The Distal Regulatory Region of the Human Involucrin Promoter Is Required for Expression in Epidermis*

Human involucrin (hINV) is a precursor of the keratinocyte cornified envelope that is specifically expressed in the suprabasal layers of stratifying squamous epithelia. The promoter distal (DRR) and proximal regulatory regions (PRR) are required for optimal in vitro expression (Welter, J. F., Crish, J. F., Agarwal, C., and Eckert, R. L. (1995) J. Biol. Chem. 270, 12614–12622; and Banks, E. B., Crish, J. F., Welter, J. F., and Eckert, R. L. (1998)Biochem. J. 331, 61–68). We now present the complete sequence of these regions and evaluate their ability to drive in vivo transcription. Transgenes containing 5000 or 2473 base pairs of upstream regulatory region drive tissue- and differentiation-appropriate expression in stratifying surface epithelia. In contrast, transgenes containing 1953, 1333, 986, or 41 base pairs of upstream regulatory region are not expressed in surface epithelia, indicating that loss of the DRR (nucleotides −2474/−1953) results in loss of expression. Fusing the isolated DRR region directly to the hINV minimal promoter restores surface epithelial expression. Sequences downstream of the transcribed gene are not required for appropriate expression. The −1953/−41 segment influences the pattern of differentiation-dependent expression. The −986/−41 region, which includes the PRR, drives expression in internal epithelia.

Epidermal keratinocytes undergo a program of differentiation that results in assembly of the epidermis (1). This differentiation process involves a series of morphological and biochemical changes that are tightly controlled and involve specific temporal and spatial changes in gene expression (1)(2)(3). These changes include activation of the gene that encodes human involucrin (hINV). 1 Involucrin is not expressed in the basal epidermal layer, but expression is activated in the late spinous layer and continues in the granular layer (4,5). Involucrin is an ␣-helical, rod-shaped, 68 kilodalton, glutamine-and glutamic acid-rich structural protein that is an efficient trans-glutaminase substrate (6 -9). During the final stages in keratinocyte differentiation, involucrin is incorporated, via the formation of interprotein ⑀-(␥-glutamyl)lysine cross-links with other proteins, into the keratinocyte cornified envelope (10). This envelope provides an essential protective barrier (11)(12)(13)(14). Involucrin play a similar role in other stratifying epithelia, including esophagus, cornea, ectocervix, vagina, etc. (15,16). In each tissue, expression is confined to the suprabasal layers. We are interested in the regulation of hINV expression and identification of the regulatory elements responsible for the tissuespecific and differentiation-appropriate expression in these tissues. Recent studies indicate that a 6-kb segment of the hINV gene, including the hINV upstream regulatory region, the intron, and the coding sequence, targets expression to the appropriate tissues (16). Targeted expression is also observed when the upstream regulatory region is linked to the SV40 intron and ␤-galactosidase (17). In vitro studies show that the promoter is expressed in cells derived from stratifying epithelia (18,19), is regulated by agents that modulate differentiation (18 -20), and is not expressed in fibroblasts (18). Deletion mapping and point mutation studies identify two regions of the upstream regulatory region, the proximal regulatory region (PRR) and the distal regulatory region (DRR), that are required for optimal expression in cultured cells (18,19). Both the DRR and PRR contain AP1 sites, AP1-5 and AP1-1, respectively, that are required for optimal promoter activity (18). In addition, the PRR contains an Sp1 site that synergistically activates expression in conjunction with the AP1-5 site (19). In the present study, we evaluate the role of the DRR in mediating in vivo expression using a transgenic mouse model. Among other findings, we demonstrate that the DRR is necessary and sufficient for expression of the transgene in epidermis and other stratifying epithelia.

MATERIALS AND METHODS
Construction of hINV Transgenes-E13E was constructed by EcoRI digestion of phage, Charon 4AI-3 (21). A 13-kb EcoRI fragment was then subcloned into pBKS(ϩ) to yield pBKS-E13E. The EcoRI insert from this plasmid is shown in Fig. 1. The H6B transgene is a 6-kb HindIII/BamHI fragment that was derived by restricting Charon 4AI-3 with HindIII/BamHI and subcloning the resulting 6-kb fragment into HindIII/BamHI-restricted pSP64 to yield pSP64I-3 H6B (21). Promoter deleted transgenes were constructed by taking advantage of unique restriction sites located upstream of the transcription start site. Consequently, the Ha5.5B transgene was generated by digesting pSP64I-3 H6B with HaeII/BamHI and isolating the HaeII/ BamHI. Likewise, the A4.3B and K4B transgenes were isolated from pSP64I-3 H6B by digesting with AccI/BamHI and KpnI/BamHI, respectively. P3.4B was generated by complete BamHI and partial PstI digestion. DRR-P3.4B was created by cloning the HindIII/HaeII DRR segment immediately upstream of the minimal promoter in P3. * This work utilized the facilities of the Skin Diseases Research Center of Northeast Ohio (National Institutes of Health Grant AR39750) and was supported by grants from the National Institutes of Health (to R. L. E.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AF085346.
Generation of Transgenic Mice-Mouse embryos from a B6CBA ϫ B6CBA mating were injected with each hINV gene construct and implanted into surrogate mothers using standard methodology (22). The offspring were characterized for the presence of the human involucrin (hINV) transgene by blotting of tail DNA (16,22).
Detection of hINV Expression-To detect hINV expression in mouse tissues, total protein extracts were prepared from tissue samples in Laemmli sample buffer, electrophoresed on acrylamide gels, and transferred to nitrocellulose for immunoblot. The blot was incubated with a primary antibody prepared against human involucrin, diluted 1:8000 as described previously (16), followed by visualization using a chemilumi-nescent detection system. To detect hINV in tissue sections, samples were fixed in buffered formalin, embedded in paraffin, and sectioned (16). The sections were deparaffinized, blocked, incubated with primary anti-hINV antibody and secondary detection agents exactly as described previously (16). The involucrin antibody was used at a dilution of 1:1000 and was preabsorbed on 3T3 fibroblast cells prior to use.
Detection of hINV mRNA-Expression of hINV mRNA in mouse tissues was detected by reverse transcription polymerase chain reaction (RT-PCR). Two micrograms of total RNA, isolated from mouse tissue, was reverse transcribed in 10 mM Tris-HCl buffer (pH 8.3) containing 50 mM KCl, 5 mM MgCl 2 , 1 mM of each dNTP, 1.6 g of oligo-p(dT) 15 primer, 50 units of RNase inhibitor, and 20 units of reverse transcriptase (Boehringer Mannheim) for 10 min at 25°C and for 60 min at 42°C in a 20-l reaction. The reverse transcriptase was inactivated by heating, and 20 l of the reverse transcription reaction was added to a 100-l PCR amplification reaction containing 10 mM Tris-HCl, pH 8.3, 50 mM KCl, 1.5 mM MgCl 2 , 1.5 mM each dNTP, 0.2 M each upstream (5Ј-CTC CAC CAA AGC CTC TGC, in exon 1) and downstream (5Ј-CTG CTT AAG CTG CTG CTC, exon 2) primers, and 2.5 units of Taq DNA polymerase. The PCR cycling reactions were 96°C for 1 min, 57°C for 1 min, and 72°C for 2 min for 35 cycles. These primers amplify a 380 bp segment of the hINV mRNA sequence. Because the primers are in different exons, PCR amplification of contaminating genomic DNA can be distinguished by production of a much larger band. ␤-actin was amplified in parallel reactions as a control.
DNA Sequencing-The upstream regulatory region from the hINV gene (21) was isolated and sequenced using Maxam-Gilbert (23) and dideoxy sequencing (24). The complete sequence was determined in both directions.

RESULTS
Structure of hINV Transgenes-The hINV transgenes are shown in Fig. 1. E13E includes approximately 5,000 bp of upstream sequence, the transcribed gene, and 4.5 kb of downstream sequence. All other constructs are truncated at a BamHI site located just downstream of the transcription stop sequence. H6B, Ha5.5B, A4.3B, and K4B are progressively truncated from the 5Ј end and contain 2473, 1953, 1333, and 986 bp of upstream regulatory region, respectively.
A 520-bp Segment of the hINV Upstream Regulatory Region Is Required for hINV Expression in Epidermis-To detect expression of the hINV protein in mice, we assayed epidermis and kidney for expression of the hINV protein by immunoblot of whole cell extracts as described previously (16). Equivalent quantities of protein extract (15 g) were immunoblotted. As shown in Fig. 2A, hINV protein is detected in epidermis in E13E and H6B mice, but not in Ha5.5B, A4.3B, or K4B mice. In contrast to the differences in hINV expression in epidermis, hINV transgene expression is retained in the kidney of all transgenic lines (Fig. 2B). These results indicate that the 520-bp DNA segment located between Ϫ2473 to Ϫ1953 is required for expression of the hINV transgene in epidermis. Nontransgenic mice did not produce hINV.
hINV Is Synthesized by Epidermal and Kidney Cells-The above experiments show that hINV protein is present in the epidermis and kidney of E13E and H6B mice, but not in Ha5.5B, A4.3B, K4B, or P3.4B mice. However, these results do not demonstrate that hINV is synthesized in these tissues. To directly demonstrate synthesis, we assayed for hINV mRNA by RT-PCR. Fig. 3 shows that mRNA encoding hINV is produced in the epidermis (E) and kidney (K) of H6B mice, but not in the epidermis of Ha5.5B mice. Thus, hINV is synthesized in these FIG. 4. Immunodetection of hINV in epidermis and cervix. Tissue sections were prepared from the epidermis (EPI) and ectocervix (EC) of representative animals from the E13E, H6B, Ha5.5B, and K4B transgenic lines. The epidermal samples were derived from the footpad. Sections were incubated with the hINV-specific primary antibody followed by treatment with a horseradish peroxidaselinked secondary antibody. No signal was detected in nontransgenic mice or in transgenic mice for which the incubation with primary antibody was omitted. The arrows indicate the surface of the epithelium, and the arrowheads indicate the basal (proliferative) layer of the epithelium.
tissues, and loss of hINV expression in the Ha5.5B as compared with H6B is associated with loss of hINV mRNA.
Differentiation Appropriate Expression of the hINV Transgene-We used immunological techniques to evaluate the differentiation-dependence of expression (Fig. 4). hINV is detected in the upper spinous and granular layers in footpad epidermis (EPI) in E13E and H6B mice, but no expression is detected in the basal layer (arrowheads). Suprabasal expression is also observed in the ectocervical epithelium (EC) in these mice. In contrast, no expression is observed in epithelium or epidermis of transgenic strains Ha5.5B or K4B (Fig. 4), and no expression is observed in nontransgenic mice (not shown).

FIG. 5. Immunodetection of hINV in kidney.
Kidney tissue sections were prepared from representative K4B transgenic animals and from nontransgenic (NT) animals. Sections were incubated with the hINV-specific primary antibody followed by treatment with a horseradish peroxidase-linked secondary antibody.
The arrowheads indicate the lumenal epithelium lining the distal convoluted tubules. The asterisks identify the renal corpuscles.
FIG. 6. Sequence of the hINV upstream regulatory region. The sequence of the hINV upstream regulatory region from Ϫ2473 to Ϫ1 is shown. The map displays key restriction enzymes sites (boxed) that were used in transgene construction (i.e. HindIII, HaeII, AccI, KpnI, and Cel II) and also identifies the activator protein-1 (AP1-1, -2, -3, -4, and -5) (18,19) and Sp1 transcription factor binding sites. The AP1 sites are indicated in bold, the Sp1 site is underlined, and the TATA box is double underlined in bold (21). Other transcription factor binding sites are also present, but are not shown. Sequence of the hINV Upstream Regulatory Region-Our results suggest that important regulatory elements are localized within the Ϫ2473/Ϫ1953 segment. To identify these potential elements, the entire 2473-bp hINV upstream regulatory region (Ϫ2473/Ϫ7) was sequenced (Fig. 6).
The DRR Region Is Sufficient to Drive Transcription in Epidermis-The experiments shown in Figs. 2, 3, and 4 suggest that the Ϫ2473/Ϫ1953 segment is required for expression in stratifying epithelia. The minimal hINV promoter construct, P3.4B, which contains only 41 bp of upstream sequence (Fig.  7A), shows no expression in epidermis (Fig. 7B). Fusion of the DRR-containing segment (Ϫ2473/Ϫ1953) immediately upstream of the minimal promoter restores expression in stratifying epithelia (Fig. 7B). The DRR-P3.4B construct is expressed at a level compared with that observed with H6B (Fig. 7B). Immunohistological examination of the pattern of expression reveals that the DRR-P3.4B drives expression in the suprabasal layers in the ectocervical epithelium (Fig. 8).

Regulation of hINV Expression in Epidermis-Identifying
mechanisms that govern tissue-specific and differentiation-appropriate gene expression in stratifying epithelia, such as epidermis, is an area of intense interest. In these epithelia, stem cells give rise to daughter cells that then differentiate to form the suprabasal layers of the tissue. This process produces profound changes in cell morphological and biochemistry. Several marker proteins have been identified that are differentially regulated during this process, including loricrin (25-28), filaggrin (29), cornifin (30, 31), transglutaminase (32), and intermediate filament proteins (29,(33)(34)(35). Involucrin, a precursor of the cornified envelope, is an early marker of keratinocyte dif-ferentiation. In epidermis, involucrin is expressed in the late spinous layer and granular layer (4,15,16,36,37); in ectocervix it is expressed in the layer immediately above the basal layer (16). Other genes, such as loricrin, are expressed later in the differentiation process (26,38).
Identification of a 520-bp Segment That Is Required for hINV Expression in Epidermis-Previous studies implicate activator protein 1 (2,18,19,39,40), Sp1 (19), POU domain (41), and other factors (42) as regulators of hINV promoter activity in cultured cells, and a previous transgenic study showed that the H6B construct (Fig. 1) can drive production of hINV in the epidermis and ectocervix of mice (16). H6B includes 2.5 kb of DNA upstream of the transcription start site, the two exons, the intron, and a short sequence segment downstream of the polyadenylation signal. Although the expression of this construct appeared physiologic, we wanted to test larger constructs to determine whether additional sequence would alter expression. The E13E results show that the presence of additional DNA does not change the regulation. Thus, neither the downstream region (ϩ4500/ϩ8000) nor the most distal upstream region (Ϫ5000/Ϫ2473) is required for appropriate expression in stratifying epithelia. To identify regions that are important for expression, we tested promoter deletions. In contrast to E13E and H6B, Ha5.5B does not express in epidermis, cervix, or other stratifying epithelia, suggesting that elements within the 520-bp Ϫ2473/Ϫ1953 DRR segment (19) are essential for this expression. Truncation of additional sequence, in constructs A4.3B and P3.4B, does not restore expression. The loss of expression was not because of a general inactivation of the promoter, as, except for P3.4B, these constructs retain expression in kidney. Only P3.4B was off in all tissues, suggesting that the 41 bp of upstream region present in this construct represents the basal promoter. This region includes only the TATA box and some associated sequences (Fig. 6). Thus, the promoter appears to be organized into modular units including a "stratifying epithelial module" located in the DRR, and a kidney module located between Ϫ986/Ϫ41, and a basal promoter (Ϫ41/Ϫ7). However, additional experiments, discussed below, suggest that these regions cannot function completely independently.
The DRR Region Is Both Necessary and Sufficient for Expression in Epidermis-Splicing the DRR segment upstream of P3.4B, the minimal promoter construct, restores expression in stratifying epithelia. This result suggests that the DRR segment is necessary for expression and that the Ϫ1953/Ϫ41 segment is not required. However, it could be argued that the DRR region may be acting in conjunction with DNA in the transcribed region of the gene (e.g. the intron). However, the hINV promoter also drives stratifying epithelia-specific expression of heterologous genes (17). The smallest construct tested, the Ϫ2473/Ϫ41 segment, drives expression of heterologous genes in stratifying epithelia. 2 Taken together, these results argue that the DRR is both necessary and sufficient for expression in stratifying epithelia.
Although the DRR is necessary and sufficient for expression, other sequences appear to be necessary to produce authentic differentiation-dependent expression. In the absence of the Ϫ1953/Ϫ41 segment, the DRR produces expression only in the extreme suprabasal cell layers. This suggests that interaction between elements in the DRR and in the Ϫ1953/Ϫ41 segment may be required for expression in the layers immediately suprabasal to the stem cell layer.
Expression in Other Stratifying Epithelia-The present studies focus on the ectocervical and epidermal epithelia. However, we have also checked several other surface epithelia to determine whether the DRR is generally required for expression. We have studied esophagus, epidermis, footpad, ectocervix, and trachea. In each case, the DRR is required for expression, indicating that the DRR is a generally required element for surface epithelial expression.
DRR Sites and Gene Regulation-DNA sequence analysis indicates that the DRR contains binding sites for several transcription factors, including functionally important Sp1 and AP1 sites (18 -20). Our previous in vitro studies show that the minimal promoter does not drive expression in keratinocytes; however, addition of the DRR restores hINV promoter activity (19). Moreover, mutation of the AP1 site (AP1-5) inactivates the promoter. AP1 has been shown to be an important regulator in several genes that are expressed in a differentiation-dependent manner in surface epithelia (38,40,(43)(44)(45)(46)(47)(48)(49)(50)(51)(52)(53)(54)(55). Based on these results and the observation that the DRR is necessary and sufficient for expression, we hypothesize that the DRR AP1 site (AP1-5) is required for expression. We are currently performing experiments to test this hypothesis in vivo.