KERATIN SYNTHESIS DURING DEVELOPMENT OF THE EMBRYONIC CHICK FEATHER

The synthesis of keratin proteins during development of the embryonic chick feather was studied by quantitative gel electrophoresis of the reduced and carboxymethylated proteins. The results demonstrated a coordinated synthesis of the major keratin proteins, during and after the onset of keratin synthesis. The results from gel electrophoresis correlated well with electron microscope visualization or keratin fibrils in the developing feathers. Autoradiography at the electron microscope level indicated that the feather cells lose the ability to synthesize DNA before keratin synthesis begins, but retain the ability to synthesize RNA after keratin synthesis begins.     

Effects of keratin filament disruption on exocrine pancreas-stimulated secretion and susceptibility to injury

Disruption or absence of hepatocyte keratins 8 and 18 is associated with chronic hepatitis, marked hepatocyte fragility, and a significant predisposition to stress-induced liver injury. In contrast, pancreatic keratin disruption in transgenic mice that express keratin 18 Arg89 --> Cys (K18C) is not associated with an obvious pancreatic pathology. We compared the effects of keratin filament disruption on pancreatic acini or acinar cell viability, and on cholecystokinin (CCK)-stimulated secretion, in transgenic mice that overexpress wild-type keratin 18 and harbor normal extended keratin filaments (TG2) and K18C mice. We also compared the response of these mice to pancreatitis induced by a choline-deficient ethionine-supplemented diet or by caerulein. Despite extensive cytoplasmic keratin filament disruption, the apicolateral keratin filament bundles appear intact in the acinar pancreas of K18C mice, as determined ultrastructurally and by light microscopy. No significant pancreatitis-associated histologic, serologic, or F-actin/keratin apicolateral redistribution differences were noted between TG2 and K18C mice. Acinar cell viability and yield after collagenase digestion were lower in K18C than in TG2 mice, but the yields of intact acini and their (125)I-CCK uptake and responses to CCK-stimulated secretion were similar. Our results indicate that keratin filament reorganization is a normal physiologic response to pancreatic cell injury, but an intact keratin cytoplasmic filament network is not as essential in protection from cell injury as in the liver. These findings raise the possibility that the abundant apicolateral acinar keratin filaments, which are not as evident in hepatocytes, may play the cytoprotective role that is seen in liver and other tissues. Alternatively, identical keratins may function differently in different tissues.     

Morphology, behavior, and interaction of cultured epithelial cells after the antibody-induced disruption of keratin filament organization

The organization of intermediate filaments in cultured epithelial cells was rapidly and radically affected by intracellularly injected monoclonal antikeratin filament antibodies. Different antibodies had different effects, ranging from an apparent splaying apart of keratin filament bundles to the complete disruption of the keratin filament network. Antibodies were detectable within cells for more than four days after injection. The antibody-induced disruption of keratin filament organization had no light-microscopically discernible effect on microfilament or microtubule organization, cellular morphology, mitosis, the integrity of epithelial sheets, mitotic rate, or cellular reintegration after mitosis. Cell-to-cell adhesion junctions survived keratin filament disruption. However, antibody injected into a keratinocyte-derived cell line, rich in desmosomes, brought on a superfasciculation of keratin filament bundles, which appeared to pull desmosomal junctions together, suggesting that desmosomes can move in the plane of the plasma membrane and may only be 'fixed' by their anchoring to the cytoplasmic filament network. Our observations suggest that keratin filaments are not involved in the establishment or maintenance of cell shape in cultured cells.     

Genomic organization and amplification of the human keratin 15 and keratin 19 genes

Keratin intermediate filaments are the major components of the cytoskeleton in epithelial cells. Mutations in keratin genes have been documented in many disorders of the skin, nails, hair, and mucous membranes. Although no mutations have been described in either keratin 15 or keratin 19, they are good candidates for other as yet uncharacterized genetic disorders of keratinization, particularly as the skin, nails, hair, and conjunctiva are sites of keratin 15 and 19 expression. To facilitate future mutation detection analyses, we have therefore characterized the intron-exon organization of the human keratin 15 and keratin 19 genes. The keratin 15 gene comprises 8 exons spanning approximately 5.1 kb on 17q21, and the keratin 19 gene consists of 6 exons covering approximately 4.7 kb on 17q21. We have also developed a PCR-based mutation detection strategy using primers placed on flanking introns followed by direct sequencing of the PCR products.     

A role for phosphorylation in the dynamics of keratin intermediate filaments

Keratins undergo highly dynamic events in the epithelial cells that express them. These dynamic changes have been associated with important cell processes. We have studied the possible role of keratin phosphorylation-dephosphorylation processes in the control of these dynamic events. Drugs that affect the protein phosphorylation metabolism (activators or inhibitors of protein kinases or protein phosphatases) have been used in two different dynamic experimental systems. First, the behaviour of keratins after the formation of cell heterokaryons, and second, the assembly of a newly synthesised keratin after transfection into the pre-existing keratin cytoskeleton. The main difference between these two systems stems on the alteration of the amount of keratin polypeptides present in the cells, since in heterokaryons this amount was unaltered whilst in transfection experiments there is an increase due to the presence of the transfected protein. We observed in both systems that the inhibition of protein kinases led to a delayed dynamic behaviour of the keratin polypeptides. On the contrary, the inhibition of protein phosphatases by okadaic acid or the activation of protein kinases by phorbol esters promoted a substantial increase in the kinetics of these processes. Biochemical studies demonstrate that this behavioural changes can be correlated with changes in the phosphorylation state of the keratin polypeptides. As a whole, present results indicate that the highly dynamic properties of the keratin polypeptides can be modulated by phosphorylation.     

The transfection of epidermal keratin genes into fibroblasts and simple epithelial cells: evidence for inducing a type I keratin by a type II gene

Through gene transfection studies, we have discovered that the forced expression of a foreign type II epidermal keratin in fibroblasts can trigger the expression of an endogenous type I epidermal keratin. Both the transfected and the induced proteins participate in the formation of filamentous structures. Interestingly, this regulation appears to be unidirectional: the expression of a transfected type I keratin does not induce type II expression. Rather, nonfilamentous aggregates of type I protein accumulate in the cytoplasm. In contrast, simple epithelial cells transfected with either a type I or a type II epidermal keratin gene do not respond by inducing the expression of a host epidermal keratin. In this case, the foreign protein is incorporated into the endogenous keratin network. These results suggest the possibility that type I keratin expression may be dependent on the accumulation of unpolymerized type II keratin.     

Rearrangement of the keratin cytoskeleton after combined treatment with microtubule and microfilament inhibitors

In addition to containing microtubule and microfilament systems, vertebrate epithelial cells contain an elaborate keratin intermediate-filament cytoskeleton. Little is known about its structural organization or function. Using indirect immunofluorescence microscopy with an antikeratin antiserum probe, we found that destabilization of microtubules and microfilaments with cytostatic drugs induces significant alterations in the cytoskeletal organization of keratin filaments in HeLa and fetal mouse epidermal cells. Keratin filament organization was observed to undergo a rapid (1-2 h) transition from a uniform distribution to an open lattice of keratin fibers stabilized by membrane-associated focal centers. Since addition of any one drug alone did not elicit significant organizational change in the keratin cytoskeleton, we suggest that microfilaments and microtubules have a combined role in maintaining the arrangement of keratin in these cells. Vimentin filaments, the only other intermediate-sized filaments found in HeLa cells, did not co-distribute with keratin in untreated or drug-treated cells. These findings offer a new way to approach the study of the dynamics and functional roles of the keratin cytoskeleton in epithelial cells.     

Structure and site of expression of a murine type II hair keratin

We present here a 1770 bp-long cDNA which encodes a murine type II keratin. Sequence comparisons of the keratin with those of various type II keratins expressed in mouse epidermis and internal stratified epithelia reveal that the new keratin is unrelated to epithelial keratins. Rather the structural organization of its amino- and carboxyterminal domains and the high content of cysteine and proline residues in these regions suggest that the keratin represents a murine type II hair keratin. This assumption was confirmed by in situ hybridization which localized the mRNA of the keratin in upper cells of the hair cortex and in suprabasal cells of the central core unit of filiform papillae of the tongue. Hybrid selection analyses revealed that the keratin has a molecular weight of 58 kD. It remains to be seen whether the keratin corresponds to MHb 3 or MHb 4.     

Sequence of a cDNA encoding human keratin No 10 selected according to structural homologies of keratins and their tissue-specific expression

We present here the nucleotide sequence of a 1700 bp-long cDNA encoding human epidermal keratin No. 10 (56.5 kDa). cDNA clones of the acidic keratin family were first isolated from a pBR322 human epidermal cDNA library by hybridization with a probe coding for keratin No. 14. Differential hybridization using total cDNA probes prepared from poly(A)⁺ RNA extracted either from epidermis (which contains keratin No. 10) and from squamous carcinoma or hepatoma cell lines (which do not express keratin No. 10) made possible the selection of clones potentially coding for keratin No. 10. The 1.7 kb sequence exhibits the characteristic features of an acidic keratin with a constant central rod domain and C-terminal variable structures. Moreover, the sequence shows extensive homologies with the cDNA of murine keratin No. 10.     

Phosphorylation of keratin intermediate filaments by protein kinase C, by calmodulin-dependent protein kinase and by cAMP-dependent protein kinase.

Keratins, constituent proteins of intermediate filaments of epithelial cells, are phosphoproteins containing phosphoserine and phosphothreonine. We examined the in vitro phosphorylation of keratin filaments by cAMP-dependent protein kinase, protein kinase C and Ca2+/calmodulin-dependent protein kinase II. When rat liver keratin filaments reconstituted by type I keratin 18 (molecular mass 47 kDa; acidic type) and type II keratin 8 (molecular mass 55 kDa; basic type) in a 1:1 ratio were used as substrates, all the protein kinases phosphorylated both of the constituent proteins to a significant rate and extent, and disassembly of the keratin filament structure occurred. Kinetic analysis suggested that all these protein kinases preferentially phosphorylate keratin 8, compared to keratin 18. The amino acid residues of keratins 8 and 18 phosphorylated by cAMP-dependent protein kinase or protein kinase C were almost exclusively serine, while those phosphorylated by Ca2+/calmodulin-dependent protein kinase II were serine and threonine. Peptide mapping analysis indicated that these protein kinases phosphorylate keratins 8 and 18 in a different manner. These observations gave the way for in vivo studies of the role of phosphorylation in the reorganization of keratin filaments.     

Dynamic light scattering and circular dichroism studies on heat-induced gelation of hard-keratin protein aqueous solutions

Animal hairs consist of aggregates of dead cells filled with keratin protein gel. We succeeded in preparing water-soluble hard-keratin proteins and reconstructing the keratin gels by heat-induced disulfide linkages in vitro. Here, the roles of intermolecular hydrophobic interaction and disulfide bonding between the proteins in the gel were discussed. Water-soluble keratin proteins consisting of mixtures of type I ( approximately 48 kDa) and type II ( approximately 61 kDa) were prepared from wool fibers as S-carboxymethyl alanyl disulfide keratin (CMADK). The gelation was achieved by heating an aqueous solution containing at least 0.8 wt % CMADK at 100 degrees C. CMADK solutions with different urea or N-ethylmaleimide concentrations or pH were exposed to dynamic light scattering (DLS) and circular dichroism (CD). DLS clarified the gelation point of CMADK solutions and provided information on the changes in keratin cluster size. DLS suggested two types of gelation mechanism. One was the regenerated chemical disulfide bonding between keratins from CMAD parts of chains. After the gel formed, this bond became important to maintain the gel structure. The other was the physical assembly due to hydrophobic interaction between alpha-helix parts of keratin chains. This hydrophobic assembly also played an important role during gelation. CD confirmed a conformational change in the keratin protein, resulting heat-induced gelation. CD clarified the relationship between keratin protein conformation and gelation, i.e., a rodlike conformation with many alpha-helix structures was necessary to associate keratin chains and form a gel network.     

The epidermal intermediate filament proteins of tunicates are distant keratins; a polymerisation-competent hetero coiled coil of the Styela D protein and Xenopus keratin 8

Two novel cytoplasmic intermediate filament (IF) proteins (C and D) from the tunicate (urochordate) Styela are characterised as putative keratin orthologs. The coexpression of C and D in all epidermal cells and the obligatory heteropolymeric IF assembly of the recombinant proteins argue for keratin orthologs, but the sequences do not directly reveal which protein behaves as a keratin I or II ortholog. This problem is solved by the finding that keratin 8, a type II keratin from man or Xenopus, forms chimeric IF when mixed with Styela D. Mutant proteins of Styela D and keratin 8 with a single cysteine in equivalent positions show that these chimeric IF are, like vertebrate keratin filaments, based on the hetero coiled coil. We propose that Styela D retains, in spite of its strong sequence drift, important molecular features of type I keratins. By inference Styela C reflects a type II ortholog. We discuss that type I to III IF proteins are expressed along the chordate branch of metazoa.     

Role of dehydration in changing the permeability of erythrocyte plasma membranes by freeze-thawing

The phosphorylation of keratin polypeptides was examined in calf snout epidermis. When slices of epidermis were incubated in the medium containing 32Pi, the radioactivity was incorporated into several proteins. The predominant phosphorylated proteins migrated in SDS-polyacrylamide gels with apparent molecular weights between 49000 and 69000 and coincided with keratin polypeptides. The extent of keratin phosphorylation was not altered in the presence of dibutyryl cyclic AMP or reagents which elevate intracellular cyclic AMP. When homogenates of epidermis were incubated with [gamma-32P]ATP, keratin polypeptides were the predominant species phosphorylated as was also observed in epidermal slices. The presence of cyclic AMP or heat-stable inhibitor of cyclic AMP-dependent protein kinase in the reaction mixture did not affect the phosphorylation of keratin polypeptides, although the phosphorylation of exogenously-added histone was stimulated and inhibited, respectively, by these additions. Keratin polypeptides extracted from calf snout epidermis by 8 M urea were phosphorylated by incubation with [gamma-32P]ATP and cyclic AMP-dependent protein kinase from calf snout epidermis or bovine heart. No proteins were phosphorylated without the addition of the enzymes. The presence of cyclic AMP in the reaction mixture stimulated the keratin phosphorylation, and further addition of heat-stable protein kinase inhibitor reduced this stimulation.     

Keratin and carcinoembryonic antigen (CEA) in human melanoma cells.

Human melanomas are known to contain vimentin intermediate filaments but there has been some dispute about their expression of cytokeratins. The cytoplasm of human M21 melanoma cells maintained in culture reacted with a rabbit anti-keratin antibody and two monoclonal anti-keratin antibodies AE1 and AE2. Cells derived directly from subcutaneous xenografts of M21 melanoma in nude mice, however, failed to express cytokeratins. The presence of keratin filaments in cultured M21 cells was confirmed by electronmicroscopic and immuno-electronmicroscopic examinations of cell extracts. Polyacrylamide gel electrophoresis (PAGE), revealed 46 KD keratin proteins in cultured M21 cells. Small amounts of these low molecular weight keratins were detected by PAGE in M21 melanoma xenografts even though immunofluorescence and immunoperoxidase assays failed to demonstrate keratin at the light microscopic level. Immunofluorescence revealed keratin and carcinoembryonic antigen (hitherto undetected in human melanomas) first on the 9th day of culture of xenograft-derived M21 cells. The appearance of keratin and CEA in M21 melanoma cells in vitro was not affected by inhibition of cellular proliferation or as a result of exposure to methotrexate or adriamycin. However, adriamycin altered the cytoplasmic distribution of keratin.     

Identification of novel principles of keratin filament network turnover in living cells

It is generally assumed that turnover of the keratin filament system occurs by exchange of subunits along its entire length throughout the cytoplasm. We now present evidence that a circumscribed submembranous compartment is actually the main site for network replenishment. This conclusion is based on the following observations in living cells synthesizing fluorescent keratin polypeptides: 1) Small keratin granules originate in close proximity to the plasma membrane and move toward the cell center in a continuous motion while elongating into flexible rod-like fragments that fuse with each other and integrate into the peripheral KF network. 2) Recurrence of fluorescence after photobleaching is first seen in the cell periphery where keratin filaments are born that translocate subsequently as part of the network toward the cell center. 3) Partial keratin network reformation after orthovanadate-induced disruption is restricted to a distinct peripheral zone in which either keratin granules or keratin filaments are transiently formed. These findings extend earlier investigations of mitotic cells in which de novo keratin network formation was shown to originate from the cell cortex. Taken together, our results demonstrate that the keratin filament system is not homogeneous but is organized into temporally and spatially distinct subdomains. Furthermore, the cortical localization of the regulatory cues for keratin filament turnover provides an ideal way to adjust the epithelial cytoskeleton to dynamic cellular processes.     

Comparison of expression patterns of keratin 6, 7, 16, 17, and 19 within multiple independent isolates of As+3- and Cd+2-induced bladder cancer

This laboratory has generated a series of seven cadmium (Cd(+2))- and six arsenite (As(+3))-transformed urothelial cancer cell lines by exposure of parental UROtsa cells to each agent under similar conditions of exposure. In this study, the seven Cd(+2)-transformed cell lines were characterized for the expression of keratin 6, 16, and 17 while the six As(+3) cell lines were assessed for the expression of keratin 7 and 19. The results showed that the series of Cd(+2)-transformed cell lines and their respective transplants all had expression of keratin 6, 16, and 17 mRNA and protein. The expression of keratin 6, 16, and 17 was also correlated with areas of the urothelial tumor cells that had undergone squamous differentiation. The results also showed that four of the six As(+3)-transformed cell lines had expression of keratin 7 and 19 mRNA and protein and produced subcutaneous tumors with intense focal staining for keratin 7 and 19. The other two As(+3)-transformed cell lines had very low expression of keratin 7 mRNA and protein and produced subcutaneous tumors having no immunoreactivity for keratin 7; although keratin 19 expression was still present. The peritoneal tumors produced by one of these two cell lines regained expression of keratin 7 protein. The present results, coupled with previous studies, indicate that malignant transformation of UROtsa cells by Cd(+2) or As(+3) produce similar patterns of keratin 6, 7, 16, 17, and 19 in the resulting series of cell lines and their respective tumors.     

Keratin Polypeptide Analysis in Fetal and in Terminally Differentiating Newborn Mouse Epidermis

The keratin polypeptide pattern of neonatal mouse epidermis consists of eight individual polypeptides having molecular weights of between 46,000 and 67,000. Unlike the keratin patterns in adult mouse epidermis, this pattern is not subject to body site-specific alterations regarding the specific content of distinct polypeptides or the absolute number of keratin constituents.
At day 16 of fetal development the neonatal keratin pattern is only partially expressed, it being fully completed just prior to birth at day 19 of gestation. A comparative analysis of the sequential changes in epidermal morphology and keratin pattern during the last third of embryonic development confirms the view that, independent of the species, keratin polypeptides below 60,000 mol. wt. are generated by basal cells, whereas the biosynthesis of high molecular weight keratin members take place in the suprabasal cell compartments of keratinizing epithelia. The site of origin of five polypeptides (60,000, 58,000, 52,000, 49,000, 46,000) could therefore be attributed to the basal cell layer, the remaining three polypeptides (67,000, 64,000, 62,000) being synthesized in the outer metabolically active epidermal layers. Similar conclusions could be drawn after subfractionation of neonatal epidermis into living (basal, spinous, and granular) and dead cell layers (stratum corneum), and investigation of the corresponding keratin patterns.
During their progression through the epidermis, two proteins (60,000, 58,000) undergo a hitherto undescribed type of posttranslational modification characterized by a slight increase in size and a change in electrical charge. The mechanism underlying this modification is unknown and it is unclear whether the modification if functional or trivial. The largest keratin polypeptide (67,000) of the protein family - probably a product of spinous cells - disappears from the cornified layer without any evidence that it serves as a precursor for smaller keratin subunits.     

Epidermolysis bullosa simplex: a keratin 5 mutation is a fully dominant allele in epidermal cytoskeleton function.

To explore the relationship between abnormal keratin molecules, 10-nm intermediate filament (IF) organization, and epidermal fragility and blistering, we sought to determine the functional consequences of homozygosity for a dominant keratin defect. We describe a family with an autosomal dominant skin-blistering disorder, epidermolysis bullosa simplex, Koebner subtype (EBS-K), that has a novel point mutation, occurring in the keratin 5 gene (KRT5), that predicts the substitution of an evolutionarily conserved lysine by an asparagine residue (K173N). Unlike previous heterozygous mutations located within the initial segment of domain 1A of keratin molecules, K173N heterozygosity did not result in severe disease or clumping of keratin filaments. One family member was found to be homozygous for the K173N allele, having inherited it from each of her affected first-cousin parents. Despite a lack of normal keratin 5 molecules, and an effective doubling of abnormal molecules, available for heterodimerization with keratin 14 during IF formation, there were no significant differences in the clinical severity or the ultrastructural organization of the keratin IF cytoskeleton of the homozygous individual. These data demonstrate that the K173N mutation behaves as a fully dominant allele and indicate that a limited number of abnormal keratin molecules are sufficient to impair cytoskeletal function and elicit epidermal fragility and blistering.     

Modulation of keratin intermediate filament distribution in vivo by induced changes in cyclic AMP-dependent phosphorylation.

Treatment of PtK1 cells with 5 mM acrylamide for 4 hr induces reversible dephosphorylation of keratin in concert with reversible aggregation of intermediate filaments (Eckert and Yeagle, Cell Motil. Cytoskeleton 11:24-30, 1988). We have examined this phenomenon by 1) in vitro phosphorylation of isolated PtK1 keratin filaments and 2) combined treatments of PtK1 cells with both acrylamide and agents which elevate intracellular cAMP levels. PtK1 keratins were incubated in gamma-32P-ATP in the presence or absence of cAMP-dependent kinase (A-kinase) and cAMP. Levels of phosphorylation were analyzed by electrophoresis and autoradiography. Phosphorylation of keratin polypeptides (56 kD, 53 kD, 45 kD, 40 kD) occurred without added kinase, suggesting the presence of an endogenous kinase which remains with intermediate filaments in residues of Triton X-100 extracted cells. Phosphorylation levels were increased by A-kinase but not by cAMP alone, indicating the presence of cAMP-dependent phosphorylation sites in addition to sites phosphorylated by the endogenous kinase. To study the possible role of cAMP-dependent phosphorylation in acrylamide-induced aggregation of keratin filaments, we treated cells with acrylamide in the presence of 8-bromo-cAMP (brcAMP), pertussis toxin (PT), isobutylmethylxanthine (IBMX), or forskolin, which increase intracellular cAMP levels. The distribution and phosphorylation levels of keratin filaments, as well as intracellular cAMP levels, were determined for each of these treatments. In addition to aggregation and dephosphorylation of keratin filaments reported previously, treatment of cells with acrylamide alone also results in reduced levels of intracellular cAMP. 8-bromo-cAMP, IBMX, and forskolin prevent acrylamide-induced aggregation of keratin filaments and result in both normal levels of keratin phosphorylation and normal intracellular cAMP levels. PT was apparently ineffective. These observations suggest that 1) PtK1 keratins are phosphorylated by cAMP-dependent kinase and an endogenous, cAMP-independent kinase and 2) alteration of levels of cAMP-dependent phosphorylation may be involved in aggregation of keratin filaments in response to acrylamide.     

Keratin polypeptide analysis in fetal and in terminally differentiating newborn mouse epidermis

The keratin polypeptide pattern of neonatal mouse epidermis consists of eight individual polypeptides having molecular weights of between 46,000 and 67,000. Unlike the keratin patterns in adult mouse epidermis, this pattern is not subjects to body site-specific alterations regarding the specific content of distinct polypeptides or the absolute number of keratin constituents. At day 16 of fetal development the neonatal keratin pattern is only partially expressed, it being fully completed just prior to birth at day 19 of gestation. A comparative analysis of the sequential changes in epidermal morphology and keratin pattern during the last third of embryonic development confirms the view that, independent of the species, keratin polypeptides below 60,000 mol. wt. are generated by basal cells, whereas the biosynthesis of high molecular weight keratin members take place in the suprabasal cell compartments of keratinizing epithelia. The site of origin of five polypeptides (60,000, 58,000, 52,000, 49,000, 46,000) could therefore be attributed to the basal cell layer, the remaining three polypeptides (67,000, 64,000, 62,000) being synthesized in the outer metabolically active epidermal layers. Similar conclusions could be drawn after subfractionation of neonatal epidermis into living (basal, spinous, and granular) and dead cell layers (stratum corneum), and investigation of the corresponding keratin patterns. During their progression through the epidermis, two proteins (60,000, 58,000) undergo a hitherto undescribed type of posttranslational modification characterized by a slight increase in size and a change in electrical charge. The mechanism underlying this modification is unknown and it is unclear whether the modification if functional or trivial. The largest keratin polypeptide (67,000) of the protein family -- probably a product of spinous cells -- disappears from the cornified layer without any evidence that it serves as a precursor for smaller keratin subunits.