2705Journal of Cell Science 108, 2705-2713 (1995)Printed in Great Britain © The Company of Biologists Limited 1995

Expression and glycosylation of the filamentous brush border glycocalyx

(FBBG) during rabbit enterocyte differentiation along the crypt-villus axis

J. Maury, A. Bernadac, A. Rigal and S. Maroux*

Laboratoire de Biochimie et Biologie de la Nutrition, CNRS-URA 1820, Faculté des Sciences de Saint Jérôme, Case 342, 13397Marseille Cedex 20, France

*Author for correspondence

The filamentous brush border glycocalyx forming the‘enteric surface coat’ of the intestinal epithelium iscomposed in rabbits of a 400 kDa mucin-type glycoprotein,which was purified using the 3A4 monoclonal antibody.This monoclonal antibody recognizes a filamentous brushborder glycocalyx-specific glycosidic structure containingan O-acetylated sialic acid, which is absent from all theother glycoproteins in the epithelium, with the exception ofcertain goblet cell mucins. Here we establish that only 50%of the rabbits tested synthesized this glycosidic structure.Upon immunolabeling surface epithelia and sections ofjejunum from these rabbits, the carbohydrate epitope rec-ognized by the 3A4 mAb was found to be present on the fil-amentous brush border glycocalyx of a variable number ofenterocytes, which were patchily distributed over all thevilli. This heterogeneous expression of 3A4 antigenicity,which was also observed in the crypts, suggests theexistence of differences between the patterns of differen-tiation of enterocytes, which results in the expression ofdifferent pools of glycosyltransferases and/or acetyl trans-ferases. In mature enterocytes, the 3A4 determinants were

present only on the filamentous brush border glycocalyx,which is anchored solely to the membrane microdomain atthe tip of brush border microvilli. However, expression of3A4 antigenicity begins in the median third of crypts, inenterocytes with a short, thin brush border devoid of apicalfilamentous brush border glycocalyx. Here the 3A4epitopes were present over the whole brush bordermembrane. A few cells higher up, the hyperpolarizedexpression of filamentous brush border glycocalyx, i.e. itssegregation at the tip of the microvilli, began and both theapical filamentous brush border glycocalyx and the wholebrush border membrane were labeled with the 3A4 mAb.The labeling of the lateral membrane of microvilligradually decreased and then disappeared during themigration of the cells to the upper part of the crypts. The3A4 structure was also detected in some of the granules ofsome rare goblet cells in the vicinity of the most deeplylocated labeled enterocytes, whereas no labeled goblet cellswere ever observed at higher cell locations.

Key words: enterocyte, differentiation, glycocalyx, mucin, brush border

SUMMARY

INTRODUCTION

Absorbing cells (enterocytes) and mucus-secreting cells(goblet cells) are the two main cell types present in the smallintestinal epithelium. They originate from the same stem cellsthat proliferate near the bottom of the crypts, where the differ-entiation of each cell type begins (Cheng and Leblond,1974a,b). The differentiation and maturation processescontinue while the cells move up along the crypt-villus axis(Cheng and Leblond, 1974a,b,c; Wright and Irwin, 1982).

Goblet cells secrete the mucus which, thanks to the presenceof gel-forming glycoproteins called mucins, forms a protectivegel covering the epithelium. Enterocytes synthesize an ‘entericsurface coat’, which was first described by Ito (1965, 1974) onthe basis of electron microscopic data. We recently establishedthat this structure, which has been called the ‘filamentous brushborder glycocalyx’ (FBBG), is composed of one 400 kDamucin-type glycoprotein, which is anchored to the membranemicrodomain at the tip of the microvilli. A specific monoclonal

antibody (3A4) against this membrane-bound mucin wasproduced and used to immunopurify it. It recognized a FBBG-specific glycosidic structure containing a terminal O-acetylatedsialic acid (Maury et al., 1995).

Mucins are high molecular mass (>100 kDa) glycoproteinsin which the glycosidic moiety amounts to 50-80% of themolecular mass (Strous and Dekker, 1992). The oligosaccha-ridic structures of mucins undergo changes during embryoge-nesis, cellular differentiation and cancer development, andtherefore constitute appropriate antigenic determinants withwhich to monitor these phenomena using specific antibodies(Feizi and Childs, 1987; Feizi et al., 1984).

The glycosidic structures of highly glycosylated glycopro-teins are very complex (Podolsky, 1985a,b; Roussel et al.,1988) and depend mainly on the glycosyltransferase pool ofthe producing cells (Paulson and Colley, 1989; Ichikawa et al.,1992). It is therefore to be expected that all the glycoproteinssynthesized by the same cell will bear common glycosidicstructures. This is indeed the case in rabbit enterocytes, as

2706 J. Maury and others

regards the human blood group ABH type 1 and/or 2 glyco-sidic structures, which are present in almost all the plasmamembrane glycoproteins (Feracci et al., 1982; Gorvel et al.,1982, 1987b). The same structures are also synthesized bygoblet cells, and are consequently present on mucins(Zweibaum and Bouhou, 1973; Oriol and Dalix, 1977; Gorvelet al., 1987b). Rabbits termed A+ express the A structureswhereas A− rabbits express the H structures. In A+ rabbits thehuman blood group A type 3 (A3) determinant is present, butonly in a few of these glycoproteins (Gorvel et al., 1987a,b).A specific monoclonal antibody (14A4) directed against thisA3 structure was used to label the FBBG on jejunum sectionsof A+ rabbits and to distinguish it for the first time fromunlabeled mucins stored in the goblet cells (Gorvel et al.,1987a).

We recently reported (Maury et al., 1995) that in both A+

and A− rabbits, the FBBG was the sole glycoconjugate in smallintestinal mucosa that bears the glycosidic antigenic determi-nant recognized by the monoclonal antibody 3A4. O-acetylatedsialic acid is an essential component of the 3A4 epitope. The3A4 antibody was used here to study the expression of the 3A4epitope during small intestinal epithelial cell differentiationalong the crypt-villus axis.

MATERIALS AND METHODS

MaterialsBiodyne B membrane was from Pall (Glen Cove NY 11542). Rabbitanti-mouse Igs and goat anti-mouse Igs coupled to either peroxidaseor fluorescein were obtained from Cappel (Malvern PA, USA).

Monoclonal antibodiesMonoclonal antibody 3.3A, which was previously called Cl 3.3(Gorvel et al., 1987a), was a gift from Dr Bara (IRSC, Villejuif). Itspecifically recognized human blood A glycosidic structures of types1, 2, and 3 (Bara et al., 1988; Gane et al., 1987).

The procedures used to produce and characterize 14A4 monoclonalantibody with a restricted specificity for the human blood group Atype 3 (A3) glycosidic structure (Gane et al., 1987; Gorvel et al.,1987a,b) and of the anti-FBBG 3A4 monoclonal antibody (Maury etal., 1995) have been described in detail.

Partial purification of the FBBG from the brush bordermembraneThis technique has been described previously (Maury et al., 1995).Briefly, a crude brush border membrane preparation devoid of mucincontamination was prepared as described by Schmitz et al. (1973).After brush border membrane solubilization by Triton X-114 andphase separation (Pryde, 1986), the FBBG present in the aqueousphase was separated from any contaminants with a lower molecularmass by filtration through a Sepharose CL-2B column.

SDS-polyacrylamide gel electrophoresis (SDS-PAGE) andimmunoblottingThese techniques have been described previously (Feracci et al.,1986). Here a poorly crosslinked (monomer to crosslinker ratio of200:1) 7.5% polyacrylamide gel and a 3.5% stacking gel were used.The proteins were transferred onto Biodyne B membrane instead ofnitrocellulose. Either undiluted culture supernatants of 14A4 and 3A4hybridomas or 3.3A ascites fluid (1:105 dilution) and goat anti-mouseIg coupled to peroxidase (1:500 dilution) were used for theimmunoblotting experiments.

ImmunoprecipitationA rat monoclonal antibody against the κ chain of mouse IgG, the 1395211 mAb (Immunotech, Marseille, France), was coupled to Affi-Gel10 (Bio-Rad, Hercules, California) as recommended by the manufac-turer. The immunogel obtained was then incubated overnight with alarge excess of 3A4 in culture supernatant (1 ml of supernatant with60 µl of beads). The 3A4-beads obtained were used for the immuno-precipitation of the recognized antigen, as described previously(Massey et al., 1987).

Surface immunoperoxidase labelingThis technique was performed as described by Schmidt et al. (1985)with a few modifications. Segments (4 cm) of jejunum and ileum werecut out, opened longitudinally and pinned on a paraffin layer with theluminal surface uppermost. The preparations were then fixed for 1hour in 4% formaldehyde solution in phosphate buffer, pH 7.3, con-taining 8.1 mM Na2HPO4, 1.47 mM KH2PO4, 137 mM NaCl, and 2.7mM KCl (PBS). To remove the mucus, the preparations wereincubated for 2 hours in 20 mM dithiothreitol (DTT) added to 15 mMTris-HCl buffer, pH 8.2, containing 0.1 M NaCl and 20% ethanol.The intestinal sheets were then washed in PBS (several times) andfixed again in the 4% formaldehyde solution for 1 hour. After beingextensively washed with PBS, the preparations were incubated: (i) for90 minutes in 0.1% sodium azide in PBS to block the endogenousperoxidase; (ii) for 30 minutes in PBS containing 10% bovine serum(PBS-BS); (iii) overnight with either culture supernatant of 3A4hybridoma or 1:500 dilution of 3.3A ascites fluid in PBS-BS. Afterfour washes with PBS-BS, the preparations were incubated for 5 hourswith anti-mouse Igs coupled to peroxidase (commercial solutiondiluted 1:200 in PBS-BS) and washed again with PBS-BS. The per-oxidase was revealed using 1.3 mM diaminobenzidine (DAB) in PBScontaining 0.01% H2O2. The fixation of the 3A4 and 3.3A antibodieson the cell surface was observed under a dissecting microscope as abrown reaction product. The villi were then dissected out, depositedon a glass slide, mounted in 50% glycerol, and examined with a Zeissphotomicroscope III.

Immunoelectron microscopyImmunocolloidal gold labeling of Epon-embedded sections wasperformed as described previously (Bernadac et al., 1984; Chambraudet al., 1989). Briefly, immediatly after the animal was killed, 1-2 mmthick pieces of jejunum were excised at a distance of about 1 metrefrom the pylorus. They were then fixed in 2.5% glutaraldehyde in PBSfor 1 hour at 4°C. After being washed with PBS, they were post-fixedin 2% OsO4 in PBS for 1 hour at room temperature, and then dehy-drated in a graded ethanol series and embedded in Epon 812. The post-embedding technique used was then performed as follows: thinsections (0.06 µm) on grids were treated with 1% H2O2 in water for5 minutes, rinsed in water and then in PBS. In order to minimize anynon-specific adsorption, the sections were floated on 1.5% BSA inPBS for 10 minutes and incubated first with 3A4 hybridoma culturesupernatant for 90 minutes at room temperature. After washing, 3A4was revealed using a double-sandwich technique with rabbit anti-mouse Igs (dilution 1:1000; incubation 30 minutes) and then ProteinA-gold (incubation 1 hour). Protein A-gold (15 nm) was prepared aspreviously described (Bernadac et al., 1984). After being washed withwater, the sections were then lightly stained with uranyl acetate for 3minutes and lead citrate for 2 minutes.

ImmunofluorescenceThe immunofluorescence technique was performed as described pre-viously (Feracci et al., 1982), except that 0.25 µm sections of Epon-embedded rabbit jejunum were used instead of ultra-thin frozensections, in which the FBBG was lost (Gorvel et al., 1987a). Briefly,the sections were deposited on a glass slide and incubated for 5minutes in PBS containing 1.5% bovine serum albumin, and then for

2707FBBG and enterocyte differentiation

15 minutes with either the 3A4 mAb in undiluted culture supernatantor the 3.3A mAb in ascites fluid diluted 1:1000 in PBS containing1.5% BSA. After being washed with PBS, the sections were incubatedfor 30 minutes with fluorescein-labeled goat anti-mouse immunoglob-ulins diluted 1:100, washed, mounted in 50% glycerol, and examinedwith a Zeiss Photomicroscope III equipped with a III RS fluorescenceepicondenser.

RESULTS

Immunolabeling of rabbit jejunum sections with themonoclonal antibody 3A4Fig. 1 shows that along the whole villi and the upper part ofthe crypts, the 3A4 mAb specifically labeled the small intesti-nal surface coat. This coat is formed as shown in Fig. 2, by thefine filaments radiating from the tip of the brush border

Fig. 1. Localization of the antigen recognized by the monoclonalantibody 3A4, as shown by performing immunofluorescence labelingon longitudinal sections (0.2-0.25 µm) of Epon-embedded jejunumfrom an A+ rabbit. Two sections obtained successively wereimmunolabeled as follows: in A, with the 3.3A mAb directed againstthe human blood group A structures, which consequently revealed allthe goblet cell granules and the brush border membrane; in B, withthe 3A4 mAb that revealed only the components coating the brushborder.

microvilli that we called ‘filamentous brush border glycocalyx’(FBBG).

In the lower third of the crypts, the Paneth cells, goblet cellsand poorly differentiated enterocytes with a few short, irregularmicrovilli were never labeled. In the median third of the crypts,labeling occurred in more differentiated crypt enterocytes, suchas cell 1 in Fig. 3A and C, which are characterized by thepresence of numerous microvilli about 1-1.2 µm in length (ascompared with 2.5 µm in mature enterocytes), forming a shortbrush border devoid of any apical FBBG. In these cells, thewhole brush border membrane was labeled. Three to five cellshigher up the crypts, the apical FBBG was present, and boththe FBBG and the whole brush border membrane were labeled(cell 2 in Fig. 3A). The labeling of the lateral part of the brushborder membrane disappeared during the migration of the cellsto the upper part of the crypts. In the vicinity of the most deeplylocated labeled crypt enterocytes, the secretory granules ofsome goblet cells were labeled (Fig. 3D). Above this restrictedregion, no goblet cells were ever labeled, as shown in Figs 1and 3A.

Immunogold labeling using 3A4 mAb was performed onjejunum sections from four rabbits. Electron microscopy ofvilli present in these sections revealed that the FBBG of someenterocytes was devoid of any labeling (Fig. 4). Enterocyteswith unlabeled FBBG were rare in the case of one rabbit (seeFigs 1, 3 and 4), but constituted the majority of the enterocytesin the case of one other rabbit. In two other rabbits, the FBBGof about 50% of the enterocytes was unlabeled. In all fourrabbits, heterogeneous expression of the 3A4 antigenicityoccurred in the enterocytes of all the crypts examined. Nocrypts that were either unlabeled or uniformly labeled wereever detected.

Surface immunolabeling of rabbit small intestinalsheets with the 3A4 mAbInvestigation of the heterogeneous expression of 3A4 epitopesby the villus enterocytes was facilitated by performing surfacelabeling on the epithelium. As shown in Fig. 5, it was possibleafter most of the mucus has been eliminated, to label the villiwith 3A4 mAb, except in some rabbits (see below). At lowmagnifications (Fig. 5) the labeling seemed to be uniform,whereas at high magnifications (Fig. 6) two types of labelingwere observed: either all the enterocytes were labeled (Fig. 6A)or a patchy pattern of labeled and unlabeled enterocytes wasdetected (Fig. 6B). The proportion of labeled to unlabeledenterocytes on the various villi seemed to be uniform in anyone rabbit, but was highly variable from one rabbit to another(compare Fig. 6A and B). Identical results were obtained forthe jejunum and ileum.

When the labeling was performed using the 3.3 A mAb rec-ognizing the human blood group A structures, the pattern ofhomogeneous labeling shown in Figs 5B and 6C was alwaysobserved in A+ rabbits.

Immunoblotting analysis of the expression of 3A4and human blood group A antigenicities in rabbitsmall intestineTo test the presence of the 3A4 glycosidic structure in theFBBG of a large number of rabbits, high molecular masssoluble glycoproteins from 5 µg of mucosa obtained byscraping of the entire small intestine from each of 132 rabbits,

2708 J. Maury and others

Fig. 2. Specific labeling obtained on the FBBG of a mature rabbitenterocyte after performing immunocolloidal gold labeling with 3A4mAb on ultra-thin sections (600 Å) of Epon-embedded jejunum.Note the absence of any labeling on the lateral membrane ofmicrovilli.

were separated by SDS-PAGE, transferred onto Biodyne B andimmunostained with 3A4 mAb. In 50% of the rabbits, the 3A4mAb specifically labeled the 400 kDa band corresponding tothe FBBG (Maury et al., 1995), whereas in the other 50% no3A4 antigenicity was expressed in the small intestinal epithe-lium. The same blots were then immunostained with 3.3A mAbto test the expression of the human blood group A antigenic-ity in the same rabbits. The results obtained showed that: 19%of the rabbits were 3A4+ and A+, 31% were 3A4+ and A−, 26%were 3A4− and A+, and 24% were 3A4− and A−. It wastherefore concluded that no correlation seemed to existbetween the expression of 3A4 and human blood group A gly-cosidic structures.

Immunoprecipitation of FBBG from A+ rabbit withthe 3A4 mAbThe presence or absence of 3A4 epitopes in FBBG was alsotested by carrying out immunoprecipitation with the 3A4 mAbof FBBG from A+ rabbits, in which the FBBG can also bespecifically labeled by the 14A4 mAb which recognized thehuman blood group A type 3 structure (Gorvel et al., 1987a;Maury et al., 1995).

Complete immunoprecipitation of material recognized bythe 3A4 mAb present in the FBBG partially purified from A+

rabbit brush border membrane vesicles was performed by suc-cessively incubating the material with three samples of 3A4mAb beads. The bound and unbound material was thenanalysed using the immunoblotting technique with 3A4, 14A4and 3.3A monoclonal antibodies. The data given in Fig. 7 showthat, as expected, only the immunoprecipitated FBBG wasrevealed by these three antibodies. The results obtained with14A4 mAb showed that FBBG devoid of 3A4 epitopes waspresent in the unbound material.

DISCUSSION

Techniques involving aggregation chimeras (Ponder et al.,1985; Schmidt et al., 1985), transgenic mice (Gordon, 1989;Sweetser et al., 1988) and somatic mutations (Winton et al.,1988), yielding mice composed of two well defined intestinalgenotypes, have shown that in adult mice, the epithelium ofone crypt is always composed of a single clone of cells that

Fig. 3. Part of the median third of one crypt observed afterperforming immunocolloidal gold labeling with the 3A4 mAb onultra-thin sections (600 Ä) of Epon-embedded jejunum. In A, fivecells can be easily delimited thanks to the presence of desmosomes,indicated by black arrowheads. Observations at higher magnificationshowed that the dark spot indicated by a white arrowhead was not adesmosome. The four enterocytes were numbered from 1 to 4, cell 1being the lowest in the crypt. The appearance of the apical FBBG,forming the enteric surface coat, can be observed in cell 2. Note thatit shows up on only one half of the microvilli of this cell. In B,magnification of part of cell 3. Note the labeling of both the FBBGand the lateral membrane of microvilli. In C, magnification of thepart of cell 1 not shown in A. Note the absence of apical FBBG andthe labeling of the whole membrane of short (1-1.2 µm) microvilli.In D, a goblet cell located in the same zone as the enterocyte shownin C, in which some of the mucus granules are labeled, whereas, asshown in A, the goblet cell located four cells higher up wasunlabeled.

2709FBBG and enterocyte differentiation

2710 J. Maury and others

Fig. 4. Brush border of three adjacent mature enterocytes after immunocolloidal gold labeling with the 3A4 mAb of ultra-thin section of Epon-embedded jejunum. Note the absence of any labeling of the FBBG synthesized by the central cell.

migrate in straight lines along the crypt-villus axis (Schmidt etal., 1985; Winton et al., 1988). The epithelium of one villustherefore originates from several crypts (Wright and Irwin,1982). In these experimental systems, the epithelium coveringone villus is formed of ribbons of cells expressing one of thetwo intestinal genetic markers introduced into the mice (forreview see Gordon, 1989).

It was recently suggested that a heterogeneous differen-tiation of enterocytes originating from one crypt might occur,since different patchy patterns of expression were observed inthe case of either lactase in hypolactasic adult humans (Maiuriet al., 1991), and adult rats and rabbits (Maiuri et al., 1992), orblood group A structure in A non-secretor humans (Maiuri etal., 1993). Here we provide another example of this heteroge-neous pattern of expression in the case of enterocytes origi-nating from the same crypt. This finding has to do with thesynthesis of the FBBG-specific glycosidic 3A4 structure rec-ognized by the 3A4 monoclonal antibody.

Electron microscopy observations carried out on Epon-embedded jejunum sections have shown that in the upper partof crypts and along the villi, all the enterocytes synthesizeFBBG, a mucin-like glycoprotein (Maury et al., 1995) whichforms the enteric surface coat first described by Ito (1965).Previous studies (Bernadac et al., 1984; Gorvel et al., 1987b)have shown the presence on FBBG of human blood group gly-

cosidic structures of types A in A+ rabbits and H in A− rabbits.We recently established that the FBBG from both of thesetypes of rabbit also bear a glycosidic epitope containing at leastone O-acetylated sialic acid recognized by the 3A4 mAb(Maury et al., 1995). Immunoblotting analysis and immunola-beling experiments on jejunum epithelium from several rabbitsshowed however that the 3A4 structure was synthesized byenterocytes in only 50% of the rabbits tested.

Whereas the A antigenicity is always present on the FBBGof all the enterocytes in A+ rabbits, the 3A4 epitopes arepresent on the FBBG of either all or only a variable number ofenterocytes in rabbits expressing the 3A4 antigenicity. In thislast case, the 3A4 immunolabeled enterocytes have a patchydistribution over all the villi. Apparently, two programs ofdifferentiation take place during differentiation in the crypts,where labeled and unlabeled immature enterocytes wereobserved. This leads to the expression of different glycosyl-transferase(s) and/or O-acetyl transferase(s), which participatein the synthesis of the glycosidic moiety of the FBBG (Corfieldet al., 1976; Paulson and Colley, 1989; Ichikawa et al., 1992).

In the vicinity of the less differentiated labeled enterocytes,the mucus granules of some goblet cells were also labeled,whereas no goblet cells were ever labeled above this level. Inthis restricted region of the crypts, the expression of enzymesinvolved in the synthesis of the 3A4 epitope might occur not

2711FBBG and enterocyte differentiation

Fig. 5. In A, indirect immunoperoxidase surface labeling of jejunumvilli, using the 3A4 mAb. In B, jejunum from A+ rabbit labeled withthe 3.3A mAb. Note the presence of small pieces of labeled mucussheet remaining on the epithelial surface. In C, a control experimentin which only the second antibody coupled to peroxidase was used.

Fig. 6. Patterns of labeled enterocytes on a villus dissected out afterindirect immunoperoxidase surface labeling of jejunum from tworabbits. After performing 3A4 mAb labeling, a uniform patern oflabeling was observed on all the enterocytes in the rabbit shown inA, whereas a patchy pattern of labeled and unlabeled enterocytes wasobserved in the rabbit shown in B. The uniform pattern of labelingshown in C was always obtained with the 3.3A mAb in A+ rabbits.

only in enterocytes but also in goblet cells where a rapidchange in the pool of enzymes involved in the mucin glyco-sylation might result in the disappearance of the 3A4 structure.The possibility cannot be excluded that the enterocytes whichexpress no 3A4 antigenicity on their FBBG may, as in the caseof the goblet cells, transiently synthesize the 3A4 antigenicdeterminant. Modifications (Corfield et al., 1976; Schauer etal., 1988) such as further acetylation, deacylation, glycosyla-tion instead of acetylation, and/or cleavage of the O-acetylatedsialic acid of the 3A4 epitope might destroy the 3A4 anti-genicity. A further glycosylation of either the chain bearingthis O-acetylated sialic acid or a neighbouring chain might alsomask the 3A4 epitope.

The differentiation of the enterocytes along the crypt-villusaxis has been investigated by studying the expression ofvarious markers. Hydrolases such as aminopeptidase N(Gorvel et al., 1986; Norén et al., 1989) and sucrase-isomal-tase (Hauri et al., 1980; Traber, 1990), which are integrated

into the whole brush border membrane, have been used asmarkers of terminal differentiation although some monoclonalantibodies have also served to detect inactive sucrase-isoma-ltase and dipeptidylpeptidase IV in the microvilli of cryptenterocytes (Beaulieu et al., 1989; Gorvel et al., 1991). Therealso exist specific markers of crypt cells (Quaroni, 1985). Likesome brush border membrane glycoproteins (Gorvel et al.,1986), brush border cytoskeleton proteins are expressed alongthe whole crypt-villus axis and, except for myosin I, are con-centrated in the apical pole of even very poorly differentiatedcrypt cells (Fath et al., 1990; Heintzelman and Mooseker,1990; Louvard et al., 1992). Myosin I is associated with thebrush border cytoskeleton only in villus enterocytes (Heintzel-man and Mooseker, 1990). From the bottom to the upper partof the crypts, however, the increase in the number and lengthof microvilli, which is correlated with the cytoskeleton organ-

2712 J. Maury and others

Fig. 7. Immunoprecipitation of FBBG induced by 3A4 mAb. Afterpartial purification from brush border membrane vesicles, thematerial recognized by the 3A4 mAb was completelyimmunoprecipitated upon performing successive incubations withthree identical samples of 3A4-beads. The material bound to eachbead sample (lanes 1 to 3), as well as the unbound material (lane 4),was analysed by carrying out immunoblotting experiments using thevarious mAbs indicated at the tops of the columns.

ization, is the most obvious index to the progressive entero-cyte differentiation (Fath et al., 1990; Leblond, 1981). In thelower region of the crypts, the microvilli are sparse, short andrun at acute angles to the cell surface. During cell migration,the microvilli become more numerous, longer and straighter.In the rabbit, we observed the formation of a short brushborder composed of numerous regular microvilli 1-1.5 µmlong, in the mid part of the crypt; whereas the brush bordermicrovilli reached a length of 2.5 µm in the mid part of thevilli.

Here the pattern of immunolabeling observed in the cryptenterocytes using 3A4 mAb showed the existence of threesteps in the differentiation of the brush border membrane. Theexpression of 3A4 antigenicity began in the median third of thecrypts. In these cells, 3A4 epitopes were present over the wholemembrane of the short microvilli, and there was no FBBG con-centrated at the tips. Three to five cells higher up, the apicalFBBG was visible at the tips of the short microvilli of entero-cytes which were in an intermediate state of differentiationcharacterized by the presence of 3A4 epitopes on the apicalFBBG at microvillar tips as well as on the whole brush bordermembrane. The 3A4 antigenicity of the lateral membrane ofthe microvilli disappeared gradually during the migration ofcells to the upper part of the crypts and no longer existed inmature enterocytes. If the 3A4 determinants were specificallyborne by FBBG in immature cells, as in the mature enterocytes,their presence along the whole brush border membrane wouldsuggest that these poorly differentiated enterocytes had not yetacquired their hyper-polarity, i.e. that the FBBG had not yetformed clusters at the tips of the microvilli. This phenomenonmight be induced by a specific step in the assembly of sub-membrane cytoskeletal component(s). By conducting parallelstudies on the two events it should be possible to test thishypothesis.

This study was supported by grants from Association Française deLutte contre la Mucovisidose and Institut National de la Santé et dela Recherche Médicale (CRE 930204). The authors thank JessicaBlanc for revising the English manuscript.

REFERENCES

Bara, J., Gautier, R., Le Pendu, J. and Oriol R. (1988). Immunochemicalcharacterization of mucins. Polypeptide (m1) and polysaccharide (A andLeb) antigens. Biochem. J. 245, 185-193.

Beaulieu, J. F., Nichols, B. and Quaroni, A. (1989). Posttranslationalregulation of sucrase-isomaltase expression in intestinal crypt and villuscells. J. Biol. Chem. 264, 20000-20011.

Bernadac, A., Gorvel, J. P., Feracci, H. and Maroux, S. (1984). Humanblood group A-like determinants as marker of the intracellular pools ofglycoproteins in secretory and absorbing cells of A+ rabbit jejunum. Biol.Cell 50, 31-36.

Chambraud, L., Bernadac, A., Gorvel, J. P. and Maroux, S. (1989).Renewal of goblet cell mucus granules during the cell migration along thecrypt-villus axis in rabbit jejunum: an immunolabeling study. Biol. Cell 65,151-162.

Cheng, H. and Leblond, C. P. (1974a). Origin, differentiation and renewal ofthe four main epithelial cell types in the mouse small intestine. I. Columnarcells. Am. J. Anat. 141, 461-480.

Cheng, H. and Leblond, C. P. (1974b). Origin, differentiation and renewal ofthe four main epithelial cell types in the mouse small intestine. II. Mucouscells. Am. J. Anat. 141, 481-502.

Cheng, H. and Leblond, C. P. (1974c). Origin, differentiation and renewal ofthe four main epithelial cell types in the mouse small intestine. V. Unitariantheory of the origin of the four epithelial cell types. Am. J. Anat. 141, 537-562.

Corfield, A. P., Ferreira Do Amaral, C., Wember, M. and Schauer, R.(1976). The metabolism of O-acyl-N-acylneuraminic acids. Biosynthesis ofO-acylated sialic acids in bovine and equine submandibular glands. Eur. J.Biochem. 68, 597-610.

Fath, K. R., Obenauf, S. D. and Burgess D. (1990). Cytoskeletal protein andmRNA accumulation during brush border formation in adult chickenenterocytes. Development 109, 449-459.

Feizi, T., Gooi, H. C., Childs, R. A., Picard, J. K., Uemura, K., Loomes, L.M., Thorpe, S. J. and Hounsell, E. F. (1984). Tumour-associated anddifferentiation antigens on the carbohydrate moieties of mucin-typeglycoproteins. Biochem. Soc. Trans. 12, 591-596.

Feizi, T. and Childs, R. A. (1987). Carbohydrates as antigenic determinants ofglycoproteins. Biochem. J. 245, 1-11.

Feracci, H., Bernadac, A., Gorvel, J. P. and Maroux, S. (1982). Localizationby immunofluorescence and histochemical labeling of aminopeptidase N inrelation to its biosynthesis in rabbit and pig enterocytes. Gastroenterology82, 317-324.

Feracci, H., Rigal, A. and Maroux, S. (1986). Biosynthesis and intracellularpool of aminopeptidase N in rabbit enterocytes. J. Membr. Biol. 83, 39-46.

Gane, P., Mollicone R., Rauger Ph. and Oriol R. (1987). Inhibition ofhaemagglutination with synsorbs and salivas of anti-A monoclonalantibodies. Rev. Fr. Transfusion et Immuno-Hématologie 30, 435-442.

Gordon, J. I. (1989). Intestinal epithelial differentiation: new insights fromchimeric and transgenic mice. J. Cell Biol. 108, 1187-1194.

Gorvel, J. P., Wisner-Provost, A. and Maroux S. (1982). Identification ofglycoproteins bearing human blood group A determinants in rabbitenterocyte plasma membranes. FEBS Lett. 143, 17-20.

Gorvel, J. P., Rigal, A., Olive, D., Mawas, C. and Maroux, S. (1986).Identification of an early expressed marker of the luminal membrane ofrabbit small intestinal columnar cells. Presence of a homologous antigen inkidney proximal tubules and glomeruli. Biol. Cell 56, 121-126.

Gorvel, J. P., Bourges, F., Ferrero, A., Bernadac, A., Rasongles, P. andMaroux, S. (1987a). Expression of a subclass of human blood group Aantigenicity in A+ rabbit jejunum: specific glycosylation of the glycocalyxand a 140 K brush border glycoprotein. Biol. Cell 59, 145-152.

Gorvel, J. P., Joly, I., Rigal, A., Chambraud, D., Bourges, F. and Maroux,S. (1987b). Expression of the human A, B and related antigens in the jejunumof human and rabbit small intestines in particular on the brush bordermembrane glycoproteins. Rev. Fr. Transfusion et Immuno-Hématologie 30,471-483.

Gorvel, J. P., Ferrero, A., Chambraud, L., Rigal, A., Bonicel, J. and

2713FBBG and enterocyte differentiation

Maroux, S. (1991). Expression of sucrase-isomaltase anddipeptidylpeptidase IV in human small intestine and colon.Gastroenterology 101, 618-625.

Hauri, H. P., Quaroni, A. and Isselbacher K. J. (1980). Monoclonal antibodyto sucrase/isomaltase: Probes for the study of postnatal development andbiogenesis of the intestinal microvillus membrane. Proc. Nat. Acad. Sci. USA77, 6629-6633.

Heintzelman, M. B. and Mooseker, M. S. (1990). Assembly of the brushborder cytoskeleton: changes in the distribution of microvillar core proteinsduring enterocyte differentiation in adult chicken intestine. Cell Motil.Cytoskel. 15, 12-22.

Ichikawa, Y., Look, G. C. and Wong, C. H. (1992). Enzyme-catalysedoligosaccharide synthesis. Anal. Biochem. 202, 215-238.

Ito, S. (1965). The enteric surface coat on cat intestinal microvilli. J. Cell Biol.27, 475-491.

Ito, S. (1974). Form and function of the glycocalyx on free cell surfaces. Phil.Trans. R. Soc. Lond. B, 268, 55-66.

Leblond, C. P. (1981). The life history of cells in renewing systems. Am. J.Anat. 160, 114-158.

Louvard, D., Kedinger, M. and Hauri, H. P. (1992). The differentiatingintestinal epithelial cell: Establishment and maintenance of functionsthrough interactions between cellular structures. Annu. Rev. Cell Biol. 8, 157-195.

Maiuri, L., Raia, V., Potter, J., Swallow, D., Ho, W. M., Fiocca, R., Finzi,G., Cornaggia, M., Capella, C., Quaroni, A. and Auricchio, S. (1991).Mosaic pattern of lactase expression by villous enterocytes in human adult-type hypolactasia. Gastroenterology 100, 359-369.

Maiuri, L., Rossi, M., Raia, V., D’Auria, S., Swallow, D., Quaroni, A. andAuricchio, S. (1992). Patchy expression of lactase protein in adult rabbit andrat intestine. Gastroenterology 103, 1739-1746.

Maiuri, L., Raia, V., Fiocca, R., Solcia, E., Cornaggia, M., Norén, O.,Sjöström, H., Swallow, D., Auricchio, S. and Dabelsteen, E. (1993).Mosaic differentiation of human villus enterocytes: patchy expression ofblood group A antigen in A non secretors. Gastroenterology 104, 21-30.

Maury, J., Nicoletti, C., Guzzo-Chambraud, L., and Maroux, S. (1995).The filamentous brush border glycocalyx, a mucin-like marker of enterocytehyper-polarization. Eur. J. Biochem. 228, 323-331.

Massey, D., Feracci, H., Gorvel, J. P. Rigal, A., Soulié, J. M. and Maroux, S.(1987). Evidence for the transit of aminopeptidase N through the basolateralmembrane before it reaches the brush border of enterocytes. J. Membr. Biol.96, 19-25.

Norén, O., Dabelsteen, E., Hoyer, P. E., Olsen, J., Sjöström, H. and Hansen,G. H. (1989). Onset of transcription of the aminopeptidase N (leukemiaantigen CD 13) gene at the crypt/villus transition zone during rabbitenterocyte differentiation. FEBS Lett. 259, 107-112.

Oriol, R. and Dalix, A. M. (1977). Differences in the maturation of theimmune response of A− and A+ rabbits. Good and poor respondersrespectively for the A antigen. Immunology 33, 91-99.

Paulson, J. C. and Colley, K. J. (1989). Glycosyltransferases. Structure,localization, and control of cell type-specific glycosylation. J. Biol. Chem.264, 17615-17618.

Podolsky, D. K. (1985a). Oligosaccharide structures of human colonic mucin.J. Biol. Chem. 260, 8262-8271.

Podolsky, D. K. (1985b). Oligosaccharide structures of isolated human colonicmucin species. J. Biol. Chem. 260, 15510-15515.

Ponder, B. A., Schmidt, G. A., Wilkinson, M. J., Wood Monk, M. and Raid,A. (1985). Derivation of mouse intestinal crypts from single progenitor cells.Nature 313, 689-691.

Pryde, J. G. (1986). Triton X-114: a detergent that has come in from the cold.Trends Biochem. Sci. 11, 160-163.

Quaroni, A. (1985). Crypt cell development in newborn rat small intestine. J.Cell Biol. 100, 1601-1610.

Roussel, Ph., Lamblin, G., Lhermitte, M., Houdret, N., Lafitte, J. J.,Prerini, J. M., Klein, A. and Scharfman, A. (1988). The complexity ofmucins. Biochimie 70, 1471-1482.

Schauer, R., Reuter, G., and Stoll, S. (1988). Sialate O-acetylesterases: keyenzymes in sialic acid catabolism. Biochimie 70, 1511-1519.

Schmidt, G. H., Wilkinson, M. H. and Ponder, B. A. (1985). Cell migrationpathway in the intestinal epithelium: An in situ marker system using mouseaggregation chimeras. Cell 40, 425-429.

Schmitz, J., Preiser, H., Maestracci, D., Ghosh, B. K., Cerda, J. J. andCrane, R. K. (1973). Purification of human intestinal brush bordermembrane. Biochim. Biophys. Acta 323, 98-112.

Strous, G. J. and Dekker, J. (1992). Mucin-type glycoproteins. Crit. Rev.Biochem. Mol. Biol. 27, 57-92.

Sweetser, D. A., Hauft, S. M., Hoppe, P. L., Birkenmeier, E. H. andGordon, J. I. (1988). Transgene mice containing intestinal fatty acid-binding protein-human growth hormone fusion genes exhibit correctregional and cell-specific expression of the reporter gene in their smallintestine. Proc. Nat. Acad. Sci. USA 85, 9611-9615.

Traber, P. G. (1990). Regulation of sucrase-isomaltase gene expression alongthe crypt-villus axis of rat small intestine. Biochem. Biophys. Res. Commun.173, 765-773.

Winton, D. J., Blaunt, M. A. and Ponder, B. A. (1988). A clonal markerinduced by mutation in mouse intestinal epithelium. Nature 333, 463-466.

Wright, N. A. and Irwin, M. (1982). The kinetics of villus cell populations inthe mouse small intestine. I. Normal villi: the steady state requirements. CellTiss. Kinet. 15, 595-609.

Zweibaum, A. and Bouhou, E. (1973). The alloantigen-alloantibody system inrabbits. Transplantation 15, 291-293.

(Received 14 October 1994 - Accepted, in revised form,29 March 1995)