Abstract
A Xenopus laevis mRNA encoding a cytokeratin of the basic (type II) subfamily that is expressed in postgastrulation embryos was cDNA-cloned and sequenced. Comparison of the deduced amino acid sequence of this polypeptide (513 residues, calculated mol. wt 55454; Mr ∼ 58 000 on SDS–PAGE) with those of other cytokeratins revealed its relationship to certain type II cytokeratins of the same and other species, but also remarkable differences. Using a subclone representing the 3′-untranslated portion of the 2·4kb mRNA encoding this cytokeratin, designated XenCK55(5development of n blot experiments, we found that it differs from the only other Xenopus type II cytokeratin known, i.e. the simple epithelium-type component XenCKl(8), in that it is absent in unfertilized eggs and pregastrulation embryos. XenCK55(5/6) mRNA was first detected at gastrulation (stage 11) and found to rapidly increase during neurulation and further development. It was also identified in Xenopus laevis cultured kidney epithelial cells of the line A6 and in the adult animal where it is a major polypeptide in the oesophageal mucosa but absent in most other tissues examined. The pattern of XenCK55(5/6) expression during embryonic development was similar to that reported for the type I polypeptides of the ‘XK81 subfamily’ previously reported to be embryo-specific and absent in adult tissues. Therefore, we used a XK81 mRNA probe representing the 3′-untranslated region in Northern blots, SI nuclease and hybrid-selection-translation assays and found the ∼ 1·6kb XK81 mRNA and the resulting protein of Mr∼ 48 000 not only in postgastrula embryos and tadpoles but also in the oesophagus of adult animals. Our results show that both these type II and type I cytokeratins are synthesized only on gastrulation and are very actively produced in early development. However, their synthesis is not restricted to developmental stages but is continued in at least one epithelium of the adult organism. These observations raise doubts on the occurrence of Xenopus cytokeratins that are strictly specific for certain embryonic or larval stages and absent in the adult. They rather suggest that embryonically expressed cytokeratins are also produced in some adult tissues, although in a restricted pattern of tissue and cell type distribution.
Introduction
In studies of the synthesis of intermediate filament (IF) proteins during human and murine development, it has been noted that the expression of the individual cytokeratin genes in the diverse epithelia is developmentally regulated, in position and in time, but that all cytokeratins found in embryonic or fetal tissues are also synthesized in one or several tissues of the adult animals. While both early embryonic epithelia, i.e. endoderm and embryonic ectoderm, contain only IFs formed by cytokeratins of the ‘simple epithelium type’, i.e. cytokeratins 8 and 18 and their equivalents in non-human species (Brûlet et al. 1980; Jackson et al. 1980, 1981; Franke et al. 1982a,c; Lehtonen et al. 1983; Oshima et al. 1983; Regauer et al. 1985; Chisholm & Houliston, 1987), apparently also with low amounts of cytokeratin 19 (see Jackson et al. 1981), the formation of stratified, pseudostratified and complex glandular epithelia is accompanied by the synthesis of other members of the cytokeratin family (see Moll et al. 1982a,b;Dale et al. 1985; Regauer et al. 1985). In summary, all cytokeratins found in mammalian embryos and fetuses have also been detected in one of the epithelia of the adult, unlike expression patterns of other multigene families such as the globins in which specific members are synthesized only in certain embryonal stages (for review, see Collins & Weismann, 1984).
Studies of amphibian embryogenesis have suggested a more complicated situation. In oocytes and early embryos of the African clawed toad, Xenopus laevis, three cytokeratins have been identified that are equivalent to human cytokeratins 8,18 and 19 and are also expressed in several organs of the adult animal such as liver and intestine (Franz et al. 1983; Franz & Franke, 1986; for localization see also Gall et al. 1983; Godsave et al. 1984; Wylie et al. 1986; Klymkowsky et al. 1987). Similarly, advanced larval stages, including tadpoles, express some cytokeratins that are also found in the epidermis of adult animals (Franz & Franke, 1986; Hoffmann & Franz, 1984; Hoffmann et al. 1985; Ellison et al. 1985). In contrast, Dawid and colleagues have described two different groups of cytokeratins, the ‘XK81’ subfamily comprising at least four different genes and the ‘XK70’ subfamily with two genes identified so far, which are expressed on gastrulation, continually synthesized during embryonal and larval development but have not been detected in significant amounts after metamorphosis (Dawid et al. 1985; Jonas et al. 1985; Winkles et al. 1985; Dawid & Sargent, 1986; Miyatani et al. 1986; Sargent et al. 1986; Jamrich et al. 1987). From their mRNA analyses using ‘dot blot’ techniques, these authors have concluded that the genes for cytokeratins of the XK81 and XK70 subfamilies belong to those expressed only during certain developmental stages.
Because of the fundamental importance of the concept of embryonic stage-specific IF proteins in relation to patterns and mechanisms of tissue formation and morphogenesis and, in view of the apparent variance in this respect between mammalian and amphibian embryogenesis, we have examined the expression of cytokeratin genes in Xenopus in greater detail. Therefore, we have cDNA-cloned certain cytokeratin mRNAs that are expressed at defined embryonic stages and examined their expression during subsequent development and in various adult tissues. In our analyses, we have also taken into account observations made in mammals that certain cytokeratins are synthesized only in one or a few cell types (Moll et al. 1982a; Tseng et al. 1982; Quinlan et al. 1985) so that the amounts of a given polypeptide produced in an animal or a given tissue may represent only a minuscule fraction of the total cytokeratins present. Here we describe a novel basic type II cytokeratin of Xenopus laevis expressed at high levels in early embryogenesis, and we show that this protein and cytokeratin XK81, the prototype of one of the embryo-specific cytokeratin groups described by Jonas et al. (1985), are synthesized in at least one tissue of the adult, i.e. the oesophagus.
Materials and methods
Animals and cells
Females of Xenopus laevis were kept as described (Krohne et al. 1981). After injection of human chorion gonadotropin, eggs were stripped and fertilized in vitro; embryos were incubated in 5 % DeBoers medium (Herrmann et al. 1989,a) and staged according to Nieuwkoop & Faber (1967).
Various tissues from adult animals were obtained as described (Franz et al. 1983; Benavente et al. 1985; Franz, 1987; Herrmann et al. 1989a).
For enrichment of oesophageal epithelium, the oesophagus was removed, opened by a longitudinal incision, the mucosal tissue was scraped off and collected in phosphate-buffered saline (PBS).
Conditions for growing XLKE-A6 cell cultures have been given (for refs see Franke et al. 1979; Herrmann et al. 1989a).
Characterization of cytoskeletal proteins
Cytoskeletal proteins were prepared from various adult tissues and whole embryos and analysed by two-dimensional gel electrophoresis as described (Franz et al. 1983; Herrmann et al. 1989a). In addition, in several cases, immunoblotting using a monoclonal antibody of broad specificity and interspecies cross-reactivity (Kspan 1-8.136; Achtstätter et al. 1986; Jahn et al. 1987), and cytokeratinbinding assays on nitrocellulose paper, using 125I-labelled purified rat cytokeratins 8 and 18 (Hatzfeld et al. 1987) were performed.
Isolation of cDNA clones and DNA sequencing
A stage-17 Xenopus cDNA library in λgt10 (kindly provided by Dr D. A. Melton, Harvard University, Cambridge, MA, USA) was screened with a hamster vimentin DNA probe under conditions of reduced stringency (for details see the companion paper by Herrmann et al. 1989a). Two clones with significant cross-hybridization, although less intense than that obtained with the authentic Xenopus vimentin clones, were selected, sequenced and shown, by comparison with the XenCKl(8) sequence (Franz & Franke, 1986), to code for a basic (type II) cytokeratin. The complete sequence was determined according to Sanger et al. (1977). In addition, one strand was sequenced using the chemical modification method (Maxam & Gilbert, 1977).
RNA preparations
Total RNA was obtained from eggs, whole embryos, tadpoles and adult tissues as previously described (Magin et al. 1983). Alternatively, the method of Chirgwin et al. (1979) was used. In brief, the material was homogenized in 5 M-guanidinium thiocyanate (100mm-Tris-HCl, pH7 · 5, 10mm-EDTA, 20mm-dithiothreitol, 1% Sarkosyl) with a Dounce homogenizer and RNA was pelleted by ultracentrifugation (180000g for 17h) through a 6 M-CsC1 gradient. The pelleted RNA was redissolved in 10mm-Tris-HCl (pH7·5, 0·1 mm-EDTA, 0 · 1 % SDS), extracted three times with phenol/chloroform, and finally precipitated in 0·3 m-sodium acetate, followed by addition of 2 vols of ethanol. Poly(A)+RNA was prepared with Hybond mAP paper (Medac, Hamburg, FRG) as described by Werner et al. (1984).
RNA blot analysis
Total RNA or poly(A)+RNA was analysed by electrophoresis on agarose gels after denaturation with glyoxal or on formaldehyde/agarose gels (Herrmann et al. 1989a), followed by RNA blot hybridization as described (Jorcano et al. 1984). For detection of specific mRNAs, the following 32P-labelled probes were used: (i) an antisense RNA prepared from the cDNA clone pKX11/8 (Franz & Franke, 1986) using the Bluescribe system (Stratagene, La Jolla, CA, USA) and [α-32P]UTP; (ii) a random-primed probe of a 800 bp Wind III fragment at the 3′ end of the clone pXenCK55(5/6) using [α-32P]dATP (for details see Results) and (iii) a synthetic 60-mer polynucleotide complementary to a large portion of the 3′ noncoding region (ATCCAACAGAATGCGAAATA AATGCACAAATAAGCAGAAATCTTCCCTGAGATCCTGAAG) of clone pC8128 (Jonas et al. 1985) encoding the Xenopus cytokeratin XK81 (this is the region in which mRNAs of different cytokeratins, even those of the same subfamily, do not reveal sequence homologies). This latter probe was end-labelled with [γ-32P]ATP.
Hybrid selection-translation
In order to isolate mRNA coding for the XK81 polypeptide, the 60-mer polynucleotide of pC8128 was cloned into the Bluescribe vector (Stratagene, La Jolla, CA, USA). For the newly identified cytokeratin, XenCK55(5/6), a 3′ untranslated Wmdlll fragment of the sequenced clone (see Fig. 1) was subcloned into the Bluescript vector (Stratagene, La Jolla, USA), and this construct was used for hybrid selection. 300 μg of total RNA were hybridized to the immobilized probe at 37°C, washed with 2 × SSC, 0 · 2 % SDS at room temperature and with 0 · 1 × SSC, 0 · 2 % SDS (at 37 °C for the polynucleotide and at 65 °C for the cDNA clones). The selected mRNAs were used for translation in vitro in the reticulocyte lysate system and the products obtained were examined by gel electrophoresis (Jorcano et al. 1984).
S1 nuclease protection assay
RNA was also examined by the SI nuclease protection assay (Berk & Sharp, 1977). 5 μg poly(A)+RNA were hybridized with 6×lCHctsmin−1 5′-labelled synthetic polynucleotide in 20 μl of formamide hybridization buffer (80 % formamide, 0·4 M-NaC1, 40mm-Pipes buffer, pH7 · 4, lmm-EDTA) at 36°C for 18 h. Digestion was performed by adding 250 μl SI nuclease buffer (250mm-NaCl, 30mm-sodium acetate, 5mm-ZnCl2, 400i.u.ml−1 SI nuclease (Sigma, St Louis, MO, USA)) and protected nucleotides were recovered by ethanol precipitation with 20 μg yeast tRNA added as carrier. Pellets were taken up in 3 μl formamide sample buffer and radioactivity was determined by Cerenkov counting. 1 μl of each sample was analysed on 8% sequencing gels. To control stability of the probe, samples were processed in parallel except that the SI nuclease was omitted from the SI nuclease buffer. After determination of radioactivity the samples were taken up in formamide sample buffer to 300 cts min−1μl−1.
Microscopy
Cryostat sections through snap-frozen tissues or embryos were processed for immunofluorescence microscopy (Jahn et al. 1987) using monoclonal murine antibodies to cyto-keratins such as KL1 (Viac et al. 1983), lu-5 (Franke et al. 1987), Kspanl-8.136 (Jahn et al. 1987) or the vimentin antibody VIM-3B4 (Herrmann et al. 1989a,b). Alternatively, we used guinea pig antibodies to cytokeratins or vimentin (Franke et al. 1979).
For electron microscopy, small pieces of tissue were fixed with sodium-cacodylate-buffered 2 · 5 % glutaraldehyde and processed as described (Franke et al. 1976).
Results
Isolation of a cDNA clone for a new type II cytokeratin
In the course of screening an embryonic stage-17 Xenopus cDNA library for Xenopus vimentin cDNA clones, we noted two clones that hybridized less intensely to the hamster vimentin cDNA probe used originally and to the identified Xenopus vimentin and desmin cDNAs (for details see Herrmann et al. 1989a,b) and had restriction maps completely different from those of Xenopus vimentin and desmin. In hybrid selection experiments, mRNA was enriched, which directed the synthesis of a polypeptide that, in SDS-PAGE, showed a mobility slightly lower than that of hamster vimentin (data not shown) and which, on two-dimensional gel electrophoresis, appeared as a pair of polypeptide spots slightly more basic than the known simple epithelial equivalent to human cytokeratin 8, i.e. component XenCK1(8) (Franz & Franke, 1986). However, the PstI restriction map of these clones was different from that of XenCKl(8). Sequencing of the four PstI fragments obtained then showed, besides a considerable sequence similarity to XenCKl(8), that both clones coded for the same polypeptide of the basic (type II) cytokeratin subfamily different from all known cytokeratins.
Nucleotide sequence and deduced amino acid sequence
Fig. 1 shows the nucleotide sequence of the cloned mRNA and the amino acid sequence deduced therefrom. The sequence of 2329 nucleotides contains the polyadenylation signal of the corresponding mRNA but the 5′-untranslated region contains only 14 nucleotides. From comparison with the mRNA size, estimated in Northern blots to be 2 · 4 kb (see below), we suspect that approximately 200 nucleotides of the mRNA are not represented in this clone, and we assume that the first methionine codon, which is preceded at —3 with A and followed in +4 with G, thus meeting the requirement for a ribosome initiation site (Kozak, 1986), defines the translation start. The sequence TCACT preceding the start codon is not an ‘ideal’ consensus sequence CCACC, but it conforms to the more general scheme of PyCACPy that seems to provide a sufficient environment for an initiation site in many other mRNAs (Kozak, 1986).
The corresponding amino acid sequence represents a polypeptide of 513 amino acids, including the initial methionine which is probably lost in the mature protein, corresponding to a total molecular weight (mol. wt.) of 55 454 and an estimated pl of 6 · 0. These data have to be compared to the 502 amino acids, mol. wt. 55 688, and pI5 · 8 determined for the only other Xenopus type II cytokeratin sequenced, i.e. the simple epithelial protein XenCKl(8) as described by Franz & Franke (1986).
In Fig. 2, the deduced amino acid sequence is compared with those of some other type II cytokeratins selected because of certain similarities. Since the corresponding in vitro translation product migrates, on SDS–PAGE, with an apparent Mr of 58000 and contains some sequence features characteristic of human cytokeratins 5 and 6 (see below), we designated this polypeptide XenCK55(5/6). In its central tv-helical rod region, it is highly homologous to other type II cytokeratins such as Xenopus component cytokeratin 1(8) and human cytokeratins 5 and 6 (Hanukoglu & Fuchs, 1983; Tyner et al. 1985; Lersch & Fuchs, 1988), except for the short first ‘spacer’ region between coils 1A and IB. Altogether, the degree of amino acid homology of XenCK55(5/6) in the rod domain is 79% with the Xenopus protein XenCKl(8) and approximately 75% with both human cytokeratins 5 and 6. Shortly after the start of the ohelical rod, XenCK55(5/6) shows the FASFI motif characteristic for type II cytokeratins, which is replaced by LASYL in the type I cytokeratins.
Both the head and the tail domains are considerably shorter than in human cytokeratins 5 and 6. Especially, the repeated GGGX tetrapeptides found in the head and tail domains of many other type II cytokeratins, though not in all (see the tails of human and bovine cytokeratins 4, 7 and 8; Glass et al. 1985; Magin et al. 1986; Leube et al. 1986,1988; Sémat et al. 1988), do not exist in XenCK55(5/6). In the head portion, however, motifs occur in tandem, similar to those in bovine epidermal cytokeratin III (M. Blessing & W. W. Franke, unpublished data), the murine equivalent to cytokeratin 4 (Knapp et al. 1986) and XenCKIII (Hoffmann et al. 1985). Both the head and tail domain are very rich in hydroxyamino acids, most strikingly in the tail (32 out of a total of 77 residues), notably its end (12 out of the terminal 20 residues are hydroxyamino acids). Interestingly, neither XenCK55(5/6) nor XenCKl(8) contain any cysteine residue, compared to six in human cytokeratin 6 and at least three in human cytokeratin 5.
Remarkably, in the otherwise rather diverged sequences of the head and tail domains, the Xenopus type II cytokeratin XenCK55(5/6) also shows some sequence similarities with certain epidermal cytokeratins of other species. Three amino acids after the initial methionine a decapeptide sequence (SRQSSFSTRS) occurs which is very similar to the aminoterminus of bovine cytokeratin III (SRQSTVSFRS; see M. Blessing & W. W. Franke, unpublished data). At the carboxy terminus, XenCK55(5/6) contains several motifs that are also present in human cytokeratins 5 and 6 (Fig. 2), and partly also in human cytokeratin I and bovine epidermal cytokeratins III and IV, but are absent in several other type II cytokeratins (Fig. 3). This and the pattern of its expression in adult animals suggest that XenCK55(5/6) is related to human cytokeratins 5 and 6.
Identification of polypeptide XenCK55(5/6)
Because of the high degree of homology between XenCK55(5/6) and XenCKl(8) in large parts of their mRNAs, it was necessary to use, in mRNA hybrid selection experiments, a subclone representing a sequence-divergent portion of the 3′-untranslated region probe. With this specific probe we selected mRNA coding for a polypeptide that, on two-dimensional gel electrophoresis, migrated at a position corresponding to Mr ∼ 58000 and a pH almost identical to that of bovine serum albumin (Fig. 4A–C). In such translation experiments, the same polypeptide was obtained with mRNA from embryonic stage 18 (Fig. 4B) and from adult oesophageal mucosa (Fig. 4C).
Differential expression of the type II cytokeratins XenCKl(8) and XenCK55(5/6) during embryogenesis and in adult animals
When we examined, by Northern blot analysis, the expression of different cytokeratins in unfertilized eggs and in various stages of development, we observed drastically different patterns of mRNA synthesis and accumulation. Fig. 5 presents the mRNA contents of cytokeratin XenCKl(8), i.e. the equivalent to human cytokeratin 8, which is already present in unfertilized eggs and early blastulae, although in relatively low concentrations, and shows an increase at neurulation (e.g. stage 14; Fig. 5, lane 6). Because we had previously shown that this mRNA also occurs in adult tissues such as liver and intestine (Franz & Franke, 1986; Franz, 1987), we used this probe as a general positive cytokeratin expression control in our studies.
The expression of XenCK55(5/6) during embryogenesis, however, showed a different situation (Fig. 6A,B). Unfertilized eggs were completely negative (see Fig. 6B for an autoradiograph after prolonged exposure), and the first positive signal was seen at gastrulation (lane 4, stage 11), followed by a strong increase at neurulation (lane 5, stage 14). Thus, the time course of expression of the type II cytokeratin XenCK55(5/6) appeared to be similar to that of the type I cytokeratins XK70 and XK81 (Dawid et al. 1985; Jonas et al. 1985; Winkles et al. 1985; Dawid & Sargent, 1986: Miyatani et al. 1986).
When the expression of cytokeratin XenCK55(5/6) in adult animals was examined using RNA from various tissues, only oesophageal mucosa was positive whereas ovary, liver, skeletal and cardiac muscle were negative (Fig. 7). Cultured XLKE-A6 cells were also positive but gave a much weaker signal (Fig. 7, lane 6), and an even weaker signal was obtained with RNA from skin on prolonged exposure.
Expression of cytokeratin XK81
Because the pattern of expression of the type II cytokeratin XenCK55(5/6) during development resembled that described for the type I cytokeratin XK81 (Jonas et al. 1985; Dawid & Sargent, 1986; Miyatani et al. 1986), we used a cloned synthetic polynucleotide as a probe for XK81 mRNA in Northern blot experiments. The results obtained confirmed those of Jonas et al. (1985 ; Miyatani et al. 1986) in that this ∼l·6kb mRNA was detected at gastrulation and in postgastrulation stages, including tadpoles (Fig. 8A), but was absent in adult epidermis. However, in contrast to Dawid and colleagues (Jonas et al. 1985; Winkles et al. 1985; Dawid & Sargent, 1986; cf. Sargent et al. 1986) we observed a positive, albeit relatively weak, reaction in an adult tissue, i.e. oesophageal mucosa (Fig. 8A, lane 5). Other internal tissues such as liver (data not shown) and muscle, as well as cultured epithelial cells of line XLKE-A6, were negative (Fig. 8A).
Because of the importance of the identification of a cytokeratin previously believed to be embryospecific, we used a sensitive SI nuclease protection assay to further characterize the mRNA detected in oesophageal epithelium. As shown in Fig. 8B, XK81 mRNA was not detected in embryonal stages prior to stage 9 but was already abundant in the stage 18 (neural groove stage; lane 6). It was also found in tadpoles (Fig. 8B, lane 7) and in lower, but significant, concentrations in the oesophageal epithelium of adult animals (Fig. 8B, lane 8). It was not detected in adult epidermis (lane 9), XLKE-A6 cells (lane 10), ovarian tissue including follicular epithelium (lane 12), or in skeletal muscle tissue (lane 11).
Microscopy
The finding that both the type II cytokeratin XenCK55(5/6) and the type I cytokeratin XK81 mRNA are selectively expressed in the oesophageal epithelium of the adult animal stimulated our interest in the morphology and the cytoskeletal composition of this tissue. Immunofluorescence microscopy of frozen sections of the oesophagus of adult toads showed that the mucosal epithelium of this organ is a complex (‘pseudostratified’) epithelium of mostly columnar cells which are rich in IFs of the cytokeratin type (Fig. 9) but are negative for vimentin and desmin (data not shown; see also Jahn et al. 1987).
Detailed light and electron microscopy (Fouquet, 1987) revealed a remarkable cell-type complexity of this epithelium which histologically differs considerably from the organization of mammalian oesophageal epithelium (Bronn & Hoffmann, 1878). Four major cell types were readily distinguished.
The most abundant cells are the fundamental columnar epithelial cells which are rich in cytokeratin IFs, desmosomes and mitochondria, form numerous intercellular bridges with desmosomes and attached IF bundles (tonofibrils) resembling spinous cell layers of epidermis, as well as ‘hemidesmosomes’. In these cells, the cytokeratin is not exclusively arranged into regular IF bundles but is also found in cytoplasmic aggregates of thinner filaments, which are reminiscent of the spheroidal aggregates transiently formed during mitosis in diverse cell cultures (Franke et al. 19826; Lane et al. 1982) and in certain normal and tumorous tissues (e.g. Brown et al. 1983; Geiger et al. 1984). Such spheroidal aggregates of cytokeratin material seem to be more common in amphibian tissues as they have been reported for larval epidermis (‘figures of Eberth’; Fox, 1986; Fox & Whitear, (1986) and for endothelia (Jahn et al. 1987; Fouquet, (1987).
Mucous cells containing subapical aggregates of secretory vesicles are occasionally met in situations suggestive of apical discharge.
Small, dark staining, basally located, cells, resembling the ‘reserve cells’ of diverse complex mammalian epithelia are seen.
Neuroendocrine cells, which are characterized by ‘dense-core’ as well as ‘empty-looking’ neurotransmitter vesicles, are less frequent and mostly located in basal positions.
Identification of cytokeratins XenCK55(5/6) and XK81 in oesophageal cells
On two-dimensional gel electrophoresis of cytoskeletal proteins from oesophageal mucosa of adult animals, nine major polypeptides were resolved (Fig. 10A), eight of which were positively identified as cytokeratins by immunoblotting (Fig. 10B) and complementary cytokeratin binding in vitro (Fig. 10C). Of these eight oesophageal cytokeratins, four (nos 1–4) showed reactions typical of type II cytokeratins (not shown) whereas the other four reacted in a mode typical of type I cytokeratins (Fig. 10C).
In order to identify unequivocally the oesophageal cytokeratin polypeptides encoded by the XenCK55(5/6) and XK81 mRNAs, we performed hybridization-selection experiments, followed by in vitro translation of the selected mRNAs. The polypeptides obtained after translation of the mRNA selected by the specific 3′-untranslated region of clone pXenCK55(5/6) comigrated with the polypeptides numbered 1 and 2 in the cytoskeletal preparation of oesophageal mucosa (compare Fig. 4B,C). At present we cannot decide whether (i) the oesophageal cells contain only cytokeratin 1, i.e. cytokeratin XenCK55(5/6), as the only genuine product whereas component 2 is a secondary degradation or modification, or (ii) whether these results reflect the presence of two very similar, but not identical, cytokeratins (for ‘microheterogeneity’ or possible allelic differences of cytokeratins see Hoffmann et al. 1985; Tyner & Fuchs, 1986; Franz, 1987).
The mRNA selected by the XK81 probe, i.e. the synthetic 60-mer polynucleotide cloned into Bluescribe vector, was translated into a polypeptide that comigrated with the component designated 6 in the cytoskeletal preparation of oesophageal mucosa (Fig. 11A,B).
Discussion
The various polypeptides of epithelial cytokeratins are differentially expressed in the various epithelial cell types. In mammals, four major cytokeratin categories can be distinguished according to their histological distribution (Moll et al. 1982a; Sun et al. 1985). (i)The ‘simple epithelial cytokeratins’, i.e. components 7, 8, 18 and 19 of the human cytokeratin catalogue, are the only cytokeratins present in certain one-layered epithelia, and at least cytokeratins 8 and 18 can also be expressed in certain cell types or layers of complex and stratified epithelia (Moll et al. 1984; Lane et al. 1985; Bartek et al. 1986; Bosch et al. 1988) and in some non-epithelial tissues (Quinlan et al. 1985; Franke & Moll, 1987; Jahn et al. 1987). (ii) Cytokeratins typical of certain epithelia of high cell complexity, i.e. glandular, ductal, pseudostratified and non-epidermal stratified epithelia as well as the ‘transitional epithelium’ of the urinary tract, include components 4–6, 13–15 and 17 of the human catalogue (Moll et al. 1982a; Banks-Schlegel & Harris, 1983; Achtstätter et al. 1985; Grace et al. 1985; Nagle et al. 1985; Quinlan et al. 1985; Sun et al. 1985; Van Muijen et al. 1986). (iii) Human cytokeratins 1, 2, 5, 9–11 and 16 are typical of the suprabasal differentiation of epidermis (e.g. Fuchs & Green, 1980; Steinert et al. 1985; Sun et al. 1985; Knapp et al. 1986; Roop et al. 1987) and certain other stratified epithelia such as gingiva, vagina, exocervix, penile mucosa and certain parts of the amnion epithelium (Moll et al. 1983; Ouhayoun et al. 1985; Quinlan et al. 1985; Regauer et al. 1985; Lane et al. 1985; Morgan et al. 1987). (iv) Cytokeratins 3 and 12 have so far only been found in corneal epithelium (Moll et al. 1982 a; Schermer et al. 1986).
The specific cytokeratins of these different categories are also characterized by certain features of their amino acid sequences (e.g. Steinert et al. 1985; Leube et al. 1986, 1988; Oshima et al. 1986; Fuchs et al. 1987; Sémat et al. 1988), and certain species differences between orthologous cytokeratins have also been noted (e.g. Franz & Franke, 1986). Our present study on Xenopus laevis cytokeratin XenCK55(5/6) presents the first sequence of a nonmammalian type II cytokeratin typical of complex epithelia, corresponding to the group (ii) of human cytokeratins, i.e. components 4–6. It is difficult, however, to relate precisely this Xenopus cytokeratin to a specific human, bovine or murine cytokeratin of this category. While the XenCK55(5/6) rod domain displays similarly high homologies to the human cytokeratins 4 (Leube et al. 1988), 5 (Lersch & Fuchs, 1988) and 6 (Hanukoglu & Fuchs, 1982; Tyner et al. 1985) as well as to the murine cytokeratin of Mr57000 (Knapp et al. 1986), its head and tail regions differ markedly from all these mammalian counterparts (see Results). Clearly, in the case of cytokeratin XenCK55(5/6) these interspecies differences are greater than those observed between the Xenopus (Franz & Franke, 1986), murine (Sémat et al. 1988) and bovine (Magin et al. 1986) equivalents of human cytokeratin 8 (Leube et al. 1988) which show homologous sequences of considerable lengths in the head and the tail, although the overall homology is lower in these domains than that in the rod piece.
In mammalian embryogenesis, the first tissues formed are polar simple epithelia containing cytokeratins only of category (i), i.e. cytokeratins 8 and 18, sometimes in combination with some cytokeratin 19 (for refs see Introduction). Subsequently in development, some epithelia remain at this level of organization whereas others differentiate into complex pseudostratified or stratified epithelia, concomitant with the advent of cytokeratins of the second category (Moll et al. 1982b; Regauer et al. 1985; Quinlan et al. 1985). Many of the cytokeratins of this category are also found in certain cell layers of fetal epidermis which, however, then produces additional cytokeratins of category (iii) and restricts its cytokeratin pattern to that typical of mature epidermis (e.g., Banks-Schlegel, 1982; Moll et al. 1982b; Dale et al. 1985; Lane et al. 1985). Important in the context of the present study is the fact that all cytokeratins identified in some embryonic or fetal stage also occur in one of the adult tissues.
In contrast, Winkles et al. (1985) have proposed that in Xenopus development three groups of cytokeratins can be classified according to their developmental pattern of synthesis: (i) egg- and embryospecific; (ii) embryo-specific; and (iii) adult-specific. Moreover, Ellison et al. (1985) have shown that the Xenopus epidermis expresses different combinations of cytokeratins in embryonic stages, tadpoles and metamorphosed animals. However, our previous (Franz & Franke, 1986) and present data lead to a different concept of expression of cytokeratin genes during Xenopus development, which is not too dissimilar to the patterns of cytokeratin synthesis during epithelial differentiations in mammalian development. To correspond with the grouping of mammahan cytokeratins according to their expression in histogenesis, we classify the Xenopus cytokeratins in relation to their patterns of developmental appearance into three major categories.
(i) Oocytes, eggs, blastula, gastrula and postgastrulation epithelia all synthesize cytokeratins of the simple epithelium type, i.e. the amphibian equivalents of human cytokeratins 8,18 and 19, and these are continued to be expressed in various simple epithelia of the adult animal (Franz et al. 1983; Franz & Franke, 1986; this study), in endothelium (Jahn et al. 1987; compare with Godsave et al. 1986), retinal pigment epithelium (Owaribe et al. 1988), and certain types of smooth muscle (Jahn et al. 1987).
(ii) During and after gastrulation another category of cytokeratins is newly synthesized in some embryonic epithelia, most prominently in the ectoderm and embryonic epidermis. This category includes the type I cytokeratins of the XK.81 and XK70 ‘subfamilies’ (Jonas et al. 1985; Miyatani et al. 1986) as well as the type II cytokeratin XenCK55(5/6) described in the present study. The latter protein may be related to the type II cytokeratin encoded by clone DG76 mentioned by Dawid & Sargent (1986; see also Jamrich et al. 1987) but a direct comparison is not possible because of the lack of sequence information. The XK81 and XK70 cytokeratins resemble human cytokeratins of the group 14-16 (Hanukoglu & Fuchs, 1982; Tyner et al. 1985; Leube et al. 1988), and Xenopus cytokeratin XenCK55(5/6) shows a relatively close relationship to human cytokeratins 5 and 6. It is probable that cytokeratin XenCK55(5/6) is a natural ‘partner’ of cytokeratins XK81 and XK70, forming the typical type I-type II heterotypic tetramer subunits with each other (Hatzfeld & Franke, 1985; Quinlan et al. 1985; Sun et al. 1985; Fuchs et al. 1987).
Our detection of both XenCK55(5/6) and XK81 in the oesophageal mucosa of the adult animal shows that the genes encoding these proteins are not totally inactivated during - or after - metamorphosis. Rather, their expression is only restricted to certain cell types and tissues. Clearly, these cytokeratins are not ‘embryo-specific’ in general terms. Restriction of expression of XK81 and XK70 cytokeratins to a small subpopulation of cells has already been discussed by Sargent et al. (1986) as one of the possible alternative explanations for the drastic decrease of XK81 mRNA synthesis upon metamorphosis.
(iii) The third cytokeratin category of Xenopus comprises those epidermal cytokeratins that appear later in development, concomitant with epidermal differentiation (for examples, see Franz et al. 1983; Hoffmann & Franz, 1984; Ellison et al. 1985; Hoffmann et al. 1985; Franz & Franke, 1986). Like the corresponding mammalian cytokeratins, many of the Xenopus cytokeratins of this category are also characterized by repeated oligoglycine clusters in the head and tail domains (Hoffmann et al. 1985; Steinert et al. 1985).
We agree with Dawid and colleagues (Dawid & Sargent, 1986; Sargent et al. 1986) that the type I cytokeratins of the XK81 group are among the early expressed IF protein genes showing a rapid increase of synthesis after stage 11 (for later stages such as neurulae see also Slack, 1984), and now we add the type II cytokeratin XenCK55(5/6) to this list. Using in situ hybridization, Jamrich et al. (1987) have detected newly synthesized mRNA for certain cytokeratins of this group in the ectoderm of blastulae of stages 9 and 10. Only a few other zygotic genes are detectably induced in such early stages (e.g. Jonas et al. 1985; Akers et al. 1986; Dawid & Sargent, 1986; Dworkin-Rastl et al. 1986; Gurdon, 1987; Kintner & Melton, 1987; Sharpe et al. 1987). It is hoped that the availability of probes for IF protein mRNAs such as those described in this study will help to elucidate the mechanisms that control these early expression programs of such cell architectural elements and also to identify, in combination with in situ hybridization methods, the cell types in which certain cytokeratins continue to be synthesized in the adult tissues.
ACKNOWLEDGEMENTS
We thank Dr Martina Schnoelzer-Rackwitz for the synthesis of the polynucleotide, Irmgard Purkert for carefully typing the manuscript, and we gratefully acknowledge the expert technical assistance of Monika Brettel and Ralf Zimbelmann. The work has been supported in parts by the Deutsche Forschungsgemeinschaft.