due to unbroken cells or large fragments precipitating with nuclei and mitochondria under centrifugation. This would therefore raise the ratio above the low values normally expected for pure preparations of these organelles. The molar ratios of the individual monosaccharides of the pig lymphocyte membrane show remarkable similarities to those reported by Kornfeld et al. [26] for calf thymocytes. The major difference between the two determinations was the higher amounts of glucose in our preparation. This may be due to a higher amount of glycolipid in pig lymphocyte membranes, or probably arises from contaminating sucrose in the preparation after sucrose density centrifugation. This finding was also reported by Kornfeld et al. in their paper [26] although extensive dialysis of our preparation could not reduce the level of glucose below that which is reported, suggesting that if sucrose is still present it must be bound to the membrane. It is interesting to compare the results obtained from pig lymphocyte membranes with that of pig erythrocyte membrane glycoproteins, as similarities also exist between these two cells. The major difference between these two cells is the lower content of galactosamine and sialic acid in the lymphocyte and this may reflect differences between the amounts of alkali-labile material of the two cells. The occurrence of the disaccharide b-D-galactosyl(1+3)-N-acetyl-~-galactosamine on the pig lymphocyte membrane is another example of the widespread occurrence of this moiety in nature, although it is usually found to be further substituted by neuraminic acid. This is the case in the human erythrocyte mem-
R. A. Newman, W. M. Glockner, and G. G. Uhlenbruck
brane glycoprotein [35], the milk fat globule membrane [20] as well as erythrocytes from other species (unpublished results). The structure is also contained within oligosaccharides from non-membrane proteins such as pig submaxillary gland glycoprotein [36], fetuin [37], antifreeze glycoprotein [21], as well as being found in human urine [38], and human brain gangliosides [9]. Further work however must be carried out on pig lymphocyte membrane oligosaccharides to establish the number and linkages of the neuraminic acid residues, Rogentine et a / . [3] and Reisner [4] have shown that in normal human serum there are complementdependant antibodies that can bind, and lyse, neuraminidase-treated human peripheral lymphocytes. The antibodies were also able to be adsorbed by neuraminidase-treated erythrocytes, which are known to possess the T-antigen on their surface [27]. Rogentine was able to show in addition that the disaccharide b-Dgalactosyl-( 1+3)-N-acetyl-~-galactosaminewas capable of inhibiting the reaction. An alternate approach was made by Kornfeld et a/. [39] who used a lectin preparation from the mushroom Agaricus hisporus and found that it strongly bound to calf thymocytes. The binding was inhibited by the tetrasaccharide from human erythrocyte membrane glycoprotein, the structure of which was determined by Winzler [35], although the protein-linked free disaccharide inhibited to a greater extent. The occurrence of the disaccharide P-D-galactosyl(1 +3)-N-acetyl-~-galactosamine,which we demonstrated chemically to be present on the pig lymphocyte membrane, had previously been hypothetically suggested by the work of Rogentine et a/. [3], Reisner [4] and Kornfeld et a/. [26]. The finding that desialylated lymphocyte membrane inhibits the lectin Arachis hypogoea, which reacts with the immunodominant carbohydrate group of the T-antigen [5], suggests that the alkali-labile disaccharide found by us, is identical with the T-antigen detected by the other authors. Moreover, the results suggest that the disaccharide is the major alkali-labile oligosaccharide of the pig lymphocyte membrane, although it is usually substituted by neuraminic acid. It is known in certain tumour cells, that the Tantigen is exposed on the cell surface [40, 411. It has also been shown that on certain mammary tumour cells, neuraminyl transferase is present in very low amounts [42], suggesting and explaining, that one of the important differences of tumour transformed cells is the exposure of cryptic antigens on the cell surface. This was a concept we had already developed in order to interpret the glycolipid alterations in brain tumours [40]. We hope to explore these differences further in future by using labelled lectins such as that from A . hypogoea which can bind to the T-antigen and which
379
may be of importance in labelling transformed cells in uiuo. This work was supported by the Deutsche Forschungsgemeinschafi (SFB/68/04). R. A. N. is an Awardee of the Alexander ljon Humboldt Stifrung. Mrs Dorit Karduck and Miss Gisela Peters, both supported by the Deutsche Forschungsgemeinschuft, are thanked for expert technical assistance,
REFERENCES 1. Uhlenbruck, G., Wernet, P. & Schumacher, K. (1973) Klin. Wochenschr. 51, 1210-1231. 2. Hammarstrom, S., Hellstrom, U., Dillner, M. L. & Perlmann, P. (1973) J . Exp. Med. 138, 1270-1275. 3. Rogentine, G . N., J r & Plocinik, B. A. (1974) J . Immunol. 113, 848 - 858. 4. Reisner, E. G. (1975) ~ansplantation,I Y , 357 - 360. 5. Kim, Z . & Uhlenbruck. G . (1966) Z. Immunitae!for.sch. 130, 88 - 99. 6. Allan, D. & Crumpton, M. J. (1970) Biochem. J . 120,133- 143. 7. Ferber, E.. Resch, K., Wallach, D . F. H. & Imm, W. (1972) Biochim. Biophys. Acta, 266, 494- 504. 8. Greaves, M. F., Owen. J. J. T., Raff, M. C. (1974) T and B Lymphoc.vtes, p. 77, American Elsevier, N . Y. 9. Klenk, E., Hendriks, U . W. & Gielcn, W. (1962) Hopp+Seyler'k Z . Physiol. Chem. 330, 140 - 144. 10. Dahr, W., Uhlenbruck, G. & Bird, G. W. G. (1974) Vox Sung. 27, 29-42. 11. Aronson, N . N., Jr & Touster, 0. (1974) Methods Enzymol. 31, 90 - 102. 12. Earl, D. C. N. & Korner, A. (1965) Biochem. J . Y4, 721 -734. 13. Trouet, A. (1974) Methods Enzymol. 31, 323-329. 14. Aminoff, B. (1961) Biochem. J . 81, 384-392. 15. Warren. L. (1959) J . B i d . Chem. 234, 1971 - 1975. 16. Zlatkis, A,, Zak, B. & Boyle, A. J. (1953) J . Luh. Clin. Med. 41, 486 - 492. 17. Lowry, 0. H., Roberts, N. R., Leiner, K. Y., Wu, M.-L. & Farr. A. L. (1954) J . Biol. Chem. 207, 1-17. 18. Chen, P. S., Toribara, T. Y. & Warner, H. (1956) Anal. Chem. 28, 1756-1758. 19. Lowry, 0. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J . (1951) J . Biol. Chem. 193, 265-275. 20. Newman, R. A,, Harrison, R. & Uhlenbruck, G . (1976) Biochim. Biciphys. Actu, in press. 21. Glockner. W. M., Newman. R. A. & Uhlenbruck, G. (1975) Biochem. Biophys. Res. Commun. 66, 701 705. 22. Boyum, A. (1968) Scund. J . Clin. Lab. h e s t . 21, Suppl. 97. 23. Boyum, A. (1974) Tissue Antigens, 4, 269-274. 24. Hammarstrom. S. & Kabat, E. A. (1969) Biochemistry, 8, 2696 - 2705. 25. Hammarstrom, S. & Kabat, E. A. (1971) Biochemi.stry, 10,1684 - 1692. 26. Kornfeld, R . & Siemers, C. (1974) J. Biol. Chem. 24Y, 12951301. 27. Uhlenbruck, G., Pardoe, G. I. & Bird. G. W. G . (1969) Z. Immunitaetsforsch. 138. 423 -433. 28. Dahr, W., Uhlenbruck. G. & Bird, G. W. G. (1975) Vox SanR. 28, 133-148. 29. Wallach. D. F. H. & Kamat, V. B. (1966) Methods Enzymol. 8, 164-172. 30. Demus, H. (1973) Biochim. Biophys. Acta, 291, 93- 106. 31. Smith, W. I . , Ladoulis. C. T., Misra, D. N., Gill, 111. T. J. & Bazin, H. (1975) Biochim. Biophys. Aria, 382, 506- 525. 32. Warley, A. & Cook. G . M. W. (1973) Biochim. Biophys. Actu, 323, 55 -68. -
Thomsen-Friedenreich Antigen on Pig Lymphocyte Membrane
380 33. Van Blitterswijk, W. J., Emmelot, P. & Feltkamp, C. A. (1973) Biochim. Biophys. Actu, 298, 577 - 592. 34. Schmidt-Ullrich, R . , Ferber, E., Knufermann, H., Fischer, H. & Wallach, D. F. H. (1974) Biochim. Biophys. Acta, 332, 175- 191. 35. Thomas, D. B. & Winder. R. J. (1969) J . Biol. Chem. 244,5943 5946. 36. Carlson, D. M. (1968) J. Biol. Chem. 243, 616-626. 37. Spiro, R. G. & Bhoyroo, V. D. (1974) J . B i d . Chem. 249, 5704 - 571 7. ~
R. A. Newman, W. M. Glockner, and G. G. Uhlenbruck Abteilung Immunbiologie, Medizinische Universitltsklinik, Kerpener Stral3e 15. D-5000 Koln 41, Federal Republic of Germany
38. Huttunen, J . K. (1966) Ann. Med. Exp. Bml. Fenn. 44, Suppl. 12. 39. Kornfeld, R. & Kornfeld, S. (1974) Ann. N . Y . Acud. Sci. 234, 276 - 282. 40. Uhlenbruck, G. & Gielen, W. (1970) Fortschr. Neurol. 38,202 218. 41. Springer, G . F. & Desai, P. R. (1974) Ann. Clin. Lub. Sci. 4 , 294- 298. 42. Keenan, T. W. & Morre, D . J . (1973) Science (Wash. D.C.) 182,935-931.
Immunochemical Detection of the Thomsen-Friedenreich Antigen (T-antigen) on the Pig Lymphocyte Plasma Membrane Roland A. NEWMAN, Wolfgang M. GLOCKNER, and Gerhard G. UHLENBRUCK Department of Immunobiology, Medical University Clinic, Cologne (Received October 20, 1975/January 26, 1976)
A method for the isolation of lymphocytes from pig peripheral blood and preparation of plasma membranes was developed. The method resulted in a ten-fold increase in 5’-nucleotidase activity and neuraminic acid, relative to protein. A cholesterol/phospholipid ratio of 0.93 was obtained. Alkaline borohydride treatment of isolated plasma membrane, after desialylation, released the disaccharide ~-D-galactosyl-(l+3)-N-acetyl-~-galactosaminitol (the immunodominant group of the T-antigen), which was identified by gas chromatography using a crystalline standard. The disaccharide was not found without prior desialylation, indicating that all the disaccharide units were substituted by neuraminic acid. Serological evidence using the agglutinins from Arachis hypogoea and Helix pomatia confirmed the presence of this disaccharide (T-antigen) in the lymphocyte membrane and indicated that it was the major alkali-labile oligosaccharide of the pig lymphocyte membrane.
It is generally accepted that the cell surface structure determines many functional properties within the cell, as well as reactions between itself and its environment or neighbour. Earlier investigations by our group have already demonstrated that human T and B lymphocytes can be differentiated by the use of the agglutinin from Helix pomatia, which reacts specifically with neuraminidase-treated human T cells [l]. This finding was also observed by Hammarstrom [2]. During the course of this work we have continued a line of research to investigate the presence of neuraminic acid-covered cryptic antigens which, if present, can be used as secondary markers for the characterisation of different types of cells. Different authors, using serological methods, have suggested the presence of receptors for anti-T (Thomsen-Friedenreich phenomenon) antibodies on the surface of lymphocytes from human peripheral blood after treatment with neuraminidase [3,4]. As the immunodominant group of the T-antigen is a disaccharide of defined structure [ 5 ] ,we decided to develop a method to demonstrate the presence of this disacEnzymes. 5’-Ribonucleotide phosphohydrolase (EC 3.1.3.5); glucose-6-phosphatase (EC 3.1.3.9); succinic oxidoreductase (EC 1.3.99.1); orthophosphate monoester phosphohydrolase (EC 3.1.3.2).
charide, by chemical means, on the surface of lymphocyte membranes. In order to obtain sufficient amounts of lymphocyte plasma membranes we isolated lymphocytes from the peripheral blood of pigs. Other authors also have published methods for the isolation of lymphocyte plasma membranes [6,7] but from pig mesenteric lymph nodes, which are known to contain a different subpopulation of lymphocytes to that of peripheral blood [S]. In future, chemical characterisation methods will be applied to different subpopulations of normal and leukaemic human lymphocyte plasma membranes.
MATERIALS AND METHODS
Materials Blood was obtained from freshly slaughtered pigs and prevented from clotting by the addition of sodium citrate solution (1 vol. 3 . 8 % sodium citrate solution to 9 vol. of blood). Dextran T-250 and Ficoll were obtained from Pharmackd Ltd (Uppsala, Sweden). Uromiro-300, a 65% aqueous solution of the methyl glucamine salt of acetylamino-methylacetylitmino triiodobenzoic acid, was obtained from D r Fwnz Kohler-Chemie (Alsbach, West Germany). Enzyme substrates, glucose-6-
374
phosphate, adenosine monophosphate sodium salt, glycerol 3-phosphate and dichlorophenol-indophenol, as well as trypan blue for viability staining, were obtained from Serva Ltd (Heidelberg). Crystalline P-D-galactosyl-(l+ 3)-N-acetyl-~-galactosamine isolated from human brain gangliosides [9] was a gift from Prof. W. Gielen (Pharmakologisches Tnstitut der Universitat Koln). All other general reagents were obtained from Merck Ltd (Darmstadt). Arachis hypogoea lectin was prepared as previously described by Dahr et al. [lo]. Enzyme Determinations
5'-Nucleotidase was used as a marker for plasma membrane and glucose-6-phosphatase for endoplasmic reticulum. Both enzymes were assayed as described by Aronson and Touster [I 11. Succinic oxidoreductase was assayed as described by Earl and Korner [12] as a mitochondria1 enzyme and acid phosphatase as a marker for lysosomes [13].
Analyticd Procedures Neuraminic Acid. Total neuraminic acid, released by hydrolysis in 0.1 M H,SO, for 1 h at 80 "C, was assayed by the method of Aminoff [14] and corrections made for any interfering nucleic acids as described by Warren [15]. Lipids. Cholesterol and phospholipids were determined after chloroform/methanol extraction as described byzlatkisetal, [16] and Lowry [I71 respectively. Phosphate. Free phosphate after enzymatic release was determined by the method of Chen et al. [18]. Protein. Protein was measured using the FolinCiocalteau reagent as described by Lowry [19]. Hexoses and Hexosamines. Individual hexoses and hexosamines in plasma membrane fractions were determined by hydrolysis followed by gas chromatography. Samples (containing approx. 400 pg of membrane protein) were hydrolysed in 3 M HCI for 4 h at 100 "C, after which erythritol (7.5 pg) was added as an internal standard. Neutralisation with Ag,CO, , reacetylation and preparation for gas chromatography was carried out as previously described [20, 211 and chromatographed on columns of SE-30 or OV-17 on Gas Chrom Q (Serva Ltd). A temperature programme from 125-230 "C (4 "C per min) with a nitrogen flow rate of 45 ml per minute was used. Disaccharides. Alkali-labile oligosaccharides were released from the membrane by dissolving plasma membrane samples (approx. 1 mg membrane protein) in 0.05 M NaOH containing 1.0 M NaBH, (2 ml) as described previously [20, 211. The released material was then desialylated (0.1 M H,SO, for 1 h at 80 "C) and prepared for gas chromatography. Free disac-
Thomsen-Friedenreich Antigen o n Pig Lymphocyte Membrane
charide was assayed by isothermal chromatography at 250 "C using trehalose as an internal standard [20,21]. Isolation ? f Lymphocytes,from Peripheral Blood
The isolation procedure was based on that described by Boyum [22,231. 10 I of citrate-containing blood was mixed with 6 % dextran solution (1100 ml) in Tris/saline buffer (0.01 M Tris, pH 7.4 containing 0.15 M NaCI) and allowed to stand at room temperature for 45 min. The resulting leukocyte-rich supernatant was removed from the sedimented erythrocytes, centrifuged (500 x g , 20 min) and then resuspended in phosphate-buffered saline. The resuspended cells were again centrifuged (500 x g, 10 min) and resuspended once again in phosphate-buffered saline. This process was repeated until no thrombocytes, as judged by phase-contrast microscopy, remained in the supernatant. By differential counting, using May-Griinwald and Giemsa solution to stain smears taken from the suspension, the method was found to remove 100% of the thrombocytes and reduce the erythrocyte count to approx. 0.5 %, of the original blood suspension. The thrombocyte-free cell pellet was then resuspended in Tris/saline (0.01 M Tris, pH 7.4 containing 0.15 M NaCl), in one tenth of the original blood volume, and aliquots (30 ml) layered onto 12 ml of a Uromiro/Ficoll mixture (Uromiro-300 mixed with 8 % Ficoll to give a solution with a specific weight of 1.077) in 50-ml centrifuge tubes. After centrifugation (800 x g, 20 min, MSE 6L centrifuge) a layer of white cells was obtained at the interface of the two layers, the white cell count of which contained 99.7% lymphocytes, the other cells being granulocytes. No thrombocytes or macrophages were present in this layer when measured by differential staining, although erythrocytes were still present. The lymphocyte layer was washed by resuspension in Tris/saline buffer followed by centrifugation (500 x g, 10 min). Cell viability as judged by the trypan blue exclusion test was always found to be greater than 90%. Residual erythrocytes were removed from the suspension by resuspending the cells in a solution of 0.155 M ammonium chloride containing 0.01 M NaHCO, and 0.01 mM EDTA at pH 7.4. After incubation for 10 min in an ice-bath, the suspension was centrifuged (500 x g , 10 min) and the lymphocyte pellet washed by resuspension in Tris/saline buffer. Throughout the isolation procedure cells were kept at 4 "C or below. Membrane Preppara t ion
The purified lymphocyte preparation, containing a total of 2 x lo1* cells was suspended in Tris/saline buffer (110 ml) and placed in an ice-cooled nitrogen cavitation bomb. (Artisan Industries Inc., Massachu-
R. A. Newman. W. M. Glockner, and G. G. Uhlenbruck
315
Lymphocyte Suspension 1. N
2 2 . 500
cavitation bomb 30 atmos., 1 5 min.
xg, 1 5 min.
I
I
sediment (nuclei)
supernatant 1 7 500
I
sediment
x g , 20 min.
supernatant
(crude mitochondria1 fraction)
I
I
sediment
supernatant
(microsomal fraction)
(cytosol fraction)
I
hypotonic shock 0.01 M Tris 135 000 x g, 60 min.
I
sediment
supernatant
(microsomes)
(soluble protein)
I
20-40%
discontinuous sucrose gradient 135 000 xg, 17h.
membrane fraction
Scheme 1
setts, U.S.A.). A pressure of 30 atmospheres for 15 min was found to be optimal for preparation of plasma membranes with least breakage of nuclei. The suspension of broken cells released from the bomb was centrifuged (500 x g, 15 min) to remove the nuclei and the resulting supernatant further centrifuged, (17 500 x g , 20 min) to obtain a mitochondria1 fraction. After pelleting of the mitochondria, the supernatant from this fraction was centrifuged (135000 x g, 60 min) using a Beckman L5/65 ultracentrifuge equipped with a swing-out rotor (SW 27).
Ultracentrifugation gave a pellet containing the microsomes and left soluble proteins in solution (Scheme 1). In order to remove any trapped proteins from vesiculated membrane fragments, the microsomes were resuspended with 3-4 strokes of a Dounce homogeniser, in hypotonic buffer (0.01 M Tris, pH 7.4). This suspension was then recentrifuged as above to repellet the microsomal membranes. Sucrose gradient centrifugation of the microsomal fraction was carried out by resuspending the pellet in 1 vol, of Tris buffer (0.01 M, pH 7.4) and mixing with
316
Thomsen-Friedenreich Antigen on Pig Lymphocyte Membrane
Table 1. Enzymatic content of the suhceNular,fractions Values are averages of 3 determinations. Variation was < 15 Specific activity is expressed as nmol ofproduct liberated . mg protein- I . min-' at 37 'C. Figures in parentheses are enrichment factors. Microsomal fractions are (1) before and (2) after hypo-osmotic shock
x.
~
~~~
Subcellular fraction
Protein
5'-Nucleotidase
Acid phosphatase
Glucose-6-phosphatase
Succinic oxidoreductase
specific activity
total activity
specific activity
specific activity
specific activity
mg
nmol x min-' x mg-'
nmol/min nmol x min-' x mg-'
nmol/min nmol x min-l x mg-'
nmol/min nmol x min-' xmg-'
nmol/min
366.0 61 .O 54.0 19.0 10.3 211.1 3.5
23.4 20.4 43.5 115.7 204.3 8.5 236.1 (10.1)
8 564 1238 2349 2 198 2104 1794 826
17677 2946 7339 1453 699 3 968 277
3 587 287 464 217 62 1836 57
9 809 1452 6 674 216 45 0 0
total activity
total activity ~
Homogenate Nuclear fraction Mitochondria1 fraction Microsomes 1 Microsomes 2 Cytosol fraction Plasma membrane
48.3 48.3 135.9 76.5 67.3 18.8 79.3 (1.6)
1 vol. of Tris buffer that contained 40% sucrose. This suspension was then layered onto a discontinuous sucrose gradient from 20 - 40 % and centrifuged (135000xg, 17 h at 4 "C in a Beckman L5/65 ultracentrifuge with SW 27 swing-out rotor). A band was obtained at the interface of the two sucrose layers containing the plasma membrane fragments. In addition, a small amount of material was found above the 20% sucrose layer, as well as a pellet. Before performing the various chemical assays, the plasma membrane fraction was washed twice with distilled water.
9.8 4.1 8.6 11.4 6.0 8.7 16.2 (1.7)
total activity
~
26.8 23.8 123.6 11.4 4.4 0 0
activity by a factor of 10 between homogenate and the purified plasma membrane fraction. Succinic oxidoreductase was confined mainly to the mitochondrial fraction with only small amounts present in the microsomal fraction. After sucrose density centrifugation however, no detectable activity was found in the plasma membrane fraction, suggesting that mitochondrial fragments are removed by this step. Acid phosphatase and glucose-6-phosphatase were distributed throughout all the fractions although the most was removed in the mitochondria1 fraction.
Chemical Analysis of Membrane Fraction Immunological Assays Arachis hypogoea lectin was used to detect the presence of the immunodominant disaccharide of the T-antigen in lymphocyte plasma membrane fragments, as described previously [2]. The agglutinin from Helix pornatia was used to detect the presence of terminal N-acetyl-galactosamine residues [24, 251. Desialylated erythrocytes were prepared by incubation of a 2 % suspension in physiological saline (20 ml) with 200 p1 of neuraminidase solution for 1 h at 37 "C, following which the cells were then washed three times with saline. 2% suspensions were used for immunological assays.
RESULTS
Isolation of Plasma Membranes The enzymic content of the various subcellular fractions outlined by scheme 1 is shown in Table 1. 5'-Nucleotidase, which was used as a marker for plasma membrane, was found to increase in specific
Chemical analysis carried out on the various subcellular fractions showed that, as well as a ten-fold increase in nucleotidase activity, there was a corresponding increase in the relative amounts of neuraminic acid, cholesterol and phospholipid, expressed per milligram of protein (Table 2). It is generally agreed that the sialic acid content of a cell is located almost exclusively at the cell surface and the increase in sialic acid found in the membrane fraction over that of the homogenate, when expressed per milligram of protein, provides a convenient measure of the percentage recovery (Table 2). Cholesterol and phospholipid, relative to protein, are also found in increased amounts in the plasma membrane fraction. The cholesterol/phospholipid ratio was found to be 0.93 in the plasma membrane fraction. The enrichment in 5'-nucleotidase activity was paralleled by a similar increase in sialic acid content between homogenate and plasma membrane fractions. Hexoses and hexosamines determined individually by gas chromatography are shown in Table 3. For comparison we have included the results obtained by Kornfeld et al. [26] for calf thymocytes and results
311
R. A. Newman, W. M. Glockner, and G . G. Uhlenbruck
Table 2. Chemical composition ojssubcellulur ,fractions Values arc averages of 3 determinations with variation i 37; for neuraminic acid determination and < 10% for cholesterol and phospholipid determination. Figures in parentheses arc enrichment factors. Microsomal fractions are (1) before and (2) after hypo-osmotic shock
Neuraminic acid
Homogenate Nuclear fraction Mitochondria1 fraction Microsomal 1 Microsomal 2 Cytosol fraction Plasma membrane
Cholesterol
specific
total
specific
total
specific
total
nmol/ mg protein
pmol
nmol/ mg protein
pmol
nmol/ mg protein
kmol
2.9 0 7.5 16.2 20.7 1.6 49.3 (17)
1.06 0 0.41 0.31 0.21 0.34 0.17
63.1 59.4 143.7 372.4 324.8 25.7 663.8 (10.5)
23.1 3.6 7.8 7.1 3.3 5.4 2.3
93.8 94.3 206 450 430 16.2 717 (7.7)
34.3 5.8 11.1 8.6 4.4 3.4 2.5
Table 3. CurboJ7ydrate composition of plasma membrane fruction Values arc averages of 3 batches obtained by gas chromatography of their trimethylsilyl derivatives after acid hydrolysis. Variation of mono and disaccharide determinations were < 10% and for neuraminic acid < 3%. Glucosamine was standardised as 3.0 for comparison of molar relationships. P-Gal(1 - 3)GalN, P-o-galactosyl( 1-3)N-acetyl-o-galactosaminitol ~~~
Sugar
Pig lymphocyte membrane molar ratio
Fucose Mannose Galactose Glucose N-Acetyl-Dgalactosamine N-Acetyl-Dglucosamine Neuraminic acid p-Gal(l 3)GalN
Phospholipid
~
~~
Pig erythrocyte Calf thymocyte membrane membrane molar ratio molar ratio (Kornfeld et al. W1)
ms/ 100 mg protein 0.5 1.4 3.6 3.6
0.6 1.7 4.3 4.3
0.7 1.2 5.7 1.3
0.34 1.8 3.7 1.6
0.6
0.6
2.0
0.7
3.1
3.0
3.0
3.0
1.5
1.1
1.6
0.7
0.42
0.24
0.34
lymphocyte membranes and erythrocyte membrane glycoprotein, both showed approximately the same absolute values when expressed per milligram of protein or glycoprotein respectively. Following desialylation of the plasma membrane fraction and subsequent alkaline borohydride treatment, a peak was obtained on gas-chromatography which corresponded to that of a standard of fi-D-galactosyl-( 1-+ 3)-N-acetyl-~-galactosaminitol.A single peak was found on columns of either OV-17 or SE-30 and the disaccharide only observed if desialylation was first carried out. The galactosaminitol content of the disaccharide fi-D-galactosyl-( 1+3)-N-acetyl-11galactosaminitol represented 40 % of the total galactosamine content of the membranes. The above results are consistent with the presence, in the pig lymphocyte membrane, of the disaccharide /h-galactosyl-( 1 3)-N-acetyl-~-galactosamine linked to protein via an alkali-labile linkage (serine or threonine) and substituted by sialic acid. -+
Serological Assays
~
obtained by us for pig erythrocytes (unpublished results). It can be seen that the molar ratio of monosaccharides in pig lymphocyte membranes show striking similarities to those obtained by Kornfeld er al. [26] for calf thymocytes, the major difference in the two determinations being the glucose content, which could not be removed by extensive dialysis. The molar ratio of monosaccharides from pig lymphocyte membrane also resembles that of pig erythrocyte membrane glycoproteins except that in the lymphocyte, sialic acid and N-acetylgalactosamine are lower. As well as similarities in the molar ratios of sugars between pig
Desialylated lymphocyte plasma memb -anes showed inhibitory activity towards the lectin from Arachis hypogoeu but not towards the agglutinin from Helix pomutiu. The inhibition of the lectin from A . hypogoea, reported to be specific for the disaccharide /3-Dgalactosyl-(1 -+ 3)-N-acetyl-~-galactosamine[27], confirms the chemical evidence for the presence of this structure in the lymphocyte plasma membrane. (Table 4). The Failure of the membrane suspension, native or desialylated, to inhibit the H . pomutiu agglutinin suggests that no monosaccharide units of galactosamine, either free or subslituted by neuraminic acid, are linked to protein. The results of serological tests, together with chemical evidence, suggests that the
378
Thomsen-Friedenreich Antigen on Pig Lymphocyte Membrane
Table 4. Inhihition of‘u~glutinarionhy A. hypogoea and H. pocnatia Inhibition titres are expressed as their reciprocal against 4 agglutination doses of agglutinin. Membrane preparations of 5 mg protein/ml were used against desialylated erythrocytes of blood group 0. Disaccharide was used at a concentration of 3 mg/ml ~~
Fraction
Inhibition of agglutination by Aruchis
by Hcliw pornutiu
hypogoea
Native membrane (4 Desialylated membrane z4 p-Galactosyl-(l - 3)-Nacetyl-galactosamine 2’
9 0
9
major alkaline-labile oligosaccharide of the lymphocyte membrane is confined to structures containing the disaccharide b-D-galactosyl-( 1+3)-N-acetyl-~-galactosamine and is further substituted by neuraminic acid.
DISCUSSION By using the nitrogen cavitation method of plasma membrane preparation, originally developed by Wallach and Kamat [29], we were able to obtain a 10-fold increase in the specific activity of 5’-nucleotidase. The protein content of the plasma membrane fraction represented 1 of the original homogenate and is in good agreement with other authors, notably Allan and Crumpton [6] and Ferber et ul. [7],who found recoveries of 1.2% and 1.7% respectively. The enrichment of 5’-nucleotidase in plasma membrane fractions is usually reported to be within the range 8 - 1Sfold [6, 30, 311 although there are reports of recoveries as high as 25-fold [7] from pig lymph node lymphocytes and as low as 1.7-fold from mouse leukaemic cells [321. Preliminary investigations by us showed that homogenisatioii with a Dounce homogeniser produced indiscriminate breakage of cells with a higher percentage of ruptured nuclei than by the nitrogen cavitation method. Lymphocytes have a large nucleus/cell volume ratio and the shear forces produced by the Dounce homogeniser are too variable to allow reproducible results. Cell fractions obtained using this method led to variable amounts of marker enzymes being present, as well as greater overlapping of fraction markers. For preparation of purified plasma membrane fractions by sucrose density gradient centrifugation, we used a step from 20”/;’,-40%. Preliminary investigations using smaller differences or a multi-step gradient gave the same results as those obtained with a 20?-40‘;/, gradient.
.Ox
The observed 10-fold increase in 5’-nucleotidase activity between homogenate and the plasma membrane fraction together with a absence of succinate oxidoreductase and only marginal increases in acid phosphatase and glucose-6-phosphatase, indicates our plasma membrane fraction to be relatively pure. Van Blitterswijk et al. [33] and Ferber et at. [7] also found no pronounced difference in glucose-6-phosphatase levels between the different fractions and the views of these authors, that in the case of lymphocytes this enzyme is a poor marker for endoplasmic reticulum as it may be adsorbed to plasma membranes, were confirmed by us. The cholesterol/phospholipid of cell membranes is a convenient measure of the purity of plasma membranes. It is usually agreed that plasma membranes have a value of about 1.0 although this can vary between different types of cell. A value of 0.93 in our preparation compares favourably with those reported by other authors, although van Blitterswijk et a / . [33] found the ratio to be as low as 0.61 for calf thymocytes and Schmidt-Ullrich et al. [34] as high as 1.26 for the same cells. The relatively high cholesterol/phospholipid ratios in our nuclear and mitochondria1 fractions are probably due to unbroken cells or large fragments precipitating with nuclei and mitochondria under centrifugation. This would therefore raise the ratio above the low values normally expected for pure preparations of these organelles. The molar ratios of the individual monosaccharides of the pig lymphocyte membrane show remarkable similarities to those reported by Kornfeld et al. [26] for calf thymocytes. The major difference between the two determinations was the higher amounts of glucose in our preparation. This may be due to a higher amount of glycolipid in pig lymphocyte membranes, or probably arises from contaminating sucrose in the preparation after sucrose density centrifugation. This finding was also reported by Kornfeld et al. in their paper [26] although extensive dialysis of our preparation could not reduce the level of glucose below that which is reported, suggesting that if sucrose is still present it must be bound to the membrane. It is interesting to compare the results obtained from pig lymphocyte membranes with that of pig erythrocyte membrane glycoproteins, as similarities also exist between these two cells. The major difference between these two cells is the lower content of galactosamine and sialic acid in the lymphocyte and this may reflect differences between the amounts of alkali-labile material of the two cells. The occurrence of the disaccharide b-D-galactosyl(1+3)-N-acetyl-~-galactosamine on the pig lymphocyte membrane is another example of the widespread occurrence of this moiety in nature, although it is usually found to be further substituted by neuraminic acid. This is the case in the human erythrocyte mem-
R. A. Newman, W. M. Glockner, and G. G. Uhlenbruck
brane glycoprotein [35], the milk fat globule membrane [20] as well as erythrocytes from other species (unpublished results). The structure is also contained within oligosaccharides from non-membrane proteins such as pig submaxillary gland glycoprotein [36], fetuin [37], antifreeze glycoprotein [21], as well as being found in human urine [38], and human brain gangliosides [9]. Further work however must be carried out on pig lymphocyte membrane oligosaccharides to establish the number and linkages of the neuraminic acid residues, Rogentine et a / . [3] and Reisner [4] have shown that in normal human serum there are complementdependant antibodies that can bind, and lyse, neuraminidase-treated human peripheral lymphocytes. The antibodies were also able to be adsorbed by neuraminidase-treated erythrocytes, which are known to possess the T-antigen on their surface [27]. Rogentine was able to show in addition that the disaccharide b-Dgalactosyl-( 1+3)-N-acetyl-~-galactosaminewas capable of inhibiting the reaction. An alternate approach was made by Kornfeld et a/. [39] who used a lectin preparation from the mushroom Agaricus hisporus and found that it strongly bound to calf thymocytes. The binding was inhibited by the tetrasaccharide from human erythrocyte membrane glycoprotein, the structure of which was determined by Winzler [35], although the protein-linked free disaccharide inhibited to a greater extent. The occurrence of the disaccharide P-D-galactosyl(1 +3)-N-acetyl-~-galactosamine,which we demonstrated chemically to be present on the pig lymphocyte membrane, had previously been hypothetically suggested by the work of Rogentine et a/. [3], Reisner [4] and Kornfeld et a/. [26]. The finding that desialylated lymphocyte membrane inhibits the lectin Arachis hypogoea, which reacts with the immunodominant carbohydrate group of the T-antigen [5], suggests that the alkali-labile disaccharide found by us, is identical with the T-antigen detected by the other authors. Moreover, the results suggest that the disaccharide is the major alkali-labile oligosaccharide of the pig lymphocyte membrane, although it is usually substituted by neuraminic acid. It is known in certain tumour cells, that the Tantigen is exposed on the cell surface [40, 411. It has also been shown that on certain mammary tumour cells, neuraminyl transferase is present in very low amounts [42], suggesting and explaining, that one of the important differences of tumour transformed cells is the exposure of cryptic antigens on the cell surface. This was a concept we had already developed in order to interpret the glycolipid alterations in brain tumours [40]. We hope to explore these differences further in future by using labelled lectins such as that from A . hypogoea which can bind to the T-antigen and which
379
may be of importance in labelling transformed cells in uiuo. This work was supported by the Deutsche Forschungsgemeinschafi (SFB/68/04). R. A. N. is an Awardee of the Alexander ljon Humboldt Stifrung. Mrs Dorit Karduck and Miss Gisela Peters, both supported by the Deutsche Forschungsgemeinschuft, are thanked for expert technical assistance,
REFERENCES 1. Uhlenbruck, G., Wernet, P. & Schumacher, K. (1973) Klin. Wochenschr. 51, 1210-1231. 2. Hammarstrom, S., Hellstrom, U., Dillner, M. L. & Perlmann, P. (1973) J . Exp. Med. 138, 1270-1275. 3. Rogentine, G . N., J r & Plocinik, B. A. (1974) J . Immunol. 113, 848 - 858. 4. Reisner, E. G. (1975) ~ansplantation,I Y , 357 - 360. 5. Kim, Z . & Uhlenbruck. G . (1966) Z. Immunitae!for.sch. 130, 88 - 99. 6. Allan, D. & Crumpton, M. J. (1970) Biochem. J . 120,133- 143. 7. Ferber, E.. Resch, K., Wallach, D . F. H. & Imm, W. (1972) Biochim. Biophys. Acta, 266, 494- 504. 8. Greaves, M. F., Owen. J. J. T., Raff, M. C. (1974) T and B Lymphoc.vtes, p. 77, American Elsevier, N . Y. 9. Klenk, E., Hendriks, U . W. & Gielcn, W. (1962) Hopp+Seyler'k Z . Physiol. Chem. 330, 140 - 144. 10. Dahr, W., Uhlenbruck, G. & Bird, G. W. G. (1974) Vox Sung. 27, 29-42. 11. Aronson, N . N., Jr & Touster, 0. (1974) Methods Enzymol. 31, 90 - 102. 12. Earl, D. C. N. & Korner, A. (1965) Biochem. J . Y4, 721 -734. 13. Trouet, A. (1974) Methods Enzymol. 31, 323-329. 14. Aminoff, B. (1961) Biochem. J . 81, 384-392. 15. Warren. L. (1959) J . B i d . Chem. 234, 1971 - 1975. 16. Zlatkis, A,, Zak, B. & Boyle, A. J. (1953) J . Luh. Clin. Med. 41, 486 - 492. 17. Lowry, 0. H., Roberts, N. R., Leiner, K. Y., Wu, M.-L. & Farr. A. L. (1954) J . Biol. Chem. 207, 1-17. 18. Chen, P. S., Toribara, T. Y. & Warner, H. (1956) Anal. Chem. 28, 1756-1758. 19. Lowry, 0. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J . (1951) J . Biol. Chem. 193, 265-275. 20. Newman, R. A,, Harrison, R. & Uhlenbruck, G . (1976) Biochim. Biciphys. Actu, in press. 21. Glockner. W. M., Newman. R. A. & Uhlenbruck, G. (1975) Biochem. Biophys. Res. Commun. 66, 701 705. 22. Boyum, A. (1968) Scund. J . Clin. Lab. h e s t . 21, Suppl. 97. 23. Boyum, A. (1974) Tissue Antigens, 4, 269-274. 24. Hammarstrom. S. & Kabat, E. A. (1969) Biochemistry, 8, 2696 - 2705. 25. Hammarstrom, S. & Kabat, E. A. (1971) Biochemi.stry, 10,1684 - 1692. 26. Kornfeld, R . & Siemers, C. (1974) J. Biol. Chem. 24Y, 12951301. 27. Uhlenbruck, G., Pardoe, G. I. & Bird. G. W. G . (1969) Z. Immunitaetsforsch. 138. 423 -433. 28. Dahr, W., Uhlenbruck. G. & Bird, G. W. G. (1975) Vox SanR. 28, 133-148. 29. Wallach. D. F. H. & Kamat, V. B. (1966) Methods Enzymol. 8, 164-172. 30. Demus, H. (1973) Biochim. Biophys. Acta, 291, 93- 106. 31. Smith, W. I . , Ladoulis. C. T., Misra, D. N., Gill, 111. T. J. & Bazin, H. (1975) Biochim. Biophys. Aria, 382, 506- 525. 32. Warley, A. & Cook. G . M. W. (1973) Biochim. Biophys. Actu, 323, 55 -68. -
Thomsen-Friedenreich Antigen on Pig Lymphocyte Membrane
380 33. Van Blitterswijk, W. J., Emmelot, P. & Feltkamp, C. A. (1973) Biochim. Biophys. Actu, 298, 577 - 592. 34. Schmidt-Ullrich, R . , Ferber, E., Knufermann, H., Fischer, H. & Wallach, D. F. H. (1974) Biochim. Biophys. Acta, 332, 175- 191. 35. Thomas, D. B. & Winder. R. J. (1969) J . Biol. Chem. 244,5943 5946. 36. Carlson, D. M. (1968) J. Biol. Chem. 243, 616-626. 37. Spiro, R. G. & Bhoyroo, V. D. (1974) J . B i d . Chem. 249, 5704 - 571 7. ~
R. A. Newman, W. M. Glockner, and G. G. Uhlenbruck Abteilung Immunbiologie, Medizinische Universitltsklinik, Kerpener Stral3e 15. D-5000 Koln 41, Federal Republic of Germany
38. Huttunen, J . K. (1966) Ann. Med. Exp. Bml. Fenn. 44, Suppl. 12. 39. Kornfeld, R. & Kornfeld, S. (1974) Ann. N . Y . Acud. Sci. 234, 276 - 282. 40. Uhlenbruck, G. & Gielen, W. (1970) Fortschr. Neurol. 38,202 218. 41. Springer, G . F. & Desai, P. R. (1974) Ann. Clin. Lub. Sci. 4 , 294- 298. 42. Keenan, T. W. & Morre, D . J . (1973) Science (Wash. D.C.) 182,935-931.
