Development 103, 725-731 (1988)Printed in Great Britain © The Company of Biologists Limited 1988

725

Pepsinogen gene transcription induced in heterologous epithelial—

mesenchymal recombinations of chicken endoderms and glandular

stomach mesenchyme

KENSUKE HAYASHI*, SADAO YASUGIt and TAKEO MIZUNO*

Zoological Institute, Faculty of Science, University of Tokyo, Hongo, Tokyo 113, Japan

* Present address: Division of Cell Biology, National Institute of Neuroscience, National Center of Neurology and Psychiatry, 4-1-1,Ogawa-higashimachi, Kodaira, Tokyo 187, Japant Author for reprint requests$ Present address: Faculty of Pharmaceutical Sciences, Teikyo University, Sagamikomachi, Kanagawa 199-01, Japan

Summary

Proventricular (glandular stomach) mesenchyme ofchicken embryos can induce endoderms of some partsof embryonic digestive tract to produce embryonicchicken pepsinogen (ECPg), a marker protein for thedifferentiation of embryonic proventricular epi-thelium. In the present study, we investigated theproduction of ECPg mRNA in the course of epithelial—mesenchymal interactions between endoderms of di-gestive tract and proventricular mesenchyme. ECPgmRNA was detected by Northern hybridization withECPg cDNA as a probe. In normal development of theproventriculus, ECPg mRNA was first detected at day7 of incubation, and it ceased to be produced by day21. Embryonic esophagus, gizzard and small intestinedid not contain ECPg mRNA. When 6-day esophagus,gizzard or proventricular endoderm was associated

and cultured with 6-day proventricular mesenchyme,the recombinates formed proventricular-gland-likecomplex glands and produced ECPg mRNA in almostequal quantity. However, 6-day small intestinal or 3-5-day allantoic endoderm did not produce pepsinogenmRNA under the same conditions, though the recom-binates formed complex glands. These results indicatethat the proventricular mesenchyme can induce denovo transcription of ECPg gene in esophagus, pro-ventricular and gizzard endoderms, and that ECPggene in small intestinal and allantoic endoderms failsto react to the inducing signal.

Key words: embryonic chicken pepsinogen, digestivetract, epithelial-mesenchymal interaction, geneexpression, transcription, Northern hybridization.

Introduction

Epithelial-mesenchymal interactions play an import-ant role in various kinds of organogenesis of ver-tebrate embryos (reviewed by Wessells, 1977; Kra-tochwil, 1982; Sawyer & Fallon, 1983). Theinteractions between the two tissues often inducemorphogenesis and differentiation of one or bothtissues. For instance, morphogenesis and differen-tiation of avian proventricular epithelium are inducedby the surrounding mesenchyme. The inducing actionof proventricular mesenchyme has been demon-strated in heterologous recombinations. If endo-derms of various parts of embryos, such as esopha-gus, gizzard, small intestine, allantois and yolk sac,

are recombined and cultured with proventricularmesenchyme, morphogenesis characteristic of pro-ventricular epithelium is induced (Sigot, 1971; Yasugi& Mizuno, 1974, 1978; Gumpel-Pinot et al. 1978;Masui, 1981). In some cases, specific cytodifferentia-tion is also demonstrated to be induced. If gizzardendoderm or hypoblast is recombined with proven-tricular mesenchyme, the endoderm produces embry-onic chicken pepsinogen (ECPg), a marker proteinfor cytodifferentiation of the proventricular epi-thelium (Yasugi & Mizuno, 1984; Takiguchi et al.1986). These investigations suggest that proventricu-lar mesenchyme influences the pattern of gene ex-pressions in the endoderm, although the molecularmechanisms working in the process are not yet well

726 K. Hayashi, S. Yasugi and T. Mizuno

understood.The molecular mechanisms involved in cell-type

specific production have been studied intensively. It isknown that cell-type-specific production is regulatednot only at the level of gene transcription, but also atthe levels of mRNA processing, modulation ofmRNA stability, translation of the mRNA and pro-tein processing. It is interesting to investigate atwhich level the production of ECPg is induced inendoderms by the proventricular mesenchyme.

In contrast to the cases in which both heterotypicmorphogenesis and cytodifferentiation are induced,there are several cases in which heterotypic cytodif-ferentiation is not induced in spite of heterotypicmorphogenesis (Sakakura et al. 1976; Lawson, 1972;Tyler & Koch, 1977; Yasugi, 1984; Yasugi etal. 1985).Yasugi (1984) showed that allantoic endoderm im-planted into the presumptive stomach area of chickenembryos fails to synthesize ECPg, even if it isincorporated into the host proventriculus and formsproventricular glands. Endoderm of the small intes-tine is equally unable to synthesize ECPg under theinfluence of proventricular mesenchyme, despite itsproventricular-gland-like morphogenesis (Yasugi etal. 1985).

Although these findings suggest that small intesti-nal and allantoic endoderms cannot react to thepepsinogen-inducing influence of proventricular mes-enchyme, we must examine the possibility that ECPggene is transcribed in these endodermal cells but themRNA cannot be translated. This possibility is notinconceivable, because Agata & Eguchi (1984)showed that a precursor of 6-crystallin mRNA can bedetected in the dedifferentiated chicken pigmentedepithelial cells, but it is not translated. This RNA isprocessed and translated when the cells 'transdiffer-entiate' into lens cells under appropriate cultureconditions (Itoh & Eguchi, 1986).

The purpose of the present study is to examine (1)at what level the expression of ECPg is controlled inthe course of epithelial-mesenchymal interactions,and (2) whether or not the small intestinal or allantoicendoderm can express any ECPg mRNA under theinfluence of the proventricular mesenchyme. Theavailability of a cDNA clone for ECPg (Hayashi et al.1988a) has allowed us to address directly thesequestions.

Materials and methods

Tissue recombination and graftingTissue recombination and grafting methods were the sameas described by Yasugi et al. (1985). Esophagus, proven-triculus, gizzard and small intestine were isolated from 6-day chicken embryos and allantois from 3-5-day embryos.The endoderms were separated from the mesenchyme by

6 dav 3-5 day

ES PV GZ SI AL

I9 day

Fig. 1. Scheme of recombination experiments. Four partsof digestive endoderm of 6-day chicken embryos andallantoic endoderm of 3-5-day embryos were isolated, andassociated with the proventricular mesenchyme of 6-dayembryos. The recombinates were cultured for 9 days onthe chorioallantoic membrane (CAM) of 9-day embryos.ES, esophagus; PV, proventriculus; GZ, gizzard; SI,small intestine; AL, allantois.

B

J—II — | cBg150

3 1 ECP1.1K

Fig. 2. cDNA probes used for Northern blot analysis.The full-length cDNA for ECPg is shown at the top.Open bars indicate the protein-coding region for ECPg.Solid boxes indicate poly(A) tails. ECP1.1K (1110 bp andpoly(A)) contains 87 % nucleotides of full-length cDNA.cBgl50 (110bp and poly (A)) contains 3' untranslatedregion and was used as a ECPg mRNA-specific probe. B,BglU cleavage sites.

collagenase digestion. These endoderms were associatedwith proventricular mesenchyme of 6-day embryos on anagar medium containing horse serum (Wolff & Haffen,1952). After incubation of the recombinates on the agarmedium overnight, they were grafted onto the chorioallan-toic membrane of 9-day chicken embryos and cultured for 9days at 37°C (Fig. 1).

ProbesECPg cDNA was cloned and the nucleotide sequence wasalready determined (Hayashi et al. 1988a). The full-lengthcDNA contains 1281 base pairs. There is a flg/n cleavagesite just at the end of the protein-coding region (Fig. 2).

Tissue interactions induce pepsinogen mRNA 727

The probes used for Northern blot analysis in this studywere a BglW fragment (cBgl50) that contains 110 bp of the3' untranslated region and the poly(A) tail, and a 1-1 kbcDNA clone (ECP1.1K, refer to Hayashi et al. 1988a),which contains about 87 % of the full-length ECPg cDNA.

mRNA extraction and Northern blot analysisSamples were homogenized with a Polytron homogenizerand total nucleic acids were extracted with phenol-chloro-form. Poly(A)+ RNA was isolated by oligo(dT)-cellulosechromatography according to Aviv & Leder (1972). RNApreparations were denatured in glyoxal, subjected to elec-trophoresis in 1 % agarose gels and blotted on a nylonmembrane (Pall) in 20x SSC (lx SSCcontains0-15 M-NaCland 0-015 M-sodium citrate). Radiolabelling of the probeswas performed by nick translation (Nick Translation Kit,Amersham) with a-[32P]dCTP. When cBgl50 was used as aprobe, the blots were hybridized in 6x SSC, 6x Denhardt'ssolution (1 x Denhardt's solution contains 1 % bovineserum albumin, 1 % polyvinylpyrrolidone and 1 % Ficoll),0 1 % SDS (sodium dodecyl sulphate), lOO^gml"1 de-natured salmon sperm DNA at 68 CC for 18 h, and werewashed twice in 2x SSC, 0-1 % SDS at 37 °C for 15 min, andtwice in 0-lx SSC, 0-1% SDS at 37°C for 30min. WhenECP1.1K was used as a probe, the blots were hybridized inthe same hybridizing solution at 50°C for 18h, and werewashed twice in 2x SSC, 0-1% SDS at 37°C for 15 min.Rehybridization was performed after washing the mem-brane in 5mM-Tris-HCl (pH8-0), 0-2mM-EDTA, 0-05%sodium pyrophosphate, and 0-002x Denhardt's solution at65°C for longer than 2h. Hybrids were detected by auto-radiography with an intensifying screen at —80°C.

Histological techniquesSamples were fixed in ice-cold 95 % ethanol for 4 h andembedded in paraffin. 5^m sections were examined byindirect immunofluorescence or by staining with PAS-hae-matoxylin. Indirect immunofluorescence was performedaccording to Sainte-Marie (1962) with monoclonal antibodyagainst ECPg (Y37; Yasugi et al. 1987) and the anti-mouse

IgG antiserum conjugated with fluorescein isothiocyanate(Miles). The fluorescence was observed with an Olympusepifluorescent microscope.

Results

ECPg mRNA in the normal development of thedigestive tract

First we investigated the time course of ECPg mRNAproduction during normal development of the pro-ventriculus. Poly(A)+ RNAs extracted from chickenproventriculi of various stages were electrophoresedin an agarose gel and blotted on a nylon membrane.The membrane was hybridized to cBgl50 (see Fig. 2).Autoradiograms showed that ECPg mRNA was firstdetectable at day 7 (Fig. 3A), a day after the proven-tricular glands appear. The most abundant transcriptwas observed at day 15 (Fig. 3B; Note that panel Awas exposed five times longer than panel B). Thequantity of mRNA decreased rapidly after day 18.Adult proventriculus possessed no detectable ECPgmRNA. This time course of ECPg mRNA productionmatched well with the change of peptic activity inchicken embryonic proventriculus, which was shownto be highest at day 15 and decreased at day 20(Yasugi & Mizuno, 1981), suggesting that ECPgmRNA is translated and degraded within not so longa period. It was also shown that ECPg mRNA was1-6 kb in size throughout the embryonic stages inwhich it existed.

The membrane was then rehybridized to ECP1.1K,which contains a large part of the protein-codingregion (see Fig. 2). A sharp cross-hybrid was ob-served at 1-6 kb in the lane of adult proventriculus(Fig. 3C). This may be mRNA for adult chickenpepsinogen, which has a considerable homology in

6 7 8 9 6 7 8 9 12 15 18 21 Ad 18 Ad

28 S-

18S- tB

Fig. 3. The time course of ECPgmRNA expression during the normaldevelopment of chickenproventriculus. The numerals at thetop indicate the days of incubation.1/ig of poly(A)+ RNA extracted fromproventriculi of embryos or adultchicken (lane Ad) were subjected toNorthern blot analysis. Blots werehybridized first with cBgl50 (A,B). InA the membrane was exposed longer(75 h) than in B (15 h exposure). Theblots were then rehybridized toECP1.1K (C). The positions of 18Sand 28S rRNA were indicated.

728 K. Hayashi, S. Yasugi and T. Mizuno

15D 6,P 6G 15G AG 15E 15S

Fig. 4. Tissue specificity of ECPg mRNA expression inchicken digestive tract. \y.g of poly(A)+ RNA extractedfrom 15-day proventriculus (15P), 6-day gizzard (6G), 15-day gizzard (15G), adult gizzard (AG), 15-day esophagus(15E) and 15-day small intestine (15S) were subjected toNorthern blot analysis with cBgl50 as a probe. The spotseen at the lane 15G is an artifact.

amino acid sequence with ECPg (Hayashi et al.1988a). This result indicates that the failure to detectECPg mRNA in adult proventriculus with cBgl50was not due to degradation of the RNA preparationof adult proventriculi. The result also indicates thatcBgl50 hybridizes to ECPg mRNA specifically.

Second, ECPg mRNA expression in the other partsof the digestive tract of chicken embryos was investi-gated. Poly(A)+ RNAs from 6-day, 15-day and adultgizzard, 15-day esophagus and 15-day small intestinewere subjected to Northern blot analysis. Theseendoderms did not exhibit any hybrids with cBgl50(Fig. 4)

ECPg mRNA in endoderms associated and culturedwith proventricular mesenchymeEndoderms of esophagus, proventriculus, gizzardand small intestine of 6-day chicken embryos andallantoic endoderm of 3-5-day embryos were isolatedand associated with proventricular mesenchyme of 6-day embryos. These recombinates were grafted on

the chorioallantoic membrane of 9-day chicken em-bryos and cultured for 9 days (Fig. 1).

In all kinds of recombinations, the endodermsformed compound glands characteristic of the pro-ventriculus. Indirect immunofluorescence with themonoclonal antibody against ECPg revealed the pres-ence of ECPg in the glandular cells developed fromendoderms of esophagus, proventriculus and gizzard(Fig. 5A-C). On the contrary, endoderms of smallintestine and allantois which formed complex glandsunder the influence of proventricular mesenchyme(Fig. 5D,F) did not synthesize any detectable ECPg(Fig. 5E,G).

We investigated the expression of ECPg mRNA inthese recombinates. More than 25 recombinates ofeach combination were pooled and poly(A)+ RNAsextracted from them were subjected to Northern blotanalysis. When cBgl50 was used as a probe, endo-derms of esophagus, proventriculus and gizzard as-sociated and cultured with proventricular mesen-chyme exhibited distinct bands at 1-6 kb, which wasthe same size as the ECPg mRNA expressed in thenormal proventriculus (Fig. 6). The intensity of thesebands was measured with a densitometer and theamount of ECPg mRNA per ng of poly(A)+ RNAtranscribed in each of the three recombinates wasestimated as about one sixth of the amount of thattranscribed in 15-day proventriculus. In contrast,recombinates of intestinal and allantoic endodermsexhibited no RNA species that hybridized to cBgl50.No signals appeared even when 10/ig of poly(A)+

RNA of each recombinate were applied (data notshown). To examine the possibility that pepsinogenmRNA other than that for ECPg was transcribed inthese recombinates, we rehybridized the membranewith ECP1.1K, a probe that hybridized to adultpepsinogen mRNAs. No hybrids appeared on thelanes of recombinates of small intestinal or allantoicendoderm with proventricular mesenchyme (data notshown).

Discussion

Many studies have revealed that tissues can beinduced to differentiate heterotypically and expressnew products in the course of epithelial-mesenchy-mal interactions (Sawyer & Fallon, 1983). However,it is still unclear whether the transcription of newgenes is indeed induced. A report by Snead et al.(1984) suggested that de novo transcription of amelo-genin gene occurs as a result of ectomesenchymalinteraction in normal organogenesis of mouse molartooth. However, they did not prove it by recombi-nation experiments. As to the mesodermal inductionin very early amphibian development, Gurdon et al.

Tissue interactions induce pepsinogen mRNA 729

Fig. 5. Histology and immunofluorescence for ECPg of recombinates. Endoderms of esophagus (A), proventriculus(B), gizzard (C), small intestine (D,E) and allantois (F,G) were associated with the proventricular mesenchyme andcultured. D and F show the histology by PAS—haematoxylin staining. The others show the immunofluorescence withanti-ECPg monoclonal antibody. Bar,

may be responsible for proventriculus-specific ex-pression or the induction of ECPg gene.

Endoderms of small intestine and allantois werenot induced to produce ECPg mRNA. This resultexcludes untranslatable ECPg mRNA as a cause forthe failure to produce the enzyme (See Introduction).Furthermore, no hybrids were detected in theserecombinates even with ECP1.1K as a probe (datanot shown), which cross-hybridized with adult-typepepsinogen mRNA (Fig. 3C). This suggests thatthese recombinates did not possess mRNA for anyisoforms of pepsinogen, whereas they showed pro-ventricular-gland-like morphogenesis. Thus, dis-sociation of morphogenesis and cytodifferentiation atthe transcriptional level was demonstrated. Morpho-genesis of proventricular gland was studied pre-viously (Hayashi, 1987), and elevated mitotic activityat the tip of the elongating glands was observed. Theglandular cell proliferation is supposed to be indis-pensable for proventricular gland elongation into themesenchyme. Although the mechanisms involved inthe glandular cell proliferation are unknown, the cellproliferation may not be coregulated with pepsinogengene activation.

6-day endoderms of esophagus and gizzard couldbe induced to express ECPg mRNA in similaramounts to proventricular endoderm, indicating thatthe developmental fate of these endoderms on day 6

(1985) demonstrated that transcription of cardiacactin gene is induced in the conjugates of animal andvegetal cells. In the present study, we demonstratedthat proventricular mesenchyme can induce de novotranscription of ECPg mRNA in some heterologousendoderms.

Our data indicated that in normal development 6-day proventricular and gizzard endoderms do notproduce detectable ECPg mRNA. The recombi-nation experiments showed that 6-day endoderms ofesophagus, proventriculus and gizzard produce ECPgmRNA in almost the same quantity (per figpoly(A)+RNA) when combined and cultured withproventricular mesenchyme. The induction of ECPgsynthesis in gizzard endoderm by proventricular mes-enchyme was previously reported by Takiguchi et al.(1986, 1988), who detected ECPg by immunohisto-chemistry and enzymic assays. The present findingsindicate the induction of de novo transcription of theECPg gene by proventricular mesenchyme. How-ever, the molecular process taking place during thetranscriptional induction of ECPg gene is unknown.To elucidate this process, we have to study factorsinvolved in the ECPg gene regulation. Recently weisolated an ECPg genomic clone and found twoenhancer core consensus sequences within 400 bpupstream of the transcription start site (Hayashi et al.19886). It is possible that the sequences around them

730 K. Hayashi, S. Yasugi and T. Mizuno

15 P E/P P/P G/P S/P A/P

Fig. 6. Northern blot analysis for ECPg mRNA ofrecombinates. Endoderms of the esophagus (E/P),proventricular (P/P), gizzard (G/P), small intestine (S/P)and allantois (A/P) were associated and cultured with theproventricular mesenchyme. 1 ^g of poly(A)+ RNAextracted from each recombinate was subjected toNorthern blot analysis with cBglSO as a probe. 1/igpoly(A)+ RNA of 15-day proventriculus was also loadedon far left lane (I5P) to indicate the normal size andquantity of ECPg mRNA.

were not yet determined as regards the ECPg ex-pression. In contrast, the developmental fate of 6-daysmall intestinal endoderm seems to be already deter-mined. It was previously shown that 6-day smallintestinal endoderm has an intensive tendency todifferentiate into small intestinal epithelium evenunder the influence of mesenchymes from other partsof the digestive tract (Yasugi & Mizuno, 1978). Thisdetermination might occur in very early stages ofdevelopment. It has been shown by mesenchyme-freeculture that some endodermal cells at stage 4 ofHamburger & Hamilton (1951) can selfdifferentiateinto small intestinal epithelium (Sumiya & Mizuno,1976; Mizuno & Sumiya, 1977; Ishizuya-Oka, 1983),although it has not yet been tested whether theseyoung cells can differentiate into small intestinal cellsunder the influence of heterologous mesenchyme. Asto the allantoic endoderm, Matsushita (1984)reported that 3-5-day allantoic endoderm differen-

tiates into the small intestinal and/or cloacal epi-thelium under the presence of any kind of digestive-tract mesenchyme.

Why did the endoderms of the small intestine andallantois fail to react to the ECPg-inducing action ofthe proventricular mesenchyme, while the endo-derms of organs in the anterior part of the digestivetract (esophagus, proventriculus and gizzard)reacted? One possible answer is that transcription ofthe ECPg genes in the small intestinal and allantoicendodermal cells is blocked by methylation or unde-methylation of DNA (Keshet et al. 1985), or byformation of inactive chromatin structure of ECPggene (Wu & Gilbert, 1981). It may also be possiblethat some factors taking part in the process of theinduction of ECPg expression, such as pre-existingtrans-acting transcription factors (Ranjan & Balti-more, 1986) that activate ECPg gene after they areactivated, or receptor molecules that receive inducingsubstances at the surface of the reacting cells (Takataet al. 1981), are missing in small intestinal andallantoic endodermal cells. The mechanisms respon-sible for the inability of the small intestinal andallantoic endoderms to be induced by the proven-tricular mesenchyme may be profoundly correlatedwith the nature of the determination of the develop-mental fate of these endoderms.

We are grateful to Professor G. Eguchi, Doctor K. Agataand Doctor M. Mochii of National Institute for BasicBiology for their technical advice.

References

AGATA, K. & EGUCHI, G. (1984). <5-crystallin gene istranscribed in multipotent dedifferentiated pigmentedepithelial cells. Dev. Growth and Differ. 26, 385.

Avrv, H. & LEDER, P. (1972). Purification of biologicallyactive globin messenger RNA by chromatography onoligothymidylic acid-cellulose. Proc. natn. Acad. Sci.U.S.A. 69, 1408-1412.

GURDON, J. B., FAIRMAN, S., MOHUN, T. J. & BRENNAN,S. (1985). Activation of muscle-specific actin genes inXenopus development by an induction between animaland vegetal cells of a blastula. Cell 41, 913-922.

GUMPEL-PINOT, M., YASUGI, S. & MIZUNO, T. (1978).Differentiation d'6pith61iums endodermiques associ£sau m6soderme splanchnique. C.r. hebd. Sianc. Acad.Sci. Paris 286, 117-120.

HAMBURGER, V. & HAMILTON, H. L. (1951). A series ofnormal stages in the development of the chick embryo.J. Morph. 88, 49-92.

HAYASHI, K. (1987). Thinning of the basement membraneand localized cell proliferation during gland formationof the chick proventriculus. Dev. Growth and Differ.29, 285-295.

HAYASHI, K., AGATA, K., MOCHII, M., YASUGI, S.,EGUCHI, G. & MIZUNO, T. (1988a). Molecular cloning

Tissue interactions induce pepsinogen mRNA 731

and the nucleotide sequence of cDNA for embryonicchicken pepsinogen: Phylogenetic relationship withprochymosin. J. Biochem. 103, 290-296.

HAYASHI, K., YASUGI, S. & MIZUNO, T. (19886).Isolation and structural analysis of embryonic chickenpepsinogen gene: Avian homologue of prochymosingene. Biochem. Biophys. Res. Commun. in press.

ISHIZUYA-OKA, A. (1983). Electron microscopical study ofself-differentiation potency in the chick embryonicendoderm cultured in vitro. Wilhelm Roux's Arch, devlBiol. 192, 171-178.

ITOH, Y. & EGUCHI, G. (1986). In vitro analysis ofcellular metaplasia from pigmented epithelial cells tolens phenotypes: a unique model system for studyingcellular and molecular mechanisms of"transdifferentiation". Devi Biol. 115, 353-362.

KESHET, I., YISRAELI, J. & CEDAR, H. (1985). Effect ofregional DNA methylation on gene expression. Proc.natn. Acad. Sci. U.S.A. 82, 2560-2564.

KRATOCHWIL, K. (1982). Embryonic induction. In CellInteractions and Development; Molecular Mechanisms.pp. 99-122. New York: John Wiley & Sons.

LAWSON, K. A. (1972). The role of mesenchyme in themorphogenesis and functional differentiation of ratsalivary epithelium. J. Embryol. exp. Morph. 27,497-513.

MASUI, T. (1981). Differentiation of the yolk-sacendoderm under the influence of the digestive-tractmesenchyme. J. Embryol. exp. Morph. 62, 277-289.

MATSUSHITA, S. (1984). Appearance of brush-borderantigens and sucrase in the allantoic endoderm culturedin recombination with digestive-tract mesenchymes.Willhelm Roux's Arch, devl Biol. 193, 211-218.

MIZUMO, T. & SuMrYA, M. (1977). Differentiation latentede l'endoderme de jeune embryon de Poulet. C.r.hebd. Sianc. Acad. Sci., Paris 284, 473-476.

RANJAN, S. & BALTIMORE, D. (1986). Inducibility ofArlmmunoglobulin enhancer-binding protein NF-R:B bya posttranslational mechanism. Cell 47, 921-928.

SAINTE-MARIE, G. (1962). A paraffin embeddingtechnique for studies employing immunofluorescence.J. Histochem. Cytochem. 10, 250-256.

SAKAKURA, T., NISHIZUKA, Y. & DAWE, C. J. (1976).Mesenchyme-dependent morphogenesis andepithelium-specific cytodifferentiation of mousemammary gland. Science 194, 1439-1441.

SAWYER, R. H. & FALLON, J. F. (1983). EpithelialMesenchymal Interactions in Development. New York:Praeger Publishers.

SIGOT, M. (1971). Organogenese de l'estomac del'embryon de Poulet, Analyse des m6canismes de ladifferentiation. Archs Anat. microsc. Morph. exp. 60,169-204.

SNEAD, M. L., BRINGAS, JR.- P., BESSEM, C. & SLAVXIN,H. C. (1984). De novo gene expression detected byamelogenin gene transcript analysis. Devl Biol. 104,255-258.

SUMIYA, M. & MIZUNO, T. (1976). Development of theautodifferentiation potencies in the endoderm of thechick embryo. Proc. Japan Acad. 52, 587-590.

TAKATA, K., YAMAMOTO, K. Y. & OZAWA, R. (1981). Use

of lectins as probes for analyzing embryonic induction.Wilhelm Roux's Arch, devl Biol. 190, 92-96.

TAKIGUCHI, K., YASUGI, S. & MIZUNO, T. (1986). Gizzardepithelium of chick embryos can express embryonicpepsinogen antigen, a marker protein ofproventriculus. Wilhelm Roux's Arch, devl Biol. 195,475-483.

TAKIGUCHI, K., YASUGI, S. & MIZUNO, T. (1988).

Developmental changes in the ability to expressembryonic pepsinogen in the stomach epithelia of chickembryos. Wilhelm Roux's Arch, devl Biol. 197, 56-62.

TYLER, M. S. & KOCH, W. E. (1977). In vitrodevelopment of palatal tissues from embryonic mice.in. Interactions between palatal epithelium andheterotypic oral mesenchyme. / . Embryol. exp. Morph.38, 37-48.

WESSELLS, N. K. (1977). Tissue Interactions andDevelopment. Menlo Park: W. A. Benjamin.

WOLFF, ET. & HAFFEN, K. (1953). Sur une m6thode de

culture d'organes embryonnaires "in vitro". Texas Rep.Biol. Med. 10, 463-472.

Wu, C. & GILBERT, W. (1981). Tissue-specific exposureof chromatin structure at the 5' terminus of the ratpreproinsulin II gene. Proc. natn. Acad. Sci. U.S.A.78, 1577-1580.

YASUGI, S. (1984). Differentiation of allantoic endodermimplanted into the presumptive digestive area in avianembryos. A study with organ-specific antigens. /.Embryol. exp. Morph. 80, 137-153.

YASUGI, S., HAYASHI, K., TAKIGUCHI, K., MIZUNO, T.,

MOCHII, M., KODAMA, R., AGATA, K. & EGUCHI, G.

(1987). Immunological relationships among embryonicand adult chicken pepsinogens: A study withmonoclonal and polyclonal antibodies. Dev. Growthand Differ. 29,85-91.

YASUGI, S., MATSUSHITA, S. & MIZUNO, T. (1985). Glandformation induced in the allantoic and small-intestinalendoderm by the proventricular mesenchyme is notcoupled with pepsinogen expression. Differentiation 30,47-52.

YASUGI, S. & MIZUNO, T. (1974). Heterotypicdifferentiation of chick allantoic endoderm under theinfluence of various mesenchymes of the digestive tract.Wilhelm Roux's Arch. EntwMech. Org. 174, 107-116.

YASUGI, S. & MIZUNO, T. (1978). Differentiation of the

digestive tract epithelium under the influence of theheterologous mesenchyme of the digestive tract in thebird embryos. Dev. Growth and Differ. 20, 261-267.

YASUGI, S. & MIZUNO, T. (1981). Developmental changesin add proteases of the avian proventriculus. J. exp.Zool. 216, 331-335.

YASUGI, S. & MIZUNO, T. (1984). Differentiation

he'teiotypique de Phypoblaste d'Oiseau sous l'influencedu mdsenchyme proventriculaire. C.r. Sianc. Soc. Biol.178, 580-583.

{Accepted 22 April 1988)