HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Jang, S.-I. Articles by Markova, N. G. Search for Related Content PubMed PubMed Citation Articles by Jang, S.-I. Articles by Markova, N. G. Social Bookmarking                      What's this?

J Biol Chem, Vol. 275, Issue 20, 15295-15304, May 19, 2000

Complex Interactions between Epidermal POU Domain and Activator Protein 1 Transcription Factors Regulate the Expression of the Profilaggrin Gene in Normal Human Epidermal Keratinocytes^*

Shyh-Ing Jang, Nevena Karaman-Jurukovska^§, Maria I. Morasso, Peter M. Steinert, and Nedialka G. Markova^§¶

From the  Laboratory of Skin Biology, NIAMS, National Institutes of Health, Bethesda, Maryland 20892 and the ^§ Living Skin Bank, SUNY-Stony Brook, Stony Brook, New York 11794

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The human profilaggrin gene is expressed in the granular layer during the late stages of the epidermal differentiation. The proximal promoter region of the gene confers high levels of keratinocyte-specific transcription via interactions with c-Jun/c-Fos heterodimers. Here we provide evidence for another level of complexity in the regulation of the profilaggrin promoter activity. The POU domain proteins Oct1, Skn1a/i, and Oct6, which are abundantly expressed in the epidermal cells, act to both stimulate and repress transcription in a general and a cell type-specific mode. While binding to specific recognition elements within the promoter region, they exert their effects by either stimulating or antagonizing the c-Jun-dependent activity of the promoter. The response of the promoter to forced expression of the POU domain proteins reflects the effect of these transcription factors on the endogenous profilaggrin mRNA synthesis and suggests that the latter requires a fine balance in the amounts and the activities of the individual activator protein 1 and POU domain proteins.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The epidermis is a stratified epithelium composed of a single layer of proliferating basal keratinocytes and several layers of terminally differentiating suprabasal keratinocytes. As cells migrate outwards from the basal layer, keratinocyte differentiation proceeds with the synthesis of sets of biochemical markers, such as specific keratins and their associated proteins, cell envelope precursors, and processing enzymes (1-3). The late stages of epidermal keratinocyte differentiation are characterized by the robust synthesis of the intermediate filament aggregating protein profilaggrin. The primary control of profilaggrin production is at the level of transcription (4). Transcription factors belonging to the AP1¹ family are key factor in this control. Interactions of c-Jun/c-Fos heterodimers with an element located 77 base pairs upstream of the transcription initiation site of the profilaggrin gene are indispensable for maintaining high levels of promoter activity, whereas DNA binding-independent interactions of JunB serve to repress the promoter. Despite the essential role of AP1, it is unlikely that these factors are solely responsible for optimal levels of profilaggrin transcription. In fact, we have previously reported that the effect of c-Jun is modulated in cooperation with Ets transcription factors, including the dual Ets/HMG protein Jen (5). By using transgenic mice, we have also shown that the homeodomain containing protein Dlx3, which recognizes AT-rich motifs in the vicinity of the initiation site, has the potential to profoundly influence profilaggrin expression (6). The homology between the Dlx3 recognition sequences within the profilaggrin promoter and the recognition sequences for the POU domain transcription factors (7) and the well documented ability of POU domain proteins to interact with AP1 to both activate and repress target genes (8-10) prompted us to investigate the role of the POU domain transcription factor in the regulation of profilaggrin expression.

POU domain proteins form a subfamily of the homeodomain proteins with a divergent POU-type homeodomain (POU_HD). Amino-terminal to POU_HD there is a second conserved domain, the POU-specific (POU_S) domain. Whereas the POU_HD domain is related to the classic homeodomain encoded by the homeobox, the POU_S domain is unique to the POU protein family. Outside of the POU domain, there are no structural characteristics common to all POU domain gene family members. The entire POU domain is required for DNA binding (reviewed in Refs. 11-13). The most frequent targets of the POU domain proteins are the "octamer" motifs (ATGCAAAT), which have been shown to be involved in both ubiquitous and cell type-specific regulation of various genes (11, 13). Furthermore, these proteins have the flexibility to bind heterogeneous sequences, such as the so-called TAATGARAT motif. The more degenerated the core sequence is, the more important the flanking bases for the affinity and the specificity of the binding (11). POU domain proteins have been implicated in development; control of replication, growth, and cell cycle arrest; and differentiation (Refs. 7 and 11-13 and references therein). They have been shown to exert both activation and repression. The complexity of POU domain protein functions is further increased by the diversity of their interactions with other transcription factors and cell type-specific co-regulators (8-10, 14-23).

The POU domain proteins reported to date in epidermis are Oct1, Skn1a/i (Epoc Oct11), and Oct6 (24-27), which are abundantly expressed, and Oct2, which is present at very low levels (26). The ubiquitous Oct1 is expressed in both mitotic and postmitotic keratinocytes (26). The expression of both Skn1a/i and Oct6 is restricted to cells of ectodermal origin, and especially for Skn1a/i, it appears to be limited almost exclusively to interfollicular epidermis and cortical cells of the hair. Both proteins have been found predominantly in the differentiating epidermal layers (24, 26, 27) and have been shown to transregulate promoters in normal epidermal keratinocytes (27, 28).

In this paper, we demonstrate that the three major epidermal POU domain proteins, Oct1, Skn1a/i, and Oct6, regulate the transcription of the epidermal late differentiation marker profilaggrin in a distinct manner. Their effect on the endogenous gene activity is faithfully reproduced over the profilaggrin promoter in transiently transfected keratinocyte cultures. Two closely located octamer-like binding sites mediate the POU domain protein function. We provide evidence that all three proteins function in conjunction with c-Jun containing AP1 complexes. Our data demonstrate the complexity of the profilaggrin promoter activity and suggest that its ultimate regulation may require a fine balance between the activities of these epidermally expressed POU domain proteins.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Construction of Profilaggrin Reporter Vectors-- All recombinant DNA work was done by standard procedures (29). The wild type and AP1-mutant profilaggrin constructs were described previously (4). The Oct mutants were generated with a polymerase chain reaction-based commercial site-directed mutagenesis kit from Stratagene (La Jolla, CA), as suggested by the manufacturer. The wild type motif OctA (CATTATCTCAG) was mutated into CACTCACAGAG; OctB (TGGAGAAATAGAG) was mutated into TGGAGAACAAGAG.

RNA Isolation and RNase Protection-- Total RNA was isolated from monolayers of submerged NHEKs by using Trizol reagent (Life Technologies, Inc.). RNase protection experiments were done with total cell lysates from about 10⁵ cells and the commercially available Direct Protect lysate RNase protection assay kit (Ambion, Austin, TX). Profilaggrin, Oct1, Skn1a, and Oct6 templates were created by subcloning polymerase chain reaction-generated fragments of the corresponding cDNAs in the original TA cloning vector (Invitrogen, San Diego, CA). The template for 28 S RNA was purchased from Ambion. Antisense riboprobes were synthesized with T7 RNA polymerase to an average specific activity of 1-5 × 10⁸ dpm/µg. The protection assays for profilaggrin, 28 S RNA, and each of the transcription factors were carried out in a single tube. The protected fragments were resolved on 6% denaturing polyacrylamide gels and detected by radiography after exposure at 70 °C for 2 days for profilaggrin, overnight for 28 S RNA and 5 days for the POU domain proteins. The profilaggrin/CAT templates were synthesized by polymerase chain reaction with profilaggrin and CAT-specific primers. A T7 promoter sequence was included at the 5'-end of the CAT primer to allow for in vitro transcription of the polymerase chain reaction fragments with a MAXIscript kit (Ambion). Similarly, the -galactosidase templates were synthesized using thymidine kinase and -galactosidase-specific primers.

Cell Cultures and Transfections-- Neonatal foreskin NHEKs were grown in serum-free keratinocyte medium, transfected with Lipofectin reagent (Life Technologies, Inc.), and assayed for CAT activity as described previously (4, 30). HeLa cells were purchased from ATCC (Manassas, VA), grown, and transfected as described previously (4). For each set, at least five independent experiments were carried out. The transcription factor expression vectors were generous gifts from the following sources: Oct1, W. Herr (Cold Spring Harbor Laboratory); Skn1a and Skn1i, B. Andersen (San Diego, CA); Oct6, E. Fuchs (Chicago, IL); c-Jun, M. Karin (La Jolla, CA); JunD, L. Lau (Chicago, IL). To enrich for cells transfected with each of the POU domain proteins, the Capture-Tec system (Invitrogen, Carlsbad, CA) was used. NHEKs were co-transfected with the corresponding expression vectors and the pHook-1 plasmid and selected on magnetic beads as instructed by the manufacturer. In a typical experiment, the selected cells made up about 15-20% of the total cell population.

Mobility Shift Assays-- The mobility shift experiments were done with about 10 µg of NHEK nuclear extracts and 5 × 10⁴ cpm of end-labeled, gel-purified, double-stranded oligonucleotides as already published (4). Where indicated, the binding reactions were preincubated with 2 µg of the designated antibodies. With the exception of the antibody against Oct6 (a gift from G. Lemke), all antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). The sequences of the double stranded oligonucleotides were as follows: OctA, 5'-GGGGAAGCTTGCCATTATCTCAGCATGCATGT; OctB, 5'-GGGGAAGCTTGCATGTGGAGAAATAGAGTGCATGCTAG-3'; Ets/HMG/AP1, 5'-TGGTTAGGAATGAATCAGACCATCCCACAG-3'.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Profilaggrin Promoter Region Encompasses Two Functional POU Domain Binding Motifs-- The proximal promoter region of human profilaggrin gene (Fig. 1a) encompasses two AT-rich motifs that are homologous to the consensus recognition sequence TAATGARAT of the POU domain transcription factors (10). The first one, CA*TTATCTCAG (designated OctA), is located between positions -16 and -6. Recently we demonstrated that this motif binds the recombinant homeodomain protein Dlx3 (5). The second, TGGAGA*AATAGAG, was identified between positions -124 and -112 and was designated OctB. The two Oct sequences share nine complementary nucleotides (shown in boldface) with just one mismatch (marked with asterisks).

View larger version (28K): [in this window] [in a new window]   Fig. 1.   Effect of OctA and OctB binding sites on the activity of the profilaggrin proximal promoter region. a, sequence of the profilaggrin gene between position -150 and +49 encompassing the OctA, OctB, and Ets/HMG/AP1 recognition motifs. The arrows mark the transcription initiation sites, the one with the asterisk being used in the presence of Skn1a. b, NHEK (black bars) and HeLa cells (open bars) were transfected with the designated wild type and mutant profilaggrin constructs, and in each cell type, the transient activity of CAT was expressed as a percentage of the activity of the wild type construct -354/+9.

To explore the role of the Oct motifs for the activity of the profilaggrin promoter, we transfected into NHEKs reporter constructs in which the expression of a CAT gene is under the control of either wild type, OctA, or OctB mutant promoter sequences. Invariably, mutation of the AT-rich core of OctA reduced the activity of CAT about 3-fold (Fig. 1b, black bars). Significantly, construct -41/+9, which encompasses only the OctA motif and the TATA box of the profilaggrin gene, did not show any marked CAT activity, suggesting that the OctA site was necessary but not sufficient to maintain the expression. Apparently, the positive interactions at the OctA site were not specific for epidermal keratinocytes, because the OctA mutant constructs showed similar behavior in HeLa cells (Fig. 1b, open bars).

In NHEKs, mutation of OctB within construct -354/+9 caused a 50% increase in activity, compared with the wild type. Gradual deletions of the sequences between -354 and -125 resulted in a 3-fold reduction in the levels of CAT. A further deletion to -116, which destroyed the OctB motif, abolished its inhibitory effect and resulted in activity that was similar to that of construct -354/+9. Mutations in the AT-rich core of OctB within construct -125/+9 restored the activity to the level of construct -116/+9. Together these results indicated that interactions at OctB have a silencing effect on the profilaggrin promoter activity and that elements residing between -125 and -354 may serve to overcome this effect. In contrast to OctA-, the OctB-mediated repression was observed only in NHEKs. In HeLa cells (Fig. 1b, open bars) and in several other cell lines tested (e.g. dermal fibroblasts, neuroblastoma, HepG2, and embryonic carcinoma F9 cells; data not shown), the levels of expression of the wild type and OctB mutant constructs were comparable.

The interactions between Oct sites and keratinocyte nuclear proteins were explored in mobility shift assays. Two strong, slowly migrating and several fast migrating complexes were formed between OctA and NHEK nuclear extracts (Fig. 2a, lane 2). Their specificity was ascertained in competition experiments with an excess of unlabeled OctA and other oligomers. Invariably, excess of OctA prevented the detection of the slowly migrating complexes A1 and A2 and of the faster complex A3 (Fig. 2a, lane 3). In contrast, incubation with a number of irrelevant oligonucleotides (such as AP1, AP2, and EF2) or with dI/dC did not interfere with the formation of complexes A1-A3 (data not shown). We concluded that complexes A1, A2 and A3 represented specific interactions between oligonucleotide OctA and nuclear proteins. The nature of the other complexes could not be unambiguously determined, and therefore, they were not further investigated. To elucidate the protein content of complexes A1, A2, and A3, the bandshift reactions were carried out in the presence of antibodies against the epidermal POU domain proteins. Preincubation with the antibody against Oct1 prevented the formation of complex A1 (Fig. 2a, lane 4). The formation of complexes A2 and A3 was perturbed in the presence of the antibody against Skn1a/i (Fig. 2a, lane 5) and to a lesser extent by the antibody against Oct6 (Fig. 2a, lane 7). Similar co-migration of Skn1a/i and Oct6 complexes formed in epidermal nuclear extracts has been reported previously by Andersen et al. (26). None of the complexes was affected by antibodies against Oct2 (Fig. 2a, lane 6) or by an antibody against the transcription factor NF1, which also recognizes AT-rich DNA elements (Fig. 2a, lane 8). Because of the ability of recombinant Dlx3 to bind the OctA motif (6), we were surprised to observe that none of the OctA complexes was sensitive to an antibody against Dlx3 (data not shown).

View larger version (55K): [in this window] [in a new window]   Fig. 2.   a, binding profiles of oligonucleotide OctA (lane 1) with NHEK nuclear proteins (lane 2). The specific complexes A1, A2, and A3 were successfully competed by a 50-fold molar excess of unlabeled oligonucleotide OctA (lane 3). The formation of complex A1 was perturbed by preincubation of the extracts with the antibody against Oct1 (lane 4). Both Skn1a/i and Oct6 antibodies interfered with the formation of complexes A2 and A3 (lanes 5 and 7, respectively). None of the complexes was affected by preincubation with an antibody against Oct2 (lane 6) or against the unrelated transcription factor NF1 (lane 8). b, binding profile of NHEK nuclear proteins with oligonucleotide OctB (lane 1) and in the presence of antibodies against Oct1 (lane 2), Oct2 (lane 3), Skn1a/i (lane 4), Oct6 (lane 5), and NF1 (lane 6). The material marked with asterisks was resistant to competition with unlabeled binding probes but was sensitive to preincubations with a number of irrelevant oligonucleotides. Most probably, it contained nonspecific DNA-protein complexes that were not investigated. In both panels P marks the migration of the free oligonucleotide probe

In bandshift experiments, OctB formed Oct1-sensitive (B1 and B2), Skn1a/i-sensitive (B3), and Oct6-sensitive (B3) complexes (Fig. 2b, lanes 2, 4, and 5, respectively). Similar to OctA, the OctB complexes did not include Oct2 (Fig. 2b, lane 3) and were not sensitive to antibodies against non-POU domain proteins, such as Dlx3 (data not shown) and NF1 (Fig. 2b, lane 6).

Distinct Effects of Homeodomain Proteins on the Activity of the Profilaggrin Promoter-- We explored the role of the interactions between the Oct motifs and the POU domain proteins by assessing the activity of CAT obtained from constructs carrying wild type or mutant Oct sequences in response to forced expression of Oct1, Skn1a/i, and Oct6.

Oct1-- The ubiquitously expressed Oct1 transactivated profilaggrin construct -354/+9 at least 2-fold. This effect was exerted mainly through the OctA site, because mutation of the OctB site did not alter the degree of transactivation (Fig. 3a). We were not able to detect differentiation or cell type specificity in the effect of Oct1 (data not shown).

View larger version (40K): [in this window] [in a new window]   Fig. 3.   Effect of epidermal POU domain proteins on the activity of profilaggrin promoter region. The designated profilaggrin constructs were co-transfected with vectors expressing Oct1 (a), Skn1a and Skn1i (b), and Oct6 (c) or with the corresponding empty expression vectors. In each set of experiments, the black bars represent the activity of the corresponding CAT constructs after co-transfection with the respective empty expression vectors into NHEKs and are presented as a percentage of the activity of construct -354/+9. The light striped bars show the activity in NHEKs grown in 1.2 mM Ca2+ in the presence of Oct1 (a), Skn1a (b), or Oct6 (c). The shaded bars show the activity in the presence of Skn1i (b) or C (c). The dark striped bars show the activity in NHEKs grown at 0.05 mM Ca2+ in the presence of Skn1a (b) or Oct6 (c). In b, the open bar represents the activity in HeLa cells of construct -354/+9 in the presence of Skn1a. The activity of the same construct in the presence of the empty cytomegalovirus expression vector is set at 100%. In c, the open bars represent the activity in HeLa cells of the reporter constructs co-transfected with the corresponding empty expression vectors.

Skn1a/i-- Co-transfection of construct -354/+9 with Skn1a expression vector down-regulated the activity of the reporter (Fig. 3b). The effect was observed both in HeLa cells (Fig. 3b, open bar) and in NHEKs, in which it was most pronounced (5-fold) at elevated calcium concentrations (compare dark and light striped bars). Under these conditions, the "inhibitory" splice variant Skn1i (23) was one-half as efficient but still repressed the activity by more than 2-fold (compare the light striped and shaded bars). To a large extent, the effect of Skn1a/i was mediated through the OctA motif, because its mutation reduced the repression to only 40%, whereas mutation of OctB still resulted in a 5-fold inhibition. However, mutation in the OctA motif was not able to completely prevent the repression. A similar effect of Skn1a was recorded, with constructs extending as little as 116 base pairs upstream of the transcription initiation site (e.g. -116/+9) and which did not encompass other octamer-like sequences. Notably, Skn1i, which binds DNA very poorly (25), was still able to repress transcription, to a level comparable to that obtained with Skn1a when binding to OctA was perturbed. These results suggested that the repression was exerted both through interactions of Skn1a with DNA over the OctA motif and with other transcriptional regulators. The latter did not require direct binding of Skn1a/i to DNA and were probably mediated through interfaces common for both splice variants. The more pronounced effect at higher calcium concentrations was also an indication that the Skn1a/i interactions were modulated by the differentiation state of the keratinocytes.

Oct6-- In NHEKs, forced expression of Oct6 down-regulated the activity of profilaggrin construct -354/+9 by 30-50%, the effect being more pronounced in cultures induced to differentiate by calcium (Fig. 3c, compare dark and light striped bars). The repression was specific for epidermal keratinocytes, because when co-transfected in HeLa cells, Oct6 increased the activity by more than 60%. These effects likely required interactions of the transcription factor with both DNA and other proteins, because a mutant variant of Oct6 with a truncated POU domain (C, Ref. 26) was not able to transactivate in HeLa, but in NHEKs acted as a dominant negative mutant, reducing the activity below the level elicited by the empty vector (Fig. 3c, shaded bars). The mutation of the OctA site in the frame of construct -354/+9 reduced Oct6-mediated activation in HeLa cells but could not interfere with Oct6 repression in NHEKs. In contrast, mutations in the OctB motif did not perturb transactivation in HeLa cells and in NHEKs caused activation by more than 50%. The profiles with the OctB mutants were very similar to the profile of construct -116/+9, which did not encompass the OctB motif. In this case Oct6 activated profilaggrin promoter in NHEKs through binding at the OctA motif, in a way similar to that in HeLa cells. Together, these data indicated that Oct6 could act, via site OctA, as a general positive regulator of the profilaggrin promoter. However, in the keratinocytes, this positive effect was overcome by repression through cell type-specific interactions over site OctB.

The specificity of the interactions between the Oct sites and the POU domain proteins was further substantiated by the response of the profilaggrin reporters to forced expression of the non -POU homeodomain protein Dlx3. Co-transfection of constructs -354/+9, -125/+9, and -116/+9 with a Dlx3 expression vector invariably resulted in about 6-fold increase in the reporter activity.² Consistent with the lack of Dlx3 in the complexes of the Oct sites with the keratinocyte nuclear extract, this activation could not be perturbed by mutations in the Oct motifs, despite the fact that they were the only sequences within the promoter region that were able to bind recombinant Dlx3 (6). Thus, the activation of the profilaggrin promoter by Dlx3 was not mediated by the Oct sites. Rather, it might result from direct protein-protein interactions of Dlx3 with the basal transcription machinery, as previously reported for other homeodomain proteins (31), and/or with other transcription activators of the profilaggrin promoter.

Effect of Homeodomain Proteins on the Endogenous Profilaggrin Expression-- We wanted to ascertain whether the observed effects of POU domain proteins on the profilaggrin promoter faithfully reproduced their effect on the transcription of the endogenous profilaggrin gene. By using the Capture-Tec system (Invitrogen), we were able to select NHEKs that were transfected with vectors expressing the POU domain proteins and compare the mRNA levels of profilaggrin and the respective transcription factors to those present in parallel cultures co-transfected with the empty expression vectors. As judged by RNase protection (Fig. 4), the cells transfected with Oct1 contained more profilaggrin mRNA than the control, whereas transfection with Skn1a or Oct6 resulted in fewer profilaggrin transcripts. In contrast, forced expression of the POU domain proteins did not alter the levels of 28 S RNA, which was used as an internal control. Thus, in NHEKs, the behavior of the profilaggrin promoter constructs in response to the POU domain proteins reflected the effect of these factors on the endogenous profilaggrin expression.

View larger version (76K): [in this window] [in a new window]   Fig. 4.   Effect of the POU domain proteins on the endogenous profilaggrin expression in NHEKs. NHEKs were transfected with the designated empty (-) or loaded (+) expression vectors and with the plasmid pHook-1, which expresses a single chain antibody against a specific hapten. The co-transfected cells (between 15 and 20% of the entire population), displaying the antibody on their surface, were selected by binding to hapten-coated magnetic beads and lysed, and the cell lysates were used directly to assess the content of profilaggrin (PF), 28 S RNA (28s), and the designated POU domain protein transcripts (TF) in RNase protection assays. The signals for profilaggrin, 28 S RNA, and the POU domain proteins were obtained after different times of exposure, as indicated under "Materials and Methods." However, the exposure times for each set were identical, thus allowing direct comparison between the controls and the test samples.

Interactions with the POU Domain Proteins Have a Distinct Effect on the AP1-dependent Activity of the Profilaggrin Promoter-- The complex effects of the POU domain proteins reported above indicated interactions with other transcriptional regulators. Our previous studies have established that the keratinocyte-specific activity of the profilaggrin promoter depends critically on binding of c-Jun/c-Fos heterodimers over an AP1 site at position -77 (4). Interactions between AP1 and POU domain factors have been documented previously to affect both the level and the specificity of expression (8-10). Accordingly, we examined the interdependence of profilaggrin AP1 and POU domain protein interactions by transfecting wild type and mutant constructs -354/+9, together with expression vectors for c-Jun or JunD. Consistent with our previous data, in NHEKs c-Jun activated the promoter about 4-fold through binding at the AP1 site at position -77, whereas the 2-fold activation by JunD was AP1 site-independent (Fig. 5a). Mutation of the OctB site did not interfere with the extent of either c-Jun or JunD transactivation. When the OctA site was mutated, however, in NHEKs, an 8-fold transactivation by c-Jun was obtained, whereas the extent of JunD transactivation remained unaltered. Significantly, in HeLa cells, c-Jun transactivated the wild type and OctA mutant constructs equally, thus ruling out the possibility that a cryptic AP1 site might have been created by the mutation.

View larger version (3246K): [in this window] [in a new window]   Fig. 5.   Effect of POU domain protein interactions on the AP1-dependent activity of the profilaggrin promoter. a, effect of Oct mutations on c-Jun and JunD activation. The designated wild type and mutant profilaggrin constructs were transfected into NHEKs (black bars) or HeLa cells (open bars) together with empty expression vectors for c-Jun or JunD. The dark striped bars show the activity after co-transfection with c-Jun, and the widely striped bars show the activity after co-transfection with JunD. The CAT values are presented relative to the levels obtained with construct -354/+9 co-transfected with the corresponding empty vectors. b, NHEK cultures were co-transfected with profilaggrin reporter constructs and expression vectors for either c-Jun (dark striped bars), the indicated POU domain protein (light striped bars), or c-Jun and the POU domain protein (shaded bars) or with the respective empty expression vectors (black bars). The CAT values are related to the activity of construct -354/+9 in the presence of the empty expression vectors: panel A, co-transfections with c-Jun and Oct1; panel B, co-transfections with c-Jun and Skn1a; panel D, co-transfection with c-Jun and Oct6. To avoid repetition, the results with the constructs co-transfected with c-Jun alone are shown only in panel A. Panel C illustrates the usage of the cryptic transcription initiation site in response to forced expression of Skn1a. Wild type and OctA mutant constructs -354/+9 were co-transfected into NHEKs with loaded and unloaded Skn1a expression vectors, and the amount and the length of the profilaggrin/CAT transcripts were assayed by RNase protection. Lane 1 represents the mobility of labeled 1-kb DNA ladder (Life Technologies, Inc.); lanes 2 and 3 show the protected profilaggrin/CAT sequences in the absence and in the presence, respectively, of excess Skn1a; lane 4 shows the effect of the OctA mutation, which prevented the usage of the cryptic site even in the presence of Skn1a. In each case, a thymidine kinase/-galactosidase-expressing plasmid was included in the transfection mixtures to monitor the efficiency of the transfection and the amount of the input RNA. Note that a 5-fold greater amount of cell lysate was analyzed in lane 4. Panel E shows the binding profile of the Ets/HMG/AP1 oligonucleotide with vector-programmed (lane 1) and Oct6-programmed (lanes 2-4) nuclear extracts; in the presence of antibodies against Oct6 (lane 3) and c-Jun (lane 4). The arrowhead indicates the AP1 complex (4), and the asterisk marks the position of the new, faster complex.

Oct1 and c-Jun acted over profilaggrin promoter cooperatively. In NHEKs, the level of transactivation conferred by each of these transcription factors alone was about 2-3-fold lower than the level obtained when c-Jun and Oct1 were present simultaneously (Fig. 5b, panel A). Activation by either factor was abolished by mutation of the AP1 site. Consistent with the results in Figs. 3a and 5a, mutation in the OctA site prevented Oct1-dependent activation and resulted in c-Jun transactivation that could not be further enhanced by the presence of Oct1. As expected, mutation of the OctB site did not alter the c-Jun/Oct1 effect.

c-Jun transactivation was completely blocked by the presence of Skn1a (Fig. 5b, panel B, shaded bars). As expected, the effect of c-Jun on the OctB mutant construct was inhibited to an equal extent. Mutation of the OctA site, however, partially released the inhibitory effect of Skn1a over c-Jun transactivation and resulted in CAT activity that was 3-fold higher than the activity of the same construct in the presence of Skn1a alone. The Skn1a repression was observed with the AP1 mutant construct as well, suggesting that the inhibitory effect of Skn1a was mediated not solely through its interactions with c-Jun. In fact, Skn1a repressed the activity of a series of mutant constructs spanning the entire region between -116 and +9 (data not shown). These findings and the close proximity of OctA to the profilaggrin transcription initiation site suggested that the inhibition might result from interference with the basal transcription machinery. This notion was supported by the observations presented in Fig. 5b, panel C. In NHEKs co-transfected with construct -354/+9 and the empty cytomegalovirus expression vector (Fig. 5b, panel C, lane 2), the transcription of the profilaggrin/CAT gene was initiated at the legitimate profilaggrin start site (Fig. 1a and Ref. 32). Co-transfection of the same construct with the Skn1a expression vector, however, resulted in a marked decrease in the amount of the chimeric transcripts, and moreover, they were initiated at a cryptic start site at position +7 (Fig. 5b, panel C, lane 3). The sequence CAGCAGG (Fig. 1a) has 6 out of 7 nucleotides homologous to the initiation consensus I (33). Significantly, mutation of OctA still resulted in lower levels of transcripts but prevented the aberrant initiation in the presence of Skn1a (Fig. 5b, panel C, lane 4).

Oct6 did not interfere with the effect of c-Jun on the wild type promoter (Fig. 5b, panel C). The relative level of activation in the presence of c-Jun and Oct6 simultaneously was comparable to the transactivation exerted by c-Jun alone (shown in Fig. 5b, panel A, dark striped bars). Mutation of either Oct site did not alter this profile. On the other hand, mutation of the AP1 site invariably abolished the effect of Oct6. The ability of excess of c-Jun to counteract the repression by Oct6 suggested that the latter might be mediated through altering the c-Jun interactions over the promoter. To explore this possibility, we prepared RNA and nuclear extracts from NHEKs transfected with Oct6. RNase protection and Western blots analysis using c-Jun- and c-Fos-specific probes and antibodies did not reveal marked changes in the levels of the corresponding mRNAs and proteins compared with vector-transfected cells (data not shown). However, when combined with an oligonucleotide that encompassed the Ets/HMG and the AP1 sites of the promoter region (Fig. 1a), the Oct6-programmed extracts were barely able to form AP1-containing complexes (Ref. 4 and Fig. 5b, panel E). Instead, there was an increase in the intensity of a faster-migrating complex (Fig. 5b, panel E, compare lanes 1 and 2). Preincubation of the binding reactions with the Oct6-specific antibody partially restored the AP1 complex and reduced the amount of the faster complex (Fig. 5b, panel E, lane 3), whereas an antibody against c-Jun had no effect on the binding profile (lane 4). The fact that the faster complex was sensitive to preincubation with the anti-Oct6 antibody and preliminary bandshift experiments with oligonucleotides carrying wild type or mutant AP1 and Ets/HMG motifs³ suggest that the faster complexes contain Oct6 bound at the Ets/HMG portion of the oligonucleotide. Thereby, the reduced AP1 binding may be a direct consequence of Oct6 binding to the Ets/HMG recognition sequences immediately adjacent to the AP1 site, and/or it may reflect displacement of a factor(s) essential for the stability of the AP1 complexes. Earlier data from our laboratory have documented that c-Jun and Ets transcription factors, such as Ets1 and Jen, bind at their adjacent recognition motifs and cooperate in activation of the profilaggrin promoter (5). Physical interactions between Ets or HMG-like factors and different POU domain proteins, through both DNA-protein and protein-protein binding, have been reported to influence the function of these proteins in both positive and negative fashion (15-18, 22, 23). Therefore, it may be speculated that by compromising the cooperation between Ets or HMG and AP1 factors bound at the composite Ets/HMG/AP1 element, Oct6 may markedly reduce the promoter activity.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The assembly of transcriptionally competent chromatin structure over the epidermal genes, as in other cell systems, requires a fine balance in the amounts and the functional availability of multiple transcription factors, the effects of which have to be further tuned by the dynamic recruitment of so far unidentified co-activators and repressors. Recently it has become evident that transcription factors belonging to the family of the POU domain proteins may be involved in this process as both positive and negative regulators (24-28, 34).

In this work, we have investigated the role of the major epidermal POU domain proteins, Oct1, Skn1a/i, and Oct6, in the transcription control of the human profilaggrin gene. Our results demonstrate that in NHEKs Oct1 acts as a positive regulator, whereas Skn1a/i and Oct6 down-regulate profilaggrin expression. Significantly, the effect of these factors on the endogenous profilaggrin transcription was reproduced over the profilaggrin promoter constructs in the transient transfection experiments, allowing parallels between endogenous profilaggrin expression and promoter construct activity in response to the POU domain proteins.

Mutation and in vitro binding analyses enabled us to identify two homologous DNA sequences, OctA and OctB, which mediate the interactions of the POU domain proteins with the profilaggrin promoter. Despite the high degree of sequence homology, OctA and OctB exhibit entirely different activities, and in the context of the promoter region, they are involved in distinct interactions. Constructs in which OctA has been destroyed exhibit only one-third of the activity of their wild type counterparts in both epidermal keratinocyte and HeLa cells, suggesting that OctA is required for full basal activity of the promoter. The increased activity of OctB mutants in NHEKs but not in other cell types, on the other hand, indicates that OctB is involved in keratinocyte-specific silencing of the promoter. Strong specificity in the activities of seemingly closely related octamer recognition motifs has been reported previously and appears to be an important characteristic of the interactions between POU domain proteins and their cognate binding sites (for reviews, see Refs. 11-13). Although all POU domain proteins harbor the conserved homeodomain and bind similar AT-rich DNA sequences, the unique POU-specific domains play a critical role in sequence-specific DNA binding (see Ref. 11 and references therein). Binding of one and the same POU domain protein at slightly different sequences has been reported to change its conformation over the DNA and allow interactions with different co-regulators (35-38). Finally, the specificity of a particular octamer binding site can be further modulated by interactions with other transcription factors bound at adjacent recognition motifs (e.g. see Refs. 8-10, 14-18, 22, 23). Our data suggest that each of these mechanisms may be utilized in the control of the profilaggrin promoter. The comparison between the bandshift profiles of OctA and OctB reveals that in NHEKs, OctA is occupied preferentially by Oct1 and Skn1a/i, bound mostly as dimers, whereas OctB binds preferentially to monomeric Oct6. Consistent with these binding profiles, site OctA mediates predominantly the effects of Oct1 and Skn1a/i, whereas site OctB is involved predominantly in interactions with Oct6.

The control each of these POU domain proteins exerts over the profilaggrin promoter is interconnected with the AP1 site-dependent activation, albeit in a distinct manner. Oct1 activates the promoter in cooperation with c-Jun. Binding of the two factors at their respective recognition sites is essential for this cooperation. Skn1a completely blocks c-Jun activation, most probably by preventing binding of Oct1 at site OctA and the initiation of transcription at the legitimate start site. This block is mediated to a large extent through binding of Skn1a at OctA and cannot be overcome by excess of c-Jun. Oct6 down-regulates the promoter by interfering with the binding of c-Jun at the AP1 site and compromising the Ets/AP1 cooperation. In this case, an excess of c-Jun can compensate for the negative effect of Oct6.

The fact that the inhibitory splice variant Skn1i and the POU domain-truncated variant of Oct6 C are both capable to markedly reduce the activity of the promoter indicates that the POU domain proteins also operate via protein-protein interactions in a DNA binding-independent manner. Down-regulation of epidermal gene activity by Oct1, Skn1a, and Oct6 via the basal transcription machinery has been previously proposed for involucrin (28). Strong repression by Oct6 and to a lesser extent by Oct1, seemingly not dependent on DNA binding, have also been observed with keratin K5 and K14 promoters (27). These findings are consistent with the properties of the POU domain proteins extensively studied in other cell systems (39-43) and indicate that interactions of these proteins outside of the POU domain can be utilized in a cell type-specific fashion and are essential for the transcription activity in the epidermal keratinocytes.

In view of our data, several speculations can be put forward to explain the apparent lack of effect of the Skn1a/1 and Oct6 knockouts on the profilaggrin expression in transgenic mice (44). It is possible that in the knockout animals, or in the mouse system in general, the repressive functions of Skn1a/i and Oct6 are taken over by another POU domain protein. In this respect, it is noteworthy that targeted mutations in the Caenorhabditis elegans POU domain gene ceh-18, the vertebrate counterpart of which has not yet been identified, seriously perturb epidermal differentiation (45). Alternatively, we may speculate that knocking out Skn1a/i and Oct6 may also repress the expression of yet another transcription factor that is required for activation of profilaggrin transcription in spinous cells, an event that would neutralize the knockout of the repressors. In any event, taking into account the profound effect the ectopic expression of profilaggrin provokes (46), it may be assumed that the correct temporal and spatial expression of this gene is assured by multiple, independent and/or interdependent mechanisms.

To date, the role of three families of transcription factors in the control of profilaggrin promoter activity has been demonstrated, namely AP1 (4), Ets (5), and POU domain (this paper). The combinatorial interactions between AP1, Ets, and POU domain proteins provide an elegant example of how a limited number of DNA elements can mediate a complex temporal and spatial regulation. The proximal promoter region of the profilaggrin gene encompasses closely situated recognition motifs for these transcription factors and binding of particular members of each family to their respective sites can coordinately regulate the activity of the promoter. Similar mechanisms have been recognized as operational over other epidermal promoters, such as involucrin (28, 47), keratin K5 (48, 49), and SPRRs (34, 50, 51) (reviewed in Refs. 52 and 53). The AP1, Ets, and POU domain families comprise a great number of proteins. Although some of them are ubiquitously expressed throughout the epidermis, many others are preferentially found in particular differentiation layers (5, 26, 27, 54, 55). Thus, replacement even of a single transcription factor over any one of the binding sites while the cells undergo differentiation may swiftly affect the transcription of the target genes by altering the associated co-repressors or co-activators. Moreover, such a mechanism can explain why Skn1a, for example, probably activates the SPRR2A (34) but at the same time strongly represses involucrin (28) and profilaggrin (this paper) promoters, all of which have similar differentiation specificity.

The mRNA levels of the late differentiation markers in the epidermis respond to extracellular signals such as elevated calcium concentrations and phorbol esters (see Refs. 52 and 53 and the references therein). It has been proposed (5, 28, 50, 51) and in the case of involucrin proven (56, 57) that the activation takes place at the level of transcription and is mediated through mitogen-activated protein kinase pathways that converge over AP1 and/or Ets interactions over the promoter regions. AP1 and Ets transcription factors have been recognized as major targets for these signal transduction pathways in a variety of cell systems (Refs. 58 and 59 and the references therein). Thus, the ability of the POU domain proteins to interact with AP1 and Ets transcription complexes, on one hand, and to closely associate with components of the basal transcription machinery within different promoter regions, on the other, provides a mechanism through which the signals on the surface of the keratinocytes can be amplified and transmitted to the transcription initiation sites of several epidermal genes in a fast and coordinated manner.

In conclusion, we propose a model (Fig. 6) whereby in the basal keratinocytes in the absence of c-Jun profilaggrin basal promoter activity is suppressed through interactions of JunB (4), basal cell-specific Ets proteins such as Elk1 (5) and/or other unidentified factors (Fig. 6a). In the spinous layers, increasing amounts of Oct6 prevent the assembly of Ets/AP1 complexes at their composite binding site, whereas association of Skn1a and specific co-repressors with the basal transcription machinery interferes with the initiation of transcription at the legitimate start site (Fig. 6b). In the granular layer, following the synthesis of high amounts of a variety of signal-dependent transcription factors, such as c-Jun, c-Fos, Ets1, and Jen (33, 51, 52), as well as the appearance of specific co-activators, the cooperative interaction between these factors leads to displacement of the repressors, assembly of transcriptionally competent complex over the initiation site, and high levels of profilaggrin expression (Fig. 6c). Finally, a dynamic interchange between Ets/AP1 and Oct6 complexes over the composite Ets/AP1 site could provide a switch-on and switch-off mechanism for optimal levels of profilaggrin expression in the granular layer (Fig. 6d).

View larger version (24K): [in this window] [in a new window]   Fig. 6.   A model of the interactions of the basal transcription machinery (BTM) with the AP1, Ets, and POU domain proteins over the profilaggrin promoter in the basal (a), spinous (b) and granular (c and d) layers of the epidermis. Shown are the TATA box and the Oct and Ets/HMG/AP1 recognition motifs. X, Y, Co-act, and Co-rep represent putative co-regulators mediating protein-protein interactions.

	ACKNOWLEDGEMENTS

We are grateful to W. Herr, B. Andersen, E. Fuchs, M. Karin, and L. Lau for the expression vectors and to G. Lemke for the Oct6 antibody. We thank M. Simon and C. Chipev for discussions and support.

	FOOTNOTES

^* The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^¶ To whom correspondence should be addressed: The Living Skin Bank, Dept. of Oral Biology and Pathology, School of Dental Medicine, SUNY-Stony Brook, Westchester Hall, Rm. 1112, Stony Brook, NY 11790-8702. Tel.: 516-632-7420; Fax: 516-632-9707; E-mail: nmarkova@epo.som.sunysb.edu.

² S.-I. Jang, M. I. Morasso, and N. G. Markova, unpublished results.

³ S.-I. Jang and N. G. Markova, unpublished results.

	ABBREVIATIONS

The abbreviations used are: AP1, activator protein 1; NHEK, normal human epidermal keratinocyte; Oct, octamer; CAT, chloramphenicol acetyltransferase.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

K.-Y. Kam, K.-H. Jeong, E. R. Norwitz, E. M. Jorge