Rat intestinal glycolipids

123

Biochimiea et Biophysics Aeta, 441 (1976) 123-133 @ Elsevier Scientific Publishing Company, Amsterdam

- Printed in The Netherlands

BBA 56814

RAT INTESTINAL

GLYCOLIPIDS

II. DISTRIBUTION AND BIOSYNTHESIS OF GLYCOLIPIDS CERAMIDE IN VILLUS AND CRYPT CELLS

LEAN-FRA~CO~S

BOUDOIRS

* and ROBERT M. ~~~C~AN

Thorndike Laboratory and Department Israel Hospital, Boston, Mass. (U.S.A.) (Received

December

AND

of

Medicine, Harvard Medical School and Beth

17th, 1975)

Intestinal epithelial cells were isolated from rat intestine and grouped into villus and crypt cell fractions. Glycolipids were purified from each cell fraction and quantitated by fluorimetric determination of glycolipid sphingosine. Significant quantities of ceramide were found in all cell fractions and accounted for approximately 15% of total glycolipid sphingosine. While villus and crypt cell fractions quantitatively contained differing amounts of sphingosine, all cell fractions contained proportionally similar quantities of sphingosine when compared to cellular cholesterol or phospholipid. Individual glycolipids, however, showed significant differences in distribution between villus and crypt cells. Hematoside and glucosylceramide were proportionally increased in villus cells, while crypt cells showed an increase in trihexosylceramide and ceramide content. The rate of UDPglucose : hydroxy fatty acid ceramide glucosyltransferase was higher in villus cells while the rate of UDPgalactose : lactosylceramide galactosyltransferase was 3-4 times increased in crypt cells. These studies demonstrate that si~ific~t differences in both the dist~bution and biosynthesis of individual glycolipids occur in crypt and villus cells of rat intestine and are of possible importance in the process of intestinal cell differentiation.

Introduction The intestinal epithelium is a tissue which undergoes rapid renewal [l] with cells at all stages of differentiation present in the mucosa. Undifferentiated cells * Cherg4 de Recherche INSERM. Fteoent address: Laboretolm de Phyelopathologle Digestive UR-45. Pavilion Hbis. H6~ital Edouard-Herriot. 89374 LYON Cedex 2 (France). Thea work ir a put of P thesis for the degree of Doctsurts Sciences to be submitted to the Unlverslt4 Claude Bernard, Lyon
124

in the crypts undergo progressive maturation as they migrate up the villus and develop a well defined brush border [ 2 ] containing a variety of digestive enzymes [3]. Lipid analysis of rat intestinal mucosa and microvillus membranes [4,5] reveal an unusually high percentage of glycolipids, with as much as 20% of microvillus membrane lipid present as glycolipid. The recent demonstration that major changes in glycolipid composition could accompany changes in cell growth [6] (i.e. division, transformation) suggests that similar changes might accompany cellular growth and differentiation in the intestinal epithelium. The ability to separate cell populations in varying stages of differentiation (i.e. villus and crypt cells) [7] offers the opportunity to study changes in glycolipid composition and synthesis in contiguous population of cells in varying stages of differentiation isolated from a nonmalignant tissue. In a previous study from this laboratory [8] it was shown that the main k&c-acid containing glycolipid of rat intestinal mucosa, hematoside, was virtually absent from crypt cells and that the enzyme required for sialic acid addition (lactosylceramide sialyltransferase) was not present in the crypt cell fraction. In the present study, total glycolipids from villus and crypt cells were extracted and the distribution of individual glycolipids was examined. In addition to confirming previous results on hematoside distribution, the present study demonstrates significant differences between villus and crypt cells in the distribution and biosynthesis of the other major glycolipids of rat intestinal mucosa. Experimental

procedure

Materials UDP[ 14C]glucose and UDP[ 14C] galactose were purchased from New England Nuclear (Boston, Mass.). Non-radioactive nucleotides were purchased from Schwarz/Mann (Orangeburg, N.Y .). Ceramides containing hydroxy and nonhydroxy fatty acids were purchased from Applied Science Laboratories, Inc. (State College, Pa.). Glucosyl ceramide standards were purchased from Analabs, Inc. (North Haven, Conn.). Dihexosylceramide, employed as a precursor for trihexosylceramide synthesis, was purified from commercial heavy cream after silicic acid column chromatography and preparative thin-layer chromatography [9 1. Fluorescamine (Fluram-Roche) was obtained from Fisher. Male Holtzman rats (200-250 g) were obtained from the Holtzman Company (Madison, Wk.) and maintained on standard laboratory feed. Methods Separation of villus and crypt cells. For each experiment, two rats fasted for 24 h were sacrificed by cervical dislocation. The entire small intestine was removed, flushed with ice cold saline (0.154 M) containing dithiothreitol (1 mM) and intestinal cells were sequentially isolated in 10 fractions as described by Weiser [7]. Cells were pooled into four fractions: 1, villus tip; 2, villus intermediate; 3, villus base; 4, crypt zone, containing respectively about 30,40, 20 and 10% of the isolated cells as determined by protein. Validation that a gradient of cells from villus to crypt was achieved was obtained by determination

125

of the specific activity of alkaline phosphatase\on each fraction. Lipid extraction. Cell fractions were homogenized with a Polytron homogenizer in 7 vol. of methanol to which 14 ~01s. of ~hlorofo~ were added. Lipids were extracted overnight at room temperature. The protein residue was collected by filtration, dispersed and re-extracted in 10 ~01s. of chloroform/ methanol (1 : 1, v/v) for 1 h at room temperature. The procedure was then repeated with chloroform/methanol (1 : 2, v/v). The filtrates were combined, dried in a rotary evaporator, redissolved in chloroform/meth~ol and filtrated on a small fritted glass funnel with a ~atrn~ 3 MM filter paper disc. Glycolipid purification. The total lipid extract was dried on a rotary evaporator with repeated additions of dry methanol. Acetylation and subsequent separation of the glycolipids from neutral and phospholipids was performed as described by Saito and Hakomori [lo] on a small column of Florisil(3 g) eluted successively with 60 ml of the following solvents: di~hloroeth~e, dichlorethanelacetone (1 : 1, v/v), dichloroethane~methanol~water (80 : 20 : 1,by vol). The glycolipid fraction was deacetylated and analyzed by thin-layer chromatography without dialysis. Thin-layer chromatography. Glycolipids were separated on thin-layer plates in the solvent system chloroform/methanol/water (60 : 35 : 8, by vol) and visualized by the ~-naphthol-sul~c acid spray. Sialic acid con~n~g glycolipids were detected by resorcinol spray Ill]. For quantitative determination and structural analysis, individual compounds were scraped directly from thinlayer plates after comparison with a reference lane. Long chain base assay. Ceramide and glycolipids were quantitatively determined as sphingosine, using a fluorimetric determination of the primary amine as described by Naoi et al. [12]. This assay could be performed on thin-layer scraping without prior extraction; however, the volume of silicagel present necessitated an increase in the proportion of the reagents as follows: about 100 nmol of a glycolipid mixture was applied as a streak in a 3cm wide lane on the plate and individual spots were scraped directly into teflon-lined screw cap tubes. 1 ml of methanol/cone. HClfwater (83 : 8.6 : 9.4, by vol.) was added and hydrolysis was performed for 18 h at 70°C. Tubes were cooled in ice and the following reagents added: 0.5 ml of NaOH, 1.5 ml of borate buffer, pH 8, and 2 ml of diethylether. The tubes were thoroughly shaken on a vortex mixer. 1 ml of a solution of fluorescamine in diethylether (0.15 mg/ml) was added iess than 30 min before determining the fluorescence. The ether phase was assayed in a Perkin-Elmer 204 Spe~trofluor~eter with excitation at 395 nm and emission at 465 nm, settings giving maximal fluorescence. The assay utilizing commercially available cerebroside (mol. wt., 801) was linear up to 100 nmol. The reading given by a corresponding area in a blank lane was subtracted from each reading and was less than 0.5-l nmol equivalent. Sphingosine recovery from thin-layer plates was found to be quantitative (104 f 2% on 15 determinations). Gas-liquid chro~a~o~aphy. Identity of the sugar moieties of glycolipids was determined as described previously [ 8] . Chemical and enzymatic assays. Phospholipids [13], sphingosine [lo] and cholesterol [ 141 were determined on total lipid extracts. Protein was determined by the method of Lowry et al. [ 151 and alkaline phosphatase according to Forstner et al. 1161.

126

Enzyme preparations for glycosyltransferase determinations. Cell fractions were isolated as described above and then homogenized in EDTA, 5 mM, pH 7.4, with a Polytron homogenizer (setting no. 5 for 20 s). The homogenates were centrifuged at 500 X g for 10 min to remove intact cells and large frements. The supernatants were centrifuged at 10 000 X g for 20 min to eliminate mitochondria and the postmitochondrial supematants were centrifuged at 40 000 rpm for 1 h in a Spinco rotor 40 to produce a microsomal pellet. The microsomal pellet was either suspended in 0.32 M sucrose solution containing Z-mercaptoethanol (14 mM) and frozen, or suspended in water and lyophilized. The suspension in sucrose solution was used as enzyme source for trihexosylceramide synthesis while the lyophilized pellet resuspended in benzene was utilized for glycosylceramide synthesis, Glucosylceramide synthesis. This was carried out as described by Constantine-Ceccerini and Morel1 [17] with minor modifications, A solution of lecithin in chloroform/methanol (50 pg) was evaporated to dryness in the bottom of the reaction tube. Then a benzene solution of the substrate, 50 pg of ceramide containing hydroxy fatty acids or non-hydroxy fatty acids was added with an ahquot of the enzyme suspension in benzene. Substrate and enzyme were dried together under a stream of nitrogen to ensure intimate contact of both major p~icipants of the reaction. The other components were then added and the complete reaction mixture contained the following: enzyme protein, 0.2-0.3 mg; hydroxy or non-hydroxy fatty acid, 50 pug; phosphatidylcholine, 50 fig; MgCl*, 0.6 pmol (instead of 0.3 pmol); ATP, 0.3 gmol; Tris - HCI, pH 8.0, 75 pmol; EDTA, 0.15 E.cmol;UDP[“4C]glucose 20 nmol at a specific activity of 10 Ci/mol. The final reaction volume was 130 ~1. Incubations were carried out at 3’7°C for 2 h with shaking. ~ithiothreitol was omitted from the reaction since its addition did not stimulate the activity of the enzyme. The reaction was stopped by the addition of 2.5 ml of chloroform/methanol (2 : 1, v/v) and then partitioned after the addition of 0.4 ml of water. The lower phase was then washed three times with 0.5 ml methanol/water (1 : 1, v/v), evaporated under nitrogen and applied to thin-layer plates with a standard of glucosylceramide. The developed plates were sprayed with a-napthol reagent and the glucosylceramide scraped. The radioactivity was determined directly on thinlayer scraping using dioxane based scintillation fluid. Trihexosylceramide synthesis. Acceptor and detergents were dissolved in chloroform/methanol (2 : 1, v/v) and dried under a stream of nitrogen. The other reaction components were then added. Complete reaction mixtures contained the following (in micromoles, unless otherwise specified) in a final volume of 100 ~1: Triton CF-54/Tween 80 (2 : 1, v/v) 0.6 mg; lactosylceramide, 0.05; MnCl*, 1.0; cacodylate-HCI buffer, pH 6.5, 15.0: UDP[‘4C]galactose (5 Ci/mol) 50 nmol; protein 100-300 pg. Incubation was carried out for 1 h at 37°C with vigorous shaking and reactions were terminated by the addition of 2 ml of chlorofo~/methanol (2 : 1, v/v). These solutions were p~itioned with 0.4 ml of water. The lower phases were then washed three times with methanol/water (1 : 1, v/v) and dried under a stream of nitrogen. The samples were then spotted on thin-layer plates together with a trihexosylceramide. Radioactivity in trihexosylceramide was determined directly from thin-layer scraping.

127

Results The stepwise isolation of intestinal cells and their subsequent pooling in four fractions was monitored by alkaline-phosphatase specific activity, an enzyme found in the intestinal microvillus membrane [18]. As described previously [ 81, and confirmed in the present study, a decreasing gradient of alkaline phosphatase activity was found from villus tip to crypt. Lipids were extracted from each fraction as described in Methods. Further extraction of the residue in chloroform/methanol (1 : 1, v/v) for 4 h under reflux yielded only small amounts of additional lipid (0.5% phosphorus and cholesterol, 1.5% sphingosine). Therefore, reflux extraction was not carried out routinely. Total sphingolipids in villus and crypt cells. The sphingosine content of total lipid extracts prepared from indicated cell fractions was determined and compared to cholesterol and phospholipid content (Table I). It is apparent that although the absolute amount of each lipid class decreases from villus to crypt, the molar ratio of sphingosine to either cholesterol or phospholipid is relatively constant. While cholesterol and phospholipid determinations on total lipid extracts may represent slight overestimations due to water-soluble products (i.e. bile salts, non-lipid phosphorus) similar results were obtained when these lipid classes were quantitated after Florisil column chromatography. As shown in Table I, both villus and crypt cells contain approximately 1 mol of sphingosine to 16 mol of phospholipid or 5 mol of cholesterol. Distribution of glycosphingolipids and cemmide in villus and crypt cells. Glycosphingolipids were separated by Florisil chromatography of total lipid extracts of isolated cell fractions, and their pattern analyzed by thin-layer chromatography. Individual glycolipids were detected by cr-naphthol/sulfuric acid spray. Gangliosides were detected by resorcinol spray. Fig. 1 is a schematic

TABLE I SPHINGOSINE

CONTENT

OF VILLUS AND CRYPT CELLS OF RAT INTESTINE

Isolated intestinal cells were prepared and grouped into indicated fracttons (Methods). Total Iipid extracts were prepared and aliquots taken for the indicated analyses. Sphingosine content was compared with cholesterol and phosphorus in each fxaction. Results are expressed as means C3.E.M. of 6 individual cell gradients. VillUS

Crypt ___-

-___

(4)

Tip

Intermediate

Base

(1)

(2)

(3)

Sphingosine Total pmol

4.6 ? 0.6

5.1 ? 0.6

3.3 * 0.4

55.3 2 4.2 14.0 ? 1.7

61.6 * 5.7 16.6 + 1.3

50.6 f 7.4 15.7 +_1.6

24.0 t 4.0 5.0 * 0.4

27.1 t 3.4 5.4 + 0.7

15.4 * 1.9 4.6 f 0.4 __-

1.6 f

0.4

Total phosphorus Total pm01 Mol/mol spingosine

30.2 +_10.6 16.0 2 1.6

Cholesterol Total pmol Mollmol of sphingosine

_.-.

7.0 f 4.0 ? -._._

1.7 0.6

Solv

front

11 10

CMH

9

Fig. 1. Schema of a thin-layer chromatoaam of rat intestinal glycollpids. Solvent system: chloroform/ methanol/water (60 : 35 : 8. by vol.). Detection: cY-naphthol spray: individual glycollpids were identified as indicated. Numbers refer to individual glycolipids which were quantitated directly from thin layer scrapings (Table II). CMH: monohexosylcersmide; CTH: trihexosylceramide; Hem: hematoside.

representation of a typical glycolipid separation by thin-layer chromatography. AS we have demonstrated earlier, the major sialic-acid containing glycolipid in rat intestinal mucosa is a hematoside. The two other major glycolipids of rat intestinal mucosa are monohexosylceramide and trihexosylceramide (Fig. 1). The identity of these compounds was confirmed by gas-liquid chromatography of hydrolyzed thin-layer scrapings; they agree with the findings of Forstner and Wherret [ 51. In addition to these three major glycolipids, small amounts of other glycolipids were visible as indicated in Fig. 1. Also shown is a compound directly beneath the solvent front which did not exhibit a colored reaction with cu-naphthol, indicating that it does not contain sugar. This compound was isolated from thin-layer scrapings and found to contain sphingosine. Hydrolysis of the compound yielded sphingosine and a fatty acid, suggesting that it is acylsphingosine (ceramide). Precise identification of the fatty ‘acid and sphingosine moiety is currently in progress. Individual glycolipids were scraped from thin-layer plates as shown in Fig. 1 and quantitated by sphingosine assay (Table II). As shown in Table II, 8245% of the total glycolipid in all cell fractions was accounted for by hematoside, mono- and trihexosylceramide, and ceramide. Other glycolipids were present in amounts too small to characterize but qualitatively are similar to the glycolipid pattern of rat intestinal mucosa described by Forstner and Wherret [ 51. While villus and crypt cell fractions qualitatively contain similar glycolipids, major differences in the distribution of individual glycolipids are apparent. In confirmation of earlier results [S], a sharp gradient of hematoside content is evident from villus to crypt. Similarly, while monohexosylceramide is the major glycolipid present in all cell fractions, this compound is present in smaller quantities in crypt cells. The decreased monohexosylceramide content of crypt cells is compensated for by a reciprocal increase in the content of trihexosylceramide and ceramide which is approximately 150% of the levels seen in villus tip cells. Significantly, in all cell frac-

129

TABLE II DISTRIBUTION TESTINE

OF GLYCOLIPIDS

AND CERAMIDE

IN VILLUS

AND CRYPT CELLS OF RAT IN-

Glyeolipids were purl&d from lipid extracts prepared from the indicated ceB fractions and separated by Wn-Iayer chromatography. Individual gtycoilpids were scraped and glycoiipid spbingodne auantitated directly tram thin layer scrapings. Results are expressed aa nmol per 100 mnoI of total spbinwsine PP plied to the thin layer plate. Individual gIycoIipids were scraped according to the schema in Fig. 1. Results are mean +S.E.M. of 5 experiments. Results are in mnol %. -__-----.-_--_.___- -~-. VillUS __ _-___

Thin layer number

.- ___

Tip (1)

.._ --

Intermediate (2)

.-..-.-____.

_

__-- __._

Hematoside

Trihexosylceramide Lactosylckramide Monohexosylceramide Ceramide -___

4 5 6 7 8 9 10 11

--...

* P < 0.01. referring to significant l * P < 0.02. the same.

~_

__.._

--_-

?: 0.2 t 0.4 + 2.8 f 0.9 t 1.0 ?r1.3 + 0.4 + 0.4 ? 1.2 f 1.3

differences

of the indicated compound

_._-____

<0.5

? 0.2 t 0.3 + 1.7 f 0.7 f 0.4 t 1.3 + 0.3 + 0.7 + 1.8 2 1.8

--.-

Cvpt (41

Base w

co.5 <0.5 1.2 1.1 13.5 6.1 3.6 15.3 1.9 2.2 37.0 17.3

<0.5

<0.5 1.0 1.5 17.7 7.2 3.8 13.6 2.3 1.6 37.7 16.9

_._____.

<0.5 1.0 0.8 8.6 6.0 4.5 19.5 2.4 1.3 36.2 21.5

* 0.0 ‘_ 0.2 * 0.9 f 0.5 f 0.7 ? 1.5 * 0.6 + 0.1 f 1.5 t 2.5

co.5 co.5 1.0 0.8 3.8 5.0 5.7 21.6 2.9 2.0 31.1 26.0 between

t 0.3 + 0.31 f 0.7 * *_0.6 f 1.2 f 1.1 * * 0.3 f 0.3 + 1.7 ** + 1.6 *

Fractions

1 and 4.

tions lactosylceramide is present in small amounts only. The change in the distribution pattern of sphingolipids was found in all experiments and seemed to be progressive as cell8 migrate from crypt zone to villus tip. In order to investigate possible mechanisms responsible for these differences, the biosynthesis of mono- and trihexosylceramide was studied in villus and crypt cells. Glucosyfceramide synthesis. Monohexosylceramide of rat intestinal mucosa has been shown to be a glucosylceramide [ 51. This was confirmed in the present study by gas chromatographic analysis of the liberated sugar moiety of rat intestinal monohexosylceramide. Therefore, UDP[ “C]glucose was employed as the sugar donor. Two acceptors, hydroxy fatty acid ceramide and nonhydroxy fatty acid ceramide were compared: hydroxy fatty acid ceramide gave a consistently higher incorporation (20-35%) of f 14C]glucose into monohexosylceramide and was therefore used for further studies. The pH optimum of the reaction was found to be 8. The reaction rate was linear for 2 h and with increasing protein content (up to 300 gg per assay). The synthesis of glucosylceramide was examined in microsomal preparations (see Methods) from villus and crypt cells. The results presented in Table III show a decreased synthesis of glucosylcer~ide in crypt cells compared to villus cells. The level of incorporation was high in all preparations without added acceptor, reflecting perhaps the presence of precursor ceramide (Table II). A decrease in activity of synthesis from villus to crypt is apparent both for endogenous (without acceptor) or exogenous synthesis (acceptor added). Similar

130

TABLE

III

GLUCOSYLCERAMIDE

SYNTHESIS

BY VILLUS

AND CRYPT

CELLS

Isolated cell fractions were prepared, homogenized and 100 000 X K pellets prepared (Methods). Complete Incubation mixtures contained the following in a final volume of 130 ~1: phosphatidylcholine b0 I.rg. hydrow fatty acid ceramide 50 pg. enzyme protein 0.34.4 mg, MgCI2 0.5 PmoI, Tris . HCI, PH 8.0. 75 I.cmOI, EDTA 0.15 pmol, UDP[‘4C1gIucose 20 run01 (SP. act. 10 Cilmol). Incubations were carried out at 37OC with shaking for 2 h. CeU fraction

[14Cl

Glucose

incorporated

(nmol/mg

per h)

Acceptor:

Hydroxy fatty acid ceramide

None

Difference

ViIlus Tip Intermediate Base

(1) (2) (3)

1.40 1.59 1.27

0.46 0.72 0.57

0.96 0.87 0.70

Crypt zone

(4)

0.70 _

0.32

0.38

* Difference

-._.____

refers to values with and without

acceptor

*

present.

results were obtained when homogenates from villus and crypt cells were centrifuged at 500 X g for 10 min to remove intact cells and large fragments, and the supematant centrifuged at 100 000 X g for 1 h to prepare a total membrane preparation. The agreement of these results with the data shown in Table III suggests that the observed differences between villus and crypt cells are not due to differences in the subcellular fractionation of these two cell types. Trihexosylceramide synthesis. The optimum pH for the reaction was found to be 6.5 and the presence of Mn” in the incubation mixture was an absolute requirement. The reaction rate was linear for 1 h and with increasing protein concentration (up to 400 pg per assay). When the transfer of galactose from UDP[ “C]galactose to exogenous lactosylceramide was studied (Table IV), a continuous increase in tribexosylceramide synthesis was observed from villus to crypt. Since the conditions for subcellular fractionation of crypt cells have

TABLE

IV

TRIHEXOSYLCERAMIDE

SYNTHESIS

BY VILLUS

AND CRYPT

CELLS

CeII pellets from individual gradient &actions were prepared as indicated. Complete incubation mixtures contained the following in a tinal reaction volume of 100 ~1: Triton CF64. Tween 80 ( 1 : 2. v/v) 0.6 mg, lactosylceramide 0.06 pmol. MnCI2 1.0 pmol, cacodylate-HCI buffer, pH 6.5. 15 pmol, UDP[ “C]gaIactose (5 Cilmol) 50 nmol. protein 100-300 pg. Incubations were carried out for 1 h at 37OC with vigorous shaking. Cell fraction

I t4ClGaIactose

incorporated

(nmol/mg

per h)

500 X L!

10 000

x ,q

140 000

ViIIuS Tip Intermediate B&W?

(1) (2) (3)

0.18 0.70 1.31

0.49 1.39 2.64

0.51 0.90 1.18

Crypt

(4)

1.86

2.66

1.65

zone

x g

131 5-r In mofe-‘h

-f/V mq Profeinl

tip

lntemediate

-8

-6

-4

f/[uoP

-2

2

GaLacrosE]

Fig. 2. Effect of UDPgalactose concentratfon rat intestine.

4

tm

6

8

M-9

on trihexosylceramfde

synthesis by vfllus and crypt cells of

not been clearly defined, synthesis was studied in three subcellular fractions as indicated in Table IV. In all fractions, a similar gradient of synthesis is observed; however, maximal rates of synthesis were found in the 10 000 X g pellet. In all experiments, endogenous synthesis (without acceptor present) was less than 10% of the activity with acceptor present. This is in agreement with the extremely low levels of precursor lactosylceramide present in all fractions (Table II). The possibility that the observed differences in trihexosylceramide synthesis were due to differing properties of the enzyme in crypt and villus cells was investigated. The pH optimum for the reaction in villus and crypt cells was identical. No significant difference in the affinity of the enzyme for the donor could be demonstrated (Fig. 2). The K, for UDPgalactose was 1.35 - low4 M, in good agreement with published results [ 19). Discussion The intestinal mucosa is a tissue characterized by a rapid rate of cellular proliferation with migration of cells from crypt to villus tip requiring 20-30 h in the rat [ 11. As cells move upward from the crypt, it is known that major modifications in morphology (brush border formation) [.2], enzyme content [ 181, and surface properties such as lectin agglutinability [20,21] occur. In this process, cells pass from a mitoticahy active state to become non-mitotic, functionally mature cells. The present study demonstrates that significant changes in glycolipid content and biosynthesis also occur during this maturation process. The present results on the lipid composition of rat intestinal epithelium are in agreement with previous analyses of mucosal extracts [ 51: 4.9 mol of sphingosine for 23 of cholesterol and 60 of phospholipids. These results confirm that the intestinal epithelium is enriched in glycosphingolipids. The distribution of individual glycosphingolipids (Table II) in villus cells closely reflects previously reported

132

values in whole mucosa [5], that is, slightly less hematoside than trihexosylceramide and twice as much mono- as trihexosylceramide. The present studies quantitated individual glycolipids by fluorimetric determination of sphingosine directly from thin-layer scrapings rather than by gas-liquid chromatography of liberated sugars, with comparable results. This method of glycolipid quantitation, in addition to its rapidity, also permits the quantitation of non-sugar containing compounds such as ceramide. The present study demonstrated for the first time that ceramide is an important component of intestinal cell lipids and comprises approx. 15-2076 of total sphingosine (Table II). Complete identification of this compound is in progress. Although villus and crypt cells contain proportionally similar amounts of glycolipid sphingosine when compared to other lipid components (Table I), significant differences in the distribution of individual glycolipids occur in crypt and villus cells (Table II). In a previous study [8], we have demonstrated a marked decrease in the hematoside content of crypt cells, a finding confirmed in the present study. In addition, the present study demonstrates that crypt cells contain decreased amounts of monohexosylceramide as well, which are compensated for by increased amounts of trihexosylceramide. While villus cells contain increased amounts of monohexosylceramide and hematoside, there is a compensatory decrease in the amount of trihexosylceramide resulting in no change in total sphingolipid concentration. To determine whether the differences in the concentration of individual glycolipids in crypt and villus cells might be due to differences in biosynthesis, monohexosylceramide synthesis was studied in crypt and villus cells (Table III). A gradient of synthesis was found with villus cells exhibiting a synthesis twofold greater than the crypt cell fraction. Of note was the relatively high rate of synthesis when no acceptor was added, reflecting perhaps the presence of significant amounts of ceramide precursor in all cell groups. The observed differences in synthesis cannot be explained on the basis of differing amounts of ceramide present, since crypt cells contain relatively more ceramide than villus cells. It is unknown, however, whether the ceramide demonstrated in the present studies is available as a precursor for glycolipid synthesis or whether it serves as an important component of cell structure. Of interest is the discrepancy between the two-fold increase in ceramide glucosyltransferase activity found in villus cells and the more modest increase in glucosylceramide found in these cells (Table II). Glycolipid degradation studies were not undertaken; however, a glucosylceramide hydrolase, identical to phlorizin-hydrolase has been demonstrated in the brush border of rat intestinal cells [ 221. It is possible that despite the high rate of glycosylceramide synthesis demonstrated in villus cells, the final content of glucosylceramide is in part determined by the activity of degradative enzymes. The present studies also demonstrate that the undifferentiated crypt cell contains increased quantities of trihexosylceramide (Table II) correlating with an increased biosynthesis (Table IV). Since the intestinal crypt cell is known to be a mitotically active cell with many features in common with fetal intestinal cells [ 20-231, this result is of interest in view of studies in other cell types which demonstrate active trihexosylceramide synthesis in functionally immature and dividing cells. Developmental studies in rat kidney have shown

133

an increase in monohexosylcer~ide synthesis 1171 and a decrease in trihexosylceramide synthesis [19] with increasing age. Similarly, increased trihexosylceramide synthesis has been demonstrated during cell division (G and S phases) in hamster cells [24]. The present study suggests that a similar reduction of trihexosylceramide synthesis occurs as cells leave their mitotically active site, the crypt zone. The present study demonstra~s an increased synthesis of t~hexosylcer~ide in crypt cells coupled with earlier results [8] which indicated an absence of hematoside synthesis in crypt cells. This is of interest since both compounds are directly derived from a common precursor, lactosylcer~ide. Although the functional importance of the glycolipid changes in villus and crypt cells remains to be determined, the present study suggests that the addition of a terminal sialic acid instead of a terminal galactose to lactosylceramide is perhaps an important event in the differentiation and maturation process. Acknowledgement This work was supported by NIH Grant AM 16453 and General Research Support, Grant 491.94. References 1 Lipkin, M. (1973) Phydol. Rev. 53,891-915 2 ~e~~tb, NJ., Van Dongen. J.M.. Van Hofwegen. B.. Keulemans, J., Visser, W.J. and GaUaard, H. (1974) Dev. Biol. 38.119-137 3 Holmes, R. (1971) Gut 2,668-+77 4 Forstner, G.G.. Tanaka, K. and Isselbacher. K.J. (1988) Biochem. J. lOS. Bl--59 5 Forstner, G.G. and Wherrett, J.R. (1973) Biochim. Biophys. Acta 306.446459 6 Hakomori, S.I. (1975) Biocbhn. Biophys. Acta 417.55-89 ‘7 Weiser, M.M. (1973) J. Biol. Chem. 248,2536-2541 8 Glickman, R.M. and Bouhours, J.F. (1975) Biochim. Biophys. Acta (1976) 424,17-26 9 Kayser, S.G. and Patton, S. (1970) Biochem. Biophys. Res. Commun. 41,1672-1578 10 Saitq, T. and Hakomori. S.1.. (1971) J. Lip. Res. 12.257-259 11 MacMillan, V.H. and Wherrett. J.R. (1969) J. Ncurochem. 16.1621-1624 12 Naoi. M., Lee, Y.C. and Roseman, S. (1974) Anal. Biochem. 58,571--577 13 Bartlett, G.R. (1959) J. Biol. Chem. 234.466468 14 Zlatkis, A., Zak, B. and Boyle. A.J. (1953) J. Lab. Clin. Med. 41.486492 15 Lowry, P.H., Roscbrougb, N.J., Farr, A.L. and RandaIl. R.J. (1951) J. Biol. Chem. 193, 265-275 16 Forstner, G-G.. Sabesin, S.M. and Isselbacher, K.J. (1968) Biochem. J. 106,381-390 17 CostantinoCeccarini, E. and Mar&, P. (1973) 3. Biol. Chem. 248.8240-8246 18 Nordstrom, C., Dablqvist. A. and Josefsson. L. (1967) J. Hist. Citochem. 15.713-721 19 Martensson, E., Ohnan, R., Graves, M. and Svennerhobn, L. (1974) J. Biol. Chem. 249.41324137 20 Podolsky, D.K. and Weiscr, M-M. (1973) J. Cell. Biol. 58,497-500 21 Etzler, M.E. and Branstrator, M.L. (1974) J. Cell. Bid. 62.329-343 22 Leese, H.J. and Semenza. G. (1973) J. Biol. Chem. 248,8170-8173 23 Weiser, M.M. (1972) Science 177. 526-526 24 Wolf, B.A. and Robbins, P.W. (1974) J. Cell. Bid. 61,676--687