Development 104, 1-14 (1988)Printed in Great Britain © The Company of Biologists Limited 1988

The expression of epidermal antigens in Xenopus laevis

KEIJI ITOH, AKIKO YAMASHITA* and H1ROSHI Y. KUBOTA

Department of Zoology, Faculty of Science, Kyoto University, Kyoto 606, Japan

'Present address: Department of Physiology, Primate Research Institute, Kyoto University, Inuyama, Aichi-ken 484, Japan

Summary

Five kinds of monoclonal antibodies that are specificfor the epidermis of Xenopus embryos were produced.Epidermis-specific antibodies were used to investigatethe spatial and temporal expressions of epidermalantigens during embryonic and larval development.The cells that were recognized by the antibodies at thelarval stage are as follows: all of the outer epidermalcells and cement gland cells were recognized by theantibody termed XEPI-1, all of the outer and innerepidermal cells, except the cement gland cells, wererecognized by XEPI-2 antibody, the large mucusgranules and the apical side of the outer epidermalcells, except for the ciliated epidermal cells, wererecognized by XEPI-3 antibody, the large mucusgranules and basement membrane were recognized byXEPI-4 antibody, and the small mucus granulescontained in the outer epidermal cells as well asextracellular matrices were recognized by the anti-body termed XEPI-5. All of the epidermal antigens,

except XEPI-4, were first detected in the epidermalregion of the late gastrula or early ncurula. TheXEPI-4 antigen was first detected in stage-26 tail-budembryos. None of these antigens were expressed by theneural tissues at any time during embryonic develop-ment. Only the XEPI-2 antigen continued to beexpressed after metamorphosis, while the expressionof the other antigens disappeared during or beforemetamorphosis. The specificity of the antibodies al-lowed us to classify the epidermal cells into four typesin early epidermal development. The four types ofepidermal cells are (1) the outer epidermal cells thatcontain small mucus granules, (2) the ciliated epider-mal cells, (3) the outer epidermal cells that containlarge mucus granules and (4) the inner sensorial cells.

Key words: Xenopus laevis, epidermal differentiation,epidermal cells, monoclonal antibodies, epidermalantigens, metamorphosis.

Introduction

In the early development of Xenopus, the epidermaltissues are derived primarily from the animal-ventralblastomeres and partially from the animal-dorsal andthe vegetal-ventral blastomeres of an 8-cell-stageembryo (Masho & Kubota, 1986). After severalcleavages, the prospective epidermal area occupiesmore than half of the animal hemisphere and spreadstoward the ventral side of the early gastrula (Keller,1975). During gastrulation, the surface of the embryois gradually covered by the presumptive ectodermdue to morphogenetic movements. The dorsal ecto-derm, in contact with inductive chordamesoderm,thickens to form the neural plate which is sub-sequently folded into the neural tube. After thecompletion of neurulation the epidermis covers theentire surface of the embryo.

Previously, a detailed SEM study of gastrulationshowed that during epiboly the multiple layers of thepresumptive epidermal region interdigitate radiallyinto two distinct layers (Keller, 1980), which aretermed the outer epithelial layer and the innersensorial layer (Nieuwkoop & Faber, 1956). Theepidermis consists of two layers until metamorphosis,a period when the number of epidermal layers in-creases to about five layers. Later during metamor-phosis, the outer layers of the larval skin are shed.After metamorphosis, at stage 66, the adult epidermiscomprises more than five layers.

In the course of epidermal development as de-scribed above, several kinds of epidermal cells aredifferentiated (Fox, 1984), therefore, it is likely thatmany molecules must participate in the process. Toinvestigate epidermal differentiation, molecularmarkers specific for the various kinds of epidermal

2 K. Itoh, A. Yamashlta and H. Y. Kubota

cells are needed.Several epidermal markers in amphibian embryos

have been reported thus far: epimucin, a majorglycoprotein in Ambystoma that is an epidermalreceptor for peanut agglutinin (Slack, 1985) and agroup of epidermis-specific antigens identified bymonoclonal antibodies (Jones, 1985; Jones & Wood-land, 1986; Akers et al. 1986). Furthermore, thetranscription of cytokeratin genes has been used toinvestigate the differentiation of epidermal cells(Jamrich et al. 1987).

In the present investigation, we have obtained fivekinds of monoclonal antibodies that recognize differ-ent antigens expressed by epidermal cells. Immuno-histochemical observations on the spatial and tem-poral expression of epidermal antigens showed thatfour types of epidermal cells can be identified inXenopus embryos.

Materials and methods

Collection of eggsEggs of Xenopus laevis were obtained through artificialmating after the injection of human chorionic gonadotropininto the dorsal lymph sac at a dose of 300 i.u. for femalesand of 200 i. u. for males. Embryos were staged according toNieuwkoop & Faber (1956).

Preparation of antigensWe used two kinds of antigens for immunization. Embryosat stages 35-38 were homogenized in modified Steinberg'ssolution buffered to pH 7-0 by 3 mM-Hepes-NaOH in placeof Tris-HCl. The centrifugation of the homogenate at1600g for 10 min resulted in the separation of the homogen-ate into five layers. The five layers, from the top of thecentrifuge tube to the bottom of the tube are as follows: thelayer that contained lipid droplets, the clear layer ofcytoplasm, the grey coloured layer of cytoplasm, the layercontaining pigment granules and the layer that containedyolk platelets. The layer of clear cytoplasm was centrifugedagain at 8000 g for 10 min to pellet any residual yolkplatelets and to discard the top layer of the lipid droplets.After the lipid and the pellet were discarded, the remainingclear cytoplasm was used as immunogen.

For the second source of antigens, neural and epidermaltissues were dissected from stage-33 to -38 larvae usingforceps. Neural and epidermal tissues collected from 200embryos were placed in modified Steinberg's solution andforced through a 25-gauge syringe. The suspension ofdisrupted neural and epidermal tissues was used as thesecond kind of immunogen.

Production of hybridomasBALB/c mice were injected intraperitoneally three timeswith 0-4 ml of the clear cytoplasm or the suspension ofdisrupted neural and epidermal tissues every two to fourweeks. Three days after the last immunization, murinespleen cells were fused with myeloma cells (P3X63Ag8U-l)

in the presence of 50 % (w/v) polyethylene glycol 4000(Nakarai Chem. Co. Ltd, Kyoto). Cells were suspended inDMEM (Dulbecco's modified Eagle medium) after cellfusion and plated into 96-well plastic dishes. After approxi-mately 12h, DMEM that contains HAT (hypoxanthine-aminopterin-thymidine) was added to the wells to raisehybridomas according to the method of Littlefield (1964).Approximately one week later, antibodies were screenedby staining histological sections of stage-35/-36 embryosusing indirect immunofluorescent microscopy. Hybridomasthat secreted epidermis-specific antibodies were cloned onthe feeder cells (thymus or spleen cells) by picking up asingle cell under an inverted microscope with a fine-tippedPasteur pipette.

Fixation and embeddingWhole embryos at various stages or excised epidermaltissues were fixed in 100% methanol for 40 min at —20°C.Next, the fixative was replaced with 100 % ethanol at—20°C. Soon after the replacement, the samples were leftat room temperature for 15 min to warm gradually thesamples. The ethanol was replaced with fresh absoluteethanol at room temperature. After 15 min the sampleswere incubated in 50% polyester wax (Steedman, 1957;BDH Chem. Ltd) in absolute ethanol for 30 min followedby impregnation with 100 % polyester wax for 1 h at 40°C.Samples were embedded in the polyester wax and stored at4°C.

Sectioning and stainingEmbedded samples were sectioned at 18°C using a micro-tome. 10 jim sections were adhered to cover slips(4-5x24mm) using 0 4 % amylopectin. After the sectionsdried overnight at room temperature, they were soaked in100 % ethanol for 15 min and washed with phosphate-buffered saline (PBS). The sections were incubated with100n\ of monoclonal supernatant for 1 h, washed with PBSand incubated with 10 /.tl of FITC-conjugated rabbit anti-mouse igG (Miles-Yeda Ltd) for 30 min. After the sectionswere washed with PBS for 15 min, they were mounted with80% glycerol in PBS, and examined under an epifluor-escence microscope.

Immunoelectron microscopyImmunoelectron microscopy was performed as describedby Asada-Kubota (1988).

ImmunoblottingThe proteins that were contained in the clear cytoplasm,obtained by centrifugation of homogenized stage-35/-36embryos, and contained in homogenized epidermal tissues(stage 54) were separated on SDS-polyacrylamide gels andblotted onto nitrocellulose sheets. The nitrocellulose sheetswere soaked in 0-5 % skimmed milk and cut into strips.Strips were incubated with the hybridoma culture super-nates, washed three times with PBS containing 0-1%Tween 20 and washed three times with PBS without Tween20. Strips were incubated with 35S-labelled anti-mouse IgG(Amersham), washed with PBS, dried and exposed for 5days to Fuji RX X-ray film at -80°C.

Peanut lectin blockingPeanut lectin blocking was carried out by incubating sec-tions with peanut lectin (Hohnen Oil Co. Ltd, Tokyo) at100j<gml"' for 30min and washing them with PBS prior tostaining the sections with the monoclonal antibodies andFITC-conjugated anti-mouse IgG.

Neuraminidase treatmentSections were incubated overnight at 37°C with 4mi.u.neuraminidase (Seikagaku Kogyo Co. Ltd, Tokyo) in0-4 ml 50mM-sodium acetate (pH6-0), 80mM-NaCl, and5mM-CaCl2, before the sections were incubated with themonoclonal antibodies and FITC-conjugated anti-mouseIgG.

Results

Five kinds of monoclonal antibodies (XEPI-1,XEPI-2, XEPI-3, XEPI-4 and XEPI-5) that recog-nized Xenopus epidermal antigens were obtained.The staining pattern of each antibody was examinedat various stages from the fertilized egg (stage 1) tothe frog at postmetamorphosis (stage 66). Each of themonoclonal antibodies showed a characteristic stain-ing pattern that made it possible to locate specificantigens in epidermal cells. The spatial and temporalexpression of the antigens recognized by the fiveantibodies are shown in Figs 1-11 and Table 1.

Expression and localization of epidermal antigensduring embryonic developmentTransverse or sagittal sections of Xenopus embryosfrom stage 1 to stage 40 were stained with theantibodies. None of the antibodies reacted withneural tissues at any stages of development that wereexamined.

Epidermal antigens in Xenopus 3

XEPI-1XEPI-1 antibody did not bind to sections until afterstage 12. A weak but distinct binding to the epidermalregion of the embryo was first observed at stage 12i.The staining became stronger as development pro-ceeded (Fig. 1). XEPI-1 was never bound to anytissue other than the epidermis. The antigen recog-nized by XEPI-1 was localized exclusively in the outerepithelial layer. Furthermore, the apical side of theouter epithelial cells was the most heavily stained partof the cell. All of the cells that constitute the outerepithelial layer were stained at each developmentalstage up to the time of metamorphosis. Asada-Kubota (1988) observed by immunoelectron micro-scopic study the binding of XEPI-1 on the moderatelyelectron-dense bodies and the cortical dense ma-terials in the outer epithelial cells. Cement gland cellswere also recognized by the XEPI-1 antibody(Fig. ID). Among five antibodies that were tested,only the XEPI-1 stained the cement gland cells.

XEPI-2The antigen recognized by XEPI-2 first appeared inthe epidermal region at stage 13. In the neurula, theantigen recognized by XEPI-2 was found mainly inthe apical side of the outer layer and very weakly inthe inner layer (Fig. 2A,B). In the tail-bud embryos,the antigen could be clearly detected in both layers(Fig. 2C). The apical side of the outer layer wasstained more strongly at this stage. In the larvae(stage 36), both layers were intensely stained withXEPI-2 (Fig. 2D,E).

Weak binding of the XEPI-2 antibody to theintersegmental region of somites from stage 26 tostage 34 was also observed (Fig. 2C).

XEPI-3The antigen that was recognized by XEPI-3 first

Table 1. Temporal expression of the XEPI-1, -2, -3, -4, and -5 antigens in Xenopus

stage

antibodies location 1 8 10 12 12.5 13 14 16 18 20 24 26 28 36 40 46 48 49 52 54 56 58 60 62 64 66

XEPI-1XEPI-2

XEPI-3

XEPI-4

XEPI-5

o. e. c.e. c.

o. e. c.

- - - - + + + +- - - - - + + +

++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ - -+ + + ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ + +

- - - - - + + + + + ++ ++ ++ ++ + + + + + + + + - - -1. m. g. - - - - - - - - - - - - - + ++ ++ ++ + ++ ++ ++ ++ ++ + - -1. m. g. - - - - - - - - - - - + + + + + + + + + + + + + + + + + + + + + + + +b. m. - - - - - - - - + + + + + + + + + + + + + + + + + +s. m. g. - - - - - + + + + + + + + + + + + + + + + - - - - - - - +ECM - + + + + + + + + + + + + - _ _ _ _ _ _ _ _ _ _ _ _

The temporal expression of the antigens recognized by five antibodies are summarized. Transverse or sagittal sections of whole embryos from stage 1to stage 49 were stained with the five antibodies and FITC-conjugated rabbit antimouse IgG. Transverse sections of dorsal epidermis from stage 52 tostage 66 were also stained with these antibodies. - : negative staining, +: positive staining, + + : heavy staining, b. m., basement membrane; ECM,extracellular matrices; e. c , epidermal cells; i. s., intersegmental region of somites; 1. m. g., large mucus granules; o. e. c , outer epidermal cells;s. m. g., small mucus granules.

K. Itoh, A. Yamashita and H. Y. Kubota

Fig. 1. Expression of the XEPI-1 antigen in Xenopus embryos. Transverse polyester wax sections were stained byindirect immunofluorescence with XEPI-1 and FITC-conjugated rabbit antimouse IgG. (A) Stage 13; (B) stage 18;(C) stage 24; (D) stage 40. Bar, 100 ;um.

appeared on the apical side of the outer epidermallayer at stage 13 and was strongly expressed up to thelarval stage (Fig. 3). It was characteristic that thestaining pattern was interrupted by the ciliated epi-dermal cells that are scattered throughout the outerlayer (Fig. 3C,D,E). Beginning at stage 36 and con-tinuing up to the time of metamorphosis, there wasalso detected binding of XEPI-3 to the large mucusgranules contained in some of the epidermal cells(Fig. 3D.E).

XEP1-4The epidermal antigen that was recognized by XEPI-4 could not be detected until stage 26. At stage 28,there was binding of XEPI-4 to intermediate-sizedgranules contained in some of the outer epithelialcells (Fig. 4A). In stage-36 larvae, these granulesincreased in size and were stained more heavily withXEPI-4 (Fig. 4B,C). The granules seemed to leak outfrom the cell (Fig. 4C).

The XEPI-4 antibody also bound to the noto-chordal sheath from stage 18 and to the basementmembrane from stage 28 (Fig. 4).

XEPI-5The epidermal antigen was first detected in the

epidermal region at stage 13 (Fig. 5B). The apicalside of the outer epithelial cells was heavily stained(Fig. 5B-E) whereas the ciliated epidermal cells werenot stained (Fig. 5D,E).

The XEPI-5 antibody recognized not only epider-mal tissues but also extracellular matrices (ECM).XEPI-5 was bound to the ECM in the blastocoel ofblastulae (Fig. 5A). Also, this antibody was bound tothe ECM that surrounds the notochord, the neuraltube, and the somites at later stages (Fig. 5B-D).

Immunoelectron microscopic observation showedthat XEPI-5 was bound to small mucus granulescontained in the apical side of the outer epithelialcells (Fig. 5F) and the ECM.

Comparison of distribution of antigens at stage 36The localization of the five different epidermal anti-gens described above were compared in the stage-36larva (Fig. 6).

The distribution of the XEPI-1 antigen was con-fined to the outer epithelial cells (Fig. 6A). TheXEPI-2 antigen was detected in both the outerepithelial cells and the inner sensorial cells (Fig. 6B).The XEPI-3 antibody was bound to the apical side ofthe outer epidermal cells, except for the ciliated cells.Also, XEPI-3 was bound to the large granules con-

Epidermal antigens in Xenopus

Fig. 2. Regional distribution of antigens recognized by XEPI-2 in Xenopus embryos. Transverse polyester wax sectionswere stained with XEPI-2 and FITC-conjugated rabbit antimouse IgG. (A) Stage 16; (B) stage 20; (C) stage 26;(D,E) stage 36. Bar,

tained in some of the outer epithelial cells (Fig. 6C).Fig. 6D shows the staining pattern after sections werestained with the XEPI-4 antibody. Fig. 6C and D arephotomicrographs of the adjacent sections of serialsections. These photomicrographs show that theXEPI-4 antibody was bound to the same large gran-ules as those that were stained using the XEPI-3antibody. Also, the XEPI-4 antibody was bound tothe basement membrane (Fig. 6D). The XEPI-5antigen was localized in the small mucus granules onthe apical side of the outer epithelial cells. Ciliatedepidermal cells were not stained with XEPI-5 anti-body (Fig. 6E).

Thus, we could identify at least four kinds ofepidermal cells in Xenopus larvae at stage 36 bycombining the staining patterns of these monoclonalantibodies. The four kinds of epidermal cells were (1)the outer epidermal cells that contained small mucusgranules, (2) the ciliated epidermal cells, (3) the outerepidermal cells that contained large mucus granulesand (4) the inner sensorial cells.

Expression and localization of epidermal antigensduring and after metamorphosisTransverse sections of the larval body from stage 46to stage 49 and those of the dorsal skin from stage 52

to stage 66 were stained with the antibodies describedabove. At stage 46, the epidermis still comprises twoepidermal layers. The number of epidermal layersincreases during metamorphosis. At stage 66, theepidermis comprises more than five layers.

XEPI-1The XEPI-1 antigen was detected in the outermostlayer of the epidermis until stage 61 (Fig. 7A-D). Inall larvae between stage 52 and 56, the expression ofthe XEPI-1 antigen was suppressed in some of thecells in the outermost layer (Fig. 7B). At stage 62, theoutermost epidermal layer that expressed the XEPI-1antigen, was being shed (Fig. 7E). At stage 64, therewas no binding of the XEPI-1 antibody to theepidermis (Fig. 7F).

XEPI-2From stages 46 to 49, both the outer epithelial and theinner sensorial layer were stained with the XEPI-2antibody. The latter was stained more heavily thanthe former (Fig. 8B). After stage 49, the number ofepidermal layers increased. All the layers continuedto express XEPI-2 up to stage 60 (Fig. 8C). At stage58, the primordia of the multicellular granular andmucus glands were seen beneath the epidermis

K. Itoh, A. Yamashita and H. Y. Kubota

Fig. 3. Expression of the XEPI-3 antigen in Xenopus embryos. Transverse polyester wax sections were stained withXEPI-3 and FITC-conjugated rabbit antimouse IgG. (A) Stage 16; (B) stage 22; (C) stage 24; (D,E) stage 36. One ofthe ciliated epidermal cells is indicated by the arrowhead. Bar, 100 fim.

Fig. 4. Regional distribution of antigens recognized by XEPI-4 in Xenopus embryos. Transverse polyester wax sectionswere stained with XEPI-4 and FITC-conjugated rabbit antimouse IgG. (A) Stage 28; (B,C) stage 36. Bar, KJO/zm.

(Fig. 8C). The cells that constituted these glands werealso stained. Thereafter the staining became weaker,while it persisted in the cells that faced the epidermallayers up to stage 64 (Fig. 8C-E). From stage 62, the

outer layers of epidermis gradually lost the XEPI-2antigen (Fig. 8D-F). At stage 66, only the innermostlayer of epidermis could be stained (Fig. 8F).

Among the five monoclonal antibodies tested in the

Epidermal antigens in Xenopus

Fig. 5. Regional distribution of the antigens that were recognized by XEPI-5 antibody in Xenopus embryos. Transversepolyester wax sections (A-E) were stained with XEPI-5 and FFTC-conjugated rabbit antimouse IgG. (A) Stage 8;(B) stage 13; (C) stage 20; (D) stage 28; (E) stage 40. (F) Ultrastructural localization of the XEPI-5 antigen at stage 41in a section of epidermal cells that were fixed with 2 % paraformaldehyde and embedded in LR gold. (A-E) Bar,100urn. One of the ciliated epidermal cells is indicated by an arrowhead, bl, blastocoel. Bar,

present investigation, only the XEPI-2 antibodystained the adult epidermis.

XEPI-3The antigen that was recognized by the XEPI-3antibody situated on the apical side of the outerepithelial layer continued to be expressed but thestaining was very weak (Fig. 9A-D). After thesloughing of the outer epidermis, the XEPI-3 antigencould no longer be detected (Fig. 9E).

The large mucus granules contained in the outerepithelial layer were also stained with the XEPI-3antibody in stage-36, -40, -46 and -48 larvae(Fig. 9A). At stage 50, vacuolation began in the cellsthat were located between the outermost layer andthe basal layer of the epidermis. Within the vacuo-lated cells, large mucus granules were stained withthe XEPI-3 antibody (Fig. 9B) and the number ofstained granules increased up to stage 60 (Fig. 9C,D).The stained vacuolated cells were located beneath theoutermost cell layer. At the time that the larval skinwas shed the stained vacuolated cells disappeared(Fig. 9D,E).

XEPl-4The large mucus granules that were stained with theXEPI-4 antibody (Fig. 10) were the same as thosestained with XEPI-3 (Fig. 9).

Also, the XEPI-4 antibody was bound to thebasement membrane during and after metamorphosis(Fig. 10).

XEPI-5The antigen that was identified by the XEPI-5 anti-body on the apical side of the outer epithelial layerdisappeared after stage 40 (Fig. 11A-C). However,the XEPI-5 antibody was bound transiently to someof the outermost epidermal cells (Fig. 11D). Afterthe sloughing of the outer epidermal layers, there wasno binding of the XEPI-5 antibody to the epidermis(Fig. 11E.F).

Molecular aspects of epidermal antigensThe result of Western blotting analysis of the solubleproteins that were obtained from stage-35/-36 larvaeshowed that the XEPI-1 antibody recognized onemajor protein band having an estimated relativemolecular mass of 250X103 (Fig. 12B). The soluble

K. hoh, A. Yamashita and H. Y. Kubota

Fig. 6. Comparison of distribution of antigens recognized by five antibodies at stage 36. Transverse polyester waxsections were stained with XEPI-1 (A), -2 (B), -3 (C), -4 (D). or -5 (E) and FITC-conjugated antimouse IgG. C and Dwere serial sections. Two of the ciliated epidermal cells are indicated by arrowheads. Bar, l(X)jum.

proteins extracted from stage-54 larval skin were usedfor Western blotting analysis using the XEPI-2 anti-body. A total of six protein bands having estimatedMrs between 45 and 67X103 were resolved by theXEPI-2 antibody (Fig. 12D). As for the other threeantibodies, we could not detect any bands by theWestern blotting technique. The reason for this maybe (1) these antigens are not proteins but glycolipidsor other molecules, (2) the antigens could not bedetected using Western blots because their anti-genicity was altered during electrophoresis or by

blotting technique or (3) the antigens could not besolubilized using the sample buffer we employed.

In order to determine whether the five monoclonalantibodies recognized peanut lectin receptors (seeSlack, 1985), sections of stage-36 embryos werepreincubated with peanut lectin prior to being stainedwith monoclonal antibodies. Peanut lectin blockedthe binding of only the XEPI-4 antibody to the largemucus granules, while the binding of the XEPI-4antibody to the basement membrane was not blocked(data not shown). This suggests that the XEPI-4

Epidermal antigens in Xenopus

Fig. 7. Expression of the XEP1-1 antigen before and during metamorphosis. Transverse polyester wax sections werestained as described in Fig. 1. B-F were dorsal epidermal sections. (A) Stage 40; (B) stage 54; (C) stage 58; (D) stage60; (E) stage 62; (F) stage 64. Bar, 100[im.

antigen in the large mucus granules is the same as oneof the peanut lectin receptors reported by Slack(1985).

Neuraminidase treatment was carried out prior tostaining sections with monoclonal antibodies in orderto determine whether there were any antigensmasked with sialic acid that could be recognized bythe five antibodies (see Slack, 1985). The results showthat none of the antigens were masked with sialic acidthat could be recognized by these antibodies.

Discussion

Epidermal cell types in the embryonic epidermisWe have obtained five monoclonal antibodies againstepidermal antigens. The spatial and temporal ex-pressions of these antigens in normal embryos showthat at least four types of epidermal cells exist in thelarval epidermis. The four types of epidermal cells are(1) the outer epidermal cells that contain small mucusgranules, (2) the ciliated epidermal cells, (3) the outerepidermal cells that contain large mucus granules and(4) the inner sensorial cells. Among these cell types,the cells that contained small mucus granules wereidentified by the XEPI-5 antibody (Fig. 5). Theappearance of the XEPI-5 antigen (stage 13) was

slightly earlier than that of the small mucus granulesthat were observed at the ultrastructural level (stage14) (Billett & Gould, 1971). The ciliated epidermalcells were detected by the XEPI-1 antibody and bythe interruption in the binding of the XEPI-3 and theXEPI-5 antibodies that stain the outer epithelial layer(Fig. 6). The cells that contained large mucus gran-ules were recognized by the XEPI-3 antibody and theXEPI-4 antibody (Fig. 6). Since their temporal pat-terns of staining were slightly different and thebinding of only the XEPI-4 antibody to the largemucus granules could be blocked by peanut lectin, itmay be concluded that the XEPI-3 and XEPI-4antibodies recognize different antigens contained inthe same granules. These cells that contained largemucus granules appeared in the outer epithelial layerat the tail-bud stage and these granules became largeras development proceeded (Fig. 4). The stainingpattern of the granules recognized by the XEPI-3 andXEPI-4 antibodies at later stages suggests that thesegranules may be secreted (Figs 3, 4). The innerepidermal cells were recognized by the XEPI-2 anti-body, but not by any of the other antibodies at anystages of development (Fig. 6). Among the fiveantibodies tested, the XEPI-2 antibody is the mostuseful tool for investigating the differentiation of the

10 K. Itoh, A. Yamashita and H. Y. Kubota

Fig. 8. Expression of the XEPI-2 antigen before, during and after metamorphosis. Transverse polyester wax sectionswere stained as described in Fig. 2. C-F were dorsal epidermal sections. (A) Stage 40; (B) stage 46; (C) stage 58;(D) stage 62; (E) stage 64; (F) stage 66. Bar, 100nm.

epidermis in early Xenopus development because thisantibody recognizes all four types of epidermal cells.The monoclonal antibodies described in the presentinvestigation were used to study the differentiation ofthe four types of epidermal cells in isolated andexplanted outer and inner ectoderm (Itoh & Kubota,in preparation).

Differentiation of epidermisThe mechanism that is responsible for the differen-tiation of the epidermis can be discussed in terms ofseveral models such as localization of cytoplasmicdeterminants, cell-cell interaction, and induction.Ectodermal explants produced from around the ani-mal pole region of the early gastrulae were shown todifferentiate into epidermal cells and express epider-mal markers (Slack, 1984, 1985). Jones & Woodland(1986) showed that animal pole explants of Xenopuscultured from the 8-cell stage onwards express aspecific epidermal antigen and that embryos that wereeither disaggregated or incubated in cytochalasin Bafter the midblastula stage do not require cell interac-tions, Ca2+ and cell divisions for epidermal differen-tiation to occur. Furthermore, Jones & Woodland(1987) showed that animal cap cells at stage 10 or laterdevelop exclusively into ectodermal tissues even if

they are transplanted into ectopic positions such asthe blastocoel or the vegetal pole of the host em-bryos. These investigations suggest that the animalpole region of the embryo at cleavage stage differen-tiates autonomously into epidermis when cultured inisolation, although cellular interactions and cell div-isions up to the midblastula stage are necessary.Further, the cells constituting the animal pole regionof the embryo at stage 10 or later are determined todevelop into the ectodermal tissues.

Cytokeratin gene transcripts have been detected inanimal pole cells of stage-9 blastulae and in the entireectoderm, including prospective neural area of theearly gastrula (Jamrich etal. 1987). The transcriptionin the neural area was suppressed when in contactwith the involuting chordamesoderm during gastru-lation. Recently, neural differentiation has beeninvestigated using neural-specific markers, N-CAMgene transcripts (Kintner & Melton, 1987) or using apolyclonal antibody that was produced against puri-fied N-CAM (Jacobson & Rutishauser, 1986; Balak etal. 1987). These studies demonstrated that the level ofN-CAM RNA increases during gastrulation when themesoderm comes in contact with the ectoderm andthat N-CAM was first detected at the neural platestage (stage 14/15). Together, these investigations

Epidermal antigens in Xenopus 11

Fig. 9. Expression of the XEPI-3 antigen before and during metamorphosis. Transverse polyester wax sections werestained as described in Fig. 3. C-F were dorsal epidermal sections. (A) Stage 40; (B) stage 52; (C) stage 58; (D) stage60; (E) stage 62; (F) stage 64. mu, muscles. Bar, 100jwn.

show that the cells that constitute the animal poleregion become specified at the late blastula stage toexpress cytokeratin genes in the presumptive epider-mal and neural cells. Furthermore, if the ectoderm isinduced by the underlying mesoderm, transcriptionof the cytokeratin genes is suppressed, whereas theN-CAM genes are activated.

In the present study, four of the epidermal cellmarkers (XEP1-1, 2, 3, and 5) were detected in theepidermal region of the late gastrula (stage 12i) orearly neurula (stage 13), but never in the neuralregion. This result is consistent with those previouslyobtained using other kinds of epidermal markers(Slack, 1985; Jones & Woodland, 1986; Akers et al.1986). Therefore, it is suggested that the epidermis ofthe late gastrula embryo expresses molecular markersthat are characteristic of differentiated cells before itexpresses morphological features that are typical tothe epidermal cells (Billett & Gould, 1971).

Expression of epidermal antigens duringmetamorphosisFrom the feeding tadpole stage and onwards, thenumber of epidermal layers increases from two toapproximately five. In all cases, the XEPI-1 antigen

was expressed in the outermost layer and then disap-peared with the sloughing of the outer epidermallayers before the adult stage (Fig. 7). In contrast, theXEPI-2 antigen was expressed in both layers of theepidermis during early development and in all of theepidermal layers up to stage 60. From stage 60 andlater, the outer layers did not express the XEPI-2antigen. In the adult epidermis, only the innermostlayer expressed the XEPI-2 antigen (Fig. 8). Theseresults suggest that the increase in the number ofepidermal layers was due to proliferation of the innerepidermal cells at earlier stages. The cells that consti-tuted the granular and mucus glands that are locatedunderneath the innermost layer also expressed theXEPI-2 antigen (Fig. 8). This result is consistent withthe description that the gland cells originate from theinner epidermal cells (Nieuwkoop & Faber, 1956).

The vacuolated cells (termed Leydig cells) werelocated between the outermost layer and the innerlayers before metamorphosis and these cells con-tained large mucus granules that were recognized bythe XEPI-3 and XEPI-4 antibodies (Figs 9, 10). Atearlier larval stages, both the XEPI-3 and XEPI-4antibodies recognized the large mucus granules con-tained in some of the cells that constitute the outerepithelial layer (Figs 3,4). However, it remains to be

12 K. Itoh, A. Yamashita and H. Y. Kubota

Fig. 10. Regional distribution of the antigens recognized by the XEPI-4 antibody before and during metamorphosis.Transverse polyester wax sections were stained as described in Fig. 4. C-F were dorsal epidermal sections. (A) Stage40; (B) stage 52; (C) stage 58; (D) stage 60; (E) stage 62; (F) stage 64. mu, muscles. Bar, 100^m.

investigated whether these granules are the same asthose present in the Leydig cells.

Molecular nature of the epidermal antigensIn the present investigation, Western blotting analysisshowed that the XEPI-1 antibody detected one majorprotein band having an estimated Mr of 250X103

(Fig. 12B). Jones (1985) showed, using the method ofimmunoprecipitation, that an epidermis-specificmonoclonal antibody termed 2F7-C7 recognizes onemajor protein band that has an Mr greater than220x lfA The staining patterns of XEPI-1 and 2F7-C7are similar because (1) both antibodies detectedepidermal antigens starting at the late gastrula stage,(2) the antigens recognized by both these antibodieswere restricted to the outer epithelial layer of epider-mis and (3) the cement gland cells were recognized byboth XEPI-1 and 2F7-C7 antibodies. In order todetermine whether the XEPI-1 antibody recognizesthe same antigen as the 2F7-C7 antibody, we com-pared them directly in the same conditions by stainingpolyester wax sections, by Western blotting and byimmunoelectron microscopy. The results of theseexperiments indicate that both antibodies showed thesame staining patterns at light microscopic and ultra-microscopic levels and both antibodies recognized the

same band in Western blotting (data not shown).Therefore, it can be concluded that the antigenrecognized by the XEPI-1 antibody is the same as thatrecognized by the 2F7-C7 antibody. The XEPI-1antibody is different from a monoclonal antibody(Epi-1) that was produced by Akers et al. (1986). ByWestern blot analysis, Epi-1 recognized an epidermalglycoprotein that has an MT of approximately300X103. Furthermore, Epi-1 did not bind to cementgland cells. Although the function of the XEPI-1antigen is not known, it is a component of MEB(moderately electron-dense bodies) and is secreted tothe surface of the embryo (Asada-Kubota, 1988).

Using the Western blotting technique, the XEPI-2antibody recognized a total of six protein bandshaving estimated M,s between 45 and 67X103 fromproteins obtained from stage-54 larval epidermis(Fig. 12D). This pattern is similar to that of keratins(Ellison et al. 1985). The distributions of cytokeratinsin early Xenopus development were investigatedusing polyclonal antibodies that recognized most ofthe type I embryonic cytokeratins (Jamrich et al.1987). The temporal distribution of the type I cyto-keratins is different from that of the XEPI-2 antigens,but both antibodies are alike in that they recognizedthe inner and outer epidermal two layers. These

Epidermal antigens in Xenopus 13

Fig. II. Expression of the XEPI-5 antigen before and during metamorphosis. Transverse polyester wax sections werestained as described in Fig. 5. C-F were sections of dorsal epidermis. (A) Stage 40; (B) stage 46; (C) stage 58; (D) stage60; (E) stage 62; (F) stage 64. Bar, 100/im.

HX 1 0292 _2 2 0 -195-

9 7 -

- 3 A B C D T 3

J

x i (

-220

- 94

- 67

- 47- 43

results suggest that the XEPI-2 antigen may becomposed of cytokeratins.

The XEPI-3 and XEPI-4 antigens could not bedetected using the method of Western blotting. How-ever, peanut lectin blocking studies showed that theXEPI-4 antigen in the large mucus granules is one ofthe peanut lectin receptors. Slack (1985, Fig. 4b)showed that these granules as well as the outersurface of epidermis were recognized by FITC-PNA(peanut agglutinin).

Slack (1985) has shown that PNA receptors exist in

Fig. 12. Proteins contained in the homogenate of stage-35/-36 embryos were separated using SDS-PAGE on a10% polyacrylamide gel (lane A). These proteins wereblotted on a nitrocellulose sheet and incubated with theXEPI-1 antibody and 35S-labelled anti-mouse IgG (laneB). The XEPI-1 antibody recognized one major proteinband having an estimated Mr of 250xl03 (lane B).Proteins contained in the homogenate of stage-54epidermal tissues were separated using SDS-PAGE on a4-20% polyacrylamide gradient gel (lane C). Theseproteins were blotted on a nitrocellulose sheet andreacted with the XEPI-2 antibody and 35S-labelledantimouse IgG (lane D). The XEPI-2 antibodyrecognized a total of six protein bands having estimatedJW,S between 45 and 67X103 from epidermal proteins.

14 K. ltoh, A. Yamashita and H. Y. Kubota

the ECM as well as in the epidermis. In the axolotl,the epidermal receptor is a glycoprotein, termedepimucin, and has an estimated Mr of 170xl03.Furthermore, the ECM receptor was shown to con-tain fibronectin and other components. In Xenopus,the epidermal receptor is known to be a highlypolydisperse glycoprotein. The present investigationshowed that the XEPI-5 antibody was bound not onlyto the epidermis but also to the ECM (Fig. 5).However, the binding of the XEPI-5 antibody wasnot blocked by preincubation of peanut lectin andneuraminidase treatment had no effect on the stain-ing pattern of the XEPI-5 antibody. Furthermore, thecement gland cells and the vitelline membrane wererecognized by FITC-PNA but not by the XEPI-5antibody. These results show that the XEPI-5 antigenis most likely different from the peanut lectin recep-tor.

The five monoclonal antibodies that we have de-scribed in this report are being used to investigate thespecification and differentiation of the epidermis.

We wish to thank Dr N. Satoh for his encouragementduring the course of this work. We thank Professor M.Yoneda for his critical reading of this manuscript. We thankDrs H. Fujisawa, S. Takagi, T. Nishikata and I. Mita-Miyazawa for their technical advice for raising monoclonalantibodies. Thanks are also due to Dr M. Asada-Kubota forher kind gift of immunoelectron micrographs. We thank DrE. A. Jones for her kind gift of the 2F7-C7 antibody. Wethank Dr W. R. Bates for his helpful comments on themanuscript.

References

AKERS, R. M., PHILLIPS, C. R. & WESSELLS, N. K.(1986). Expression of an epidermal antigen used tostudy tissue induction in the early Xenopus laevisembryo. Science 231, 613-616.

ASADA-KUBOTA, M. (1988). A monoclonal antibodyspecific for an epidermal cell antigen of Xenopus laevis:electron microscopic observations using a gold-labelingmethod. J. Histochem. Cytochem. 36, 515-552.

BALAK, K., JACOBSON, M., SUNSHINE, J. & RUTISHAUSER,U. (1987). Neural cell adhesion molecule expression inXenopus embryos. Devi Biol. 119, 540-550.

BILLETT, F. S. & GOULD, R. P. (1971). Fine structuralchanges in the differentiating epidermis of Xenopuslaevis embryos. J. Anat. 108, 465-480.

ELLISON, T. R., MATHISEN, P. M. & MILLER, L. (1985).Developmental changes in keratin patterns during

epidermal maturation. Devi Biol. 112, 329-337.Fox, H. (1984). Amphibian Morphogenesis. Clifton: The

HUMANA Press Inc.JACOBSON, M. & RUTISHAUSER, U. (1986). Induction of

neural cell adhesion molecule (NCAM) in Xenopusembryos. Devi Biol. 116, 524-531.

JAMRICH, M., SARGENT, T. D. & DAWID, I. B. (1987).Cell-type-specific expression of epidermal cytokeratingenes during gastrulation of Xenopus laevis. Genes &Development 1, 124-132.

JONES, E. A. (1985). Epidermal development in Xenopuslaevis: the definition of a monoclonal antibody to anepidermal marker. J. Embryol. exp. Morph. 89,Supplement, 155-166.

JONES, E. A. & WOODLAND, H. R. (1986). Developmentof the ectoderm in Xenopus: Tissue specification andthe role of cell association and division. Cell 44,345-355.

JONES, E. A. & WOODLAND, H. R. (1987). Thedevelopment of animal cap cells in Xenopus: the effectsof environment on the differentiation and the migrationof grafted ectodermal cells. Development 101, 23-32.

KELLER, R. E. (1975). Vital dye mapping of the gastrulaand neurula of Xenopus laevis. 1. Prospective areasand morphogenetic movements of the superficial layer.Devi Biol. 42,222-241.

KELLER, R. E. (1980). The cellular basis of epiboly: AnSEM study of deep-cell rearrangement duringgastrulation in Xenopus laevis. J. Embryol. exp.Morph. 60, 201-234.

KiNTNER, C. R. & MELTON, D. A. (1987). Expression ofXenopus N-CAM RNA in ectoderm is an earlyresponse to neural induction. Development 99,311-325.

LITTLEFIELD, J. W. (1964). Selection of hybrids frommatings of fibroblasts in vitro and their presumedrecombinants. Science 145, 709-710.

MASHO, R. & KUBOTA, H. Y. (1986). Developmentalfates of blastomeres of eight-cell-stage Xenopus laevisembryos. Devi Growth Differ. 28, 113-123.

NIEUWKOOP, P. D. & FABER, J. (1956). Normal Table ofXenopus laevis (Daudin). Amsterdam: North-Holland.

SLACK, J. M. W. (1984). In vitro development of isolatedectoderm from axolotl gastrulae. J. Embryol. exp.Morph. 80, 321-330.

SLACK, J. M. W. (1985). Peanut lectin receptors in theearly amphibian embryo: Regional markers for thestudy of embryonic induction. Cell 41, 237-247.

STEEDMAN, H. F. (1957). Polyester wax. A new ribboningembedding medium for histology. Nature, Lond. 179,1345.

(Accepted J5 June 1988)