Probing Keratinocyte and Differentiation Specificity of the

MOLECULAR AND CELLULAR BIOLOGY, June 1993, p. 3176-3190 0270-7306/93/063176-15$02.00/0 Copyright © 1993, American Society for Microbiology

Vol. 13, No. 6

Probing Keratinocyte and Differentiation Specificity of the Human K5 Promoter In Vitro and in Transgenic Mice CAROLYN BYRNE AND ELAINE FUCHS* Department of Molecular Genetics and Cell Biology, Howard Hughes Medical Institute, The University of Chicago, Chicago, Illinois 60637 Received 13 November 1992/Returned for modification 11 January 1993/Accepted 17 February 1993

Intermediate filament proteins form cytoskeletal networks which display cell type specificity. Keratins, the intermediate filament proteins of epithelial cells, are subdivided into two groups, type I and type II, which are expressed as pairs, forming obligatory heteropolymers (reviewed references 1 and 70). There are -15 individual keratin pairs which are differentially expressed in epithelial tissues at various stages of development and differentiation (43, 71). For example, Kl and K10 are expressed suprabasally in the epidermis and in the inner root sheath of the hair follicle (16, 43, 68), K4 and K13 are expressed in the suprabasal cells of esophagus and other internal stratified epithelia (73a), K3 and K12 are expressed in differentiating corneal epithelium (57, 71), and K6 and K16 are expressed in the suprabasal cells of the outer root sheath (ORS) of the hair follicle (28, 67). K5 and K14 are interesting in that they are broadly expressed in most stratified epithelial tissues (45). In these tissues, the pair is expressed in mitotically active basal keratinocytes, as judged by in situ hybridizations and/or immunohistochemical localization of epidermis (32, 42, 51), rodent forestomach (60), tongue (11), and various other stratified and pseudo-stratified epithelia (42). In addition, K5 has been detected histologically in a number of epithelial cells, including human thymic reticulum, the ORS, sebaceous glands, and both basal and myoepithelial cells of eccrine sweat glands and salivary glands (42). Indeed, an important and unifying feature of basal keratinocytes appears to be their unique ability to express K5 and K14, irrespective of body location (45). Given that cell-typespecific expression of KS and K14 genes appears to be at the transcriptional level (33, 69), the promoters of these genes

*

seem to display an insensitivity to the varied environments of keratinocytes. Little is known about the molecular mechanisms which govern the complex expression of genes in the keratinocyte. Given that (i) K5 and K14 account for up to 30% of total protein of mitotically active keratinocytes and (ii) the tissuespecific expression of these proteins is regulated transcriptionally (33, 69), K5 and K14 genes provide ideal tools for investigating these regulatory mechanisms. The human KS and K14 genes have been cloned and sequenced (33, 37). Cell type specificity of the KS promoter has been suggested from tissue culture studies (22, 46), although the K14 promoter is promiscuously expressed in a number of cell types in vitro (30, 75). While the in vivo behavior of the K5 promoter has not yet been explored, 2,100 bp of the K14 promoter is sufficient to direct expression of heterologous genes to the basal cells of stratified epithelia in transgenic mice (7, 73, 74). A number of recent studies have begun to focus on identifying putative regulatory domains and transcription factors that may be involved in expression of either endogenous or viral genes in keratinocytes (3, 4, 8, 9, 20, 22, 24, 29, 30, 35, 37, 46, 52, 54, 65, 66, 72, 75). Thus far, only a few nuclear factors that both are restricted in their tissue specificity and appear to participate in conferring keratinocytespecific gene expression have been identified. One of these, AP2, has been implicated broadly in the regulation of endogenous and viral genes that are typically expressed in keratinocytes (29, 30, 40, 54, 65, 66). However, AP2 does not appear to be sufficient for keratinocyte specificity on its own (29). The aim of this work was to explore the regulatory mechanisms involved in controlling keratinocyte-specific gene expression in culture and in transgenic animals. We

Corresponding author. 3176

Downloaded from http://mcb.asm.org/ on December 18, 2017 by guest

Keratins KS and K14 form the extensive intermediate filament network of mitotically active basal cells in all stratified epithelia. We have explored the regulatory mechanisms governing cell-type-specific and differentiation stage-specific expression of the human K5 gene in transiently transfected keratinocytes in vitro and in transgenic mice in vivo. Six thousand base pairs of 5' upstream K5 sequence directed proper basal cell-specific expression in all stratified epithelia. Surprisingly, as few as 90 bp of the K5 promoter still directed expression to stratified epithelia, with expression predominantly in epidermis, hair follicles, and tongue. Despite keratinocyte-preferred expression, the truncated K5 promoter displayed departures from basal to suprabasal expression in epidermis and from outer root sheath to inner root sheath expression in the follicle, with some regional variations in expression as well. To begin to elucidate the molecular controls underlying the keratinocyte specificity of the truncated promoter, we examined protein-DNA interactions within this region. A number of keratinocyte nuclear proteins bind to a K5 gene segment extending from -90 to +32 bp and are functionally involved in transcriptional regulation in vitro. Interestingly, several of these factors are common to both the K5 and K14 promoters, although they appear to be distinct from those previously implicated in keratinocyte specificity. Mutagenesis studies indicate that factors binding in the vicinity of the TATA box and transcription initiation site are responsible for the cell type specificity of the truncated K5 promoter.

VOL. 13, 1993

KERATINOCYTE-SPECIFIC TRANSCRIPTIONAL REGULATION

MATERIALS AND METHODS Construction of recombinant clones. pK5,Bgal6000 (Fig. 1A) contains 6,000 bp of the human K5 promoter, extending from an EcoRI site 5' to a SacI site 32 bp 3' to the transcription initiation site (33). This fragment was cloned into the eukaryotic expression vector pNass, (Clontech, Palo Alto, Calif.), referred to as p1gal in this report. Other pKSgal constructs were engineered similarly to pK5,Bgal6000, but they contain various amounts of K5 promoter sequences. Plasmid pTK1gal contains the truncated thymidine kinase promoter SalI-XhoI fragment from pBLCAT2 (34) cloned into pNass,B. pK5CAT constructs were similarly generated from pSVOcatPL1 (17, 30). All recombinant DNA technology was performed by standard methodology (36), and DNA sequencing was performed with a Sequenase kit (United States Biochemical, Cleveland, Ohio). Cell culture, transfections, 13-Gal assays, and CAT assays. Mouse NIH 3T3 fibroblast cells (American Type Culture Collection, Rockville, Md.) and human HepG2 hepatoma cells (American Type Culture Collection) were cultured in a 3:1 mixture of Dulbecco's modified Eagle medium and Ham's F-12 nutrient mixture, supplemented with 10% newborn calf serum. The human SCC-13 squamous cell carcinoma line (a gift from J. Rheinwald [76]) were grown as described previously (30). Transfections were always performed in triplicate (30), using the following per 100-mmdiameter dish of cells: 6 x 10-12 mol of K5 promoterreporter gene plasmid, 5 ,ug of pSV2cat (chloramphenicol acetyltransferase [CAT] reporter) (17) or pCH110 (v-Gal reporter) as an internal standard to control for variation in transfection efficiencies, and the appropriate amount of Bluescript-pKS+ (Stratagene, La Jolla, Calif.) to standardize the total amount of DNA used to 50 ,ug per reaction. At 65 h following transfection, cells were harvested, extracts were prepared, and total protein was quantitated (5). CAT assays were performed by using an enzyme-linked immunosorbent assay (Boehringer Mannheim, Indianapolis, Ind.). 1-Gal assays were performed as described previously (12). As a positive control, pTK,Bgal was tested in parallel (and in triplicate) with all K5 constructs. Preparation of nuclear extracts, EMSA, and methylation interference assays. Nuclear extract preparations, the electrophoretic mobility shift assay (EMSA), and methylation

pK50gaI6000

A

Sphl Hindlil BamHl Sall

K5 promoter p-galactosidase gene EcoFi HindiliSphl Hindlil BamHlNheISphl

ri e VI 4

=:: : .

:

,

SV40 splice donor/splice acceptor

r

M

pUC19

SV40 polyadenylation

4800 3400 800 250 90

B 100,

I

I

o

SCC3 * HepG2

0

NIH3T3

0

co 0 0

0o 4L 6000

.AN- I

4800

3400

800 250 promoter iength

90

M.P= 0

FIG. 1. (A) DNA constructs used in transient transfection analysis and generation of transgenic mice. Plasmids containing serial deletions of the human K5 promoter linked to a reporter gene, either 1-Gal (shown) or CAT (not shown), were prepared for transient transfection assays. Open bar, the 6,000-bp K5 promoter sequence; hatched bar, the 1-Gal gene; closed bars, simian virus 40 (SV40) sequences; thin line, bacterial sequences from plasmid pUC19. Deletions were made by using restriction endonucleases to remove 5' portions of the K5 promoter. (B) Histograms showing the cell type specificity of 1-Gal expression in cultured cells. Keratinocytes (SCC-13), liver epithelial cells (HepG2), and fibroblasts (NIH 3T3) were transfected with K5 reporter gene (1-Gal and CAT) constructs, and 65 h later, cells were harvested and assayed for 1-Gal or CAT activity as described in Materials and Methods. Error bars represent 1 standard deviation. The reporter activity of the 6,000-bp KS promoter in SCC-13 cells was set at 100%.

interference assays were performed as described previously (14, 30) except that for Spl binding reactions, ZnSO4 (0.067 ,uM) was included, and glycerol and KCI concentrations were increased to 20% (vol/vol) and 100 p,M, respectively. Oligonucleotides containing Spl and AP2 consensus binding sites (Stratagene) and purified Spl and AP2 protein (Promega Corp., Madison, Wis.) were purchased. Mutagenesis of the K5 promoter. Mutations were engineered in pK5-KS+ containing K5 promoter DNA inserted into pKS+ (Stratagene). Mutants were generated by sitedirected mutagenesis (27, 39), and a fragment containing the mutation was then cloned back into pNass1 as follows: CAGCCC (-88 to -83) was replaced with CTCGAG, CAGCCC (-48 to -43) was replaced with CTCGAG, GCATCA (-18 to -13) was replaced with CACGTG, and CTGGGT (-6 to -1) was replaced with AGATCT. Preparation of transgenic mice, immunohistochemistry, and 13-Gal histochemistry. Transgenic mice were prepared as described previously (75) and were identified by using transgene-specific primers in a polymerase chain reaction of ear DNA. Expression studies were performed on F1 generation


began by engineering a fusion gene containing 6,000 bp of 5' upstream sequence of the human KS gene linked to the 13-galactosidase (,B-Gal) reporter gene. In transiently transfected keratinocytes and in transgenic mice, the K5 sequences were sufficient for appropriate tissue-specific and differentiation-specific gene expression. In delineating the limit sequences necessary for proper tissue-specific gene expression, we learned that a mere 90 bp of KS promoter sequence could still confer largely cell type specificity, albeit not differentiation specificity, in culture and in transgenic mice. Surprisingly, this sequence did not contain binding sites for AP2 or any other transcription factor thus far implicated in keratinocyte-specific gene expression. Our studies further revealed that factors binding in the vicinity of the TATA box and transcriptional initiation site of the K5 gene play an important role in keratinocyte specificity. Taken together, our observations provide important new insights into our understanding of differentiation-specific and tissue-specific gene expression in stratified epithelial tissues and suggest that there may be multiple ways of achieving this specificity.

3177

3178

BYRNE AND FUCHS

RESULTS The K53gal gene is expressed in a cell-type-specific fashion in vitro. The parent K5,Bgal fusion gene construct used in this study, along with restriction sites used for truncated versions, is illustrated in Fig. 1A. When 6,000 bp of the human K5 promoter was used to drive expression of the Escherichia coli lacZ (or CAT) gene in tissue culture cells, appreciable 1-Gal (or CAT) activity was detected in the human epidermal keratinocyte line SCC-13 (K5+) (Fig. 1B). In contrast, little or no expression was detected in HepG2 liver hepatocytes, which express K8 and K18 (K5- [29]), or in NIH 3T3 fibroblasts, which express vimentin (K5- [15, 43]). These results suggested that 6,000 bp of the human KS gene was sufficient to confer cell-type-specific gene expression in vitro. This finding agreed with the previous results of Jiang et al. (22), who showed that 560 bp of the human KS promoter was sufficient to drive appreciable CAT expression in cultured esophageal and corneal keratinocytes, with only a very low expression in simple (HeLa) epithelial cells and no detectable expression in NIH 3T3 fibroblasts. To define the minimal KS sequences necessary to confer this cell type specificity, we made and tested a nested series of pK5Sgal6000 deletion mutants, each containing sequentially less of the 5' KS promoter sequence (Fig. 1A). Surprisingly, deletions down to 90 bp of the K5 promoter failed to result in a loss of cell type specificity (Fig. 1B). Thus, even though pK513gal90 was expressed in keratinocytes at -30% the level of pKSBgal6000, expression was nevertheless promoted preferentially in epidermal keratinocytes. These results extended the observations of Ohtsuki et al. (46), who found that 542 bp of the human K5 promoter conferred higher CAT expression in cultured human epidermal cells than in 3T3 fibroblasts. Generation of human K5 promoter-driven 13-Gal-expressing transgenic mice. While the cell type specific activity of K513gal90 in our three cell lines seemed promising, mammals have hundreds of cell types, and consequently it is impractical, if not impossible, to conduct a thorough analysis of cell type specificity by using cell culture studies. Therefore, we introduced our fusion genes into the germ line of mice. Twelve transgenic K5S1gal6000 founder animals were generated. Of these, nine exhibited expression of the transgene, as judged by direct detection of 13-Gal activity in frozen tail sections. Two K5Sgal6000 mice were bred for analysis of the F1 generation. The patterns of transgene expression for these two lines were similar, suggesting that expression patterns were dictated by the K5 regulatory sequences rather than transgene integration site. Ten transgenic

K51gal9O founder animals were generated, only two of which exhibited 1-Gal expression. Again, two independently derived transgene-expressing animals were bred for further analysis of the F1 generation, and with a few exceptions (see below), patterns of transgene expression were similar for the two lines. The low percentage of animals expressing KSgal90 is most likely reflective of the reduced ability of the highly truncated promoter to drive appreciable transgene expression (Fig. 1A). Under conditions in which promoter strength is reduced, transgene expression would be expected to be more sensitive to position and copy number variations. The K50gal6000 gene is expressed in a keratinocyte-specific fashion in vivo. In most tissues, the pattern of expression of the K5Sgal6000 gene was indistinguishable from that of the endogenous KS gene, whose expression was monitored in parallel, using an anti-K5 antiserum (33). Examples are shown in Fig. 2A (1-Gal) and B (anti-K5), in which tailskin of an animal expressing K5Sgal6000 exhibited transgene expression in the basal epidermal layer and in the hair follicle ORS. Similar to KS activity, 13-Gal activity was found in the sebaceous and eccrine glands of skin (not shown). Overall, a K5/K14 pattern of transgene expression was found in tailskin of all nine KSBgal6000 founders. The pattern of transgene expression paralleled KS expression in other stratified squamous epithelia as well. For example, the basal layer of both dorsal (shown) and ventral (not shown) tongue epithelium coexpressed 13-Gal and K5 (Fig. 2C and D, respectively). These patterns coincided with the known pattern of K14 mRNA and protein expression in tongue (11, 48). Similarly in cornea, 1-Gal activity was confined to the limbal region (Fig. 2E, left side of corneal section), where the K5- and K14-expressing mitotically active stem cells are thought to reside (57). The extension of anti-K5 staining to the differentiating cells of the nonlimbal cornea (Fig. 2F, right side of section) is most likely a reflection of the extraordinary stability of K5 in 10-nm filaments (33). In a few epithelial tissues, 13-Gal activity was absent in a subset of basal cells. This was most pronounced in the esophagus (shown) and the forestomach (not shown), where only a few basal cells expressed the transgene (Fig. 2G), even though anti-K5 staining was detected in all basal cells (Fig. 2H) (42). This severely restricted pattern of 1-Gal relative to KS was seen in one other tissue, the thymus. Here, only an occasional reticular epithelial cell exhibited 1-Gal activity (Fig. 2I and J; see also reference 42). The restriction of 1-Gal activity to a subset of K5-positive cells in these three tissues did not appear to be due to uneven 1-Gal detection, since (i) we assayed frozen sections of tissues for X-Gal cleavage and generation of the diagnostic blue color and (ii) 13-Gal was detected throughout the basal cells of epidermis, tongue, and cornea. Rather, it was reminiscent of the heterogeneity in basal keratin mRNA expression reported for some epithelial basal populations that exhibited homogeneous basal keratin protein presence (60). Overall, basal cell-preferred activities of the transgene and endogenous KS promoters in stratified squamous epithelial tissues appeared to be largely similar. Moreover, when nonstratified epithelial tissues were examined, only a single abnormality in transgene-driven 1-Gal was found. This was in brain, where the transgene was expressed in a subset of neurons, including the pyramidal cell and CA1, CA2, and CA4 layers of the hippocampus (Fig. 2K) and one of the layers of the cortex (Fig. 2L). As for transgene expression patterns in other tissues, expression in the brain was observed in both lines of


transgenic animals. Immunohistochemistry was performed on 5-,um sections of paraffin-embedded, Bouin's-fixed tissue samples. Immunogold enhancement detection of tissuebound antibodies was performed with an Auroprobe LM GAR kit (Amersham, Arlington Heights, Ill.). For 1-Gal detection, frozen tissues were embedded in optimal cutting temperature (OCT) compound at -120°C, sectioned (20 ,um), and fixed in 0.5% glutaraldehyde for 2 min. 3-Gal activity was detected by its ability to convert colorless 5-bromo-4-chloro-3-indolyl-13-D-galactopyranoside (X-Gal) to a blue end product (reference 77 and references therein). Occasionally, tissues were prefixed for 1 h in 1% formaldehyde-0.2% glutaraldehyde, stained for 13-Gal, postfixed in 2% formaldehyde-1% glutaraldehyde, paraffin embedded, and sectioned. Sections were counterstained with hematoxylin and eosin.

MOL. CELL. BIOL.


VOL. 13, 1993

A

B

*C

9%

.D

3179

7)%. ~a

jp~~~~~~~~~u

H

F

~

~

~

~

~

~

1"

FIG. 2. Expression of K5,gal6000 compared with endogenous K5 in transgenic mouse tissues. Tissues of transgenic animals expressing KS,gal6000 were processed and sectioned as described in Materials and Methods. Tissue sections were assayed for P-Gal activity (A, C, E, G, I, K, and L) or anti-KS staining (B, D, F, H, and J). The blue-green P-Gal reaction product denotes activity of the K5 promoter; the brown-black color is the immunogold labeling of anti-K5 binding. (A and B) Tail; (C and D) dorsal tongue; (E and F) limbal region of cornea; (G and H) esophagus; (I and J) thymus; (K and L) brain. The four stages of epidermal differentiation are indicated in panels A and B. bl, basal layer; sp, spinous layer; gr, granular layer; sc, stratum corneum; hf, hair follicle. Bars: 20 ,um (A and B), 30 p.m (C to F, I, and J), 60 p.m (G, H, and L), and 120 pum (K).

K5jgal6000 F1 mice. A summary of the overall expression patterns of K50gal6000 relative to K5 is provided in Table 1. Maintenance of keratinocyte specificity but altered differentiation specificity of severely truncated K5 promoters. The K5,gal90 mice exhibited lower transgene expression than did the K5Pgal6000 mice, as judged by the ability of tissues to convert X-Gal to the blue end product. This difference correlated with the reduced expression levels of K5,Bgal90 relative to K5Pgal6000 in cultured keratinocytes. Surprisingly, despite reduced expression levels, the two different K5,Bgal9O-expressing lines both exhibited patterns of transgene expression that were largely restricted to stratified epithelia and their appendages. Among epithelial tissues, p-Gal detection appeared restricted to the epidermis, hair follicles, and tongue (Fig. 3).

The keratinocyte specificity of the 90-bp promoter was perhaps most readily apparent in skin, where at least 20 other cell types are present (Fig. 3A to J). Remarkably, while expression of the truncated promoter in the skin was primarily restricted to the epidermis and hair follicles, it was accompanied by a striking switch in differentiation specificity. Instead of expression in mitotically active keratinocytes, expression was confined to terminally differentiating cells. For example, in the thickened plantar and palmar epidermis, the truncated promoter directed n-Gal activity to the granular layer (Fig. 3A). This pattern was similar to that of filaggrin, an intermediate filament-associated protein (10), and somewhat similar to that of K9, a keratin often expressed in the later stages of terminal differentiation in thick-skinned regions of the body (41). The pattern was in


E

3180

MOL. CELL. BIOL.

BYRNE AND FUCHS

TABLE 1. Summary of expression patterns of endogenous K5, K5,Bgal6000, and K50gal90 in transgenic mice Expression

Location

Endogenous

K5pgal90

Stratified epithelial cells and appendages Skin Tongue Esophagus Forestomach Cornea (limbal region) Hair follicle Sebaceous gland Eccrine gland

+++++ +++++ +++++ +++++ +++++

+++++ +++++ +++ +++ +++++

+++ +++

+++++ ++

+++++

+++

++ ++

+ NDa

Other tissues Liver Kidney Pancreas Gallbladder Gut Heart Smooth muscle Striated muscle Body wall muscle Ovary Oviduct Testis Pituitary Brain

++ -

-

-

-

+b

-

-

-

-

-

-

-

-

-

-

-

-

-

-

+b +b

-

-

ND

-

+

-

+

+

a ND, not done. Expression was apparent in one line of 90 mice only.

b

striking contrast to K5,Bgal6000, expression of which was confined to the basal epidermal layer (Fig. 3B). The differentiation specificity of the truncated promoter varied markedly with body location. In contrast to palmar and plantar epidermis, the thinner epidermis of the ear and backskin exhibited K5Pgal90 expression in the spinous and granular layers (Fig. 3C). Again, this was diametrically opposite the strict basal expression of K5Pgal6000 (Fig. 3D). This switch from basal to suprabasal expression did not appear to be due to positional effects of chromosomal integration; both K5Pgal90 mice exhibited this pattern, as did transgenic mice that were expressing 13-Gal transgenes driven by another severely truncated KS promoter, K5Pgal250 (Fig. 3E). Interestingly, this pattern of suprabasal expression in ear and backskin epidermis paralleled that of Kl and K10, expressed naturally in the differentiationspecific cells of these tissues (16, 51). In tail, expression of K5Pgal9O was unusual in that a lateral heterogeneity was observed, with alternating patches of suprabasal expression interspersed with regions of no expression (Fig. 3F [low magnification] and G [higher magnification]). Murine tail consists of concentric rings of scales interposed by nonscaly regions. Scale epidermis has a parakeratotic morphology, with eosinophilic, nucleated cells in cornified layers and a notably absent granular layer; interscale epidermis has orthokeratotic morphology more typical of other epidermal regions (59). Transgene expression of K5RSgal9O was uniform throughout orthokeratotic segments and absent in parakeratotic regions. Interestingly, while Kl

expression. While the few cases of aberrant expression outlined above

were obvious, they were detected only in the higher-express-


KSjgal6000

K5

and K10 are expressed suprabasally throughout tail epidermis, a novel 70-kDa type II mouse keratin exhibits a pattern of expression indistinguishable from that of K51gal90 (49). Anti-K6 staining also revealed an orthokeratotic-specific pattern in tail (not shown), suggesting that the promoter mimicked expression of the 70-kDa keratin and K6/K16 in this body location. The differentiation specificity of K5,Bgal90 was also altered in epidermal appendages. Thus, in contrast to the endogenous K5 gene or the K5S,gal6000 transgene, K51gal90 was inappropriately expressed in the inner root sheath cuticle of the follicle rather than in the ORS (Fig. 3H and I, K5Pgal90; compare with K51gal6000 expression in Fig. 3J). Interestingly, whereas K6 and K16 are expressed in inner layers of the ORS, Kl and K10 expression is restricted to the inner root sheath (19, 68). Thus, in tail and follicle, where both the K6/K16 and K1/K1O pairs of suprabasal keratins are expressed, K51gal90 expression mimicked K6 and K16 expression in tailskin and Kl and K10 expression in the follicle. The pattern of expression of K5,gal90 in the tongue was similar in some respects to that seen in the tail (Fig. 3K). The filiform papillae of the tongue consist of two morphologically distinct compartments, an anterior, orthokeratinized epithelium and a posterior epithelium which gives rise to cornified spines (21, 48). In sections of dorsal tongue epithelium, the 90-bp K5 promoter was expressed in the interpapillary suprabasal epithelium (thin arrow) and the anterior orthokeratinized epithelium of the filiform papilla (arrowheads). This was in striking contrast to K5,gal6000, which was appropriately expressed in the basal and parabasal cells (Fig. 3L), similar to the endogenous K5 promoter (48). Keratin expression in the differentiating cells of the tongue is complex. K6 and K16 are expressed in the posterior segments (5a, 48). K13 is expressed in the orthokeratinized anterior segments, and K4 and K13 are both expressed in the interpapillary suprabasal epithelium (11, 49). While expression of K5Pgal90 paralleled that of K13 in tongue, ,B-Gal was not detected in other internal stratified epithelia, e.g., esophagus and forestomach, where this keratin is also expressed (60, 71). Thus, although transgene expression driven by the truncated KS promoter shared certain similarities with expression patterns of differentiation-specific keratin genes of skin and tongue, the overall pattern of K5,Bgal90 did not perfectly coincide with that of any known keratin. While expression of K5,gal90 was largely restricted to keratinocytes, some expression was detected in nonkeratinocyte cell types. In the dorsal region of the tongue, expression of 3-Gal was detected in cells accumulating in papilla regions of the underlying mesenchyme (thick arrow in Fig. 3K; compare with K51gal6000 expression in Fig. 2C). A few cells just beneath the basement membrane of tail and palmar/ plantar skin also exhibited 13-Gal activity (not shown). Electron microscopy indicated that the blue-stained mesenchymal cells were fibroblasts (not shown), indistinguishable from those unstained cells residing deeper in the mesenchyme. The close proximity of these cells to the basal keratinocytes was intriguing and suggested the possibility that some factor produced by keratinocytes is involved in induction of this expression. However, aberrant expression was also detected in some smooth muscle cells, in body wall muscle, in brain, and also in epithelial cells of the thick ascending limb of the kidney tubules. Thus, close proximity to keratinocytes was not essential for aberrant transgene

VOL. 13, 1993

A*lv

I

"

i

~

,. . i le

00

If

.

'.

,

...

16

*.

,;

*,EW

Aw

4

&-

6

p

-0

-

*-

.

,s

r_j

g

we C' 4i

.

4.q '4

'ok

.0

0

z.-,

v, .,

v

.

F

E

Ie

.1

D

i-u.'-

-

--:.M7"

0

-.-,

..

c

.Mvf uo; iI .4

..

f

B

3181

w.

I

v

w,

I\,,.

.

P.;. .o

.41

.

iol. V

I

!klt V

A

1.

I,~

~ ~ ~~T A

4p. ._

FIG. 3. Alteration of differentiation specificity without compromising stratified squamous epithelial expression in severely truncated K5 promoter constructs. Tissues from two K5,gal90, two K5,gal6000, and one K5pgal250 mouse were processed and sectioned (see Materials and Methods). Sections were assayed for 13-Gal activity. (A, C, F, G, H, I, and K) Sections from a K5,Bgal90 mouse; (E) section from a K5Pgal250 mouse; (B, D, J, and L) sections from a K5,Bgal6000 mouse. (A and B) Palmar epidermis; (C and D) ear epidermis; (E to G) tail epidermis; (H to J), hair follicle; (K and L) dorsal surface of tongue. co, cortex; c, cuticle; i, inner root sheath; o, outer root sheath. Other abbreviations are as in the legend to Fig. 2. Bars: 60 p.m (A, B, K, and L), 20 p.m (C, D, I, and J), 30 pLm (E), 120 p.m (F), 12 p.m (G), and 9 p.m (H). The patterns of 1-Gal expression in stratified epithelia were similar for F1 offspring of different founder mice expressing the same transgene, indicating that differences in keratinocyte expression patterns were not due to variations in chromosomal integration sites.

ing K53gal9O mouse line, and hence it is possible that the aberrations arose from an insertion site artifact. Alternatively, it could be that the aberrant expression was inherent within the truncated promoter and that the levels of 1-Gal produced in the lower-expressing K5S,gal90 mouse line were simply below the limits of detection. Given that detection of transgene expression in the nonstratified squamous epithelial tissues of the higher-expressing K5,Bgal90 mouse required prolonged incubations with X-Gal, this possibility was plausible. Nevertheless, given the severely truncated nature of this promoter, it was still surprising that K5S,gal90 was as faithful as it was in its expression pattern, appearing nearly exclusively in the tongue and skin of the two transgenic lines examined. A summary of the cell type and tissue distribution of K5S,gal90 expression is provided in Table 1.

The 90-bp KS promoter binds a number of keratinocyte nuclear proteins: preliminary characterization. Since our transgenic mouse studies indicated that K5Pgal90 was still largely restricted to keratinocytes in vivo, we focused on the nuclear proteins that might be involved in conferring this specificity. We first examined the ability of this DNA segment to bind proteins isolated from human keratinocyte (SCC-13) nuclear extracts. Our DNA probe was a 122-bp SphI-SacI fragment, spanning the sequence -90 (5') to +32 (3') from the K5 transcription initiation site (Fig. 4A). When radiolabeled double-stranded probe was combined with nuclear extract isolated from SCC-13 keratinocytes, a series of complexes were produced that were resolved by nondenaturing acrylamide gel electrophoresis and showed significantly slower mobility than did the probe itself (Fig. 4B and


X; tip.

.1

1


3182

BYRNE AND FUCHS

MOL. CELL. BIOL.

A SphI

Sacl

-0

GAGCTCACAGWTItjVtCCCCAGG TGCCCAGOOCACTTAATCATTCACAGCTCGACAGCTCTCTCGCCCAGCCG K5-6 SpI

K5-5

K5-4

WI

SP'. protein1 c

x 3

transcriotiorn star

Sacl

t

pt

+

AGTTCTGGAAGGGATAAAAAGGGGGCATCACCGTTCCTGGGTAACAGAGCCACCTTCTGCGTCCTGCTGAGCTC K5-3 protein complex 4

m

K5-2

Drotein(s) 1-2

K5-i

C

K5 -116/-22

B K5 -90/+32 COm

epidermal extract

+

_

+

+

_

KS-1

+

+

+

+

+

+

+

+

K5-2 K5-3 KS 4 KS-5 K5-6

+

.

-

U)

-

+

SRil

SP1

0o

e x t r a c t ;D Q) ~Qc a) ...... extract

Q

Comp. _K 5h ......... .4.....

.

..4t4 ......

'#4

...........

MiIAj

-4--

4

SPi AP2 5Ki

Sp,

FIG. 4. Identification of nuclear proteins which bind to the truncated K5 promoter. (A) DNA sequence of the human KS promoter extending from -116 to +32 (33). Sequences used for double-stranded oligonucleotide competitors are indicated by black bars. Sequences with homology to Spl sites are shaded grey, and a sequence shown to bind AP2 is boxed. Arrowheads indicate the nucleotides which bind AP2 and protein(s) 1-2 as judged by methylation interference assays (see text). (B) EMSA identifying nuclear proteins which bind to the truncated KS promoter. A fragment of the KS promoter from 90 bp 5' to the transcription initiation region (-90) to 32 bp 3' to the transcription initiation region (+32) was used as the radiolabeled probe. EMSAs using human epidermal cell nuclear extracts as a source of protein were carried out. Competitor (Comp.) sources were the -90+32 K5 promoter fragment (K5), K5-1 to K5-6 oligomers, and oligomers containing an Spl binding site (see Materials and Methods). In most cases, competitors were used at two concentrations, 100- and 200-fold molar excess (first and second lanes, respectively, of each set). Unidentified protein-DNA complexes are labeled 1 to 4, and their approximate locations on the DNA are shown in panel A. (C) Identification of Spl binding to the KS promoter. Conditions for these EMSAs were adjusted to optimize Spl binding (see Materials and Methods). EMSAs were conducted with use of (i) a radiolabeled probe consisting of a fragment of the KS promoter extending from -116 to -22, (ii) oligonucleotides containing Spl and AP2 binding sequences, and (iii) protein sources of either epidermal (epid.) nuclear extracts or purified Spl. Purified Spl binding to the probe consistently generated two bands; the mobilities of these complexes were indistinguishable from those generated with epidermal nuclear extracts.


+3

t

TATA box

VOL. 13, 1993


white box). Given the importance of AP2 in other epidermal promoters (29, 30, 65, 66), we examined this site in greater detail (Fig. 5). When radiolabeled probe encompassing this site was combined with epidermal extracts, four complexes were formed (Fig. 5A, WT probe, epid. protein). One of these complexes did not form when keratinocyte extracts were combined with a probe encompassing a mutated AP2 site (Fig. 5A, AP2 mutant probe, epid. protein). This complex was also specifically competed for by the wild-type probe when a cold AP2 competitor oligomer was added to the reaction, and this complex was one of two formed when purified AP2 was substituted for the epidermal extract (Fig. 5A). The presence of an extra complex with purified AP2 was likely due to a degradation product in the AP2 preparation, since mutation of the single AP2 site at -104 to -94 obliterated formation of both AP2 complexes. Finally, methylation interference assays revealed the G nucleotides that were protected by AP2 binding (Fig. 5B). Collectively, these data verified the presence of a bona fide AP2 site in the KS promoter. As shown in Fig. 1, expression levels were influenced by deletion of 160 bp of the KS promoter that encompassed the AP2 binding site. When the AP2 site was mutated in the context of KSgal6000, the mutation also affected the overall level of gene expression in cultured keratinocytes (not shown). In neither case was cell type specificity altered, a result which was consistent with our transgenic mice studies (Fig. 3E). Thus, although AP2 has been implicated in epidermal cell-specific expression and was functional in our K5 promoter, it appeared to be dispensable for the stratified epithelial cell-specific expression displayed by the 90-bp promoter in transgenic animals and for the cell type specificity in the context of the 6,000-bp KS promoter. With the exception of AP2 (30), all of the proteins or protein complexes that we described here were also detected in nuclear extracts from primary human epidermal cells as well as a number of nonkeratinocyte cell types, including human lung fibroblasts (WI38), mouse fibroblasts (NIH 3T3), human epithelial cells (HepG2), and human cervical carcinoma cells (HeLa) (data not shown). Thus, the cell type specificity conferred by the 90-bp K5 promoter in culture could not be explained by a DNA-binding protein that is uniquely expressed in cultured keratinocytes. Role of nuclear proteins that bind to the 90-bp K5 promoter in mediating cell type specificity. To further delineate DNA sequences involved in mediating cell type specificity, we made sequential 5' deletions to remove the protein binding sites within KSBgal90 (Fig. 6A) and tested the effects of these mutations in transfected keratinocytes, hepatocytes, and fibroblasts. Deletions of the Spl binding sites could be made without obliterating cell type specificity (Fig. 6B), suggesting that at least in vitro, the Spl sites do not contribute appreciably to cell type specificity. Moreover, these data further narrowed the sequences involved in cell type specificity to the region between -40 and +32. Within the context of KSBgal90, individual mutations in the binding sites for both protein complex 4 and protein complex 1-2 resulted in a dramatic loss of cell type specificity without compromising greatly the overall levels of expression (Fig. 6C). Gel retardation assays revealed that the mutation in the protein complex 4 binding site prevented the binding of a component of this complex to the vicinity of the TATA box, while the other mutation prevented the binding of protein complex 1-2 to the region of transcriptional initiation (not

shown). Collectively, our results suggested strongly that the kera-


C; compare the second lane [probe alone] with the third lane [probe and nuclear extract]). Protein-DNA complex formation was not influenced by the presence of the nonspecific competitor DNA (poly(dI-dC) (Fig. 4B and C, all lanes) but was greatly reduced by addition of 100-fold molar excess of specific competitor DNA identical to the radiolabeled probe (Fig. 4B, first lane). These data suggested that factors present in SCC-13 nuclear extracts bound specifically to the K5 promoter, within the region from -90 to +32. To localize further the DNA sequences involved in the binding of these proteins, we repeated our EMSAs, this time using annealed (i.e., double-stranded) synthetic oligomers as unlabeled competitors. Our six sets of complementary oligomers spanned the length of the 122-bp SphI-SacI probe, and each set overlapped the next in sequence by 10 nucleotides (Fig. 4A). Two of the oligomer sets, K5-6 and K5-4, competed for the same epidermal extract protein (Fig. 4B; listed over the appropriate gel lanes is each oligomer set used as competitor, at 100-fold excess in the first lane and 200-fold excess in the second lane of each set). This protein appeared to be the known transcription factor Spl. Our first indication was based on the presence of the sequence 5'-TGGGC TGGGC-3' in both the K5-4 and K5-6 oligomers (Fig. 4A, grey boxes). This sequence is similar to the consensus Spl sequence, 5'-(G/T)GGGCGG(G/A)(G/A)(C/T)-3' (25). Indeed, these complexes were competed for by an oligomer set containing an Spl consensus site (Fig. 4B). When the EMSA was repeated, this time with the K5 promoter, purified Spl, and optimal conditions for Spl binding, prominent complexes with mobilities similar to those in the keratinocyte nuclear extract were generated (Fig. 4C). Moreover, these bona fide Spl-K5 promoter complexes were competed for with both the consensus Spl oligonucleotide and the K5-6 and K5-4 oligonucleotides. In addition to Spl, three other protein complexes bound to the 122-bp SphI-SacI K5 probe (Fig. 4B). On the basis of competition with unlabeled oligomers, the positions of these three complexes on the K5 promoter were mapped to the regions outlined in Fig. 4A. Protein complexes 1 and 2 were both competed for equally and effectively by oligomer K5-2. Proteins 3 and 4 are probably multiprotein complexes, since prolonged electrophoresis resolved these into multiple bands and, in addition, mutation of the binding regions resulted in loss of single components of the complex (data not shown). Oligomers K5-3 and K5-4 competed for binding of the protein 3 complex, and K5-3 competed for the binding of protein 4 complex (Fig. 4B). These data indicated that these protein complexes 3 and 4 bound near or at the KS ATAAA (TATA) site (33), while protein complexes 1 and 2 bound near or at the transcription initiation site. Role of transcription factor AP2 in the K5 promoter. Transcription factor AP2 is expressed in high levels in epidermal keratinocytes in vivo and in vitro (29, 40, 66), and in vitro studies have implicated AP2 in expression of a number of epidermal genes (29, 30, 40, 54, 66). In fact, for KS's partner K14 gene, an AP2 element in the proximal promoter seemed to act in concert with a distal element to generate keratinocyte-specific gene expression in transgenic mice (30). Thus, we were initially surprised that the 90-bp K5 promoter did not contain an AP2 binding site, nor did an AP2 oligomer compete for the binding of any of the protein complexes that associated with this segment (data not shown). Interestingly, an AP2 consensus binding site, 5'GC CCCCCAGGC3', is located at nucleotides -104 to -94 5' upstream from the K5 transcription initiation site (Fig. 4A,

3183

MOL. CELL. BIOL.

BYRNE AND FUCHS

3184

A

-22

K5 -116 N

-

protein

m

a)

o

C'J

CL

comp.

ct(0 (0 E E C

H

probe

HQ

-
100-fold (26, 33). However, seemingly ectopic expression in mouse brain has been observed with a variety of transgenic promoters (55). Since brain is a postmitotic tissue, transcript and protein can both accumulate (38), thereby amplifying low levels of aberrant (ectopic) expression. Overall, our findings indicate that 6,000 bp of human K5 promoter/enhancer is sufficient to nearly, if not exclusively, direct faithful expression of genes to cells in which the endogenous KS gene is expressed. Comparison of the behavior of the 90-bp truncated promoter with that of the 6,000-bp promoter indicates that the 6,000-bp KS promoter/ enhancer must contain regulatory sequences that control not only keratinocyte specificity but also (i) basal specificity, (ii) suppression of suprabasal keratinocyte expression in epidermis, tongue, and hair follicle, and (iii) suppression of nonkeratinocyte expression in smooth muscle, dermal fibroblasts, and a few other cell types. K513gal90: switching a basal promoter to a novel differentiation-specific promoter and restricting tissue specificity without appreciably reducing keratinocyte specificity. Despite reduced expression levels of KSBgal90, the severely truncated KS promoter was remarkable in its ability to direct expression to keratinocytes of skin and tongue, to the near exclusion of a myriad of other cell types in the animal. However, the differentiation specificity of the truncated promoter was strikingly and unexpectedly altered. Most notable was transgene expression in suprabasal, differentiating epidermal cells rather than basal cells and in the hair follicle cuticle rather than the ORS. These findings imply that (i) stratified epithelial cell-specific and basal cell-specific gene expression can be uncoupled and (ii) information necessary to direct differentiation-specific gene expression in skin and tongue is contained within the promoter encoding a basal cell-specific keratin. The truncated KS promoter was unusual in that its expression was not analogous to that of any known keratin gene but rather appeared to be a region-dependent subset of the combined activities of a number of suprabasal genes, including those encoding K1/K10, K6/K16, K4/K13, and K9. It


FIG. 7. Evidence that protein complex 1-2 is present in both the KS and K14 promoters but binds to the transcription initiation site of one promoter and the translation initiation region of the other. (A) Evidence that a K14 oligomer encompassing the translation initiation region competes for the binding of protein complex 1-2 to the K5 promoter. A radiolabeled K5 promoter probe spanning the region from -90 to +32 bp was competed for with an oligonucleotide spanning the K5 transcription initiation region (K5-2; see Fig. 4A), two oligonucleotides spanning the K14 transcription and translation start sites (K14-1 and K14-2), an oligonucleotide containing the transcription initiation region of the TdT gene (64), which was used as a source of binding sites for Inr proteins, and an oligonucleotide from the p-fibrinogen gene (P28), which was used as a source of binding sites for the liver-specific transcription factor HNF-1 (2). The oligonucleotides used were K14-1 (CCCl-l-CCAATF-IACCCGAGCACCTTCT), K14-2 (CITCCCTCCTCTGCACCATGACTACCTGCAG) (30), TdT (GCATCAGAGCCCT CATTCTGGAGACACCA) (64), and P28 (CAAACTGTCAAATATTAACTAAAGGGAG) (2). Oligonucleotide K14-2 encompassing the K14 translation initiation region effectively competed for binding of protein complex 1-2 to the KS promoter. (B) Methylation interference analysis of protein 1 binding to the K5 promoter. To precisely define the location of protein complex 1-2, methylation interference assays were conducted (see the legend to Fig. 5). Results for the noncoding (left) and coding (right) strands are shown for protein complex 1. Protein complex 2 gave identical results (not shown). Guanine (G) residues whose methylation interfered strongly with the binding of protein 1 are marked with a closed triangle. Lanes: F, free probe DNA which failed to bind protein 1; B, probe DNA which was bound to protein 1; C, a control reaction showing the normal pattern of G residues. The DNA probes used were a 96-bp TaqI-SacI DNA fragment (left) and a 59-bp AvaI-HphI DNA fragment (right). The AvaI fragment which was used for DNA labeling derives from polylinker DNA adjacent to the SacI site. (C) Methylation interference analysis of protein 1 interaction with the K14 promoter. Results for the noncoding (left) and coding (right) strands are shown for protein complex 1 interaction with the K14 promoter. Protein complex 2 gave identical results (not shown). The DNA probes used were a 108-bp AvaI-BanII DNA fragment (left) and a 153-bp AvaII-EarI DNA fragment (right).

3188

BYRNE AND FUCHS

(50). We do not yet know whether the specificity of the truncated KS promoter is governed by cell-type specific variations in the relative levels or modifications of the factors that bind to the truncated K5 promoter or in the non-DNAbinding proteins that might associate with these DNAbinding proteins. However, our functional studies in vitro suggest the possibility that protein complex 1-2 is involved in conferring keratinocyte specificity. The protein 1-2 binding sequence (GTTCCTGGGTAAC) in KS is similar to the consensus binding sequence (GTTAATNATTAAC) for the POU-homeodomain transcription factors HNF-lao and HNF-113, involved in liver-specific gene expression (reviewed in reference 23). While proteins 1 and 2 share neither the cell type specificity nor the precise binding specificity of HNF-lao and HNF-113, further studies will be necessary to ascertain whether they might nevertheless be related members of the same family. It is also possible that this complex interacts with factors that control initiation of transcription. Members of a class of transcription-initiating factors have been variously termed 8 (18), YYI (63), TFII-I (53), NF-E1 (47), and UCRBP (13) and are either identical or similar binding factors (53, 62). These factors interact with TFIID, and they bind DNA and initiate transcription even in the absence of a TATA box (64). It is intriguing that these factors have been found both on and distal to the initiation site (44), a feature shared with protein complex 1-2. While our preliminary evidence does not support the notion that protein complex 1-2 belongs to this group of initiators, we cannot rule out the possibility that our complex is involved with such factors. In this regard, it may

be relevant that multiple mutations throughout the region encompassing the KS TATA box and transcription initiation site interfered with the specificity of the promoter. As additional studies are conducted, the role of proteins 1 and 2 in keratinocyte specificity should become more apparent. Finally, our studies suggest that the regulatory elements controlling basal cell-specific keratinocyte gene expression are likely to be complex and quite possibly redundant, since we have observed proper cell-type-specific and differentiation-specific expression in transgenic mice even when regulatory elements that are indispensable in shorter promoter segments are mutated in the context of the 6,000-bp promoter. Our findings further suggest that while some factors, such as AP2, both are expressed in high levels in the epidermis (30, 40, 66) and bind to many epidermal promoters (29, 30, 65, 66), they can be eliminated at least in certain contexts without compromising epidermal cell-specific gene expression. In a similar fashion, it may be relevant that human K14 promoter segments missing the binding site(s) for proteins 1 and 2 were nevertheless able to drive keratinocyte-specific gene expression in transgenic mice (30). That we have thus far been able to identify factors sufficient for keratinocyte-specific gene expression only when they are in the context of a minimal promoter suggests that a keratinocyte-specific transcription complex may involve a number of stabilizing factors, and it is only when the number of these factors becomes limiting that their importance is appreciated. As further studies are conducted to examine the multiplicity of binding sites for AP2 and proteins 1 and 2 throughout the keratin promoter/enhancers, the extent to which keratin gene elements controlling keratinocyte specificity are redundant versus compensatory should become more apparent. In closing, our studies reveal that the mechanisms governing cell-type-specific and differentiation-specific gene expression in stratified epithelia are complex, involving both positive and negative regulation. We discovered that the differentiation-specific cues directing expression of a subset of keratin gene pairs to suprabasal stratified squamous epithelia are also located within the basal cell-specific keratin pair, the only pair expressed broadly among stratified squamous epithelia. In addition, while there appear to be additional redundant and/or compensatory tissue-specific elements between -90 and -6000 of the human KS gene, it is remarkable that the small segment from -90 to +32 of the K5 promoter can confer largely keratinocyte-specific expression in a subset of stratified epithelial tissues. It is also intriguing that the promoters for the two major genes expressed exclusively in keratinocytes share a number of the same transcription factors. At least two of these, AP2 and proteins 1 and 2, appear to participate in tissue specificity, even if they may dispensable in certain contexts for tissuespecific gene expression. Thus, although some of the regulatory proteins governing keratinocyte-specific gene expression remain to be identified and characterized, the pathways involved in this complex process are beginning to emerge. ACKNOWLEDGMENTS We acknowledge Robert Lersch for construction of CAT gene clones, preliminary cell culture, and gene transfection studies on the human K5 promoter, which we later confirmed and extended. We thank Linda Degenstein and Debra Dugger for making the transgenic mice used in these studies. We thank Andrew Leask for helpful advice and for citing his unpublished data on the role of Spl in K14 gene regulation, Lia Meisinger for excellent assistance in histological and photographic procedures and for identification of the neurons that express K50gal6000, and Grazina Traska for expert technical assistance in cell culture.


seems relevant that mouse epidermis displays remarkable histological heterogeneity as well as considerable regional diversity in suprabasal keratin gene expression (58, 61). Interestingly, the truncated promoter appeared capable of responding to regional cues and displayed an environmental sensitivity that is absent in the intact promoter from which it is derived. It may be equally significant that many tissues that express K1/K10, K6/K16, or K4/K13 did not seem to express the K5,gal90 transgene. While we cannot unequivocally exclude the possibility that transgene expression was merely below our limits of detection in these tissues, the peculiar behavior of the transgene in skin suggests a more likely possibility, namely, that regulatory elements controlling expression of a keratin in one tissue can differ from those governing expression in a different tissue. If this is true, the regulation of keratin gene expression is significantly more complex than previously realized. Insights into the molecular mechanisms controlling keratinocyte specificity. Our studies with K5,1gal90 enabled us to focus on a relatively small segment of DNA which is capable of orchestrating expression predominantly in a keratinocytespecific fashion. A series of proteins, some of which are common to both the KS and K14 promoters, have been found to bind to this segment. These include Spl, which has a ubiquitous but not necessarily homogeneous distribution in vitro (56), and protein complex 1-2, which binds in the vicinity of the transcription and translation initiation regions of these two promoters. The protein complexes that bind to the 90-bp promoter segment are present in a variety of cultured cell types, and thus, at least in vitro, the restricted expression conferred by the truncated K5 promoter may not be achieved by an epidermal cell-specific, DNA-binding transcription factor. Whether this is also the case in vivo remains to be shown. In this regard, it is notable that in vivo, Spl seems to be expressed in a tissue-specific fashion, while in vitro, it is not

MOL. CELL. BIOL.

VOL. 13, 1993


E.F. is an investigator of the Howard Hughes Medical Institute. C.B. is a research associate funded by the Howard Hughes Medical Institute.

19:1033-1042. 17. Gorman, C. M., L. F. Moffat, and B. H. Howard. 1982. Recombinant genomes which express chloramphenicol acetyltransferase in mammalian cells. Mol. Cell. Biol. 2:1044-1051. 18. Hariharan, N., D. E. Kelley, and R. P. Perry. 1991. Sigma, a transcription factor that binds to downstream elements in several polymerase II promoters, is a functionally versatile zinc

finger protein. Proc. Natl. Acad. Sci. USA 88:9799-9803. 19. Heid, H. W., I. Moll, and W. W. Franke. 1988. Patterns of expression of trichocytic and epithelial cytokeratins in mammalian tissues. Differentiation 37:215-230. 20. Hirochika, H., R. Hirochika, T. R. Broker, and L. T. Chow. 1988. Functional mapping of the human papillomavirus type II transcriptional enhancer and its interactions with the transacting E2 proteins. Genes Dev. 2:54-67. 21. Hume, W. J., and C. S. Potten. 1983. Proliferative units in stratified squamous epithelium. Clin. Exp. Dermatol. 8:95-106. 22. Jiang, C. K., H. S. Epstein, M. Tomic, I. M. Freedberg, and M. Blumenberg. 1991. Functional comparison of the upstream regulatory DNA sequences of four human epidermal keratin genes. J. Invest. Dermatol. 96:162-167. 23. Johnson, P. F. 1990. Transcriptional activators in hepatocytes. Cell Growth Differ. 1:47-52. 24. Jonas, E. A., A. M. Snape, and T. D. Sargent. 1989. Transcriptional regulation of a Xenopus embryonic epidermal keratin gene. Development 106:399-405. 25. Kadonaga, J. T., K. A. Jones, and R. Tjian. 1986. Promoterspecific activation of RNA polymerase II transcription by Spl. Trends Biol. Sci. 11:20-23. 26. Kulesh, D. A., and R. G. Oshima. 1988. Cloning of the human keratin 18 gene and its expression in nonepithelial mouse cells. Mol. Cell. Biol. 8:1540-1550. 27. Kunkel, T. A. 1985. Rapid and efficient site-specific mutagenesis without phenotypic selection. Proc. Natl. Acad. Sci. USA 82:488-492. 28. Kopan, R., and E. Fuchs. 1989. A new look into an old problem: keratins as tools to investigate determination, morphogenesis, and differentiation in skin. Genes Dev. 3:1-15. 29. Leask, A., C. Byrne, and E. Fuchs. 1991. Transcription factor AP2 and its role in epidermal-specific gene expression. Proc. Natl. Acad. Sci. USA 88:7948-7952. 30. Leask, A., M. Rosenberg, R. Vassar, and E. Fuchs. 1990. Regulation of a human epidermal keratin gene: sequences and nuclear factors involved in keratinocyte-specific transcription. Genes Dev. 4:1985-1998. 31. Lvask, A., A. Weng, and E. Fuchs. Unpublished data. 32. Lersch, R., and E. Fuchs. 1988. Sequence and expression of a type II keratin, K5, in human epidermal cells. Mol. Cell. Biol. 8:486-493. 33. Lersch, R., V. Stellmach, C. Stocks, G. Giudice, and E. Fuchs. 1989. Isolation, sequence, and expression of a human keratin K5 gene: transcriptional regulation of keratins and insights into pairwise control. Mol. Cell. Biol. 9:3685-3697. 34. Luckow, B., and G. Schutz. 1987. CAT constructions with multiple unique restrictions sites for the functional analysis of eukaryotic promoters and regulatory elements. Nucleic Acids Res. 15:5490. 35. Mack, D. H., and L. A. Laimins. 1991. A keratinocyte-specific transcription factor, DRF-1, interacts with AP-1 to activate expression of human papillomavirus type 18 in squamous epithelial cells. Proc. Natl. Acad. Sci. USA 88:9102-9106. 36. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1989. Molecular cloning: a laboratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y. 37. Marchuk, D., E. Fuchs, and S. McCrohon. 1985. Complete sequence of a type I human keratin gene: presence of enhancerlike elements in the regulatory region of the gene. Proc. Natl. Acad. Sci. USA 82:1609-1613. 38. Matsuoka, M., F. Nagawa, K. Okazaki, L. Kingsbury, K. Yoshida, U. Muller, D. T. Larue, J. A. Winer, and H. Sakano. 1991. Detection of somatic DNA recombination in the transgenic mouse brain. Science 254:81-86. 39. McCormick, A., H. Brady, J. Fukushima, and M. Karin. 1991. The pituitary-specific regulatory gene GHF1 contains a minimal cell type-specific promoter centered around its TATA box. Genes Dev. 5:1490-1503. 40. Mitchell, P. J., P. M. Timmons, J. M. Hebert, P. W. J. Rigby, and R. Tjian. 1991. Transcription factor AP-2 is expression in neural crest cell lineages during mouse embryogenesis. Genes Dev. 5:105-119.


REFERENCES 1. Albers, K., and E. Fuchs. 1992. The molecular biology of intermediate filament proteins. Int. Rev. Cytol. 134:243-279. 2. Baumhueter, S., G. Courtois, and G. R. Crabtree. 1988. A variant nuclear protein in dedifferentiated hepatoma cells binds to the same functional sequences in the 0-fibrinogen gene promoter as HNF-1. EMBO J. 8:2485-2493. 3. Blessing, M., J. L. Jorcano, and W. W. Franke. 1989. Enhancer elements directing cell-type-specific expression of cytokeratin genes and changes of the epithelial cytoskeleton by transfections of hybrid cytokeratin gene. EMBO J. 8:117-126. 4, Blessing, M., H. Zentgraf, and J. L. Jorcano. 1987. Differentially expressed bovine cytokeratin genes. Analysis of gene linkage and evolutionary conservations of 5'-upstream sequences. EMBO J. 6:567-575. 5. Bradford, M. M. 1976. A rapid and sensitive method for the quantitation of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72:249-254. 5a.Byrne, C., and E. Fuchs. Unpublished data. 6. Cereghini, S., M. Blumenfeld, and M. Yaniv. 1988. A liverspecific factor essential for albumin transcription differs between differentiated and dedifferentiated rat hepatoma cells. Genes Dev. 2:957-974. 7. Cheng, J., K. Turksen, Q. C. Yu, H. Schreiber, M. Teng, and E. Fuchs. 1992. Cachexia and graft-vs.-host-disease-type skin changes in keratin promoter-driven TNFa transgenic mice. Genes Dev. 6:1444-1456. 8. Chin, M. T., T. B. Broker, and L. T. Chow. 1989. Identification of a novel constitutive enhancer element and an associated binding protein: implications for human papillomavirus type II enhancer regulation. J. Virol. 63:2967-2976. 9. Cripe, T. P., T. H. Haugen, J. P. Turk, F. Tabatabai, P. G. Schmid II, M. Durst, L. Gissmann, A. Roman, and L. P. Terek. 1987. Transcriptional regulation of the human papillomavirus-16 E6-E7 promoter by a keratinocyte-dependent enhancer, and by viral E2 trans-activator and repressor gene products: implications for cervical carcinogenesis. EMBO J. 6:3745-3753. 10. Dale, B. A., K. A. Holbrook, J. R. Kimball, M. Hoff, and T.-T. Sun, 1985. Expression of epidermal keratins and filaggrin during human fetal skin development. J. Cell Biol. 101:1257-1269. 11. Dhouailly, D., C. Xu, M. Manabe, A. Schermer, and T.-T. Sun. 1989. Expression of the hair-related keratin in a soft epithelium: subpopulations of human and mouse dorsal tongue keratinocytes express keratin markers for hair- skin- and esophagealtypes of differentiation. Exp. Cell Res. 181:141-158. 12. Eustice, D. C., P. A. Feldman, A. M. Colberg-Poley, R. M. Buckery, and R. H. Neubauer. 1991. A sensitive method for the detection of beta-galactosidase in transfected mammalian cells. BioTechniques 11:739-742. 13. Flanagan, J. R., M. Murata, P. A. Burke, Y. Shirayoshi, E. Appelia, P. A. Sharp, and K. Ozato. 1991. Negative regulation of the major histocompatibility complex class I promoter in embryonal carcinoma cells. Proc. Natl. Acad. Sci. USA 88:3145-3149. 14. Fong, T. C., and R. M. Emerson. 1992. The erythroid-specific protein CGATA-1 mediates distal enhancer activity through a specialized beta-globin TATA box. Genes Dev. 6:521-532. 15. Franke, W. W., E. Schmid, S. Winter, M. Osborn, and K. Weber. 1979. Widespread occurrence of intermediate-sized filaments of the vimentin-type in cultured cells from diverse vertebrates. Exp. Cell Res. 123:25-46. 16. Fuchs, E., and H. Green. 1980. Changes in keratin gene expression during terminal differentiation of the keratinocyte. Cell

3189

3190

BYRNE AND FUCHS

41. Moll, I., H. Heid, W. W. Franke, and R. Moll. 1987. Distribution

of a special subset of keratinocytes characterized by the expression of cytokeratin 9 in adult and fetal human epidermis of various body sites. Differentiation 33:254-265. 42. Moll, R., D. Dhouailly, and T.-T. Sun. 1989. Expression of keratin 5 as a distinctive feature of epithelial and biphasic mesotheliomas. Virchows Archiv. B 58:129-145. 43. Moll, R., W. W. Franke, D. L. Schiller, B. Geiger, and R.

Krepler. 1982. The catalog of human cytokeratins: patterns of expression in normal epithelia, tumors, and cultured cells. Cell 31:11-24. 44. Nakatani, Y., M. Horikoshi, M. Brenner, T. Yamamoto, F. Besnard, R. G. Roeder, and E. Freese. 1990. A downstream

45.

47.

48.

49.

50.

51.

52.

S. H. Yuspa, and D. R. Roop. 1991. A human epidermal

53.

54.

55. 56.

differentiation-specific keratin gene is regulated by calcium but not negative modulators of differentiation in transgenic mouse keratinocytes. Cell Growth Differ. 2:107-113. Roy, A. L., M. Meisterernst, P. Pognonec, and R. G. Roeder. 1991. Cooperative interaction of an initiator-binding transcription factor and the helix-hoop-helix activator USF. Nature (London) 354:245-248. Royer, H.-D., M. P. Freyaldenhoven, I. Napierski, D. D. Spitkovsky, T. Bauknecht, and N. Dathan. 1991. Delineation of human papillomavirus type 18 enhancer binding proteins: the intracellular distribution of a novel octamer binding protein p92 is cell cycle regulated. Nucleic Acids Res. 19:2363-2371. Russo, A. F., E. B. Crenshaw III, S. A. Lira, P. M. Simmons, L. W. Swanson, and M. G. Rosenfeld. 1988. Neuronal expression of chimeric genes in transgenic mice. Neuron 3:311-320. Saffer, J. D., S. P. Jackson, and M. B. Annarella. 1991. Developmental expression of Spl in the mouse. Mol. Cell. Biol. 11:2189-2199.

57. Schermer, A., S. Galvin, and T.-T. Sun. 1986. Differentiationrelated expression of a major 64K corneal keratin in vivo and in culture suggests limbal location of corneal epithelial stem cells.

J. Cell Biol. 103:49-62. 58. Schweizer, J., I. Baust, and H. Winter. 1989. Identification of murine type I keratin 9 (73kDa) and its immunolocalization in neonatal and adult mouse foot sole epidermis. Exp. Cell Res. 184:193-206. 59. Schweizer, J., and F. A. Marks. 1977. A developmental study of the distribution and frequency of Langerhans cells in relation to formation of patterning in mouse tail epidermis. J. Invest.

Dermatol. 69:198-204. 60. Schweizer, J., M. Rentrop, R. Nischt, M. Kinjo, and R. Winter. 1988. The intermediate filament system of the keratinizing mouse forestomach epithelium: coexpression of keratins of internal squamous epithelia and of epidermal keratins in differentiating cells. Cell Tissue Res. 253:221-229. 61. Schweizer, J., and H. Winter. 1982. Changes in regional keratin polypeptide patterns during phorbol ester-mediated reversible and permanently sustained hyperplasia of mouse epidermis. Cancer Res. 42:1517-1529. 62. Seto, E., Y. Shi, and T. Shenk. 1991. YY1 is an initiator sequence-binding protein that directs and activates transcription in vitro. Nature (London) 352:241-245. 63. Shi, Y., E. Seto, L.-S. Chang, and T. Shenk. 1991. Transcriptional repression by YY1, a human GLI-Kruppel-related protein, and relief of repression by adenovirus ElA protein. Cell 67:377-388. 64. Smale, S., and D. Baltimore. 1989. The "initiator" as a transcription control element. Cell 57:103-113. 65. Snape, A. M., E. A. Jonas, and T. C. Sargent. 1990. KTF-1, a transcriptional activator of Xenopus embryonic keratin expression. Development 109:157-165. 66. Snape, A. M., R. S. Winning, and T. D. Sargent. 1991. Transcription factor AP-2 is tissue-specific in Xenopus and is closely related or identical to keratin transcription factor 1 (KTF-1). Development 113:283-293. 67. Stark, H.-J., D. Breitkreutz, A. Limat, P. Bowden, and N. E. Fusenig. 1987. Keratins of the human hair follicle: "hyperproliferative" keratins consistently expressed in outer root sheath cells in vivo and in vitro. Differentiation 35:236-238. 68. Stark, H.-J., D. Breitkreutz, A. Limat, C. M. Ryle, D. Roop, I. Leigh, and N. Fusenig. 1990. Keratins 1 and 10 or homologues as regular constituents of inner root sheath and cuticle cells in the human hair follicle. Eur. J. Cell Biol. 52:359-372. 69. Stellmach, V., A. Leask, and E. Fuchs. 1991. Retinoid-mediated transcriptional regulation of keratin genes in human epidermal and squamous cell carcinoma cells. Proc. Natl. Acad. Sci. USA 88:4582-4586. 70. Steven, A. C. 1990. Intermediate filament structure, p. 233-263. In R. D. Goldman and P. M. Steinert (ed.), Cellular and molecular biology of intermediate filaments. Plenum Press, New York. 71. Sun, T.-T., R. Eichner, A. Schermer, D. Cooper, W. G. Nelson, and R. A. Weiss. 1984. The transformed phenotype, p. 169-176. In A. Levine, W. Topp, G. van de Woude, and J. D. Watson (ed.), The cancer cell, vol. 1. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. 72. Tomic, M., C.-K. Jiang, H. S. Epstein, I. M. Freeeberg, H. H. Samuels, and M. Blumenberg. 1990. Nuclear receptors for retinoic acid and thyroid hormone regulate transcription of keratin genes. Cell Regul. 1:965-973. 73. Turksen, K., T. Kupper, L. Degenstein, I. Williams, and E. Fuchs. 1992. Interleukin 6: insights to its function in skin by overexpression in transgenic mice. Proc. Natl. Acad. Sci. USA 89:5068-5072. 73a.van Mujen, G. N. P., D. J. Ruiter, W. W. Franke, T. Achtstatter, W. H. B. Haasnoot, M. Ponec, and S. 0. Warnaar. 1986. Cell type heterogeneity of cytokeratin expression in complex epithelia and carcinomas as demonstrated by monoclonal antibodies specific for cytokeratins nos 4 and 13. Exp. Cell Res. 162:97-113. 74. Vassar, R., and E. Fuchs. 1991. Transgenic mice provide new insights into the role of TGF-alpha during epidermal development and differentiation. Genes Dev. 5:714-727. 75. Vassar, R., M. Rosenberg, S. Ross, A. Tyner, and E. Fuchs. 1989. Tissue-specific and differentiation-specific expression of a human K14 keratin gene in transgenic mice. Proc. Natl. Acad. Sci. USA 86:1563-1567. 76. Wu, Y.-J., and J. G. Rheinwald. 1981. A new small, (40 kd) keratin filament protein made by some cultured human cell carcinoma. Cell 25:627-635. 77. Yu, C. C.-K., L.-C. Tsui, and M. L. Breitman. 1990. Homologous and heterologous enhancers modulate spatial expression but not cell-type specificity of the murine lambdaF-crystallin promoter. Development 110:131-136.


46.

initiation element required for efficient TATA box binding and in vitro function of TFIID. Nature (London) 348:86-88. Nelson, W., and T.-T. Sun. 1983. The 50- and 58-kdalton keratin classes as molecular markers for stratified squamous epithelia: cell culture studies. J. Cell Biol. 97:244-251. Ohtsuki, M., M. Tomic-Canic, I. M. Reedberg, and M. Blumen berg. 1992. Nuclear proteins involved in transcription of the human K5 keratin gene. J. Invest. Dermatol. 99:206-215. Park, K., and M. L. Atchison. 1991. Isolation of a candidate repressor/activator, NF-E1 (YY-1, sigma) that binds to the immunoglobulin k 3' enhancer and the immunoglobulin heavychain uEi site. Proc. Natl. Acad. Sci. USA 88:9804-9808. Rentrop, M., B. Knapp, H. Winter, and J. Schweizer. 1986. Differential localization of distinct keratin mRNA-species in mouse tongue epithelium by in situ hybridization with specific cDNA probes. J. Cell Biol. 103:2583-2591. Rentrop, M., R. Nischt, B. Knapp, J. Schweizer, and H. Winter. 1987. An unusual type-II 70-kilodalton keratin protein of mouse epidermis exhibiting postnatal body-site specificity and sensitivity to hyperproliferation. Differentiation 34:184-200. Robidoux, S., P. Gosselin, M. Harvey, S. Leclerc, and S. L. Guerin. 1992. Transcription of the mouse secretory protease inhibitor p12 gene is activated by the developmentally regulated positive transcription factor Spl. Mol. Cell. Biol. 12:3796-3806. Roop, D. R., H. Huitfeldt, A. Kilkenny, and S. H. Yuspa. 1987. Regulated expression of differentiation-associated keratins in cultured epidermal cells detected by monospecific antibodies to unique peptides of mouse epidermal keratins. Differentiation 35:143-150. Rosenthal, D. S., P. M. Steinert, S. Chung, A. Huff, J. Johnson,

MOL. CELL. BIOL.