Printer Friendly

Structural-mechanical integration of keratin intermediate filaments with cell peripheral structures in the cornified epidermal keratinocyte.

The outermost visible layer of the skin consists of terminally differentiated, flattened, dead, cornified cells (squames). These squames consist predominantly of keratin intermediate filaments (KIF) embedded in a matrix of filaggrin, and these are contained within a specialized thickened insoluble cell peripheral structure termed the cornified cell envelope (CE). The primary function of these cells is to provide a barrier against the environment. We are interested in how these components are assembled and integrated, since it now seems that any defect in either component will cause a serious epidermal disorder, generally an ichthyosiform disease.

One approach has been to study the structure of the CE. The CE is a 15-20 nm-thick layer of insoluble protein formed on the inside of the plasma membrane during terminal differentiation in the epidermis (as well as in other stratified squamous epithelia). It constitutes about 10% of the mass of the cornified cell. Its insolubility is largely attributable to cross-linking of certain structural proteins by a series of transglutaminase (TGase) enzymes which form an [N.sup.[Epsilon]]([Gamma]-glutamyl) lysine isopeptide bond between the [Gamma]-amide sidechain of a donor glutamine residue and the [Epsilon]-N[H.sub.2] sidechain group of an acceptor lysine residue. Since this bond cannot be cleaved in nature, the result is an insoluble macromolecular complex. In addition, a series of [Omega]-hydroxy ceramides become attached by ester linkages to the cross-linked proteins on the extracellular surface of the cornified cells. This composite protein-lipid CE structure replaces the plasma membrane, and its integrity is vital for barrier function.

Using limited proteolysis procedures, we have isolated and sequenced large numbers of peptides that contain cross-links, and have addressed such questions as which proteins are involved; which glutamine and lysine residues are used on what parts of the proteins; what is the temporal order of protein deposition; and what is the mechanism of assembly of the CE. Quantitatively, most cross-links are intra- and inter-chain links between loricrin, which is the most abundant CE protein (70%-80%). This protein is unusual in that its content of glycine residues is the highest of any known protein in biology. The glycine sequences are clustered into domains that are thought to be configured as highly flexible loops, interspersed by glutamine and lysine rich domains that are the sites of cross-linking. In addition, some cross-links are between loricrin and representatives of the small proline-rich protein (SPR) class (about 5% of total). Most of these linkages suggest that the SPRs serve as cross-bridging proteins. The SPRs contain more proline than any proteins known in biology, and include least 10 different proteins that are differentially expressed in the epidermis of different body sites. The prolines are distributed among peptide repeats of 8 or 9 residues; the repeats are located in a central domain, and each domain comprises from about 3 (smallest SPR protein) to 23 (largest SPR protein) repeats. These peptide repeats are thought to form a relatively stiff structure. The glutamines and lysines used for cross-linking are located on the head (amino-terminal) and tail (carboxy-terminal) domains of the SPRs. Together, loricrin and the SPRs constitute 85% of the protein mass of the CE of the epidermis. Protein expression data show that loricrin and the SPR proteins are located on the intracellular side of the CE; thus these proteins must have been accumulated onto the CE structure during the last stages of its assembly. Thus the bulk of the CE consists of a "solution" of SPRs in loricrin. Although the total amount of loricrin + SPRs remains constant in the epidermis of different body sites, we find the content of SPRs is highest in the epidermis of the palms and soles, with decreasing amounts in elbows, lips, foreskin, and trunk. This series is correlated with decreasing epidermal thickness and with the relative amount of mechanical stress or trauma that these epidermal sites undergo during their normal function.

This variability in composition is familiar to materials science, wherein variations in the amounts of a minor cross-bridging component can greatly change the mechanical properties of a composite material. We therefore speculate that the differential expression of the SPRs reflects the required biomechanical properties of the epidermis of different body sites.

When the ceramide lipids were removed by mild alkaline hydrolysis, we found that a number of other CE proteins were exposed for immunogold electron microscopy. These proteins presumably represent the innermost components of the CE, corresponding to its initial stages of assembly. By additional proteolysis and sequencing, we have found many cross-links within involucrin and between such proteins as involucrin and desmoplakin, or cystatin [Alpha], or envoplakin, or loricrin and SPRs.

Some cross-linking also involved the type II keratins K1, K2e, K5, and K6. Only one lysine residue in these keratin chains was used for cross-linking: it is located in a conserved stretch of sequences in the V1 head domain. This residue is important in two ways. First, in in vitro crosslinking reactions, TGase enzymes specifically utilize this lysine residue to cross-link synthetic peptides. Second, we have discovered a family with autosomal dominant non-epidermolytic palmar plantar keratoderma whose affected members have a single point mutation that results in the loss of this residue. The patients have a severe scaling disorder of their palms and soles and of other thickened epidermal sites. When viewed by electron microscopy, the affected cornified cells are highly irregular in shape, and there is an abnormal accumulation and distribution of the lamellar granules that export the lipids. The failed distribution of the lipids is presumably the cause of the hyperkeratosis and scaling phenotype.

Together, these data suggest that CE assembly is initiated at, or near, the site where KIF meet desmosomes. In addition, they suggest that certain proteins can mediate an indirect association between KIF and desmoplakin. Moreover, these studies offer a mechanism by which the KIF cytoskeleton of the cornified cell is mechanically integrated with the CE. We conclude that the terminal differentiation program exploits the existing KIF cytoskeleton-desmosome machinery to build the novel and vital CE barrier structure. Failure to implement proper attachment by cross-linking has resulted in severe mechanical and biological consequences.


Candi, E., E. Tarcsa, J. J. DiGiovanna, J. G. Compton, P.M. Elias, L.N. Marekov, and P.M. Steinert. 1998. Identification of a highly conserved lysine residue on type II keratins which is essential for the structural coordination of keratin intermediate filaments and the cornified cell envelope through isopeptide crosslinking by transglutaminases. Proc. Natl. Acad. Sci. USA 95: 2067-2072.

Hohl, D., U. Lichti, M. L. Turner, D. R. Roop, and P.M. Steinert. 1991. Characterization of human loricrin: structure and function of a new class of epidermal cell envelope proteins. J. Biol. Chem. 266: 6626-6636.

Jarnik, M., T. Kartasova, P.M. Steinert, U. Lichti, and A. C. Steven. 1996. Differential expression and cell envelope incorporation of small proline rich protein 1 in different cornified epithelia. J. Cell Sci. 109: 1381-1391.

Kartasova, T., Y. Kohno, H. Koizumi, S. Osada, N. Huh, U. Lichti, P.M. Steinert, and T. Kuroki. 1996. Sequence and expression patterns of mouse SPRI: correlation of expression with epithelial function. J. Invest. Dermatol. 106: 294-304.

Kimonis, V., J. J. DiGiovanna, J.-M. Yang, S. Z. Doyle, S.J. Bale, and J. G. Compton. 1994. A mutation in the V1 end domain of keratin 1 causing non-epidermolytic palmar-plantar keratoderma. J. Invest. Dermatol. 103: 764-769.

Steinert, P.M., and L. N. Marekov. 1995. The proteins elafin, filaggrin, keratin intermediate filaments, loricrin and SPRs are isodipeptide crosslinked components of the human epidermal cornified cell envelope. J. Biol. Chem. 270: 17702-17711.

Steinert, P.M., and L. N. Marekov. 1997. Involucrin is an important early component in the assembly of the epidermal cornified cell envelope. J. Biol. Chem. 272: 2021-2030.

Steinert, P.M., E. Candi, T. Kartasova, and L. N. Marekov. 1998. Small Proline Rich Proteins are cross-bridging proteins in the cornified cell envelopes of stratified squamous epithelia. J. Struct. Biol. (in press).

Steinert, P.M., T. Kartasova, and L. N. Marekov. 1998. Biochemical evidence that the small proline rich proteins and trichohyalin function in epithelia by modulation of the biochemical properties of their cornified cell envelopes. J. Biol. Chem. 273:11758-11769.


COULOMBE: Am I correct that non-epidermal PPK features hyper-proliferation in the basal compartment?


COULOMBE: Since the structural defect involved in this disease is a late event, in terms of the differentiation pathways, there must be some means of signaling between suprabasal cells and basal cells in the absence of lysis. Would you care to speculate on this?

STEINERT: I can only speculate. Dr. Peter Elias, a skin biologist and lipid expert, believes that there is a complex interplay between lipid synthesis in these lamellar granules and cell proliferation. He is particularly interested in ceramides, a major lipid component of the lipid envelope, which are known to be second messengers.

GOLDMAN: This is remarkable remodeling of already existing cellular components. The speed at which a cell becomes enucleated, presumably by going through some part of apoptosis, and then reutilizes all of these proteins by cross-linking to become a part of that fortified envelope structure is quite remarkable. Desmosomal components, and keratin, and other normal components are being reorganized into a highly cross-linked structure that bears very little resemblance to desmosomes and other structures in the super basal cells. Can you speculate on how long it takes to go through the process?

STEINERT: On the basis of our studies of cross-linking and immunogold electron microscopy, I believe that we can make intelligent guesses as to how assembly of this complex structure is initiated. We think that this occurs at the site of the desmosome. What probably happens first is association of things like envoplakin and involucrin to the site of the desmosome by cross-linking - more specifically, to the tail of desmoplakin, which extends 100 nm from the main plaque of the desmosome into the cytoplasm. Proteins such as involucrin and envoplakin then form interdesmosomal sheets at the focal points of envelope assembly. This produces a layer of protein in the granular layer of the epidermis. At about this time mayhem breaks loose and everything within the cell is dissolved. This includes all of the cytoplasmic constituents and the plasma membrane. These events are finalized by the addition of loricrin to the existing involucrin-envoplakin scaffold. We do not yet understand when and how the lipid granules, which are of critical importance to both extra- and intracellular lipids, can escape before the whole structure closes down.

GREEN: You mentioned a compromised keratin attachment, but you place more emphasis on the role of the lipid organization problem. Am I correct in presuming that these are two separate functions? If so, what are their related roles for the amino terminus?

STEINERT: I assume you are referring to the amino terminus of desmoplakin? We have no information about its cross-linking to the cell envelope. Perhaps this is buried and may be lost before the envelope is assembled. We have seen only certain desmoplakin sequences near its carboxy terminus, which you have predicted project far into the cytoplasm. However, I want to answer a slightly different question that is important because it bears on what you ask. The amount of keratin cross-linked to the cell envelope represents only about 0.1% of the envelope protein mass. We have calculated what this means in terms of the extent of cross-linking of keratin filaments. Assuming that keratin filaments are 15 nm wide, and assuming a model in which the filaments completely line the intracellular surface of the cell envelope, we calculate about one crosslink per 100 nm of filament length. This is actually a high level, and it suggests that the keratin filament cytoskeleton really gets glued tightly and permanently onto the cell periphery. We have also seen that the amount of keratin cross-linked to the cell envelope of mouse forestomach tissue is about 10 times higher, which is really an incredible amount. This must reflect the biomechanical requirements of a tissue that is subjected to extraordinarily rigorous stresses and trauma. Also, the disease we talk about is autosomal dominant, so we don't know what would happen if both alleles were knocked out - perhaps not a viable fetus. To check this experimentally is of course difficult because, in cell culture, you get upregulation of K5 and K6. A mouse with this lysine residue knocked out in both alleles could be made. We believe that to make only 50% of the cross-links seriously compromises the structural interface between the cytoplasmic keratin filament network and the cell periphery, or the growing cell envelope. It is highly convoluted, which increases the surface area of the cellular connections. Peter Elias hypothesizes that this greater surface area prevents lipids from distributing properly, resulting in decreased barrier function and thus a hyperkeratotic response.

JANMEY: Is the transglutaminase, which crosslinks the proteins together, constitutively active?

STEINERT: Yes the transglutaminase enzymes are very active during the stages when the cell envelope barrier is assembled. The transglutaminase story is complicated. There are a number of different enzymes functioning simultaneously. Some of these enzymes exist in multiple different post-translationally modified forms, each possibly with a different function. What we are trying to unravel now is which enzyme does what. Preliminary data suggest that transglutaminase 3 strongly favors cross-linking of loricrin and SPRs. Apparently one or two forms of transglutaminase 1 are involved in the earlier stages of cross-linking of involucrin and envoplakin to the tails of desmoplakin. We also have data that transglutaminase 1 works synergistically with transglutaminase 3 to cross-link loricrin and SPRs. The studies needed here involve systematic in vitro analyses of the preferences of the enzymes for the numerous substrate proteins, and comparisons with the in vivo data. If the transglutaminase 1 enzyme is knocked out you get the naturally occurring, autosomal recessive human disease called lamellar ichthyosis. This is a life-threatening disease because it produces a serious loss of skin barrier function. The mouse knockout model results in death of the newborn within 6 hours of birth because of dehydration; that is, there has been complete breakdown of skin barrier function.

BORISY: We are cheered by this presentation. If intermediate filaments have no purpose in life, they now have a purpose after death.
COPYRIGHT 1998 University of Chicago Press
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1998 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:The Cytoskeleton: Mechanical, Physical, and Biological Interactions; includes panel discussion
Author:Steinert, Peter M.
Publication:The Biological Bulletin
Date:Jun 1, 1998
Previous Article:Type I keratin 16 forms relatively unstable tetrameric assembly subunits with various type II keratin partners: biochemical basis and functional...
Next Article:What links laminin-5 to the keratin cytoskeleton in epithelial cells?

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters