The Novel Murine Ca2+-binding Protein, Scarf, Is Differentially Expressed during Epidermal Differentiation*

Calcium (Ca2+) signaling-dependent systems, such as the epidermal differentiation process, must effectively respond to variations in Ca2+ concentration. Members of the Ca2+-binding proteins play a central function in the transduction of Ca2+ signals, exerting their roles through a Ca2+-dependent interaction with their target proteins, spatially and temporally. By performing a suppression subtractive hybridization screen we identified a novel mouse gene, Scarf (skin calmodulin-related factor), which has homology to calmodulin (CaM)-like Ca2+-binding protein genes and is exclusively expressed in differentiating keratinocytes in the epidermis. The Scarf open reading frame encodes a 148-amino acid protein that contains four conserved EF-hand motifs (predicted to be Ca2+-binding domains) and has homology to mouse CaM, human CaM-like protein, hClp, and human CaM-like skin protein, hClsp. The functionality of Scarf EF-hand domains was assayed with a radioactive Ca2+-binding method. By Southern blot and computational genome sequence analysis, a highly related gene, Scarf2, was found 15 kb downstream of Scarf on mouse chromosome 13. The functional Scarf Ca2+-binding domains suggest a role in the regulation of epidermal differentiation through the control of Ca2+-mediated signaling.

tiation-specific markers (2). The differentiation process can be recapitulated partially in mouse keratinocytes cultivated in vitro by increasing the Ca 2ϩ concentration in the culture medium (3) and is associated with the activation of protein kinase C (PKC) (4). This produces a situation that mimics the endogenous Ca 2ϩ gradient present in the skin, with low extracellular levels surrounding the basal cells and increasing levels toward the upper granular layers (5). The ability to induce differentiation in primary cultured keratinocytes, along with the presence of a Ca 2ϩ gradient, underscores the importance of this ion and the Ca 2ϩ -dependent signaling pathways in the epidermis.
A crucial role for the transduction of the Ca 2ϩ signal is accomplished by members of the Ca 2ϩ -binding proteins, which are characterized by the presence of a common helix-loop-helix structural motif in their Ca 2ϩ -binding domain, termed EFhand (6). These proteins function by undergoing conformational changes upon binding of Ca 2ϩ , allowing the association and regulation of activity of a range of specific target proteins. EF-hand-containing proteins have been described in epidermis; these include the large proteins profilaggrin and repetin (7,8), which present EF-hand motifs in the NH 2 -terminal region followed by multiple tandem repeats and members of the dimeric EF-hand S100 multigenic family (9 -11). The best studied of the Ca 2ϩ -signaling proteins is CaM, a small, highly conserved, and ubiquitous protein that is crucial to many Ca 2ϩ -dependent processes in eukaryotes (6,12). The functions of many proteins involved in cell signaling by phosphorylation/dephosphorylation or in the modulation of intracellular levels of second messengers are reportedly CaM-dependent (12). Recently, several Ca 2ϩ -binding proteins with structural homology to CaM have been reported: calcium-binding proteins (CaBPs) (13,14), hClp (15), and the human skin-specific hClsp (16,17).
We present the cloning, genomic structure, expression, and functional assay for a novel mouse Ca 2ϩ -binding protein that we have termed Scarf. The Scarf open reading frame (ORF) codes for a small protein with four EF-hand Ca 2ϩ -binding domains. The Scarf gene and protein are differentially expressed in vivo in the spinous and granular layers of the epidermis. We also present the identification of a second highly homologous gene, Scarf2, which localizes 15 kb downstream from Scarf in the mouse chromosome 13. We propose that Scarf and Scarf2 are new members of the CaM-like proteins with potential roles in the Ca 2ϩ -dependent epidermal differentiation process.

EXPERIMENTAL PROCEDURES
Suppression Subtractive Hybridization (SSH)-An SSH screen was performed following the instructions of the manufacturer (Clontech) using basal cell RNA as a "driver" and suprabasal cell RNA as a "tester" (18,19). The primary mouse basal and suprabasal keratinocytes were obtained from neonatal skins that were trypsinized overnight at 4°C and then separated by a discontinuous Percoll gradient (20). From this screen, we identified a partial cDNA sequence (231 bp), which corresponded to a partial coding region of a novel mouse gene we called Scarf * 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 (113-344 bp). The upstream sequence was obtained by performing rapid amplification of cDNA ends (5Ј-RACE, Clontech) utilizing a genespecific oligonucleotide (5Ј-tgaacacagcccgcagctcccctgctctgtgccc-3Ј). The complete mRNA sequence was deposited in GenBank TM (accession no. AY293058).
Cloning and Genomic Sequences Analysis of Scarf and Scarf2-A genomic DNA region of ϳ120 kb was cloned through screening of a mouse VJ/129 BAC library (Genome Systems Inc.) using an 0.9-kb Scarf cDNA as a probe. Comparison of the genomic and cDNA sequences determined the 5Ј-and 3Ј-end boundaries, the absence of intronic regions in the Scarf gene, and the identification of a Scarf-related gene that we termed Scarf2 (GenBank TM accession no. AY314008). For sequence homology searches we used a BLAST analysis (basic local alignment search tool), made available through the National Center for Biotechnology Information (www.ncbi.nlm.nih.gov/BLAST).
Cell Culture-Primary mouse keratinocytes were isolated from trypsinized newborn BALB/c mouse skins and cultured as reported previously (19). Keratinocytes were cultured in medium with 0.05 mM Ca 2ϩ to maintain a basal-like population of undifferentiated cells (3). Cells were treated with different inhibitors at 10 M concentration: a PKC inhibitor, GF109203X; a PKA inhibitor, H89; and a CaM kinase II inhibitor, KN62. Differentiation was achieved by increasing the Ca 2ϩ concentration in the medium to 0.12 mM.
Northern Blots and Gene Expression-Total RNA was isolated from basal and suprabasal keratinocytes and mouse primary keratinocytes differentiated in vitro using TRIZOL (Invitrogen). The RNA samples (1-3 g) were electrophoresed in 1.2% agarose-methymercuryhydroxide gels, electroblotted to nylon membranes, and hybridized according to Church and Gilbert (21). An mRNA dot blot and Northern blot for mouse tissues were purchased from Clontech and used according to the instructions of the manufacturer. The mouse-specific probes used were: the 3Ј-untranslated regions for loricrin (LOR), keratin 5 (K5), and keratin 1 (K1) (kindly provided by Dr. Yuspa and Dr. Compton). For the early differentiation marker involucrin (INV), the expressed sequence tag clone AA798100 was utilized. The coding region of CaM used as probe was obtained by PCR with the following gene-specific oligonucleotides: Forward, 5Ј-tccgttcttccttcttcgctcgcaccatggc-3Ј, and Reverse, 5Ј-gtgtgtggacagaggggcttctgacatcag-3Ј. The Scarf coding region was obtained by PCR using: Forward, 5Ј-gccggatccagatgtctcacgggtttactaaggag-3Ј, and Reverse, 5Ј-gccggatccttttaaatgtggaggcgcacaaactc-3Ј. A glyceraldehyde-3-phosphate dehydrogenase cDNA probe was used to control for RNA integrity (22).
For the RT-PCR reactions, cDNA was synthesized by Superscript II kit (Invitrogen) using 0.2 g of mRNA isolated from the total RNA of suprabasal keratinocytes with an Oligotex kit (Qiagen). The oligonucleotides used were: Forward Scarf2, 5Ј-tactaaggaggagaataaggatggcc-3Ј, and Reverse Scarf2, 5Ј-gtcagtcagatatgtggaccttgtat-3Ј. As a negative control, an RT-PCR reaction was performed with no cDNA input. PCR reactions were 40 cycles of denaturation at 94°C for 30 s, annealing at 60°C for 30 s, and extension at 72°C for 1 min.
Radioactive in Situ Hybridization-RNA probes corresponding to the sense and antisense strands of mouse Scarf partial cDNA (202-447 bp) were prepared using T7 and Sp6 RNA polymerase. In situ hybridization was performed at high stringency on consecutive sagittal sections of 8-, 9-, 15-, and 16-day mouse embryos and neonatal skin as described by Mackem and Mahon (23).
Protein Expression and Purification-Scarf coding sequence was obtained by PCR and subcloned into the BamHI site of the pET-28c(ϩ) vector (Novagen). The recombinant protein was expressed in BL21(DE3) and contained a His tag and a T7 tag at the NH 2 terminus of the protein. The soluble recombinant protein was purified over a nickel-nitrilotriacetic acid agarose column (Ni-NTA, Qiagen). We analyzed the predicted protein sequence using the Swiss Institute of Experimental Cancer Research tools (www.isrec.isb-sib.ch/), and Mac-Vector version 6.5.3. Mutations in the EF-hand motifs to generate the pET-28c/Scarf mutants were performed using the QuikChange sitedirected mutagenesis kit (Stratagene). The conserved first Asp residue of each EF-hand motif (sequence GAC) was replaced with a Gly residue (GCC sequence). Mutant recombinant proteins were purified on nickelnitrilotriacetic acid spin columns (Qiagen). mtEF1-4 represent the mutated forms of Scarf in each EF-hand 1-4 motif, respectively. Anti-Scarf antibody against the full-length protein was generated in hens (Aves Laboratories).
Recombinant hClsp was obtained by subcloning a PCR fragment containing the ORF into the BamHI site of pET-28a(ϩ) vector (Novagen). The oligonucleotides used were: Forward, 5Ј-gccggatccatggccggtgagctgact-3Ј, and Reverse, 5Ј-cggggatcctcactcctgggcgagcat-3Ј. PCR reactions were performed for 30 cycles of denaturation at 94°C for 30 s, annealing at 58°C for 30 s, and extension at 72°C for 1 min. Purification of the recombinant hClsp was performed as mentioned above. hClp recombinant protein was kindly provided by Dr. Strehler.
Immunohistochemistry and Western Blot Analysis-Immunohistochemical analysis was performed on paraffin-embedded 8-m sections of murine neonatal skin. The deparaffinized sections were treated with a 1:2000 dilution of the anti-Scarf primary antibody overnight at 4°C and a 1:200 dilution of anti-IgY fluorescein isothiocyanate-conjugated secondary antibody (Aves Laboratories) for 1 h at room temperature. Sections were visualized by fluorescence microscopy.
Calcium Binding Assay-Calcium binding was assayed by the micro method described by Kawasaki et al. (24). Briefly, 3 g of the recombinant proteins were blotted onto a polyvinylidene difluoride membrane and washed with two changes of methanol for 5 min, after which the membrane was soaked in 10 M CaCl 2 , 50 mM Tris-HCl (pH 7.5),100 mM NaCl, and incubated for 1 h with 10 Ci of 45 Ca 2ϩ /ml in the same buffer. The filter was washed twice for 30 s at room temperature in a solution containing 10 M CaCl 2 , 50 mM NaCl, 20 mM Tri-HCl (pH 7.5), and 10% methanol, dried, and exposed to an x-ray film. The autoradiograms were quantitated on a densitometer. The ability to bind Ca 2ϩ by the recombinant proteins was assayed in three independent experiments.

Scarf Cloning and Genomic and Protein Characterization-A
cDNA differentially expressed in suprabasal keratinocytes was obtained from an SSH screen. Scarf 899-bp cDNA (Fig. 1A) contains an ORF that encodes a protein of 148 amino acids with four EF-hand domains (Fig. 1A, bold letters) and a predicted molecular mass of 16.7 kDa. As is characteristic of other members of the CaM-like Ca 2ϩ -binding proteins, Scarf is an acidic protein (pI ϭ 4.59) with no cysteine or tryptophan residues. A particular difference with calcium-binding proteins (13,14) is that Scarf does not present putative sites for myristoylation. Comparison of the four EF-hand motifs of Scarf and other CaM-like proteins showed high homology in the EF-2, -3, and -4 motifs, with EF-1 being the most divergent (Fig. 1B). Protein sequence comparisons determined that Scarf presented 56.7% similarity to mouse CaM (mCaM), 64.2% to hClp, and 64.9% hClsp (Fig. 1C). A phylogenic analysis of several Ca 2ϩ -binding proteins shows that Scarf and hClsp belong to the same subgroup and are more distantly related to CaM and hClp (data not shown).
A genomic DNA region of ϳ120 kb was cloned by screening of a mouse VJ/129 BAC library using an 0.9-kb Scarf cDNA probe. Comparison of the genomic and cDNA sequences determined the 5Ј-and 3Ј-end boundaries and the absence of intronic regions in the Scarf gene. Southern blot analysis of the BAC clone helped us identify the existence of a second gene with homology to Scarf. We termed this second gene Scarf2.
Computational analysis of the Scarf2 genomic sequence proposed that a complete ORF would be derived after the splicing of a putative small intronic region. To determine whether the gene was expressed in keratinocytes and whether the transcript was the product of a splicing event, Scarf2-specific oligonucleotides encoding sequences surrounding the putative intronic region ( Fig. 2A) were utilized to perform RT-PCR with mouse keratinocytes mRNA. As shown in Fig. 2A (lower panel), a fragment was obtained, and after subcloning and subsequent sequencing we determined that the predicted intronic sequence had been spliced to generate the Scarf2 mRNA transcript. As a control reaction, RT-PCR was performed with the same set of primers and no input cDNA. Therefore, Scarf2 is expressed in keratinocytes; the mRNA sequence and putative ORF is presented in Fig. 1A.
The Scarf2 gene localized 15.3 kb downstream of Scarf (Fig.  2B), and the high degree of homology between Scarf and Scarf2 extended to regions upstream and downstream of the ORF for both genes. These regions contain potential proximal promoter and regulatory sequences (Fig. 2B). The upstream regions (Fig.  2B, striped boxes) are 1.6 kb in length and 96% homologous. The downstream regions (Fig. 2B, gray boxes) are 1.2 kb in length and 89.5% homologous.
A mouse gene with homology to hClp (mClp) localized ϳ33 kb upstream from the Scarf gene. Comparison of the genomic organization of the CaM-like genes in human chromosome 10 and the syntenic region on mouse chromosome 13 is shown in Fig. 2B. Human and mouse show similar genomic arrangement, with Scarf and hClsp being potential homologs and the differences being that hClp is in reverse transcriptional orientation and the apparent absence of a Scarf2 homolog of human chromosome 10.
Scarf Expression during Development and in Cultured Primary Keratinocytes-Using Northern blot analysis and radioactive in situ hybridization, we examined the pattern of Scarf expression during development and in mouse primary keratinocytes. In situ hybridization of sagittal sections of 15-day (Fig. 3, A and B) and 16-day (Fig. 3, C and D) mouse embryos with antisense Scarf probes (Fig. 3, A and C), showed epidermis-specific expression at both developmental stages. The expression was not detected in the basal layer, but was clearly seen within the spinous and granular layers in the stratified epidermis of 16-day embryos (Fig. 3, E and F). Expression was also detected in the dorsum of the tongue (Fig. 3, H and I) and in the vibrissae of a 16-day mouse embryo (Fig. 3, J and K). Scarf expression was restricted to the differentiated layers of the stratified epidermis in neonatal skin (Fig. 3, L and M). No expression was detected with the use of Scarf sense probe (Fig.  3, B, D, G, and N).
Using a commercial Northern blot of total embryo mRNA at different developmental stages, Scarf was first detected at 15 days of embryogenesis and dramatically increased by day 17 (Fig. 4A). Sites of expression detected on a dot blot and Northern blot of different adult tissues (Clontech) were thyroid (data not shown) and skeletal muscle (Fig. 4B). Using the commercial blot, no expression was detected in heart, brain, spleen, lung, liver, kidney, or testis (Fig. 4B). With the hybridization of the commercial blot to a glyceraldehyde-3-phosphate dehydrogenase probe to determine loading amounts and integrity of the mRNA, it became apparent that there was less input from the spleen, lung, and testis samples. For this reason, it is not possible to clearly assess the expression level of Scarf in these tissues.
In the epidermis, Scarf expression was restricted to the suprabasal differentiated cells obtained by Percoll gradient of preparations from neonatal skins (Fig. 4C). Several epidermisspecific markers were utilized to validate the Percoll separation method for basal (B) and suprabasal (SB) cells (Fig. 4C). The K5 marker is restricted to the basal fraction, whereas the markers INV, LOR, and K1 are primarily in the suprabasal fraction. The ubiquitously expressed CaM mRNA was detected in both proliferating and differentiated keratinocytes (Fig. 4C).
The expression of Scarf was also studied in cultured primary keratinocytes induced to differentiate in vitro. Scarf expression was detectable in low Ca 2ϩ (0.05 mM) and showed a 4-fold increase 24 h after induction of differentiation by raising the Ca 2ϩ concentration to 0.12 mM and a further increase by 48 h (Fig. 4D). Northern blots were also hybridized with basal cellspecific marker (K5) and suprabasal differentiation-specific markers (K1, LOR, and INV). The Scarf expression pattern in cultured keratinocytes mimics that of early differentiation markers such as INV, whereas the late epidermal differentiation marker, LOR, was detected 48 h after Ca 2ϩ switch.
Treatment of cultured cells with different kinase (PKC, PKA, CaM kinase II) inhibitors showed a specific down-regulation of Scarf expression by treatment with a PKC inhibitor, GF109203X (Fig. 4E). We also determined the down-regulation for other epidermal factors (INV, KI, and LOR), where expression is dependent on the PKC signaling pathway. No effect on Scarf expression was seen with treatment with PKA inhibitor H89 and CaM kinase II inhibitor KN62.
Scarf Ca 2ϩ -binding Protein in Differentiated Keratinocytes-To assess the expression of Scarf protein in epidermis, we performed immunohistochemistry on paraffin sections of mouse neonatal skin with a polyclonal antibody generated in hens against the full-length recombinant protein (rScarf). Scarf protein was clearly detected from the spinous layers to anucleated cells of the stratum corneum in neonatal skin sections using a fluorescein isothiocyanate-conjugated secondary antibody against hen (shown in green, Fig. 5A). Propidium iodide counter-stain determined the localization of the nuclei (Fig. 5A, shown in red).
In Western blot assays, anti-Scarf polyclonal antibody detected rScarf and the endogenous Scarf protein in the soluble fraction of suprabasal cells (Fig. 5B). No Scarf was detected in the insoluble fractions of either basal or suprabasal keratinocytes. Using several recombinant CaM-like proteins in a Western blot analysis, Scarf antibody weakly detected recombinant hClsp but had no cross-reactivity to either hClp or CaM (Fig. 5C).
To investigate the capability of Scarf to bind Ca 2ϩ and the functionality of each of the EF-hand motifs, we performed Ca 2ϩ binding assays. Scarf mutants were obtained by substituting the conserved first Asp residue of each EF-hand motif (sequence GAC) with a Gly residue (GCC sequence) to generate mtEF1, mtEF2, mtEF3, and mtEF4. Parallel binding assays were performed with CaM as control, several CaM-like Ca 2ϩbinding proteins (CaM, hClsp, and hClp), rScarf, and each recombinant mutant. Quantitation of the autoradiograms with a densitometer show that Scarf is able to bind Ca 2ϩ (Fig. 6A) and that mutations in each of the EF-hand motifs lead to a diminished ability to bind Ca 2ϩ , with mutation of the EF-2 domain showing the highest effect (Fig. 6B). The data presented are the mean from three independent experiments Ϯ S.D. (Fig. 6B).

DISCUSSION
Identification of differentially expressed proteins during epidermal stratification will help us elucidate the necessary precursors and the series of complex steps that are required for the formation of a functional water barrier in the skin. By performing an SSH screen between the undifferentiated and differentiated cells of the neonatal mouse epidermis, we identified a novel murine Ca 2ϩ -binding protein gene we termed Scarf. Sequencing of the full-length cDNA determined that Scarf contained regions that would translate into putative EF-hand Ca 2ϩ -binding motifs. The presence of four EF-hand motifs and the degree of homology to CaM indicate that Scarf belongs to the family of CaM-like Ca 2ϩ -binding proteins.
In the in situ hybridization data, we showed that Scarf mRNA was expressed throughout the differentiated layers of the embryonic developing epidermis. In neonatal epidermis, the Scarf gene was also differentially expressed in the spinous and granular layers. An 0.9-kb mRNA Scarf transcript was detected in suprabasal cells obtained from neonatal epidermis by discontinuous Percoll gradients and in primary keratinocytes differentiated in vitro. Scarf expression was similar to that of other early differentiation markers such as INV and was dependent on PKC signaling.
The Scarf transcript encodes a small acidic protein of 148 amino acids detected with anti-Scarf-specific antibodies in the suprabasal layers of the epidermis. Scarf EF-hand motifs maintain the highly conserved residues present in the canonical 12 residue EF-hand motif (DXDX(D/N)GX(I/V)XXXE) (12). The Ca 2ϩ -binding ability and functionality of each Scarf EF-motif were tested. Our results indicate that Scarf is able to bind Ca 2ϩ and mutations in each EF-hand lead to a decreased ability to bind Ca 2ϩ , with mutations in the EF-2 hand motif being the most effective. For the CaM-like protein hClsp, it has recently been reported that its four EF-hand motifs bind Ca 2ϩ , but the motifs have been defined as two Ca 2ϩ -Mg 2ϩ high affinity binding motifs (EF-1 and EF-2) and two low Ca 2ϩ -binding motifs (EF-3 and EF-4) (25). In CaM, mutations of the EF-2 motif decreased its affinity for several of its known target proteins: smooth and skeletal muscle myosin light chain kinases, adenyl cyclase, and plasma membrane Ca 2ϩ -ATPase (26). It remains to be determined whether Scarf EF-hands have differential affinity for Ca 2ϩ , whether the motifs are independent or mutually interactive, and what physiological relevance this has for its capacity to interact with its specific target proteins in the mouse epidermis. Targets for other members of the CaM-like binding-protein family have been reported, with hClp binding a human unconventional myosin X (27), CaBPs binding the inositol trisphosphate receptor (Ins-P 3 -R) (28), and hClsp binding TGase3 (17). TGases are Ca 2ϩ -dependent enzymes that catalyze the N-(␥-glutamyl)lysine isopeptide bonds that occur during the cross-linking of precursors to form the specialized cornified cell envelope in the epidermis (1). The Ca 2ϩ -dependent binding of hClsp to TGase3 is a potential way of regulating epidermal differentiation through the modulation of TGase3 enzyme activity (16). Alternatively, members of the S100 EFhand-containing protein family are TGase enzymatic targets (11,29), leading to a mechanism for regulating the S100 protein functions in the epidermis.
Based on sequence homology and genomic localization, we propose that Scarf is a potential mouse homolog of hClsp. However, certain relevant distinctions between Scarf and hClsp have been determined: (a) hClsp was reported as a late granular differentiation marker expressed only in the very late stages of human keratinocyte differentiation and detected in FIG. 6. Functional Ca 2؉ binding assay of CaM-like binding proteins. A, a Ca 2ϩ binding assay was performed with rScarf and CaM with increasing amounts of protein (g). The results were quantitated and plotted as concentration of protein (g) versus relative intensity as determined by densitometer. B, Ca 2ϩ binding assays were performed with wild-type and mutant proteins. The data presented are the mean from three independent experiments Ϯ S.D. mtEF1-4 represent the mutated forms of Scarf in each EF-hand 1-4 motif, respectively. lung tissue (16,17), whereas Scarf is expressed early in the differentiation process throughout the spinous and granular layers in the stratified epidermis and is also detected in thyroid and skeletal muscle, and (b) Scarf has a slightly extended central helix when compared with hClsp (16), making it more similar to CaM. For CaM, the central helix serves a fundamental role, allowing the amino-and carboxyl-terminal domains to swing around and permitting interactions with specific target proteins (6,12). The focus of our future studies will be to establish whether the slightly extended central helix in Scarf allows for the binding to targets other than the factors that interact with hClsp and to determine the role of these interactions in the control of the Ca 2ϩ -dependent epidermal differentiation.
In this study we have also identified a gene with high homology to Scarf, which we termed Scarf2, that localizes ϳ15 kb downstream on mouse chromosome 13. Sequence analysis of the genomic regions 5Ј and 3Ј to each gene showed that the homology extended to the putative proximal promoter and 3Јregions. We hypothesize that the bigenic cluster arose by recent duplication in the mouse, because we were unable to localize by genomic sequence scanning a counterpart to Scarf2 in the human genome. We detected a transcript for Scarf2 in mouse keratinocytes by RT-PCR but not in Northern blots with samples of total RNA from basal and suprabasal cells isolated from neonatal skins or in commercial blots (data not shown), suggesting that Scarf2 is expressed at lower levels than Scarf. Analysis of the Scarf2 putative ORF indicated that it encodes a Ca 2ϩ -binding protein, and based on a comparison with the canonical consensus sequence for the EF-hand motif and the critical role of the residues at positions 1 and 12, we hypothesize that the Scarf2 EF-1 and -4 motifs will be inactive or present altered Ca 2ϩ binding. How these differences characterize Scarf2 function remains to be determined, as does the role of the highly homologous Scarf and Scarf2 and their interaction with target proteins in the transduction of the Ca 2ϩ -dependent regulatory signals during the epidermal differentiation process.