An antibody directed against the DNA-binding region of c-fos was used to localize the distribution of cells positive for Fos protein in epithelial tissues. The antibody consistently bound to the nuclei of epithelial cells in the late stages of differentiation, just prior to cornification. The epidermis, palate, buccal mucosa, gingiva, tongue, forestomach and vagina in estrus all produced this type of labelling, suggesting a burst of expression immediately before cell death and cornification. The differentiating cells of the hair follicle, including the hair and inner root sheath, were also labelled. Non-keratinized tissues including junctional epithelium, embryonic epidermis and diestrus vaginal epithelium showed little or no Fos labelling. With the onset of keratinization at 18 days gestation or with induction of estrus in ovariecto-mized mice with estradiol benzoate, the epidermis and vagina expressed Fos protein in the manner typical for keratinized tissues. The Erf Er mutant epidermis, a tissue that is blocked in its ability to keratinize, overexpresses Fos with Fos-positive cells appearing in virtually every cell layer. Gel shift analysis demonstrates the presence of a functional AP-1 complex in epidermal extracts that is recognized by our antibody. Our data suggest that the expression of Fos is intricately related to epithelial cell differentiation, specifically in relation to the process of cornification and cell death.

Keratinization is an orderly process involving cellular stratification and differential regulation of epithelial-specific gene products. Keratinizing cells arise from an epithelial basal cell layer and migrate toward the tissue surface. During the course of this migration, the cells undergo a variety of morphological changes that reflect the differential expression of epithelial genes (Fuchs and Green, 1980; Sun et al. 1984; Fisher et al. 1987a,b). The final event in this process of cellular differentiation is the cornification of the cell resulting in the elaboration of complexes of proteins and the degradation of cellular organelles (Brody, 1959; Lavker and Matoltsy, 1970; reviewed in Holbrook, 1989). The protein complexes consist primarily of cytoskeletal elements (keratin intermediate filaments) embedded in an electron-dense matrix of protein (Brody, 1959, 1960; Dale et al. 1978), and an insoluble, membrane-associated envelope (Rice and Green, 1979). The assembly of these materials is accomplished by the enzymatic modification of proteins present in the viable cells of the epithelium (Resing et al. 1984; Bowden et al. 1984; Rice and Green, 1979).

The transition from a viable cell, synthesizing epithelial structural proteins, to a dead, cornified cell is dramatic. This process occurs rapidly and likely requires the sudden induction of a host of genes, including nucleases and proteases, in response to an as yet unknown stimulus. The Fos family of proteins are transcriptional activators that are rapidly induced by extracellular stimuli (Turner and Tjian, 1989; Zerial et al. 1989; Cohen and Curran, 1988). We are interested in localizing the expression of protooncogenes in epithelia in order to identify the cell types and periods during development that are critical to the appearance of these genes. This paper reports that Fos protein is expressed immediately prior to cell death and cornification in keratinizing tissues. The data suggest that Fos plays an important role in the transcriptional activation of genes mediating cornification and, also, that the cells expressing Fos in keratinizing epithelia are the target of an as yet unknown differentiation signal.

Animals and tissues

Embryonic and newborn Swiss-Webster mouse and Sprague-Dawley rat tissues were collected and fixed in 4 % paraformaldehyde in phosphate-buffered saline (PBS). Tissues were embedded in paraffin by routine procedures. In the case of adult vaginal tissue, mice were ovariectomized, allowed to recover for 2 weeks, and then administered 10 μg estradiol benzoate in com oil i.p. This allowed us to collect nonkeratinized (ovariectomized) and keratinized (72 h postestradiol) vaginal tissue (Barker and Walker, 1966). For examination of adult oral tissues, 4-week-old Sprague-Dawley rats were anesthetized, perfused for 10min with 4% paraformaldehyde in 0.1M phosphate buffer, and jaws removed and fixed for another 1–2 h. Jaws were decalcified for 3–5 days in 4 N formic acid in 0.5 M sodium formate at 4°C.

Fos antibodies

A number of antibodies directed against c-fos peptides were used in these studies. Two antibodies directed against NH2-terminal peptides gave high backgrounds in immunohistochemical preparations. These peptides were subsequently found to share significant homology (approximately 40–60%) with basic keratins and to cross-react with keratins on Western blots. The antibodies that gave the best results were prepared against a 25 amino acid synthetic peptide from the DNA binding region (the M-peptide region; Franza et al. 1987) of c-fos as previously described (Quinn et al. 1989). This peptide is 100 % conserved in the mouse, chicken and human c-fos protein (Van Straaten et al. 1983; Van Beveren et al. 1983; Molders et al. 1987). Briefly, the peptide (KVEQLS-PEEEEKRRIRRERNKMAAA) was conjugated with l-ethyl-3(3-dimethylaminopropyl)-carbodiimide to succinic anhydride-reacted keyhole limpet hemocyanin and injected intradermally in rabbits in an emulsion containing Freund’s complete adjuvant. The antibodies were purified by affinity chromatography with the peptide.

Immunohistochemistry

Immunohistochemistry by the avidin-biotin-peroxidase technique was performed as previously described (Fisher et al. 1987a) except that the primary antibody (1:1000 dilution) step was carried out at 37 °C for 2–3 h and sections were stained briefly with hematoxylin. Adult rat jaws were equilibrated in 30% sucrose, serially sectioned at 50pm with a freezing microtome, and the sections incubated in a 1:2500 dilution of the M peptide antibody for 60 h at 4°C. Antibody binding was localized by the avidin-biotin-peroxidase technique and sections were counterstained with cresyl violet. Controls, consisting of elimination of primary antibody and competition of antibody with 10–6M peptide, were routinely negative.

Epidermis extracts

Epidermises of newborn (1-to 2-day-old) mice were extracted by a procedure modified from Dignam et al. (1983). The buffers were modified as in Quinn et al. (1989). In summary, the epidermis was separated from mouse skin after incubation in 10 mM EDTA at 55°C for 2 min. After separation, all subsequent steps were at 4 °C. The epidermis was minced in the extraction buffer containing 20 mM Hepes (pH 7.2), 20% glycerol, 0.42M sodium chloride, 1.5mM magnesium chloride, 0.2 mM EDTA, 0.5 mM phenylmethylsulfonyl fluoride, 0.5 mM dithiothreitol and 2.1μgml–1 aprotinin. The tissue was homogenized with 10 strokes, twice, in a Dounce homogenizer with the B pestle and centrifuged at 25000g for 20 min. The supernatant was dialyzed for 3–4 h against dialysis buffer containing 80 mM potassium chloride instead of the sodium chloride and magnesium chloride in the extraction buffer. The supernatant was rapidly frozen in ethanol/dry ice and stored at –70°C. These extracts were used in gel retardation studies.

Gel retardation analysis

Gel retardation analysis was performed as previously described (Singh et al. 1986; Quinn et al. 1989). Briefly, 32P-labelled 3′-end double stranded oligonucleotide representing the gibbon ape leukemia virus (GALV) enhancer (CGA-GAATAGATGAGTCAACAGCG) was reacted with varying concentrations of epidermal extracts, with and without anti-M peptide antibody, and run on polyacrylamide gels. Controls consisted of competition with cold GALV oligonucleotide and competition with a random-mer oligonucleotide.

Immunohistochemistry

A number of rodent epithelia were examined for the localization of Fos including skin and hair, oral epithelia, forestomach and vagina from estradiol-induced, ovariectomized mice. Results were identical in every case in which both rat and mouse tissues were examined.

Skin and hair

The localization of Fos-positive nuclei in skin was restricted to the epidermis and the epithelial component of the hair follicle (Fig. 1A,B). Within the newborn epidermis of mice and rats Fos-positive cells were found in the basal layers and the upper granular layers of the epidermis (Fig. IB). The cells in the upper granular layers showed a more intense immunoreactivity over the immunoreactive cells in the deeper epidermal layers. While the nuclear labelling of the basal cell layer was of variable intensity depending upon the preparation, the labelling of the upper granular layer was consistently and strongly positive.

Fig. 1.

Localization of Fos protein in rodent skin. (A) Neonatal rat skin shows labelling in epithelial cells including basal cell layer (arrowheads), granular cell layer (small arrows) and hair follicles (large arrows). Bar=50 μm. Higher magnification of interfollicular epidermis demonstrating typical Fos labelling in epidermis. Note labelling of basal cell layer (arrowheads) and intensely positive cells (arrows) in the upper stratum granulosum (sg), just below the cornified cells of the stratum comeum (sc). Bar=50 μm. (C) The upper granular layer Fos-positive cells (large arrows) first appear in interfollicular epidermis of 17–18 day gestation embryonic mice. The Fos-negative epidermal basal cells are indicated by arrowheads. Non-specifically stained cells of the dermis, probably mast cells, are indicated by small arrows. Bar=50 μm. (D) The hyperplastic epidermis of the 18 day gestation Er/Er mouse shows Fos-positive cells throughout its thickness, in most cell layers. Basal cell layer indicated by arrowheads. Bar=50 μm. (E) The newborn mouse vibrissae follicle has Fos-positive cells in almost all layers of the hair cortex and inner root sheath, beginning (arrows) at the level of the apex of the dermal papilla (dp). Bar=50 μm. (F) Higher magnification of E showing that positive labelling of cells for Fos begins at about the level of the apex of the dermal papilla in all layers of the hair cortex (HC), and in the inner root sheath cuticle (cu) and Henle (He) layers. Fos is not detected in the inner root sheath layer of Huxley (Hu). Bar=50 μm.

Fig. 1.

Localization of Fos protein in rodent skin. (A) Neonatal rat skin shows labelling in epithelial cells including basal cell layer (arrowheads), granular cell layer (small arrows) and hair follicles (large arrows). Bar=50 μm. Higher magnification of interfollicular epidermis demonstrating typical Fos labelling in epidermis. Note labelling of basal cell layer (arrowheads) and intensely positive cells (arrows) in the upper stratum granulosum (sg), just below the cornified cells of the stratum comeum (sc). Bar=50 μm. (C) The upper granular layer Fos-positive cells (large arrows) first appear in interfollicular epidermis of 17–18 day gestation embryonic mice. The Fos-negative epidermal basal cells are indicated by arrowheads. Non-specifically stained cells of the dermis, probably mast cells, are indicated by small arrows. Bar=50 μm. (D) The hyperplastic epidermis of the 18 day gestation Er/Er mouse shows Fos-positive cells throughout its thickness, in most cell layers. Basal cell layer indicated by arrowheads. Bar=50 μm. (E) The newborn mouse vibrissae follicle has Fos-positive cells in almost all layers of the hair cortex and inner root sheath, beginning (arrows) at the level of the apex of the dermal papilla (dp). Bar=50 μm. (F) Higher magnification of E showing that positive labelling of cells for Fos begins at about the level of the apex of the dermal papilla in all layers of the hair cortex (HC), and in the inner root sheath cuticle (cu) and Henle (He) layers. Fos is not detected in the inner root sheath layer of Huxley (Hu). Bar=50 μm.

Embryonic epidermis ⩽16 days gestation exhibited no or weak immunoreactivity (data not shown). The intense labelling associated with keratinocytes in the upper granular layers was not evident in embryonic epidermis until the onset of keratinization at 17–18 days gestation in the mouse (Fig. 1C). In order to test whether this pattern of nuclear labelling was altered in an epidermis that is blocked in its ability to differentiate, we examined the distribution of Fos in the epidermis of the repeated epilation (Er/Er) mutant mouse (Fig. ID). The Er/Er mutant epidermis presented an altered pattern of labelling with virtually all cell layers showing nuclear labelling throughout the thickened mutant epidermis.

Cells immunoreactive with the M-peptide antibody were also detected in both pelage (Fig. 1A,B) and vibrissa (Fig. 1E,F) hair follicles, both in the infundibulum of the hair as well as the hair bulb (Fig. 1A). Preparations of individually embedded vibrissa follicles were useful for accurately localizing the Fos-positive cells in the hair and inner root sheath (Fig. 1E,F). The cells of the hair first expressed Fos as they moved from the proliferative compartment into the zone immediately superior to the dermal papilla. The nuclei remained positive throughout the period of differentiation until the cells cornified higher in the hair follicle (Fig. IE); by the time the hair keratinized the nuclei became elongated and parallel to the axis of the cornifying hair cell. The cells of the cuticle showed a similar pattern of Fos expression during their differentiation. The small cuticle cell nuclei appeared as a string of small nuclei that were Fos-positive above the level of the dermal papilla and ended with cornification (Fig. 1E,F). Layers of the inner root sheath, on the other hand, showed a more variable pattern of staining with positive cells appearing in the suprabulbar region in the layer of Henle but not the layer of Huxley (Fig. 1E,F).

Occasionally cells in the dermis of embryonic and newborn skin were stained (see Fig. 1C) but this labeling was determined to be non-specific as it also appeared in controls.

Forestomach

The mouse forestomach is a keratinized tissue that, like epidermis, elaborates granular and cornified layers. Localization of Fos in this tissue demonstrated positive cells immediately beneath the cornifying cells (Fig. 2A,B). As with the other tissues examined, incubation of the antibody with the 25 amino acid synthetic M-peptide eliminated this immunoreactivity (Fig.2C).

Fig. 2.

Localization of Fos in neonatal mouse forestomach. (A,B) Micrographs demonstrating the distribution of Fos-positive cells in mouse forestomach show positive cells in the granular layer, immediately adjacent to the cornified cell layers lining the lumen (L). The basal cell layers are indicated by arrowheads. Bars=50 μm. (C) Control demonstrating the elimination of immunoreactivity by competition of the antibody with the peptide. The basal cells (arrowheads) and lumen (L) are indicated. Bar=50 μm.

Fig. 2.

Localization of Fos in neonatal mouse forestomach. (A,B) Micrographs demonstrating the distribution of Fos-positive cells in mouse forestomach show positive cells in the granular layer, immediately adjacent to the cornified cell layers lining the lumen (L). The basal cell layers are indicated by arrowheads. Bars=50 μm. (C) Control demonstrating the elimination of immunoreactivity by competition of the antibody with the peptide. The basal cells (arrowheads) and lumen (L) are indicated. Bar=50 μm.

Oral epithelium

Adult rat oral epithelia displayed (Fig. 3) similar associations of Fos-positive cells with cornification. In buccal (Fig. 3A), gingival (Fig. 3A,B), tongue (Fig. 3C) and palatal (data not shown) epithelia, Fos-positive cells appeared high in the viable layers of these epithelia, just prior to cornification. Changes in Fos labelling were associated with abrupt epithelial transitions. With the transition from buccal to gingival in the adult rat oral epithelium, an enhancement of Fos labelling in the basal cell population was noted (Fig. 3A). The transition from the keratinized gingival epithelium to the nonkeratinized junctional epithelium was associated with an abrupt decrease of Fos labelling (Fig. 3A,B). While some Fos labelling is found in the nonkeratinized junctional epithelium, it was less intense and more random than labelling in keratinizing tissues. The tongue (Fig. 3C) also displayed a similar association of Fos labelling with cornification, particularly in the filiform papilla. As with all other tissues examined, competition of the antibody with the M-peptide resulted in the elimination of nuclear binding (Fig. 3D).

Fig. 3.

The localization of Fos-positive cells in rat oral epithelia. (A) Section through oral epithelia of 4-week-old rat demonstrating the localization of Fos-positive cells in buccal (b), gingival (g) and junctional (j) epithelia adjacent to tooth enamel (T). The keratinizing buccal (cut tangentially) and gingival epithelia show numerous Fos-positive cells just below the stratum comeum (sc) while the nonkeratinized junctional epithelium (j) is largely negative. Note the switch in Fos labelling in transition zones between epithelia (arrows). Bar= 100 μm. (B) Higher magnification of gingival (g)-junctional (j) epithelial transition showing the lack of Fos labelling in the nonkeratinized junctional epithelia. Bar=100 μm.(C) Micrograph of tongue epithelium demonstrating Fos-positive cells associated with keratinization in outer epithelium. Some basal cell staining is associated with the fungiform papilla housing a taste bud (arrow), but the strongest labelling is associated with keratinization in the filiform papillae flanking the fungiform papilla. Bar=100 μm. (D) Incubation of the antibody with the peptide eliminates labelling in the gingival (g) and junctional (j) epithelia. Bar=60 μm.

Fig. 3.

The localization of Fos-positive cells in rat oral epithelia. (A) Section through oral epithelia of 4-week-old rat demonstrating the localization of Fos-positive cells in buccal (b), gingival (g) and junctional (j) epithelia adjacent to tooth enamel (T). The keratinizing buccal (cut tangentially) and gingival epithelia show numerous Fos-positive cells just below the stratum comeum (sc) while the nonkeratinized junctional epithelium (j) is largely negative. Note the switch in Fos labelling in transition zones between epithelia (arrows). Bar= 100 μm. (B) Higher magnification of gingival (g)-junctional (j) epithelial transition showing the lack of Fos labelling in the nonkeratinized junctional epithelia. Bar=100 μm.(C) Micrograph of tongue epithelium demonstrating Fos-positive cells associated with keratinization in outer epithelium. Some basal cell staining is associated with the fungiform papilla housing a taste bud (arrow), but the strongest labelling is associated with keratinization in the filiform papillae flanking the fungiform papilla. Bar=100 μm. (D) Incubation of the antibody with the peptide eliminates labelling in the gingival (g) and junctional (j) epithelia. Bar=60 μm.

Vagina

The nonkeratinized vaginal epithelium of 0 and 24 h post-estradiol-treated ovariectomized mice exhibited no detectable immunoreactivity with the M-peptide antibody. However, 72 h after administration of estradiol, the cells in the upper, viable layers of the keratinizing vaginal epithelium exhibited an intense positive reaction with the antibody (Fig. 4C,D). These Fos-positive cells were located immediately beneath a well-formed stratum corneum.

Fig. 4.

Fos localization in vaginal epithelium of ovariectomized, estradiol benzoate treated mice. (A) Ovariectomized mice receiving no estradiol show no signs of Fos labelling in the mucous-secreting epithelium. Bar=50 μm. (B) 24 h after receiving estradiol benzoate the mucous-secreting epithelium is still negative for Fos. Bar=50 μm. (C) 72 h after receiving 10 ug estradiol benzoate the vaginal epithelium is now keratinized and shows Fos-positive cells in the outermost layers of the granular layer. Bar=50 μm. (D) Higher magnification of C showing the keratinized vaginal epithelium of an ovariectomized mouse, 72 h after estradiol benzoate injection. Bar=50 μm.

Fig. 4.

Fos localization in vaginal epithelium of ovariectomized, estradiol benzoate treated mice. (A) Ovariectomized mice receiving no estradiol show no signs of Fos labelling in the mucous-secreting epithelium. Bar=50 μm. (B) 24 h after receiving estradiol benzoate the mucous-secreting epithelium is still negative for Fos. Bar=50 μm. (C) 72 h after receiving 10 ug estradiol benzoate the vaginal epithelium is now keratinized and shows Fos-positive cells in the outermost layers of the granular layer. Bar=50 μm. (D) Higher magnification of C showing the keratinized vaginal epithelium of an ovariectomized mouse, 72 h after estradiol benzoate injection. Bar=50 μm.

Gel retardation analysis

Extracts of newborn murine epidermis caused a marked retardation of oligonucleotides representing the GALV enhancer AP-1 binding site (Fig. 5). The M-peptide antibodies produced a further retardation of migration of the AP-l-complex (Fig. 5). Competition with cold GALV oligonucleotides completely eliminated the AP-1 shift, while a random oligonucleotide did not compete for AP-1 (data not shown).

Fig. 5.

Gel shift analysis demonstrates an AP-1 shift elicited by 10 μl of extract of newborn mouse epidermis. Addition of M-peptide antibodies causes a retardation in migration of the AP-1 complex (+c-fos AB).

Fig. 5.

Gel shift analysis demonstrates an AP-1 shift elicited by 10 μl of extract of newborn mouse epidermis. Addition of M-peptide antibodies causes a retardation in migration of the AP-1 complex (+c-fos AB).

The data presented demonstrate a distinctive association between the expression of Fos protein and keratinization. The process of keratinization results in the destruction of cellular organelles, and the appearance of the keratin pattern and cornified cell envelopes typical of a dead, cornified cell. This cataclysmic process is mediated by a variety of mechanisms involving the processing and assembly of keratins and keratohyalin (Dale et al. 1988; Fisher et al. 1987a; Bowden et al. 1984), and the degradation of cell organelles, presumably by catalytic enzymes including proteases and nucleases (Brody, 1959; Lavker and Matoltsy, 1970). Furthermore, this process is rapid and must be regulated so that the destructive mechanisms involved do not interfere with the assembly of the components of the cornified cell. In the light of these facts, it is anticipated that a number of genes might be activated or repressed immediately prior to entering the destructive phase of the keratinization process. We have demonstrated in neonatal and adult rodent tissue that Fos protein is expressed just prior to cornification and cell death in a variety of diverse keratinizing epithelia including epidermis and hair, oral tissues including tongue, palate, gingiva and buccal mucosa, the vagina in estrus and the forestomach. Nonkeratinized epithelia including embryonic epidermis, junctional epithelium, and the vagina in diestrus express Fos protein in reduced or undetectable levels. Our results suggest that Fos expression plays an important role in keratinization, possibly in the activation or repression of genes important to cornification. It is not clear what genes are activated or repressed in these cells but reports have suggested an important role for c-fos in the activation of proteases including collagenase (Schonthal et al. 1988) and transin (Kerr et al. 1988).

The transition from a viable to cornified cell is presumably rapid because cells bearing ultrastructural features of-both viable and cornified cells, so-called ‘transitional cells’, are surprisingly rare. The Fos-positive cells that we find high in the epidermal granular layer may represent cells that are primed to undergo this transition. It is tempting to speculate that the apparent even spacing of Fos-positive cells in the upper granular layer (Fig. 1) is related to the columnar organization of cells described for mouse trunk epidermis (MacKenzie, 1969). It may be that cells are triggered to keratinize within these so-called epidermal proliferative units (Potten, 1981) in precisely the same position, possibly in the center of the column, resulting in the apparent non-random distribution of Fos-labeling.

While the labelling of cells immediately prior to keratinization with anti-Fos antibodies is a consistently reproducible finding, the association of Fos-labeling with proliferative (basal) cell populations is more variable. Labelling of interfollicular epidermal basal cells occurs in some preparations and not in others. In the hair follicle, labelling was restricted to post-mitotic cell populations. Perhaps the most informative observation on the relationship between proliferation and differentiation can be made in oral epithelia, where well-defined borders divide one epithelial type from another (Fig. 3). Differences in basal cell labelling in these tissues are often associated with the transition from one epithelium to another (Fig. 3A) These results suggest that Fos-expression in basal cell populations may have less to do with regulation of cell proliferation than with control of cell phenotype. Some caution, however, should be exercised in the interpretation of these results. Failure to identify Fos within certain populations of cells may be due to the stability of the protein or limitations in sensitivity of our technique.

The antibodies employed in the immunohistochemical studies were prepared against a 25 amino acid synthetic peptide from a conserved region of c-fos responsible for mediating specific DNA binding (Franza et al. 1987; Nakabeppu and Nathans, 1989; Quinn et al. 1989). The span of 25 amino acids against which the antibodies were prepared is highly conserved for both fra-1 and fos-R (Zerial et al. 1989) and the antibodies will likely recognize any of these Fos family members. All of these Fos proteins are rapidly induced by growth factors and are capable of interacting with Jun proteins (Rauscher et al. 1988; Zerial et al. 1989) to form transacting complexes. Several lines of evidence indicate that the antibodies employed in these studies are Fos-specific: (1) the antibodies recognize AP-1 enhancer element complexes, (2) the antibodies bind primarily to the cell nucleus, and (3) elimination of antibody or competition with the Fos peptide result in the elimination of nuclear binding. As c-fos is rapidly induced in epidermal keratinocytes stimulated to differentiate (Dotto et al. 1986), we believe that c-fos is the major Fos protein of the intact epidermis.

The cumulative data suggest a fundamental role for Fos in regulation of late stages in the process of keratinization. Normal, keratinizing rodent tissues from diverse and different origins such as epidermis, oral epithelia, vagina and forestomach all show similar patterns of Fos expression. The primary similarity among these tissues is the appearance of Fos-positive cells just prior to cell death and cornification. Normal, nonkeratinized epithelia such as the junctional epithelium (Fig. 4), the vaginal epithelium of mice in diestrus (Fig. 5) and the embryonic epidermis (data not shown) show levels of Fos expression that are greatly reduced or undetectable by the methods employed.

The epidermis of the Er/Er mutant mouse fails to keratinize (Holbrook et al. 1982; Fisher, 1987; and Fisher et al. 1987a) and shows alterations in expression of Fos (Fig. 2D). Unlike normal, nonkeratinized epithelia, the Er/Er epidermis shows an abnormal expression of Fos, with virtually every nuclei staining positively. These results are enigmatic when held up against our results for Fos localization in normal tissues but we do not believe they exclude a central role for Fos in regulating keratinization. Fos expression probably represents an early step in a complex cascade of events that ultimately results in keratinization. The Er mutation may result in a block of this process downstream from Fos expression. Alternatively, there is evidence to suggest that the Er mutation acts systemically (Fisher, 1987; Fisher etal. 1987a). This would indicate that aberrant expression of Fos in the mutant epidermis may be due to abnormal systemic delivery of factors to the mutant epidermis. The aberrant expression of Fos in the Er/Er epidermis may explain, in part, the abnormalities in regulation of genes for epidermal structural proteins that have been reported for this mutation (Holbrook et al. 1982; Fisher, 1987; Fisher et al. 1987a). These results also suggest that other elements, such as the stage of differentiation of the cell in which Fos is expressed, may be important for cornification.

The Fos family of genes are all rapidly induced by growth factors in the presence of protein synthesis inhibitors (Greenberg and Ziff, 1984; Cohen and Curran, 1988; Zerial et al. 1989) and participate in the formation of a DNA binding, transcription activating complex (Rauscher et al. 1989; Turner and Tjian, 1989; Nakabeppu and Nathans, 1989). While these genes have been primarily studied as immediate-early genes in the mitogenic response, it is well established that c-fos is transiently induced in cells, including epidermal keratinocytes, stimulated to differentiate (Kruijer et al. 1984; Dotto et al. 1986). The Fos-positive cells in keratinizing tissues are not only well past competence for mounting a mitogenic response but are entering the final stages of differentiation and cell death.

Our results suggest an important role for Fos in regulating the terminal step in the process of keratinization. The data presented suggest that Fos expression is a highly conserved, fundamental mechanism by which a variety of epithelial cells from diverse embryological origin may control their final stages of differentiation. In order to address this possibility, it will be necessary to identify the genes upon which Fos is acting in these terminally differentiated cells.

The work was initiated in the Department of Biological Structure, University of Washington where C.F. was supported by NIH Grant HD 24443. M.B. is supported by NIH Grant DE 05159. The authors acknowledge the excellent technical assistance of Jude Rosenthal, Sue Gilbertson-Beadling, and Kelly Mecifi. We thank Tom Kawabe for providing sectioned mouse vibrissae follicles and Dr Allen Buhl for demonstrating the ovariectomy procedure. Margaret A. Kornacker and Barbara A. Moody are gratefully acknowledged for preparation of the manuscript, and Drs Karen A. Holbrook and Arthur Diani are thanked for helpful comments on the manuscript.

Barker
,
T. E.
and
Walker
,
B. E.
(
1966
).
Initiation of irreversible differentiation in vaginal epithelium
.
Anat. Rec
.
154
,
149
160
.
Bowden
,
P. E.
,
Quinlan
,
R. A.
,
Brettkreutz
,
D.
and
Fusenig
,
N. E.
(
1984
).
Proteolytic modification of acidic and basic keratins during terminal differentiation of mouse and human epidermis
.
Eur. J. Biochem
.
142
,
29
36
.
Brody
,
I.
(
1959
).
The keratinization of epidermal cells of normal guinea pig skin as revealed by electron microscopy
.
J. Ultrastruct. Res
.
2
,
482
511
.
Brody
,
I.
(
1960
).
The ultrastructure of the tonofibrils in the keratinization process of normal human epidermis
.
J. Ulrastruct. Res
.
4
,
264
297
.
Cohen
,
D. R.
and
Curran
,
T.
(
1988
).
Fra-1: a serum-inducible, cellular immediate-early gene that encodes a fos-related antigen
.
Molec. cell. Biol
.
8
,
2063
2069
.
Dale
,
B. A.
,
Holbrook
,
K. A.
and
Steinert
,
P. M.
(
1978
).
Assembly of stratum comeum basic protein and keratin filaments into macrofibrils
.
Nature
276
,
729
731
.
Dale
,
B. A.
,
Resing
,
K. A.
,
Haydock
,
P. V.
,
Fleckman
,
P.
,
Fisher
,
C.
and
Holbrook
,
K. A.
(
1988
).
Intermediate filament associated protein of the epidermis
.
In The Biology of Wool and Hair
(ed.
G. E.
Rogers
et al. 
)
Chapman and Hall
,
London and New York
.
Dignam
,
J. D.
,
Martin
,
P. L.
,
Shastry
,
B. S.
and
Roeder
,
R. G.
(
1983
).
Eukaryotic gene transcription with purified components
.
Methods Enzymol
.
101
,
582
598
.
Dotto
,
G. P.
,
Gilman
,
M. Z.
,
Maruyama
,
M.
and
Weinberg
,
R. A.
(
1986
).
C-myc and c-fos expression in differentiating mouse primary keratinocytes
.
EMBO J
.
5
,
2853
2857
.
Fisher
,
C.
(
1987
).
Abnormal development in the skin of the pupoid fetus mutant mouse; abnormal keratinization, recovery of a normal phenotype, and relationship to the repeated epilation (Er/Er) mutant mouse
.
Curr. Top. devl Biol
.
22
,
209
234
.
Fisher
,
C.
,
Jones
,
A.
and
Roop
,
D. R.
(
1987a
).
Abnormal expression and processing of keratins in pupoid fetus (pf/pf) mutant mice
.
J. Cell Biol
.
105
,
1807
1819
.
Fisher
,
C.
,
Haydock
,
P. V.
and
Dale
,
B. A.
(
1987b
).
Localization of profilaggnn mRNA in newborn rat skin by in situ hybridization
.
J. invest. Derm
.
88
,
661
664
.
Franza
,
B. R.
,
Sambucetti
,
L. C.
,
Cohen
,
D. R.
and
Curran
,
T.
(
1987
).
Analysis of Fos protein complexes and Fos-related antigens by high-resolution two-dimensional gel electrophoresis
.
Oncogene
1
,
213
221
.
Fuchs
,
E.
and
Green
,
H.
(
1980
).
Change in keratin gene expression during terminal differentiation of the keratinocyte
.
Cell
19
,
1033
1042
.
Greenberg
,
M. E.
and
Ziff
,
E. B.
(
1984
).
Stimulation of 3T3 cells induces transcription of the c-fos proto-oncogene
.
Nature
311
,
433
438
.
Holbrook
,
K. A.
(
1989
).
Biologic structure and function, perspectives on morphologic approaches to the study of the granular layer keratinocytes
.
J. invest. Derm
.
92
,
845
1045
.
Holbrook
,
K. A.
,
Dale
,
B. A.
and
Brown
,
K. S.
(
1982
).
Abnormal epidermal keratinization in the repeated epilation mutant mouse
.
J. Cell Biol
.
92
,
387
397
.
Kerr
,
L. D.
,
Holt
,
J. T.
and
Matrisian
,
L. M.
(
1988
).
Growth factors regulate transin gene expression by c-Jos-dependent and c-/os-independent pathways
.
Science
242
,
1424
1427
.
Kruuer
,
W.
,
Cooper
,
J. S.
,
Hunter
,
T.
and
Verma
,
I. M.
(
1984
).
Platelet-derived growth factor induces rapid but transient expression of the c-fos gene protein
.
Nature
312
,
711
716
.
Lavker
,
R. M.
and
Matoltsy
,
A. G.
(
1970
).
Formation of homy cells: the fate of cell organelles and differentiation products in ruminal epithelium
.
J. Cell Biol
.
144
,
501
520
.
Mackenzie
,
I. D.
(
1969
).
Ordered structure of the stratum corneum of mammalian skin
.
Nature
222
,
881
882
.
Menon
,
G.
,
Grayson
,
S.
and
Elias
,
P.
(
1985
).
Ionic calcium reservoirs in mammalian epidermis: ultrastructural localization by ion-capture cytochemistry
.
J. invest. Derm
.
48
,
508
512
.
Molders
,
H.
,
Jenuweln
,
T.
,
Adamkiewiez
,
J.
and
Moller
,
R.
(
1987
).
Isolation and structural analysis of a biologically active chicken c-fos cDNA: Identification of evolutionarily conserved domains in fos protein
.
Oncogene
1
,
377
385
.
Nakabeppu
,
Y.
and
Nathans
,
D.
(
1989
).
The basic region of the Fos mediates specific DNA binding
.
EMBO J
.
8
,
3833
3841
.
Potten
,
C. S.
(
1981
).
Cell replacement in epidermis (keratopoiesis) via discrete units of proliferation
.
I nt. Rev. Cytol
.
69
,
271
318
.
Quinn
,
J. P.
,
Masato
,
T.
,
Iadarola
,
M.
,
Holbrook
,
N.
and
Levens
,
D.
(
1989
).
Distinct factors bind the AP-1 consensus sites in gibbon ape leukemia virus and simian virus 40 enhancers
.
J. Virol
.
63
,
1737
1742
.
Rauscher
,
F. J.
,
Cohen
,
D. R.
,
Curren
,
T.
,
Bos
,
T. J.
,
Vogt
,
P. K.
,
Bohmann
,
D.
,
Tjian
,
R.
and
Franza
,
B. R.
(
1988
).
Fos-associated protein p39 is the product of the jun protooncogene
.
Science
240
,
1010
1016
.
Resing
,
K. A.
,
Walsh
,
K. A.
and
Dale
,
B. A.
(
1984
).
Identification of two intermediates during processing of profilaggrin to filaggrin in neonatal mouse epidermis
.
J. Cell Biol
.
99
,
1372
1378
.
Rice
,
R. H.
and
Green
,
H.
(
1979
).
Presence in human epidermal cells of a soluable protein precursor of the cross-linked envelope: activation of the cross-linking by calcium ions
.
Cell
18
,
681
694
.
Schonthal
,
A.
,
Horlich
,
P.
,
Rahmsdorf
,
H. J.
and
Ponta
,
H.
(
1988
).
Requirement for fos gene expression in the transcriptional activation of collagenase by other oncogenes and phorbol esters
.
Cell
54
,
325
334
.
Singh
,
H.
,
Sun
,
R.
,
Baltimore
,
D.
and
Sharp
,
P. A.
(
1986
).
A nuclear factor that binds to a conserved sequence motif in transcriptional control elements of immunoglobulin genes
.
Nature
319
,
154
158
.
Sun
,
T.-T.
,
Eichner
,
R.
,
Schermer
,
A.
,
Cooper
,
D.
,
Nelson
,
W. G.
and
Weiss
,
R. A.
(
1984
).
Classification, expression, and possible mechanism of evolution of mammalian epithelial keratins: a unifying model
.
In Cancer Cells I, The Transformed Phenotype
.
Cold Spring Harbor Laboratory
,
Cold Spring Harbor, NY
, pp.
169
171
.
Turner
,
R.
and
Than
,
R.
(
1989
).
Leucine repeats and an adjacent DNA binding domain mediate the formation of functional cFos-cJun heterodimers
.
Science
243
,
1689
1694
.
Van Beveren
,
C.
,
Van Straaten
,
F.
,
Curran
,
T.
,
Moller
,
R.
and
Verma
,
I. M.
(
1983
).
Analysis of FBJ-MuSV provirus and c-fos (mouse) gene reveals that viral and cellular for gene products have different carboxy termini
.
Cell
32
,
1241
1255
.
Van Straaten
,
F.
,
Moller
,
R.
,
Curran
,
T.
,
Van Beveren
,
C.
and
Verma
,
I. M.
(
1983
).
Complete nucleotide sequence of a human c-onc gene: deduced amino acid sequence of the human c-fos protein
.
Proc. natn. Acad. Sci. U.S.A
.
80
,
3183
3187
.
Zerial
,
M.
,
Toschi
,
L.
,
Ryseck
,
R.-P.
,
Schuermann
,
M.
,
Muller
,
R.
and
Bravo
,
R.
(
1989
).
The product of a novel growth factor activated gene, fos interacts with JUN proteins enhancing their DNA binding activity
.
EMBO J
.
8
,
805
813
.