l~.ochimicaet Biophj,'~caActa. !!3611992)23-27
23
© 1992ELsevierScience FublishcrsB.V. All rightsreserved0167-4889/92/$05.00
Altered ganglioside composition in virally transformed rat embryo fibroblasts H o n g w e i Bai, J o s e p h O r l a n d o a n d T h o m a s N. S e y f r i e d Department of Biology; Boston College, Chestnut Hill, MA (USA)
(Rece/ved 13January 1992l
Keywords: Gangl/osid¢;Pob'omatransformant;SV40transformant;Cultured cell; (Rat embryofibrobtast) The composition of ganglinsides was examined in a normal rat embryo fibroblast cell line (REF52) and in two viral transformants: a Imllanna trar~foi'~ua~'at(REF52-PyMLV) and a simian viral 40 transformant (REF52-SV40). The distribution of gangl/osides in the cell lines was determined using gas-liquid chromatography and high-performance thin-layer chromatography. N-acetylneuranlialc acid was the predominant sialic acid spe~es detected in the three cell lines. The total ganglinside concentration (ttg/100 mg d~ weight of cells) in the normal, I~MLV, and SV40 lines was 144.7 4-10.4. 153.84- 9.2, and 86.1 +_68, respect~'eb-. Ganglinsides GM3, GM2, GMI, and GDIa were the major species in the normal and transformed lines. The distribution of these gangJiasides, however, differed markedly between the normal and the transformed lines and also between the transformed lines themselves.The transformed cells also differed from the normal cells in growth rate, morphology, and social behavior. The cell line with highest GM3 content (PyMLV) formed islands, whereas the nornmt and SV40 cell lines, which had lower GM3 levels, grew as monelayers. The findings suggest that PyMLV and SV40 transformation can have multiple and differem effects on cellular ganglinside distribution and growth behavior.
latruductiea Gangliesides are siafic acid-containing glycnsphingolipids that are recognized as ubiquitous constituents of mammalian plasma membranes [1-3]. These molecules are thought to participate in cell recognition, growth control, differentiation, and tumorigenesis [4-
9].
Since it was first discovered that ganglioside alterations are closely associafed wi.*h cell transformation [10], considerable work has been done to characterize these alterations. The influence of various transforming agents, such as DNA or RNA viruses, =hemitr,d carcinogens, and X-ray irradiation on gaoglioside composition were studied on a variety of cell types [11-18]. There is considerable controversy, however, on the ganglioside changes associated with viral transformation of mammalian cells. The ganglioside changes observed can depend on both the transforming virus used and on the cell type transformed. In general, cellular transformation with DNA viruses, either the polyoma marine leukemia virus (I~MLV) or the simian virus 40
Conespondence to: T.N. Se~'fried.Department of Bidogy. Boston College. ChestnutHill MA 02167. USA.
(SV40), is associated with an elevation of ganglioside GM3 [11-13,16] and reductions of the more complex gangliosides, e.g., GM2 [13,14], GMI and G D l a [1114]. There are, however, exceptions to this general trend [10-16]. To better define ganghoside changes associated with viral transformation, we have characterized the ganglinside composition in a normal rat embryo fibroblast cell line (REF52) and in two viral transformants of the cell line: REF52-PyMLV and REF52-SV40. Ours is the first study of ganglioside changes associated with polyoma and SV40 viral transformation in rat embryo fibroblasts. Materials and Methods Cells and cell culture
The REF52 cell line was established from fibroblasts of a 14-day gestation Fischer rat embryo [19]. This cell line expresses contact inhibition and is nontumorigenic. The REF52-1~MLV variant was produced after transformation ~ the polyoma virus middle T gene [20] and the REF52-SV40 variant was produced after transformation by the SV40 virus [21]. Both the REF52PyMLV and the REF52-SV40 transformants lack contact inhibition and are tumorigenic. Wild-type REF52 and transformants were obtained from Dr. J.IC Cben (Chang Gun Medical College, Taipei, Taiwan).
Ceils (1- |04/ore 2 for REF52 variant and 2104/cm 2 for REF52-PyMLV arid REF52-SV40 variants) were seeded in 150 cm 2 flasks and grown in 30 nd of a basal medium consisting of 25% Ham's !:-12 and 75% DMEM and supplemented with 10% fetal bovine serum. The ceils were cultured in a humidified atmosphere of 95% air and 5% CO 2 at 37~C and the medium was changed every two days during the growing period. V~en the ceils reached confluence (about 6 days), they were treated with 0.1% trypsin, harvested by cenzrifugadon and washed three times with cold phosphate-buffered saline. The washed ceils were bophilized and stored at - 2 0 ~ C . Gang//as/de ana6,s/s Gan~ioside~ were ;:~c~!a:cd and purified from lyhophifized ceD puwders according to the methods described previously [22-24]. The gangliosides were treated with mild base (0.1 M NaOH for 1 h) and desalted using Sephadex G-50 column chromatography [23,25]. The concentration o£ ganglioside Nacctylneura.,~c acid (NeuAc) and N-glycolylneuraminic acid (NeuGc) ~ s deter~ained using the gas-liquid chromatographic method of Yu and Ledeen [26]. The sialic acid concentration was expressed as p g p e r 109 mg dry weighL The quantitative anab~-is of individual ganglioside species was performed according to the procednre of Ando et al. [27]. In this procedure, 10 × 10 cm thin-layer silica gel plates were used (Whatman HP-K high-performance silica gel). The conditions for HPTLC development are gk'en in Fig. 1. The ganglioside bands were visualized by spraying the dried plates with a resorcinobHCl reagent [28] and by baking the plates at 95°C for 30 rain. The percent gangliuside distribution was determined by the direct densitometric scanning method of Ando et al. [27]. The gangliosides of the normal and transformed rat embryo fibrublasts were identified by comparison with ganglinside s~tandards isolated from the cerebral cortex of normal adult mice and with GM4, GM3, GM2, GM!, GD3, GDIa, GTIa, GDlb, G F l b and G O l b purified and isolated from mouse, bovine and, human tissues by the method of Ando and Yu [29,30]. A preliminary structural characterization of the ganglios[des in both the normal and transformed rat embryo fibroblasts was performed by a neuraminidase treatment (EC 3.2.1.18, type VI from Clostridimn p e r fr/ngens, Sigma), according to the procedure of Ando and Yu [30]. Purified ganglioside samples (containing 5 /tg of sialic acid) from each cell line were mixed with 100 gl of a fresh neuraminidase solution (1 unit in 1 ml of 0.1 M sodium acetate buffer, pH 5) and incubated at 3"PC for 24 h. After base ~reatment (0.1 M NaOH for 1 h) and desalting, the treated gangliusides were spotted on HPTLC plates and developed as described above.
.....7
,
GM4 GM3 GM2
W
GMI
i!~!!~:!:~:¸¸
GO3 GDla GTla
:!!:!!!:!!!!~
GDlb GTlb G~lb
1
$
2
4
S
~g. I. Th~t-la3,~*r ~ of ganglioskles in normal and ,~ams-Uansform~rat eml~3~ fi~robLasls.Apprmdmately 1.5 ~g of g~m~ Neu_Ac~ s spoued for each sample. The plate ~;~s de~etoped by one ascendh~ elufion ~ l h chlorofmm/iaelhanol/ ~r (ff0:45:!0, v/v) th~ Lonlained 03~% CaCI2-2H20. The ban~ were i ~ a l i z ~ by ~ n o l spry'.. Lane !. SV40 zransform~at: lane ~_ p o ~ a uansfonnant: lane 3. normal rat embryo
bne 4. ad~h meese brain; lane 5. purified ganglioside s~am~'ds.
N-Acetylneuraminic acid (NeuAc) was the m a j o r siaHc acid species detected in the n o r m a l and transformed rat embryo fibroblasts. Only trace amounts of N-gl)~obflneuraminic acid (NeuGc) were found. The total ganglioside Neu~x concentration was similar in the normal (REF52) and the REF52-PyMLV transforman'. but was significantly reduced in the REF52-SV40 tran~Yonnant (Table !). GM3, GM2, G~.I1 and G D I a were the mawr ganglinside species detected in the three cell lines (Fig- 1). T h e ganp~ioside patterns of REF32-1~MVL and REF52-SV40 transformed variants were markedly altered omnpared to the normal TABLE I Toca/~ t'~r,cm~zf/on/n a nomud rat embryo ~ cd/ t3ne IREF52) and in tn~) ~ Iramf~ of Ih~ ~ a p o l ~ a transfonmam eREF52-Ph'MLiO and a SV,.16tranff'onnam (REF525V40)
Cell line
n
REF52 REF52-~.,tLV REF52-SV40
3 5 5
Gan~ NeuAc (tzg/100 mgd~ ~ighz) 144.7+ 10.4 153.8± 9.2 86.I +_ 6.8 *
n, Number c~ separate ~ anab'zed, ~here each sample conrained 48 to 64 mg d~- weight of ~eHs. Values are expressedas mean+-$.F_
* SigniF~cantb"different fromnormalcells at P < 0.01 by t-lest.
R E F 5 2 cells (Fig. ! a n d T a b l e !i). in comparison with the n o r m a l cells, the p e r c e n t a g e distribution of G M 3 was signif'mantly elevated, w h e r e a s the distribution of G M ! a n d G D i a w e r e r e d u c e d in both transformed
variants (Table !1). T h e distribution of G M 2 , however, was reduced in the REF52-PyMLV cells, but was elevated in the REF52-SV40 cells. Although abnormal in p e r c e n t a g e distribution, the concentrations o f G M 3
IVJg. 2. ~ allqlmm-ancee4~ ral embl3a~ I'~roblasls. A, Normal RIEF52 cells, which grow as an orderly aligned monolayer: B. REF52-PyMLVcells, which grow as dumps cr islands: C. REF52-SV40 cells, which grow in a somewhat disorderly manner but are more similar in appt:arance to the normal REF52 cells than to REF52-PyMLVcells (45 × ).
26 TABLE !i Percent d ~ ' ~ / o n and c~'ourat/on of gang//o~des/na norma/rat
Discussien
embryo~ ceg ~ (REF52) and in r*-o dral tramformants of tld~ ~ ; a p o ~ a ~ranyfo~mmu (REF52-PyMLV'~and a SV40 transformara (REF52-SV40)
The influence of viral transformation on cell gang l ~ d e compositinn is not yet clear. In general, the transformation of mouse, hamster, and human cells with DNA paporavimses (either polyoma o r SV40) is associated with elevated levels of GM3 and reduced levels of the more complex ganglioside structures, e.g., GM1 and G D I a [11-14]. Our analysis of the ganglioside compositinn of the PyMLV and SV40 transformants (Table ID does not conform to this general pattern. Transformation of REF52 with PyMLV virus resulted in a variant with an increased concentration of GM~ and decreased copx.entrations of GM2, GMI, and G D Ia, whereas transformatinn with SV40 virus had no effect on GM3 er GM2 concentration, but signifieangy reduced G MI and G D I a concentration. These differences in ganglinside concentration between the variants suggest ,.hat the transformant~ are quite different is some of their cellular characteristics and that these differences may be dependent upon the transforming virus emp|oyed. Varied effects of oncogene transfmmatinn on the ganglinside composition of the 3Y1 rat embryo fibroblast line were also observed [32,33]. in addition to having a differential effect on cell ganglk~side ~ompnsitinn, the SV40 and PyMLV viruses also had a differential effect on cell social behavior, it is interesting that the island-forming PyMLV cells also expressed the highest concentratinn of GM3. A similar phenomenon was obseP,e d by Hirabayashi et al. [34] and Taki et al. [35] in an island-forming rat ascites hepatomas cell line, AH'/974, where GM3 was the sole ganglimide. We also observed a similar phenomenon in a cell fine cultured from a chemically-induced mouse neural tumor, epend)Taoblastuma [36]. This cell line expressed GM3 as the only major ganglioside and also formed d u m p s or islands when grown in tissue culture. In contrast, another chemically-induced mouse neural tumor cell line, CT-2A, expressed more complex ganglinside structures, had little GM3, and grew as a fnsiform m o r ~ w j e r [36]. These findings suggest that cellular adhesion or island formation may be mediated in part by the concentration of GM3.
G~f,~s~le
Perc~ GM3 GM2 GMI GDIa
~
CcH REF52 (n = 3) ~ 29.1+_0.4 29.0_+3.3 7.1+_0.5 43.8+-3.2
REF52-PyMLV REF52-SV40 (n = 51 (n = 5) 57.3_+1.3* 9.4+_0-5* 4.4_+0.3* 28.8+-I.3 *
33.4+_0.9 * 54.5+_0.7 * 3.5+0.3 * 8.6-+ !-2 *
Couccnlxat/on (/~gNcuAc/lfiO w,gdlrywl)
GM3 GM2 GMI GDla
29.2+_2.6 4~-6+7-5 10.1-+O.1 62.8_+0.7
88.6+_7.1* 14.5+-!.2 * 6.8-+0.6* 43.9+-!-0 *
28.9_+2.9 47.0+4.1 3-0+_0.2 * 7-2+_I.! *
Values represem the perc=ntageof toud NeuAc detected by densi~netdc ~ ~ d are expressedas the mean+_S.E. * SigJ,A.~c~n~different titan ~ ceUsat P < 0.01 by t-tesl. and GM2 were actually similar in the normal REF52 and REF52-SV40 ceil lines. In the REF52-PyMLV transformanL however, there was a three-fold elevation of GM3 and a three-fo[d decrease of GM2 concenu'at/on compared to the normal REF52 and REF52-SV40 cell l/nes (Tab!e ID. The structures of the four major ganglinsides in the ceil lines were determined from comparative TLC mobiiit~ with known standards (Fig. 1) and from neuraminidase treatment of the purified ceil line gangliosides. After neuraminidase treatment, the GM3 bands of the three cell lines disappeared and the GD1 a bands were converted to GM1 (data not shown). The neuraminidase treatment had no influence on the GM2 bands. It is important to mention that the ganglinsides in each line migrated as double bands on the HPTLC plates (Fig. 1). ~ can not result from differences in sialic structure since NeuAc is the predominant sialic species in each cell line. Instead, the double bands most likely arise from structural heterogeneity in the ceramide portion o f the molecule, as p r e v i o n s l y described [22,24,31].
Acknowledgment G r o w t h characteristics
Compared to the no..-mal REF52 cells, which grce~ as an orderly aligned monolayer, the REF52-PyMLV ceils grow as clumps or islands (Fig 2). Tlie REF52SV40 cells, on the other hand, grow in a somewhat disorderly manner but are more similar in appearance to the normal REF52 cells than to the REF52-PyMLV cells. The rate of cell growth was approx. 1.5-times faster in the two transformed cell lines than in the normal line.
This research was supported by NIH Grant 24826.
1 Ledeen, R.. Salsman, K. and Cabrera. M. (1968) Biochemistry 7, ~_287-2_795. 2 Ledeen~ R.W. (1985) Tremts Neurosci. 8,169-174.
3 Ledeen, R.W. and Yu. R.K. (19821Methods ~ 1 . 191.
83. 139-
4 Menitl. W.D.. Richardson. C.L. Keenan. T.W. and Morre. D.J. (|978) J. Natl. Cancer Inst. 60. 1313-1327. 5 Hakomori. S. 41981) Annu. Res. B~hem. 50. 733-764. 6 Ando, S. (19831 Neurodlew- InL 5, 507-537. 7 Brgmer. E.G.. Hakomori. S . Bo-~-en-Pop~, D.F. Raincs, E. and Ros~ R. (19'34) J. Biol. Chem. 259.6818-6825. 8 Spiegel S. and Fishman. P.H. (19871 Proc. Natl. Acid. Sci. USA 84,141-145. 9 S~'rh:d. T.N. and Yn. IlK. (19851 Mol. Cell Biochem. 63, 3-102. 10 Hakomori. S. and MurakamL W.T. 41968) Proc. Natl. Acad. Sci. USA 59. 234-261. I I Mora, P.T., Brady. R.O. Bradley. R.M. and McFarland. V.W. 41969) Proc. Natl. A~td. Sci. USA 63. 1290-1296. 12 Brady. R.O. and Mora. P.T. (19"/0) B i ~ h a . Biophys. Aeta 218, 308-319. 13 Yogeeswaran~ G~ Sheinin. R_ ~Vherrelt. J . i i and Murray. R.FL 41972) J. B-~I. Chew_ 247. 5146--5158. 14 Mora. P.T.. Cumar. F.A. and Brady. RO. (19711 Virology 46. 60-72. 15 HakomorL S. Teallmr. C. and Andrews. H. 41968) Biochem. Bk~l~'s. Res. Con'anun. 33. 563-568. 16 Brooks. S.E. Amsterdam. D ~ Hoffman. L.M_ Adachi. M. and Schneck, L (Iq79} $. Cell Sci. 38, 211-?.223. 17 Coleman. EL.. F"~hman. P.H.. Brady. I l O . and Todaro. G J . (19751J. Biol.Chew_ 250, 55-60. 18 HakomorL, S . Saito. 1". and Vogt. P.K. (19711 Virology 44. 609621. 19 Logan. J.. Nh:olas, J.C~ Topp. W.C~ Girard. M_ Shenk. T. and I..~'~t~,/~J. (itJ~i) _Virology 115. 419--422.
2(I Donoghue. D J . Anderson. C., Hunter. T. and Kaplan, P.L. (19841 Nature (London) 308, 74~-750. 21 McClure, D.B.. Hightower. MJ. and Topp. W.C. (19821 Cold Spring Harbor Conf. Cell Proliferation. 9. 345-364. 22 Seyfried. N.T., Glascr, G.H. and Yu, R.K. 41978) J. Neurochem. 31.21-27. 23 Seyfried. T.N. IIoh. T.. Glaser0 G.H. Miyazawa, N. and Yo. R.K. (1981) Brain Res. 261,429-436. 24 Seyfried. T.N. (1987) Dev. Biol. 123. 286-291. 25 Ueno. K.. Ando. S. and Yu. R.K. (1978) ]. Lipid Res. 19. 863-871. 26 Yu, IlK. and Ledecn, R.W. (1970) J. !Apid Res. II. 506-516. 27 Ando. S.. Chang, N.C., and Yu. R.K. (1078) Anal. Biocbem. 89. 437-450. 28 Svcnnerholm. L. (1957) Biochim. Bk,,phys. Acta 24. e~4-612. 29 Ando. S. and Yu. IlK. (19771J. Biol. Chem. 252. 6247~,250. 30 Ando. S. and Yu, R.IC (19791J. Biol. Chem. 254. ! ~ 4 - 1 ~ ' ~ 9 . 31 Chnu. K.H., Ambers, S.A., and Jungalwala, EB. (1979) L New rochew- 33. 8~,3-873. 32 Nakaishi. H . Sanai. Y.. Shiroki. K. and Nagai. Y (19881BiochewBiophys. Res. Commun. 15t!. 760-765. 33 Nakaishi. H . Sanai. Y., Shibuya. M. and Nagai, Y (19881BiocbewBiophys. Res. Commun. 150, 766-774. 34 Hirabay&shi, Y., Taki, T., Matsumoto, M. and Kojima, !~ (19781 Biochiw- Biophys. Acta 529. 96-105. 35 Taki, T., Hirabayashi, Y.. Ishiwata, Y.. Matsumoto, M. and Kojima. [~ (1979,~ Biochim. Bioph)~. Acta 5 7 ! 113-120. 36 Seyfried. T.N.. EI-Abbadi, M. and Roy. M.L (19921 Mol. Chem. Neuropathol.. in press.
23
© 1992ELsevierScience FublishcrsB.V. All rightsreserved0167-4889/92/$05.00
Altered ganglioside composition in virally transformed rat embryo fibroblasts H o n g w e i Bai, J o s e p h O r l a n d o a n d T h o m a s N. S e y f r i e d Department of Biology; Boston College, Chestnut Hill, MA (USA)
(Rece/ved 13January 1992l
Keywords: Gangl/osid¢;Pob'omatransformant;SV40transformant;Cultured cell; (Rat embryofibrobtast) The composition of ganglinsides was examined in a normal rat embryo fibroblast cell line (REF52) and in two viral transformants: a Imllanna trar~foi'~ua~'at(REF52-PyMLV) and a simian viral 40 transformant (REF52-SV40). The distribution of gangl/osides in the cell lines was determined using gas-liquid chromatography and high-performance thin-layer chromatography. N-acetylneuranlialc acid was the predominant sialic acid spe~es detected in the three cell lines. The total ganglinside concentration (ttg/100 mg d~ weight of cells) in the normal, I~MLV, and SV40 lines was 144.7 4-10.4. 153.84- 9.2, and 86.1 +_68, respect~'eb-. Ganglinsides GM3, GM2, GMI, and GDIa were the major species in the normal and transformed lines. The distribution of these gangJiasides, however, differed markedly between the normal and the transformed lines and also between the transformed lines themselves.The transformed cells also differed from the normal cells in growth rate, morphology, and social behavior. The cell line with highest GM3 content (PyMLV) formed islands, whereas the nornmt and SV40 cell lines, which had lower GM3 levels, grew as monelayers. The findings suggest that PyMLV and SV40 transformation can have multiple and differem effects on cellular ganglinside distribution and growth behavior.
latruductiea Gangliesides are siafic acid-containing glycnsphingolipids that are recognized as ubiquitous constituents of mammalian plasma membranes [1-3]. These molecules are thought to participate in cell recognition, growth control, differentiation, and tumorigenesis [4-
9].
Since it was first discovered that ganglioside alterations are closely associafed wi.*h cell transformation [10], considerable work has been done to characterize these alterations. The influence of various transforming agents, such as DNA or RNA viruses, =hemitr,d carcinogens, and X-ray irradiation on gaoglioside composition were studied on a variety of cell types [11-18]. There is considerable controversy, however, on the ganglioside changes associated with viral transformation of mammalian cells. The ganglioside changes observed can depend on both the transforming virus used and on the cell type transformed. In general, cellular transformation with DNA viruses, either the polyoma marine leukemia virus (I~MLV) or the simian virus 40
Conespondence to: T.N. Se~'fried.Department of Bidogy. Boston College. ChestnutHill MA 02167. USA.
(SV40), is associated with an elevation of ganglioside GM3 [11-13,16] and reductions of the more complex gangliosides, e.g., GM2 [13,14], GMI and G D l a [1114]. There are, however, exceptions to this general trend [10-16]. To better define ganghoside changes associated with viral transformation, we have characterized the ganglinside composition in a normal rat embryo fibroblast cell line (REF52) and in two viral transformants of the cell line: REF52-PyMLV and REF52-SV40. Ours is the first study of ganglioside changes associated with polyoma and SV40 viral transformation in rat embryo fibroblasts. Materials and Methods Cells and cell culture
The REF52 cell line was established from fibroblasts of a 14-day gestation Fischer rat embryo [19]. This cell line expresses contact inhibition and is nontumorigenic. The REF52-1~MLV variant was produced after transformation ~ the polyoma virus middle T gene [20] and the REF52-SV40 variant was produced after transformation by the SV40 virus [21]. Both the REF52PyMLV and the REF52-SV40 transformants lack contact inhibition and are tumorigenic. Wild-type REF52 and transformants were obtained from Dr. J.IC Cben (Chang Gun Medical College, Taipei, Taiwan).
Ceils (1- |04/ore 2 for REF52 variant and 2104/cm 2 for REF52-PyMLV arid REF52-SV40 variants) were seeded in 150 cm 2 flasks and grown in 30 nd of a basal medium consisting of 25% Ham's !:-12 and 75% DMEM and supplemented with 10% fetal bovine serum. The ceils were cultured in a humidified atmosphere of 95% air and 5% CO 2 at 37~C and the medium was changed every two days during the growing period. V~en the ceils reached confluence (about 6 days), they were treated with 0.1% trypsin, harvested by cenzrifugadon and washed three times with cold phosphate-buffered saline. The washed ceils were bophilized and stored at - 2 0 ~ C . Gang//as/de ana6,s/s Gan~ioside~ were ;:~c~!a:cd and purified from lyhophifized ceD puwders according to the methods described previously [22-24]. The gangliosides were treated with mild base (0.1 M NaOH for 1 h) and desalted using Sephadex G-50 column chromatography [23,25]. The concentration o£ ganglioside Nacctylneura.,~c acid (NeuAc) and N-glycolylneuraminic acid (NeuGc) ~ s deter~ained using the gas-liquid chromatographic method of Yu and Ledeen [26]. The sialic acid concentration was expressed as p g p e r 109 mg dry weighL The quantitative anab~-is of individual ganglioside species was performed according to the procednre of Ando et al. [27]. In this procedure, 10 × 10 cm thin-layer silica gel plates were used (Whatman HP-K high-performance silica gel). The conditions for HPTLC development are gk'en in Fig. 1. The ganglioside bands were visualized by spraying the dried plates with a resorcinobHCl reagent [28] and by baking the plates at 95°C for 30 rain. The percent gangliuside distribution was determined by the direct densitometric scanning method of Ando et al. [27]. The gangliosides of the normal and transformed rat embryo fibrublasts were identified by comparison with ganglinside s~tandards isolated from the cerebral cortex of normal adult mice and with GM4, GM3, GM2, GM!, GD3, GDIa, GTIa, GDlb, G F l b and G O l b purified and isolated from mouse, bovine and, human tissues by the method of Ando and Yu [29,30]. A preliminary structural characterization of the ganglios[des in both the normal and transformed rat embryo fibroblasts was performed by a neuraminidase treatment (EC 3.2.1.18, type VI from Clostridimn p e r fr/ngens, Sigma), according to the procedure of Ando and Yu [30]. Purified ganglioside samples (containing 5 /tg of sialic acid) from each cell line were mixed with 100 gl of a fresh neuraminidase solution (1 unit in 1 ml of 0.1 M sodium acetate buffer, pH 5) and incubated at 3"PC for 24 h. After base ~reatment (0.1 M NaOH for 1 h) and desalting, the treated gangliusides were spotted on HPTLC plates and developed as described above.
.....7
,
GM4 GM3 GM2
W
GMI
i!~!!~:!:~:¸¸
GO3 GDla GTla
:!!:!!!:!!!!~
GDlb GTlb G~lb
1
$
2
4
S
~g. I. Th~t-la3,~*r ~ of ganglioskles in normal and ,~ams-Uansform~rat eml~3~ fi~robLasls.Apprmdmately 1.5 ~g of g~m~ Neu_Ac~ s spoued for each sample. The plate ~;~s de~etoped by one ascendh~ elufion ~ l h chlorofmm/iaelhanol/ ~r (ff0:45:!0, v/v) th~ Lonlained 03~% CaCI2-2H20. The ban~ were i ~ a l i z ~ by ~ n o l spry'.. Lane !. SV40 zransform~at: lane ~_ p o ~ a uansfonnant: lane 3. normal rat embryo
bne 4. ad~h meese brain; lane 5. purified ganglioside s~am~'ds.
N-Acetylneuraminic acid (NeuAc) was the m a j o r siaHc acid species detected in the n o r m a l and transformed rat embryo fibroblasts. Only trace amounts of N-gl)~obflneuraminic acid (NeuGc) were found. The total ganglioside Neu~x concentration was similar in the normal (REF52) and the REF52-PyMLV transforman'. but was significantly reduced in the REF52-SV40 tran~Yonnant (Table !). GM3, GM2, G~.I1 and G D I a were the mawr ganglinside species detected in the three cell lines (Fig- 1). T h e ganp~ioside patterns of REF32-1~MVL and REF52-SV40 transformed variants were markedly altered omnpared to the normal TABLE I Toca/~ t'~r,cm~zf/on/n a nomud rat embryo ~ cd/ t3ne IREF52) and in tn~) ~ Iramf~ of Ih~ ~ a p o l ~ a transfonmam eREF52-Ph'MLiO and a SV,.16tranff'onnam (REF525V40)
Cell line
n
REF52 REF52-~.,tLV REF52-SV40
3 5 5
Gan~ NeuAc (tzg/100 mgd~ ~ighz) 144.7+ 10.4 153.8± 9.2 86.I +_ 6.8 *
n, Number c~ separate ~ anab'zed, ~here each sample conrained 48 to 64 mg d~- weight of ~eHs. Values are expressedas mean+-$.F_
* SigniF~cantb"different fromnormalcells at P < 0.01 by t-lest.
R E F 5 2 cells (Fig. ! a n d T a b l e !i). in comparison with the n o r m a l cells, the p e r c e n t a g e distribution of G M 3 was signif'mantly elevated, w h e r e a s the distribution of G M ! a n d G D i a w e r e r e d u c e d in both transformed
variants (Table !1). T h e distribution of G M 2 , however, was reduced in the REF52-PyMLV cells, but was elevated in the REF52-SV40 cells. Although abnormal in p e r c e n t a g e distribution, the concentrations o f G M 3
IVJg. 2. ~ allqlmm-ancee4~ ral embl3a~ I'~roblasls. A, Normal RIEF52 cells, which grow as an orderly aligned monolayer: B. REF52-PyMLVcells, which grow as dumps cr islands: C. REF52-SV40 cells, which grow in a somewhat disorderly manner but are more similar in appt:arance to the normal REF52 cells than to REF52-PyMLVcells (45 × ).
26 TABLE !i Percent d ~ ' ~ / o n and c~'ourat/on of gang//o~des/na norma/rat
Discussien
embryo~ ceg ~ (REF52) and in r*-o dral tramformants of tld~ ~ ; a p o ~ a ~ranyfo~mmu (REF52-PyMLV'~and a SV40 transformara (REF52-SV40)
The influence of viral transformation on cell gang l ~ d e compositinn is not yet clear. In general, the transformation of mouse, hamster, and human cells with DNA paporavimses (either polyoma o r SV40) is associated with elevated levels of GM3 and reduced levels of the more complex ganglioside structures, e.g., GM1 and G D I a [11-14]. Our analysis of the ganglioside compositinn of the PyMLV and SV40 transformants (Table ID does not conform to this general pattern. Transformation of REF52 with PyMLV virus resulted in a variant with an increased concentration of GM~ and decreased copx.entrations of GM2, GMI, and G D Ia, whereas transformatinn with SV40 virus had no effect on GM3 er GM2 concentration, but signifieangy reduced G MI and G D I a concentration. These differences in ganglinside concentration between the variants suggest ,.hat the transformant~ are quite different is some of their cellular characteristics and that these differences may be dependent upon the transforming virus emp|oyed. Varied effects of oncogene transfmmatinn on the ganglinside composition of the 3Y1 rat embryo fibroblast line were also observed [32,33]. in addition to having a differential effect on cell ganglk~side ~ompnsitinn, the SV40 and PyMLV viruses also had a differential effect on cell social behavior, it is interesting that the island-forming PyMLV cells also expressed the highest concentratinn of GM3. A similar phenomenon was obseP,e d by Hirabayashi et al. [34] and Taki et al. [35] in an island-forming rat ascites hepatomas cell line, AH'/974, where GM3 was the sole ganglimide. We also observed a similar phenomenon in a cell fine cultured from a chemically-induced mouse neural tumor, epend)Taoblastuma [36]. This cell line expressed GM3 as the only major ganglioside and also formed d u m p s or islands when grown in tissue culture. In contrast, another chemically-induced mouse neural tumor cell line, CT-2A, expressed more complex ganglinside structures, had little GM3, and grew as a fnsiform m o r ~ w j e r [36]. These findings suggest that cellular adhesion or island formation may be mediated in part by the concentration of GM3.
G~f,~s~le
Perc~ GM3 GM2 GMI GDIa
~
CcH REF52 (n = 3) ~ 29.1+_0.4 29.0_+3.3 7.1+_0.5 43.8+-3.2
REF52-PyMLV REF52-SV40 (n = 51 (n = 5) 57.3_+1.3* 9.4+_0-5* 4.4_+0.3* 28.8+-I.3 *
33.4+_0.9 * 54.5+_0.7 * 3.5+0.3 * 8.6-+ !-2 *
Couccnlxat/on (/~gNcuAc/lfiO w,gdlrywl)
GM3 GM2 GMI GDla
29.2+_2.6 4~-6+7-5 10.1-+O.1 62.8_+0.7
88.6+_7.1* 14.5+-!.2 * 6.8-+0.6* 43.9+-!-0 *
28.9_+2.9 47.0+4.1 3-0+_0.2 * 7-2+_I.! *
Values represem the perc=ntageof toud NeuAc detected by densi~netdc ~ ~ d are expressedas the mean+_S.E. * SigJ,A.~c~n~different titan ~ ceUsat P < 0.01 by t-tesl. and GM2 were actually similar in the normal REF52 and REF52-SV40 ceil lines. In the REF52-PyMLV transformanL however, there was a three-fold elevation of GM3 and a three-fo[d decrease of GM2 concenu'at/on compared to the normal REF52 and REF52-SV40 cell l/nes (Tab!e ID. The structures of the four major ganglinsides in the ceil lines were determined from comparative TLC mobiiit~ with known standards (Fig. 1) and from neuraminidase treatment of the purified ceil line gangliosides. After neuraminidase treatment, the GM3 bands of the three cell lines disappeared and the GD1 a bands were converted to GM1 (data not shown). The neuraminidase treatment had no influence on the GM2 bands. It is important to mention that the ganglinsides in each line migrated as double bands on the HPTLC plates (Fig. 1). ~ can not result from differences in sialic structure since NeuAc is the predominant sialic species in each cell line. Instead, the double bands most likely arise from structural heterogeneity in the ceramide portion o f the molecule, as p r e v i o n s l y described [22,24,31].
Acknowledgment G r o w t h characteristics
Compared to the no..-mal REF52 cells, which grce~ as an orderly aligned monolayer, the REF52-PyMLV ceils grow as clumps or islands (Fig 2). Tlie REF52SV40 cells, on the other hand, grow in a somewhat disorderly manner but are more similar in appearance to the normal REF52 cells than to the REF52-PyMLV cells. The rate of cell growth was approx. 1.5-times faster in the two transformed cell lines than in the normal line.
This research was supported by NIH Grant 24826.
1 Ledeen, R.. Salsman, K. and Cabrera. M. (1968) Biochemistry 7, ~_287-2_795. 2 Ledeen~ R.W. (1985) Tremts Neurosci. 8,169-174.
3 Ledeen, R.W. and Yu. R.K. (19821Methods ~ 1 . 191.
83. 139-
4 Menitl. W.D.. Richardson. C.L. Keenan. T.W. and Morre. D.J. (|978) J. Natl. Cancer Inst. 60. 1313-1327. 5 Hakomori. S. 41981) Annu. Res. B~hem. 50. 733-764. 6 Ando, S. (19831 Neurodlew- InL 5, 507-537. 7 Brgmer. E.G.. Hakomori. S . Bo-~-en-Pop~, D.F. Raincs, E. and Ros~ R. (19'34) J. Biol. Chem. 259.6818-6825. 8 Spiegel S. and Fishman. P.H. (19871 Proc. Natl. Acid. Sci. USA 84,141-145. 9 S~'rh:d. T.N. and Yn. IlK. (19851 Mol. Cell Biochem. 63, 3-102. 10 Hakomori. S. and MurakamL W.T. 41968) Proc. Natl. Acad. Sci. USA 59. 234-261. I I Mora, P.T., Brady. R.O. Bradley. R.M. and McFarland. V.W. 41969) Proc. Natl. A~td. Sci. USA 63. 1290-1296. 12 Brady. R.O. and Mora. P.T. (19"/0) B i ~ h a . Biophys. Aeta 218, 308-319. 13 Yogeeswaran~ G~ Sheinin. R_ ~Vherrelt. J . i i and Murray. R.FL 41972) J. B-~I. Chew_ 247. 5146--5158. 14 Mora. P.T.. Cumar. F.A. and Brady. RO. (19711 Virology 46. 60-72. 15 HakomorL S. Teallmr. C. and Andrews. H. 41968) Biochem. Bk~l~'s. Res. Con'anun. 33. 563-568. 16 Brooks. S.E. Amsterdam. D ~ Hoffman. L.M_ Adachi. M. and Schneck, L (Iq79} $. Cell Sci. 38, 211-?.223. 17 Coleman. EL.. F"~hman. P.H.. Brady. I l O . and Todaro. G J . (19751J. Biol.Chew_ 250, 55-60. 18 HakomorL, S . Saito. 1". and Vogt. P.K. (19711 Virology 44. 609621. 19 Logan. J.. Nh:olas, J.C~ Topp. W.C~ Girard. M_ Shenk. T. and I..~'~t~,/~J. (itJ~i) _Virology 115. 419--422.
2(I Donoghue. D J . Anderson. C., Hunter. T. and Kaplan, P.L. (19841 Nature (London) 308, 74~-750. 21 McClure, D.B.. Hightower. MJ. and Topp. W.C. (19821 Cold Spring Harbor Conf. Cell Proliferation. 9. 345-364. 22 Seyfried. N.T., Glascr, G.H. and Yu, R.K. 41978) J. Neurochem. 31.21-27. 23 Seyfried. T.N. IIoh. T.. Glaser0 G.H. Miyazawa, N. and Yo. R.K. (1981) Brain Res. 261,429-436. 24 Seyfried. T.N. (1987) Dev. Biol. 123. 286-291. 25 Ueno. K.. Ando. S. and Yu. R.K. (1978) ]. Lipid Res. 19. 863-871. 26 Yu, IlK. and Ledecn, R.W. (1970) J. !Apid Res. II. 506-516. 27 Ando. S.. Chang, N.C., and Yu. R.K. (1078) Anal. Biocbem. 89. 437-450. 28 Svcnnerholm. L. (1957) Biochim. Bk,,phys. Acta 24. e~4-612. 29 Ando. S. and Yu. IlK. (19771J. Biol. Chem. 252. 6247~,250. 30 Ando. S. and Yu, R.IC (19791J. Biol. Chem. 254. ! ~ 4 - 1 ~ ' ~ 9 . 31 Chnu. K.H., Ambers, S.A., and Jungalwala, EB. (1979) L New rochew- 33. 8~,3-873. 32 Nakaishi. H . Sanai. Y.. Shiroki. K. and Nagai. Y (19881BiochewBiophys. Res. Commun. 15t!. 760-765. 33 Nakaishi. H . Sanai. Y., Shibuya. M. and Nagai, Y (19881BiocbewBiophys. Res. Commun. 150, 766-774. 34 Hirabay&shi, Y., Taki, T., Matsumoto, M. and Kojima, !~ (19781 Biochiw- Biophys. Acta 529. 96-105. 35 Taki, T., Hirabayashi, Y.. Ishiwata, Y.. Matsumoto, M. and Kojima. [~ (1979,~ Biochim. Bioph)~. Acta 5 7 ! 113-120. 36 Seyfried. T.N.. EI-Abbadi, M. and Roy. M.L (19921 Mol. Chem. Neuropathol.. in press.
