Cell Biology International 1999, Vol. 23, No. 2, 91–95Article No. cbir.1999.0351, available online at http://www.idealibrary.com on

GLYCOLIPID GLYCOSYLTRANSFERASE ACTIVITIES DURING EARLYDEVELOPMENT OF XENOPUS: EFFECT OF RETINOIC ACID

FEDERICA ROSSI1, ROSALBA GORNATI1, ANGELA M. RIZZO1, LORETTA VENTURINI2,GIOVANNI BERNARDINI2 and BRUNO BERRA1*

1Istituto di Fisiologia Generale e Chimica Biologica, Universita di Milano, 2 Via Trentacoste, I-20134 Milano, Italyand 2Dipartimento di Biologia Strutturale e Funzionale, Universita dell’insubria, 3 Via Dunant, I-21100 Varese,

Italy

Received 18 August 1998; accepted 20 January 1999

Retinoic acid (RA) plays an important role in differentiation stage in which it also influencesglycoconjugate metabolism. Previous work in our laboratory has shown that treatment withRA modifies glycolipid synthesis and distribution in total Xenopus embryos during develop-ment. In this study we have investigated the activity of the following anabolic enzymesinvolved in glycolipid biosynthesis: sialyltransferase-1 (SAT-1), GM3(�1,4)-N-acetyl-galactosaminyltransferase (GalNAcT-1) and LacCer(�1,3)N-acetylglucosaminyltransferase(GlcNAcT-1). These enzymes are located at the branching point of lactosylceramide (Lc2)metabolism. Enzyme activities were assayed after treatment with different doses of RA addedexogenously to the medium during the first 7 days of Xenopus embryo development. Our resultsshow that RA activates GlcNAcT-1, the enzyme that drives Lc2 to the glycolipids of thelacto-series, and SAT-1 that inserts Lc2 in the ganglio-series pathway. These data support ourprevious analysis of glycolipid pattern in Xenopus embryos after RA treatment (Rizzo et al.,1995; Cell Biol Int 19: 895–901) indicating a possible correlation between the distribution ofglycolipids and the enzymes involved in their metabolism. � 1999 Academic Press

K: Xenopus laevis; glycolipids, retinoic acid; development

*To whom correspondence should be addressed.

INTRODUCTION

Retinoic acid (RA) is a pivotal molecule for theprocesses of reproduction, growth and develop-ment as it plays an important role in the control ofcell proliferation and differentiation (Napoli,1990). The biological effects of RA are mediated bytwo classes of DNA-binding transcription factors(RAR and RXR) as well as by cytoplasmic bindingproteins (CRABP and CRBP) (Leid et al., 1992;Ellinger-Ziegelbauer and Dreyer, 1993). The differ-entiation in human embryonal carcinoma cellsinduced by RA treatment was shown to modifythe glycolipid distribution. Fenderson (1987) andWenk et al. (1994) reported a shift from themolecules of the globo-series characterized bythe sequence Cer-Glc-Gal-Gal to molecules of the

1065–6995/99/020091+05 $30.00/0

lacto-and ganglio-series (Cer-Glc-Gal-GlcNAc andCer-Glc-Gal-GalNAc respectively). In Xenopusembryos, teratogenic concentrations of RA wereshown to induce a decrease in neutral glycolipidcontent and a parallel increase of gangliosideamount (Rizzo et al., 1995). Lactosylceramide(Lc2, Cer-Glc-Gal) is the common substrate forthe biosynthesis of different members of the glyco-lipid families (both of the ganglio- and lacto-series,neutral or sialylated); for example GA2 (Cer-Glc-Gal-GalNAc), Lc3 (Cer-Glc-Gal-GlcNAc)and GM3 (Cer-Glc-Gal-NeuAc). The differentpathways are characterized by the additionof a first residue of N-acetylgalactosamine(GalNAc), N-acetylglucosamine (GlcNAc) andN-acetylneuraminic acid, respectively. In thepresent paper, the effects of RA on the activity ofthe enzymes GalNAcT-1, GlcNAcT-1 and SAT-1,

� 1999 Academic Press

92 Cell Biology International, Vol. 23, No. 2, 1999

responsible for the above-mentioned reactions arereported. Our aim was to possibly correlate theseactivities with the already observed modificationsin glycolipid distribution.

MATERIALS AND METHODS

Embryo culture

In vitro fertilization was performed using fourdifferent females (counter marked 1 to 4) as pre-viously described (Bernardini et al., 1994). Briefly,gonadotropin-injected females were made to layeggs that were immediately inseminated with spermsuspension. Successful insemination was detectedwhen the eggs were oriented with the animal poleup. All the irregularly segmented embryos wereremoved. Embryos, kept in a thermostatic chamberat 23�0.5�C, were collected every 24 h at stages22, 35/36, 41, 45, 46 and 47/48, according to theNieuwkoop and Faber (1956) Xenopus stagingtable. Previous preliminary experiments had shownthat a very large variability in glycolipid profile andenzyme activities was present in embryos fromdifferent mothers. In order to reduce this variabilityall the embryos derived from a single mother, withor without RA treatment, were grouped togetherand the comparisons were made within groups ofembryos derived from a single mother. The groupswere differentiated by the different treatments theembryos received during development (see below).

Retinoic acid treatment

Retinoic acid (Sigma Chemical Co) stock solution(about 3 m) was prepared in dimethylsulfoxide(DMSO), stored at �20�C and dissolved inFETAX solution (mg/l): NaCl 625, NaHCO3 96,KCl 30, CaCl2 15, CaSO4 · 2H2O 60 and MgSO470. The embryos obtained from each female weredivided into three groups: control (C), treated with15 n RA (T1) and treated with 30 n RA (T2).The treatment started at stage 9, about 8 h afterfertilization; control (FETAX) and RA solutionswere changed every 24 h. For each day examined,about 50 embryos were collected for each femaleand each treatment group (C, T1 and T2), freedof the excess water, and stored in 1 ml HEM(20 m HEPES, 0.1% 2-mercaptoethanol and1 m EDTA, pH 7) at �20�C until enzymedetermination.

Substrates for enzyme determinations

GM3 was purchased from INALCO (Milan, Italy).Lc , CMP-sialic acid, UDP-N-acetylglucosamine

2

and UDP-N-acetylgalactosamine were purchasedfrom Sigma (St Louis, MO, U.S.A.). CMP-[14C]sialic acid, UDP-[14C]N-acetyglucosamine andUDP-[14C]N-acetylgalactosamine were obtainedfrom NEN-Du Pont GmbH (Bad Homburg,Germany).

Preparation of crude enzymes

Enzyme preparation was performed at 0–5�Caccording to Basu (Basu et al., 1982). Briefly, about50 embryos were homogenized in 1 ml of SHEM(0.32 sucrose in 20 m HEPES). The homog-enate was centrifuged at 100,000�g for 1 h at4�C, the pellet was suspended in 0.5 ml 20 mcacodylate-HCl buffer containing 5 m benzami-dine, 0.1% 2-mercaptoethanol and 5% glycerol(pH 6.4) and assayed for protein content accordingto Peterson (1977).

Enzyme assay

The activity of glycosyltransferases was assayedaccording to Basu et al. (1987) with some slightmodifications.

Sialyltransferase-1 (SAT-1). The final incubationmixture contained the following components in50 �l 0.2 cacodylate-HCl buffer (pH 6.4): 50 �gTriton CF 54, 0.02 �moles lactosylceramide,0.03 �moles CMP[14C]NeuAc and 100 �g proteinsfrom the crude enzyme preparation. The mixturewas incubated for 2 h at 37�C and the reaction wasstopped by adding 20 �l methanol. The wholemixture was spotted on Whatman 3MM paper andresolved by descending chromatography in 1% Naborate to separate the products of the reaction.

GM3(�1,4)-N-acetylgalactosaminyltransferase(GalNAcT-1). The reaction mixture contained thefollowing components in 50 �l buffer (0.2 HEPES pH 7.6 plus 10 m MnCl2, final concen-trations): 60 �g Triton CF54, 0.02 �moles GM3,0.02 �moles [14C] UDPGalNAc, 100 �g of proteinsfrom the crude enzyme preparation. The assayprocedure for the analysis of the reaction productswas similar to that used for SAT-1 determination.

LacCer(�1,3)N-acetylglucosaminyltransferase(GlcNAcT-1). The reaction mixture contained thefollowing components in 50 �l buffer (0.2 HEPES pH 7.6 containing 5 m MnCl2, final con-centrations): 50 �g Triton CF 54, 0.02–0.04 �moleslactosylceramide, 0.02 �moles UDP[14C]GlcNAc,

Cell Biology International, Vol. 23, No. 2, 1999 93

100 �g of proteins from the crude enzyme prep-aration. The mixture was incubated for 4 h at 37�Cand the products were separated on 3MM paperdescending chromatography, using sodium borate1% as solvent system.

RESULTS

The treatments with 15 or 30 n RA caused a highpercentage of embryo malformations includingoedema, axial shrinking, acephaly or deformities tothe intestines and to the eyes (Fig. 1). However, themortality rate was low. These teratological aspectsare in agreement with the results describing theeffects of exogenously added RA in Xenopusembryos (Drysdale and Crawford, 1994); the datareported by these authors however were onlyrelated to morphological aspects of RA treatment.

The activities of GalNAcT-1, GlcNAcT-1, andSAT-1, followed during the first week of normalembryonal development, showed a biphasic curvewith a maximum around the fourth day of devel-opment as already reported in a previous paperfrom our laboratory (Gornati et al., 1997). Anexample of changes in enzyme activities due to RAtreatment, compared with the control, is reportedin Figure 2, in which SAT-1 and GlcNAcT-1activities are tested in the embryos derived frommother number 1. The study has been performedon control and treated (with two doses of RA)embryos derived from three or four differentmothers, referred to as numbers 1, 2, 3 and 4. Theshape of the enzyme activity curves varies with thedifferent mothers of each embryo batch. For thisreason, the results obtained after treatment withRA are expressed as an increase or a decrease inenzyme activity at day 4 (maximum peak), whereeach batch accounts for its own control.

Treatments with 15 n RA caused practically noeffect on GalNAcT-1 in female 1 and a decrease ofactivity in female 3 (Fig. 3, panel A, dark bars).

When the embryos were treated with 30 n RA adifferent decrease in three out of four samples wasobserved (Fig. 3, panel A, light bars). In all thetested samples, GlcNAcT-1 activity is increased bytreatment with 15 n RA (Fig. 3, panel B, darkbars), while 30 n RA caused a decrease in peakactivity in two batches and only a slight increase inthe other two (Fig. 3, panel B, light bars). Treat-ment with 15 n RA increased the activity ofSAT-1 (Fig. 3, panel C, dark bars); 30 n RA didnot cause any significant modification in twosamples; in one a decrease was observed, while inthe last the activity was higher (Fig. 3, panel C,light bars).

90

0

Mal

form

atio

ns

(%)

50

807060

40302010

Embryosmalformed

Output ofcrystalline

Irregulareye border

Absence ofcrystalline

Other eye

malformations

Microphthalmia

CT1T2

Fig. 1. Malformations induced by RA in Xenopus embryos,observed at the stage 47/48 of development (6 days post-fertilization).

7

0.05

–0.011

Days

Ch

ange

in e

nzy

me

acti

vity

(n

mol

es/m

g pr

otei

n/h

)

0.03

0.04

0.02

0.01

0

3 4 5

GlcNAcT-1

7

0.3

–0.051

Days

0.15

0.2

0.1

0.05

0

3 4 5

SAT-1

0.25 0 nM RA15 nM RA30 nM RA

Fig. 2. Changes in SAT-1 and GlcNAcT-1 activities due to thetreatment with RA, compared with the control, in female 1.

DISCUSSION

Glycolipids are involved in cell social behaviourand therefore play a pivotal role in differentiationand embryogenesis, particularly of the centralnervous system (Zeller and Marchase, 1992; Nagaiand Iwamori, 1995); cell differentiation is ac-companied by sequential changes in glycolipidexpression (Taga et al., 1995). During cell growth,

94 Cell Biology International, Vol. 23, No. 2, 1999

0.2

–0.2

(C)

Ch

ange

s in

pea

k ac

tivi

ty (

enzy

me

acti

vity

RA

– e

nzy

me

acti

vity

con

trol

)

1

0.1

0

–0.1

2 3 41 2 3

0.08

–0.08

(B)

1

0.04

0

–0.04

2 3 41 2 3

0.1

–0.1

(A)

1

0.05

0

–0.05

2 3 41 2 3

Fig. 3. Peak activity modifications relieved at the fourth dayafter RA treatment of Xenopus embryos of: (A) GalNAcT-1(nmole UDP[C14]GalNAc/mg protein/h); (B) GlcNAcT-1(nmole UDP[C14]GlcNAc/mg protein/h); (C) SAT-1 (nmoleCMP[C14]NeuAc/mg protein/h). Average peak activitiesin control embryos are: 0.113 nmole/mg protein/h forGalNAcT-1; 0.039 nmole/mg protein/h for GlcNAcT-1;0.188 nmole/mg protein/h for SAT-1. Dark and light bars referto 15 and 30 n RA, respectively. 1, 2, 3, 4: batches of embryosfrom different mothers.

CER GlcCer Gal-Glc-CerLc2

NeuAc-Gal-Glc-CerGM3 GD3 GT3

GalNac-Gal-Glc-CerGA2 GM2 GD2 GT2

GA1 GM1 GD1b GT1c

GM1b GD1a GT1b GQ1c

GD1c GT1a GQ1b GP1c

(a) (b) (c)

GalNacT-1

GalT-3

SAT-4

SAT-5

SAT-2 SAT-3SAT-1GalT-2GltTase

GalT-6 GlucNAcT-1

Gal-Gal-Glc-CerGb3

GlcNAc-Gal-Glc-CerLc3

Fig. 4. General scheme of metabolic pathways involved inglycolipid metabolism.

glycolipids or their degradation products werereported to act as potential regulatory molecules insignal transduction pathways (Hannun and Bell,1989; Okazaki et al., 1989; Hakomori, 1990;Hannun and Obeid, 1995). On this line, we havepreviously shown in our laboratory that activesynthesis of glycolipids and interesting modi-fications in their content occur during the earlydevelopment of Xenopus embryos (Gornati et al.,1995); in this model exogenous exposure to RA

modifies the glycolipid distribution inducing a shiftfrom neutral glycolipids toward gangliosides and aredistribution of the species inside a particular classof these glycoconjugates (Rizzo et al., 1995).

On the other hand, in embryonic carcinomacells, differentiation induced by RA treatment alsoalters glycolipid distribution (Fenderson et al.,1987; Wenk et al., 1994).

In the present work we have analysed the effectsof RA on the activity of the glycosyltransferasessituated at the branching point of the metabolicpathways involving lactosylceramide as a substrate(Fig. 4). Our results demonstrate, in the embryosexposed to 15 n RA, a general activation ofGlcNAcT-1, the enzyme that drives Lc2 to thelacto-series and of SAT-1 that inserts Lc2 in theganglio-series pathway (Figs 2 and 4). The datasupport our previous results (Rizzo et al., 1995),where in Xenopus embryos an increase of ganglio-side content concomitant with a decrease of neutralglycolipid was found after RA treatment. Due tothe particular pattern of glycolipid presented byXenopus embryos (mono- and di-hexosylceramidesaccount for more than 80% of the total glycolipidcontent) it is not possible to discriminate betweenglycolipids of the globo- and lacto-series. In thisrespect we cannot speculate if the previousobserved decrease of neutral glycolipid would bedue to the decrease of the globo-members partiallycounteracted by an increase of the lacto-molecules.Therefore no correlation can be made with theenzyme activity of GlcNAcT-1 and the neutralglycolipid distribution. However, the possibilitythat RA exposure would result in a diminutionof glycolipid of the globo-series with a parallelincrease of the members of the lacto- and ganglio-series due to the increased activity of GlcNAcT-1

Cell Biology International, Vol. 23, No. 2, 1999 95

and of SAT-1 is supported by the data ofFenderson et al. (1987) and Osanai et al. (1997)obtained on embryonal carcinoma cells afterdifferentiation induced by RA.

As shown in Figure 3 and as previously indi-cated, the activities of the enzymes analysed in theembryos exposed to the treatment with 15 n RA(dark bars) are differently influenced by the treat-ment itself. In other words the data are not homo-geneous, indicating relevant differences amongembryos generated from different females. Thesebiochemical differences could be related tothe morphological and teratogenic variabilityobserved.

The exposure of the embryos to 30 n RA causesa general decrease in the enzyme activities. Thisconcentration is probably highly toxic for Xenopusembryos as indicated also by the relevant degree ofmalformations (Fig. 1). For this reason it is difficultto draw the same possible correlation betweenglycolipid pattern and glycosyltransferase activitiesin the animals treated with this dosage as for theanimals exposed to the 15 n RA; moreover, thedifferences with the control animals for glycolipid(neutral and acidic, i.e. gangliosides) distribution isnot so evident.

Work is now in progress in our laboratory tostudy the possible modifications of other biologicalparameters, such as motility and behavior, aftertreatment of Xenopus embryos with RA.

ACKNOWLEDGEMENTS

Supported by CNR (94.00235.CT14) (to G.Bernardini), MURST 40% and 60% and ASIgrants (to B. Berra).

REFERENCES

B M, B S, S A, S P, 1982. Biosynthesisin vitro of sialyl(�2-3)neolactosylceramide by a sialyltrans-ferase from embryonic chicken brain. J Biol Chem 257:12,765–12,769.

B M, D KK, K JW, C HC, S R, B S,1987. Glycolipids. Methods Enzymol 138: 575–607.

B G, V C, B P, C M, 1994.Lethality, teratogenicity and growth inhibition of heptanolin Xenopus assayed by a modified frog embryo teratogenesisassay-Xenopus (FETAX) procedure. Sci Total Environ 151:1–8.

D TA, C MJ, 1994. Effects of localizedapplication of retinoic acid on Xenopus laevis development.Dev Biol 162: 394–401.

E-Z H, D C, 1993. The pattern ofretinoic acid receptor g (RARg) expression in normaldevelopment of Xenopus laevis and after manipulation of themain body axis. Mech Dev 41: 33–46.

F BA, A PW, N E, C H,H S, 1987. Glycolipid core structure switchingfrom globo- to lacto- and ganglio-series during retinoicacid-induced differentiation in TERA-2 derived humanembryonal carcinoma cells. Dev Biol 122: 21–34.

G R, R AM, T XW, B B, B G,1995. Glycolipid pattern during Xenopus embryo develop-ment. Cell Biol Int 19: 183–189.

G R, B S, B G, R AM, R F,B B, 1997. Activities of glycolipid glycosyltransferasesand sialidases during the early development of Xenopuslaevis. Mol Cell Biochem 166: 117–124.

H S, 1990. Bifunctional role of glycosphingolipids:modulators for transmembrane signalling and mediators forcellular interactions. J Biol Chem 265: 18,713–18,716.

H YA, B RM, 1989. Functions of sphingolipids andsphingolipid breakdown products in cellular regulation.Science 243: 500–507.

H YA, O LM, 1995. Ceramide: an intracellularsignal for apoptosis. Trends Biol Sci 20: 73–77.

L M, K P, C P, 1992. Multiplicity generatesdiversity in the retinoic acid signalling pathways. Trends BiolSci 7: 427–433.

N Y, I M, 1995. Cellular biology of gangliosides.In: Rosemberg A, ed. Biology of Sialic Acids. Plenum Press,New York. 197–241.

N JL, 1990. The biogenesis of retinoic acid: a physiologi-cally significant promoter of differentiation. In: Dawson MI,Okamura WH, eds. Chemistry and Biology of SyntheticRetinoids. CRC Press, Florida. 229–249.

N PD, F J, 1956. Normal table of Xenopuslaevis (Daudin). A systematical and chronological surveyof the development from the fertilized egg till the endof metamorphosis. North-Holland Publishing Company,Amsterdam.

O T, B RM, H YA, 1989. Sphingomielinturnover induced by vitamin D3 in HL-60 cells: role in celldifferentiation. J Biol Chem 264: 19,076–19,080.

O T, W Y, S Y, 1997. Glycolipid glycosyl-transferases are enhanced during neural differentiation ofmouse embryonic carcinoma cells, P19. Biochem BiophysRes Commun 241: 327–333.

P GM, 1997. A simplification of the protein assaymethod of Lowry et al. which is more generally applicable.Anal Biochem 83: 346–356.

R AM, G R, R F, B G, B B,1995. Retinoic acid induces changes in Xenopus embryoglycolipid pattern. Cell Biol Int 19: 895–901.

T S, T C, M M, T T, W J, 1995.Sequential changes in glycolipid expression during human Bcell differentiation: enzymatic bases. Biochim Biophys Acta1254: 56–65.

W J, A PW, C J, H J, P MF, VK A, D I, F BA, 1994. Glycolipids ofgerm cell tumors: extended globo-series glycolipids are ahallmark of human embryonal carcinoma cells. Int J Cancer58: 108–115.

Z CB, M RB, 1992. Gangliosides as modulatorof cell functions. Am J Physiol 262: C1341–C1355.