EXPERIMENTAL

CELL

RESEARCH

202,132-141

(19%)

Characterization of Human Cytokeratin 2, an Epidermal Cytoskeletal Protein Synthesized Late during Differentiation’ CHRISTINE *Institute

COLLIN,*,~

ROLANDMOLL,~ STEFANKUBICKA,~ ANDWERNER W. FRANKE*,'

of Cell and Tumor Biology, German Cancer Research Biologie-Odontologie, Institut Biomt!dical des Cordeliers, and *Institute of Pathology, University

Center, D-6900 Heidelberg, Federal Republic of Germany; TLaboratoire 15-21, rue de 1’Ecole de Mt!decine, F-75270 Paris Cedex 06, France; of Mainz, Mainz, Federal Republic of Germany

tion of epidermal utes to terminal

Among the more than 30 different human proteins of the cytokeratin (CK) group of intermediate filament (IF) proteins, the significance of the epidermal polypeptide CK 2 (Mall et cd., 1982, Cell 31, 11-24) has been repeatedly questioned in the literature. Here, we show, by in vitro translation and protein gel electrophoresis, that human epidermis from various body sites does indeed contain relatively large amounts of mRNA encoding a distinct polypeptide comigrating with native epidermal CK 2. We also report the isolation of a cDNA clone encoding the complete sequence of CK 2, which is a type II CK different from-but related to-epidermal CKs 1 and 5 on the one hand and cornea1 CK 3 on the other. The mRNA of -2.6 kb encodes a polypeptide of 645 amino acids and iV= 65,852, in good agreement with the value of 65.5 kDa previously estimated from gel electrophoresis. This human CK, the largest so far known, displays several features typical of CKs of stratified epithelia, including numerous repeats of glycinerich tetrapeptides in the head and tail domains. Northern blot and in situ hybridizations have shown that CK 2 is expressed strictly suprabasally, usually starting in the third or fourth cell layer of epidermis, and this was confirmed at the protein level by immunohistochemistry using CK 2-specific antibodies. The protein has been detected as a regular epidermal component in skin samples from different body sites, albeit as a minor CK in “soft skin” (e.g., breast nipple, penile shaft, axilla), but not in foreskin epithelium and in other epithelia, in squamous metaplasias and carcinomas, or in cultured cell lines derived therefrom. We propose that CK 2 is a late cytoskeletal IF addition synthesized during matura-

$5.00

Copyright 0 1992 by Academic Press, Inc. All rights of reproducGon in my form reserved.

keratinocytes cornification.

which 0 1992

probably Academic

de

contribPress,

Inc.

1NTRODUCTION The intermediate filament (IF)3 cytoskeleton of epithelial cells consists predominantly, if not exclusively, of cytokeratins (CKs) which therefore are widely regarded as hallmarks of epithelial differentiation ([l-5]; for reviews see [6-lo]). Two subfamilies of CKs can be distinguished (type I and type II; e.g., [11-14]), and the basic subunit (“building block”) of CK IFS is a heterotypic tetramer of two polypeptide chains of each type I and type II (e.g., [14-171; for review see [X3]). The specific polypeptide composition of CK IFS, however, varies from one cell type to another, and a certain CK polypeptide complement is characteristic for a given epithedial tissue or carcinoma (e.g., [6-8, 191). However, although in vitro CK complexes and IFS can be formed from many, probably all, stoichiometric combinations of type I and type II CKs [20,21], only certain combinations of type I and II CKs occur in Go, whether in tissues or in cell cultures [6, 8, 22, 231. Epidermal proliferation and differentiation are characterized by the expression of three major sets of CKs: (i) type II CK 5 occurs together with type I CK 14 in normal keratinocytes of the basal cell layer; (ii) type II CK 6 is usually synthesized together with type I CK 16 in keratinocytes particularly active in proliferation (“hyperproliferative states”); and (iii) type II CK 1 appears, mostly together with type I CK 10 and/or 11, during suprabasal differentiation but the synthesis of

’ This work was supported by the Ministry for Research and Technology of the FRG and the Deutsche Forschungsgemeinschaft. C. Collin received a stipend from the French-German Projets de Coop&ration et d’Echanges (PROCOPE). ‘To whom correspondence and reprint requests should be addressed at Institute of Cell and Tumor Biology, German Cancer Research Center, Im Neuenheimer Feld 280, D-6900 Heidelberg, Federal Republic of Germany. Fax: ++49-6221-402598.

0014-4827/92

JEAN-PIERRE~UHAYOUN,?

’ Abbreviations used: 2D GE, two-dimensional gel electrophoresis; CK, cytokeratin; CK 2e, epidermal cytokeratin 2; IF, intermediate filament; NEPHGE, nonequilibrium pH gradient electrophoresis; PBS, phosphate-buffered saline; PCR, polymerase chain reaction; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis. 132

HUMAN

EPIDERMAL

these CKs often already starts in certain cells of the basal layer [8, 9, 24-301. In addition, the synthesis of type I CK 9 has een described in suprabasal cell layers of epidermis of special body sites, most prominently in palmar and plantar epidermis, where it is abundant [3lj 321. Another type II buman CK, of Mr -65,500 on sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), has been listed among the epidermal CK polypeptides (CK 2; [6, 331)~ IIowever, other authors have regarded this polypeptide as a product of proteolytic degradation from CK 1 (e.g~, [9,28, 34-381; cf. [39]; see also [4Q]). To decisively answer the question of the significance of this CK polypeptide and to examine whether it is a distinct gene product or is proteolytically derived from CK lY we have characterized it at both the protein and the mRNA level. Gur demonstration that CK 2 is indeed a unique and genuine cytoskeletal protein, regularly specific for epidermis is primarily based on the analysis of the corresponding cDNA and on antibodies specific for this major cytoskeletal protein. MATERXALS

AND

METHQDS

T&sues. Specimens of buman skin from various sites (e.g., foot sole, face, breast skin and nipple, axillary pit, penile shaft, foreskin, groin, back of the knee, thigh) were obtained after surgery for various medical reasons or during autopsy (within 8 h postmortem) and were immediately frozen in isopentane, cooled with liquid nitrogen to -130°C, or directly frozen in liquid nitrogen and usually stored at -7Q’C (cf. [6, 33, 41]). Small biopsies of mucosae from hard palate and gingiva were obtained during periodontal surgery (cf. [42, 431). Samples from several other nonepidermal human tissues (cf. [6,33]) were taken from autopsies or upon large excision operations, frozen, and stored as described (e.g., [6, 19, 44, 451). Human cornea1 epithelium was obtained during autopsy by gentle scraping with a scalpel, leaving the body of the eye intact. The scraped off tissue material was washed as a loose suspension in phosphate-buffered saline (PBS), collected by gentle centrifugation, and frozen as pellets. Purijication of epidermal cytokeratin 2 (CK 2e) and preparation of antibodies. CK 2e was purified from cytoskeletal preparations of human thigh epidermis. Slices of fresh buman thigh skin from a leg amputated because of arteriosclerosis and foot gangrene were obtained with a fine scalpel. The epidermis was separated from the dermis by incubation at 6O’C for 3 min in PBS and was frozen in liquid nitrogen. The frozen tissue was ground in a cold mortar with a cold pistil, and cytoskeletal material was prepared therefrom by subsequent extractions of soluble material in a buffer containing 1.5 &f KC1 and Triton X-100 [6,19]. After an additional wash in a low-salt buffer [19], the cytoskeletal polypeptides were separated by preparative two-dimensional gel electrophoresis (2D GE), using nonequilibrium pH gradient eiectrophoresis (NEPHGE) in the first dimension and SDS-PAGE in the second [46, 471. The spot series representing CK Ze, visualized by negative staining with 4 &f sodium acetate (cf. [46]), was excised from a total of 180 gels, and protein was eluted by diffusion using 0.1% SDS and concentrated by vacuum dialysis. The purity of the CK 2e preparation was confirmed by 2D GE followed by si!ver staining. After addition of 4.5 mg murine albumin (Sigma, Munich Germany) per 100 fig of CK 2e, the purified protein was precipitated with acetone [46]. Antibodies against CK 2e were raised in a female BALB/c mouse (6

CYTOKERATIN

2

133

weeks old). For the first immunization, an aliquot of the acetone-precipitated and dried protein powder> containing 20 pg of CK 2e, was finely suspended in 50 ~1 sterile 0.9% NaCl, supplemented with 100 ,ul comp!ete Freund’s adjuvant (Merck, Darmstadt, Germany) ernul.. gated by sonication, and injected intraperitoneally. For boosting, the same amount of antigen, but with incomplete Freund’s adjuvant, was injected at Days 29, 43, and 111. Serum obtained by bleedings from the retrobulbar plexus at Days 121 and 127 was tested by immunoblotting on human epidermal cytoskeletal proteins separated by 2D GE. This test showed the presence of high-titer antibodies selectively reactive with CKs 1, 2, and 5. Therefore, the antiserum was further affinity-purified by repeated incubations with nitrocellulose membrane strips loaded with CK 2e by electrotransfer after SDS-PAGE [46, 481. Subsequently, the serum was absorbed with similar strips containing CKs 1 and 5 to remove cross-reacting antibodies. The specificity of the CK 2e antibodies thus enriched was demonstrated by immunoblot reaction with cytoskeletal or total human epidermai proteins separated by 2D GE. Isolation of RNA asd in vitro translation. The protocol used was essentially as described by Kreis et al. [49]. A431 cells, originally derived from a vulvar carcinoma [50], HaCaT cells, derived from keratinocytes of adult buman back skin [51]$ or buman tissue samples were homogenized for 4-5 mm in 5 M GTC btiffer (5 M guanidine thiocyanate, 0.1 J4 Tris-HCl? pH 7.5, 10 rnM DTT? 10 miW EDTA) with a Polytron homogenizer (Kinematica, Luzern, Switzerland). After centrifugation at 5000g for 10 min at 4’C to remove cellular debris, the supernatant was precipitated with 0.5 vol of ethanol and stored at -2O’C overnight. The precipitated RNA was collected by centrifugation at lO,OOOg for 30 min at 1O’C and resuspended in 0.5 vol (relative to tbe initial volume of homogenization) of 7 AJ “guanidium buffer” (7 &f guanidium-HCl, 0.1 M sodium acetate, pH 5.5> 10 rnM DTT, 10 m&’ EDTA) with a Polytron homogenizer. RNA was again precipitated with 0.5 vol of ethanol for 5 mm in ).iqaid nitrogen and, after thawing, collected by centrifugation. This procedure was repeated witb smaller volumes of guanidium buffer. The final pellet was dissolved in 10 m&f ‘Iris-HCl (pH 7.5), 1 rnlV EDTA, 1% SDS> extracted three times with phenol:chloroform (l:l), and precipitated with 2 vol of ethanol in the presence of 0.3 1!4 sodium acetate (pH 5.5). Total RNA was translated in uitro using the rabbit reticulocyte lysate system (Promega, Madison, WI) and [?S]methionine (1200 Wmmol; Amersham-Buchler, Braunschweig~ Germany). The fs5S]methionine-labeled translation products were cbaracteri.zed by 2D GE using appropriate cytoskeletal proteins for comigration [52, 53]. Cloning of partial ciliVA by polymerax chain reuction (PCR). The procedure used was that of Frobman et al. [54] with minor modifications. Briefly, 5 gg of total RNA from thigh epidermis was beated at 65°C for 3 mm, chilled on ice, and reverse-transcribed in 29 ~1 of “reverse transcription buffer” (50 m&f Tris-HCl, pH 8.15, at 41°C, 6 rnM MgC&, 40 rnM KCl, 1 rnM DTT, each dKE? at 1.5 mM) containing 20 units of “RNasin” (Pharmacia, F&burg i. Br., Gernxny), 0.50 &f oligo(dT)is primer, and 20 units of avian myelo-blastosis virus reverse transcriptase (Boehringer-Mannheim, Mannbeim, Germany). The reaction mixture was incubated for 2 b at 4l’C and diluted t,o 1 ml with “TE b&e? (10 rnM ‘Tris-HCl, pH ?.5g 0.1 rn>d EDTA). One- to ten-microliter aliquots of the first cDNA strand were slubjected to amplification by PCR, using an oligo(dT)1s primer and a synthetic primer (5’.dAGATCGCCACCTACCGCAAA-3’1, corresponding to a sequence located eight bases upstream of the amino acid sequence motif TYRKLLEGEE that is conserved in all IF proteins and situated at the carboxy-terminal end of the a-helical “rod” domain (see [18]). Amplification was performed in 100 ~1 of “PCR b&e? (50 rnM KCl, 1.5 m&f MgC&> 10 m&f Tris-HCl, pH 8.3, 0.01% BSA) containing 200 ),& of each dNTP, 0.50 +iW of both primers, and 2.5 units of Ampli-Taq (Perkin-Elmer Cetus, Norwalk, CT). After 30 cycles of amplification (denaturation, 0.5 min at 94’C; annealing, 0.5 mm at

134

COLLIN

55°C; extension, 1 min at 72’C), followed by 10 min elongation at 72’C, PCR products were purified by electrophoresis on low-meltingpoint agarose gel and cloned into the Not1 site of Bluescript vector (Stratagene, Heidelberg, Germany). The recombinant plasmids were digested by NotI, and the insert-containing constructs in the size range of 0.8-1.2 kb were studied by sequence analysis. On the basis of its characteristic sequence, notably its similarity with human CK 1, one of the cDNA clones (pEK2TYR) was selected for further studies. Preparation of radioactively labeled selective hybridization probes. A clone, derived from clone pEK2TYR by deletion from the 3’-end using exonuclease III, was digested with BglII and religated to itself. The resulting clone, termed pEK2TYR-2AC8, contains a 259-bp DNA insert representing a portion of the 3’-untranslated region (nucleotides 213222391 of pEK2; see below) and was used to detect the expression of CK 2e. To compare the level of expression of CK 2e with that of another terminal differentiation-specific CK (e.g., CK 1), a clone, termed pTZl9-pKH1 (cf. [43]), encompassing a 450-bp ,S’ulI/XhoI DNA fragment of exon 1 of the human CK 1 gene inserted into the pTZl9 R vector (Pharmacia), was used. To prepare antisense riboprobes, the linearized templates were transcribed in the presence of [a-3zP]CTP (NEN, Dreieich, Germany; 800 Ci/mmol, 10 pCi/pl) and T3 or T7 RNA polymerase, following the “pBluescript” protocol (Stratagene). A double-stranded DNA probe from clone pEK2TYR-ZAC8 was prepared using the “Random Primed DNA labeling” kit (Boehringer-Mannheim). Cloning of the full-length cDNA ECK2. A “mini-cDNA library” was constructed from RNA specifically released after hybrid selection performed as follows. Briefly, 200 pg of total RNA extracted from thigh epidermis was hybridized to the 3’-end probe, pEK2TYR-2AC8, bound to nitrocellulose filter. After washing at 65’c and boiling, the specifically released RNAs were resuspended in 10 ~1 of “TE buffer”. To construct the library, 2 ~1 of the selected RNAs was used for double-strand cDNA synthesis with the help of the “you-prime” cDNA kit (Pharmacia). The products were cloned into the EcoRI site of the A Zap11 vector (Stratagene), and the phage DNA was then packaged into phage particles with the “Giga Gold” packaging extract (Stratagene), Library screening was done using 32P-labeled cRNA derived from pEK2TYR-2ACS and the same hybridization and washing conditions as those used for Northern blot analysis (see below). Positive phages were plaque-purified and the inserts excised in vivo into the pBluescript vector by infection with the helper phage R408 according to the “A Zap11 protocol” (Stratagene). All the recombinant plasmids analyzed showed sequence identity with the partial clone pEK2TYR. A plasmid (pEK2) containing an insert of the expected full mRNA size (2.4-2.6 kb) was further studied by coupled in vitro transcription/ translation. DNA sequencing. Both strands of the cDNAs were sequenced with the “T7 sequencing” kit (Pharmacia). For the sequence of inserts longer than 0.5 kb, the Bluescript constructs were linearized and unidirectionally deleted using the “double-stranded nested deletion” kit (Pharmacia), and the deletion clones obtained were sequenced with the same protocol. Hybrid selection of mRNA. Total RNA from thigh epidermis was hybridized to filter membrane-bound cloned cDNA (clone pEK2TYR-2AC8) and the specific mRNAs were released and translated essentially as reported (cf. [53, 551). The [?S]methioninelabeled translation products from the selected mRNAs were separated by 2D GE and identified by electrophoresis with unlabeled cytoskeletal proteins from thigh epidermis. Southern blot analysis. Human genomic DNA isolated from placenta was processed for Southern analysis according to the protocol described by Sambrook et ul. [56]. After overnight digestion with different enzymes, DNA was separated by electrophoresis, transferred to “N-bond” membrane (Amersham-Buchler), and hybridized to the

ET

AL.

“P-labeled riboprobe from pEK2TYR-2ACS. Hybridization was carried out overnight as described for Northern blot analysis, but at 42’C. Washing of the filter was performed by two washing steps at 7O’C for 30 min in 0.1X SSC-1% SDS. Northern blot anaZysis. Total RNA was electrophoresed on 0.9% agarose gels containing formaldehyde [57] and blotted onto N-bond membrane (Amersham-Buchler). The filters were hybridized for 1218 h with 3-5 X 106 cpm/ml of specific radioactive RNA or DNA probe (see above) in 50% formamide, 5X SSPE (1X SSPE is 0.18 MNaCl, 20 rnM NaHzPOd, 1 rnM EDTA, pH 7.4), 5X Denhardt’s solution, 1% SDS, and 100 pg/ml yeast tRNA [47]. The hybridization temperature and the washing conditions were different according to the type of probe used. For riboprobes, the hybridization temperature was 68’C and one 20-min wash was carried out in 0.1X SSC (1X SSC is 0.15 &f NaCl, 0.015 A!f sodium citrate, pH 7.0), 1% SDS at 72’C, followed by RNase treatment in 2X SSC-0.1% SDS containing 30 fig/ml of RNase A for 20 min at room temperature. The filters were washed again in 0.1X SSC-1% SDS for 20 min at 72’C and then processed for autoradiography. When double-stranded DNA probes were used, the hybridization temperature was 42’C and the washing temperature 65’C (in 0.1x SSC, 1% SDS). In situ hybridization. The procedure used for in situ hybridization on tissue sections was based on the work of Cox et al. [58] modified as described [43, 571, and [35S]labeled CK cRNA probes were used. Zmmunohistochemistry. Immunofluorescence and immunoperoxidase microscopy on acetone-fixed cryostat sections were performed as described [7,59], sometimes with the additional use of a detergent, saponin (cf. [60]). Monoclonal antibody AE5, reported to be specific for cornea1 CK 3 [61], was a generous gift of Dr. T.-T. Sun (Department of Dermatology, New York University Medical Center).

RESULTS

Significance

of CK 2 as a Genuine

mRNA

Product

Epidermal tissues from various sites contain a slightly basic cytoskeletal protein (Fig. la) with typical properties of a type II CK, termed “CK 2” [6], including cross-immunoreactivity with several authentic CK antibodies and in vitro binding of type I CKs in the blot assay introduced by Hatzfeld et al. [62]. As the relative amounts and the number of isoelectric variant spots of this CK were found to vary considerably in different preparations from various sources of skin, and they were not detected at all in some preparations of epiderma1 material such as from foreskin, it was suggested that this component is not a genuine protein but is proteolytically derived from CK 1 (for discussion see, e.g., [9, 28, 34-381). Therefore, we decided to examine the significance of this polypeptide by in vitro translation of RNA isolated from normal thigh epidermis. The result (Fig. lb) showed that CK 2 was among the prominent translation products of this tissue’s RNA and was exceeded among the type II CKs only by CKs 1 and 5. When the cDNA clone pEK2TYR-2AC8 (see Materials and Methods) was used for mRNA enrichment in combined hybrid selection-in vitro translation experiments, a single mRNA product was selected (Fig. lc) in the CK 2 position, with some minor isoelectric variants indicative of modifications during and after translation. In vitro transcription of clone pEK2, followed by in vitro

HUMAN

EPIDERMAL

CYTCKERATIN

2

135

FIGI. Expression of human epidermal cytokeratin 2 (CK 2e) in human thigh epidermis (a, b) and identification of the cDNA clone pEK2 as coding for CK 2e by hybrid selection/translation (c) and in uitro transcription/translation (d). (a) Coomassie blue-stained cytoskeletal polypeptides from thigh epidermis separated by 2D GE (horizontal arrow: direction of first-dimension NEPHGE; vertical arrow, SDS-PAGE in the second dimension). Reference poiypeptides used for coelectrophoresis are B, bovine serum albumin; P, yeast phospboglycerokinase; A, a-actin. (b) Autoradiography of a gel similar to that in (a), revealing [?S]methionine-labeled products of in vitro translation of total RNA from thigh epidermis after coelectrophoresis with unlabeled cytoskeletal proteins from human epidermis for identification. (c) Autoradiograph showing the [%]methionine-labeled product of in vitro translation of mRNA selected by hybridization of 200 gg total RNA from thigh epidermis to the 3’-end-specific cDNA subclone pEK2TYR-2AC8. Note that CK 2e is a genuine translation product (b) and that only the mRNA coding for CK 2e was specifically bound to clone pEK2TYR-2AC8 (c). The arrowhead denotes the major isoelectric variant spot of CK 2e, probably post-translationally modified. (d) Autoradiograph of [?3]methionine-labeled products obtained after in vitro transcription and translation of clone pEK2 coelectrophoresed with cytoskeletal proteins from thigh epidermis to show that the encoded polypeptide comigrates with authentic CK 2e.

translation of the resulting RNA (Fig. Id), also yielded a polypeptide with an isoelectric variant series comigrating with authentic CK 2 in 2D GE. Sequence Features

of CK 2 and Its mRNA

The complete nucleotide sequence of clone pEK2 is shown in Fig. 2, together with the encoded amino acid sequence. The clone comprises 2427 nucleotides, including the polyadenylation signal sequence AATAAA.

As the originally obtained IX ceeded clone pEK2 at the 3’ en another ATTCAG, followed by the poly(A) primer, assume that the total mRNA comprises 2433 nucleotides plus the poly(A) tail? i.e.? probably exceeding 2.5 kb. The first possible initiation codon starts at position 34. The homology of the resulting amino acid sequence at the amino terminus to the sequences of other CKs indicated that this is the act in used. The start total polypeptide encoded by K2 co ns 645 amino

136

COLLIN

1 1

ET

AL.

111 21

bGCCTGTGACTTTCCTCCCTGGACAAAGGCATCATGAGTTGTCAGATCTCTTGCAAATCTCGAGGAAGAGGA~AGGTGGAGGAGGATTCCGGGGCTTCAGCAGCGGCTc GGGGGGFRGFSSGS MSCQISCKSRGR + AGCTGTGGTGTCTGGT~AAGCC~AGATCAACTTCCAGCTTCTCCTGCTTGAGCCGCCATGGTGGTGGT~CGGGGGCTTCGGTGGAGGCGGCTTTGGCAGTCGGAGTC AVVSGGSRRSTSSFSCLSRHGGGGGGFGGGGFGSRSL

221 64

TTGTTGGCCTTGGAGGGACCAAGAGCATCTCCATTAGTGTGGC~GGAGGAGGTGGTGGCTTTGGCGCCGCTGGTGGATTTGGTGGCAGAGGAGGTGGTTTTGGAGGCGGC VGLGGTKSIS ISVAGGGGGFGAAGGFGGRGGGFGGG

331 100

AGCG~TTTGGAGGCGGCAGCGGCTTTGGAGGTG~AGCGGCTTCAGTGGTGGTGGTTTCGGTGGAGGCGGCTTTGGTGGAG~CGCTTTGGAGGTTTTGGGGGCCCTGG SGFGGGSGFGGGSG FSGGGFGGGGFGGGRFGGFGGPG

441 137

TGGTGTTGGAGGTT'rAGGGGGTCCTGGTGGcTTT~GCCTGGAGGATACCCTGGTGGCATCCACGAAGTCTCTGTCMCCAGAGCCTCCTGCAGCCTCTC~CGTGMhG GVGGLGGPGCFGPGG YPGGIHEVSVNQS LLQPLNVKV

551 114

TTGACCCAGAGATCCAGAATGTG~GGCCCAAGAGCGTGAGCAGATCAMACTCTCAACAACAAATTTGCCTCCTTCATTGACAAGGTGCGGTTCTTGGAGCAGCAGAAC DPEIQNVKAQEREQ IKTLNNKFASFIDKVRFLEQQN . . CAGGTGTTACAGACCAMTGGGAGCTGCTACAAC~ATGAATGTTGGCACCC~CCCATCAACCTGGAGCCCATCTTCCAGGGGTATATCGACAGCCTCAAGAGATATCT IDSLKRYL QVLQTKWELLQQMNVGTRP INLEPIFQGY

661 210 771 247 881 284

GGATGGGCTCACTGCAGAAAGAAC.~TCACAGAATTCAGAGCTGAATMCATGCAGGATCTTGTGGAGGATTATAAG~GAAGTATGAGGATGAAATCAAT~GCGCACAG LNNMQDLVEDYKKKYEDEINKRTA DGLTAERTSQNSE * C~GCTGAGAATGATTTTGTGACGCTTAA~AGGACGTGGACAATGCCTACATGATAAAGGTGGAGTTGCAGTCCAAGGTGGACCTGCTGAACCAGGAAATTGAGTTTCTG AYMIKVELQSKVDLLNQEIEFL AENDFVTLKKDVDN

991 320

~AGTTCTCTATGATGCGGAGATATCCCAGATACATCAGAGTGTCACTGACACCAACGTCATCCTCTCCATGGACAACAGCCGCAACCTGGACTTGGATAGCATCATCGC LSNDNSRNLDLDSIIA KVLYDAEISQIHQSVT5TNVI

1101 357 1211 394

CGAGGTCAAGGCCCAGTATGAGGAGATCGCCCAGAGGAGC~GGAAGAAGCGGAGGCCCTGTACCACAGC~GTATGAGGAGCTCCAGGTGACTGTCGGGAGACATGGAG EVKAQYEEIAQR SKEEAEALYHSKYEELQVTVGRHGD . * * ACAGCCTGAAAGAGATCAAGATAl~AGATCAGCGAGCTGAACCGCGTGATCCAGAGGCTGCAGGGGGAGATCGCACATGTGAAGAAGCAGTGTAAGAATGTGCAAGATGCC SLKEIKIEISELNRV IQRLQGEIAHV KKQCKNVQDA

1321 430

ATCGCAGATGCCGAGCAGCGTG~GAGCATGCCCTCAAGGATGCCA~AACAAGTTGAATGACCTGGAGGAGGCCCTGCAGCAGGCCAAGGAGGACTTGGCGCGGCTGCT IADAEQRGEKALKDAR NKLNDLEEALQQAKEDLARLL

1431 461

GCGTGACTACCAGGAGCTGATG~CGTGAAGCTGGCCCTAGATGTGGAGATCGCCACCTACCGC~ACTGCTGGAG~CGAGGAGTGCAGGATGTCTGGAGACCTCAGCA RDYQELMNVKLAL DVEIATYRKLLEGEECRMSGDLSS

1541 504

GCAATGTGAC~TGTCTGTGAC~GCAGCACCATTTCATC~ATGT~CATCCAAGGCTGCCTTTGGAGGTTCTGGAGGTAGAGGGTCCAGTTCCGGAGGAGGATACAGC NVTVSVTSSTISSNVASKAAFGGSGGRGSSSGGGYS

1651 540

TCTGGAAGCAGCAGTTATGGCTCTGGAGGCC~ACAGTCTG~TCCAGAGGCGGTAGTGGAGGAGGAGGTTCTATCTCTGGAGGAGGATATGGCTCTGGCGGTGGTTCTGG SGSSSYGSGGRQSGSRGGSGGGGSISGGGYGSGGGSG

1761 517

AGGAAGATAC~ATCT~TGGTffiCTCT~GGGA~GTCCATCTCT~AGGA~ATAT~CTCTGGAGGT~AAAACACAGCTCTGGAGGTGGCTCTAGA~AGGCTCCA GRYGSGGGSKGGSISGGGYGSGGGKHSSGGGSRGGSS

1871 614 1981

G~TCTGGAGGAGGATATGGCTCTGGAGG~GGGGTTCTAGCTCTGT~GGGTAGCTCAGGTG~GCTTTTGGTTCCAGCGTGACCTTCTCTTTTAGATA~GATGAGCC SGGGYGSGGGGSSSVKGSSGEAFGSSVTFSFIt* . CCCACCACCACCGACTCTCCCAACCCAGACTCTCCCACTCCAGAATGTAG~GCCTGTCTCTGTACCTCTMCTGGCAGCAAGTTAAATTTTTGTcATTTATCTcTGATG

2091

GCACTTTGAG~GMTGTCCACATACAGTTTTTG~AGATCTTCTCTCCA~CCAGTTAGTTAGAGCCAGTGACGCCTCTGTGTTCTGGGGCGGAATCTGTGCTGT

.

.

2201 2311 2421

ATTTGCA

FIG. 2. Complete (one-letter code). The

2427 nucleotide stop codon

secluence of cDNA clone pED2 coding for human is designated by an asterisk and the polyadenylation

epidermal signal

CK 2 and the deduced is underlined.

amino

acid sequence

HUMAN 10

20

EPIDERMAL

30

40

CYTQKERATIN 50

60

H Kl H K2e H K3

~RQ-------FSSGSG-Y~SGGGFSFGFAGIINYQRRTTSSST~~~SSSGGGGGSFGAG~FGSRSLAGS~SIASIS-------------~CQISCKSR~RGG~F~-~SSGSAWSGGS~STSSFSCLS~~GFGG~GS~LVGL~T~ISX ~RQ------ASKTS~SQ-~§GRSAWSGSS~CVAHSGGAGGGAYGFRSGA~G~LYNL~~~S~~~SRAGGFGGGRSS~AFAGGY

H Kl H K2 H K3

GARGG-GGGS~~YGG~FGGGGFG~-~-G~I~-~~FGSGG~GGGGFG~GGYG~Y~VCS~~IQE~INQ~~

137

2 70

80

90

m(=&FGA------

100

---..--

--

IDPEXQ

78 85 93

174

G~GSGYG~~FGGGRGMGGGF~AG~GGA~~A~~P-GGF~SGGF~PGS~SP~~~F~~IQE~T CZ 1A 9

H Kl H K2e H K3

272 276 292

NVKAQEWQIK’l!LN?IRFASFIDKVRJ?I.EQQNQVLQTKUE LWVGTE-PINLEPIFQGYIDSLKRYLDGLTAEDTSQNSEL~DL NILGmGRLDSEL QVKAQEXSQIFX~itSF~D~~QQNKVLETKWNL~GTSSISGTNNLEPLRENHINYLRS C 1B .+-I

H Ki H K2e H K3

~~~T~ND~~~NA~I~LQS~LLN~IE~LKVL~~ISQIHQS-VTD~~L~NS~~~SII~~~~I EXNRBTAAENEFVTLlGtDVDSAYMNlCVE LQAKVDALIDEIDFLRTLYDAELSQM-QSHISDTSVVLSXDmS

E Kl H K2e H K3

LYQS~~iTA~GDSVRNSKIgIS~~~~RSgID~ISNLQQSISD~~N~~~D LYHSKYUZQVTVGBBGDSLKEIKIEISELNRVIQRLQGEIAHV?XQCKNVQDAIADAEQRGEHALKDARBKLNDLE~D LYQTKLG~LQTTAGXUiGDDLRNTKS~~IELNRMIQRLRABIEGVRKQNANLQTAIAQAJ3QH GEMXXDANAXLQELQAAL

H Kl H K2e H K3

~S~G~CAF~S~STSH~SISGGGSRG~~GGYGSGG~S--YGSGGG~YGs~GGGG~~SY~S ~~~T~~~-C~GDLSS~VS~SSTISS~ASK~F~S~RGSSS~SSGSSSYGS~QSGS~SG ~~~~~~YS~GECPSAVSISWSSSTTSASAGGYGGGYG~MGGGL~FSAGGGGGIGF~GG6GGIGG~

H Kl H K2e 3 K3

~G~HGSYGSGSSSGGYRGGS~GGGGSSGGRGS~SSGGSS~RGSSS~VKSSGGSSSVXFVSTTYSGV--T~ ~S~--~YGS~SK~SISG~YGSGGGKHSSG~SR~GSSS~GYGSG~GSSSVKGSSGE~GSSVTFS--FR SISGA---~~~S~GFSSASNR~-------------------------------------sIKFSQssQSSQRySR

FIG. 3* Amino acid sequence comparison of human epidermal maximal alignment. deletions (hyphens) have been introduced. subdomains: coils IA, ll$ and 2. Residues are identical in HKZe

* * *

YQNW

411 475 431

GiYGSGGSSY&

568

644 645 629

CK 2 (HK2e) and human CKs 1 (HK1 [64,65]) and 3 (HK3 [66]). To obtain The arrows indicate the a-helical rod domain, which can be divided into and at least one of the other two sequences is printed in boldface Ietters.

acids with a calculated MT of 65,861, which makes CK 2 the largest human CK known, exceeding CK 1 (Mr 65,600) by one amino acid although CK 1 is clearly slower in AGE (apparent Mr -68 kDa; determined MT Comparison of the sequence of epidermal CK 2 with those of other CKs (Figs 3) showed that it is a characteristic type II CK (cf~ [9, l2-14j) with a particularly high sequence homology to human CK 1, a hallmark of suprabasal differentiation of epidermal keratinocytes (cf. [24,25,28,29,63-65])> CK 3, which is abundant in-and characteristic for-the cornea1 epithelium of the eye ([6, 6l> 661; for bovine CKs see [19]), and CK 5, the major type II CK of basal cell layer keratinocytes (for references see the Introduction). Like several other CKs typically synthesized in stratified epitheha, the human epidermal CK 2 exhibits an exceptionally high glycine content in both the head and the tail domain (F’ig. 3)> including many repeats of the

basic type (G)GGF in the head and (G)GGY? GGGS, or GGR(K) in the tail domain. In the ro domain, known for its high sequence conservation in diRerent IF proteins during evolution (e.g., [l3, l4> 67-69j)> epidermal CK 2 shows 72% amino acid sequence identity with human CKs 1,3, and 5. Southern blot analysis of human genomic onstrated few and specific restriction fragment bands hybridizing with the specific CK 2 hybridization probe, i.e., a riboprobe corresponding to the daeletion subclone pEK2TYR-2AC8 (Fig. 4). This suggests that this CK exists in a single gene copy or in alleIes wit.h extremely low sequence diversification.

When RNA from different lmman tissues and cultured cell lines was examined by Northern blot hybridization analysis (Fig. 5) using the same 3’sublarobe (see

138

COLLIN

ET

AL.

FIG. 4. Southern blot hybridization of human genomic DNA (20 wg per lane) from placenta digested with EcoRI (lane l), Hind111 (lane 2), and BumHI (lane 3), separated by electrophoresis in a 0.9% agarose gel, transferred to “N-bond” membrane, and hybridized to a radioactively labeled riboprobe derived from clone pEK2TYR-2AC8 specific for CK 2e. In each digest, only a single band of equal intensity is detected. Positions of DNA size standards (Hind111 digest of phage A DNA) in kilobase pairs are indicated in the right margin.

above) of epidermal CK 2 (CK 2e), strongly reactive bands were seen in a position corresponding to a mRNA size of about 2.6 kb present in epidermal tissues from such diverse sites as foot sole and thigh. Moreover, a weak but still significant Northern blot band was observed in some but not all samples of hard palate (Fig. 5b) and gingiva (not shown) whereas several other strat1

a

23

4

512

a’

34

567

b

FIG. 5. Northern blot analysis of total RNA from human tissues or epithehal cell lines hybridized to probes for CK 2e (a, b) or CK 1 (a’). Lane 1, foot sole epidermis (10 +g); lane 2, vulvar carcinoma cell line A431 (20 pg); lanes 3 and 7, thigh epidermis (2 and 4 pg, respectively); lane 4, keratinocyte cell line HaCaT (20 gg); lane 5, exocervix (10 Kg); lane 6, hard palate (4 pg). Total RNA was electrophoretically fractionated in a 0.9% agarose gel containing formaldehyde and transferred to N-bond membrane. Filters were hybridized with subclone pEK2TYR-2AC8, representing the 3’-untranslated region (nucleotides 2132-2391) of pEK2, labeled by random priming (a) or by reverse transcription in the presence of [3zP]CTP (b). (a’) Same filter as that in (a), but hybridized with the antisense riboprobe specific for CK 1 used as control. Using rRNAs as markers, the estimated size of the mRNAs coding for CK 1 and CK 2e is about 2.6 kb.

FIG. 6. Autoradiomicrographs of frozen sections of epidermis, showing the result of in situ hybridization CTP-labeled antisense RNA specific for CK 1 (a, a’) or mRNAs. (a, b’, bright field; a’, b, dark field). Bars, 30 pm v tb, W,

human thigh with [w?S]CK 2e (b, b’) (a, a’) and 12

ified epithelia such as exocervix (Figs. 5a and 5a’) and cell lines derived from stratified epithelia, including epidermis, such as the vulvar carcinoma-derived line A431 and the back-skin-derived keratinocyte line HaCaT (Fig. 5a), were consistently negative. In the epidermal tissue, CK 2e is expressed clearly suprabasally, as demonstrated by in situ hybridization (Fig. 6). Comparison with CK 1 mRNA (Figs. 6a and 6a’), which can be detected already in the first suprabasal layer and in many places in certain basal cells (see [30] and references in the Introduction), revealed significant levels of mRNA encoding CK 2e in the third layer, i.e., maturing stratum spinosum cells, and in more distal strata (Figs. 6b and 6b’). A corresponding result was obtained by immunohistochemistry using antibodies specific for CK 2e, in comparison with other antibodies against other CKs, including monoclonal antibody AE5, which reacts with cornea1 CK 3 as well as with a specific type II CK polypeptide present in gingiva and palate (not shown). Fig-

HIJMAN

EPIDERMAL

CYTQKERATIN

2

139

All nonepithelial tissues as well as simple epithelia, urothelium, and other squamous epitbelia were negative for epidermal-type CK 2e, including cornea, esophagus, tongue? mucosa, and vagina (data not &own), except for occasional, though inconsistent5 reactions in gingival and palatal epithelium (for cytoskeletal characteristics of these tissues see [42,43])* Likewise, samples of bronchial metaplasia (cf. ]70]) an a diversity of squamous cell carcinomas (cf. [6> 411) were negative. Such negative tissues included some samples in wbicb epidermal CKs 1 and/or 10 were detected.

FIG. 7. Micrographs of frozen sections, showing immunocytochemical results (immunoperoxidase technique) obtained with antibodies specific for CK Ze (a, e-f) on human cornea (a), hairy skin from the ventral side of the thigh (c, d), and foot sole skin (e, f). Note that the antibodies specific for CK 2e are negative on cornea (a; b shows a positive control reaction with CK 3 antibody AE5), but intensely react with cells of the suprabasal layers of epidermis from both locations. S, stroma; D, dermis; SD, sweat gland duct near transition to acrosyringium; E, epithelium. Magnifications: 19O:l (a, b), 18O:l (c, d), 1OO:l (e), and 12O:l (f).

ure 7a shows that anti-CK 2e is negative on human cornea, which reversely is strongly immunostained with AE5 (Fig. 7b). I-lowever, CK 2e antibodies react specifically and strongly with suprabasal layers of epidermis from various sites and different individuals, including hairy body skin such as from thigh (Figs. 7c and 7d), knee, and groin (all not shown) as well as foot sole epidermis (Figs. 7e and 7f). Here, the reaction was usually seen in the third or fourth cell layer, i.e., in the distal spinous layer and in part of the granular layer. Positive, although sparse and heterogeneous, immunoreactions restricted to certain cells were seen in skin from axillae, penile shaft, and breast nipple. Epidermal stratum corneum always appeared negative, but we could not decide whether this reflects the absence of the protein or, more likely, its masking in this layer due to the abundance of dense horny material in the dead cells forming squames. ln palmar and plantar epidermis (e.g., Figs. 7e and 7f), we also observed a weak CK 2e immunoreaction in occasional cells of the first suprabasal layer. Comparison with immunostaining ofthe type I CK 9 (cf. [31,32]) &owed the latter often to occur somewhat “earlier”, Le., one layer below.

The results of this study will u~do~b~ed~,y end the long-lasting dispute over the significance of the epiderma1 CK 2. We have shown that it is a genuine polypeptide with a characteristic type 11 CK sequence, encoded by a unique gene. We have no satisfactory explanations for the failure of several other groups to d tein in their preparations of skin material observation by Tyner and Fuchs [34] ~who nize this polypeptide among the Gz G&o translationai products of mRNAs isolated from human foreskin is now readily explained by our finding that this is the only type of epidermis found so far which appears to be totally devoid of CK 2e (for other dih!eren plement between foreskin epidermis from other body sites see, e.g.? [3$, 45]). 2e is very difficult to extract in S-containing buffers probably due to its association h epiderrnd differentiation and “hardening” (cor~ificatio~)~ While it is now clear that in ~~ere~ti~~ng epidermal keratinocytes from diverse body 2 is a distinct cytoskeletal polypeptide, it is als that this mol ecule is subject to a series of pas lational modifications, as it uwa&y appears in a series of isoelectric variants, at least some of w&h might represent products of phosphorylation Like other type 11 CKs, notably CK I, CK 2e is also subject to partial degradation by endogenous proteases, and minimizing such degradation during preparation may ah be crucial to preserve detectable amounts of the intact, relatively large polypeptide chain. The amino acid sequence of shows that the relationship of this molecule w 1 is not closer (72% identity in the rod domain) than that with CK 3> CK 5, or a gingival p te CK of similar size (C~ Collin J.-P. Ouhayoun, and We Franke, manusc aration). ~Tbis, as we as our results wit Southern blot studies7 seems to exclude th that CK 2e is allelic representative of genetic polymorphism of 1 (cf. [38, 711)~ The sequence of human CK 2e displays remarkable homologyY in parts of the molecule, to the sequence reported for a partial cDNA clone byb~~~~~i~gto an -2.8-

140

COLLIN

kb epidermal mRNA and encoding an as yet uncharacterized murine epidermal type II CK of -70 kDa [72]. While this homology, and the restriction of the expression of the murine CK to certain body sites and late stages of keratinocyte differentiation, could suggest that the murine epidermal 70-kDa CK is an ortholog polypeptide of human CK 2, other features indicate that the two large type II CKs differ biologically. The sequence of the carboxy-terminal end of the tail domain is completely different in the murine polypeptide (LGQGKVVAQV) and in the corresponding terminal decapeptide FGSSVTFSFR of human CK 2e. Moreover, the mRNA encoding the murine -7O-kDa CK is not restricted to suprabasal differentiation as it has also been localized to the basal cell layer of epidermis from various body sites. Also, human CK 2e is not a major component of fetal skin epidermis but is readily seen in skin of the newborn baby [44], in contrast to the mouse CK protein of -70 kDa, which has been detected only later in the postnatal development of this laboratory animal. However, characterization of the complete cDNA sequence and studies of the regulation of the expression of the corresponding murine and human genes are necessary to decide the evolutionary and functional relationship of the largest human and murine type II CKs known so far. Our in S&U hybridization and immunolocalization data have revealed a very restricted tissue pattern of expression of CK 2e, which is essentially specific for epidermis of various, but not all, body sites and always occurs in upper cell strata that represent advanced stages of keratinocyte differentiation. It is definitely a later product than CK 1. Only proteins such as filaggrin that are exclusively expressed in the granular layer are induced even later in keratinocyte maturation and death (for review see [73]). So far we have not learned of any treatment which might induce the synthesis of epidermal CK 2 in keratinocytes growing in culture. With the availability of DNA probes and antibodies specific for this protein, as reported in this article, more direct studies of the possible function and the regulation of expression of this human cytoskeletal protein are now possible, including transfections of the gene into cultured cells or into transgenic animals, be it the cDNA form (this study) or the recently isolated corresponding genomic clone (C. Collin and W. W. Franke, unpublished results). We thank Drs. B. L. Bader (this Institute) and T. M. Magin (University of Edinburgh) for good advice and helpful discussions, R. Zimbelmann (this Institute) for the sequencing work, and E. Gundel for the typing.

REFERENCES 1. 2.

Franke, W. W., Schmid, E., Osborn, M., and Weber, K. (1978) Proc. Natl. Acad. Sci. USA 75,.5034-5038. Franke, W. W., Weber, K., Osborn, M., Schmid, E., and Freudenstein, C. (1978) Exp. Cell Res. 116, 429-445.

ET

4L.

3.

Franke, Osborn,

W. W., Appelhans, B., Schmid, E., Freudenstein, C., M., and Weber, K. (1979) Eur. J. CeZl Biol. 19,255-268.

4.

Sun,

5.

Sun, T.-T., 83, 1.

6.

Moll, R., Franke, W. W., Schiller, R. (1982) CeZZ 31, 11-24.

7.

Mall,

T.-T.,

and Green, Shih,

R., Schiller,

H. (1978)

C. H., and

Cell 14,469-476.

Green,

H.

(1979)

D. L., Geiger,

D. L., and Franke,

Cell. Immunol. B., and Kepler,

W. W. (1990)

J. CelZ Biol.

111,567~580. 8.

Cooper, 243-256.

9.

Fuchs, E., Tyner, A. L., Giudice, G. J., Marchuk, D., RayChaudhury, A., and Rosenberg, M. (1987) Curr. Top. Deu. BioZ. 22, 5-34.

10. 11.

D., Schermer,

Heid, H. W., Moll, 37,137-157. Schiller,

A., and Sun, T.-T.

I., and Franke,

D. L., Franke,

(1985)

Lab. Inuest.

W. W. (1988)

W. W., and Geiger,

52,

Differentiation

B. (1982)

EMBO

J.

and

Parry,

1,

761-769. 12. 13. 14.

15.

Crewther, W. G., Dowling, L. M., Steinert, P. M., D.A. D. (1983) Int. J. Biol. Macromol. 5, 267-275. Hanukoglu, I., and Fuchs, E. (1983) CelZ 33,915-924.

Weber, K., and Geisler, N. (1984) in Cancer Cells. 1. The Transformed Phenotype (Levine, A. J., Van de Woude, G. F., Topp, W. C., and Watson, J. D., Eds.), pp. 153-159, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Gruen, L. C., and Woods, E. F. (1983) Biochem. J. 209,587-598.

16.

Quinlan, Franke,

17.

Fraser, R. D. B., MacRae, T. P., Parry, (1986) Proc. Natl. Acad. Sci. USA 83, Steinert, P. M., and Roop, D. R. (1988) 593-625.

18.

R. A., Cohlberg, J. A., Schiller, D. L., Hatzfeid, W. W. (1984) J. MoZ. Biol. 178, 365-388.

M., and

D. A. D., and Suzuki, E. 117991183. Annu. Reu. Biochem. 57,

19.

Franke, W. W., Schiller, D. L., Mall, R., Winter, S., Schmid, E., Engelbrecht, I., Denk, H., Krepler, R., and Platter, B. (1981) J. Mol. Biol. 153, 933-959.

20.

Hatzfeld, M., and Franke, 1841. Eichner, R., Sun, T.-T., 1767-1777.

21.

W. W. (1985) and Aebi,

J Cell Biol.

U. (1986)

101, 1826-

J. Cell Biol.

102,

22.

Sun, T.-T., and Weiss, Phenotype Watson, J. tory Press,

23.

Quinlan, R. A., Schiller, D. L., Hatzfeld, M., Achtstatter, T., Mall, R., Jorcano, J. L., Magin, T. M., and Franke, W. W. (1985) Ann. N. Y. Acad. Sci. 455, 282-306.

24.

Fuchs,

25. 26.

Fuchs, E., and Green, H. (1981) CeZl25,617-625. Woodcock-Mitchell, J., Eichner, R., Nelson, W. G., and Sun, T.-T. (1982) J. Cell Biol. 95, 580-588. Schweizer, J., Kinjo, M., Fiirstenberger, G., and Winter, H. (1984) Cell 37, 1599170. Bowden, P. E., Stark, H. J., Breitkreutz, D., and Fusenig, N. E. (1987) Curr. Top. Deu. Biol. 22, 35-68.

27. 28.

Eichner, R., Schermer, A., Cooper, D., Nelson, J. G., R. A. (1984) in Cancer Cells. 1. The Transformed (Levine, A. J., Van de Woude, G. F., Topp, W. C., and D., Eds.), pp. 1699176, Cold Spring Harbor LaboraCold Spring Harbor, NY.

E., and Green,

H. (1980)

Cell

19, 1033-1042.

29.

O’Guin, W. M., Galvin, S., Schermer, Cm-r. Top. Dev. Biol. 22, 97-125.

A., and Sun,

30.

Roop, D. R., Krieg, S. H. (1988) Cancer

31.

Knapp, A. C., Franke, W. W., Heid, H., Hatzfeld, M., Jorcano, J. L., and Mall, R. (1986) J. Cell Biol. 103, 657-667.

T. M., Mehrel, T., Cheng, Res. 48, 3245-3252.

T.-T.

(1987)

C. K., and Yuspa,

HUMAN W. W., and Mall,

EPIDERMAL

32.

Mall, I., Heid, H., Franke, tiution 33, 254-265.

33.

Mall, R., Franke, W. W., VolePlatzer, B., and Kepler, R. (1982) J, Cell BioL 95? 285-295. Tyner, A. L., and Fuchs, E. (1986) J, Cell %oZ. 103,1945-1955. G., and Fuchs,

R. (1987)

CYTOKERATIN

E. (1987)

Difieren-

Kopan, 427-440.

R., Traska,

J, Cell Viol.

Kopan, Staler,

R., and Fuchs, E. (1989) J, CelZ Bioi. 109, 2955307. A., Kopan, R., Duvic, M., and Fuchs, E. (1988) J. Cell

Jorcano, J. L., Franz9 t&ion 28, 155-163.

54.

Frohman, M. A., Dush, M. IS., and Martin, G. R. (1988) &oc. NatZ. Acad. Sci. USA $5,8998-9002. Jorcano, J. L., Magin, ‘I’. M., and Frank@, W, W. (1984) J. MO!. Eiol. 1’76, 21-37. Sambrook, J., Fritsch, E. F., and Maniatis T. (1989) Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor NY.

55. 56.

J. K., and Franks,

W. W. (1984)

Differen-

57.

Leube, R. E., Bosch, F. X., Romano, V., Zimbelmann, R., Hofler, H., and Franke, W. W. (1986) Di,#erentiution 33? 69-85.

58.

Cox, K. II., De Leon, D. V., Angerer, (1984) Deu. Biol. 101,485-5Q2.

59. 60.

Franke, Cowin, Hynes,

Res.

61.

Ouhayoun, J.-P., Gosselin, F., Forest, N., Winter, S., and Franke, W. W. (1985) Diferentiation 30, 1233129. Bosch, F. X., Ouhayoun, J‘-P., Bader, B. L., Collin, C., Grund, G., Lee, I., and Franke, W. W. (1989) Virchom Arch. B 58,5977.

62.

Schermer, A., Galvin, 49-62. Hatzfeld, M., Maier, 197, 237-255.

Boukamp, P., Breitkreutz, D., Stark, H.-J., and Fusenig, N. (1990) Diferentiatiorz 44, 150-161. Van Muijen, G. N. P., Warnaar, S. O., and Ponec, M. (1987) E.xp. Cell Res. 171, 331-345. Bowden P. E., Quinlan, R. A., Breitkreutz, N. E. (1984) Eur. J. Biochem l423 29-36. Moll, 276,

R., Mall, 349-363.

I., and Franke,

D.,

W. W. (1984)

Arch.

and

Fusenig,

Dermutol.

Moll, R., Mall; I., and Wiest, W. (1983) Differentiution 23, 170178. Kuruc, N., Leube, R. L., Mall, I., Bader, B. L., and Franke, W. W. (1989) Differentiation 42, 111-123. Achtstatter, T., Hatzfeld, M., Quinlan, R. A., Parmelee, D. C., and Franke, W. W. (1986) in Methods in Enzymology (Vallee, R. B., Ed.), Vol. 134, pp. 355-371, Academic Press, San Diego. Bader, B. L., Jahn, BioL 47, 300-319.

48.

53.

105,

Biol. lQ'7,427-446.

46.

141

2

Kapprell,

L., and

H. P., Owaribe,

Franke,

W. W. (1988)

K., and Franke,

Eur.

J. Cell

W. W. (1988)

J. CelZ

L.

., and Angerer,

W. W., and Mail, R. (1987) ~@erent&on, 36? 145-163. P., Kappreil, H.-P., Franke, W. W.> Tamkun, J., and R. 0. (1986) Cell 46,1063-1073. S., and Sun, T.-T. G., and Franke,

(1986)

J. CeLi Viol.

W. W. (1987)

Roop, (1983)

64.

Johnson, Steinert,

65.

Steinert, P. M., Parry, D. A. D., Idler, W. W., Johnson, L. D., Steven, A. C., and Roop, D. R. (1985) J. Viol. Chem. 260, i’l427149. Klinge, E. M., Sylvestre, Y. R., Freedberg, I. IM. and Blumenberg, M. (1987) J. MoZ. Euol. 24> 319-329~

66. 61.

D. R., Hawley-Nelson, P., Cheng, C~ K., andYuspa, Proc. N&l. Acad. Sci. USA l30* 716-720.

68.

Geisler, N., and Weber, K. (1981) Proc. N&l. Acad. Sci. USA 4120-4123. Hanukoglu, I., and Fuchs, E. (1982) CelZ 31$ 243-252. Blumenberg,

Kreis, T. E., Geiger9 B., Schmid, E., Jorcano, J. L., and Franke, W. W. (1980) Cell 32, 1125-1137. Fabricant, R. N., Delarco, J. E., and Todaro, G. (1977) Proc. Natl. Acad. Sci. USA 74y 565-569.

71.

Leube, R. E., and Rustad, T. J. (1991) %,rchozs Arch. 221-253. Mischke, D., and Wild, G. (1987) J, Inuest. Der.matol. 197.

Magin, T. M., Jorcano, 29 1387-1392. Received

April

28, 1992

J. L., and Franke,

W. W. (1983)

72.

J.,

761-

E&Q?OJ.

S. H.

L. D., Idler, W. W., Zhou, X.-M, Roop, D. R., and P. M. (1985) Proc. N&l. Acad. Sci. U§A 82,1896-1900.

70.

R. T., Breitkreutz, D., Hornung, N. E. (1988) J. Cell Biol. 106,

&ol.

63.

69.

P., Petrussevska, A.> and Fusenig,

IQ39

J. Mol.

Biol. 106,1679-1691.

Boukamp, Markham, 771.

R. C.

73.

M.

(1988)

J. MoZ.

&ol.

‘?‘$?

2+7,203-211. B 61,

Rentrop, M., Niscbt, R., Knapp, B., Sehweizer> J., and Winter> H. (1987) Diflerentiation 34g 189-200. Dale, B. A., Resing, K. A., and Haydock, P. V. (1990) in Cellular and Molecular Biology of Intermediate Filaments (Goldman R. D., and Steiner& P. M., Eds.), pp. 393-412, Plenum Fress, New York.