If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
To whom correspondence should be addressed: German Cancer Research Center, Division of Cell Biology, A0100, Im Neuenheimer Feld 280, D-69120 Heidelberg, Germany. Tel.: 49-6221-42-3436; Fax: 49-6221-42-3404
* This work was supported in part by Deutsche Forschungsgemeinschaft Grant Schw 539/4-1. This is Paper II in the series “Catalog of Human Hair Keratins.” Ref. 46 is Paper I in the series.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The human type II hair keratin subfamily consists of six individual members and can be divided into two groups. The group A members hHb1, hHb3, and hHb6 are structurally related, whereas group C members hHb2, hHb4, and hHb5 are rather distinct. Specific antisera against the individual hair keratins were used to establish the two-dimensional catalog of human type II hair keratins. In this catalog, hHb5 showed up as a series of isoelectric variants, well separated from a lower, more acidic, and complex protein streak containing isoelectric variants of hair keratins hHb1, hHb2, hHb3, and hHb6. Both in situ hybridization and immunohistochemistry on anagen hair follicles showed that hHb5 and hHb2 defined early stages of hair differentiation in the matrix (hHb5) and cuticle (hHb5 and hHb2), respectively. Although cuticular differentiation proceeded without the expression of further type II hair keratins, cortex cells simultaneously expressed hHb1, hHb3, and hHb6 at an advanced stage of differentiation. In contrast, hHb4, which is undetectable in hair follicle extracts and sections, could be identified as the largest and most alkaline member of this subfamily in cytoskeletal extracts of dorsal tongue. This hair keratin was localized in the posterior compartment of the tongue filiform papillae. Comparative analysis of type II with the previously published type I hair keratin expression profiles suggested specific, but more likely, random keratin-pairing principles during trichocyte differentiation. Finally, by combining the previously published type I hair keratin catalog with the type II hair keratin catalog and integrating both into the existing catalog of human epithelial keratins, we present a two-dimensional compilation of the presently known human keratins.
basic hair keratins
acidic hair keratins
polyacrylamide gel electrophoresis
in situ hybridization
The large keratin multigene family comprises the epithelial keratins (also designated as “cytokeratins” or historically “soft keratins”), which are differentially expressed in the various types of epithelia, and the hair keratins (historically designated as “hard keratins”), which are involved in the formation of hard keratinized structures such as hairs, nails, claws, etc. These keratins can be divided into acidic type I and basicto-neutral type II members, which form the 10-nm intermediate filament network of the cytoskeleton of epithelial cells through the obligatory association of equimolar amounts of type I and type II keratins (for review see Refs.
), hair keratins were collectively designated “H” for hair, “b” for the basic members (Hb),1 and “a” for the acidic members (Ha), with the two-dimensionally resolved and Coomassie-stained protein spots of each subfamily being numbered from 1 to 4 in a counterclockwise manner. In addition to the “major” hair keratins hHa1–hHa4 and hHb1–hHb4, a weakly expressed additional pair was designated Hax/Hbx (
). The hair keratin family was therefore suggested to comprise 10 individual members and thus to be considerably smaller than the family of epithelial keratins. Recent investigations in our laboratory have shown, however, that the human hair keratin family is distinctly more complex than previously assumed. Subsequent to the characterization of seven type I hair keratins and four type II hair keratins isolated from a human scalp cDNA library (7-11) and the demonstration that the genes of type I and II hair keratins are located on chromosomes 17q12-21 and 12q13, respectively (
), the screening of a P1 artificial chromosome library with polymerase chain reaction primers for two randomly selected type I hair keratins yielded a 190-kilobase pair P1 artificial chromosome contig that contained nine functional hair keratin genes, hHa1–hHa8(including two highly related hHa3 genes, hHa3-Iand hHa3-II), and one transcribed pseudogene,ϕhHaA, within a 140-kilobase region (
). By means of both specific cRNA probes and specific antibodies for the individual type I hair keratins, we subsequently elucidated the complex expression patterns of the members of this subfamily in the hair follicle (
Recently, we used P1 artificial chromosome cloning to unravel the organization of the human type II hair keratin gene locus. This resulted in the characterization of an ∼200-kilobase pair DNA domain that also contained 10 hair keratin genes (
). However, unlike the type I hair keratin genes, only six of these genes,hHb1–hHb6, were functional, whereas the remaining members either represented nontranscribed (ϕhHbA,ϕhHbB, and ϕhHbC) or transcribed (ϕhHbD) pseudogenes. We were also able to show that the type II hair keratin domain is flanked by the genes for the epithelial keratins K6hf and K7, respectively (
). In the present study, we have once again used in situ hybridization with specific cRNA probes and indirect immunofluorescence studies with specific antibodies to determine the expression sites of the individual type II hair keratins in the human hair follicle. A comparison with the type I hair keratin expression pattern suggests complex dynamics of hair keratin pair formation during the process of hair shaft differentiation. We also provide a two-dimensional catalog of the presently known human type I and II hair keratins.
After the elucidation of the human type I hair keratin genes and their expression patterns in the hair follicle as well as the assembly of the encoded proteins into a catalog (
), the extension of these studies to type II hair keratins became mandatory. The type I hair keratin gene subfamily on chromosome 17q12–21 comprises nine genes, which in terms of sequence homologies and gene organization can be divided into three groups. Although two of these groups, A and B, each encode highly related hair keratins, group C codes for less related hair keratins (Ref.
and Fig.7a). In contrast, the type lI hair keratin gene subfamily located on chromosome 12q13 lacks functional equivalents of type I group B genes, and its six members are organized into type I group A and C counterparts, respectively (Ref.
), we met with substantially less difficulties in designing specific oligopeptides for the individual type II hair keratins. Consequently, the individual antisera all detected only single protein bands in one-dimensional Western blots of hair keratin extracts. Although we previously noticed considerable discrepancies between calculated molecular mass values and position in one-dimensional gels for some members of the type I hair keratin subfamily (
), the migration properties and mass calculations of type II hair keratins agreed much better. The assignment of the type II hair keratins in Western blots of two-dimensionally resolved hair keratins was facilitated in that hair keratin hHb5, the largest and most alkaline member of the Coomassie-stainable proteins, could be used as “marker” for the positioning of the remaining, distinctly more acidic proteins. Of these, hHb1, hHb3, and hHb6 all showed similar mobilities, whereas hHb2 was located slightly above the bulk of hHb1, hHb3, and hHb6. Intriguingly, despite the use of antisera raised against four different oligopeptides of the hair keratin hHb4, this protein could not be detected in Western blots of hair keratin extracts. However, in cytoskeletal extracts of human dorsal tongue,i.e. another anatomical site known to express hair keratins in its filiform papillae (
), all hHb4 antisera unambiguously identified this hair keratin as the largest and most alkaline member of the type II hair keratin family.
The catalog of human type II hair keratins resulting from the compilation of these two-dimensional data is shown in Fig.7b. A comparison of this catalog, in which gray dots indicate the most alkaline isoelectric variants of the various hair keratins, with the data proposed earlier by Heid et al. (
) (Fig. 7c) reveals that none of the earlier designations agree with the present ones. A priori, this can not be expected because the previous nomenclature for type I and type II hair keratins relies on a counterclockwise numerical designation of Coomassie-stainable hair keratins (
), whereas the recent designation is based on the immunodetection of defined hair keratin gene products. Similar to our study, one “minor” member of the type II hair keratin subfamily, designated Hbx by Heid et al. (
). Here we have demonstrated the identity of Hbx with hHb4 using two-dimensional Western blots of human dorsal tongue cytoskeletal extracts with specific hHb4 antisera (Fig. 2, g and h, see also Fig. 7, b and c).
In Fig. 7d we have combined the type II hair keratin catalog with the previously published type I hair keratin catalog (
). As it stands, this human keratin catalog now comprises 22 type II members (16 epithelial keratins and six hair keratins) and 20 type I members (11 epithelial keratins and nine hair keratins). Although we are fairly convinced that this catalog contains the complete number of hair keratins (
), there is strong evidence that the number of keratins expressed in epithelia is far from being complete. Recent estimates from the human genome sequencing project predict the existence of as many as 111 keratin genes (
) and for two pseudogenes of each K14, K16, and K17, as well as one K19 pseudogene.3 Moreover, up to 20 pseudogenes have been suggested for K18.3 This would bring the total number of type I keratin genes/pseudogenes to 55. Similarly, 27 human type II keratin genes/pseudogenes (17 keratin genes/pseudogenes (K1,K2e, K2p, K3, K4,K5, K6a-f, K6hf, K6irs,K7, K8, and ϕK8)3 and 10 hair keratin genes/pseudogenes (hHb1, hHb2,hHb3, hHb4, hHb5, hHb6,ϕhHbA, ϕhHbB, ϕhHbC, andϕhHbD (
). If a comparable gene density on the remaining 390 kilobases of the YAC contig exists as for the 240-kilobase contig, 21 instead of 14 genes should be expected.
the prediction of ∼111 human keratin genes seems realistic, although the number of functional genes will certainly be much lower.
The catalog of human keratins reveals a numerical imbalance between type I and type II members. In the case of hair keratins, this is mainly because of the lack of type II counterparts of the type I group B hair keratin genes hHa7 and hHa8 (Fig.7a). Both hair keratins show unusual expression restrictions to central cortex cells of vellus hairs (hHa7) or single cortex cells of terminal hairs (hHa8) (
). More importantly, not only the hair keratin encoded by the chimpanzee ortholog of ϕhHaA but also those encoded by the chimpanzee orthologs of hHa7 and hHa8 were found to constitute major components of cortex cells of terminal chimpanzee hairs (
L. Langbein, H. Winter, M. A. Rogers, S. Praetzel, and J. Schweizer, manuscript in preparation.
These striking differences between human and chimpanzee may indicate that since theHomo-Pan divergence, the entire human type I group B hair keratin gene cluster is under low evolutionary pressure and possibly on its way to being completely eliminated from the genome. This fate might have already happened to its type II counterpart, from which apparently only a nontranscribed pseudogene, ϕhHbB, is left (see Fig. 7a).
To collectively describe the extremely complex expression pattern of human type I and type II hair keratins in the hair follicle, we have summarized the corresponding expression data schematically (Ref.
and this paper) in Fig. 8, a andb, for better understanding. For the sake of a comprehensive interpretation of these data, in both schemes only mRNA expression profiles of the various hair keratins are reported. To visualize the sequential expression of the various hair keratins, the designations of the individual type I (red) and type II (blue) hair keratins in Fig. 8a stand for the site of onset of their mRNA synthesis in the hair-forming compartments, with the size of the letters reflecting the degree of expression. In Fig.8b, vertical columns indicate the mRNA expression zones of the individual type I (red) and type II (blue) hair keratins within the hair cuticle and matrix/cortex, respectively. Because our IIF studies have shown that independent of the onset of mRNA expression, virtually all hair cuticle and matrix/cortex keratin proteins can be demonstrated up to the zone of fiber hardening (thin vertical arrows in Fig.8b), this implies that any hair cuticle cell leaving the living cell compartment contains four different hair keratins, whereas any cortex cell may accumulate an unprecedented number of up to 12 different hair keratins (Fig. 8b). In this respect, it is highly astonishing that the impact of a disruptive mutation in a cortex keratin, hHb1 or hHb6, is obviously not attenuated by the high number of paralleling unaffected keratin pairings but entails such dramatic alterations as the beaded-hair phenotype in Monilethrix patients (
It is evident that this complex scenario of hair keratin expression in the hair-forming compartments makes it difficult to clearly decipher the formation of specific type I and type II heteropolymeric hair keratin pairs as observed previously for keratins in various types of epithelia (
). The hair cuticle with its less complex expression pattern may serve as an example for this (Fig. 8b). In the lowermost portion of the cuticle (level 2 in Fig. 8b), the presence of two type I hair keratins hHa2 and hHa5 versusthe presence of only one type II hair keratin, hHb5, clearly suggests a competition of hHa2 with hHa5 for filament formation with hHb5. Further up (level 3 in Fig. 8b), the type I hair keratins hHa2 and hHa5 are coexpressed temporarily with type II hair keratins hHb2 and hHb5 before hHa5 expression ceases and hence confronts hHa2 with hHb2 and hHb5 (level 4 in Fig. 8b), thus creating an inversion of the situation in the lowermost cuticle. Only in the uppermost region of active cuticular hair keratin expression is hHa2 left to specifically pair with hHb2 (level 5 in Fig.8b). Collectively, this scenario, which resembles that of the complex epithelial keratin K1, K10, K9, and K2e expression pattern in suprabasal human plantar epidermis (
), suggests competitive, complex, and unique type I and type II hair keratin-pairing patterns. The same dynamics hold true for the potential pairing possibilities of matrix keratins and in particular cortex keratins (Fig. 8b), the unprecedented complexity of which makes it difficult to assume the formation of specific pairs. It remains to be seen whether the assessment of the morphology (cf. 11) or viscosity of intermediate filaments assembled in vitro from various recombinant type I and type II hair keratins or biophysical investigations on affinity strengths between various in vitro combinations of type I and type II hair keratins (cf. 37) may be suitable approaches to ultimately solve the problem of specific or “promiscuous” hair keratin pairing in vivo.
It is also evident that the complex scenario of hair keratin expression requires highly stringent control mechanisms at the gene level, in particular if further studies should confirm specific rather than random pairing patterns. At present, we know next to nothing about the regulation of hair keratin expression in the anagen follicle. The only regulatory factor for which the activation of hair keratin genes via direct interaction with promoter elements has been demonstrated clearly is the forkhead transcription factor Foxn1 (formerly Whn, winged helix nude), the defective gene of which is responsible for the nude mouse phenotype (
). Recent transgenic experiments using a reporter gene under the control of a sheep wool keratin promoter containing a mutated LEF-1 binding site suggested that LEF-1 may enhance transcriptions of hair keratin genes by introducing changes in DNA conformation, thus allowing other transcription factors to assemble into an active complex (
). This view is contrasted by our own in vitrotransfection experiments, which clearly showed that the activation of various hair keratin promotors by β-catenin, generally thought to act in combination with LEF-1 because of the absence of a DNA binding domain, was completely independent of the presence or absence of the LEF-1 binding site in the promotors.
L. F. Jave-Suarez, H. Winter, L. Langbein, M. A. Rogers, and J. Schweizer, unpublished data.
It is evident that these data represent only the tip of an iceberg and that the basic control mechanisms governing hair keratin expression have yet to be elucidated. In this context, it is worth mentioning that recent studies in our laboratory have shown that the ordered hair keratin expression seen in the hair follicle is largely maintained in pilomatricomas,i.e. a tumor type thought to originate from the hair-forming compartment of the hair follicle (Ref.
and references therein). Therefore, pilomatricomas, and more importantly cell lines derived from such tumors, should be suitable tools to address the compelling question of how hair keratin gene expression is controlled.
We are grateful to Werner W. Franke for his continued interest in and support of this work and Harald Herrmann and Ilse Hofmann for helpful discussion. We thank Irene M. Leigh (London) for monoclonal hair keratin antibodies, Herbert Spring for help with confocal laser microscopy, and Ulrike Beckhaus for technical assistance.