INTRODUCTION

Adult epidermis is a stratified epithelium in which cells in theinnermost, basal layer cease to divide concomitantly with aprocess of upward movement and terminal differentiation.Dead squames sloughed from the skin surface are proteina-ceous, cross-linked envelopes, or sacs, filled with highlybundled, macrofibrils of hundreds of keratin filaments. Thebiochemical program of terminal differentiation involves thesequential expression of different proteins, including keratinsin the basal and spinous layers, and loricrin, a component ofthe cornified envelope, in the granular layer.

Type I (K9-K20) and type II (K1-K8) keratins form obligateheterodimers that constitute the subunits of epithelial interme-diate filaments (IFs) (reviewed by Fuchs and Weber, 1994).Simple epithelia express K8 and K18 (Wu and Rheinwald,1982), while basal cells of adult epidermis express K5 and K14(Fuchs and Green, 1980; Nelson and Sun, 1983). As epidermalcells become committed to terminal differentiation, they switchfrom expression of K5 and K14 to K1 and K10 (Fuchs andGreen, 1980; Roop et al., 1987). Hair follicles display more

elaborate programs of differentiation and keratin expression,including K5 and K14 in the outer root sheath (ORS), K1 andK10 in the inner root sheath (IRS) and hair-specific keratins inthe cortex and medulla (Heid et al., 1986; Lynch et al., 1986;Stark et al., 1990).

Much of what is known about embryonic epidermal geneexpression in mammals comes from characterization of skinproteins, and from immunohistochemistry of skin sections,often stained with antibodies of multiple keratin specificities.Some studies have suggested that early in development, whenthe epidermis is only a single layer, cells express simple epithe-lial keratins (Jackson et al., 1981; Franke et al., 1981). Whilesubsequent studies with monoclonal antibodies have notdetected K8 and K18 at very early stages (E7.5), there isagreement that K8 and K18 are detected when the epidermisbecomes bi-layered (Jackson et al., 1981; Franke et al., 1981;Moll et al., 1982; Thorey et al., 1993). At this stage, K8 andK18 localize to the flattened upper layer of periderm cells(Moll et al., 1982; Dale et al., 1985). Intriguingly, peridermalso has desmosomes and can express involucrin, suggestingcertain similarities to epidermal as well as simple epithelial

2369Development 120, 2369-2383 (1994)Printed in Great Britain © The Company of Biologists Limited 1994

As the major proteins of adult keratinocytes, keratinsprovide biochemical markers for exploring mouseepidermal embryogenesis. Here, we used a modifiedmethod of whole-mount in situ hybridization to track skin-specific expression of endogenous keratin mRNAs through-out embryogenesis. To monitor transcriptional regulation,we coupled this with

β−galactosidase expression of ahuman epidermal keratin promoter-driven transgene.These studies have radically changed our perception ofhow the program of gene expression becomes establishedduring epidermal development. Specifically, we have dis-covered that (1) basal keratin (K5 and K14) genes are firstdetected at E9.5 in a highly regional fashion, and surpris-ingly as early as the single layered ectodermal stage; (2) theearly patterns do not correlate with morphogenesis per se,but rather with regional variations in the embryonic originof underlying mesenchyme, supporting morphogeneticcriteria that early inductive cues are mesenchymal; (3)epidermal keratin genes are expressed in periderm, sup-porting the notion that this layer arises from ectodermal

stratification, even though it is simple epithelial-like inmorphology and is subsequently sloughed during develop-ment; (4) later embryonic patterns of K5 and K14 geneexpression parallel proliferative capacity and not stratifi-cation; and (5) K1 and K10 mRNAs are first detected asearly as E13.5, and their patterns correlate with differen-tiation and not stratification.

These patterns of epidermal gene expression led us toexplore whether potential transcriptional regulators ofthese genes are expressed similarly. We show that AP2 (butnot Sp1) cRNAs hybridize in a pattern similar to, butpreceding that of basal keratin cRNAs. Finally, using geneexpression in cultured cells, we demonstrate that AP2 hasa strong inductive effect on basal keratin expression in acellular environment that does not normally possess AP2activity.

Key words: embryonic ectoderm, keratins, loricrin, AP2, geneexpression, skin

SUMMARY

Programming gene expression in developing epidermis

Carolyn Byrne1, Michael Tainsky2 and Elaine Fuchs1,*1Howard Hughes Medical Institute, Department of Molecular Genetics and Cell Biology, The University of Chicago, 5841 S.Maryland Avenue, Chicago, IL 60637, USA2The University of Texas MD Anderson Cancer Center, Department of Tumor Biology, Houston, Texas 77030, USA

*Author for correspondence

2370

cells (Watt et al., 1989). In later stages of development,periderm is apparently sloughed (Sengel, 1976), making itsorigin even more difficult to assess.

Expression of K5 and K14 has been detected as early as thebi-layered epithelial stage (Moll et al., 1982; Schweizer andWinter, 1982; Banks-Schlegel, 1982; Dale et al., 1985; Fisherand Holbrook, 1987). Immunohistochemistry using a mono-specific anti-K14 antiserum and in situ hybridization with aK14 cRNA probe has localized expression to the inner,embryonic basal layer of back skin sections (Kopan and Fuchs,1989). After stratification, K14 seemed to be restricted to theinnermost layer, in agreement with earlier studies using lessspecific probes. When combined with protein synthesis dataand morphological correlations, these findings have led to theassumption that K5 and K14 genes are downregulated uponstratification. Researchers have uniformly stressed the intimatecorrelation between these keratins and a single layer ofcolumnar, basal cells.

It is well-accepted that K1 and K10 induction occurs postK5 and K14 (Moll et al., 1982; Banks-Schlegel, 1982;Schweizer and Winter, 1982; Dale et al., 1985; Kopan andFuchs, 1989), and times as early as E15.5 have been reported(Greer and Roop, 1991). K1 and K10 have often been detectedconcomitantly with the appearance of a third, middle epidermallayer, suggesting a link between these keratins and stratifica-tion (Moll et al., 1982; Schweizer and Winter, 1982; Dale etal., 1985; Kopan and Fuchs, 1989). Collectively, keratinstudies have led investigators to conclude that there are strongparallels between embryonic and adult epidermis, with K5 andK14 gene expression restricted to the inner layer, and K1 andK10 confined to stratifying layers.

Despite the relative consistency of these findings, we foundtheir logical predictions difficult to reconcile in light of earliermorphological studies and tissue reconstitution experiments.For many years, developmental biologists have underscoredthe importance of underlying mesenchyme in epidermalembryogenesis, emphasizing that mesenchyme is highlyregional and stems from different origins (for reviews, seeSengel, 1976, 1990; Hardy, 1992). Mesenchymal influence hasbeen implicated in proliferation, maintenance and differen-tiation of embryonic epidermis and its appendages (Wessels,1963; Noden, 1988; Richman and Tickle, 1989; reviewed bySengel, 1990). Differences in underlying mesenchyme alsoseem to account for certain regional morphological hetero-geneities in ectoderm (Hanson, 1947; Sengel, 1976). This said,epidermal morphologies are sometimes difficult to distinguisheven when cells overlie different mesenchymes and willdevelop differently. Such findings have led us to wonderwhether K5 and K14 expression in early development strictlycorrelates with a single layer of columnar basal cells, orwhether environmental cues might also be involved.

Another distinguishing aspect of embryonic epidermis isthat upon stratification, suprabasal cells often retain basal-likemorphology and mitotic activity. This feature may satisfy theneed for rapid surface expansion, something not required ofadult epidermis. It then seems antithetical that K5 and K14expression could both be governed by morphology and simul-taneously restricted to a single layer of developing cells.

In this report, we address many of the apparent discrepan-cies that have arisen between studies on epidermal morphol-ogy and those on the biochemistry of skin development. To

reevaluate these issues, we adapted whole-mount in situhybridization methods to study mouse embryogenesis from E9to E16. While this method is not applicable to internal organs,the unique accessibility and architecture of skin provided anopportunity to investigate induction of late stage geneexpression in epidermis. We analyzed a series of mRNAsencoding basal keratins, suprabasal keratins and the cornifiedcell envelope precursor, loricrin. We paralleled these analyseswith sensitive assays for a human K5 promoter-driven β-galac-tosidase (β-gal) transgene. Our studies uncovered a highlyspecific patterning of keratin expression, which has hithertogone unrecognized.

We have also investigated transcription factors that might beinvolved in establishing these patterns of gene expression. Inparticular, we focused on AP2, a transcription factor ofepidermal and neural lineages (Mitchell et al., 1991). Althoughthe precise nature of AP2’s role in epidermal gene expressionis not clear, functional AP2 binding sites exist in the 5′upstream region of many epidermally expressed genes (Leasket al., 1990, 1991; Snape et al., 1990, 1991; Behrens et al.,1991; Byrne and Fuchs, 1993; Rothnagel et al., 1993;Magnaldo et al., 1993; Powell et al., 1992). Here, we reportthat an AP2 cRNA hybridized in patterns similar to those ofbasal cell keratin cRNAs. Moreover, we show that expressionof AP2 can impart to hepatocytes the ability to expressepidermal keratin genes in culture.

MATERIALS AND METHODS

Embryo preparation and β-galactosidase histochemistryMice from 3 different heterozygotic lines (Byrne and Fuchs,1993) weremated, and embryo ages were estimated from the time of the appear-ance of the vaginal plug (=E0.5). Embryos were fixed in 4%paraformaldehyde in PBS for 10-30 minutes depending on age, andwashed extensively in PBS. To assay for β-gal activity, whole embryoswere incubated in 5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside(X-Gal) to convert the colorless reagent to a blue end-product. Incuba-tion times for embryos were: 12 hours (E9.5), approx. 2 hours (E10.5-14.5) or ≤1 hour (E16.5). After post-fixing in 4% paraformaldehyde theembryos were photographed, embedded in paraffin and sectioned (5µm). Some embryos were embedded in epoxy resin, sectioned (3 µm)and counterstained with toluidine blue. As controls, stainings weresometimes done post-sectioning rather than pre-sectioning.

Preparation of hybridization probesDigoxigenin cRNA probes were synthesized according to the manu-facturer (Boehringer Mannheim Biochemicals, Indianapolis, IN).Biotin cRNA probes were prepared as described by Kopan andWeintraub (1992). A 1900 nucleotide AP2 cRNA probe was made bylinearizing plasmid RP1 (Moser et al., 1993; supplied by R. Buettner)with

EcoRI and transcribing with SP6 polymerase. An approx. 200nucleotide AP2B cRNA was made by asymmetric PCR from a PstI-SpeI AP2B-specific restriction fragment from pNAP2B (Buettner etal., 1993) using the primer AAGAAGGAAGAAGGATGGAG-GTAT. An approx. 1000 nt K5 cRNA was made from EcoRI-lin-earized pK5 and T7 RNA pol (Lersch and Fuchs, 1988). An approx.1300 nt K14 cRNA was made from HindIII-linearized p3SP and SP6RNA pol (Tyner and Fuchs, 1986). An approx. 350 nt K1 cRNA wasmade from BsteII-linearized pMK1-3′NC (provided by S. Yuspa). Anapprox. 220 nt K10 cRNA was made from EcoRI-linearized pMK10-3′NC (provided by S. Yuspa) and SP6 RNA pol. An approx. 800 ntloricrin cRNA was made from BamHI-linearized pGZskinEE2.9 (H.Hsu and E. Fuchs, unpublished data) and T7 pol.

C. Byrne, M. Tainsky and E. Fuchs

2371Gene expression in developing epidermis

Whole-mount situ hybridizationIn situ hybridizations with digoxigenin probes were as described byConlan and Rossant (1992) except that proteinase K treatments wereoptimized to preserve skin morphology without sacrificing penetra-tion of skin by the probes. Treatments ranged from 3.5 minutes in 5mg/ml proteinase K (E9.5 embryos) to 20 minutes in 20 mg/ml pro-teinase K (E16.5 embryos). Detection of biotinylated probes was asdescribed (Herrmann, 1991).

RNA analysisEpidermis was separated from dermis by incubation in 10 mM EDTAin PBS for 10-30 minutes depending on age. RNA was prepared(Chomczynski and Sacchi, 1987), and RT-PCR was performed(Leonard et al., 1993). Primers used were: mouse ribosomal S16primers (Leonard et al., 1993); AP2A-specific primers, CGG-GTCGCCTGTCGCTCCTCA, TTTCTTGCCACTTGCTCATTG;AP2B-specific primers, CGGGTCGCCTGTCGCTCCTCA, TAT-TATCCATTTTCCAAGA; primers for the region shared by AP2Aand AP2B, CCCAGCCTCAGCCGCAGCAC, TTACCACGCCAC-CGAAGAGG.

Transfections, protein estimations and enzyme assaysHepG2 (ATCC, Rockville, MD) and SCC-13 (gift of Dr J. Rheinwald)were cultured (Leask et al., 1990), and transfections were performedusing the following per 100 mm dish: 2 pmol of either pK5cat6000(Lersch et al., 1989), pK14cat2300 (Leask et al., 1990), or5xTREcolTK-CAT (Angel et al., 1988); 1, 2 or 4 pmol pNAP2A orpNAP2B or the equivalent of empty expression vector (Buettner etal., 1993); and Bluescript-pKS+ (Stratagene, La Jolla, Calif.) toequalize DNA levels. 5 µg of pTKβ-gal was used internally to adjustfor variations in transfection frequencies. Transfections and protein,cat and β-gal assays were performed as described (Leask et al., 1990).

RESULTS

Basal epidermal transcription is induced early inpost-gastrulation mouse embryos: evidence formesenchymal cuesPreviously, we used a promoter from a basal cell epidermalkeratin to examine expression of a reporter gene in adult trans-genic mice (Byrne and Fuchs, 1993). In three independentlyderived lines, 6000 bp of human K5 promoter faithfullydirected expression of β-gal to basal cells of stratifiedsquamous epithelia, including epidermis and the ORS of hairfollicles (Fig. 1). We took advantage of the extreme sensitiv-ity of this system to assay activity of the K5 promoter duringmouse embryogenesis. All results outlined below were highlyreproducible, not only for the three lines, but in addition, fordifferent litters of the same age within a single line.

K5 promoter activity was not detectable before E9.5. At thistime, faint β-gal activity was detected in the dorsolateral regionof what appeared to be ectoderm over posterior somites (Fig.2A). Although difficult to demonstrate by photography, faintexpression over the head was also detected. By E10.5,expression was intense in the first branchial arch (Fig. 2B). Oneday later, β-gal activity included the maxillary and mandibu-lar region of the arch, and also broadened rostrally and caudallyin the putative dorsolateral ectoderm over the somites (Fig.2C,D). Expression was also induced in limb buds, but was notdetected dorsally or ventrally (Fig. 2D,E). The timing of β-galexpression was consistent with PCR studies, which detectedbasal keratin mRNAs at low levels as early as E9.5 (notshown).

To verify that the patterns of K5 promoter activity reflectedthose of the endogenous K5 gene, we conducted whole-mountin situ hybridization using a digoxigenin-labeled riboprobespecific for K5 (Lersch and Fuchs, 1988). This technique wasnot sufficiently sensitive to localize the E9.5 K5 mRNAsdetected by PCR. However, beginning at E10.5 and thereafter,K5 hybridization was detected, and it increased with age. Incontrast, a sense strand K5 control cRNA remained negativethroughout embryogenesis (not shown). Throughout E10.5 andsubsequent stages of embryogenesis, human K5β-gal andmouse K5 mRNA patterns were indistinguishable. An exampleof this is shown in Fig. 2H, where K5 cRNA hybridization inE11.5 ectoderm over somites paralleled β-gal staining(compare with Fig. 2F,G). For most data from E10.5-E12.5,we present β-gal staining, since this method was significantlymore sensitive.

Sections of stained E9.5 mouse embryos confirmed that β-gal expression was in the ectoderm overlying somites. Mor-phology was best visualized in semithin sections (Fig. 2I),while K5 promoter activity was best detected in frozen sections(Fig. 2J). The striped pattern of β-gal in this region did notreflect obvious differences in morphology within the ectodermnor did it appear to be an artifact of the folding of theembryonic surface (Fig. 2K, embryo surface; Fig. 2L, after sec-tioning; arrowheads provide reference points for comparisons).Rather, expression appeared to be in specific zones, at a timewhen mitoses were still occurring parallel to the basal lamina(Fig. 2M, bracket; see also Smart, 1970), and when the somiticectoderm was still single-layered (Fig. 2I).

In contrast to somitic ectoderm, E9.5 head ectoderm was bi-layered (Fig. 2N; arrowheads denote basement membrane).Surprisingly, cells from both layers expressed K5β-gal,although initial expression was extremely low and mosaic(bracket). These data revealed for the first time that peridermexpresses basal epidermal keratins. While we cannot unequiv-ocally rule out the possibility that the mosaicism was artifac-tual, we think this is unlikely, since (1) expression in bothperiderm and ectoderm soon became uniform and persistedthroughout development (see below), and (2) mosaicism wouldbe predicted at a time when critical levels of positive actingfactors are beginning to accumulate.

A review of expression patterns revealed that K5 wasdetected at times when mesodermal cells were beginning topopulate skin. Expression was seen first in posterior somiticectoderm, where it then expanded anteriorly and posteriorly.Expression did not seem to be in synchrony with the underly-ing wave of migration of newly differentiated somitic der-mamyotome. Rather, it seemed that it was the nature of thematuring dermamyotome, not migration per se, that providedK5 inductive cues. This was consistent with recombinationexperiments showing that dermatome cells from differentsomites possess distinct properties (Sengel, 1990).

Somites had migrated and differentiated into dermamy-otome by the time K5 was detected (Fig. 2I, the 4-5 mes-enchymal layers under the ectoderm). Intriguingly, the stripedpattern of β-gal expression in the E9.5 ectoderm overlying thisregion resembled that of underlying homeobox Mox1 geneexpression in the E8.5 dermamyotome (Candia et al., 1992).Such mirror image patterns of gene expression in dermamy-otome and ectoderm further strengthened the notion that the

2372

striped patterns of K5 promoter activity were genuine andreflective of dermamyotome signals.

K5 induction in the first branchial arch ectoderm alsofollowed soon after derivation and migration of underlyingmesenchyme, in this case stemming from neural crest (E10.5;see Fig. 2B; see Nichols, 1981; Couly et al., 1992). In contrast,K5 had not yet been induced in ventral ectoderm, whose under-lying mesenchyme derived from somatopleural mesoderm(Sengel, 1976; 1990). Collectively, these data predicted that (1)the K5 promoter in ectoderm receives inductive signal(s) fromdermal cells of specific embryonic origin, (2) the inductivecue(s) is acquired in a temporal fashion, and (3) induction doesnot correlate with morphology, but rather can occur at eitherthe single or bi-layered epithelial stage, and in both ectodermand periderm, depending on location.

Patterning of K5 gene expression at E12.5:correlation with proliferative potential and notstratificationK5 expression in the maxillary region appeared before the pre-cocious development of whisker pads, and by E12.5, it wasprominent in upper jaw (Fig. 3A,D). Expression also occurredlaterally along the body, but was still reduced ventrally anddorsally (Fig. 3B,C). At E12.5, K5 was not detected in coat,i.e. pelage follicles, whose development is known to lag behindthat of tactile vibrissae.

In E12.5 whisker pads, there was no strict correlationbetween ectodermal morphology and K5 expression. Traces ofmaxillary and frontonasal fusion were accentuated by a markedabsence of K5 in the nasolacrimal groove (arrowhead in Fig.3D). Both epidermis and the groove side of vibrissae remainednegative in a slightly older littermate (arrowhead in Fig. 3E),and yet were morphologically indistinct from surrounding β-gal positive regions.

Vibrissae follicles display a hierarchical developmentalsequence, so that a series of embryonic stages can be observedwithin a single whisker pad (Van Exan and Hardy, 1980 andreferences within). Vibrissae placodes in decreasing stages ofmaturation are situated on mesenchymal ridges (Van Exanand Hardy, 1980), and these stained positive for β-gal (left toright in Fig. F). Surprisingly, K5 promoter activity wasintense throughout all layers of newly stratified ectoderm(Fig. 3F-I). Numerous suprabasal mitoses were seen, oftenwith a plane of cleavage parallel to the basal lamina(arrowhead in Fig. 3H). In contrast, in single layeredectoderm (see Fig. 2M) and in adult epidermis, the plane ofmitotic cleavage is vertical to the basal lamina (Smart, 1970;Weiss and Zelickson, 1975).

Over other body regions, E12.5 ectoderm displayed repro-ducible, but varied patterns of β-gal activity (Fig. 3J; horizon-tal section of mid-body). Some areas of forelimb ectodermdisplayed equal periderm and ectoderm staining (Fig. 3Ja),while others showed higher periderm than ectodermal staining(Fig. 3Jb). In transitional areas, β-gal was higher in someregions than others (Fig. 3Jc), and absent in others (Fig. 3Jd-e). In these regions, there was greater morphological hetero-geneity in underlying mesenchyme than in ectoderm.

To summarize our E12.5 data, while K5 induction seemedto correlate with mesenchymal cues, maintenance of K5expression correlated with proliferative capacity, and a basal-like cell morphology, but not with stratification per se. This

parallel could only be recognized through whole-mountstudies, and had not been appreciated previously.

Basal keratin gene expression during pelage follicleinitiation: early signs of epidermal differentiationBy E13.5, K5 promoter activity occurred in most lateral areas(Fig. 4A), but it still lagged dorsally (Fig. 4B) and ventrally(not shown). Intensity of activity continued to be greatest overvibrissae, and the first signs of activity in pelage folliclesappeared in the upper dorso-lateral region (arrowheads in Fig.4B). Body ectoderm was now partially stratified (Fig. 4C-F),but mitoses and β-gal activity still occurred suprabasally, andmorphological signs of differentiation were absent.

A dramatic increase in K5 promoter activity took placebetween E13.5 and E14.5 (compare Fig. 4A with G). Dorso-lateral placode condensates stained intensely at early stages ofpelage follicle development (Fig. 4H). This gave the embryosurface a spotted appearance, which was readily visible in bothβ-gal stained embryos (Fig. 4G and inset in 4H) and in whole-mount in situ hybridizations with either K14 or K5 cRNAs(Fig. 4I; shown are K14 cRNA data). Even at this late stage,K5 and K14 cRNAs did not hybridize appreciably to dorsalepidermis over the spinal chord (Fig. 4J), nor to ventralepidermis (not shown).

From E13.5 to E14.5, vibrissae development progressed dra-matically (Fig. 4K-N). Throughout immature hair germs, K5gene expression was uniform (Fig. 4K). As development

C. Byrne, M. Tainsky and E. Fuchs

Fig. 1. pK5βgal6000 and expression in adult skin. pK5βgal6000(Byrne and Fuchs, 1993) contains 6000 bp of human K5 promoter(open box) linked to the β-gal gene (stippled box). SV40 sequences(black boxes) include synthetic splice donor and acceptor sites (Spl)and polyadenylation sequence (pA). The β-gal gene and SV40sequences are from pNassβ (Clontech, Palo Alto, CA). The constructis flanked by EcoRI (small black box) and HindIII (small triangle)sites. Mice transgenic for pK5βgal6000 express β-gal in the basalcompartment of stratified epithelia such as skin (shown), in the ORSand bulge of hair follicles (bracket), and in sebaceous glands(arrowhead). Bar represents 50 µm.

2373Gene expression in developing epidermis

proceeded, a region of K5-negative cells accumulated over theinvaginating epithelium (Fig. 4L-M). These pockets appearedto be morphologically differentiated, showing no signs ofmitoses. Thus in this region, K5 downregulation correlatedwith a loss of basal-like characteristics and with induction ofdifferentiation.

Follicle cells at the base of the hair germs, i.e. in contact

with mesenchymal aggregates, also showed a reduction in β-gal activity (Fig. 4N), a feature seen irrespective of whetherskin was sectioned prior to or after the X-gal assay. Theseresults were consistent with previous studies showing thatthese cells do not crossreact with an anti-K14 antiserum(Kopan and Fuchs, 1989). However, in contrast to K5-negativecells in the upper follicle segment, cells in the lower segment

Fig. 2. Detection of K5 promoter activity and mRNA in E9.5-E11.5 embryos. (A-G) Whole-mount detection of β-gal in K5βgal6000 embryos.(A) E9.5 ventral, (B) E10.5 lateral, (C) E11.5 lateral, (D) E11.5 ventral, (E) E11.5 dorsal, (F,G) K5 promoter activity over posterior somites ofE11.5 embryo. Arrowheads denote striped pattern of expression. (Ha,b) E11.5 K5 cRNA hybridization in same region as in G (arrowheads asin G). (I-N) K5 promoter activity in E9.5 ectoderm. (I,J) Parasagittal sections through somites; (I), semi-thin toluidine blue counterstainedsection; (J) frozen section; (K) whole mount of striped somitic region, to be compared to (L) for visualizing how the alternating pattern ofstaining comes about; (M), section to illustrate that the plane of mitotic cleavage is vertical to the basal lamina at the single-layered stage (seebracket). (N) Bi-layered head ectoderm. Note single periderm cell positive for β-gal (bracketed). Basal lamina is marked by arrowheads. Bar:38 µm for I; 105 µm for J; 123 µm for L; 15 µm for M; and 29 µm for N.

2374 C. Byrne, M. Tainsky and E. Fuchs

Fig. 3. Detection of K5 promoter activity in E12.5 embryos. (A-C) Whole-mount staining for K5 promoter activity (A, lateral; B, ventral; Cdorsal view). (D) β-gal in whisker pad. Arrowhead points to nasolacrimal groove, negative for β-gal. (E) Lack of β-gal, seen in D, ismaintained in slightly older littermate (arrowhead points to site of former nasolacrimal groove). Note presence of vibrissae placodes flankingthe groove, which are partly K5 positive and partly K5 negative. (F-I) Sections through whisker pad showing K5 positive, stratifiedinterfollicular epidermis and vibrissae placodes on mesenchymal ridges in decreasing order of maturation. White arrowhead (H) shows mitosiswhere plane of cleavage is horizontal to basal lamina. (J) Transverse section through upper body of E12.5 embryo. Regions that are magnifiedin (Ja-Je) are boxed. Note differences in β-gal staining without appreciable variation in ectodermal morphology (Ja-Jc). Note: β-gal was barelydetectable in thin epidermis above neural tube (Jd) and ventral epidermis (Je). Bar: 110 µm for F, 35 µm for G and I, 15 µm for H, 555 µm for Jand 55 µm for Ja-e.

2375Gene expression in developing epidermis

Fig. 4. K5 promoter activity and basal keratin mRNA expression from E13.5 to E14.5. β-gal activity in lateral (A) and dorsal (B) E13.5transgenic embryo. Note first indication of pelage follicles, which stain intensely (arrowheads). (C-F) Lateral body skin sections showingstratification and very weak β-gal in inner, middle and outer layers, often most prominent in periderm. Suprabasal mitoses, where plane ofcleavage is parallel to basement membrane, are denoted by arrowheads in E. (G) β-gal activity in E14.5 embryo. (H) Section of pelage follicle;inset denotes magnified embryo surface of region sectioned. (I) in situ hybridization with K14 cRNA. (J) Dorsal whole mount showing absenceof β-gal over E14.5 neural tube. (K-M) 5 µm paraffin sections through E14.5 vibrissae follicles of increasing maturation. (O-Q) Body skinsections of E14.5 β-gal embryo. Note that in O, K5 promoter activity has diminished in the middle layer of epidermis, where morphologicaldifferentiation has occurred. Note that in P, expression is still prominent in epidermal layers where stratification has occurred withoutmorphological differentiation. (Q) In situ hybridization of a K5 cRNA probe against a similar skin section as in P demonstrating theequivalence of the two procedures. Bar represents 35 µm for C, H, K-M; 22 µm for D-F; 53 µm for N and 24 µm for O-Q.

2376 C. Byrne, M. Tainsky and E. Fuchs

Fig. 5. Localization of keratin and loricrin mRNAs from E13.5 to E16.5. (A-C) E13.5 embryos hybridized with K5, K1 and loricrin cRNAs,respectively. (D-F) E14.5 embryos hybridized with K5, K1 and loricrin cRNAs, respectively. (G,H) Close-up of vibrissae follicles of E14.5embryos hybridized with K5 and K1 cRNAs, respectively. (I-K) E15.5 embryos hybridized with K5, K10 and loricrin cRNAs, respectively.(L) E16.5 embryo hybridized with loricrin cRNA. (M-P) 10 µm frozen sections from E15.5 embryos. Sectioning was performed after whole-mount in situ hybridization. (M) localization of K5 mRNA (blue) to basal layer of stratified epithelium and K1 mRNA (brown) to suprabasallayer at E15.5. (N) K5 mRNA in basal layer and in periderm of ventral skin. (O) K10 mRNA in suprabasal layers of backskin. Arrowheadsdenote basement membrane. (P) Loricrin mRNA over a pelage follicle (backskin). Expression is confined to an upper layer of cells and there is,as yet, no apparent involvement in the follicle. Bar represents 20 µm for M and 35 µm for N-P.

2377Gene expression in developing epidermis

did not appear morphologically differentiated. Thus, evenwithin a single follicle, two different regulatory mechanismsappeared to be operating on K5 gene expression.

At E14.5, interfollicular epidermis began to show signs ofsuprabasal differentiation, and in these regions, K5 becamerestricted to the innermost layer (Fig. 4H,O). At this stage,differentiation was regional, and much of the epidermis stillshowed basal-like cells in the suprabasal layers and concomi-tant K5 promoter activity (Fig. 4P). These differences werealso reflected in analyses of sections from our whole-mount insitu hybridizations (Fig. 4Q).

Expression of K1, K10 and loricrin as epidermalcells begin to differentiate into spinous andgranular-like layersPreviously, K1 and K10 expression had been detected as earlyas E15.5 (Greer and Roop, 1991), with loricrin mRNAs in E16skin (Yoneda and Steinert, 1993). To assess whether thesemRNAs might also display patterning, we subjected wholeembryos to in situ hybridizations with K1, K10 and loricrincRNAs. All three mRNAs were first detected at E13.5 in thenasal region (Fig. 5B,C). At this time, K1, K10 and loricrincRNAs did not hybridize over dorsolateral epidermis, strength-ening our previous conclusion that stratification precedesdifferentiation in this region. By E14.5, K1 and K10 cRNAshybridized to whisker pads and paws (Fig. 5E), while loricrincRNAs hybridized strongly to ventral and lateral body skin butnot appreciably to whisker pads (Fig. 5F). These hybridizationpatterns were specific, and were not observed with sense strandprobes.

In vibrissae, K5 and K14 cRNAs hybridized to outer regions(Fig. 5G), while K1 and K10 cRNAs hybridized to innerregions (Fig. 5H). Sectioning (not shown) provided additionalbiochemical data to support our prior suggestion that K5/K14negative cells overlying developing placodes were differenti-ated, and that the switch in keratin expression correlates withdifferentiation, not stratification.

Consistent with previous PCR and northern analyses (Greerand Roop, 1991; K. Turksen and E. Fuchs, unpublished data),expression of differentiation-specific keratin mRNAs rose dra-

matically at E15.5 (Fig. 5, compare J with E), and loricrinmRNA rose at E16.5 (Fig. 5, compare L with K; Yoneda andSteinert, 1993). These increases took place over most of theembryo surface, with somewhat lesser hybridization dorsally.By E15.5, ventral skin showed confinement of basal keratinmRNAs to the innermost basal layer and the outermostperiderm layer (Fig. 5N) with the differentiation-specifickeratin mRNAs confined to suprabasal layers (Fig. 5M, doublelabel, K5 in blue and K1 in brown; Fig. 5O, K10; backskin).Loricrin cRNA was confined to the layer just beneath periderm(Fig. 5P; backskin). This appeared to be the granular layer,known to develop at this time (Sengel, 1976).

We were surprised that loricrin cRNA hybridized overpelage follicles, since sections showed no hybridization withinfollicles. The patterning may be reflective of pre-inductiveevents for inner root sheath (IRS) formation: we have detectedloricrin mRNA in adult follicle IRS (data not shown), despitenegative data to this point from other groups (Hohl et al., 1991;Yoneda and Steinert, 1993).

E16.5: the last day at which whole-mount in situhybridization can be used to track epidermalmarkersBy E16.5, last vestiges of K5-negative dorsal skin disappeared,and K5 promoter activity occurred over the entire embryosurface (Fig. 6A,B). Interfollicular epithelia maintained basal-specific expression, while pelage follicles displayed K5patterns reflecting a range of stages similar to that seen earlierin whiskers (Fig. 6C-F). By this time, K5 expression wasprominent in the outer follicle layer (Fig. 6E), probably pre-dicting the eventual confinement of K5 follicle expression tothe ORS (Fig. 1). K5 gene expression was down regulated inthe vicinity of dermal papillae (Fig. 6F), in a fashion analogousto that seen in vibrissae (Fig. 4M,N). After E16.5, as a stratumcorneum developed, skin became impermeable to theseprobes.

AP2 cRNA hybridization during developmentK5 and K14 promoters share a series of transcription factors,including AP2, Sp1 and protein 1/2 (Byrne and Fuchs, 1993).

Fig. 6. Detection of K5 promoter activity in E16.5 embryos. (A) Lateral and (B) dorsal whole-mount view of β-galactosidase activity intransgenic E16.5 embryo showing intense and uniform staining over the surface of the animal. (C-F) 5 µm paraffin sections showing pelagefollicles in increasing order of maturation. Arrow in D shows K5 negative cells in center of placode. Bar represents 35 µm for C-F.

2378

AP2 was of particular interest, not only because of functionalAP2 binding sites in a variety of epidermal promoters, but alsobecause of its enriched expression in epidermis (Leask et al.,1990; Snape et al., 1990; Mitchell et al., 1991). In contrast, Sp1shows ubiquitous expression, with no apparent enhancementin skin (Saffer et al., 1991).

The known AP2 gene has multiple splice variants includingAP2A, encoding a functional AP2, and AP2B, encoding aninhibitor form of AP2 with no DNA binding capacity (Buettneret al., 1993). From PCR analyses on embryonic skin mRNAs,two PCR products were generated with oligonucleotideprimers that cannot distinguish AP2A and AP2B mRNAs (Fig.7A). When reduced cycling times were used, the lower bandcould not be detected, suggesting that it was a minorcomponent of the PCR reaction. The upper band was producedusing cloned mouse AP2A and AP2B cDNAs as controls, indi-cating that this band represents AP2A/B. We do not yet knowthe identity of the lower band, but AP2A/B paralleled the

temporal appearance of basal epidermal keratin mRNAs, whilethe lower band disappeared during development.

When a primer set specific for AP2A was used, a majorspecies was generated. This was bona fide AP2A, as judged bya control with a mouse AP2A clone. The band became moreprominent at later stages of development, when K5 and K14gene expression was rising (Fig. 7B). A lower band appearedat longer cycling times (not shown), suggesting the possibilitythat there may be additional forms of AP2. When the AP2B-specific primer set was used, one major band was generated,which corresponded to authentic AP2B (Fig. 7B). This banddiminished in intensity during development, as keratinexpression was increasing. A minor band was detected inE13.5-E16.5 skin when extended cycling times were used.Overall, while there may be additional species of AP2-likemRNAs in mouse skin during development (see also Winninget al., 1991; Buettner et al., 1993), of the two known AP2variants, only AP2A mRNAs seemed to correlate with K5 andK14 gene expression.

One way to explore further whether AP2 or AP2-likefactors might be playing a significant role in keratin geneexpression is to assess whether mRNAs hybridizing to an AP2cRNA display pattern formation similar to that of keratincRNAs. To investigate this possibility, we subjected ourembryos to whole-mount in situ hybridization with a mouseAP2 cRNA that recognizes both AP2A and AP2B forms. Pre-viously, Mitchell et al. (1991) had conducted in situhybridizations with this probe on E8.5-E12.5 embryosections, and noted hybridization in surface ectoderm at E8.5,i.e. one day before we detected K5 β-gal activity in thesecells. Beyond this observation, conventional methods were oflimited value in relating AP2 cRNA hybridization to K5 andK14 patterns.

As shown in Fig. 8, AP2 cRNA patterns were remarkablysimilar to those of K5 and K14 cRNAs. At early stages, AP2cRNA localization reflected somite organization (Fig. 8A,AP2, compare with 8B, K5), as well as prominence in the firstbranchial pouch (Fig. 8C, compare with 2B, K5). By E12.5,hybridization extended over the dorso-lateral surface, and wasdetected in vibrissae (Fig. 8D). At E13.5 and E14.5, parallelsbetween K5 and AP2 cRNA hybridizations were particularlystriking (Figs. 8E-H). A major difference was in fore andhindlimbs, where K5 hybridization was not as prominent asAP2 (Fig. 8E, compare with 8F,G). Intriguingly, AP2 cRNAhybridization to pelage follicles at E13.5 (Fig. 8G) precededhybridization of K5 and K14 cRNAs to these areas (comparewith Fig. 4A).

Sectioning of hybridized embryos revealed that AP2 RNAswere largely ectodermal over much of the body surface. Thisincluded intense ectodermal hybridization in limb skin (Fig.8I). Interestingly, the spots of AP2 mRNA that were seen overE13.5 lateral body skin (Fig. 8G) appeared as discrete patchesof ectodermal hybridization (brackets, Fig. 8J) preceding mor-phological signs of hair follicle development. Such patternshave also been observed in embryonic chick skin stained withantibodies to the cell adhesion molecule tenascin (Jiang andChuong, 1992), suggesting that a series of biochemical changesin overlying ectoderm precede follicle placode formation.

Further sectioning revealed an even more complex patternof AP2 hybridization at E13.5. In some regions, e.g. whiskerpad, AP2 patterns were remarkably similar to K5 (8K-M). In

C. Byrne, M. Tainsky and E. Fuchs

Fig. 7. PCR detection of AP2 mRNAs in embryonic skin. RNAswere isolated from either whole embryos (E9.5-E13.5) or embryonicskin (14.5 to NB). Radiolabeled AP2 cDNAs were amplified andresolved by acrylamide gel electrophoresis and subjected tophosphoimage analysis and autoradiography. (A) Primers used wereto a segment shared by AP2A and AP2B mRNAs. A major 357 bpband represents authentic AP2A and AP2B as determined by itsselective amplification from AP2A and AP2B plasmids. At increasedcycle times, a lower band was also detected (open arrowhead). Atreduced cycle times (21 cycles), co-amplification with primers forrRNA (S16) allowed quantitation, and demonstrated the equivalenceof mRNA levels. NB, newborn.(B) Primers specific for AP2A wereused and the PCRs were repeated. At reduced cycle times, only thebona fide 401 bp AP2A band was detected. Under increased cycletimes (35), AP2A primers amplified an additional band which wasdifferentially expressed (not shown). (C) AP2B primers amplifiedone lower band (298 bp bona fide AP2B, closed arrowhead) and oneminor species (open arrowhead).

2379Gene expression in developing epidermis

other regions, e.g. upper and lower jaw, paws and tail, AP2cRNAs hybridized to both mesenchymal and ectodermal cells,a feature also noted by Mitchell et al. (1991) (Fig. 8N). This

mesenchymal AP2 hybridization seemed to be due to AP2B,as judged by hybridization with a specific probe (Fig. 8O andP). Often, however, the total probe showed a hybridization

Fig. 8. AP2 cRNAhybridization to skin ofwhole embryos.(A) AP2 cRNAhybridization in theectoderm overlyingsomites at E10.5.(B) β-gal activity ofsame region as in A.(C) AP2 cRNAhybridization tomaxillary andmandibular regions offirst branchial arch atE10.5. (D) AP2 cRNAhybridization at E12.5.Note staining inplacodes of vibrissaeand supra andsuborbital follicles.(E-G) E13.5 embryoshybridized with K5 (E)and AP2 (F,G) cRNAprobes, respectively.(H) AP2 cRNAhybridization of E14.5embryo. (I-N, P, Q,S)10 µm frozen sections.(I) E13.5 limb skinfrom AP2 cRNA wholemount. (J) E13.5backskin from AP2whole mount, showingthat AP2 positive dots(brackets) precedemorphological signs ofpelage follicleformation. (K-M)E13.5 AP2 (K and M)or K5 (L) wholemounts in vibrissaeregion; dp, dermalpapilla; arrowhead,layer of matrix cells.(N) E13.5 paw skinshowing AP2epidermal andmesenchymal staining.Arrowheads denotebasement membrane.(O) E13.5 whole mountof embryo hybridizedwith a specific AP2Bprobe showing absenceof AP2B in nostrils andvibrissae, but presenceof AP2B in snout.(P) E13.5 snout skinfrom AP2B whole

mount showing confinement of signal to mesenchyme, compared to Q which shows predominant epidermal AP2 staining in adjacent vibrissaeregion. (R) Whole mount of E15.5 AP2 hybridization. (S) E15.5 skin from AP2 whole mount showing basal expression. Arrowheads in S,basement membrane. Bar represents 118 µm (I,J); 230 µm (K); 167 µm (L,M); 87 µm (N,P, Q,S).

2380

pattern similar to K5, which likely reflected the pattern ofAP2A-like factors (Fig. 8Q).

By E15.5, as the skin became both stratified and differenti-ated, hybridization detected by the general AP2 probe appearedregionally confined to basal cells (Fig. 8R and S), similar toK5. Collectively, our data suggested that AP2, but not AP2B,preceded slightly and was strikingly similar to the pattern ofK5 gene expression in the developing embryo.

AP2 can impart to hepatocytes the ability to expressK14 and K5 genes in cultureThe correlation between AP2A and basal keratin cRNA

hybridizations prompted us to explore whether AP2 or an AP2-like factor might be sufficient to impart to other cell types theability to express K5 and K14. We focused on the human liverhepatocyte line, HepG2, which normally expresses K8 andK18, but not K5, K14, or AP2 (Williams et al., 1988). As testgenes we used pK5cat6000 and pK14cat2300, containing 6000and 2300 bp, respectively, of human K5 and human K14 5′upstream sequence (Leask and Fuchs, 1990; Byrne and Fuchs,1993).

On their own, K5 and K14 promoters generated minimal catexpression in HepG2 cells (Fig. 9A and B, respectively, set 1).In contrast, these promoters produced marked expression inhuman epidermal (SCC13) keratinocytes, which expressendogenous K5 and K14 and which have abundant AP2activity (Leask et al., 1990). When a human AP2A expressionvector (Buettner et al, 1993) was co-transfected withpK5cat6000, a substantial increase (up to 18X) in catexpression was observed in HepG2, but not in SCC13 (Fig. 9A,sets 2-4). Similarly, up to a 44× increase in cat expression wasobserved when HepG2 cells were co-transfected withpK14cat2300 and the AP2A vector (Fig. 9B, sets 2-4). At highlevels of AP2A, the response was reduced, suggestive ofsquelching. The concentration needed to see squelching variedwith promoter and cell type, and the effects could not beexplained by variation in tk-β-gal control expression.

When pK5cat6000 was co-transfected with an expressionvector encoding the dominant negative AP2B, K5 promoteractivity remained low in HepG2 cells and was suppressed inSCC13 cells (Fig. 9A, set 5). For the K14 promoter, AP2B hadno apparent effect in HepG2 cells, and in SCC13 cells, itappeared to relieve the squelching described above (Fig. 9B,set 5).

When these experiments were repeated using p5xTREcolTK-CAT, containing 5 repeats of an AP1 site linked to the Herpesthymidine kinase promoter (Angel et al., 1988), no apprecia-ble AP2A-mediated enhancement in HepG2 cells wasobserved (not shown). Collectively, our results demonstratethat a simple epithelial cell, which is non-permissive forepidermal keratin gene expression, can be converted to a per-missive cell by expression of AP2A. These data are also con-sistent with our PCR results, and suggest that as ectodermaldevelopment proceeds and K5 and K14 expression rise,positive acting AP2 forms are likely to prevail in basal cells.Testing this prediction will require the cloning of all AP2-likeforms and the development of monospecific AP2 antibodies.

DISCUSSION

Our investigation of human K5-β-gal transgene activityindicate that (1) sequences involved in the complex develop-mental program of basal keratin expression are containedwithin 6000 bp of 5′ upstream region of the human K5 gene,(2) these sequences are evolutionarily conserved and (3) β-galpattern formations are bona fide representations of the patternsof K5 and K14 gene expression. The high sensitivity of the β-gal assay made it ideal for early embryonic studies, whenmRNA and protein expression levels are extremely low. Whilewe cannot rule out minor differences between β-gal andendogenous K5 gene expression, we did not detect such dif-ferences at times (E10.5-E16.5) where data with both probes

C. Byrne, M. Tainsky and E. Fuchs

Fig. 9. Effect of exogenous AP2 on keratin expression in culturedcells. (A) pK5cat6000 and AP2 expression vectors (pNAP2A,pNAP2B; Buettner et al, 1993) were transfected into SCC-13 andHepG2 cells and assayed 60 hours later for CAT activity.Set 1, pK5cat6000 co-transfected with empty expression vector;set 2, pK5cat6000 co-transfected with 0.5× molar ratio of pNAP2A;set 3, pK5cat6000 co-transfected with 1× molar ratio of pNAP2A;set 4, pK5cat6000 co-transfected with 2× molar ratio of pNAP2A;set 5, pK5cat6000 co-transfected with 1× molar ratio of pNAP2B.Expression is given as a percentage of pK5cat6000 levels in SCC13cells in the presence of empty expression vector. (B) Same as Aexcept pK14cat2300 was used as test vector. Expression is given as apercentage of pK14cat2300 levels in SCC13 cells in the presence ofempty expression vector. Note: reduction of activity in set 4 was notattributed to an AP2A-mediated enhancement of the tk-β-gal control.

2381Gene expression in developing epidermis

could be analyzed. When coupled with whole-mount in situhybridization, the approach enabled us to uncover novel andhitherto unexplored patterns of gene expression in embryonicdevelopment. The existence of epidermal patterning under-scores the importance of whole-mount analyses in evaluatingprograms of gene expression in skin, and exposes the limita-tions of conventional in situ hybridizations, and of mRNA orprotein analyses of whole embryo (or organ) extracts.

Early patterns in K5 gene expression seem to beestablished from inductive cues imparted byunderlying mesenchyme and can occur in a single-layered ectodermMany studies have focused on the role of dermis in theformation of epidermal appendages such as hair follicles,feathers and scales (for reviews, see Sengel, 1986; 1990;Hardy, 1992). The specification of these structures occurs rel-atively late in embryogenesis and depends upon signalsimparted by dermis of diverse embryonic origins (Dhouailly etal., 1978; Van Exan and Hardy, 1984). The precise nature ofthe dermal-epidermal signals has not yet been determined,though a number of proteins demonstrate regional micro-heterogeneities in the vicinity of developing appendages(Sengel, 1990; Holbrook et al., 1993: present studies).Typically, such differences have been noted for dermal devel-opment post E13, when the round, immature mesenchymalcells have differentiated (Van Exan and Hardy, 1984).

A long-standing question has been whether the capacity tosynthesize epidermal keratins is conferred by underlying mes-enchyme or whether it is acquired by self-determination (forreview, see Sengel, 1986). Our studies demonstrate that thepatterning and timing of K5 induction coincide with well-known regional heterogeneities in the origins of dermal mes-enchyme, suggesting that the major K5 inductive cues comefrom developing dermal cells. Cues appear as early as E9.5,i.e. significantly earlier in epithelial morphogenesis than hadbeen suspected. In fact, in some regions, notably the ectodermoverlying somites, K5 gene induction even precedes formationof a columnar basal layer. This sets K5 induction as a veryearly biochemical marker of epidermal development andamong the earliest consequences of dermal-epidermal interac-tions thus far described. These findings necessitate revision ofprior assumptions that epidermal keratin induction correlateswith the morphogenesis of a bilayered epithelium (Jackson, etal., 1981; Moll et al., 1982; Dale et al., 1985; Kopan and Fuchs,1989).

What might be the biochemical natures of the inductivedermal signals? At the moment, it is too early to say. However,it is interesting that a number of Hox genes show patterns ofexpression that overlap partially with and are either parallel to,or mirror image to, those of the epidermal keratin genes (Huntet al., 1991; Candia et al., 1992; Takahashi et al., 1992).Further studies will be necessary to explore these issues.

Periderm formation and K5 expressionThe periderm has been a developmental enigma for manyyears. Most researchers agree that periderm stems from upwardmigration of ectoderm. This said, histological studies revealmitoses in this layer (Herken and Schultz-Ehrenburg, 1983),and morphological distinctions between periderm and under-lying embryonic basal layer have led to suggestions that

amniotic fluid rather than ectoderm and/or dermis may orches-trate the program of gene expression in these cells (for dis-cussion, see M’Boneko and Merker, 1988). Our studies addthree findings that underscore similarities between peridermand immature ectoderm: (1) K5 gene expression in singlelayered ectoderm, (2) K5 gene expression in periderm cells,and (3) coexpression of K5 in periderm and in embryonic basalcells in a highly patterned fashion. Hence, periderm not onlyappears to arise from stratification of a single layeredectoderm, but also it receives inductive cues for geneexpression from the same source as ectoderm.

Expression of K5 and K14 correlates withproliferative activity of developing basal cells, notwith stratificationK5 and K14 expression clearly correlate with proliferativeactivity of cells in embryonic skin and not stratification, andK5 and K14 mRNA downregulation correlate with differen-tiation and not stratification. While these findings are incontrast to earlier developmental studies on keratins, they areconsistent with morphological studies noting major differencesbetween embryonic and adult stratification (Smart, 1970). Dis-crepancies from past keratin studies can be explained by (1)the failure to appreciate that keratin gene expression is highlypatterned in embryonic skin, in a fashion which at early stagesdoes not correlate with cell morphology, and (2) the frequentuse of monoclonal antibodies, which were often-times not onlycross-reacting, but also prone to antigen masking, a feature thatis now a well-known problem in anti-keratin immunohisto-chemistry.

In our studies, we compared K5 and K14 mRNA expressionfrom E13 to E16.5, and throughout this time, patterns wereindistinguishable. In human embryogenesis, K5 expression hasbeen reported to precede that of K14 (Moll et al., 1982). Dueto the relatively short gestation period of mouse compared tohuman, it was not possible to assess whether such a differencealso occurs during mouse embryogenesis.

Expression of differentiation-specific keratins andloricrinMany studies correlate K1 and K10 induction with formationof a three-layered epithelium in mammalian skin (Moll et al.,1982; Dale et al., 1985; Kopan and Fuchs, 1989). However, itis clear from whole-mount studies that K1 and K10 mRNAexpression correlates with differentiation, not stratification.This process occurs well after somitic and neural crest differ-entiation to mesenchyme, and hence, early regulatory cuesseem to be quite different for the two keratin gene pairs. Incontrast, some later cues may be shared, as judged by regionalsimilarities in expression patterns post E13.5.

The timing of differentiation-specific markers was markedlyearlier than previously suspected. This was largely because K1and K10 appear much earlier in some regions e.g. nasal epithe-lium, than in backskin, i.e. the skin source for most prior inves-tigations. While K1 and K10 showed regional inductivepatterns that were similar to those of K5 and K14, loricrindiverged considerably, suggesting cues for its expression maydiffer at least in part from those of the keratins.

AP2 and basal epidermal gene expressionOur compilation of expression patterns of epidermal mRNAs

2382

during embryonic development should be useful for dissectingout the complex mechanisms underlying epidermal develop-ment and differentiation in mammals. We have begun to testthis notion by exploring the possible physiological relevanceof AP2 binding sites in the 5′ upstream regulatory regions ofmany epidermal genes.

Multiple forms of AP2 appear during skin embryogenesis(Winning et al., 1991; Snape et al., 1991; Buettner et al., 1993),and in mouse, a cRNA corresponding to a large codingsegment of AP2 hybridizes to tissues of ectodermal and neuralcrest lineages, peaking at E11.5 (Mitchell et al., 1991). Ourstudy extends AP2 analyses up to E15.5, and further suggeststhat AP2 cRNA hybridization patterns closely mimic, andprecede slightly, those of K5 and K14 cRNAs, while AP2BmRNAs are primarily localized in underlying mesenchyme.AP2-like mRNAs appear to be in the right place and at the righttime to play a significant role in controlling basal epidermalgene expression.

Our culture studies provide the first direct demonstration thatAP2 can act in a positive fashion to control K5 and K14 geneexpression, while AP2B counteracts this positive effect.Moreover, our data show that K5 and K14 gene expression canbe induced in cultured hepatocytes when they are co-trans-fected with an AP2 expression vector. These data don’t nec-essarily imply that the absence of AP2 in non-K5/K14 express-ing cell types is what accounts universally for the absence ofkeratin gene expression. In fact, we know that HeLa cellsexpress AP2 and yet they still do not express K5-β-gal (Lerschand Fuchs, unpublished data). Additionally, we know that ker-atinocytes possess other factors required for K5 and K14expression (Leask et al., 1990; 1991). This said, our studies dosuggest that (1) the AP2 sequences that hybridized toembryonic ectoderm are likely reflective of an overall balancethat is positive rather than negative for AP2 activity, and (2)AP2 is likely to play a significant role in K5 and K14 regula-tion in vivo, both with regards to its presence in basal-likeepithelial cells and with regards to its absence in some othercell types.

Our PCR analyses indicate that AP2 mRNA patterns may bemore complex than previously realized. Further studies will benecessary before we understand fully how this family ofmRNAs controls basal keratin gene expression during devel-opment. In the future, as expression of other epidermal RNAsare reevaluated using the approaches described here, we antic-ipate that new correlations will become apparent and newinsights into mechanisms involving epidermal developmentwill emerge.

A very special thank you goes to Dr Qian-Chun Yu who preparedsemithin sections, Linda Degenstein for transgenic mouse technologyand Jennifer Jacobs who assisted with photography and preparationof embryos. We also thank Dr Brian Aneskievich for discussion andGrazina Traska for help with tissue culture, Phillip Galiga for prepa-ration of art work and Paul Gardner for his help in synthesizingoligonucleotide primers. We thank Dr Pamela Mitchell (Institute forMolecular Biology, University of Zurich, Switzerland) and DrReinhard Buettner (Institute for Pathology, University of Regensburg,Germany) for mouse AP2 cDNA clones used in whole-mount in situhybridizations, Dr Stuart Yuspa (NIH) for mouse K1 and K10cDNAs. C. B. is a research associate supported by the Howard HughesMedical Institute. E. F. is an investigator of the Howard HughesMedical Institute. This work was supported by a grant AR31737 from

the National Institutes of Health. The University transgenic facility isin part supported by grant CA 14599 from the National CancerInstitute.

REFERENCES

Angel, P., Hattori, K., Smeal, T. and Karin, M. (1988). The jun proto-oncogene is positively autoregulated by its product, jun/AP-1. Cell 55, 875-885.

Banks-Schlegel, S. P. (1982). Keratin alterations during embryonic epidermaldifferentiation: a presage of adult epidermal maturation. J. Cell Biol. 93, 551-559.

Behrens, J., Lowrick, O., Klein-Hitpass, L. and Birchmeier, W. (1991). TheE-cadherin promoter: Functional analysis of a G.C-rich region and anepithelial cell-specific palindromic regulatory element. Proc. Natl. Acad. Sci.USA 88, 11495-11499.

Buettner, R., Kannan, P., Imhof, A., Bauer, R., Yim, S. O., GlockshuberVan Dyke, R. M. W., and Tainsky, M. A. (1993). An alternatively splicedmRNA from the AP-2 gene encodes a negative regulator of transcriptionalactivation by AP-2. Mol. Cell. Biol. 13, 4174-4185.

Byrne, C. and Fuchs, E. (1993). Probing keratinocyte and differentiationspecificity of the human K5 promoter in vitro and in transgenic animals. Mol.Cell. Biol. 13, 3176-3190.

Candia, A. F., Hu, J., Crosby, J., Lalley, P. A., Noden, D., Nadeau, J. H. andWright, C. V. E. (1992). Mox-1 and Mox-2 define a novel homeobox genesubfamily and are differentially expressed during early mesodermalpatterning in mouse embryos. Development 116, 1123-1136.

Chomczynski, P. and Sacchi, N. (1987). Single-step method of RNA isolationby acid guanidinium thiocyanate-phenol-chloroform extraction. Analyt.Biochem. 162, 156-159.

Conlan, R. A. and Rossant, J. (1992). Exogenous retinoic acid rapidly inducesanterior ectopic expression of murine HOX-2 genes in vitro. Development116, 357-68.

Couly, G. F., Coltey, P. M. and Le Douarin, N. M. (1992). Thedevelopmental fate of the cephalic mesoderm in quail-chick chimeras.Development 114, 1-15.

Dale, B. A., Holbrook, K. A., Kimball, J. R., Hoff, M. and Sun, T.-T. (1985).Expression of epidermal keratins and filaggrin during human fetal skindevelopment. J. Cell Biol. 101, 1257-69.

Dhouailly, D., Rogers, G. E. and Sengel, P. (1978). The specification offeather and scale protein synthesis in epidermal-dermal recombinations. Dev.Biol. 65, 58-68.

Fisher, C. and Holbrook, K. (1987). Cell surface and cytoskeletal changesassociated with epidermal stratification and differentiation in organ culturesof embryonic human skin. Dev. Biol. 119, 231-241.

Franke, W. W., Schmid, E., Schiller, D. L., Winter, S., Jarasch, E. D., Moll,R., Denn, H., Jackson, B. W. and Illmensee, K. (1981). Differentiation-related patterns of expression of proteins of intermediate-size filaments intissues and cultured cells. Cold Spring Harbor Symp. Quant. Biol. 46, 431-453.

Fuchs, E. and Green, H. (1980). Changes in keratin gene expression duringterminal differentiation of the keratinocyte. Cell 19, 1033-1042.

Fuchs, E. and Weber, K. (1994). Intermediate filaments: structure, dynamics,function, and disease. Annu. Rev. Biochem. 63, 345-382.

Greer, J. and Roop, D. (1991). Loricrin, a major keratinocyte cell envelopeprotein is expressed late in development. J. Invest. Dermatol. 96, 553.

Hanson, J. (1947). The histogenesis of the epidermis in the rat and mouse. J.Anat. 174-197.

Hardy, M. H. (1992). The secret life of the hair follicle. Trends Genet. 8, 159-66.

Heid, H. W., Werner, E. and Franke, W. W. (1986). The complement ofnative alpha-keratin polypeptides of hair-forming cells: a subset of eightpolypeptides that differ from epithelial cytokeratins. Differentiation 32, 101-119.

Herrmann, B. G. (1991). Expression pattern of the Brachyury gene in whole-mount TWis/Twis mutant embryos. Development 113, 913-917.

Hohl, D., Lichti, U., Breitkreutz, D., Steinert, P. M. and Roop, D. R. (1991).Transcription of the human loricrin gene in vitro is induced by calcium andcell density and suppressed by retinoic acid. J. Invest. Dermatol. 96, 414-418.

Holbrook, K. A., Smith, L. T., Kaplan, E. D., Minami, S. A., Hebert, G. P.and Underwood, R. A. (1993). Expression of morphogens during human

C. Byrne, M. Tainsky and E. Fuchs

2383Gene expression in developing epidermis

follicle development in vivo and a model for studying follicle morphogenesisin vitro. J. Invest. Dermatol. 101, 39S-49S.

Hunt, P., Gulisano, M., Cook, M., Sham, M., Faiella, A., Wilkinson, D.,Boncinella, E. and Krumlauf, R. (1991). A distinct Hox code for thebranchial region of the vertebrate head. Nature 353, 861-864.

Jackson, B. W., Grund, C., Winter, S., Franke, W. W. and Illmensee, K.(1981). Formation of cytoskeletal elements during mouse embryogenesis II.Epithelial differentiation and intermediate filaments in earlypostimplantation embryos. Differentiation 20, 203-216.

Jiang, T.-X. and Chuong, C.-M. (1992). Mechanisms of skin morphogenesis.I. Analysis with antibodies to adhesion molecules tenascin, N-CAM, andintegrin. Dev. Biol. 150, 82-98.

Kopan, R. and Fuchs, E. (1989) A new look into an old problem: keratins astools to investigate determination, morphogenesis, and differentiation inskin. Genes Dev. 3, 1-15.

Kopan, R. and Weintraub, H. (1992). Mouse Notch: Expression in hairfollicles correlates with cell fate determination. J. Cell Biol. 121, 631-641.

Leask, A., Byrne, C. and Fuchs, E. (1991). Transcription factor AP2 and itsrole in epidermal-specific gene expression. Proc. Natl. Acad. Sci. USA 88,7948-7952.

Leask, A., Rosenberg, M., Vassar, R. and Fuchs, E. (1990). Regulation of ahuman epidermal gene: sequences and nuclear factors involved inkeratinocyte-specific gene regulation. Genes Dev. 4, 1985-1998.

Leonard, M. W., Lim, C. K. and Engel, J. D. (1993). Expression of thechicken GATA factor family during early erythroid development anddifferentiation. Development 119, 519-531.

Lersch, R. and Fuchs, E. (1988). Sequence and expression of a type II keratin,K5, in human epidermal cells. Mol. Cell. Biol. 8, 486-493.

Lersch, R., Stellmach, V., Stocks, C., Giudice, G. and Fuchs, E. (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.

Lynch, M. H., O’Guin, W. M., Hardy, C. M. L. and Sun, T.-T. (1986).Acidic and basic hair/nail (‘hard’) keratins: their colocalization in uppercortical and cuticle cells of the human hair follicle and their relationship to‘soft’ keratins. J. Cell Biol. 103, 2593-2606.

Magnaldo, T., Bernerd, F., Freedberg, I. W., Ohtsuki, M. and Blumenberg,M. (1993). Transcriptional regulators of expression of K16, the disease-associated keratin. DNA and Cell Biol. 12, 911-23.

M’Boneko, V. and Merker, H.-J. (1988). Development and morphology ofthe periderm of mouse embryos (days 9-12 of gestation). Acta. Anat. 133,325-336.

Moser, M., Pscherer, A., Bauer, R., Imhof, A., Seegers, S., Kersher, M. andBuettner, R. (1993). The complete murine cDNA sequence of thetranscription factor AP-2. Nucl. Acids Res. 21, 4844.

Mitchell, P. J., Timmons, P. M., Hebert, J. M., Rigby, P. W. J. and Tijan, R.(1991). Transcription factor AP-2 is expressed in neural crest cell lineagesduring mouse embryogenesis. Genes Dev. 5, 105-119.

Moll, R., Moll, I. and Wiest, W. (1982). Changes in the pattern of cytokeratinpolypeptides in epidermis and hair follicles during skin development inhuman fetuses. Differentiation 23, 170-78.

Nelson, W. G. and Sun, T. T. (1983). The 50- and 58-kdalton keratin classes asmolecular markers for stratified squamous epithelia: cell culture studies. J.Cell Biol. 97, 244-51.

Nichols, D. (1981). Neural crest formation in the head of the mouse embryo asobserved by a new histological technique. J. Embryol. Exp. Morph. 64, 105-120.

Noden, D. M. (1988). Interactions and fates of avian craniofacial mesenchyme.Development. 103 Supplement, 121-140.

Powell, B., Crocker, L. and Rogers, G. (1992). Hair follicle differentiation:expression, structure and evolutionary consequences of the hair type IIkeratin intermediate filament gene family. Development 114, 417-433.

Richman, J. and Tickle, C. (1989). Epithelia are interchangeable betweenfacial primordia of chick embryos and morphogenesis is controlled by themesenchyme. Dev. Biol. 136, 201-210.

Roop, D. R., Huitfeldt, H., Kilkenny, A. and Yuspa, S. H. (1987). Regulatedexpression of differentiation-associated keratins in cultured epidermal cells

detected by monospecific antibodies to unique peptides of mouse epidermalkeratins. Differentiation 35, 143-150.

Rothnagel, J. A., Greenhalgh, D. A., Gagne, T. A., Longley, M. A. andRoop, D. R. (1993). Identification of a calcium inducible, epidermal-specificregulatory element in the 3′-flanking region of the human keratin 1 gene. J.Invest. Derm. 101, 506-513.

Saffer, J. D., Jackson, S. P. and Annaraella, M. B. (1991). Developmentalexpression of SP1 in the mouse. Mol. Cell. Biol. 11, 2189-2199.

Schweizer, J. and Winter, H. (1982). Keratin polypeptide analysis in fetal andin terminally differentiating newborn mouse epidermis. Differentiation 22,19-24.

Sengel, P. (1976). Morphogenesis of Skin. Cambridge, London, New York:Cambridge University Press.

Sengel, P. (1986). Epidermal-dermal Interaction. Berlin, Heidelberg, NewYork, Tokyo: Springer-Verlag.

Sengel, P. (1990). Pattern formation in skin development. Int. J. Dev. Biol. 32,33-50.

Smart, I. H. M. (1970). Variation in plane of cell cleavage during the processof stratification in mouse epidermis. Br. J. Dermatol. 82, 276-282.

Snape, A. M., Jonas, E. A. and Sargent, T. D. (1990). KTF-1, a transcriptionactivator of Xenopus embryonic keratin expression. Development 109, 157-165.

Snape, A. M., Winning, R. S. and Sargent, T. D. (1991). Transcription factorAP-2 is tissue-specific and is closely related or identical to keratintranscription factor 1 (KTF-1). Development 113, 283-293.

Stark, H. J., Breitkreutz, D., Limat, A., Ryle, C.-M., Roop, D., Leigh, I. andFusenig, N. (1990). Keratin 1 and 10 or homologues as regular constituentsof inner root sheath and cuticle cells of human hair follicle. Eur. J. Cell Biol.52, 359-372.

Takahashi, Y., Monsoro-Burq, A.-H., Bontoux, M. and Le Douarin, N. M.(1992). A role for Quox-8 in the establishment of the dorsoventral patternduring vertebrate development. Proc. Natl. Acad. Sci. USA. 89, 10237-10241.

Thorey, I. S., Meneses, J. J., Neznanov, N., Kulesh, D. A., Pedersen, R. A.and Oshima, R. G. (1993). Embryonic expression of human keratin 18 andK18-beta-galactosidase fusion genes in transgenic mice. Dev. Biol. 160, 519-534.

Tyner, A. L. and Fuchs, E. (1986). Evidence for posttranscriptional regulationof the keratins expressed during hyperproliferation and malignanttransformation in human epidermis. J. Cell Biol. 103, 1945-1955.

Van Exan, R. J. and Hardy, M. F. (1980). A spatial relationship betweeninnervation and the early differentiation of vibrissa follicles in the embryonicmouse. J. Anat.. 131, 643-656.

Van Exan, R. J. and Handy, M. F. (1984). The differentiation of dermis in thelaboratory mouse. Amer. J. Anat. 169, 149-164.

Watt, F. M., Keeble, S., Fisher, C., Hudson, D. L., Codd, J. and Salisbury,J. R. (1989). Onset of expression of peanut lectin-binding glycoproteins incorrelated with stratification of keratinocytes during human epidermaldevelopment in vivo and in vitro. J. Cell Sci. 94, 355-359.

Weiss, L. W. and Zelickson, A. S. (1975). Embryology of the epidermis:ultrastructural aspects. II Period of differentiation in mouse with mammaliancomparisons. Acta Dermatol. 55, 321-329.

Wessells, N. K. (1963). Effects of extra-epithelial factors on the incorporationof thymidine by embryonic epidermis. Exp. Cell Res. 30, 36-55.

Williams, T., Admon, A., Luscher, B. and Tjian, R. (1988). Cloning andexpression of AP2, a cell-type-specific transcription factor that activatesinducible enhancer elements. Genes Dev. 2, 1557-1569.

Winning, R. S., Shea, L. J. and Sargent, T. D. (1991). Developmentalregulation of transcription factor AP-2 during Xenopus laevisembryogenesis. Nucl. Acids. Res. 19, 3709-3714.

Wu, Y.-J. and Rheinwald, J. G. (1982). A new small (40 kd) keratin filamentprotein made by some cultured human squamous cell carcinomas. Cell 25,627-635.

Yoneda, K. and Steinert, P. (1993). Overexpression of human loricrin intransgenic mice produces a normal phenotype. Proc. Natl. Acad. Sci. USA.90, 10754-10758.

(Accepted19 May 1994)