- Email: [email protected]
Functional and structural diversity of the human Dickkopf gene family
Gene 238 (1999) 301–313 www.elsevier.com/locate/gene
Functional and structural diversity of the human Dickkopf gene family Valery E. Krupnik a, John D. Sharp b, Chian Jiang c, Keith Robison d, Troy W. Chickering c, Lakshmi Amaravadi c, Diane E. Brown c,1, Deborah Guyot c,2, Gail Mays c, Kevin Leiby c, Belle Chang c, Thao Duong c, Andrew D.J. Goodearl c, David P. Gearing c, Sergei Y. Sokol a, Sean A. McCarthy c, * a Department of Microbiology and Molecular Genetics, Harvard Medical School and Division of Molecular Medicine, Beth Israel Deaconess Medical Center, 330 Brookline Avenue, Boston, MA 02215, USA b Division of Research Technologies and Proteins, Lilly Research Laboratory, Indianapolis, IN 46285, USA c Millennium BioTherapeutics Inc., 640 Memorial Drive, Cambridge, MA 02139, USA d Millennium Pharmaceuticals Inc., 640 Memorial Drive, Cambridge, MA 02139, USA Received 29 April 1999; received in revised form 23 July 1999; accepted 6 August 1999; Received by I. Verma
Abstract Wnt proteins influence many aspects of embryonic development, and their activity is regulated by several secreted antagonists, including the Xenopus Dickkopf-1 (xDkk-1) protein. xDkk-1 inhibits Wnt activities in Xenopus embryos and may play a role in induction of head structures. Here, we characterize a family of human Dkk-related genes composed of Dkk-1, Dkk-2, Dkk-3, and Dkk-4, together with a unique Dkk-3 related protein termed Soggy (Sgy). hDkks 1–4 contain two distinct cysteine-rich domains in which the positions of 10 cysteine residues are highly conserved between family members. Sgy is a novel secreted protein related to Dkk-3 but which lacks the cysteine-rich domains. Members of the Dkk-related family display unique patterns of mRNA expression in human and mouse tissues, and are secreted when expressed in 293T cells. Furthermore, secreted hDkk-2 and hDkk-4 undergo proteolytic processing which results in cleavage of the second cysteine-rich domain from the full-length protein. Members of the human Dkk-related family differ not only in their structures and expression patterns, but also in their abilities to inhibit Wnt signaling. hDkk-1 and hDkk-4, but not hDkk-2, hDkk-3 or Sgy, suppress Wnt-induced secondary axis induction in Xenopus embryos. hDkk-1 and hDkk-4 do not block axis induction triggered either by Xenopus Dishevelled ( Xdsh) or Xenopus Frizzled-8 ( Xfz8), both of which function to transduce signals from Wnt ligands. Thus, hDkks 1 and 4 may inhibit Wnt activity by a mechanism upstream of Frizzled. Our findings highlight the structural and functional heterogeneity of human Dkk-related proteins. © 1999 Elsevier Science B.V. All rights reserved. Keywords: Antagonist; Frizzled related protein; In situ hybridization; Secreted protein; Soggy; Wingless; Xenopus
1. Introduction The orchestration of early embryonic development requires the concerted actions of many secreted signaling Abbreviations: Cys-1, cysteine-rich domain-1; Cys-2, cysteine-rich domain-2; Dkk, Dickkopf; EST, expressed sequence tag; FRP, Frizzled-related protein; ISH, in situ hybridization; PCR, polymerase chain reaction; Sgy, Soggy; Wg, Wingless; Xdsh, Xenopus Dishevelled; Xfz8, Xenopus Frizzled-8. * Corresponding author. Tel.: +1-617-679-7157; fax: +1-617-374-7653. E-mail address: [email protected] (S.A. McCarthy) 1 Present address: Lilly Research Laboratory, Greenfield, IN 46140, USA. 2 Present address: Pfizer Central Research Division, Eastern Point Rd., Groton, CT 06340, USA.
molecules. The Wnt gene family comprises a large class of secreted proteins related to the Int1/Wnt1 protooncogene and Drosophila wingless ( Wg). Wnts are expressed in a variety of tissues and organs, and are required for many developmental processes, including segmentation in Drosophila, endoderm development in Caenorhabditis elegans, establishment of limb polarity, neural crest differentiation, kidney morphogenesis, sex determination and brain development in mammals (Parr and McMahon, 1994; Cadigan and Nusse, 1997; Wodarz and Nusse, 1998). Recent studies have led to the identification of several components of the Wnt signal transduction pathway in responding cells. Wnt signals are transduced by the Frizzled (Fz) family of seven transmembrane domain receptors (Bhanot et al., 1996). The
0378-1119/99/$ – see front matter © 1999 Elsevier Science B.V. All rights reserved. PII: S0 3 7 8 -1 1 1 9 ( 9 9 ) 0 0 36 5 - 0
302
V.E. Krupnik et al. / Gene 238 (1999) 301–313
resulting signal leads to the activation of the cytoplasmic protein Dishevelled (Dsh) and stabilization of the Armadillo/b-catenin protein (Perrimon, 1994). Negative regulators of the Wnt pathway include glycogen synthase kinase 3 (GSK3)/shaggy (Perrimon, 1994), the tumor suppressor gene product APC (adenomatous polyposis coli; Gumbiner, 1997) and Axin ( Zeng et al., 1997). In the absence of a Wnt ligand, these regulators promote phosphorylation and consequential degradation of b-catenin. Conversely, Wnt signaling inactivates GSK3 and therefore stabilizes b-catenin, which translocates to the nucleus and forms a complex with TCF transcription factors, thereby regulating target gene expression. Deregulation of this pathway can lead to carcinogenesis, emphasizing the long-recognized connection between Wnts, normal development and cancer (Cadigan and Nusse, 1997). This connection has been strengthened recently with the identification of the c-Myc protooncogene as a target of Wnt signaling (He et al., 1998). While the outcome of Wnt signaling may be influenced by multiple intracellular regulatory mechanisms, recent studies have also identified several classes of secreted factors which modulate Wnt action outside of the cell. These include a family of secreted Frizzledrelated proteins (FRPs; Zorn, 1997; Moon et al., 1997; Wodarz and Nusse, 1998), Cerberus (Piccolo et al., 1999) and Wnt Inhibitory Factor-1 ( WIF-1; Hsieh et al., 1999a). FRPs are structurally related to the extracellular domains of Frizzled proteins, particularly in conserved cysteine-rich domains (Moon et al., 1997). Cerberus is a multifunctional inhibitor of BMP, Nodal and Wnt signaling (Piccolo et al., 1999). WIF-1 possesses a unique structure composed of an amino terminal WIF domain and 5 carboxy terminal epidermal growth factor repeats ( Hsieh et al., 1999a, b). WIF-1 and certain FRPs inhibit Wnt function by direct binding of the ligand, thus preventing access to cell surface receptors (Hsieh et al., 1999a, b; Wang et al., 1997; Leyns et al., 1997; Rattner et al., 1997; Bafico et al., 1999). The Dickkopf (Dkk) family comprises another discrete class of secreted Wnt inhibitors. xDkk-1 was identified in a screen for factors capable of inducing Xenopus head formation in the context of inhibition of bone morphogenetic protein signaling (Glinka et al., 1998). xDkk-1 is expressed in the Spemann organizer, which is responsible for induction of head structures, and is therefore a candidate for the endogenous signal for head induction (Glinka et al., 1998). Dkk-1 inhibits the axis-inducing activity of Xwnt8, consistent with the suggestion that simultaneous inhibition of BMP and Wnt signaling is required for head formation. xDkk-1 does not inhibit secondary axis induction by Xdsh, however. Thus, like the FRPs and WIF-1, xDkk-1 is most likely to act as an extracellular factor that antagonizes Wnt function directly. A murine homolog of xDkk-1 also inhibits Xwnt8-mediated axis duplication
in Xenopus embryos. Furthermore, Dkk-1 appears to be the prototypical member of a family of Dkk-related genes (Glinka et al., 1998). Here, we describe the characterization of a family of four human Dkk-related proteins and a unique protein, Soggy, which possesses homology to hDkk-3 but not other Dkks. Our data demonstrate structural and functional heterogeneity within this novel class of secreted proteins.
2. Materials and methods 2.1. Isolation of mammalian Dkk cDNAs Human Dkk-related proteins were identified by searching of internal and public expressed sequence tag ( EST ) databases with the sequence of hDkk-3 using the TBLASTN algorithm ( WashUversion 2.0, BLOSUM62 search matrix). From this search, four distinct partial protein sequences were identified which displayed significant homology to human Dkk-3. The full-length cDNA sequence of hDkk-1 was determined by complete sequencing of an individual clone from a human fetal kidney cDNA library which was identified within an internal EST database. A full-length hDkk2 cDNA was isolated from a human fetal lung cDNA library by plaque hybridization (Sambrook et al., 1989) using a partial cDNA obtained from the IMAGE collection of cDNA clones as a probe (GenBank accession No. W55979). A full-length cDNA clone for hDkk-3 was isolated from a human fetal brain cDNA library by plaque hybridization using a partial cDNA identified by a Signal Peptide Trap procedure as a probe (manuscript in preparation). The full-length cDNA sequence of mDkk-3 was determined by complete sequencing of an individual clone identified within a murine adult brain cDNA library which was identified within an internal EST database. The full-length cDNA sequence for hDkk-4 was determined by complete sequencing of a partially sequenced cDNA clone from the IMAGE collection (GenBank accession No. AA565546). Fulllength cDNA sequences for hSoggy and mSoggy were obtained by complete sequencing, respectively, of IMAGE clones with GenBank accession Nos. AA397836 and AA497886. All IMAGE collection clones were purchased from Research Genetics. cDNA libraries for screening were prepared using the Gibco Superscript cDNA synthesis kit and the lZiplox vector (Gibco). Probes were labeled with [a-32P]dCTP (Prime-It, Stratagene). DNA sequencing was performed using automated ABI377 sequence detectors and dye terminator chemistry. 2.2. Analysis of mRNA expression For Northern analysis, all hybridizations were to Clontech Multiple Tissue Northern Blots and were
V.E. Krupnik et al. / Gene 238 (1999) 301–313
performed in ExpressHyb solution (Clontech) for 1– 20 h. All cDNA probes were prepared by random primed radiolabeling with [a-32P]dCTP (Prime-It, Stratagene). Blots were washed sequentially to a final stringency of 0.2×SSC/0.2% SDS and exposed to autoradiographic film. Hybridizations of a control b-actin cDNA probe consistently demonstrated even loading of the Northern blots (not shown). For in situ hybridization (ISH ), normal mouse embryos and adult mouse tissues were collected from C57BL/6 mice, embedded in TissueTek O.C.T Compound (Sakura Finetek USA, Torrance, CA) and frozen on dry ice. Cryostat serial sections (8 mm) were thaw mounted on Superfrost Plus slides ( VWR) and air dried on a slide warmer at 40°C for 20 min. Sections were then fixed with 4% formaldehyde in DEPC-treated 0.1 M phosphatebuffered saline (DEPC PBS, pH 7.5) at room temperature for 10 min and rinsed twice in DEPC PBS. Sections were rinsed in 0.1 M triethanolamine–HCl ( TEA, pH 8.0), incubated in 0.25% acetic anhydride–TEA for 10 min and rinsed in DEPC 2×SSC. Sections were dehydrated through a series of graded ethanols, incubated in 100% chloroform for 5 min, rinsed in 100% and 95% ethanol for 1 min and air dried. Labeling and hybridizations of radiolabeled antisense and sense RNA probes were performed as described previously (Busfield et al., 1997). Primers for template amplification were as follows: mDkk-3 forward, 5∞-AACAAACATGCAGCGGCTCG-3∞; reverse, 5∞-TGTGATAGTTGGGAGGTAAG-3∞; mSoggy forward, 5∞-ACCTGCAATGTGTCGACTGAG-3∞; reverse 5∞-CACTTACAGCTGTTGGGATG-3∞. 2.3. Protein expression Flag epitope-tagged forms of hDkk-1, hDkk-2, hDkk-3, hDkk-4 and hSgy cDNAs were constructed by PCR and ligated to the expression plasmid pM7-stop (S. McCarthy, unpublished ). The full-length open-reading frame of each clone was PCR amplified using a 5∞ primer incorporating a Kozak consensus sequence prior to the initiator methionine and a 3∞ primer including the nucleotide sequence encoding the flag epitope (DYKDDDDK ) followed by a termination codon. All DNA constructs were sequenced fully to ensure fidelity of PCR amplification. Constructs were transiently transfected into human embryonic kidney 293T cells using Lipofectamine (Gibco) according to the manufacturer’s instructions. 72 h post-transfection, serum-free conditioned media (OptiMEM, Gibco/BRL) were harvested, centrifuged and filtered. Cell lysates were prepared using lysis buffer (50 mM Tris pH 8.0, 150 mM NaCl, 1% NP-40, 0.05% SDS) supplemented with Complete Protease Inhibitor cocktail (Boehringer Mannheim). For Western analysis, samples of conditioned media were electroblotted onto PVDF (Novex)
303
after SDS–PAGE on 4–20% gradient gels, probed with M2 anti-flag antibody (1:500, Sigma) followed by HRP conjugated sheep anti-mouse IgG (1:5000, Amersham), developed with chemiluminescent reagents ( Renaissance, Dupont) and exposed to Biomax MR2 film ( Kodak). 2.4. Deglycosylation Samples of 293T cell supernatants collected 72 h after transfection with the appropriate Dkk expression plasmids were incubated with anti-flag M2 agarose beads (Sigma) for 16 h at 4°C. Beads were washed with PBS (pH 7.4) containing, sequentially, 0.1%, 0.05% and 0.01% Triton X-100. The beads were resuspended in incubation buffer (20 mM sodium phosphate, pH 7.5, 50 mM EDTA, 0.02% sodium azide) containing 0.5% SDS, 5% 2-mercaptoethanol and boiled for 2 min. Equal amounts of each immunoprecipitate were diluted with incubation buffer containing 1% NP-40 and then treated either with N-glycanase (Oxford Glycosystems) in enzyme buffer (20 mM Tris–HCl, 1 mM EDTA, 50 mM NaCl, 0.02% sodium azide pH 7.5) or with enzyme buffer alone as control for 18 h at 37°C. Samples were analyzed by reducing SDS–PAGE and Western blotting with anti-flag M2 antibody. 2.5. Affinity purification and N-terminal sequencing Flag-tagged hDkk-4 protein was purified by passing conditioned media over an anti-flag(M2)biotin/streptavidin Poros column (2.1×30 mm, PE Biosystems); the column was then washed with PBS (pH 7.4), and flag-tagged protein eluted with 200 mM glycine (pH 3.0). Eluted fractions were analyzed by SDS–PAGE and Western blot. Purified hDkk-4 protein bound to PVDF membrane was subjected to N-terminal amino-acid analysis on a PE Applied Biosystems Model 494 Procise instrument using Edman-based chemistry ( Hewick et al., 1981). The amino-acid residues were analyzed by HPLC (Spherogel micro PTH 3-micron column) and determined by separation and peak height as compared to standards. 2.6. Xenopus embryo culture and RNA microinjections Oocytes were obtained from Xenopus females injected with 700 units of human chorionic gonadotropin, fertilized in vitro and cultured in 0.1×MMR (Newport and Kirschner, 1982). Embryonic stages were determined according to Nieuwkoop and Faber (1967). All cDNAs were subcloned into the vector pCS2 and capped mRNAs were synthesized in vitro using the Message Machine kit (Ambion). Constructs for Xwnt8 (Christian et al., 1991), Xwnt2b (Landesman and Sokol, 1997), Xwnt3a ( Wolda et al., 1993), Xfz8 (Itoh et al., 1998)
304
V.E. Krupnik et al. / Gene 238 (1999) 301–313
and Xdsh (Sokol et al., 1995) have been described previously. Protein expression from all pCS2–Dkk constructs was confirmed by in vitro transcription and translation ( TNT, Promega). For tests of secondary axis induction, a single ventral blastomere of four- or eightcell embryos was injected with 10 nl of a solution containing 2–4 pg of Xwnt8 mRNA, 2.5–5 pg of Xwnt3a mRNA, 10 pg of Xwnt2b mRNA, 2 ng Fz8 mRNA or 1 ng Xdsh mRNA as described (Itoh et al., 1995). 2.5 ng of hDkk or hSgy mRNAs were co-injected as indicated. After injections, embryos were cultured in 3% Ficoll 400 (Pharmacia), 0.5×MMR. Secondary axes were scored at stage 35 as complete when they contained anterior neuroectodermal derivatives including pronounced cement gland and eyes. Axes were scored as partial when the secondary neural tube with melanocytes was apparent but head structures were not.
3. Results and discussion 3.1. Characterization of the human Dkk-related gene family Fig. 1A depicts schematically the structures of the family of human Dkk-related proteins that we have characterized in the present study. The human Dkk (hDkk) family comprises hDkk-1, hDkk-2, hDkk-3, hDkk-4, and a unique Dkk-3-related protein we refer to as Soggy (Sgy). hDkk-1 is the full-length human homolog of previously reported Xenopus and mouse Dkk-1 (Glinka et al., 1998). hDkk-2 and hDkk-3 correspond to previously described partial hDkk-2 and hDkk-3 sequences (Glinka et al., 1998). hDkk-4 and Sgy represent novel members of this gene family. Sequence identities between the various Dkk proteins described here and elsewhere (Glinka et al., 1998) are summarized in Fig. 1B. As described previously for Xenopus Dkk-1, human Dkks 1–4 each possess an N-terminal signal peptide and contain two conserved cysteine-rich domains (Cys-1 and Cys-2) separated by a linker region (Fig. 1A, C ). Each domain possesses 10 conserved cysteine residues. The positions of the 10 conserved cysteines in Cys-2 are closely similar to those in proteins of the colipase family (Fig. 1D; Aravind and Koonin, 1998). Furthermore, sequence conservation among the Dkks is greatest within Cys-2, suggestive of a conserved function. Colipases facilitate interactions of pancreatic lipases with lipid micelles, and for this reason it has been suggested that the Cys-2 domain of the Dkks may enable the proteins to interact with lipids in order to regulate Wnt function (Aravind and Koonin, 1998). Wnt proteins are known to remain tightly associated with the cell surface (Smolich et al., 1993), and a putative lipid binding function may facilitate Wnt/Dkk interactions at the
Fig. 1. The human Dkk family. (A) Schematic illustration of the human Dkk/Sgy family showing signal peptides (darkened boxes), Cysteine-rich domains (Cys-1 and Cys-2) and putative sites of N-linked glycosylation in the human Dkks (branches). (B) Percent identities between full-length protein sequences for hDkks 1–4, mDkk-1 (AF030433), mDkk-3, cDkk-3 (D26311) and xDkk-1 (AF030434). Figures were generated using the Smith–Waterman algorithm as implemented in the program Bestfit of the GCG package, with gap penalties of 8 for opening and 1 for extending. All 20 cysteines were aligned in all but five of the 28 comparisons, in which the final cysteine was mismatched. (C ) Multiple sequence alignment of different Dkk proteins. Alignments were performed with the ClustalW algorithm as implemented in the GCG program PILEUP. Predicted signal peptides (Nielsen et al., 1997) are underlined, putative N-glycosylation sites are indicated by a thick bar, Cys-1 by an open box, Cys-2 by a shaded box. The proteolytic cleavage site within hDkk-4 is indicated by an arrow. (D) Multiple sequence alignment of hDkks with human colipase (J02883). Cys-2 of each hDkk is shown and indicated by the shaded box. The human colipase sequence is shown missing eight amino acids from the N-terminus of the mature peptide. The alignment was made using gap penalties of 12 for opening and 2 for extending. A minor adjustment was necessary, since PILEUP inserts a single gap in hDkk-1 and hDkk-2 between GS at position 56–57, even with gap opening penalty of 15. hDkk-1 is shown from amino acid 181. (E ) Multiple sequence alignment of Dkk-3 and Soggy proteins. Signal peptides and N-glycosylation sites are indicated as in (C ). Cys-1 (open box) and Cys-2 (shaded box) within Dkk-3 are shown for reference. Asterisks indicate amino-acid identities between all four proteins. Soggy protein sequences terminate at #.
plasma membrane. Interestingly, sequence similarity between Wnt-1 and the lipid binding domain of secreted phospholipase A2 has recently been noted, further suggesting that Wnts may interact directly with lipids ( Reichsman et al., 1999). In contrast to the conserved cysteine-rich domains, the linker region between Cys-1 and Cys-2 is highly variable between hDkks and it is notably larger in hDkk-1, -2 and -4 (50–55 aa) compared with Dkk-3 (12 aa). Four potential sites of N-linked glycosylation in hDkk-3 are conserved in chick and
V.E. Krupnik et al. / Gene 238 (1999) 301–313
305
Fig. 1. (continued).
mouse Dkk-3 (Fig. 1A, C ). These sites are not conserved in other Dkk family members. hDkk-1 possesses one potential N-glycosylation site located close to the
C-terminus of the protein which is conserved in murine Dkk-1 but not in Xenopus Dkk-1 (Fig. 1C ). In addition, each hDkk possesses several potential sites of proteolytic
306
V.E. Krupnik et al. / Gene 238 (1999) 301–313
Fig. 1. (continued).
cleavage by furin-type proteases (Nakayama, 1997), suggesting that the proteins may be subject to posttranslational proteolytic processing. hDkk-3 is the most divergent of the four human Dkks and possesses an extended N-terminal domain which precedes the Cys-1 domain and an extended C-terminal region that follows Cys-2 (Fig. 1A, C ). Sgy is a novel protein with sequence similarity to Dkk-3 but not other Dkks (Fig. 1A, E). Human and murine Sgy cDNAs encode proteins of 242 aa and 230 aa, respectively, and are predicted to be secreted owing to the presence of N-terminal signal peptides. hDkk3 and hSgy display 22% amino-acid identity, whereas human and murine Sgy proteins display 59% overall identity. Human and mouse Sgy proteins each possess two sites of potential N-linked glycosylation which are conserved with Dkk-3 ( Fig. 1E). Sequence homology between Dkk-3 and Sgy is most pronounced within the N-terminal domain of Dkk-3, although significant amino-acid identities are seen beyond this domain that extend into the cysteine-rich domains of Dkk-3 (Fig. 1E). However, cysteine residues are not conserved
within these domains and the residues shared by Sgy and Dkk-3 are poorly conserved in other Dkks. Thus, Sgy is uniquely related to Dkk-3. 3.2. mRNA expression analysis To investigate further the relationship between different mammalian Dkks, we analyzed their tissue distribution by Northern analysis. Fig. 2 shows that mammalian Dkks are expressed in distinct tissues. A 1.8 kb hDkk-1 mRNA was detected in human placenta, but not in other tissues tested ( Fig. 2A). hDkk-2 transcripts of approx. 4 and 4.5 kb were detected in multiple tissues, including heart, brain, skeletal muscle and lung ( Fig. 2B). A 2.5 kb hDkk-3 transcript was also widely expressed in human tissues, although levels of this mRNA were substantially higher in heart, brain and spinal cord relative to other tissues tested ( Fig. 2C, D). hDkk-4 mRNA was undetectable in all adult and fetal human tissues examined by Northern analysis, although a survey of a human cDNA library panel by PCR with hDkk-4-specific PCR primers generated products from
Fig. 2. Tissue distribution of Dkk and Soggy mRNAs. Multiple tissue Northern (MTN ) blots hybridized with radiolabeled cDNA probes specific to (A) hDkk-1, (B) hDkk-2, (C, D) hDkk-3, and ( E, F ) mSgy. All MTN blots were from Clontech. (A, C ) Human MTN-1; (B) human 12-lane MTN; (D) human MTN-III; ( E) mouse MTN-1; (F ) whole mouse embryo. Sizes of molecular weight markers are indicated.
V.E. Krupnik et al. / Gene 238 (1999) 301–313 307
308
V.E. Krupnik et al. / Gene 238 (1999) 301–313
libraries prepared from cerebellum, activated human T-lymphocytes, lung and esophagus (data not shown). A 1 kb Soggy mRNA was expressed in mouse testis, but not in other adult murine tissues tested (Fig. 2E). Interestingly, mSgy mRNA was also expressed between days 11 and 15 of gestation in mouse embryos (Fig. 2F ). Thus, mSgy displays temporal regulation during mouse development. We also investigated the expression of murine Dkk-3 and murine Sgy by in situ hybridization (ISH ). Fig. 3 shows the expression pattern of mDkk-3 in the adult murine brain, eye, and heart. mDkk-3 expression in the brain was found in neurons of the cortex and hippocampus (Fig. 3B). In adult eye, mDkk-3 mRNA was found to be highly expressed in the retina, ciliary body, and lens epithelium ( Fig. 3E). Expression in the retina was limited to the integrating bipolar neurons and ganglion cells. Consistent with our findings, Dkk-3 mRNA is also expressed at high levels in the chick eye (Sawada et al,
1996). In adult heart, mDkk-3 was detected in the atrioventricular valves and also in myocytes of the atria. Expression was highly restricted to the atria and noticeably absent from ventricular tissue ( Fig. 3H ). We also observed expression of mDkk-3 mRNA in developing eye, bone and cartilage in day 14 embryos (not shown). Fig. 4 details the expression of murine Sgy analyzed by ISH. Consistent with Northern analysis, mSgy mRNA was expressed at high levels in adult testis ( Fig. 4A) in the spermatogenic epithelium of the seminiferous tubules and the spermatogonia at various stages of development. We also analyzed a series of sagittal sections of mouse embryos from E13.5 to E17.5. In E15.5 embryos, Sgy mRNA transcripts were localized to the developing dorsal root ganglia and also in the cartilage primordium of the nasal septum ( Fig. 4C ). Soggy expression was also seen in the eye from E13.5 to E16.5, as we observed for mDkk-3 (Fig. 4E). Expression of Sgy mRNA at these stages of development suggests that this novel
Fig. 3. Expression of mDkk-3 in adult mouse tissues. In situ hybridization analysis of mDkk-3 mRNA expression in adult murine tissues. (A–C ) Brain, parasagittal section; (D–F ) eye; (G–I ) heart, parasagittal section. (A, D, G) Bright field image of hematoxylin and eosin stained sections; (B, E, H ) antisense probe; (C, F, I ) sense probe. Scale bars in the bright field images shown in (A, D, and G) correspond to 1 mm.
V.E. Krupnik et al. / Gene 238 (1999) 301–313
Fig. 4. Expression of Sgy in embryonic and adult mouse tissues. In situ hybridization analysis of mSoggy mRNA expression in: (A, B) mouse testis, (C, D) E15.5 embryo sagittal sections, and (E, F ) E.16.5 embryo sagittal sections. (A, C, E ) Antisense probe; (B, D, F ) sense probe. Scale bar in (A) corresponds to 0.5 mm, and in (C, E ) to 1 mm.
secreted protein may play multiple roles in during embryogenesis. Thus, Dkks 1–4 and Sgy display distinct patterns of mRNA expression in human and murine tissues. Similarly, different Wnt family members have divergent patterns of mRNA expression in adult and embryonic mammalian tissues. For example, murine Wnts-4, -7a and -7b are expressed in brain and lung, whereas Wnt-6 is highly expressed in testis (Gavin et al., 1990). Wnts-5b and -13 are more broadly expressed (Gavin et al., 1990; Katoh et al., 1996). Given the large size of the mammalian Wnt family and the growing number of distinct families of Wnt inhibitors, it seems likely that tissuespecific expression of regulators such as the Dkks may afford an additional level of control to Wnt signaling. 3.3. Secretion and post-translational modification of hDkk proteins We next characterized human Dkk and Sgy proteins by transient overexpression of flag epitope-tagged forms in 293T cells (Fig. 5). Similar to Xenopus Dkk-1, hDkk-1
309
was readily detected in conditioned medium of transfected cells. Mature secreted hDkk-1 migrated with a molecular mass of approx. 42–50 kDa ( Fig. 5a, lane 3). hDkk-3 was also secreted by 293T cells and migrated as a heterogeneous band of 45–65 kDa ( Fig. 5a, lane 1). Soluble hDkk-4 was detected as three major immunoreactive species of approx. 40 kDa [form (i)], 30–32 kDa [ form (ii)] and 15–17 kDa [form (iii)]. We also observed additional minor forms of hDkk-4 in certain experiments, as shown, in the range of 20–21 kDa. Soluble hDkk-2 was detected as a major species of 15–17 kDa ( Fig. 5a, lane 4), closely similar in size to form (iii) of hDkk-4. hSgy was also efficiently secreted from transfected 293T cells and migrated with a molecular mass of approx. 40–50 kDa (Fig. 5a, lane 5). Thus, human Dkk/Sgy proteins are soluble, secreted factors with heterogeneous profiles suggestive of post-translational modification. We also studied the effect of N-glycanase treatment on the mobilities of secreted hDkk-1, hDkk-3, hDkk-4 and hSgy. Treatment with N-glycanase caused only a minor increase in the mobility of soluble hDkk-1 ( Fig. 5b, lanes 1 and 2). hDkk-1 contains only a single putative N-glycosylation site which is located towards the C-terminus of the protein. Soluble hDkk-3 protein displayed a substantial increase in mobility following Nglycanase treatment. The major 45–65 kDa form of soluble hDkk-3 was observed as two species of 45–55 and 40 kDa following deglycosylation (Fig. 5b, lanes 3 and 4). This finding is consistent with the presence of four potential sites of N-linked glycosylation in the hDkk-3 protein. The reason for the heterogeneity of deglycosylated hDkk-3 is unclear, although it may reflect either proteolytic processing or incomplete removal of carbohydrate from one or more attachment sites. We also observed a 30 kDa hDkk3 species in these experiments, the mobility of which was unaltered by Nglycanase treatment. This form was observed only after overnight incubation of the samples and may therefore be a non-specific degradation product. For hDkk-4, neither form (i), (ii) nor (iii) was significantly affected by N-glycanase treatment, consistent with the absence of N-glycosylation sites from the protein ( Fig. 5b, lanes 5 and 6). hSgy, however, displayed a 5–10 kDa decrease in apparent molecular mass after N-Glycanase treatment, consistent with the presence of two potential sites of N-glycosylation in the protein (Fig. 5b, lanes 7 and 8). Taken together, these results suggest that hDkk-3 and hSgy are expressed and secreted by 293T cells as glycoproteins. hDkk-1 displays minimal, if any, attachment of N-linked carbohydrate and hDkk-4 shows no evidence of such modification. Soluble hDkk-4 was consistently observed in three forms of approx. 40 kDa [ form (i)], 30–32 kDa [ form (ii)] and 15–17 kDa (form (iii), while soluble hDkk-2 was detected as a single C-terminal form of 15–17 kDa. We therefore analyzed processing of hDkk-4 by Edman
310
V.E. Krupnik et al. / Gene 238 (1999) 301–313
Fig. 5. hDkk protein expression and analysis. (a) Secretion of hDkk/Soggy proteins. Anti-flag Western blots are shown of conditioned media of 293T cells transiently transfected with expression vectors for the indicated flag-tagged hDkks. Lane 1: hDkk-3; lane 2: hDkk-4; lane 3: hDkk-1; lane 4: hDkk-2; lane 5: hSgy. For hDkk-4, bands (i), (ii) and (iii) represent the three forms of the tagged protein consistently observed in cell supernatants. Note the similarity in size of hDkk-4 form (iii) and the major form of hDkk-2 detected by this analysis ( lane 4). (b) Effect of Nglycanase treatment on the mobilities of hDkk/Soggy proteins on SDS–PAGE. Immunoprecipitated hDkk/Soggy proteins were incubated overnight either with buffer alone (−) or N-glycanase (+) and then separated on reducing SDS–PAGE. The proteins were analyzed by Western analysis with anti-flag antibodies. Lanes 1, 2: hDkk-1; lanes 3, 4: hDkk-3; lanes 5, 6: hDkk-4; lanes 7, 8: hSgy. (c) N-Terminal sequence analysis of hDkk-4 protein species. N-Terminal sequences of affinity purified material corresponding to hDkk-4 forms (i) and (iii) are shown. (d ) Effect of reducing agent (b-mercaptoethanol ) on mobility of soluble hDkk-4 species. Anti-flag Western blots of samples of conditioned media containing either hDkk-1 ( lanes 1 and 3) or hDkk-4 ( lanes 2 and 4) after separation under non-reducing or reducing SDS–PAGE, as indicated.
N-terminal sequencing of anti-flag affinity purified material corresponding to bands (i), (ii) and (iii) ( Fig. 5c). The N-terminal sequence of band (i) was found to be XVLDFNNIRS, which corresponds to the sequence that follows the predicted signal peptide cleavage site (between Ala-18 and Leu-19). Because the same band is identified by anti-flag antibodies, which recognize the C-terminal epitope tag, band (i) was thus identified as mature hDkk-4 protein with the signal sequence removed. The band (iii) N-terminal sequence was found to be SQGRKGQEGS, which corresponds to the Cys-2 domain cleaved at the dibasic site Lys132/Lys133 (Lys 114/Lys115 of mature hDkk4). We were unable to obtain conclusive data concerning the N-terminal sequence of band (ii), owing to low abundance of purified material. However, the data obtained for bands (i) and (iii) indicate clearly that hDkk-4 is proteolytically processed by 293T cells, potentially by furin type proteases (Nakayama, 1997). By inference from the hDkk-4
protein sequences, it is likely that the major 15–17 kDa form of hDkk-2 detected in these experiments also corresponds to the Cys-2 domain, although the precise site of cleavage remains to be determined. We also investigated whether the mobilities of hDkk-4 bands (i), (ii) and (iii) were altered on non-reducing SDS–PAGE. The three major species migrated similarly on SDS– PAGE conducted under either reducing and non-reducing conditions ( Fig. 5d). Thus, each of the major C-terminal (anti-flag immunoreactive) hDkk-4 species appear to exist as independent proteolytic fragments that are not covalently linked via disulfide bonds to other subunits or proteins when secreted from 293T cells. 3.4. Effects of hDkks on Wnt-induced axis duplication in Xenopus embryos The studies described above establish that Dkks are secreted proteins and that each Dkk displays a unique
V.E. Krupnik et al. / Gene 238 (1999) 301–313
pattern of proteolytic processing and N-linked glycosylation. We next analyzed the functional activities of human Dkk-related proteins. It has been shown previously that murine and Xenopus Dkk-1 block the ability of Xwnt8 to induce a secondary axis in the frog embryo (Glinka et al., 1998). We evaluated the activities of different hDkks in this assay. hDkk mRNAs were co-injected with Xwnt8 mRNA into single ventral blastomeres of four- or eight-cell embryos. Injected embryos were cultured for 2 days and frequencies of embryos with secondary axes were determined. As previously described (Christian et al., 1991), Xwnt8-injected embryos displayed complete axis duplication (Fig. 6A). Xwnt8-mediated axis duplication was inhibited by co-injection with mRNAs encoding hDkk-1 ( Fig. 6B) and hDkk-4 (Fig. 6F ). Injection of mRNAs encoding hDkk-2, hDkk-3 or hSgy had no effect on Xwnt8-induced axis duplication (Fig. 6C–E). We also investigated the effects of hDkks and hSgy on secondary axis induction by Xwnt3a and Xwnt2b. As with Xwnt8, hDkk-1 and hDkk-4 inhibited axis duplication in
311
response to these Wnt proteins whereas hDkk-3 and hSgy had no effect ( Fig. 7). To investigate the mechanism by which hDkk-1 and hDkk-4 inhibited Wnt function, we co-injected Dkk mRNAs with Xdsh, a downstream component of the Wnt signaling pathway (Sokol et al., 1995, 1996). As shown in Fig. 7, hDkks 1 and 4 did not block secondary axis induction by Xdsh, indicating that Dkks function upstream of, or in parallel with, Xdsh signaling. Similar findings have been reported previously for xDkk-1 (Glinka et al., 1998). We also tested whether hDkks could antagonize signaling by Xenopus Frizzled-8 ( Xfz8), a cell-surface Wnt receptor which can also induce secondary axes when ectopically expressed (Itoh et al., 1998). Neither hDkk-1 or hDkk-4 inhibited the axis-inducing activity of Xfz8 mRNA (Fig. 7). Taken together with the fact that hDkk-1 and hDkk-4 are secreted, these data suggest that Dkks antagonize Wnt function at a point upstream of Wnt receptors. The present study describes a family of five human Dkk-related proteins which display marked differences
Fig. 6. Effect of hDkks and Sgy on Wnt-induced axis duplication in Xenopus embryos. Effect of hDkks and hSgy on Xwnt8-induced secondary axis formation in Xenopus embryos. Four or eight cell stage embryos were injected subequatorially into a ventral blastomere with mRNAs for: (A) Xwnt8, (B) Xwnt8+hDkk-1, (C ) Xwnt8+hDkk-2, (D) Xwnt8+hDkk-3, ( E ) Xwnt8+hSgy, ( F ) Xwnt8+hDkk-4. Four representative embryos from multiple injections are shown at stage 35–37 in each panel.
312
V.E. Krupnik et al. / Gene 238 (1999) 301–313
AF177395; hDkk-3 — AF177396; hDkk-4 — AF177397; h-Soggy — AF177398; m-Soggy — AF177399; mDkk-3 — AF177400.
Acknowledgements We thank R. Moon for plasmids. We also thank Rick Clark, John Keilty, Betty Woolf, Pam Brauer, Donna Michnick, Kevin McDonald and members of the Millennium DNA sequencing, cDNA library and Full Length cDNA Cloning Groups for skilled assistance and advice. We are grateful to Joe Weber and Bill Holmes for critical comments on the manuscript and Chuck Gray for assistance with collaborations. This work was supported in part by grants from the March of Dimes Birth Defect Foundation and the National Institutes of Health to S.S. Fig. 7. Effect of human Dkk proteins on Wnt, Dsh and Fz-induced axis duplication. Four- or eight-cell stage embryos were injected subequatorially into a ventral blastomere with mRNAs for Xwnt-8, Xwnt-2b, Xwnt-3a, Xfz8 or Xdsh in the presence or absence of mRNAs for either hDkk-1, hDkk-3, hDkk-4 or hSgy as indicated. Embryos were allowed to develop until stage 33–37, then scored for secondary axis formation. Complete axes included eyes and cement glands; partial axes included partial neural tubes and hindbrain, but lacked fore/midbrain structures. n=number of embryos in the experiment; open bars, % of embryos with complete secondary axes; filled bars, % of embryos with partial secondary axes.
in sequence, mRNA expression patterns, post-translational modification and in their activities as inhibitors of Wnt signaling. The restricted mRNA expression patterns of different Dkk family members suggest that these proteins play diverse roles in embryonic and adult tissues. Further work is required to determine the mechanism by which hDkk-1 and hDkk-4 inhibit Wnt signaling, and also to investigate the relationships between post-translational processing and Wnt inhibition. The ability of hDkks to inhibit Wnt signaling was inversely correlated with the degree of proteolytic processing we observed in 293T cells, suggesting that proteolytic mechanisms may regulate Dkk activities. Recent studies indicate that hDkk-1 inhibits signaling and morphological changes induced in NIH3T3 cells by mammalian Wnt-2 (Fedi et al., 1999). It will be of interest to characterize the activities of the mammalian Dkk family in such assays, as well as to establish structure–function relationships and biochemical mechanisms of action. Recent advances in the biochemical analysis of Wnt proteins should facilitate these studies (Hsieh et al., 1999a, b).
4. Note added in proof The Genbank accession numbers of the sequences listed in this paper are: Dkk-1 — AF177394; hDkk-2 —
References Aravind, L., Koonin, E.V., 1998. A colipase fold in the carboxy-terminal domain of the Wnt antagonists — the Dickkopfs. Curr. Biol. 8 (477), 478–1997. Bafico, A., Gazit, A., Pramila, T., Finch, P.W., Yaniv, A., Aaronson, S.A., 1999. Interaction of Frizzled Related Protein ( FRP) with Wnt ligands and the Frizzled receptor suggests alternative mechanisms for FRP inhibition of Wnt signaling. J. Biol. Chem. 274, 16180–16187. Bhanot, P., Brink, M., Samos, C.H., Hsieh, J.C., Wang, Y., Macke, J.P., Andrew, D., Nathans, J., Nusse, R., 1996. A new member of the frizzled family from Drosophila functions as a Wingless receptor. Nature 382, 225–230. Busfield, S.J., Michnick, D.A., Chickering, T.W., Revett, T.L., Ma, J., Woolf, E.A., Comrack, C.A., Dussault, B.J., Woolf, J., Goodearl, A.D., Gearing, D.P., 1997. Characterization of a neuregulin-related gene, Don-1, that is highly expressed in restricted regions of the cerebellum and hippocampus. Mol. Cell. Biol. 17, 4007–4014. Cadigan, K.M., Nusse, R., 1997. Wnt signaling: a common theme in animal development. Genes Dev. 11, 3286–3305. Christian, J.L., McMahon, J.A., McMahon, A.P., Moon, R.T., 1991. Xwnt8, a Xenopus Wnt1/int-1-related gene responsive to mesoderm inducing growth factors, may play a role in ventral mesodermal patterning during embryogenesis. Development 111, 1045–1055. Fedi, P., Bafico, A., Nieto Soria, A., Burgess, W.H., Miki, T., Bottaro, D.P., Kraus, M.H., Aaronson, S.A., 1999. Isolation and biochemical characterization of the human Dkk-1 homologue, a novel inhibitor of mammalian Wnt signaling. J. Biol. Chem. 274, 19465–19472. Gavin, B.J., McMahon, J.A., McMahon, A.P., 1990. Expression of multiple novel Wnt-1/int-1-related genes during fetal and adult mouse development. Genes Dev. 4, 2319–2332. Glinka, A., Wu, W., Delius, H., Monaghan, A.P., Blumenstock, C., Niehrs, C., 1998. Dickkopf-1 is a member of a new family of secreted proteins and functions in head induction. Nature 391, 357–362. Gumbiner, B.M., 1997. Carcinogenesis: a balance between beta-catenin and APC. Curr. Biol. 7, R443–R436. He, T.C., Sparks, A.B., Rago, C., Hermeking, H., Zawel, L., da Costa, L.T., Morin, P.J., Vogelstein, B., Kinzler, K.W., 1998. Identifica-
V.E. Krupnik et al. / Gene 238 (1999) 301–313 tion of c-MYC as a target of the APC pathway. Science 281, 1509–1512. Hewick, R.M., Hunkapiller, M.W., Hood, L.E., Dreyer, W.J., 1981. A gas-liquid solid phase peptide and protein sequenator. J. Biol. Chem. 256, 7990–7997. Hsieh, J., Kodjabachian, L., Rebbert, M.L., Rattner, A., Smallwood, P.M., Harryman Samos, C., Nusse, R., Dawid, I.B., Nathans, J., 1999a. A new secreted protein that binds to Wnt proteins and inhibits their activities. Nature 398, 431–436. Hsieh, J., Rattner, A., Smallwood, P.M., Nathans, J., 1999b. Biochemical characterization of Wnt–Frizzled interactions using a soluble, biologically active vertebrate Wnt protein. Proc. Natl. Acad. Sci. USA 96, 3546–3551. Itoh, K., Jacob, J., Sokol, S.Y., 1998. A role for Xenopus Frizzled 8 in dorsal development. Mech. Dev. 74, 145–157. Itoh, K., Tang, T.L., Neel, B.G., Sokol, S.Y., 1995. Specific modulation of ectodermal cell fates in Xenopus embryos by glycogen synthase kinase. Development 121, 3979–3988. Katoh, M., Hirai, M., Sugimura, T., Terada, M., 1996. Cloning, expression and chromosomal localization of Wnt-13, a novel member of the Wnt gene family. Oncogene 13, 873–876. Landesman, Y., Sokol, S.Y., 1997. Xwnt-2b is a novel axis-inducing Xenopus Wnt, which is expressed in embryonic brain. Mech. Dev. 61, 1199–1209. Leyns, L., Bouwmeester, T., Kim, S.H., Piccolo, S., De Robertis, E.M., 1997. Frzb-1 is a secreted antagonist of Wnt signaling expressed in the Spemann organizer. Cell 88, 747–756. Moon, R.T., Brown, J.D., Yang-Snyder, J.A., Miller, J.R., 1997. Structurally related receptors and antagonists compete for secreted Wnt ligands. Cell 88, 725–728. Nakayama, K., 1997. Furin: a mammalian subtilisin/Kex2p-like endoprotease involved in processing of a wide variety of precursor proteins. Biochem. J. 327, 625–635. Newport, J., Kirschner, M., 1982. A major developmental transition in early Xenopus embryos. I. Characterization and timing of cellular changes at the midblastula stage. Cell 30, 675–686. Nielsen, H., Engelbrecht, J., Brunak, S., von Heijne, G., 1997. Identification of prokaryotic and eukaryotic signal peptides and prediction of their cleavage sites. Protein Eng. 10, 1–6. Nieuwkoop, P.D., Faber, J., 1967. Normal Table of Xenopus laevis (Daudin). North Holland, Amsterdam. Parr, B.A., McMahon, A.P., 1994. Wnt genes and vertebrate development. Curr. Opin Genet. Dev. 4, 523–528.
313
Perrimon, N., 1994. The genetic basis of patterned baldness in Drosophila. Cell 76, 781–784. Piccolo, S., Agius, E., Leyns, L., Bhattacharyya, S., Grunz, H., Bouwmeester, T., De Robertis, E.M., 1999. The head inducer Cerberus is a multifunctional antagonist of Nodal, BMP and Wnt signals. Nature 397, 707–710. Rattner, A., Hsieh, J.C., Smallwood, P.M., Gilbert, D.J., Copeland, N.G., Jenkins, N.A., Nathans, J., 1997. A family of secreted proteins contains homology to the cysteine-rich ligand-binding domain of frizzled receptors. Proc. Natl. Acad. Sci. USA 94, 2859–2863. Reichsman, F., Moore, H.M., Cumberledge, S., 1999. Sequence homology between Wingless/Wnt-1 and a lipid binding domain in secreted phospholipase A2. Curr. Biol. 9, R353–R355. Sambrook, J., Fritsch, E.F., Maniatis, T., 1989. Molecular Cloning: A Laboratory Manual. 2nd ed., Cold Spring Harbor Laboratory Press, Plainview, NY. Sawada, K., Agata, K., Eguchi, G., 1996. Characterization of terminally differentiated cell state by categorizing cDNA clones derived from chicken lens fibers. Int. J. Dev. Biol. 40, 531–535. Smolich, B.D., McMahon, J.A., McMahon, A.P., Papkoff, J., 1993. Wnt family proteins are secreted and associated with the cell surface. Mol. Biol. Cell 4, 1267–1275. Sokol, S.Y., Klingensmith, J., Perrimon, N., Itoh, K., 1995. Dorsalizing and neuralizing properties of Xdsh, a maternally expressed Xenopus homologue of dishevelled. Development 121, 1637–1647. Sokol, S., 1996. Analysis of Dishevelled signalling pathways during Xenopus development. Curr. Biol. 6, 1456–1467. Wang, S., Krinks, M., Lin, K., Luyten, F.P., Moos Jr., M., 1997. Frzb, a secreted protein expressed in the Spemann organizer, binds and inhibits Wnt-8. Cell 88, 757–766. Wodarz, A., Nusse, R., 1998. Mechanisms of Wnt signaling in development. Annu. Rev. Cell Dev. Biol. 14, 59–88. Wolda, S.L., Moody, C.J., Moon, R.T., 1993. Overlapping expression of Xwnt-3A and Xwnt-1 in neural tissue of Xenopus laevis embryos. Dev. Biol. 155, 46–57. Zeng, L., Fagotto, F., Zhang, T., Hsu, W., Vasicek, T.J., Perry III, W.L., Lee, J.J., Tilghman, S.M., Gumbiner, B.M., Costantini, F., 1997. The mouse Fused locus encodes Axin, an inhibitor of the Wnt signaling pathway that regulates embryonic axis formation. Cell 90, 181–192. Zorn, A.M., 1997. Cell–cell signaling: frog frizbees. Curr. Biol. 7, R501–R504.
Functional and structural diversity of the human Dickkopf gene family Valery E. Krupnik a, John D. Sharp b, Chian Jiang c, Keith Robison d, Troy W. Chickering c, Lakshmi Amaravadi c, Diane E. Brown c,1, Deborah Guyot c,2, Gail Mays c, Kevin Leiby c, Belle Chang c, Thao Duong c, Andrew D.J. Goodearl c, David P. Gearing c, Sergei Y. Sokol a, Sean A. McCarthy c, * a Department of Microbiology and Molecular Genetics, Harvard Medical School and Division of Molecular Medicine, Beth Israel Deaconess Medical Center, 330 Brookline Avenue, Boston, MA 02215, USA b Division of Research Technologies and Proteins, Lilly Research Laboratory, Indianapolis, IN 46285, USA c Millennium BioTherapeutics Inc., 640 Memorial Drive, Cambridge, MA 02139, USA d Millennium Pharmaceuticals Inc., 640 Memorial Drive, Cambridge, MA 02139, USA Received 29 April 1999; received in revised form 23 July 1999; accepted 6 August 1999; Received by I. Verma
Abstract Wnt proteins influence many aspects of embryonic development, and their activity is regulated by several secreted antagonists, including the Xenopus Dickkopf-1 (xDkk-1) protein. xDkk-1 inhibits Wnt activities in Xenopus embryos and may play a role in induction of head structures. Here, we characterize a family of human Dkk-related genes composed of Dkk-1, Dkk-2, Dkk-3, and Dkk-4, together with a unique Dkk-3 related protein termed Soggy (Sgy). hDkks 1–4 contain two distinct cysteine-rich domains in which the positions of 10 cysteine residues are highly conserved between family members. Sgy is a novel secreted protein related to Dkk-3 but which lacks the cysteine-rich domains. Members of the Dkk-related family display unique patterns of mRNA expression in human and mouse tissues, and are secreted when expressed in 293T cells. Furthermore, secreted hDkk-2 and hDkk-4 undergo proteolytic processing which results in cleavage of the second cysteine-rich domain from the full-length protein. Members of the human Dkk-related family differ not only in their structures and expression patterns, but also in their abilities to inhibit Wnt signaling. hDkk-1 and hDkk-4, but not hDkk-2, hDkk-3 or Sgy, suppress Wnt-induced secondary axis induction in Xenopus embryos. hDkk-1 and hDkk-4 do not block axis induction triggered either by Xenopus Dishevelled ( Xdsh) or Xenopus Frizzled-8 ( Xfz8), both of which function to transduce signals from Wnt ligands. Thus, hDkks 1 and 4 may inhibit Wnt activity by a mechanism upstream of Frizzled. Our findings highlight the structural and functional heterogeneity of human Dkk-related proteins. © 1999 Elsevier Science B.V. All rights reserved. Keywords: Antagonist; Frizzled related protein; In situ hybridization; Secreted protein; Soggy; Wingless; Xenopus
1. Introduction The orchestration of early embryonic development requires the concerted actions of many secreted signaling Abbreviations: Cys-1, cysteine-rich domain-1; Cys-2, cysteine-rich domain-2; Dkk, Dickkopf; EST, expressed sequence tag; FRP, Frizzled-related protein; ISH, in situ hybridization; PCR, polymerase chain reaction; Sgy, Soggy; Wg, Wingless; Xdsh, Xenopus Dishevelled; Xfz8, Xenopus Frizzled-8. * Corresponding author. Tel.: +1-617-679-7157; fax: +1-617-374-7653. E-mail address: [email protected] (S.A. McCarthy) 1 Present address: Lilly Research Laboratory, Greenfield, IN 46140, USA. 2 Present address: Pfizer Central Research Division, Eastern Point Rd., Groton, CT 06340, USA.
molecules. The Wnt gene family comprises a large class of secreted proteins related to the Int1/Wnt1 protooncogene and Drosophila wingless ( Wg). Wnts are expressed in a variety of tissues and organs, and are required for many developmental processes, including segmentation in Drosophila, endoderm development in Caenorhabditis elegans, establishment of limb polarity, neural crest differentiation, kidney morphogenesis, sex determination and brain development in mammals (Parr and McMahon, 1994; Cadigan and Nusse, 1997; Wodarz and Nusse, 1998). Recent studies have led to the identification of several components of the Wnt signal transduction pathway in responding cells. Wnt signals are transduced by the Frizzled (Fz) family of seven transmembrane domain receptors (Bhanot et al., 1996). The
0378-1119/99/$ – see front matter © 1999 Elsevier Science B.V. All rights reserved. PII: S0 3 7 8 -1 1 1 9 ( 9 9 ) 0 0 36 5 - 0
302
V.E. Krupnik et al. / Gene 238 (1999) 301–313
resulting signal leads to the activation of the cytoplasmic protein Dishevelled (Dsh) and stabilization of the Armadillo/b-catenin protein (Perrimon, 1994). Negative regulators of the Wnt pathway include glycogen synthase kinase 3 (GSK3)/shaggy (Perrimon, 1994), the tumor suppressor gene product APC (adenomatous polyposis coli; Gumbiner, 1997) and Axin ( Zeng et al., 1997). In the absence of a Wnt ligand, these regulators promote phosphorylation and consequential degradation of b-catenin. Conversely, Wnt signaling inactivates GSK3 and therefore stabilizes b-catenin, which translocates to the nucleus and forms a complex with TCF transcription factors, thereby regulating target gene expression. Deregulation of this pathway can lead to carcinogenesis, emphasizing the long-recognized connection between Wnts, normal development and cancer (Cadigan and Nusse, 1997). This connection has been strengthened recently with the identification of the c-Myc protooncogene as a target of Wnt signaling (He et al., 1998). While the outcome of Wnt signaling may be influenced by multiple intracellular regulatory mechanisms, recent studies have also identified several classes of secreted factors which modulate Wnt action outside of the cell. These include a family of secreted Frizzledrelated proteins (FRPs; Zorn, 1997; Moon et al., 1997; Wodarz and Nusse, 1998), Cerberus (Piccolo et al., 1999) and Wnt Inhibitory Factor-1 ( WIF-1; Hsieh et al., 1999a). FRPs are structurally related to the extracellular domains of Frizzled proteins, particularly in conserved cysteine-rich domains (Moon et al., 1997). Cerberus is a multifunctional inhibitor of BMP, Nodal and Wnt signaling (Piccolo et al., 1999). WIF-1 possesses a unique structure composed of an amino terminal WIF domain and 5 carboxy terminal epidermal growth factor repeats ( Hsieh et al., 1999a, b). WIF-1 and certain FRPs inhibit Wnt function by direct binding of the ligand, thus preventing access to cell surface receptors (Hsieh et al., 1999a, b; Wang et al., 1997; Leyns et al., 1997; Rattner et al., 1997; Bafico et al., 1999). The Dickkopf (Dkk) family comprises another discrete class of secreted Wnt inhibitors. xDkk-1 was identified in a screen for factors capable of inducing Xenopus head formation in the context of inhibition of bone morphogenetic protein signaling (Glinka et al., 1998). xDkk-1 is expressed in the Spemann organizer, which is responsible for induction of head structures, and is therefore a candidate for the endogenous signal for head induction (Glinka et al., 1998). Dkk-1 inhibits the axis-inducing activity of Xwnt8, consistent with the suggestion that simultaneous inhibition of BMP and Wnt signaling is required for head formation. xDkk-1 does not inhibit secondary axis induction by Xdsh, however. Thus, like the FRPs and WIF-1, xDkk-1 is most likely to act as an extracellular factor that antagonizes Wnt function directly. A murine homolog of xDkk-1 also inhibits Xwnt8-mediated axis duplication
in Xenopus embryos. Furthermore, Dkk-1 appears to be the prototypical member of a family of Dkk-related genes (Glinka et al., 1998). Here, we describe the characterization of a family of four human Dkk-related proteins and a unique protein, Soggy, which possesses homology to hDkk-3 but not other Dkks. Our data demonstrate structural and functional heterogeneity within this novel class of secreted proteins.
2. Materials and methods 2.1. Isolation of mammalian Dkk cDNAs Human Dkk-related proteins were identified by searching of internal and public expressed sequence tag ( EST ) databases with the sequence of hDkk-3 using the TBLASTN algorithm ( WashUversion 2.0, BLOSUM62 search matrix). From this search, four distinct partial protein sequences were identified which displayed significant homology to human Dkk-3. The full-length cDNA sequence of hDkk-1 was determined by complete sequencing of an individual clone from a human fetal kidney cDNA library which was identified within an internal EST database. A full-length hDkk2 cDNA was isolated from a human fetal lung cDNA library by plaque hybridization (Sambrook et al., 1989) using a partial cDNA obtained from the IMAGE collection of cDNA clones as a probe (GenBank accession No. W55979). A full-length cDNA clone for hDkk-3 was isolated from a human fetal brain cDNA library by plaque hybridization using a partial cDNA identified by a Signal Peptide Trap procedure as a probe (manuscript in preparation). The full-length cDNA sequence of mDkk-3 was determined by complete sequencing of an individual clone identified within a murine adult brain cDNA library which was identified within an internal EST database. The full-length cDNA sequence for hDkk-4 was determined by complete sequencing of a partially sequenced cDNA clone from the IMAGE collection (GenBank accession No. AA565546). Fulllength cDNA sequences for hSoggy and mSoggy were obtained by complete sequencing, respectively, of IMAGE clones with GenBank accession Nos. AA397836 and AA497886. All IMAGE collection clones were purchased from Research Genetics. cDNA libraries for screening were prepared using the Gibco Superscript cDNA synthesis kit and the lZiplox vector (Gibco). Probes were labeled with [a-32P]dCTP (Prime-It, Stratagene). DNA sequencing was performed using automated ABI377 sequence detectors and dye terminator chemistry. 2.2. Analysis of mRNA expression For Northern analysis, all hybridizations were to Clontech Multiple Tissue Northern Blots and were
V.E. Krupnik et al. / Gene 238 (1999) 301–313
performed in ExpressHyb solution (Clontech) for 1– 20 h. All cDNA probes were prepared by random primed radiolabeling with [a-32P]dCTP (Prime-It, Stratagene). Blots were washed sequentially to a final stringency of 0.2×SSC/0.2% SDS and exposed to autoradiographic film. Hybridizations of a control b-actin cDNA probe consistently demonstrated even loading of the Northern blots (not shown). For in situ hybridization (ISH ), normal mouse embryos and adult mouse tissues were collected from C57BL/6 mice, embedded in TissueTek O.C.T Compound (Sakura Finetek USA, Torrance, CA) and frozen on dry ice. Cryostat serial sections (8 mm) were thaw mounted on Superfrost Plus slides ( VWR) and air dried on a slide warmer at 40°C for 20 min. Sections were then fixed with 4% formaldehyde in DEPC-treated 0.1 M phosphatebuffered saline (DEPC PBS, pH 7.5) at room temperature for 10 min and rinsed twice in DEPC PBS. Sections were rinsed in 0.1 M triethanolamine–HCl ( TEA, pH 8.0), incubated in 0.25% acetic anhydride–TEA for 10 min and rinsed in DEPC 2×SSC. Sections were dehydrated through a series of graded ethanols, incubated in 100% chloroform for 5 min, rinsed in 100% and 95% ethanol for 1 min and air dried. Labeling and hybridizations of radiolabeled antisense and sense RNA probes were performed as described previously (Busfield et al., 1997). Primers for template amplification were as follows: mDkk-3 forward, 5∞-AACAAACATGCAGCGGCTCG-3∞; reverse, 5∞-TGTGATAGTTGGGAGGTAAG-3∞; mSoggy forward, 5∞-ACCTGCAATGTGTCGACTGAG-3∞; reverse 5∞-CACTTACAGCTGTTGGGATG-3∞. 2.3. Protein expression Flag epitope-tagged forms of hDkk-1, hDkk-2, hDkk-3, hDkk-4 and hSgy cDNAs were constructed by PCR and ligated to the expression plasmid pM7-stop (S. McCarthy, unpublished ). The full-length open-reading frame of each clone was PCR amplified using a 5∞ primer incorporating a Kozak consensus sequence prior to the initiator methionine and a 3∞ primer including the nucleotide sequence encoding the flag epitope (DYKDDDDK ) followed by a termination codon. All DNA constructs were sequenced fully to ensure fidelity of PCR amplification. Constructs were transiently transfected into human embryonic kidney 293T cells using Lipofectamine (Gibco) according to the manufacturer’s instructions. 72 h post-transfection, serum-free conditioned media (OptiMEM, Gibco/BRL) were harvested, centrifuged and filtered. Cell lysates were prepared using lysis buffer (50 mM Tris pH 8.0, 150 mM NaCl, 1% NP-40, 0.05% SDS) supplemented with Complete Protease Inhibitor cocktail (Boehringer Mannheim). For Western analysis, samples of conditioned media were electroblotted onto PVDF (Novex)
303
after SDS–PAGE on 4–20% gradient gels, probed with M2 anti-flag antibody (1:500, Sigma) followed by HRP conjugated sheep anti-mouse IgG (1:5000, Amersham), developed with chemiluminescent reagents ( Renaissance, Dupont) and exposed to Biomax MR2 film ( Kodak). 2.4. Deglycosylation Samples of 293T cell supernatants collected 72 h after transfection with the appropriate Dkk expression plasmids were incubated with anti-flag M2 agarose beads (Sigma) for 16 h at 4°C. Beads were washed with PBS (pH 7.4) containing, sequentially, 0.1%, 0.05% and 0.01% Triton X-100. The beads were resuspended in incubation buffer (20 mM sodium phosphate, pH 7.5, 50 mM EDTA, 0.02% sodium azide) containing 0.5% SDS, 5% 2-mercaptoethanol and boiled for 2 min. Equal amounts of each immunoprecipitate were diluted with incubation buffer containing 1% NP-40 and then treated either with N-glycanase (Oxford Glycosystems) in enzyme buffer (20 mM Tris–HCl, 1 mM EDTA, 50 mM NaCl, 0.02% sodium azide pH 7.5) or with enzyme buffer alone as control for 18 h at 37°C. Samples were analyzed by reducing SDS–PAGE and Western blotting with anti-flag M2 antibody. 2.5. Affinity purification and N-terminal sequencing Flag-tagged hDkk-4 protein was purified by passing conditioned media over an anti-flag(M2)biotin/streptavidin Poros column (2.1×30 mm, PE Biosystems); the column was then washed with PBS (pH 7.4), and flag-tagged protein eluted with 200 mM glycine (pH 3.0). Eluted fractions were analyzed by SDS–PAGE and Western blot. Purified hDkk-4 protein bound to PVDF membrane was subjected to N-terminal amino-acid analysis on a PE Applied Biosystems Model 494 Procise instrument using Edman-based chemistry ( Hewick et al., 1981). The amino-acid residues were analyzed by HPLC (Spherogel micro PTH 3-micron column) and determined by separation and peak height as compared to standards. 2.6. Xenopus embryo culture and RNA microinjections Oocytes were obtained from Xenopus females injected with 700 units of human chorionic gonadotropin, fertilized in vitro and cultured in 0.1×MMR (Newport and Kirschner, 1982). Embryonic stages were determined according to Nieuwkoop and Faber (1967). All cDNAs were subcloned into the vector pCS2 and capped mRNAs were synthesized in vitro using the Message Machine kit (Ambion). Constructs for Xwnt8 (Christian et al., 1991), Xwnt2b (Landesman and Sokol, 1997), Xwnt3a ( Wolda et al., 1993), Xfz8 (Itoh et al., 1998)
304
V.E. Krupnik et al. / Gene 238 (1999) 301–313
and Xdsh (Sokol et al., 1995) have been described previously. Protein expression from all pCS2–Dkk constructs was confirmed by in vitro transcription and translation ( TNT, Promega). For tests of secondary axis induction, a single ventral blastomere of four- or eightcell embryos was injected with 10 nl of a solution containing 2–4 pg of Xwnt8 mRNA, 2.5–5 pg of Xwnt3a mRNA, 10 pg of Xwnt2b mRNA, 2 ng Fz8 mRNA or 1 ng Xdsh mRNA as described (Itoh et al., 1995). 2.5 ng of hDkk or hSgy mRNAs were co-injected as indicated. After injections, embryos were cultured in 3% Ficoll 400 (Pharmacia), 0.5×MMR. Secondary axes were scored at stage 35 as complete when they contained anterior neuroectodermal derivatives including pronounced cement gland and eyes. Axes were scored as partial when the secondary neural tube with melanocytes was apparent but head structures were not.
3. Results and discussion 3.1. Characterization of the human Dkk-related gene family Fig. 1A depicts schematically the structures of the family of human Dkk-related proteins that we have characterized in the present study. The human Dkk (hDkk) family comprises hDkk-1, hDkk-2, hDkk-3, hDkk-4, and a unique Dkk-3-related protein we refer to as Soggy (Sgy). hDkk-1 is the full-length human homolog of previously reported Xenopus and mouse Dkk-1 (Glinka et al., 1998). hDkk-2 and hDkk-3 correspond to previously described partial hDkk-2 and hDkk-3 sequences (Glinka et al., 1998). hDkk-4 and Sgy represent novel members of this gene family. Sequence identities between the various Dkk proteins described here and elsewhere (Glinka et al., 1998) are summarized in Fig. 1B. As described previously for Xenopus Dkk-1, human Dkks 1–4 each possess an N-terminal signal peptide and contain two conserved cysteine-rich domains (Cys-1 and Cys-2) separated by a linker region (Fig. 1A, C ). Each domain possesses 10 conserved cysteine residues. The positions of the 10 conserved cysteines in Cys-2 are closely similar to those in proteins of the colipase family (Fig. 1D; Aravind and Koonin, 1998). Furthermore, sequence conservation among the Dkks is greatest within Cys-2, suggestive of a conserved function. Colipases facilitate interactions of pancreatic lipases with lipid micelles, and for this reason it has been suggested that the Cys-2 domain of the Dkks may enable the proteins to interact with lipids in order to regulate Wnt function (Aravind and Koonin, 1998). Wnt proteins are known to remain tightly associated with the cell surface (Smolich et al., 1993), and a putative lipid binding function may facilitate Wnt/Dkk interactions at the
Fig. 1. The human Dkk family. (A) Schematic illustration of the human Dkk/Sgy family showing signal peptides (darkened boxes), Cysteine-rich domains (Cys-1 and Cys-2) and putative sites of N-linked glycosylation in the human Dkks (branches). (B) Percent identities between full-length protein sequences for hDkks 1–4, mDkk-1 (AF030433), mDkk-3, cDkk-3 (D26311) and xDkk-1 (AF030434). Figures were generated using the Smith–Waterman algorithm as implemented in the program Bestfit of the GCG package, with gap penalties of 8 for opening and 1 for extending. All 20 cysteines were aligned in all but five of the 28 comparisons, in which the final cysteine was mismatched. (C ) Multiple sequence alignment of different Dkk proteins. Alignments were performed with the ClustalW algorithm as implemented in the GCG program PILEUP. Predicted signal peptides (Nielsen et al., 1997) are underlined, putative N-glycosylation sites are indicated by a thick bar, Cys-1 by an open box, Cys-2 by a shaded box. The proteolytic cleavage site within hDkk-4 is indicated by an arrow. (D) Multiple sequence alignment of hDkks with human colipase (J02883). Cys-2 of each hDkk is shown and indicated by the shaded box. The human colipase sequence is shown missing eight amino acids from the N-terminus of the mature peptide. The alignment was made using gap penalties of 12 for opening and 2 for extending. A minor adjustment was necessary, since PILEUP inserts a single gap in hDkk-1 and hDkk-2 between GS at position 56–57, even with gap opening penalty of 15. hDkk-1 is shown from amino acid 181. (E ) Multiple sequence alignment of Dkk-3 and Soggy proteins. Signal peptides and N-glycosylation sites are indicated as in (C ). Cys-1 (open box) and Cys-2 (shaded box) within Dkk-3 are shown for reference. Asterisks indicate amino-acid identities between all four proteins. Soggy protein sequences terminate at #.
plasma membrane. Interestingly, sequence similarity between Wnt-1 and the lipid binding domain of secreted phospholipase A2 has recently been noted, further suggesting that Wnts may interact directly with lipids ( Reichsman et al., 1999). In contrast to the conserved cysteine-rich domains, the linker region between Cys-1 and Cys-2 is highly variable between hDkks and it is notably larger in hDkk-1, -2 and -4 (50–55 aa) compared with Dkk-3 (12 aa). Four potential sites of N-linked glycosylation in hDkk-3 are conserved in chick and
V.E. Krupnik et al. / Gene 238 (1999) 301–313
305
Fig. 1. (continued).
mouse Dkk-3 (Fig. 1A, C ). These sites are not conserved in other Dkk family members. hDkk-1 possesses one potential N-glycosylation site located close to the
C-terminus of the protein which is conserved in murine Dkk-1 but not in Xenopus Dkk-1 (Fig. 1C ). In addition, each hDkk possesses several potential sites of proteolytic
306
V.E. Krupnik et al. / Gene 238 (1999) 301–313
Fig. 1. (continued).
cleavage by furin-type proteases (Nakayama, 1997), suggesting that the proteins may be subject to posttranslational proteolytic processing. hDkk-3 is the most divergent of the four human Dkks and possesses an extended N-terminal domain which precedes the Cys-1 domain and an extended C-terminal region that follows Cys-2 (Fig. 1A, C ). Sgy is a novel protein with sequence similarity to Dkk-3 but not other Dkks (Fig. 1A, E). Human and murine Sgy cDNAs encode proteins of 242 aa and 230 aa, respectively, and are predicted to be secreted owing to the presence of N-terminal signal peptides. hDkk3 and hSgy display 22% amino-acid identity, whereas human and murine Sgy proteins display 59% overall identity. Human and mouse Sgy proteins each possess two sites of potential N-linked glycosylation which are conserved with Dkk-3 ( Fig. 1E). Sequence homology between Dkk-3 and Sgy is most pronounced within the N-terminal domain of Dkk-3, although significant amino-acid identities are seen beyond this domain that extend into the cysteine-rich domains of Dkk-3 (Fig. 1E). However, cysteine residues are not conserved
within these domains and the residues shared by Sgy and Dkk-3 are poorly conserved in other Dkks. Thus, Sgy is uniquely related to Dkk-3. 3.2. mRNA expression analysis To investigate further the relationship between different mammalian Dkks, we analyzed their tissue distribution by Northern analysis. Fig. 2 shows that mammalian Dkks are expressed in distinct tissues. A 1.8 kb hDkk-1 mRNA was detected in human placenta, but not in other tissues tested ( Fig. 2A). hDkk-2 transcripts of approx. 4 and 4.5 kb were detected in multiple tissues, including heart, brain, skeletal muscle and lung ( Fig. 2B). A 2.5 kb hDkk-3 transcript was also widely expressed in human tissues, although levels of this mRNA were substantially higher in heart, brain and spinal cord relative to other tissues tested ( Fig. 2C, D). hDkk-4 mRNA was undetectable in all adult and fetal human tissues examined by Northern analysis, although a survey of a human cDNA library panel by PCR with hDkk-4-specific PCR primers generated products from
Fig. 2. Tissue distribution of Dkk and Soggy mRNAs. Multiple tissue Northern (MTN ) blots hybridized with radiolabeled cDNA probes specific to (A) hDkk-1, (B) hDkk-2, (C, D) hDkk-3, and ( E, F ) mSgy. All MTN blots were from Clontech. (A, C ) Human MTN-1; (B) human 12-lane MTN; (D) human MTN-III; ( E) mouse MTN-1; (F ) whole mouse embryo. Sizes of molecular weight markers are indicated.
V.E. Krupnik et al. / Gene 238 (1999) 301–313 307
308
V.E. Krupnik et al. / Gene 238 (1999) 301–313
libraries prepared from cerebellum, activated human T-lymphocytes, lung and esophagus (data not shown). A 1 kb Soggy mRNA was expressed in mouse testis, but not in other adult murine tissues tested (Fig. 2E). Interestingly, mSgy mRNA was also expressed between days 11 and 15 of gestation in mouse embryos (Fig. 2F ). Thus, mSgy displays temporal regulation during mouse development. We also investigated the expression of murine Dkk-3 and murine Sgy by in situ hybridization (ISH ). Fig. 3 shows the expression pattern of mDkk-3 in the adult murine brain, eye, and heart. mDkk-3 expression in the brain was found in neurons of the cortex and hippocampus (Fig. 3B). In adult eye, mDkk-3 mRNA was found to be highly expressed in the retina, ciliary body, and lens epithelium ( Fig. 3E). Expression in the retina was limited to the integrating bipolar neurons and ganglion cells. Consistent with our findings, Dkk-3 mRNA is also expressed at high levels in the chick eye (Sawada et al,
1996). In adult heart, mDkk-3 was detected in the atrioventricular valves and also in myocytes of the atria. Expression was highly restricted to the atria and noticeably absent from ventricular tissue ( Fig. 3H ). We also observed expression of mDkk-3 mRNA in developing eye, bone and cartilage in day 14 embryos (not shown). Fig. 4 details the expression of murine Sgy analyzed by ISH. Consistent with Northern analysis, mSgy mRNA was expressed at high levels in adult testis ( Fig. 4A) in the spermatogenic epithelium of the seminiferous tubules and the spermatogonia at various stages of development. We also analyzed a series of sagittal sections of mouse embryos from E13.5 to E17.5. In E15.5 embryos, Sgy mRNA transcripts were localized to the developing dorsal root ganglia and also in the cartilage primordium of the nasal septum ( Fig. 4C ). Soggy expression was also seen in the eye from E13.5 to E16.5, as we observed for mDkk-3 (Fig. 4E). Expression of Sgy mRNA at these stages of development suggests that this novel
Fig. 3. Expression of mDkk-3 in adult mouse tissues. In situ hybridization analysis of mDkk-3 mRNA expression in adult murine tissues. (A–C ) Brain, parasagittal section; (D–F ) eye; (G–I ) heart, parasagittal section. (A, D, G) Bright field image of hematoxylin and eosin stained sections; (B, E, H ) antisense probe; (C, F, I ) sense probe. Scale bars in the bright field images shown in (A, D, and G) correspond to 1 mm.
V.E. Krupnik et al. / Gene 238 (1999) 301–313
Fig. 4. Expression of Sgy in embryonic and adult mouse tissues. In situ hybridization analysis of mSoggy mRNA expression in: (A, B) mouse testis, (C, D) E15.5 embryo sagittal sections, and (E, F ) E.16.5 embryo sagittal sections. (A, C, E ) Antisense probe; (B, D, F ) sense probe. Scale bar in (A) corresponds to 0.5 mm, and in (C, E ) to 1 mm.
secreted protein may play multiple roles in during embryogenesis. Thus, Dkks 1–4 and Sgy display distinct patterns of mRNA expression in human and murine tissues. Similarly, different Wnt family members have divergent patterns of mRNA expression in adult and embryonic mammalian tissues. For example, murine Wnts-4, -7a and -7b are expressed in brain and lung, whereas Wnt-6 is highly expressed in testis (Gavin et al., 1990). Wnts-5b and -13 are more broadly expressed (Gavin et al., 1990; Katoh et al., 1996). Given the large size of the mammalian Wnt family and the growing number of distinct families of Wnt inhibitors, it seems likely that tissuespecific expression of regulators such as the Dkks may afford an additional level of control to Wnt signaling. 3.3. Secretion and post-translational modification of hDkk proteins We next characterized human Dkk and Sgy proteins by transient overexpression of flag epitope-tagged forms in 293T cells (Fig. 5). Similar to Xenopus Dkk-1, hDkk-1
309
was readily detected in conditioned medium of transfected cells. Mature secreted hDkk-1 migrated with a molecular mass of approx. 42–50 kDa ( Fig. 5a, lane 3). hDkk-3 was also secreted by 293T cells and migrated as a heterogeneous band of 45–65 kDa ( Fig. 5a, lane 1). Soluble hDkk-4 was detected as three major immunoreactive species of approx. 40 kDa [form (i)], 30–32 kDa [ form (ii)] and 15–17 kDa [form (iii)]. We also observed additional minor forms of hDkk-4 in certain experiments, as shown, in the range of 20–21 kDa. Soluble hDkk-2 was detected as a major species of 15–17 kDa ( Fig. 5a, lane 4), closely similar in size to form (iii) of hDkk-4. hSgy was also efficiently secreted from transfected 293T cells and migrated with a molecular mass of approx. 40–50 kDa (Fig. 5a, lane 5). Thus, human Dkk/Sgy proteins are soluble, secreted factors with heterogeneous profiles suggestive of post-translational modification. We also studied the effect of N-glycanase treatment on the mobilities of secreted hDkk-1, hDkk-3, hDkk-4 and hSgy. Treatment with N-glycanase caused only a minor increase in the mobility of soluble hDkk-1 ( Fig. 5b, lanes 1 and 2). hDkk-1 contains only a single putative N-glycosylation site which is located towards the C-terminus of the protein. Soluble hDkk-3 protein displayed a substantial increase in mobility following Nglycanase treatment. The major 45–65 kDa form of soluble hDkk-3 was observed as two species of 45–55 and 40 kDa following deglycosylation (Fig. 5b, lanes 3 and 4). This finding is consistent with the presence of four potential sites of N-linked glycosylation in the hDkk-3 protein. The reason for the heterogeneity of deglycosylated hDkk-3 is unclear, although it may reflect either proteolytic processing or incomplete removal of carbohydrate from one or more attachment sites. We also observed a 30 kDa hDkk3 species in these experiments, the mobility of which was unaltered by Nglycanase treatment. This form was observed only after overnight incubation of the samples and may therefore be a non-specific degradation product. For hDkk-4, neither form (i), (ii) nor (iii) was significantly affected by N-glycanase treatment, consistent with the absence of N-glycosylation sites from the protein ( Fig. 5b, lanes 5 and 6). hSgy, however, displayed a 5–10 kDa decrease in apparent molecular mass after N-Glycanase treatment, consistent with the presence of two potential sites of N-glycosylation in the protein (Fig. 5b, lanes 7 and 8). Taken together, these results suggest that hDkk-3 and hSgy are expressed and secreted by 293T cells as glycoproteins. hDkk-1 displays minimal, if any, attachment of N-linked carbohydrate and hDkk-4 shows no evidence of such modification. Soluble hDkk-4 was consistently observed in three forms of approx. 40 kDa [ form (i)], 30–32 kDa [ form (ii)] and 15–17 kDa (form (iii), while soluble hDkk-2 was detected as a single C-terminal form of 15–17 kDa. We therefore analyzed processing of hDkk-4 by Edman
310
V.E. Krupnik et al. / Gene 238 (1999) 301–313
Fig. 5. hDkk protein expression and analysis. (a) Secretion of hDkk/Soggy proteins. Anti-flag Western blots are shown of conditioned media of 293T cells transiently transfected with expression vectors for the indicated flag-tagged hDkks. Lane 1: hDkk-3; lane 2: hDkk-4; lane 3: hDkk-1; lane 4: hDkk-2; lane 5: hSgy. For hDkk-4, bands (i), (ii) and (iii) represent the three forms of the tagged protein consistently observed in cell supernatants. Note the similarity in size of hDkk-4 form (iii) and the major form of hDkk-2 detected by this analysis ( lane 4). (b) Effect of Nglycanase treatment on the mobilities of hDkk/Soggy proteins on SDS–PAGE. Immunoprecipitated hDkk/Soggy proteins were incubated overnight either with buffer alone (−) or N-glycanase (+) and then separated on reducing SDS–PAGE. The proteins were analyzed by Western analysis with anti-flag antibodies. Lanes 1, 2: hDkk-1; lanes 3, 4: hDkk-3; lanes 5, 6: hDkk-4; lanes 7, 8: hSgy. (c) N-Terminal sequence analysis of hDkk-4 protein species. N-Terminal sequences of affinity purified material corresponding to hDkk-4 forms (i) and (iii) are shown. (d ) Effect of reducing agent (b-mercaptoethanol ) on mobility of soluble hDkk-4 species. Anti-flag Western blots of samples of conditioned media containing either hDkk-1 ( lanes 1 and 3) or hDkk-4 ( lanes 2 and 4) after separation under non-reducing or reducing SDS–PAGE, as indicated.
N-terminal sequencing of anti-flag affinity purified material corresponding to bands (i), (ii) and (iii) ( Fig. 5c). The N-terminal sequence of band (i) was found to be XVLDFNNIRS, which corresponds to the sequence that follows the predicted signal peptide cleavage site (between Ala-18 and Leu-19). Because the same band is identified by anti-flag antibodies, which recognize the C-terminal epitope tag, band (i) was thus identified as mature hDkk-4 protein with the signal sequence removed. The band (iii) N-terminal sequence was found to be SQGRKGQEGS, which corresponds to the Cys-2 domain cleaved at the dibasic site Lys132/Lys133 (Lys 114/Lys115 of mature hDkk4). We were unable to obtain conclusive data concerning the N-terminal sequence of band (ii), owing to low abundance of purified material. However, the data obtained for bands (i) and (iii) indicate clearly that hDkk-4 is proteolytically processed by 293T cells, potentially by furin type proteases (Nakayama, 1997). By inference from the hDkk-4
protein sequences, it is likely that the major 15–17 kDa form of hDkk-2 detected in these experiments also corresponds to the Cys-2 domain, although the precise site of cleavage remains to be determined. We also investigated whether the mobilities of hDkk-4 bands (i), (ii) and (iii) were altered on non-reducing SDS–PAGE. The three major species migrated similarly on SDS– PAGE conducted under either reducing and non-reducing conditions ( Fig. 5d). Thus, each of the major C-terminal (anti-flag immunoreactive) hDkk-4 species appear to exist as independent proteolytic fragments that are not covalently linked via disulfide bonds to other subunits or proteins when secreted from 293T cells. 3.4. Effects of hDkks on Wnt-induced axis duplication in Xenopus embryos The studies described above establish that Dkks are secreted proteins and that each Dkk displays a unique
V.E. Krupnik et al. / Gene 238 (1999) 301–313
pattern of proteolytic processing and N-linked glycosylation. We next analyzed the functional activities of human Dkk-related proteins. It has been shown previously that murine and Xenopus Dkk-1 block the ability of Xwnt8 to induce a secondary axis in the frog embryo (Glinka et al., 1998). We evaluated the activities of different hDkks in this assay. hDkk mRNAs were co-injected with Xwnt8 mRNA into single ventral blastomeres of four- or eight-cell embryos. Injected embryos were cultured for 2 days and frequencies of embryos with secondary axes were determined. As previously described (Christian et al., 1991), Xwnt8-injected embryos displayed complete axis duplication (Fig. 6A). Xwnt8-mediated axis duplication was inhibited by co-injection with mRNAs encoding hDkk-1 ( Fig. 6B) and hDkk-4 (Fig. 6F ). Injection of mRNAs encoding hDkk-2, hDkk-3 or hSgy had no effect on Xwnt8-induced axis duplication (Fig. 6C–E). We also investigated the effects of hDkks and hSgy on secondary axis induction by Xwnt3a and Xwnt2b. As with Xwnt8, hDkk-1 and hDkk-4 inhibited axis duplication in
311
response to these Wnt proteins whereas hDkk-3 and hSgy had no effect ( Fig. 7). To investigate the mechanism by which hDkk-1 and hDkk-4 inhibited Wnt function, we co-injected Dkk mRNAs with Xdsh, a downstream component of the Wnt signaling pathway (Sokol et al., 1995, 1996). As shown in Fig. 7, hDkks 1 and 4 did not block secondary axis induction by Xdsh, indicating that Dkks function upstream of, or in parallel with, Xdsh signaling. Similar findings have been reported previously for xDkk-1 (Glinka et al., 1998). We also tested whether hDkks could antagonize signaling by Xenopus Frizzled-8 ( Xfz8), a cell-surface Wnt receptor which can also induce secondary axes when ectopically expressed (Itoh et al., 1998). Neither hDkk-1 or hDkk-4 inhibited the axis-inducing activity of Xfz8 mRNA (Fig. 7). Taken together with the fact that hDkk-1 and hDkk-4 are secreted, these data suggest that Dkks antagonize Wnt function at a point upstream of Wnt receptors. The present study describes a family of five human Dkk-related proteins which display marked differences
Fig. 6. Effect of hDkks and Sgy on Wnt-induced axis duplication in Xenopus embryos. Effect of hDkks and hSgy on Xwnt8-induced secondary axis formation in Xenopus embryos. Four or eight cell stage embryos were injected subequatorially into a ventral blastomere with mRNAs for: (A) Xwnt8, (B) Xwnt8+hDkk-1, (C ) Xwnt8+hDkk-2, (D) Xwnt8+hDkk-3, ( E ) Xwnt8+hSgy, ( F ) Xwnt8+hDkk-4. Four representative embryos from multiple injections are shown at stage 35–37 in each panel.
312
V.E. Krupnik et al. / Gene 238 (1999) 301–313
AF177395; hDkk-3 — AF177396; hDkk-4 — AF177397; h-Soggy — AF177398; m-Soggy — AF177399; mDkk-3 — AF177400.
Acknowledgements We thank R. Moon for plasmids. We also thank Rick Clark, John Keilty, Betty Woolf, Pam Brauer, Donna Michnick, Kevin McDonald and members of the Millennium DNA sequencing, cDNA library and Full Length cDNA Cloning Groups for skilled assistance and advice. We are grateful to Joe Weber and Bill Holmes for critical comments on the manuscript and Chuck Gray for assistance with collaborations. This work was supported in part by grants from the March of Dimes Birth Defect Foundation and the National Institutes of Health to S.S. Fig. 7. Effect of human Dkk proteins on Wnt, Dsh and Fz-induced axis duplication. Four- or eight-cell stage embryos were injected subequatorially into a ventral blastomere with mRNAs for Xwnt-8, Xwnt-2b, Xwnt-3a, Xfz8 or Xdsh in the presence or absence of mRNAs for either hDkk-1, hDkk-3, hDkk-4 or hSgy as indicated. Embryos were allowed to develop until stage 33–37, then scored for secondary axis formation. Complete axes included eyes and cement glands; partial axes included partial neural tubes and hindbrain, but lacked fore/midbrain structures. n=number of embryos in the experiment; open bars, % of embryos with complete secondary axes; filled bars, % of embryos with partial secondary axes.
in sequence, mRNA expression patterns, post-translational modification and in their activities as inhibitors of Wnt signaling. The restricted mRNA expression patterns of different Dkk family members suggest that these proteins play diverse roles in embryonic and adult tissues. Further work is required to determine the mechanism by which hDkk-1 and hDkk-4 inhibit Wnt signaling, and also to investigate the relationships between post-translational processing and Wnt inhibition. The ability of hDkks to inhibit Wnt signaling was inversely correlated with the degree of proteolytic processing we observed in 293T cells, suggesting that proteolytic mechanisms may regulate Dkk activities. Recent studies indicate that hDkk-1 inhibits signaling and morphological changes induced in NIH3T3 cells by mammalian Wnt-2 (Fedi et al., 1999). It will be of interest to characterize the activities of the mammalian Dkk family in such assays, as well as to establish structure–function relationships and biochemical mechanisms of action. Recent advances in the biochemical analysis of Wnt proteins should facilitate these studies (Hsieh et al., 1999a, b).
4. Note added in proof The Genbank accession numbers of the sequences listed in this paper are: Dkk-1 — AF177394; hDkk-2 —
References Aravind, L., Koonin, E.V., 1998. A colipase fold in the carboxy-terminal domain of the Wnt antagonists — the Dickkopfs. Curr. Biol. 8 (477), 478–1997. Bafico, A., Gazit, A., Pramila, T., Finch, P.W., Yaniv, A., Aaronson, S.A., 1999. Interaction of Frizzled Related Protein ( FRP) with Wnt ligands and the Frizzled receptor suggests alternative mechanisms for FRP inhibition of Wnt signaling. J. Biol. Chem. 274, 16180–16187. Bhanot, P., Brink, M., Samos, C.H., Hsieh, J.C., Wang, Y., Macke, J.P., Andrew, D., Nathans, J., Nusse, R., 1996. A new member of the frizzled family from Drosophila functions as a Wingless receptor. Nature 382, 225–230. Busfield, S.J., Michnick, D.A., Chickering, T.W., Revett, T.L., Ma, J., Woolf, E.A., Comrack, C.A., Dussault, B.J., Woolf, J., Goodearl, A.D., Gearing, D.P., 1997. Characterization of a neuregulin-related gene, Don-1, that is highly expressed in restricted regions of the cerebellum and hippocampus. Mol. Cell. Biol. 17, 4007–4014. Cadigan, K.M., Nusse, R., 1997. Wnt signaling: a common theme in animal development. Genes Dev. 11, 3286–3305. Christian, J.L., McMahon, J.A., McMahon, A.P., Moon, R.T., 1991. Xwnt8, a Xenopus Wnt1/int-1-related gene responsive to mesoderm inducing growth factors, may play a role in ventral mesodermal patterning during embryogenesis. Development 111, 1045–1055. Fedi, P., Bafico, A., Nieto Soria, A., Burgess, W.H., Miki, T., Bottaro, D.P., Kraus, M.H., Aaronson, S.A., 1999. Isolation and biochemical characterization of the human Dkk-1 homologue, a novel inhibitor of mammalian Wnt signaling. J. Biol. Chem. 274, 19465–19472. Gavin, B.J., McMahon, J.A., McMahon, A.P., 1990. Expression of multiple novel Wnt-1/int-1-related genes during fetal and adult mouse development. Genes Dev. 4, 2319–2332. Glinka, A., Wu, W., Delius, H., Monaghan, A.P., Blumenstock, C., Niehrs, C., 1998. Dickkopf-1 is a member of a new family of secreted proteins and functions in head induction. Nature 391, 357–362. Gumbiner, B.M., 1997. Carcinogenesis: a balance between beta-catenin and APC. Curr. Biol. 7, R443–R436. He, T.C., Sparks, A.B., Rago, C., Hermeking, H., Zawel, L., da Costa, L.T., Morin, P.J., Vogelstein, B., Kinzler, K.W., 1998. Identifica-
V.E. Krupnik et al. / Gene 238 (1999) 301–313 tion of c-MYC as a target of the APC pathway. Science 281, 1509–1512. Hewick, R.M., Hunkapiller, M.W., Hood, L.E., Dreyer, W.J., 1981. A gas-liquid solid phase peptide and protein sequenator. J. Biol. Chem. 256, 7990–7997. Hsieh, J., Kodjabachian, L., Rebbert, M.L., Rattner, A., Smallwood, P.M., Harryman Samos, C., Nusse, R., Dawid, I.B., Nathans, J., 1999a. A new secreted protein that binds to Wnt proteins and inhibits their activities. Nature 398, 431–436. Hsieh, J., Rattner, A., Smallwood, P.M., Nathans, J., 1999b. Biochemical characterization of Wnt–Frizzled interactions using a soluble, biologically active vertebrate Wnt protein. Proc. Natl. Acad. Sci. USA 96, 3546–3551. Itoh, K., Jacob, J., Sokol, S.Y., 1998. A role for Xenopus Frizzled 8 in dorsal development. Mech. Dev. 74, 145–157. Itoh, K., Tang, T.L., Neel, B.G., Sokol, S.Y., 1995. Specific modulation of ectodermal cell fates in Xenopus embryos by glycogen synthase kinase. Development 121, 3979–3988. Katoh, M., Hirai, M., Sugimura, T., Terada, M., 1996. Cloning, expression and chromosomal localization of Wnt-13, a novel member of the Wnt gene family. Oncogene 13, 873–876. Landesman, Y., Sokol, S.Y., 1997. Xwnt-2b is a novel axis-inducing Xenopus Wnt, which is expressed in embryonic brain. Mech. Dev. 61, 1199–1209. Leyns, L., Bouwmeester, T., Kim, S.H., Piccolo, S., De Robertis, E.M., 1997. Frzb-1 is a secreted antagonist of Wnt signaling expressed in the Spemann organizer. Cell 88, 747–756. Moon, R.T., Brown, J.D., Yang-Snyder, J.A., Miller, J.R., 1997. Structurally related receptors and antagonists compete for secreted Wnt ligands. Cell 88, 725–728. Nakayama, K., 1997. Furin: a mammalian subtilisin/Kex2p-like endoprotease involved in processing of a wide variety of precursor proteins. Biochem. J. 327, 625–635. Newport, J., Kirschner, M., 1982. A major developmental transition in early Xenopus embryos. I. Characterization and timing of cellular changes at the midblastula stage. Cell 30, 675–686. Nielsen, H., Engelbrecht, J., Brunak, S., von Heijne, G., 1997. Identification of prokaryotic and eukaryotic signal peptides and prediction of their cleavage sites. Protein Eng. 10, 1–6. Nieuwkoop, P.D., Faber, J., 1967. Normal Table of Xenopus laevis (Daudin). North Holland, Amsterdam. Parr, B.A., McMahon, A.P., 1994. Wnt genes and vertebrate development. Curr. Opin Genet. Dev. 4, 523–528.
313
Perrimon, N., 1994. The genetic basis of patterned baldness in Drosophila. Cell 76, 781–784. Piccolo, S., Agius, E., Leyns, L., Bhattacharyya, S., Grunz, H., Bouwmeester, T., De Robertis, E.M., 1999. The head inducer Cerberus is a multifunctional antagonist of Nodal, BMP and Wnt signals. Nature 397, 707–710. Rattner, A., Hsieh, J.C., Smallwood, P.M., Gilbert, D.J., Copeland, N.G., Jenkins, N.A., Nathans, J., 1997. A family of secreted proteins contains homology to the cysteine-rich ligand-binding domain of frizzled receptors. Proc. Natl. Acad. Sci. USA 94, 2859–2863. Reichsman, F., Moore, H.M., Cumberledge, S., 1999. Sequence homology between Wingless/Wnt-1 and a lipid binding domain in secreted phospholipase A2. Curr. Biol. 9, R353–R355. Sambrook, J., Fritsch, E.F., Maniatis, T., 1989. Molecular Cloning: A Laboratory Manual. 2nd ed., Cold Spring Harbor Laboratory Press, Plainview, NY. Sawada, K., Agata, K., Eguchi, G., 1996. Characterization of terminally differentiated cell state by categorizing cDNA clones derived from chicken lens fibers. Int. J. Dev. Biol. 40, 531–535. Smolich, B.D., McMahon, J.A., McMahon, A.P., Papkoff, J., 1993. Wnt family proteins are secreted and associated with the cell surface. Mol. Biol. Cell 4, 1267–1275. Sokol, S.Y., Klingensmith, J., Perrimon, N., Itoh, K., 1995. Dorsalizing and neuralizing properties of Xdsh, a maternally expressed Xenopus homologue of dishevelled. Development 121, 1637–1647. Sokol, S., 1996. Analysis of Dishevelled signalling pathways during Xenopus development. Curr. Biol. 6, 1456–1467. Wang, S., Krinks, M., Lin, K., Luyten, F.P., Moos Jr., M., 1997. Frzb, a secreted protein expressed in the Spemann organizer, binds and inhibits Wnt-8. Cell 88, 757–766. Wodarz, A., Nusse, R., 1998. Mechanisms of Wnt signaling in development. Annu. Rev. Cell Dev. Biol. 14, 59–88. Wolda, S.L., Moody, C.J., Moon, R.T., 1993. Overlapping expression of Xwnt-3A and Xwnt-1 in neural tissue of Xenopus laevis embryos. Dev. Biol. 155, 46–57. Zeng, L., Fagotto, F., Zhang, T., Hsu, W., Vasicek, T.J., Perry III, W.L., Lee, J.J., Tilghman, S.M., Gumbiner, B.M., Costantini, F., 1997. The mouse Fused locus encodes Axin, an inhibitor of the Wnt signaling pathway that regulates embryonic axis formation. Cell 90, 181–192. Zorn, A.M., 1997. Cell–cell signaling: frog frizbees. Curr. Biol. 7, R501–R504.