| Human Molecular Genetics | Pages |
©1999 Oxford University Press |
Ectodysplasin is a collagenous trimeric type II membrane protein with a tumor necrosis factor-like domain and co-localizes with cytoskeletal structures at lateral and apical surfaces of cells
Introduction
Results
Ectodysplasin belongs to the type II class of membrane proteins
Ectodysplasin proteins with collagenous domain form trimers
Ectodysplasin co-localizes with the actin cytoskeleton on the apical and lateral surfaces of cells
Subcellular localization of the isoforms and mutation constructs
Discussion
Ectodysplasin is a trimeric type II membrane protein
Cellular localization of ectodysplasin
Homology to TNF-like ligands
Materials And Methods
Production of antibodies
Preparation of constructs
Immunoblots
Immunofluorescence staining
Acknowledgements
References
Ectodysplasin is a collagenous trimeric type II membrane protein with a tumor necrosis factor-like domain and co-localizes with cytoskeletal structures at lateral and apical surfaces of cells
Received June 4, 1999; Revised and Accepted July 20, 1999
Anhidrotic ectodermal dysplasia (EDA) is a human genetic disorder of impaired ectodermal appendage development. The EDA gene encodes isoforms of a novel transmembrane protein, ectodysplasin. The sequence of the longest isoform includes an interrupted collagenous domain of 19 Gly-X-Y repeats and a motif conserved in the tumor necrosis factor (TNF)-related ligand family. In order to understand better the function of the ectodysplasin protein molecule and its domains, we have studied the processing and localization of wild-type and mutated isoforms in transfected human fetal kidney 293 and monkey kidney COS-1 cells. Similar to other members of collagenous membrane proteins and members of TNF-related ligands, ectodysplasin is a type II membrane protein and it forms trimers. The membrane localization of ectodysplasin is asymmetrical: it is found on the apical and lateral surfaces of the cells where it co-localizes with cytoskeletal structures. The TNF-like motif and cysteines found near the C-terminus are necessary for correct transport to the cell membrane, but the intracellular and collagenous domains are not required for the localization pattern. Our results suggest that ectodysplasin is a new member in the TNF-related ligand family involved in the early epithelial-mesenchymal interaction that regulates ectodermal appendage formation.
INTRODUCTION
Ectodermal dysplasias (EDs) are a group of >50 distinguishable hereditary disorders with impaired development of ectodermal appendages (1). All are rare, and the molecular basis for EDs is largely unknown. Since all the appendages affected in EDs-hair follicles, sweat glands, teeth and nails-arise from interactions between epithelium and mesenchyme, it has been suggested that the genes responsible for the various EDs are part of signaling pathways that regulate ectodermal organogenesis. One entry point to study the pathways is provided by the identification of the gene mutated in X-linked anhidrotic ectodermal dysplasia (EDA), one of the most common types of ED. The symptoms of EDA include absence or hypoplasia of hair, teeth and sweat glands (2). The EDA gene encodes novel protein products named ectodysplasin (3,4). The original cDNA encodes a novel 135 amino acid product with a suggested type II orientation (3), and the protein has been shown to localize to the outer cell membrane (5). Recently we have shown that the EDAgene undergoes extensive alternative splicing and produces several transcripts: the original shortest EDA-O form, longest EDA-A form and five other forms (EDA-B, -C, -D, -E and -F) (4). All the transcripts have the same first exon which encodes a short predicted intracellular domain, a single transmembrane domain and an extracellular domain of 66 amino acids. In addition, the longest 391 amino acid form, EDA-A, includes a positively charged domain, 19 repeats of Gly-X-Y with an interruption between repeats 11 and 12, two putative N-glycosylation sites and a cysteine-rich C-terminal domain (4,6).
The phenotype of mouse Tabby strain, which is highly similar to EDA, results from mutations in the murine homologue of EDA (7,8). The EDA gene is evolutionarily highly conserved, the longest encoded forms of human and mouse proteins showing an overall homology of 94.6%. The Tabby gene also undergoes alternative splicing and homologs of the human EDA-A, -B and -C have been described also in mice (8). Studies of expression of EDA in human tissues by in situ hybridization and immunohistochemistry have shown that, in line with the EDA disease phenotype, the gene is expressed in a number of tissues, including developing and mature skin appendages (3,9). The mouse Tabby gene is also expressed in various tissues, including embryonic teeth and epidermis (8).
The structure of ectodysplasin suggests that it might belong to the group of membrane-associated collagenous type II proteins, which include collagen types XIII and XVII (10,11), proteins with short collagenous domains, types I and II macrophage scavenger receptors (12,13) and MARCO protein (14). Scavenger receptors and MARCO protein have been suggested to function as trimeric receptors in host defence. Scavenger receptors can bind a variety of polyanionic ligands (15), and MARCO protein has been shown to bind bacteria (14). Since the number of uninterrupted Gly-X-Y repeats in EDA-A is only 11, and is among the smallest found in this group, it has remained unclear whether it can form collagenous triple helix.
In addition to the collagen domain, the C-terminal domain of the EDA-A protein contains a domain of homology to the members of the tumor necrosis factor (TNF) ligand family (Fig. 1). TNF-like ligands induce signals leading to cell death, proliferation or differentiation and they are also trimeric type II membrane proteins (16).
Figure 1. (A) Schematic structure of EDA isoforms and mutation constructs used in the study. EDA-O, -C, -A1 and -A2 with the shown protein product sizes are encoded by alternatively spliced transcripts of the EDA gene (4). All the forms have a common 132 amino acid N-terminus. EDA-A2 is missing two amino acids, Val307 and Glu308, found in EDA-A1. The putative transmembrane domain is shown as a hatched box, and Gly-X-Y collagenous repeats as black boxes; four repeats are found at positions 79-90 in all the forms, two repeats at the C-terminus of EDA-O and 19 interrupted repeats at positions 180-238 of EDA-A1 and -A2. Two putative N-glycosylation sites found at positions 311 and 372 are shown as hexagons. Homology to ligands of the TNF family is found in the C-terminal domain starting from amino acid 249. The most conserved motif (amino acids 293-309) is shown by a shaded box [see (B) below]. The mutation constructs are as follows: [Delta]COL is truncated before the 19 repeat collagenous domain; [Delta]CYS is truncated before the cysteine-rich C-terminus; in Cys-Ser the C-terminal cysteines are replaced by serines; [Delta]IC lacks the N-terminal intracellular domain; COL-mut has an in-frame deletion of Gly-X-Y repeat numbers 8 and 9; COL-del has an in-frame deletion of all the 19 Gly-X-Y repeats; and TNF-del has an in-frame deletion of the 17 amino acid TNF motif. The antigen recognized by the A17 antibodies is shown with a black line. The columns on the right summarize the ability of each of the products to trimerize and localize to the plasma membrane (see Results). NA, not assessed. (B) Alignment of the C-terminal domain of ectodysplasin EDA-A1 form (EDA) with representative members of the TNF ligand superfamily, human TNF-[alpha] (TNFA), TNF-[beta] (TNFB) and FAS ligand (FASL), using the CLUSTALW program. Amino acid positions are shown for each of the proteins. The main homology regions are boxed. The shaded box shows the motif conserved in all the TNF family members, with consensus pattern L/V-X-L/I/V/M-X-X-X-G-L/I/V/M/F-Y-L/I/V/M/F/Y-L/I/V/M/F/Y-X-X-Q/E/K/H/L-L/I/V/M/G/T-X-L/I/V/M/F/Y (35,36). Sites of missense mutations found in the EDA patients within the TNF homology region (H252L, G291R, G291W, D298H, G299S, A349T, A356D and R357P) are shown with asterisks (4,6).
The evidence presented here proves that ectodysplasin also has type II membrane orientation and trimeric structure, and that it has strikingly asymmetrical membrane localization. The functional importance of the portions of the EDA-A protein is shown by many EDA producing mutations located in the collagenous, TNF-like and cysteine-rich domains (4,6). We have shown previously that the EDA-O and -A forms can induce rounding and detachment of breast epithelial MCF-7 cells (4,5). In this report, we study the processing and defined cellular localization of wild-type and mutated ectodysplasin isoforms in transfected cells in order to understand better the function of the ectodysplasin protein molecule and its various domains. Based on the scenario provided by the results, we propose that ectodysplasin is a new collagen domain-containing member of the superfamily of TNF-like ligands involved in the regulation of ectodermal morphogenesis.
RESULTS
Ectodysplasin belongs to the type II class of membrane proteins
No signal peptide is found in the ectodysplasin protein sequence, and it has been suggested to belong to the type II class of membrane proteins, with their C-terminus protruding to the extracellular space. To test for its orientation, we used antibodies raised against a 17 amino acid peptide sequence from the N-terminus of the protein (A17, Fig. 1A; Materials and Methods). Anti-D-EDA antibodies, raised against the full length 135 amino acid EDA-O product (Fig. 1) (5) were used as control. When EDA-A2 (Fig. 1A) transfected 293 cells are reacted with anti-D-antibodies, surface staining is obtained both under permeabilizing and non-permeabilizing conditions (Fig. 2A and B). However, when non-permeabilized cells are reacted with A17 antibodies, no staining is visible (Fig. 2D), indicating that the N-terminal antigen recognized by the antibody truly resides inside the cell.
Figure 2. Ectodysplasin is orientated with its N-terminus intracellularly. 293 cells transfected with EDA-A2 were stained with anti-D-EDA or A17 antibodies (green) and counterstained for actin (red). When permeabilized cells (+Triton X-100) are reacted with anti-D-EDA (A) or A17 (C) antibodies, intracellular as well as membrane staining is seen. Without permeabilization (-Triton X-100), anti-D-EDA antibodies, which recognize antigens from both sides of the transmembrane domain, stain the cell surface (B), whereas the A17 antibodies recognizing an N-terminal antigen fail to show staining (D).
Ectodysplasin proteins with collagenous domain form trimers
To test whether the 19 Gly-X-Y repeats found in the EDA-A protein could support collagenous trimer formation we used expression constructs with the cDNAs corresponding to the EDA-O, -C, -A1 and -A2 transcripts (Fig. 1A), as well as mutated forms of the constructs lacking various domains. [Delta]COL has a deletion of the C-terminus starting from the collagenous domain; [Delta]CYS has deletion of the cysteine-rich C-terminus; Cys-Ser has serines instead of the cysteines in the C-terminus; COL-mut has an in-frame deletion of two Gly-X-Y repeats; and [Delta]IC lacks the intracellular domain (Fig. 1A).
Total protein lysates from 293 cells transfected with the constructs were run on SDS-PAGE and immunoblotted. When the samples were run without boiling under non-reducing conditions in 6% gel, bands corresponding to the size of trimeric forms migrating around 140 kDa were seen in the lysates of EDA-A1 and -A2 transfected cells (Fig. 3A). Although some weaker high molecular weight bands were seen also in cells transfected with the EDA-O construct, no bands corresponding to the estimated trimeric size could be detected. Of the mutated constructs, trimeric forms could be detected in the lysates of COL-mut and [Delta]IC transfected cells, but not in [Delta]COL, [Delta]CYS and Cys-Ser transfected cells (Fig. 3A). No trimer bands were detected on the 6% gel when the samples were treated with reducing agent and boiled (data not shown).
Figure 3. Trimer formation by the wild-type and mutated ectodysplasin protein isoforms. 293 cells were transfected with the indicated constructs, run on SDS-PAGE, blotted and reacted with anti-D-EDA antibodies. (A) (top) Non-boiled samples were run under non-reducing conditions on a 6% gel; (B) (bottom) boiled samples were run under reducing conditions on a 12% gel. Under non-reducing conditions (A), bands corresponding to the expected trimeric size, ~130-150, kDa are seen in the lysates of EDA-A1, EDA-A2, COL-mut and [Delta]IC transfected cells. In the other lysates, no clear bands matching the expected trimeric size (~55 kDa for EDA-O, 60 kDa for [Delta]COL, 120 kDa for [Delta]CYS and 140 kDa for Cys-Ser) are detected. Under reducing conditions (B), all the protein monomers migrate as two or three close bands around the correct putative size, EDA-O at 18 kDa, EDA-A1, EDA-A2, COL-mut and Cys-Ser at 45-50 kDa, [Delta]COL at 20-25 kDa, and [Delta]CYS and [Delta]IC at 40-45 kDa. In the lysates of EDA-A2 and COL-mut transfected cells, also smaller bands migrating around 20 kDa, possibly degradation or cleavage products, are seen. No protein is detected in the control cells transfected with the vector alone.
When the same samples were boiled and run on a 12% gel under reducing conditions, all the protein monomers showed the expected size (Fig. 3B). All the products were seen as two or three closely migrating bands, suggesting post-translational modifications. The putative two N-glycosylation sites found in the EDA-A form cannot account for the two bands seen in the EDA-O form and in [Delta]COL, since they lack these sites. In EDA-A1, EDA-A2 and COL-mut transfected cells smaller ~20 kDa products are also observed (Fig. 3B; in the EDA-A1 lysate not seen in this exposure). These could be degradation products or products left in the membrane after a specific cleavage event. No protein product was visible in control cells transfected with the vector alone (Fig. 3). Under the conditions used, no protein product could be visualized on immunoblots of the lysates from cells transfected with the EDA-C construct (data not shown). The trimer formation results of each of the proteins is summarized in Figure 1A.
Ectodysplasin co-localizes with the actin cytoskeleton on the apical and lateral surfaces of cells
We have previously reported localization of the EDA-O and -A proteins to the plasma membrane in transfected COS-1 and MCF-7 cells (4,5). The localization patterns of the various protein isoforms were now compared in COS-1 and 293 cells. Cells were transfected with the EDA-O and -A2 constructs, immunostained and analyzed with confocal laser scanning microscopy. Staining patterns obtained with EDA-O-transfected COS-1 cells and EDA-A2-transfected 293 cells are shown in Figure 4. In the basal section of the transfected cells almost no staining for the protein was detectable (Fig. 4A and E). When the cells were scanned upwards, the staining grew stronger in the lateral (Fig. 4F and G) and apical (Fig. 4B and H) surfaces of the cells. There the EDA protein co-localized with filamentous actin. On the apical surface of the cells both forms co-localized with cytoskeletal structures to give a punctated pattern. The same staining pattern was obtained with EDA-O (Fig. 5) and -A1 transfected into 293 cells, or with EDA-A1 and -A2 transfected into COS-1 cells (data not shown).
Figure 4. Ectodysplasin associates with cytoskeletal structures on the apical and lateral surfaces of cells. EDA-O-transfected COS-1 (A-D) and EDA-A2-transfected 293 (E-H) cells were immunostained for ectodysplasin (green), counterstained for filamentous actin (red) and analyzed with confocal microscopy. For COS-1 cells, double staining of the basal (A) and apical (B) surfaces, as well as single staining for ectodysplasin (C) and actin (D) in the same section as (B) are shown. For 293 cells double stainings are shown for four different sections of the same cells, going up from the basal to apical (E-H). In the basal layer where the cells are attached to the substratum (A and E), almost no staining for ectodysplasin is visible. Staining is more prominent in the lateral (F and G, arrowheads) and apical surfaces (B and H, arrowheads), where ectodysplasin co-localizes with actin, as seen by yellow color. Both the EDA-O (B, C and D) and the EDA-A2 (H) proteins associate with the punctate pattern of cytoskeletal structures in the apical surface of the cells.
Figure 5. Subcellular localization of wild-type and mutated ectodysplasin isoforms. 293 cells were transfected with EDA-O (A), EDA-C (B), EDA-A2 (C), [Delta]CYS (D and H), [Delta]COL (E), COL-mut (F) and [Delta]IC (G), stained with anti-D antibodies (green) and counterstained for nucleic acids (red). (A-D) Cells were permeabilized prior to the staining; (E-H) the permeabilization was omitted and thus only surface staining is visible. In EDA-O-transfected cells (A), both intracellular staining of endoplasmic reticulum (arrow) and apical membranal staining (arrowhead) are seen. In EDA-C-transfected cells (B), only few cells are stained (arrows) with a weaker intensity compared with the other constructs. In [Delta]CYS transfected cells, the intracellular staining differs from wild-type; the cells have more compact ring-like staining around the nucleus (D, arrows) compared with the netlike endoplasmic reticulum staining filling the whole intracellular space seen in the EDA-O- (A) and EDA-A2- (C) transfected cells. Also, with [Delta]CYS the outer membrane staining is markedly reduced (H). Proteins translated from [Delta]COL (E), COL-mut (F) and [Delta]IC (G) localize to the cell membrane and show same punctate pattern as wild-type on apical surface of the cells (arrowheads).
Subcellular localization of the isoforms and mutation constructs
In order to study which domains of the ectodysplasin protein are required for correct localization in the cell membrane, 293 and COS-1 cells were transfected with the isoform and mutation constructs shown in Figure 1A. Immunostaining and analysis by confocal laser scanning microscopy of 293 cells gave patterns seen in Figure 5. In the cells transfected with the EDA-C isoform (Fig. 5B), staining was obtained only in a few cells, and the signal was weaker compared with cells transfected with the three other isoforms EDA-O (Fig. 5A), -A2 (Fig. 5C) and -A1 (data not shown). Also, very little surface staining was seen in non-permeabilized EDA-C transfected cells (data not shown). Taken together with the immunoblot results, this suggests that the EDA-C protein is unstable.
When 293 cells transfected with the deletion constructs [Delta]COL, COL-mut or [Delta]IC were immunostained without permeabilization, distinct cell surface staining was observed, indicating that the proteins encoded by these constructs were effectively transported to the cell membrane (Fig. 5E-G). Furthermore, the staining pattern obtained with these constructs was comparable with that seen with the wild-type proteins, and again, the proteins localized in a punctate pattern on the lateral and apical surfaces of the cells. The staining obtained with the COL-del construct lacking the 19 Gly-X-Y repeat collagen domain was similarly comparable with the wild-type localization (data not shown). In contrast, in cells transfected with the [Delta]CYS construct, surface staining was remarkably reduced (Fig. 5H), and the intracellular localization differed from the wild-type proteins. Compared with the fine `network' seen with wild-type isoforms (Fig. 5A and C) this variant gave more compact staining around the nucleus (Fig. 5D). The localization patterns obtained with the Cys-Ser construct, and with the construct lacking the 17 amino acid TNF-like motif, TNF-del, were similar to that for [Delta]CYS (data not shown). Each of the constructs behaved in a similar way in transfected COS-1 cells (data not shown). The cell membrane localization results obtained with the various constructs are summarized in Figure 1A.
DISCUSSION
Ectodysplasin is a trimeric type II membrane protein
As previously suggested by analysis of the amino acid sequence, we show now that ectodysplasin is a transmembrane protein with type II orientation; i.e. its N-terminus is positioned inside the cell. Type II orientation has recently been shown also for the mouse homolog, Tabby-A protein (17). This orientation occurs in only ~5% of transmembrane proteins, but has been found for other collagenous membrane proteins, type XIII and XVII collagens (11,18), macrophage scavenger receptors types I and II (12,13) and MARCO protein (14). Also members of the TNF ligand family are known to have type II orientation (16). The orientation implies that all ectodysplasin isoforms have a short 43 amino acid intracellular domain and extracellular domains of varying length, and that the (Gly-X-Y)19 collagen-like domain and the TNF-like domain are out in the extracellular space.
Our results show that the EDA-A protein is found in cells in a complex with a molecular weight corresponding to trimers. Trimeric forms were not seen with the proteins lacking the collagenous domain, EDA-O and [Delta]COL, nor with constructs lacking the C-terminal cysteines, [Delta]CYS and Cys-Ser. The deletion of the intracellular domain did not prevent the trimer formation. The uninterrupted Gly-X-Y repeat segment in the EDA-A protein seems to be among the shortest known, but is comparable with ficolin [alpha], a secreted collagen-like protein, which has a domain of 18 uninterrupted repeats followed by two additional repeats and is also known to form trimers (19). The membrane-associated collagen types XIII and XVII have several collagen domains, type XIII having three domains of 31, 56 and 77 Gly-X-Y repeats (20) and type XVII having 15 domains varying in length between 5 and 80 repeats (21). In the MARCO protein the number of the repeats is 89 with one interruption, and for the type I and type II scavenger receptors, 24 repeats.
Apparently, any of several other protein domains can promote or stabilize collagenous trimers. The murine MARCO protein has been suggested to form trimers in which the monomers are linked to each other by disulfide bonds (14). For the type I and type II scavenger receptors, the trimers have been shown to comprise of disulfide-linked dimers and non-covalently associated monomers (22). Truncated scavenger receptors containing only eight Gly-X-Y repeats have also been shown to form trimers, suggesting that at least in the case of the scavenger receptor, other domains of the protein, probably the [alpha]-helical coiled coil domain, can trimerize independently of the collagenous domain (23). In the case of the EDA-A protein, transformation of the C-terminal cysteines to serines remarkably reduced the amount of the trimeric form, and thus the cysteines seem to be necessary for the formation of stable trimers.
The collagenous domain has been shown to mediate ligand binding by scavenger receptors (23). Moreover, an isolated collagen domain from complement component C1q, which is a secreted, complex oligomeric protein consisting of A, B and C subchains with 26 or 27 Gly-X-Y repeat collagenous domains (24), has been shown to bind polyanionic ligands (23). Functional importance for the collagenous domain of the EDA-A protein was suggested by the finding that some EDA patients carry in-frame deletions of two or four Gly-X-Y repeats (4). It remains to be seen whether the collagen domain of ectodysplasin has any ligand binding properties.
Cellular localization of ectodysplasin
As shown by double staining with filamentous actin, ectodysplasin co-localizes with cytoskeletal structures on the plasma membrane. Furthermore, the localization is strikingly asymmetrical. Almost no protein is found in the regions of contact with the substratum, some in the regions of cell-cell contact, and high levels are seen at points where cells are free of any contact. All the analyzed protein forms localized similarly, except for [Delta]CYS, Cys-Ser and TNF-del, which were not efficiently transported to the cell membrane. Thus, the C-terminal cysteines, in addition to being needed for trimerization, are needed for efficient transport to the membrane. It is possible that without the cysteines, the collagenous domain makes abnormal structures and the protein remains in the endoplasmic reticulum, accumulating in packed staining around the nucleus. The deletion of the TNF motif also interfered with the localization of the protein to the cell membrane. Possible explanations for this are that the deletion causes conformational changes or affects the glycosylation occurring on the nearby site. Thus, although the significance of the TNF-like domain for the function of ectodysplasin remains under further study, our results clearly demonstrate its importance for the proper processing of the protein. The deletion of the intracellular domain did not abolish co-localization with cytoskeletal structures, implying that the specific pattern does not depend on intracellular contact with cytoskeletal proteins. Neither does it depend on the presence of the collagenous domain, since the same pattern was seen also with the shortest EDA-O form.
We have previously suggested the involvement of ectodysplasin in cell-cell or cell-matrix interactions (4,5). However, the subcellular localization of ectodysplasin suggests that it is not involved directly in the attachment of the cells to the substratum, at least in the in vitro conditions used. The localization pattern is also quite different from that of classical cell-cell adhesion molecules such as E-cadherin, which is found in tight junctions between the cells (24,25; S. Ezer, unpublished data). Ectodysplasin might rather have a regulatory role in adhesion. Such a role would be consistent with the observed rounding up and detachment of transfected MCF-7 cells (4,5).
Homology to TNF-like ligands
Despite the similarity in quaternary structures, ectodysplasin differs from the other type II membrane collagenous proteins. Whereas the scavenger receptors and MARCO protein show sequence homology of 48.9% in their C-terminal domains, no significant sequence homology is found to ectodysplasin outside the collagenous domain. In addition, whereas both scavenger receptors and MARCO are implicated in host defence mechanisms and are expressed mainly on macrophages (14,27), ectodysplasin is involved in differentiation and is expressed in a broad range of tissues and cell types (9).
The C-terminal part of the extracellular domain of ectodysplasin reveals homology to members of TNF ligand superfamily. However, ectodysplasin is unique also among this group of proteins. Except for the conserved TNF motif, there is quite limited sequence homology to other TNF-like molecules (Fig. 1B). So far only one other protein with collagen-like structure, the adipocyte complement-related protein of 30 kDa (ACRP30), has been described to contain regions of structural homology to TNF (28). Members of the growing family of TNF-like ligands are widely expressed and are involved in cell death, survival, proliferation and differentiation (29-31). Many of the TNF-like ligands occur in cell-bound form, but proteolytically released forms are also known (30,32,33). In transfected cells, the EDA-A protein showed the expected molecular weight of 45 kDa, but in addition we detected minor amounts of 20 kDa product. Thus, it is possible that ectodysplasin is also cleaved to produce a diffusible molecule. Since most of the TNF-like ligands are known to be active as trimers (16), the trimer formation by ectodysplasin suggests that it could also function as a ligand. Importance of the TNF-like domain is also supported by the fact that many missense mutations leading to the full EDA phenotype are found there (4,6) (Fig. 1).
We suggest that ectodysplasin is a new TNF-related ligand with collagen structure that activates a signaling pathway in the early morphogenetic stages of ectodermal appendage formation. Consistent with this role, EDA is expressed in body epithelium and hair follicles during embryonic development, but continues also in adulthood, in mature skin appendages (3,4). This could be merely basal expression without functional consequences, or could be due to a possible function of EDA in additional pathways not related to skin development. Alternatively, EDA could participate also in signaling pathways involved in the continuous regeneration of adult skin.
MATERIALS AND METHODS
Production of antibodies
To obtain affinity purified polyclonal A17 antibodies, rabbits were immunized with a synthetic 17 amino acid peptide (PEVERRELLPAAAPRER) designed from the N-terminus of the EDA-A protein (Fig. 1A). Obtained antiserum was purified with the peptide antigen bound to Affi-gel 10 column (Bio-Rad, Hercules, CA), according to the manufacturer's instructions. Anti-D-EDA antibodies, which were made against the 135 amino acid EDA-O protein, have been described previously (5).
Preparation of constructs
To prepare the [Delta]COL, [Delta]CYS and [Delta]IC constructs, fragments were amplified by PCR using pCMV5-EDA-A2 (4) as a template with the following primers. For [Delta]COL: AK-1F, 5[prime]-TAGGTACCTCAGAGGTCGTGAACGGCTG-3[prime] (KpnI site underlined) and AB-2R, 5[prime]-TAGGATCCTCATGCTTTCTTTCCCTTTTTCTTG-3[prime] (BamHI site underlined); for [Delta]CYS: AK-1F, as above, and AB-3R, 5[prime]-TAGGATCCTCACTGCAGGAAGGGCTTCTC-3[prime] (BamHI site underlined); for [Delta]IC: AC-2, 5[prime]-TCATCGATTGAGGCCATGGGCGAAGGGAACATCTGC-3[prime] (ClaI site underlined) and AB-R, 5[prime]-TAGGATCCACAGCAGCACTTAGAGGT-3[prime] (BamHI site underlined). The fragments were cloned into the pCMV5 expression vector utilizing the underlined digestion sites. To prepare the COL-mut construct, pCMV5-EDA-A2 was partly digested with EcoRI and self ligated. A clone was selected in which a 18 bp fragment between two EcoRI sites located in the collagenous domain was lost. This clone has an in-frame deletion of two Gly-X-Y repeats (repeat numbers 8 and 9). To prepare the Cys-Ser construct the method of site-directed mutagenesis with PCR (34) was used. Primer cys-mut, 5[prime]-TGCAGTCCACACGCAGCATCGAGACGGGCAAGACAACTACAACACT-
TCCTATACCGCAGGCGTCTCCCTCCT-3[prime], which changes all the three cysteines to serines, and primer AB-R, as above, were used in the first round of PCR. This product was used as a primer together with primer E1AEXT, 5[prime]-CCGGGCCTCAAGAGAGTGGATGT-3[prime], in the second round. The obtained PCR product was exchanged with the insert of plasmid pCMV5-EDA-A2 utilizing ApaI and BamHI restriction sites. To prepare COL-del and TNF-del constructs, two PCR fragments from each side of the segment to be deleted were amplified from the plasmid pCMV-EDA-A1 and ligated into the pCMV-5 vector between HindIII and SalI sites. The primers used were ATGHind, 5[prime]-AAGCTTATGTCTCCGGAGGCCATGGG-3[prime] and R-Eco, 5[prime]-CTGCTTTCTTTCCCTTTTTC-3[prime] for the 5[prime] fragment of COL-del, and for the 3[prime] fragment F-Eco, 5[prime]-GATAAAGCTGGAACTCGAG-3[prime] and ECR, 5[prime]-GTCGACACGGACAGAGGCAAAATGGG-3[prime]. For the 5[prime] fragment of TNF-del the primers were ATGHind, as above, and TNF-R, 5[prime]-GGTACCCCCGCTGCGGGGATGTAG-3[prime] and for the 3[prime] fragment TNF-F, 5[prime]-GGTACCCAACTTCACTGACTTTGCCA-3[prime] and ECR, as above. All constructs used in the study were verified by sequencing.
Immunoblots
Cells were transfected using the lipofection method with Lipofectin (Gibco BRL Life Technologies, Rockville, MD) or FuGENE (Boehringer Mannheim, Mannheim, Germany) reagents according to the manufacturer's instructions. Transfected cells were grown in the presence of 75 µg/ml ascorbate (Sigma, St Louis, MO). Forty-eight hours post-transfection the cells were washed with phosphate-buffered saline (PBS) and lysed with Laemmli loading buffer. In non-reducing conditions, loading buffer without reducing agent was used and the lysates were loaded as such. For reducing conditions, [beta]-mercaptoethanol was added at a concentration of 5% and the samples were boiled for 5 min. Equal volumes from each lysate were run on SDS-PAGE and electroblotted. The proteins were visualized by incubating with anti-D-EDA serum precleared against the GST-column at a concentration of 1:8000 and peroxidase conjugated secondary antibody followed by enhanced chemiluminescence. Kaleidoscope prestained standards (Bio-Rad) were used as molecular weight markers.
Immunofluorescence staining
Cells were grown on glass coverslips, transfected and stained 48 h after transfection as described (5). Permeabilization was done with 0.2% Triton X-100 in PBS for 5 min at room temperature. Affinity purified anti-D-EDA antibodies were used at the concentration of 1.75µg/ml and A17 antibodies at 1 µg/ml. Samples were imaged with an LSM 410 Laserscan microscope (Zeiss, Oberkochen, Germany) and the images processed using Adobe Photoshop Software.
ACKNOWLEDGEMENTS
We are grateful to Marja Mikkola and Irma Thesleff for valuable comments and collaboration. We also thank Ulpu Saarialho-Kere, Johanna Pispa and Anand Srivastava for helpful discussions. This work was supported by the Sigrid Juselius Foundation and Academy of Finland. M.B. is a recipient of an EU TMR fellowship.
REFERENCES
+Present address: Department of Biological Chemistry, Weizmann Institute of Science, Rehovot 76100, Israel
§To whom correspondence should be addressed at: Finnish Genome Center, PO Box 21 (Tukholmankatu 2), 00014 University of Helsinki, Finland. Tel: +358 9 1912 6538; Fax: +358 9 1912 6789; Email: juha.kere{at}helsinki.fi
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: jnl.info{at}oup.co.uk
Last modification:
Copyright© Oxford University Press, 1999.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  S. Song, D. Han, H. Qu, Y. Gong, H. Wu, X. Zhang, N. Zhong, and H. Feng EDA Gene Mutations Underlie Non-syndromic Oligodontia Journal of Dental Research, February 1, 2009; 88(2): 126 - 131. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C.-Y. Cui, T. Hashimoto, S. I. Grivennikov, Y. Piao, S. A. Nedospasov, and D. Schlessinger Ectodysplasin regulates the lymphotoxin-beta pathway for hair differentiation PNAS, June 13, 2006; 103(24): 9142 - 9147. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C.-W. Franzke, P. Bruckner, and L. Bruckner-Tuderman Collagenous Transmembrane Proteins: Recent Insights into Biology and Pathology J. Biol. Chem., February 11, 2005; 280(6): 4005 - 4008. [Full Text] [PDF]
|
|
|
|
 |
|
 C.-Y. Cui, M. Durmowicz, C. Ottolenghi, T. Hashimoto, B. Griggs, A. K. Srivastava, and D. Schlessinger Inducible mEDA-A1 transgene mediates sebaceous gland hyperplasia and differential formation of two types of mouse hair follicles Hum. Mol. Genet., November 15, 2003; 12(22): 2931 - 2940. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Latvanlehto, A. Snellman, H. Tu, and T. Pihlajaniemi Type XIII Collagen and Some Other Transmembrane Collagens Contain Two Separate Coiled-coil Motifs, Which May Function as Independent Oligomerization Domains J. Biol. Chem., September 26, 2003; 278(39): 37590 - 37599. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Smahi, G. Courtois, S. H. Rabia, R. Doffinger, C. Bodemer, A. Munnich, J.-L. Casanova, and A. Israel The NF-{kappa}B signalling pathway in human diseases: from incontinentia pigmenti to ectodermal dysplasias and immune-deficiency syndromes Hum. Mol. Genet., October 1, 2002; 11(20): 2371 - 2375. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C.-Y. Cui, M. Durmowicz, T. S. Tanaka, A. J. Hartung, T. Tezuka, K. Hashimoto, M. S.H. Ko, A. K. Srivastava, and D. Schlessinger EDA targets revealed by skin gene expression profiles of wild-type, Tabby and Tabby EDA-A1 transgenic mice Hum. Mol. Genet., July 15, 2002; 11(15): 1763 - 1773. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. K. Srivastava, M. C. Durmowicz, A. J. Hartung, J. Hudson, L. V. Ouzts, D. M. Donovan, C.-Y. Cui, and D. Schlessinger Ectodysplasin-A1 is sufficient to rescue both hair growth and sweat glands in Tabby mice Hum. Mol. Genet., December 1, 2001; 10(26): 2973 - 2981. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. J. Stratigos and H. P. Baden Unraveling the Molecular Mechanisms of Hair and Nail Genodermatoses Arch Dermatol, November 1, 2001; 137(11): 1465 - 1471. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Priolo and C. Lagana Ectodermal dysplasias: a new clinical-genetic classification J. Med. Genet., September 1, 2001; 38(9): 579 - 585. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Chen, S. S. Molloy, L. Thomas, J. Gambee, H. P. Bachinger, B. Ferguson, J. Zonana, G. Thomas, and N. P. Morris Mutations within a furin consensus sequence block proteolytic release of ectodysplasin-A and cause X-linked hypohidrotic ectodermal dysplasia PNAS, June 19, 2001; 98(13): 7218 - 7223. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
O. Elomaa, K. Pulkkinen, U. Hannelius, M. Mikkola, U. Saarialho-Kere, and J. Kere Ectodysplasin is released by proteolytic shedding and binds to the EDAR protein Hum. Mol. Genet., April 1, 2001; 10(9): 953 - 962. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. M. Gonzalez, C. Otey, M. Edlund, and J. C. R. Jones Interactions of a hemidesmosome component and actinin family members J. Cell Sci., January 12, 2001; 114(23): 4197 - 4206. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Kumar, M. T. Eby, S. Sinha, A. Jasmin, and P. M. Chaudhary The Ectodermal Dysplasia Receptor Activates the Nuclear Factor-kappa B, JNK, and Cell Death Pathways and Binds to Ectodysplasin A J. Biol. Chem., January 19, 2001; 276(4): 2668 - 2677. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Drogemuller, O. Distl, and T. Leeb Partial Deletion of the Bovine ED1 Gene Causes Anhidrotic Ectodermal Dysplasia in Cattle Genome Res., October 1, 2001; 11(10): 1699 - 1705. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||