sign in sign up
<<
Home , Goldfish, Keratin 21, Type II keratin, Type I hair keratin, Zebrafish

Cell Tissue Res DOI 10.1007/s00441-005-0031-1

REGULAR ARTICLE

Dana M. García . Hermann Bauer . Thomas Dietz . Thomas Schubert . Jürgen Markl . Michael Schaffeld Identification of and analysis of their expression in and : comparison with the and trout catalog

Received: 21 March 2005 / Accepted: 23 May 2005 # Springer-Verlag

Abstract With more than 50 in human, keratins zebrafish, and the cyprinid being more similar to each make up a large family, but the evolutionary pres- other than to the salmonid trout. Because of the detected sure leading to their diversity remains largely unclear. Nev- similarity of keratin expression among the cyprinid fishes, ertheless, this diversity offers a means to examine the we propose that, for certain experiments, they are inter- evolutionary relationships among organisms that express changeable. Although the zebrafish distinguishes itself as keratins. Here, we report the analysis of keratins expressed being a developmental and genetic/genomic model organ- in two cyprinid fishes, goldfish and carp, by two-dimen- ism, we have found that the goldfish, in particular, is a more sional polyacrylamide gel electrophoresis, complementary suitable model for both biochemical and histological stud- keratin blot binding assay, and immunoblotting. We further ies of the , especially since goldfish cytoskel- explore the expression of keratins by immunofluorescence etal preparations seem to be more resistant to degradation microscopy. Comparison is made with the keratin expres- than those from carp or zebrafish. sion and catalogs of zebrafish and rainbow trout. The keratins among these fishes exhibit a similar range of mo- Keywords Keratin . Intermediate filaments . Goldfish lecular weights and isoelectric points, with a similar overall ( auratus) . Carp ( carpio) . Zebrafish pattern on two-dimensional gels. In addition, immunofluo- ( rerio)(Teleostei) rescence microscopy studies of goldfish and carp tissues have revealed the expression of keratins in both epithelial and mesenchymally derived tissues, as reported previously Introduction for zebrafish and trout. We conclude that keratin expression is qualitatively similar among these fishes, with goldfish (IF) belong to a large and carp patterns being more similar to each other than to multigene family with diverse members that exhibit tis- sue- and developmental-phase-specific expression patterns (Moll et al. 1982; Hoffmann and Franz 1984; Hoffmann et al. 1985; Winkles et al. 1985; Franz and Franke 1986; This work was supported by grants to J.M. from the Stiftung Miyatani et al. 1986; Fouquet et al. 1991; Markl and Rheinland-Pfalz für Innovation (836-386261/138) and the Deutsche Forschungsgemeinschaft (Ma 843/5-1) and a grant to D.G. Franke 1988; Fuchs and Weber 1994; Parry and Steinert from the National Science Foundation (INT-0078261). 1999; Watanabe et al. 2001, 2002; Herrmann et al. 2003). According to their amino acid sequence, six IF D. M. García types are generally distinguished: type I and II (keratins), Department of Biology, Texas State University-San Marcos, San Marcos, TX, 78666, USA type III (, , glial fibrillary acidic protein, e-mail: [email protected] , and plasticin), type IV ( pro- teins, α-, gefiltin, and xefiltin), type V (nuclear H. Bauer ), and the heterogeneous type VI (including , Max-Planck-Institute for Molecular , Ihnestrasse 73, , paranemin, , and tanabin). In mammals, 14195 Berlin, Germany keratins are largely restricted to epithelial cells and tissues, whereas vimentin is primarily expressed in mesenchymal T. Dietz . T. Schubert . J. Markl . M. Schaffeld (*) cells (Franke et al. 1978; Moll et al. 1982). Other IF pro- Institute of Zoology, Johannes Gutenberg University, teins also show tissue specificity. Assigning functions to Johannes-von-Müller-Weg 6, 55099 Mainz, Germany IFs beyond that of structural support has been difficult, in e-mail: [email protected] part because of the size and diversity of the keratin gene e-mail: [email protected] family. Over 50 keratin genes have been identified in humans, representing almost 80% of the IF genes dis- amined groups, including lamprey, shark, bony covered thus far (Hesse et al. 2001, 2004; Rogers et al. , and terrestrial (e.g., Schaffeld et al. 1998, 2004, 2005). 2002a,b). The tissue specificity of expression of the various IF However, for goldfish and carp, only cytoskeletal pro- proteins has been hypothesized to indicate their diversity of teins from the optic nerve and oocytes, respectively, have function. This supposition has been challenged by Owaribe been biochemically characterized in detail. In this study, we et al. (1988) who have observed that the retinal present a further evaluation of the carp and the goldfish, by epithelium of some species expresses keratins, whereas in examining many different organs representing tissues de- others, vimentin is expressed. Since the retinal pigment rived from the ecto-, endo-, and mesoderm. These fish epithelium is presumed to perform similar functions (e.g., species have been chosen because they are closely related metabolic support of the , phagocytosis of photore- to the genetic and developmental , the ceptor discs) in these various species, the idea that IF zebrafish, but are, in several respects, easier to work with diversity reflects the functional diversity of cells and tis- because of their larger size. We find the types of keratins sues seems less likely. This idea has further been called into expressed to be qualitatively similar when examined by question following Markl and Franke’s(1988) discovery two dimensional polyacrylamide gel electrophoresis (2D- that, in trout, mesenchymally derived cells express keratins PAGE), complementary keratin blot-binding (CKBB) as- as a rule and vimentin as an exception to the rule. This say, and immunoblotting. Furthermore, tissues examined finding has since been extended to zebrafish (Cerdá et al. by immunofluorescence microscopy exhibit similar label- 1998; Conrad et al. 1998) and carp (Groff et al. 1997a,b). ing patterns to those previously observed in zebrafish, with Although the hope of assigning functions to IFs based on few exceptions, making goldfish in particular an appealing their tissue-specific expression has dimmed, the possibility organism to represent cyprinid fishes in future studies of that IFs function with respect not only to cell structure, but keratins. also to, for example, cell signaling, is gaining support (Paramio and Jorcano 2002; Kirfel et al. 2003). Never- theless, IF expression patterns may yet contain information Materials and methods regarding the of IFs and the evolutionary relationships among the organisms in whose tissues they and preparation of tissues are expressed. In addition, evolutionary comparisons might elucidate structure-function relationships, e.g., by illustrat- Mature carp (Cyprinus carpio) were obtained from a local ing a role for IFs in adapting to osmotic stress (see fish dealer and killed by a blow to the head. Mature Schaffeld and Markl 2004; Schaffeld et al. 2004, 2005)or goldfish (Carassius auratus) were a gift from Carlos Mora- enabling nerve, , and limb (e.g., Ferretti et Ferrer (University of Mainz) and were killed by spinal al. 1989, 1993; Tsonis et al. 1992; Corcoran and Ferretti section. Both fish types were immediately dissected on 1997, 1999; Markl and Schechter 1998; Martorana et al. ice. Organs to be used in immunofluorescence microscopy 2001). Furthermore, the reduction in the number of keratin were immediately snap-frozen in isopentane that had been genes by loss/decrease of function in some fish lines should pre-cooled in liquid nitrogen to approximately −130°C. be considered as a possibility in evolutionary comparisons Tissues to be used in biochemical preparations were fro- (see Winter et al. 2001 for an example in hominids). With zen directly in liquid nitrogen. All tissues were stored at this in mind, the expression of IFs, particularly keratins and −80°C until use. vimentin, has been evaluated in a number of fishes representing various branches of the evolutionary tree. The fishes most extensively studied include the shark Analytical procedures Scyliorhinus stellaris (Schaffeld et al. 1998, 2001, 2004), the zebrafish Danio rerio (Cerdá et al. 1998; Conrad et al. Procedures for isolating the insoluble cytoskeletal proteins 1998; Schaffeld et al. 2003), the rainbow trout Oncorhyn- and the following two-dimensional polyacrylamide gel cus mykiss (Markl and Franke 1988; Markl et al. 1989; electrophoresis (2D-PAGE), by using isoelectric focusing Herrmann et al. 1996; Schaffeld et al. 2002a,b), the com- (IEF) for the first dimension and sodium dodecylsulfate mon goldfish Carassius auratus (Giordano et al. 1989, (SDS) for the second dimension, were carried out as pre- 1990; Cohen et al. 1991; Druger et al. 1992, 1994; Fuchs et viously described (Schaffeld and Markl 2004). For the al. 1994), the carp Cyprinus carpio (Groff et al. 1997a,b; identification of keratins, polypeptides separated by 2D- Mencarelli and Cotelli 1997), and the lungfish Protopterus PAGE were electrically blot-transferred to nitrocellulose aethiopicus (Alibardi 2001, 2002; Alibardi and Joss 2003; sheets and exposed either to various keratin antibodies Schaffeld et al. 2005). These investigations have shown (immunoblotting) or to biotinylated recombinant human that, in vertebrates generally, two keratin expression sets keratins 8 or 18 for the complementary keratin blot-binding can be distiguished: “E” keratins are expressed in the epi- (CKBB) assay (for experimental details, see Schaffeld and dermis and other stratified epithelia, and “S” keratins are Markl 2004). Molecular masses and isoelectric points for present in internal simple epithelia. The typical “S” keratins keratins were compared statistically by using a Student are represented by the heterodimeric pair K8/K18 that t-test. Values were either calculated from cDNA sequences, moreover seem to be the only keratins present in all ex- estimated from 2D-PAGE gels, or taken from the literature. Table 1 Anti-keratin antibodies and dilutions Two important points must be taken into account: (1) “ ” Antibody Dilution for Dilution for immunofluorescence minor keratin spots ( trace components ) might not be immunoblots microscopy detectable by using a CKBB assay, (2) the diverse fish keratins might bind the recombinant keratins in different GPpolya Not done 1:500 amounts. However, a comparison of the Coomassie-blue- 164.4 Undiluted Undiluted stained gels of the cytoskeletal preparations with their C04 1:200 1:20 corresponding CKBB assays strongly indicates that the 68.4 Undiluted Undiluted different fish keratins each seem to bind comparable LE64 Undiluted Undiluted amounts of their complementary binding partner. There- 26.8 Not done Undiluted fore, we infer that we have not missed any major .3.1 Undiluted Undiluted spot in our experiments. KL1 1:500 1:50 Protein spots representing the type I keratins of the fin ″ 79.14 Undiluted Undiluted tip in goldfish (Fig. 1a,a ) and carp (Fig. 1b) had a size – AE1 1:100 1:100 range of 44 51 kDa and with isoelectric points ranging from 4.9–6.2. Both species possessed a type I protein (“I?”) AE3 Not done 1:100 “ ” “ ” Pan 1–8 1:40 1:4 whose identity as an E or S keratin was uncertain. This protein was not observed in (see Fig. 1c,d) or intestine 69.16 Undiluted Undiluted – ″ 8.60 1:200 1:20 (Fig. 1i i ). It appeared in low abundance in carp (Fig. 1f) but was enriched in goldfish eye (Fig. 1e). 145.2 Undiluted Undiluted The sets of spots identified by CKBB assay as type II Lu5 1:200 1:20 keratins in goldfish and carp fin tip (skin) in each species For the source of monoclonal antibodies (mAbs) 69.16 and 145.2 exhibited a range of apparent molecular weights from and their specificity, see Markl and Schechter (1998); for the other 54 kDa to 58 kDa and pI values ranging from 5.4 to 6.1 (see antibodies, see Schaffeld and Markl (2004) – ′ aGuinea pig anti-trout keratin antibodies; all others are mAbs Fig. 1a a ,b). In carp, degradation products were a prom- inent feature in the CKBB assays from both the fin skin and eye (not shown) and in the Coomassie-blue-stained gel of Immunofluorescence microscopy on 5-μm-thick cryo- the liver preparation (Fig. 1d). Interestingly, the major stat tissue sections was performed as described (Schaffeld degradation products from skin and eye were derived from and Markl 2004). The applied antibodies and their dilutions the type I keratins, whereas those from liver appeared to be are presented in Table 1. MALDI-TOF mass spectrometry mainly derived from K8 (see Fig. 1d). Furthermore, the was performed by Christian Hunzinger (ProteoSys, Mainz) expression pattern of keratins observed in goldfish liver as recently described (Schaffeld et al. 2005). was similar to that observed in intestine (see Fig. 1c–c″,i–i″), although the latter had an abundance of an insoluble protein that was probably desmin. The cytoskeletal prep- Results arations from carp intestine were highly degraded and are not shown. Identification of keratins by 2D-PAGE To obtain a more complete picture with the strong and CKBB assay representation of both “E” and “S” keratins from a single tissue, experiments were performed on cytoskeletal prep- To identify the keratin polypeptides expressed by carp and arations derived from ocular tissue and from goldfish goldfish, cytoskeletal proteins from both species were (Fig. 1e–e″,g–g″) and carp (Fig. 1f,h). Cytoskeleton from extracted from various organs and tissues, including fin tip, forehead skin was also prepared from carp (Fig. 2c). Since forehead skin, liver, eye, gill, and intestine, and separated “S” keratins in fish are expressed in mesenchymal by 2D-PAGE (Figs. 1, 2). Preparations from fin tip (rich in tissues and in simple epithelia, and since eye, gill, and skin) in carp and goldfish revealed a strikingly similar forehead skin contain structural features derived from ec- constellation of protein spots with respect to isoelectric todermal and mesenchymal tissues, a complete representa- points and molecular masses (Fig. 1a,b). Most of these tion of keratins was expected. All of the major spots spots could be identified as keratins by CKBB assays that identified in skin and liver preparations from goldfish furthermore distinguished type I from type II keratins (Fig. 1a,c) and carp (Fig. 1b,d) were also observed in eye (Fig. 1a-a″). We defined the most prominent among these preparations from both goldfish (Fig. 2e–e″) and carp spots as “E” keratins, because they were absent from (Fig. 1f). However, in goldfish gill, two additional type I fractions taken from liver; on the other hand, the typical keratins were observed (Fig. 1g,g″; arrowheads), together liver keratins were defined as “S” keratins (Fig. 1c-c″,d). with an additional type II keratin (Fig. 1g,g′; arrow). In the The “S” keratins were also observed in skin preparations as pattern obtained from carp gill, no additional keratins were trace components. The liver preparations showed only two observed relative to skin, eye, or liver (Fig. 1h). Compared major keratins, which we identified as orthologs of human with fin tip preparations, an enrichment of proteins thought K8 and K18 (see below), together with some minor keratin to be the keratin K8 and K18 orthologs (see below) was components. noted in eye and gill from both fishes, and in the forehead Fig. 1 Two dimensional polyacrylamide gels and complementary the available keratin catalogs from zebrafish and trout (see also blot-binding assays (CCKB) of cytoskeletal preparations Fig. 3). Degradation products are demarcated by an arc (b, d, h). from goldfish (a–a″ fin, c–c″ liver, e–e″ eye, g–g″ gill, i–i″ The gel from intestine (i) exhibited a series of spots considered to be intestine) and carp (b fin, d liver, f eye, h gill). a–i Coomassie-blue– desmin (D). Spots identified as “S” keratins based on their ex- stained gels. Isoelectric focusing (IEF) was applied in the first pression in liver sometimes comigrated with proteins identified as dimension and SDS-PAGE in the second dimension. Bovine serum “E” keratins based on their abundance in skin (see Fig. 3a), and the albumin (B) and (A) were added as marker proteins. Left two could not be distinguished from one another with confidence Molecular weight is specified in kDa. Type I and type II keratins (IS+IE, relative position of these proteins in the gel after IEF, I? type were identified by CCKB with biotinylated recombinant human K8 I protein whose identity as an “E” or “S” keratin was uncertain). (HsaK8; a″, c″, e″, g″, i″) and K18 (HsaK18; a′, c′, e′, g′, i′), from Samples from goldfish and carp gill (g, h) revealed an extra type I gels similar to those shown in a, c, e, g, i. The “E” keratins were keratin spot (arrowhead) migrating close to the K18 spot, and in identified by their abundance in fin (a, b) and their absence in liver goldfish gill, two additional novel keratin spots were detected. One (c, d). Conversely, “S” keratins were identified by their presence in of these was a (MW∼45 kD; pI∼5.6; arrowhead), and liver and their enrichment in organs containing abundant mesen- the latter was a type II keratin (MW∼54 kD; pI∼5.2; arrow). With chymal tissues, such as eye (e, f), gill (g, h), and forehead skin (see the exception of gill, little degradation of proteins obtained from Fig. 2c). Identification of K8 and K18 was based on MALDI-MS goldfish tissues, including internal organs, was observed peptide mass fingerprinting and by comparison of our results with Fig. 2 Two dimensional poly- acrylamide gels and immuno- blots of cytoskeletal preparations from goldfish (a–a′ fin, d–d″ liver, e–e′ eye) and carp (b–b′ fin, c–c′ forehead skin, f–f″ eye). a–f Coomassie-blue- stained gels. Isoelectric focusing (IEF) was applied in the first dimension and SDS-PAGE in the second dimension. Bovine serum albumin (B) and actin (A) were added as marker proteins. Immunoblots were performed on blots derived from gels similar to those shown in a–f. Identifica- tion of K8 and K18 were based on MALDI-MS peptide mass fingerprinting and by compari- son of our results with the available keratin catalogs from zebrafish and trout (see also Fig. 3). Monoclonal antibody (mAb) 164.4 in goldfish recog- nized both “E” and “S” type II keratins, including K8 (a′, d′), whereas in carp, this antibody recognized all type II keratins with the exception of the K8 candidate (f′). The latter corresponded to the results ob- tained by mAb 79.14 (c′). The mAb KL1 was selective for the carp type IIE keratins (b′), whereas mAb LE64 specifically labeled type IE keratins (f″). In both carp and goldfish, mAb CO4 specifically labeled the presumed K18 ortholog and its putative isoelectric variants (b″, d″, e′)

skin from carp (Fig. 2c); no forehead skin was prepared versus type II keratins, but the range of isoelectric points from goldfish. Interestingly, a protein spot from goldfish overlap, and type I and type II keratins show no statistically gill and from carp gill and liver was observed that was significant difference. almost superimposed on the K18 spot but that migrated slightly more slowly and with a slightly more acidic isoelectric point. This protein was not recognized by Analysis of keratins by immunoblotting monoclonal antibody (mAb) C04 (see below) and was not and mass spectrometry observed in goldfish liver nor in the other carp tissues examined. We performed numerous Western analyses (Fig. 2) with the By further co-electrophoresis studies on the aforemen- two-fold purpose of characterizing the specificities of the tioned tissues, we found that the keratin spots that we had antibodies and of better defining the antigens recognized. identified as corresponding proteins in the different prep- Of the antibodies tested, only mAb C04 was specific for a arations did indeed co-migrate, whereas those that we had single keratin and its variants (Fig. 2b″,d″,e'), and this identified as distinct did not. Such co-electrophoreses proved to be the case in both carp and goldfish. The cor- represented the catalog of major keratins for goldfish and responding protein spot from goldfish liver (Fig. 2d,d″) carp (Fig. 3a,b); these keratins were also represented in was identified by MALDI-TOF mass spectrometry as ker- preparations from the gill of both species (Fig. 1g,h). atin K18, based on a comparison with the known goldfish For comparison, the keratin catalogs of zebrafish and keratin sequences GK48 (= K18), GK49, and GK50 trout are also given (Fig. 3c,d). As has been observed in the (Druger et al. 1992). Following tryptic digestion of the latter fishes, our present data reveal a statistically signif- spot, a MALDI spectrum with 52 peptide masses was icant (P<0.01) difference in the molecular weights of type I generated, and out of those, we found 21 masses matching Fig. 3 Keratin catalogs from goldfish, carp, zebrafish, and carp as products of several keratins). Molecular weight is specified in kDa. illustrated by Coomassie-blue-stained 2D-PAGE of cytoskeletal Note that the range of molecular masses and isoelectric points of the preparations from liver plus skin (goldfish), eye plus forehead skin keratins present in the four different is uniform. Never- (carp), eye (zebrafish), and RTG2-cells plus skin (trout). To make theless, the carp and goldfish keratin catalogs appear more similar to comparison of the four gels easier, the goldfish gel had been each other than either does to zebrafish, and the cyprinid fishes extended in length and thus does not show the aspect ratio of the appear more similar to one another than any does to trout (see also original gel (d cross in circle position of actin, which was barely Discussion) visible in this particular gel, open triangle putative degradation to a theoretical digest of the amino acid sequence for K8 was the prominent type II keratin in liver samples. goldfish keratin K18 (29.4% sequence coverage). The Furthermore, the migration behavior of this goldfish ker- amino acid sequences for GK49 and GK50 only matched atin in the 2D-PAGE resembled that of K8 from zebrafish 7 and 4 masses, respectively, with the obtained MALDI and trout (see Conrad et al. 1998; Markl et al. 1989; Fig. 3). spectrum. The average variance of the values for the de- Assignment of this protein spot to the putative goldfish K8 tected K18 masses compared with the theoretically cal- sequence (termed ON3; Giordano et al. 1989) by MALDI culated is only ±29 ppm. Moreover, two matching masses yielded 12 matching masses (from about 90 overall). Se- corresponded to different fragments of the K18 head do- quence coverage was only 21%, but a specific fragment main, which was expected to be highly specific for each from the head domain was included. Whether the liver type keratin. A protein migrating similarly to the protein sub- II keratin is identical with ON3 is therefore uncertain; jected to MALDI-MS was also observed in carp liver, probably the latter represents a closely related keratin, hence the previously designation of the same protein as possibly an additional K8 variant. K18 (Fig. 1d). In goldfish liver, a single major type II In carp, mAb KL1 appeared selective for the identified keratin was detected (see Fig. 1c,c′, 2d) that probably type IIE keratins (Fig. 2b′), whereas mAb LE64 appeared represented K8 because, in all vertebrates studied so far, specific for type IE keratins (Fig. 2f″); however, a possible

Table 2 Antibody specificities Antibody Keratin type specificity determined by immunoblotting Carp Goldfish Zebrafisha Troutb

C04 K18 K18 K18 K18 aDeduced from Conrad et al. KL1 IIE Not tested IIE IIE (1998) 164.4 II, except K8 II, including K8 IIE, except K8c K8a b Data from Markl et al. (1989) 79.14 II, except K8 II, including K8 IIE, except K8 K8a and Schaffeld et al. (2002a,b) LE64 IE Not tested IE Not tested cUnpublished additional reaction with co-migrating “S” keratin(s) of Immunolocalization of keratins in carp and goldfish LE64 could not be excluded. In goldfish, we observed a co-migration of “E” and “S” keratins at this position Consistent with published data from rainbow trout (Markl (Fig. 3a). Both mAb 79.14 (Fig. 2c′) and mAb 164.4 and Franke 1988, 1989) and zebrafish (Conrad et al. 1998), (Fig. 2f′) in carp recognized all identified type II kera- carp and goldfish generally expressed keratins in ecto- and tins, except the K8 candidate. Goldfish differed from carp endodermally derived epithelial tissues and in mesenchy- in that both mAb 79.14 (not shown) and mAb 164.4 mally derived tissues (Figs. 4, 5). Thus, anti-keratin anti- (Fig. 2a′,d´) recognized the K8 candidate in addition to bodies labeled both carp and goldfish epidermis (Figs. 4a–a′, the other type II keratins. The results of the immunoblot- 5a–a′), corneal epithelium (Fig. 4b–b′), gill epithelium ting analyses are summarized in Table 2 and compared (Figs. 4d–d′, 5c–c′), hepatocytes (Figs. 4e–e′, 5d–d′), renal with the situation observed in zebrafish (Conrad et al. tubules (not shown), fibroblasts (Figs. 4c–c′,d–d′,f–f′, 5a–a′, 1998).

Fig. 4 Micrographs of 5-μm- thick frozen tissue sections of carp skin (a, a′), cornea (b, b′, c, c′), gill (d, d′), liver (e, e′), intestine (f, f′), spleen (g, g′), brain (h, h′), and optic nerve (i, i′) labeled with the following mAbs: 79.14 (a′), 164.4 (b′), CO4 (c′, e′, f′), 68.4 (d′, i′). Spleen and brain were immunostained with polyclonal antibody GPpoly (g′, h′). The same fields were viewed by phase-contrast optics (a–i) and epifluorescence optics (a′–i′). The mAb 79.14 labeled the epidermis (E) in sections of carp forehead skin, in which Leydig cells (Le) were a salient feature. No labeling in the dermis (D) was detected (a, a′). In cornea, mAb 164.4 labeled the stratified epithelium of the cornea (C), leaving the underlying corneal stroma (S) unlabeled (b, b′). In contrast, mAb CO4 labeled the underlying stroma (S) but not the overlying corneal epithelium (C; c, c′). In gill, mAb 68.4 labeled the respiratory epitheli- um (R) and chondrocytes (Ch; d, d′). The CO4 mAb labeled hepatocytes (H), bile canaliculi (Ca), and blood vessels (B)in sections from the liver (e, e′), whereas in intestine (f, f′), labeling was restricted to con- nective tissue in the lamina propria (L); the columnar mu- cosal epithelium (M) was not labeled. The stroma of the spleen was labeled with the polyclonal antibody GPpoly (g, g′), which also immuno- stained blood vessels (B) and nervous tissue in the brain (h, h′). The specific type of cells in the brain could not be iden- tified, similar to the situation in optic nerve (i, i′), in which mAb 68.4 labeled vascular tissue (V) and nervous tissue (N). Bar 50 μm (in a) Fig. 5 Micrographs of 5-μm- thick frozen tissue sections of goldfish lip skin (a, a′), cornea (b, b′), gill (c, c′), liver (d, d′), intestine (e, e′), lens (f, f′), spleen (g, g′), brain (h, h′), and optic nerve (i, i′) labeled with the following mAbs: CO4 (a′, b′, d′, e′, f′), 68.4 (c′), 164.4 (i′). Spleen and brain were labeled with polyclonal antibody GPpoly (h′, i′, respectively). The same fields were viewed by phase-contrast optics (a–i) and epifluorescence optics (a′–i′). The CO4 mAb labeled dermis (D) in lip skin (a, a′) and stroma (S) in cornea (b, b′), but in neither case did it recognize the overlying stratified epithelium (E epidermis, C corneal epithe- lium). In gill, mAb 68.4 labeled respiratory epithelium (R) and connective tissue (c, c′). The mAb CO4 labeled hepatocytes (H), bile canaliculi (Ca), and blood vessels in sections from the liver (d, d′), whereas in intestine (e, e′), labeling was restricted to connective tissue in the lamina propria (L) and peri- mysium of the muscular layer (Mu); labeling of the columnar mucosal epithelium (M) was absent. The simple epithelium of the lens (Le) was strongly im- munopositive for mAb CO4 (F lens fiber cells; f, f′). Blood vessels (B) and germinal centers (G) in the spleen were labeled with the polyclonal antibody GPpoly (g, g′), together with elements in the stroma (S). GPpoly labeled blood vessels (B) and nervous tissue (N) in the brain (h, h′). The specific type of cells in the brain could not be identified, similar to the situa- tion in optic nerve (i, i′), in which mAb 164.4 labeled vas- cular tissue, nervous tissue (N), and endoneurium (En). Bar 50 μm (in a) b–b′,c–c′,e–e′,i–i′), blood vessel endothelium (Figs. 4f–f′, all differences among the three cyprinid fishes were ob- g–g′,h–h′,i–i′, 5d–d′,e–e′,g–g′,h–h′), stroma of spleen and served (Table 3). lymphatic tissues (Figs. 4g–g′, 5g–g′), and chondrocytes As noted above, mAb C04 specifically labeled K18 in (Fig. 4d–d′). Nervous tissue of the brain (Figs. 4h–h′, 5h–h′) immunoblots. Immunofluorescence microscopy demon- and optic nerve (Figs. 4i–i′, 5i–i′) were also labeled with strated the restriction of mAb C04 labeling to simple anti-keratin antibodies. Although we could not ascertain epithelia and several tissues of mesenchymal origin. An whether glial cells or were immunostained, the extensive survey of antibodies was performed on carp labeling of was much more likely on the basis of results tissues (see Table 3). In carp, mAbs 79.14 and KL1 from trout (see Markl and Franke 1988; Markl et al. 1989). exclusively labeled the stratified epithelia examined, cor- In addition, in both carp and goldfish, the dura mater of the responding to the situation observed in zebrafish by Conrad brain was positive for keratins (not shown). Whereas in et al. (1998). In addition to the labeling of many stratified some cases, the labeling patterns of individual antibodies and several simple epithelia, mAb 164.4 in carp and differed among goldfish, carp and zebrafish, no other over- goldfish also stained a variety of mesenchymal cells and Table 3 Antibody reaction patterns in carp, goldfish, zebrafisha and troutb determined by immunofluorescence microscopy Tissue Antibodies 68.4 C/G/Z/T LE64 C/G/Z/T KL1 C/G/Z/T 79.14 C/G/Z/T C04 C/G/Z/T Gppol C/G/Z/T 164.4 C/G/Z/T

Stratified epithelia Keratinocytes +/n/+/+ +/n/+/n +/n/+/+ +/n/+/−−/−/−/− +/+/+/+ +/+/+/− Anterior esophagus +/n/+/n n/n/+/n +/n/+/+ +/n/+/n −/n/−/− n/n/+/n +/n/+/n Gill mucosa +/n/+/+ +/n/+/n −/n/+/+ +/+/+/+c −/−/−/− n/+/+/n +/+/+/+c Corneal epithelium +/+/+/+ +/n/+/n +/n/+/+ +/n/+/−−/−/−/−d n/+/+/n +/+/+/−d Simple epithelia Intestinal mucosa +/n/+/+ −/n/+/n −/n/−/+ −/n/−/+ −+/+/− +/+/+/n +/+/−/+ Bile duct +/n/+/+ −/n/+/n −/n/−/+ −/n/−/+ +/n/+/ n/+/+/n −/−/−/+ Gall bladder +/n/n/n +/n/n/n n/n/n/+ −/n/n/n +/+/n/+ n/+/n/n +/+/−/n Hepatocytes +/n/+/+ −/n/−/n −/n/−/−−/n/−/+ +/+/+/+ n/+/+/n −/−/n/+ Lining of renal +/n/+/+ +/n/−/n −/n/−/−−/n/−/n −/+/+/+ +/+/+/n −/+/−/n Endothelia +/+/+/+ −/+/+/n −/n/−/−−/n/−/+ +/+/+/+ +/+/+/n +/+/−/+ Ocular lens n/+/+/− n/n/−/n n/n/−/− n/n/−/+ n/+/+/+ n/+/+/n n/n/n/+ Ocular lens fibers n/−/−/− n/n/−/n n/n/−/− n/n/−/− n/+/−/− n/+/−/n n/n/n/− Meninges −/+/n/+ −/+/n/n −/n/n/−−/−/n/+ −/n/n/+ −/n/n/+ −/n/n/+ Nonepithelial cells Chondrocytes +/n/+/+ n/n/+/n n/n/−/−−/n/−/+ +/+/+/+ n/+/+/n n/+/−/+ Fibroblasts +/+/+/+ +/n/+/n −/n/−/−−/+/−/+ +/+/+/+ +/+/+/+ +/+/−/+ Spleen stroma +/+/n/n +/+/n/n −/n/n/n −/+/n/n +/+/n/n +/+/n/n +/+/n/n Optic nerve n/+/+/+ n/n/−/n n/n/−/− n/n/−/+ n/+/+/+ n/n/+/n n/n/n/+ Smooth musclese +/n/+/+ n/n/+/n −/n/−/− n/n/−/+ −/n/+/−−/−/+/n −/−/−/+ Blood cells −/n/−/−−/n/−/n n/n/−/−−/n/−/−−/−/−/−−/−/−/n −/−/−/− Ovarial theca cells +/n/+/+ n/n/+/n −/n/−/−−/n/−/+ +/n/+/+ +/n/+/n +/n/−/+ Egg yolk granulesf −/n/−/− n/n/−/n −/n/−/− +/n/+/−−/n/−/−−/n/+/n +/n/−/− (C carp, G goldfish, Z zebrafish, T trout, + positive reaction, − negative result, n not tested) aData from Conrad et al. (1998); the reaction pattern of 164.4 evaluated in the study presented here bData from Markl et al. (1989) and Schaffeld et al. (2002a,b) plus some unpublished results cOnly basal cells were stained dOnly certain scattered cells were stained eOnly certain muscles were stained in all cases fOnly the outermost lining was stained

tissues, such as blood vessel endothelia, ovary, and con- 5.6), and which could represent modifications of a single nective tissue fibroblasts. gene product, products of different genes, or some com- bination of those possibilities. In trout, a second and more acidic K8 cluster is present (pI 5.0–5.4), which we have Discussion recently identified as a second trout K8 isoform (Omy- K8b; Schaffeld et al. 2002a). Several 2D-PAGE gels of Our analysis of the keratins from carp and goldfish has cytoskeletal preparations from carp and goldfish have revealed that the expression patterns are, for the most part, contained a vertically arranged double or triple spot with qualitatively similar to each other. How do they compare slightly different molecular masses (see Fig. 3a,b) at the with the data from zebrafish and trout? The range of position of K8. These spots may represent different K8 molecular masses and isoelectric points of the keratins variants because of tetraploidy or post-translational mod- present in the four different teleosts is mostly uniform (see ifications. Furthermore, in some gels, the K8 spot appears Fig. 3). These measurements indicate that the carp and to be slightly larger in molecular mass than the row of goldfish keratin catalogs are more similar to each other major IIE keratins, especially if the running time for the than either is to zebrafish, and the cyprinid fishes are more second dimension is extended (e.g. Fig. 3b). The latter has similar to one another than are any to trout. In all four also been observed in zebrafish (see Conrad et al. 1998; teleosts, the largest type II keratins appear as a row of Figs. 1, 2). Nevertheless, in all cyprinids and also in trout, polypeptides with apparent molecular masses of ∼58 kDa. the position of K8 is similar; the same applies to its smaller This row consists of the major IIE keratins (pIs 5.5–6.0) putative assembly partner K18. In trout, K18 migrates at a and several isoelectric variants of K8, which, in each case, similar pI as K8a, whereas cyprinid K18 is much less acidic form the most acidic spots of this 58-kDa cluster (pIs 5.5– than K8. In all three cyprinids, a second cluster of type II keratins simple epithelia and a number of different connective with molecular masses of ∼54 kDa in a pI range from 5.6 to tissues and brain cells (see Table 3, Fig. 5). 5.8 is present in addition to the major row of type II An alternative explanation supports the inference made keratins (Fig. 3a–c). In trout, a type IIE keratin spot can be by Conrad et al. (1998) that the K4 spot represents an “E” found in a corresponding position (Fig. 3d). Although this keratin. Schaffeld et al. (2003) have assigned the K4 spot second row of type II keratins is visible in preparations to the protein incorrectly described as zebrafish K8 by Ju obtained from skin of carp and goldfish, they are probably et al. (1999) and Gong et al. (2002). Ju et al. (1999) have not derived from the epidermis; they are more likely to reported that, in adult zebrafish, the protein that they come from the underlying mesenchymal tissue, from mes- studied is expressed only in the outermost layer of all enchymal cells scattered among the epidermis, and from stratified epithelia. Therefore, one would expect this pro- scale-associated cells (for details in trout, see Markl and tein to be present in skin and abundant in gill, in which Franke 1988). This inference is based on the presence of the respiratory surface is covered with a bilayered epi- this protein in preparations from the liver and intestine and thelium. Ju et al. (1999) have reported some minor reac- its abundance in gill, an organ which is replete with mes- tions in eye, gill, intestine, and muscle, based on in situ enchymal tissue (Figs. 1c,g–i, 2c). We have therefore clas- hybridization. Their in situ hybridization results have also sified these carp and goldfish keratins as expression type demonstrated some labeling in the midgut, which they IIS. have interpreted as being attributable to endogenous al- Generally, the major type I keratins detected in goldfish, kaline phosphatase activity. However, an interpretation of carp, and zebrafish also migrate in comparable positions, genuine expression of K4 in the midgut would be con- notably two main polypeptide clusters of 47–48 kDa, in sistent with our observation of the IIS row in liver and addition to a less acidic trace component (I?; see Fig. 3a,b), intestine. If this alternative accounts better for the data which, in goldfish, shows a slightly higher molecular and if K4 is expressed in both stratified and simple epi- mass than that in carp. The latter does not appear in thelia or mesenchymal tissue, then perhaps a better desig- zebrafish, but in trout a type IE keratin spot is present at a nation for expression type of K4 might be “E/S”. If this corresponding position. However, the major trout IE ker- is so, then one or more type IE/S keratins might also be atins migrate differently from those present in the three expected to exist. cyprinids in that they form a single cluster with a generally The two major spots obtained from goldfish liver and higher molecular mass and pIs from 5.3–5.5. Furthermore, intestine represent type IS and type IIS keratins and might the trout catalog includes two IS keratins of 42 kDa and be the orthologs of human K8 and K18; other spots are 45 kDa that are not observed in goldfish, carp, or considered likely to be additional “S” keratins (Fig. 2d). zebrafish. The mAb C04 was originally raised against human K18 The proteins that we designate herein as IIE keratins and has been found to label K18 orthologs specifically from carp and goldfish apparently correspond to the K1 in rainbow trout and zebrafish (Markl et al. 1989; Conrad proteins identified in zebrafish (Conrad et al. 1998). et al. 1998; Schaffeld et al. 2003) but not in shark or Therefore, we can compare the protein labeled as “K4” lungfish (Schaffeld et al. 2004, 2005). In the present study in gels derived from whole zebrafish and zebrafish skin to of goldfish and carp, mAb C04 recognizes K18 and its the second and smaller type II keratin row present in carp putative electrophoretic variants (posttranslational modifi- and goldfish (see Fig. 3a–c). As mentioned above, this cations or different genes) of the same molecular mass, protein is probably not epidermally derived and is likely to whereas an accompanying, slightly heavier keratin spot be a IIS keratin. The mAb 79.14 fails to recognize this spot observed in gill (see arrowhead in Fig. 1g,h) is not labeled. on immunoblots of zebrafish (see Fig. 1k,n in Conrad et al. Interestingly, in trout, the K18 cluster also consists of 1998), consistent with the observation that this antibody two major variants that show slightly different molecular exclusively labels stratified epithelia in zebrafish. The weights, although, in this case, both polypeptides are same antibody recognizes both the IIE and IIS keratin rows labeled by mAb C04 in immunoblots (Fig. 3c; Markl et al. in carp (except for K8) but nevertheless almost exclusively 1989; Schaffeld et al. 2002b). Thus, it remains unclear labels stratified epithelium in this fish. The mAb 164.4 whether the corresponding polypeptide in carp and goldfish shows a reaction pattern similar to that of mAb 79.14 in also represents a K18 variant. immunoblots but generally labels the IIS keratins more In carp, keratin expression in mesenchymal tissues has strongly. Interestingly, mAb 164.4 labeling as seen by im- previously been documented by Groff et al. (1997a,b), and munofluorescence microscopy includes several simple in the present study, the existence of mesenchymal kera- epithelia (e.g., intestinal mucosal and gall bladder epithe- tins has been established in both carp and goldfish (see lium and vascular endothelium) and some connective Table 3). In comparing the cytoskeletal preparations of fin tissues (e.g., the lamina propria of the intestine, the serosa, tip, which comprises mostly keratinocytes and only minor and endomysium of the heart). On immunoblots of goldfish amounts of mesenchymal tissue (dermis), with prepara- preparations, mAbs 79.14 and 164.4 also show labeling tions obtained from forehead skin, which contains much similar to one another, marking the IIE and IIS keratins, more dermis, we can see an increase in the representation but including K8 in this fish. The former antibody has not of K8 and K18 in the latter in carp. We have tried to been used extensively for immunofluorescence microsco- identify the mesenchymal keratins merely by looking for py of goldfish tissues, but the latter labels stratified and differences in the cytoskeletal preparation. From this anal- ysis, we conclude that the spots present in only minor fish, even from more perishable organs such as liver and amounts in fin tip preparations but in higher amounts in intestine, thus makes this fish an especially suitable model forehead skin arise from mesenchymal cells. These spots for the biochemical analyses of IF proteins in cyprinids. include those that we have labeled as IIS and the K8 and K18 spots. Consistent with this interpretation, mAb KL1 in Acknowledgements We thank Harald Genrich for supplementary carp labels only the identified type IIE keratins in immu- immunofluorescence experiments, Dr. Christian Hunzinger for MALDI-TOF analysis, and Prof. Dr. Werner W. Franke’s group for noblots and only immunostains stratified epithelia by im- providing several antibodies and recombinant human keratins 8 and munofluorescence microscopy. Similarly, mAb C04 labels 18. only the K18 spot, which is strongly enriched in forehead skin relative to fin skin, and fails to immunostain any of the stratified epithelia probed. References Based on the reported (Conrad et al. 1998) and observed data, the molecular masses of type I versus type II keratins Alibardi L (2001) Keratinization in the epidermis of amphibians and are statistically significantly different in the cyprinid fishes. the lungfish: comparison with amniote keratinization. Tissue Cell 33:439–449 This difference is also observed in trout (Markl et al. 1989; Alibardi L (2002) Immunocytochemical localization of keratins, Schaffeld et al. 2002a,b), lungfish (Schaffeld et al. 2005), accociated proteins and uptake of histidine in the epidermis of and shark (Schaffeld et al. 1998). In contrast, statistically fish and amphibians. Acta Histochem 104:297–310 significant differences in the isoelectric points of type I and Alibardi L, Joss JMP (2003) Keratinization of the epidermis of the – type II keratins have not been observed in any of these Australian lungfish Neoceratodus forsteri. J Morphol 256:13 22 fishes. Therefore, a designation of type I keratins as acidic Cerda J, Conrad M, Markl J, Brand M, Herrmann H (1998) or type II as basic is not supported in the fishes studied thus Zebrafish vimentin: molecular characterization, assembly prop- far. Nevertheless, the presumed partners for keratin dimer erties and developmental expression. Eur J Cell Biol 77:1–13 formation always seem to differ substantially in pI, mo- Cohen F, Shani Y, Blaugrund E, Schwartz M (1991) Isolation and sequence analysis of two intermediate filament cDNA clones lecular mass, or both. For example, the major IIE keratins from fish optic nerve. Mol Brain Res 11:181–185 of skin are substantially more basic than the major IE Conrad M, Lemb K, Schubert T, Markl J (1998) Biochemical keratins. Conversely, in cyprinids, the type I keratin K18 is identification and tissue-specific expression patterns of keratins much more basic than its presumed type II partner K8. In in the zebrafish Danio rerio. Cell Tissue Res 293:195–205 Corcoran JP, Ferretti P (1997) and 18 expression in trout, K8 and K18 are similar in pI but markedly different mesenchymal progenitor cells of regenerating limbs is asso- in mass (see Markl et al. 1989). ciated with cell proliferation and differentiation. Dev Dyn 210: In conclusion, the closer phylogenetic relationships 355–370 among the three cyprinid fishes and the more distant Corcoran JP, Ferretti P (1999) RA regulation of keratin expression position of the trout are evident not only from keratin and myogenesis suggests different ways of regenerating muscle in adult amphibian limbs. J Cell Sci 112:1385–1394 sequence alignments (e.g. Schaffeld et al. 2002a,b, 2003, Druger RK, Levine EM, Glasgow E, Jones PS, Schechter N (1992) 2004, 2005), but also from their 2D-PAGE keratin fin- of a type I keratin from goldfish optic nerve: dif- gerprints. This result implies that, for coarse examination ferential expression of keratins during regeneration. Differen- of tissue-specific keratin expression, cyprinid fishes are tiation 52:33–43 Druger RK, Glasgow E, Fuchs C, Levine EM, Matthews JP, Park interchangeable. For example, in previous studies on zeb- CY, Schechter N (1994) Complex expression of keratins in rafish, cytoskeletal preparations from internal organs such goldfish optic nerve. J Comp Neurol 340:269–280 as liver or intestine that have not suffered substantial Ferretti P, Fekete DM, Patterson M, Lane EB (1989) Transient degradation were impossible to obtain (Conrad et al. 1998). expression of simple epithelial keratins by mesenchymal cells of regenerating newt limb. Dev Biol 133:415–424 As we have pointed out above, tissues from internal organs Ferretti P, Corcoran JP, Ghosh S (1993) Expression and regulation of carp also yield highly degraded , even of keratins in the wound epithelium and mesenchyme of the when several protease inhibitors are employed. In contrast, regenerating newt limb. Prog Clin Biol Res 383A:261–269 by using goldfish for the same purpose, we have been able Fouquet B (1991) Expression von Intermediärfilament-Proteinen to isolate good quality cytoskeletal preparations that have in laevis. Doctoral Thesis, Faculty of Biology, Heidelberg enabled us to identify a pattern of keratins with two main Franke WW, Schmid E, Osborn M, Weber K (1978) Different keratin spots (the putative K8 and K18) accompanied by intermediate-sized filaments distinguished by immunofluores- several probable electrophoretic variants. Moreover, we cence microscopy. Proc Natl Acad Sci U S A 75:5034–5038 have been able to use the preparation from goldfish liver Franz JK, Franke WW (1986) Cloning of cDNA and amino acid sequence of a cytokeratin expressed in oocytes of Xenopus unambiguously to identify the goldfish K18 ortholog and laevis. Proc Natl Acad Sci U S A 83:6475–6479 characterize its migration in 2D polyacrylamide gels. Fuchs E, Weber K (1994) Intermediate filaments: structure, dy- Thus, the carp and goldfish offer a distinct advantage over namics, function and disease. Annu Rev Biochem 63:345–382 smaller cyprinid fishes such as the zebrafish in that cy- Fuchs C, Druger RK, Glasgow E, Schechter N (1994) Differential expression of keratins in goldfish optic nerve during regener- toskeletal preparations from their internal organs are much ation. J Comp Neurol 343:332–340 more easily obtained. In addition, the isolation of organs Giordano S, Glasgow E, Tesser P, Schechter N (1989) A type II for histological observation is also easier. The observation keratin is expressed in glial cells of the goldfish visual pathway. that high quality preparations can be generated from - 2:1507–1516 Giordano S, Hall C, Quitschke W, Glasgow E, Schechter N (1990) Owaribe K, Kartenbeck J, Rungger-Brändle E, Franke WW (1988) Keratin 8 of simple epithelia is expressed in glia of the goldfish Cytoskeletons of retinal pigment epithelial cells: interspecies nervous system. Differentiation 44:163–172 differences of expression patterns indicate independence of cell Gong Z, Ju B, Wang X, He J, Wan H, Sudha PM, Yan T (2002) function from the specific complement of cytoskeletal proteins. Green fluorescent protein expression in germ-line transmitted Cell Tissue Res 254:301–315 transgenic zebrafish under a stratified epithelial promoter from Paramio JM, Jorcano JL (2002) Beyond structure: do intermediate keratin8. Dev Dyn 223:204–215 filaments modulate cell signalling? BioEssays 24:836–844 Groff JM, Nayan DK, Higgins RJ (1997a) Cytokeratin-filament Parry DA, Steinert PM (1999) Intermediate filaments: molecular expression in epithelial and non-epithelial tissues of the com- architecture, assembly, dynamics and . Q Rev mon carp (Cyprinus carpio). Cell Tissue Res 287:375–384 Biophys 32:99–187 Groff JM, Naydan DK, Zinkl JG, Osburn BI (1997b) Immunolog- Rogers MA, Winter H, Langbein L, Bleiler R, Schweizer J (2004) ical cross-reactivity of type I–III intermediate filaments in the The human type I keratin gene family: characterization of new : in situ localization with use of heterologous hair follicle specific members and evaluation of the chromo- antibodies. Trans Am Fish Soc 126:948–960 some 17q21.2 gene domain. Differentiation 72:527–540 Herrmann H, Münick MD, Brettel M, Fouquet B, Markl J (1996) Rogers MA, Edler L, Winter H, Langbein L, Beckmann I, Vimentin in a cold-water fish, the rainbow trout: highly con- Schweizer J (2005) Characterization of new members of the served primary structure but unique assembly properties. J Cell human type II keratin gene family and a general evaluation of Sci 109:569–578 the keratin gene domain on 12q13.13. J Invest Herrmann H, Hesse M, Reichenzeller M, Aebi U, Magin TM (2003) Dermatol 124:536–544 Intermediate filaments: novel assembly models and exciting Schaffeld M, Markl J (2004) Fish keratins. Methods Cell Biol 78: new functions for nuclear lamins. Cell Mol Life Sci 60:1607– 627–671 1612 Schaffeld M, Löbbecke A, Lieb B, Markl J (1998) Tracing keratin Hesse M, Magin TM, Weber K (2001) Genes for intermediate evolution: catalog, expression patterns and primary structure of filament proteins and the draft sequence of the human : shark (Scyliorhinus stellaris) keratins. Eur J Cell Biol 77:69–80 novel keratin genes and a surprisingly high number of pseu- Schaffeld M, Herrmann H, Schultess J, Markl J (2001) Vimentin and dogenes related to keratin genes 8 and 18. J Cell Sci 114:2569– desmin of a cartilaginous fish, the shark Scyliorhinus stellaris: 2575 sequence, expression patterns and in vitro assembly. Eur J Cell Hesse M, Zimek A, Weber K, Magin TM (2004) Comprehensive Biol 80:692–702 analysis of keratingene clusters in humans and rodents. Eur J Schaffeld M, Haberkamp M, Braziulis E, Lieb B, Markl J (2002a) Cell Biol 83:19–26 Type II keratin cDNAs from the rainbow trout: implications for Hoffmann W, Franz JK (1984) Amino acid sequence of the carboxy- keratin evolution. Differentiation 70:282–291 terminal part of an acidic type I cytokeratin of molecular weight Schaffeld M., Höffling S, Haberkamp M, Conrad M, Markl J 51000 from Xenopus laevis epidermis as predicted from the (2002b) Type I keratin cDNAs from the rainbow trout: inde- cDNA sequence. EMBO J 3:1301–1306 pendent radiation of keratins in fish. Differentiation 70: 292– Hoffmann W, Franz JK, Franke WW (1985) Amino acid sequence 299 microheterogeneities of basic (type II) of Xenopus Schaffeld M, Knappe M, Hunzinger C, Markl J (2003) cDNA laevis epidermis and evolutionary conservativity of helical and sequences of the authentic keratins 8 and 18 in zebrafish. Dif- non-helical domains. J Mol Biol 184:713–724 ferentiation 71:73–82 Ju B, Xu Y, He J, Liao J, Yan T, Hew CL, Lam TJ, Gong Z (1999) Schaffeld M, Höffling S, Markl J (2004) Differential expression of Faithful expression of green fluorescent protein (GFP) in trans- “E” and “S” keratins in the shark Scyliorhinus stellaris, and genic zebrafish under control of zebrafish gene pro- cDNA sequence of a novel epidermal type I keratin. Eur J Cell moters. Dev Genet 25:158–167 Biol 83:359–368 Kirfel J, Magin TM, Reichelt J (2003) Keratins: a structural scaffold Schaffeld M, Bremer M, Hunzinger C, Markl J (2005) Evolution of with emerging functions. Cell Mol Life Sci 60:56–71 tissue-specific keratins as deduced from novel cDNA sequences Markl J, Franke WW (1988) Localization of cytokeratins in tissues of the lungfish Protopterus aethiopicus. Eur J Cell Biol 84: of the rainbow trout: fundamental differences in expression 363–377 pattern between fish and higher vertebrates. Differentiation 39: Tsonis PA, Mescher AL, Del Rio-Tsonis K (1992) Protein synthesis 97–122 in the newt regenerating limb. Comparative two-dimensional Markl J, Schechter N (1998) Fish intermediate filament proteins in PAGE, computer analysis and protein sequencing.Biochem J structure, function and evolution. Sub-Cell Biochem 31:1–33 281:665–668 Markl J, Winter S, Franke WW (1989) The catalog and the ex- Watanabe Y, Kobayashi H, Suzuki K, Kotani K, Yoshizato K (2001) pression complexity of cytokeratins in a teleost fish, the rain- New epidermal keratin genes from Xenopus laevis: hormonal bow trout. Eur J Cell Biol 50:1–16 and regional regulation of their expression during anuran skin Martorana ML, Tawk M, Lapointe T, Barre N, Imboden M, Joulie C, metamorphosis. Biochim Biophys Acta 1517:339–350 Geraudie J, Vriz S (2001) Zebrafish keratin 8 is expressed at Watanabe Y, Tanaka R, Kobayashi H, Utoh R, Suzuki K, Obara M, high levels in the epidermis of regenerating caudal fin. Int J Yoshizato K (2002) Metamorphosis-dependent transcriptional Dev Biol 45:449–452 regulation of xak-c, a novel Xenopus type I keratin gene. Dev Mencarelli C, Cotelli F (1997) Intermediate filament proteins im- Dyn 225:561–570 munologically related to cytokeratins in the oocyte of the fish Winkles JA, Sargent TD, Parry DA, Jonas E, Dawid IB (1985) Cyprinus carpio. Zygote 5:207–212 Developmentally regulated cytokeratin gene in Xenopus laevis. Miyatani S, Winkles JA, Sargent TD, Dawid IB (1986) Stage- Mol Cell Biol 5:2575–2581 specific keratins in Xenopus laevis embryos and tadpoles: the Winter H, Langbein L, Krawczak M, Cooper DN, Jave-Suarez LF, XK81 gene family. J Cell Biol 103:1957–1965 Rogers MA, Praetzel S, Heidt PJ, Schweizer J (2001) Human Moll R, Franke WW, Schiller DL, Geiger B, Krepler R (1982) The type I pseudogene phihHaA has functional or- catalog of human cytokeratins: patterns of expression in normal thologs in the chimpanzee and gorilla: evidence for recent epithelia, tumors and cultured cells. Cell 31:11–24 inactivation of the human gene after the Pan-Homo divergence. Hum Genet 108:37–42