234, 93–106 (2001) doi:10.1006/dbio.2001.0238, available online at http://www.idealibrary.com on

CORE Metadata, citation and similar papers at core.ac.uk Provided by ElsevierVertebrate - Publisher Connector Related to Drosophila Naked Cuticle Bind Dishevelled and Antagonize Wnt Signaling

Keith A. Wharton, Jr.,*,†,‡,1 Gregor Zimmermann,*,† Raphae¨l Rousset,*,† and Matthew P. Scott*,†,2 *Department of Developmental Biology, †Department of , and ‡Department of Pathology, Howard Hughes Medical Institute, Beckman Center, B300, 279 Campus Drive, Stanford School of Medicine, Stanford, California 94305

Wnt signals control cell fate decisions and orchestrate cell behavior in metazoan animals. In the fruit fly Drosophila, defective in signaling mediated by the Wnt Wingless (Wg) exhibit severe segmentation defects. The Drosophila segment polarity naked cuticle (nkd) encodes an EF hand protein that regulates early Wg activity by acting as an inducible antagonist. Nkd antagonizes Wg via a direct interaction with the Wnt signaling component Dishevelled (Dsh). Here we describe two mouse and human proteins, Nkd1 and Nkd2, related to fly Nkd. The most conserved region among the fly and vertebrate proteins, the EFX domain, includes the putative EF hand and flanking sequences. EFX corresponds to a minimal domain required for fly or vertebrate Nkd to interact with the basic/PDZ domains of fly Dsh or vertebrate Dvl proteins in the yeast two-hybrid assay. During mouse development, nkd1 and nkd2 are expressed in multiple tissues in partially overlapping, gradient-like patterns, some of which correlate with known patterns of Wnt activity. Mouse Nkd1 can block Wnt1-mediated, but not ␤-catenin- mediated, activation of a Wnt-dependent reporter construct in mammalian cell culture. Misexpression of mouse nkd1 in Drosophila antagonizes Wg function. The data suggest that the vertebrate Nkd-related proteins, similar to their fly counterpart, may act as inducible antagonists of Wnt signals. © 2001 Academic Press Key Words: naked cuticle; Nkd; wingless; Wnt; dishevelled; dsh; Dvl; development; signal transduction; Drosophila; mouse; human; inducible antagonist.

INTRODUCTION pathways are repeatedly used in patterning and induction events. Similarly, Wg and Hh homologs are essential for a A majority of essential for developmental pattern- vast range of signaling events during vertebrate develop- ing in the fruit fly, Drosophila melanogaster, have verte- ment. Both pathways have been linked to human cancer, brate counterparts. The segment polarity genes, which particularly colon, breast, and epidermal appendage tumors encode signal transduction components for the secreted for the Wnt pathway (Morin, 1999) and basal cell carcinoma molecules Wingless (Wg, a Wnt protein) and Hedgehog (Hh), and medulloblastoma for the Hh pathway (Goodrich and are prime examples. Early in embryonic segmentation, each Scott, 1998). signaling protein maintains the production of the other in Many segment polarity genes, such as dishevelled (dsh) adjacent cells, forming a positive regulatory loop (DiNardo for the Wnt pathway, were first discovered in flies and et al., 1994). Later in development, the two signaling subsequently isolated in vertebrates (Klingensmith et al., 1996). Other Wnt pathway components were first discov- 1 To whom correspondence may be addressed in the Depart- ered in other organisms and later found in the fly. For ments of Pathology and Molecular Biology, University of Texas example, the first vertebrate Wnt gene, Wnt1, was isolated Southwestern Medical Center, 5323 Harry Hines Blvd. NB6.440, et al., ␤ Dallas, TX 75390-9072. Fax: 214-648-4070. E-mail: keith.wharton@ as a mouse (Nusse 1984), while -catenin, utsouthwestern.edu. a homolog of fly Armadillo (Arm), was initially described as 2 To whom correspondence may be addressed. Fax: 650-723- a cell adhesion molecule-associated protein (Ozawa et al., 9878. E-mail: [email protected]. 1989; Peifer et al., 1992). The similarities between fly

0012-1606/01 $35.00 Copyright © 2001 by Academic Press All rights of reproduction in any form reserved. 93 94 Wharton et al.

FIG. 1. Drosophila and vertebrate Nkd proteins share the EFX domain. (A) Schematic alignments of Recoverin (Rec, black), Drosophila Nkd (DNkd, yellow), and the mouse (m) and human (h) Nkd-related proteins Nkd1 (orange) and Nkd2 (green). Recoverin has four EF hands, numbered 1–4. DNkd shares sequence similarity with the third EF hand of Recoverin (3) (Zeng et al., 2000). DNkd shares two regions of similarity with the four vertebrate Nkd-related proteins: a histidine-rich C-terminus (H), and the EFX domain. The N-terminus of all four vertebrate proteins has a myristoyl consensus (horizontal stipple). mNkd1 has a rare alternatively spliced form without a myristoyl consensus (vertical stipple). Linewidth designates length of 100 amino acids. Sequence alignment below shows encoded amino acid sequences in DNkd and the four vertebrate Nkd proteins; identical amino acids are in black box, while conservative changes are in gray. The consensus EFX residues (EFX cons) are aligned with the consensus EF hand residues below (EF cons). Key: capital letters, standard amino acid codes; h, hydrophobic residue; b, basic residue; a, acidic residue; O, oxygen-donating residue in EF hand; *, variable residue; J, hydrophobic residue; X, Y, Z, residue that coordinates ion binding in x, y, and z axes. (B, C) Adult multiple-tissue Northern blot hybridized to mouse nkd1 (B) or mouse nkd2 (C) probe. Transcript size in kilobases (kb) is designated to the right of each blot. Key: B, brain; H, heart; K, kidney; L, lung; T, testis; S, skin. segment polarity genes and their vertebrate counterparts mouse and the human genomes (see Wnt homepage http:// have helped elucidate common signal transduction mecha- www.stanford.edu/ϳrnusse/wntwindow.html). In the de- nisms and have greatly aided in the understanding of both veloping mouse Wnt signals are typically produced in developmental biology and cancer. partially redundant and highly dynamic patterns through- Eighteen Wnt proteins have been found in both the out embryogenesis. Mutations in a few Wnt genes have

Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved. Vertebrate Naked Cuticle-Related Proteins 95

FIG. 2. Encoded amino acid sequences of vertebrate Nkd proteins. Boxshade alignment of the four vertebrate Nkd proteins. Residues identical in all four encoded proteins are in black box and conservative changes are in gray. Nkd1-specific residues are orange, and Nkd2-specific residues are green. Red bar shows the EFX domain. Consensus residues are designated as in Fig. 1. The cDNA sequences have been deposited to Genbank under the following Accession Nos.: mNkd1, AF358134; mNkd2, AF358136; hNkd1, AF358135; hNkd2, AF358137.

Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved. 96 Wharton et al. revealed their importance for murine development. For signaling is autoregulated via positive and negative feed- example, Wnt1 is required for midbrain/hindbrain forma- back loops (Perrimon and McMahon, 1999; Freeman, 2000). tion (McMahon and Bradley, 1990), Wnt3 for early gastru- We recently showed that naked cuticle (nkd), originally lation (Liu et al., 1999), Wnt3A for tailbud development isolated in the pioneering genetic screens of Nu¨sslein– (Takada et al., 1994), Wnt4 for renal development (Stark et Volhard and Wieschaus (Ju¨rgens et al., 1984), acts in a al., 1994), and Wnt7A for dorsal limb bud development (Parr negative feedback loop with Wg (Zeng et al., 2000). nkd was and McMahon, 1995). In addition, mutations affecting the one of the last segment polarity genes to be cloned and Wnt signal transduction components ␤-catenin, APC, or encodes a cytoplasmic Wg-inducible protein that can bind Axin that result in the accumulation of ␤-catenin can cause to and antagonize the function of Dsh (Rousset et al., 2001; cancer in mice and humans (Polakis, 1997; Rubinfeld et al., Zeng et al., 2000). This elegant control system, essential for 1997; Satoh et al., 2000). limiting the early effects of Wg in Drosophila segmenta- Insights into the complexity of Wnt signal regulation tion, allows the Nkd antagonist to accumulate in propor- have been gleaned from studies in model genetic organisms tion to Wg signal strength or duration. In this paper, we such as the nematode and fruit fly. Genetic analysis in the describe two mouse and human genes related to Drosophila fly originally ordered the functions of a handful of segment nkd. Our data suggest that these genes, similar to their fly polarity genes required for Wg signal transduction. In cells counterpart, may act as inducible antagonists for vertebrate that send the signal, porcupine (porc) is required for Wg Wnt signals. secretion (Kadowaki et al., 1996). In responsive cells the Wg signal is transmitted sequentially through the action of dsh, zeste-white 3 (zw3), and arm (Noordermeer et al., 1994; MATERIALS AND METHODS Siegfried et al., 1994). Similar genetic experiments in the nematode confirmed and extended the fly epistasis results (Han, 1997). Recently the number of protein families impli- Cloning of nkd-Related Genes cated in Wnt signaling has grown from 5 to over 40 (Wnt Mouse Nkd1, identified by sequence similarity to fly Nkd, was homepage), including the identification of Frizzled and originally found as expressed sequence tag (EST) W78547. The Arrow as Wnt coreceptors and TCF/Lef family proteins as insert from this EST was used as a probe to screen ␭-phage cDNA key transcriptional effectors (Bhanot et al., 1996; Brunner et libraries from 8.5 d.p.c. mouse embryos (generously provided by al., 1997; Wehrli et al., 2000). Many of these components Brigid Hogan) and adult mouse lung (Stratagene), and the clones are well conserved between nematode and human. harboring the entire open reading frame (L84 and L85) were A common target of Wnt signaling is Armadillo/␤- sequenced. A single phage clone retrieved twice (L16, L79) from ϫ 6 catenin (Peifer et al., 1992, 1991). In the absence of a Wnt screening 2 10 phage clones in the 8.5 d.p.c. library encoded an alternate 5Ј end that was spliced in frame with amino acid 8 of signal, ␤-catenin undergoes phosphorylation-dependent mNkd1 and contained almost the entire remainder of the mNkd1 ubiquitination and rapid degradation by proteasomes coding region. These unique sequences are also present on a mouse (Aberle et al., 1997). Wnt signaling antagonizes this phos- bacterial artificial clone encoding nkd1 (K. A. Whar- phorylation, resulting in ␤-catenin accumulation (Peifer et ton, unpublished data), suggesting that this represents a rare splice al., 1994) and nuclear translocation (Schneider et al., 1996). variant and not a chimeric clone. Once in the nucleus, ␤-catenin binds TCF/Lef family DNA- Mouse Nkd2 was originally identified as EST AA756769 and binding proteins to directly regulate many Wnt-responsive AA794228 because of their similarity to mNkd1 downstream of the genes (Behrens et al., 1996; Brannon et al., 1997; Molenaar EF hand. ␭-phage screens revealed two clones with alternate 3Ј et al., 1996; Riese et al., 1997). However, some Wnt proteins untranslated regions (L2-1 and L2-5). Each clone was sequenced, Ј Ј can act independently of ␤-catenin (Kengaku et al., 1998), or and ESTs now exist that extend these clones 3 and 5 . The ϩ remainder of the 5Ј sequence of mouse nkd2 was retrieved from even independently of transcription, to promote Ca2 fluxes EST clone BE291907, which represents a partially spliced clone. (Slusarski et al., 1997), control cell polarity (Han, 1997), or Human nkd1 was originally identified from EST H55148 and regulate cell shape changes (Heisenberg et al., 2000). subsequently from genomic scaffold sequence present on high- Because excess Wnt signaling can initiate unregulated throughout genome sequence (HTGS) clones AC007608 and cell proliferation, it is critically important to understand AC007334. Human nkd2 was originally identified from EST how Wnt signals are normally regulated during develop- AA101288 and subsequently in HTGS clones AC010446 and ment and throughout the life of the organism. In principle, AC016498. The human sequences in Fig. 2 were assembled using a proteins that regulate signal transduction can act extracel- combination of EST and genomic clone sequence based on their lularly, in the cytoplasm, or in the nucleus. In Wnt signal- high degree of similarity to the mouse sequences in coding regions ing, families of secreted or membrane-bound proteins and near-conserved splice junctions (not shown). Northern analysis was performed using hybridization conditions present in the extracellular space, including Frzb, Cerberus, specified by the manufacturer on a multiple tissue blots from Dickkopf, and Wif, as well as proteoglycans, regulate Wnt Stratagene and Origene (Fig. 2). A GAPDH probe was hybridized to distribution (Pfeiffer and Vincent, 1999). In the nucleus, the blot to ensure equal loading in all lanes (not shown). Groucho acts as transcriptional corepressor with TCF/Lef The sequence alignments were made using Clustal and linked in the absence of early Wg signals (Cavallo et al., 1998). Boxshade utilities that are available at http://dot.imgen.bcm.tmc.edu: A growing body of evidence suggests that developmental 9331/multi-align/multi-align.html.

Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved. Vertebrate Naked Cuticle-Related Proteins 97

Yeast Two-Hybrid and Binding Assays Studies were performed as previously described (Rousset et al., 2001). For the yeast two-hybrid assay, the bait plasmid was pAS2-1 and the prey plasmid was pACT2 (Clontech). The Nkd-EFX plas- mids correspond to amino acids 177–253 of DNkd, 125–190 of mNkd1, and 121–186 of mNkd2. The Dsh-basic/PDZ plasmids correspond to amino acids 167–338 of Dsh, 163–337 of mDvl1, 180–353 of mDvl2, and 161–335 of mDvl3 (all homologous regions based on Clustal alignments.) The transformation of yeast strain PJ69-4A (containing the ADE reporter gene) was achieved using a variation of the lithium acetate method (Clontech). Yeast growth on media Ϯ adenine was compared after 5–6 days at 30°C. We did not test the Dsh/Dvl sequences in the bait plasmids because Dsh activates transcription by itself when present in the bait plasmid (Rousset et al., 2001). For the GST pulldown assay, full-length DNkd and mNkd1 proteins were produced and labeled with 35S-methionine using the TNT T7-coupled reticulocyte lysate system (Promega) from pBlue- script IIKS(ϩ) derived constructs. The GST-Dsh plasmid has been described previously (Willert et al., 1997).

Whole-Mount In Situ Hybridization Embryos were harvested from pregnant C57BL6 female mice and immediately fixed in ice-cold 4% paraformaldehyde in phosphate- buffered saline (PBS). The entire procedure was performed as previously described (Goodrich et al., 1996). Embryos were photo- graphed under darkfield or brightfield on a Leica M10 stereomicro- scope. Sense and antisense probes were synthesized with compa- rable specific incorporation of digoxygenin-conjugated nucleotide, and the sense probes did not show any background staining (not shown). FIG. 3. Nkd EFX domains bind Dsh basic/PDZ domains. (A) Summary of yeast two-hybrid studies showing bait and prey Section In Situ Hybridization constructs and resultant growth in the presence or absence of the auxotrophic marker adenine (ADE). Growth in the absence of Paraffin sections were rehydrated through graded alcohols and adenine (ϪADE) indicates an interaction. Key: (ϩϩ), strong growth; washed in PBS. Frozen and paraffin sections were hybridized with (ϩ), growth; (Ϫ), no growth. (B) mNkd1 and fly Nkd bind fly Dsh in the digoxygenin-labeled probes overnight at 60°C and washed in the GST pulldown assay. Glutathione sepharose beads were incu- 50% formamide, 2ϫ SSC at 60°C, and then were incubated with bated with purified glutathione-S-transferase (GST) (lanes 1, 2) or a alkaline phosphatase-conjugated antidigoxigenin antibody (Roche) GST-Dsh fusion protein (lanes 3, 4) and mixed with in vitro overnight at 4°C. Sections were washed with PBS and visualized translated 35S-labeled Drosophila Nkd (DNkd, top) or mNkd1 with BM Purple-AP substrate (Roche). The slides were dehydrated (bottom). Labeled proteins present in the pellet (P) or supernatant with 95% ethanol, counterstained with Eosin Y (Fig. 5C), and (S) were run on denaturing SDS–PAGE, dried, and examined via photographed with a blue filter on a Zeiss Axiophot. autoradiography.

Wnt-1 Reporter Assay Cos7 or HEK293T cells were cultured in DMEM containing 10% fetal calf serum, 100 u/ml penicillin, 100 ␮g/ml streptomycin, 1 Cos7 cells (5 ϫ 104) were plated in each well of a 24-well plate and mM sodium pyruvate, 2 mM L-glutamate, and 0.1 mM nonessen- transfected with 0.02 ␮g TopFlash TCF luciferase reporter con- tial amino acids. Both cell types were transfected at subconfluence struct (Upstate Biotechnology), 0.02 ␮g pCDNA-lacZ, and the with Fugene 6 (Roche) as per manufacturer’s instructions. Wnt1- indicated amounts of pCNDA-mNkd1mycHis6. After 6–8 h, both dependent TopFlash activity was obtained by coculture of cultures were washed 3 times with phosphate-buffered saline, and HEK293T cells expressing Wnt1 and Cos7 cells cotransfected with the HEK293T cells were resuspended in 2 ml media. The resus- the TopFlash reporter, lacZ expression plasmid (pCDNA-lacZ; a pended HEK293T cells (40 ␮l) were added to each well of Cos7 cells gift of L. Collier), and either mouse nkd1 or empty vector expres- and then returned to the incubator. The luciferase assay was sion plasmids. All cDNA expression vectors employed the cyto- performed 12 h later on a luminometer using the Dual-Light megalovirus promoter to drive expression, and mouse nkd1 was Reporter Gene Assay System (Perkin–Elmer, as per manufacturer’s cloned into pCDNA as a myc-His6 fusion protein. For each instructions). Data was normalized for transfection efficiency experiment, 2 ϫ 106 HEK293T cells were transfected with 10 ␮gof based on lacZ expression. The total amount of DNA was kept the Wnt1 expression plasmid (pECG; a gift of B. Yu and R. Nusse). constant for all transfections in order to obtain comparable trans-

Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved. 98 Wharton et al.

fection efficiencies. ␤-catenin-induced activity was performed similarly, except that instead of the Wnt1 expression plasmid, 0.04 ␮gof␤-catenin expression plasmid (pRK5SK; a gift of B. Yu and R. Nusse) was transfected directly into the Cos-7 cells in each well.

Misexpression of mNkd1 in the Fly Myc-tagged (clone L16) and untagged (clone L84) versions of mouse nkd1 were cloned into pUAS-T and transformed into w or yw flies using standard methods as previously described (Zeng et al., 2000). Similar phenotypes were obtained using tagged and untagged constructs in multiple independent transformant lines. All crosses were performed at 29°C. E105-Gal4 and B119-Gal4 have been previously described (Zeng et al., 2000).

RESULTS

Identification of nkd-Related Genes Searches of public expressed sequence tag databases re- vealed mouse and human EF hand-encoding cDNA se- quences related to fly nkd (Fig. 1A and Materials and Methods). Subsequent phage library and public database screens revealed two nkd-related genes each in mouse (m) and human (h), here called nkd1 and nkd2 (Fig. 2). Comparison of the fly Nkd amino acid sequence with those of the vertebrate Nkd-related proteins reveals two regions of similarity. First, both fly and vertebrate Nkd proteins share a histidine-rich C-terminus (7/10 amino acids in the fly protein, 12/15 and 16/19 amino acids in vertebrate Nkd1 and Nkd2 proteins). While His-rich regions have been described in many proteins, the His repeats at the C-terminus of the vertebrate Nkd proteins are interrupted by conserved glutamate residues (Fig. 2). Second, the fly and vertebrate proteins share an approximately 66 amino acid region of similarity including the putative EF hand. We term this region the EFX domain (Fig. 1A). The amino acid sequences of the EFX domains are 42% identical between fly Nkd and mNkd1, and 41% identical between fly Nkd and mNkd2. EF hands are modular, helix-loop-helix Ca2ϩ binding

(s) expression. Red arrowhead designates dorsal CNS staining in C that is absent in D. (E, F) Right lateral view of trunk showing dorsal CNS stain in E (red arrowheads) that is absent in F. Blue arrowhead highlights tailbud expression that is similar in nkd1 and nkd2. (G, H) Right (G) and left (H) lateral views of bisected head showing generalized subepidermal mesenchyme expression that is down- regulated in primitive whisker follicles (blue arrowhead). As whis- FIG. 4. nkd1 and nkd2 expression patterns during mouse - ker follicle developments proceed mediolaterally, more mature genesis. Whole-mount in situ hybridization to 9.5 days postcoitum medially placed follicles express nkd1 (red arrowheads) but not (d.p.c.) (A–D), 10.5 d.p.c. (E, F), and (G–J) 12.5 d.p.c. mouse embryos nkd2. (I, J) Median sagittal view of bisected embryos (dorsal up) stained for nkd1 (A, C, E, G, I) or nkd2 (B, D, F, H, J). (A, B) Right showing expression of nkd1 (I) but not nkd2 (J) in Rathke’s pouch lateral view with head up. Red arrowheads designate presence (A) (red arrowhead), but similar expression in the subepidermal mes- or absence (B) of dorsal CNS staining. (C, D) Higher power of lower enchyme of the snout, soft palate, tongue (t), and mandible (blue part of embryos in A and B showing similar forelimb (fl) and somite arrowhead).

Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved. Vertebrate Naked Cuticle-Related Proteins 99

FIG. 5. Fly and vertebrate nkd patterns. (A) Whole-mount in situ hybridization of nkd in the third instar Drosophila wing pouch shows wing margin expression (large arrowhead) tapering off in a gradient on either side of the wing margin (small arrowheads). (B) Whole-mount in situ hybridization of 11.5 d.p.c. mouse embryo showing gradient-like expression of nkd1 in branchial arch and face elements. Eye placode is in the center. Highest expression is seen at arch boundaries (large arrowhead) with expression tapering off away from the boundaries (small arrowhead). (C, D) Similar gradient expression of mNkd in snout mesenchyme (C) in P0 (newborn) mouse skin and perichondrium (D) in 15.5 d.p.c. hindlimb revealed by section in situ hybridization. Note how expression peaks at epidermal/ mesenchyme (C) or cartilage/perichondrium (D) boundary (large arrowhead) and tapers away in the adjacent mesenchyme (small arrowhead). wf, whisker follicle; lc, longbone cartilaginous condensation. (E) Side view of 10.5 d.p.c. embryo forelimb stained with nkd1 showing greater expression in dorsal (D) than ventral (V) limb mesenchyme. nkd2 does not display this bias during this stage (not shown). Arrowhead, position of apical epidermal ridge.

motifs present in many Ca2ϩ binding proteins, including hand (EF3) of the recoverin subfamily of EF-hand proteins calmodulin and recoverin (Kawasaki et al., 1998). EF hand- (Zeng et al., 2000). Recoverins are N-terminally myristoy- containing proteins typically have between two and four lated, Ca2ϩ-dependent switch proteins that regulate the consecutive EF hands, but proteins with single EF hands localization and activity of a variety of enzymes from yeast have been described (Mochizuki et al., 1996). Previous to mammals (Burgoyne and Weiss, 2001). The EF hand sequence comparisons suggested that, among published region of mNkd1 is 35% identical over 53 amino acids to protein sequences, Drosophila Nkd shared the greatest Drosophila Frequenin (Pongs et al., 1993), the recoverin sequence similarity with the high-affinity Ca2ϩ binding EF subfamily EF hand protein most closely related to mNkd1

Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved. 100 Wharton et al.

(not shown). While the Drosophila Nkd EFX is about 40% identical over 66 amino acids to all of the vertebrate EFX sequences, the vertebrate EFX sequences are all 84% iden- tical to each other over the same region (Fig. 1A). In some characterized EF hand motifs, amino acid substitutions have occurred that preclude ion binding, indicating those domains do not serve as ion sensors. Both fly and vertebrate nkd sequences encode oxygen-donating residues in EF loop positions known to coordinate ion binding (Fig. 1A), al- though fly Nkd has an unusual pair of histidine residues at the apex of the loop that could, based on comparisons with EF hand-crystal structures, alter ion binding. We do not know whether any of the Nkd proteins bind Ca2ϩ or indeed whether they adopt standard EF hand conformations. Examination of all four vertebrate Nkd-related protein sequences (Fig. 2) reveals that all four vertebrate genes form a gene family with 36% overall amino acid identity. The vertebrate proteins share blocks of amino acid conservation on either side of the EF hand that are not obviously shared with fly Nkd (Fig. 2, and not shown). Supporting the idea that nkd1 and nkd2 represent distinct orthologs, the amino acid sequence of mNkd1 is 86% identical to hNkd1, but only 43% identical to hNkd2. Conversely, mNkd2 is only 43% identical to hNkd1 but 75% identical to hNkd2. Further, mNkd1 and hNkd1 have identical amino acids in 206/309 positions (66%; orange residues in Fig. 2) that are different in mNkd2 and hNkd2, excluding amino acid residues shared among all four encoded proteins. Similarly, mNkd2 and hNkd2 have identical amino acids in 164/309 positions (53%; green residues in Fig. 2) that are different in ␤ mNkd1 and hNkd1. All four Nkd-related proteins, similar FIG. 6. Nkd blocks Wnt1-induced, but not -catenin-induced, to the recoverins, have a myristoylation consensus at their transcription in cultured cells. Luciferase activity was determined after Cos7 cells transfected with the TopFlash Luciferase reporter N-terminus (MGKxxSK), although we do not know whether were incubated with Wnt1-expressing HEK293T cells (Wnt1) (A) or the vertebrate Nkd-related proteins are indeed myristoy- when the Cos7 cells were cotransfected with ␤-catenin (B). The lated. In contrast, fly Nkd does not have an N-terminal effect of increasing amounts of cotransfected mNkd1 plasmid (␮g myristoyl consensus sequence (Zeng et al., 2000). Accord- mNkd1) into Cos7 cells was evaluated for both activators. The ing to high-throughput genome sequence mapping data, level of Wnt1 or ␤-catenin-dependent activation was set at 100% in human nkd1 maps to 16q12, and human nkd2 maps to the presence of Wnt1 or ␤-catenin in the absence of mNkd1, and is 5p15.3, the latter of which we confirmed by radiation 5 to 7 times the level of activity seen in the absence of Wnt1, hybrid analysis (not shown). ␤-catenin, and mNkd1 (Basal). (A) Expression of increasing amounts of mNkd1 results in a dose-dependent block in luciferase activity to baseline levels. (B) Expression of comparable amounts of Nkd EFX Domains Bind Dishevelled Basic/PDZ mNkd1 does not block ␤-catenin-induced activation of luciferase Domains activity. For both graphs, data are expressed as normalized lucif- erase activity. Assays were performed in triplicate (bars ϭ standard We recently showed that in Drosophila, Nkd targets the deviation) and are representative of at least three independent Wnt signaling component Dishevelled (Dsh) via a direct experiments. interaction with the Dsh basic and PDZ (basic/PDZ) do- mains (Rousset et al., 2001). If the Nkd/Dsh interaction has been important for the control of Wnt signaling during , then the region in Drosophila Nkd that binds Dsh should be conserved in a vertebrate Nkd and bind the construct artifacts or differences in expression levels be- vertebrate Dsh homologs, Dvl1, Dvl2, and Dvl3. Yeast cause mNkd1 and mNkd2 EFX domains interacted equally two-hybrid studies showed that the EFX domains of fly and well with fly Dsh (Fig. 3A). We confirmed the interaction mouse Nkd proteins interacted with the basic/PDZ do- using glutathione-S-transferase (GST) “pull down” experi- mains of fly and mouse Dsh proteins (Fig. 3A). The mNkd2- ments. In vitro translated mNkd1, similar to fly Nkd, EFX was less effective than mNkd1-EFX at interacting with specifically bound glutathione sepharose beads to which a the Dvl proteins in this assay. This is unlikely to be due to GST-Dsh fusion protein was attached (Fig. 3B). In contrast,

Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved. Vertebrate Naked Cuticle-Related Proteins 101

GST-Dsh does not bind the control protein luciferase (Rous- cardial cushion, pulmonary epithelium, renal mesenchyme, set et al., 2001), and neither mNkd1 nor fly Nkd binds GST hair follicles, and in other tissues (Fig. 5 and not shown). by itself (Fig. 3B). To our knowledge, this is the first The mouse nkd expression patterns present three striking described example of an interaction between EF hand and features. First, nkd1 and nkd2 are expressed in a partially PDZ domains. Further, the fact that fly and vertebrate overlapping fashion, with nkd1 generally being expressed proteins interact equally well suggests that the Nkd/Dsh more broadly (or at higher levels) than nkd2. Second, in interaction has been conserved during approximately 300 many cases nkd1 and/or nkd2 transcripts are distributed in million years of evolution. a gradient emanating from a tissue boundary, such as the boundaries between branchial arches (Fig. 5B), skin epithe- lium and mesenchyme (Fig. 5C), and the somites (Figs. 4C Mouse nkd-Related Genes Are Expressed in Many and 4D). Third, there is some overlap between the distribu- Tissues in Gradients tion of the nkd genes and Wnt genes in places where Wnt We used Northern and in situ hybridization to study the functions have been well-studied. For example, Wnt1, distribution of nkd transcripts in mouse embryos and Wnt3, and Wnt3A are all expressed in the mouse dorsal tissues. Since many regions of the Nkd-related proteins are neural tube and serve partially redundant roles in neural identical by amino acid sequence (Fig. 1) and highly similar development (Ikeya et al., 1997; Parr et al., 1993). nkd1, but by DNA sequence (not shown), we synthesized probes not nkd2, RNA is highly enriched in the dorsal neural tube unique for mouse nkd1 and nkd2 to minimize potential (Figs. 4A, 4C, and 4E) and extends in a gradient into the cross-hybridization. These probes, made from the coding ventricular zone (not shown). Both nkd1 and nkd2 RNAs region corresponding to amino acids 278 and 450 in mNkd1 are expressed in the tailbud (Figs. 4E and 4F), where Wnt3A and amino acids 273 to 438 in mNkd2, were used for both is expressed and plays an essential role in development Northern and in situ hybridization studies (Figs. 1, 4, and 5). (Takada et al., 1994). In 10.5 d.p.c. embryos, nkd1 RNA is A blot of RNA from multiple adult tissues revealed a present in dorsal limb bud mesenchyme, which is the tissue broad distribution of nkd1 and more limited expression of targeted by a ␤-catenin-independent dorsalization signal nkd2 (Figs. 1B and 1C). The longest mouse nkd1 cDNA that mediated by Wnt7A (Kengaku et al., 1998; Riddle et al., contained the entire open reading frame was 1731 base pairs 1995). In Drosophila, we noted a striking correlation be- (bp), consistent with the predominant 1.8 kilobase (kb) band tween nkd and Wg distribution, in a pattern that mimics seen on the Northern blot (Fig. 1B). For nkd1, an additional the Wg activity gradient (Zeng et al., 2000) (Fig. 5A). rare spliced isoform was isolated as a cDNA; it encodes a Similarly, in mouse tissues, some gradient patterns of nkd unique 5Ј untranslated region (UTR) and initiator methio- expression bear a striking resemblance to the distribution of nine. This splice form is predicted to encode a slightly known Wnt activities. larger protein of 502 amino acids without an obvious N-terminal myristoyl consensus sequence and may corre- A Nkd-Related Protein Can Antagonize Wnt spond to the 2.5-kb band seen on the Northern blot. No Signaling human genomic or EST sequence has been found with significant similarity to this region, so this isoform may be Fly Nkd expressed in the developing fly or frog embryo specific to mouse or rodents. We have not identified the full acts as a Wnt antagonist (Zeng et al., 2000). We therefore sequence of the approximately 4.2-kb band on the Northern asked whether one of the vertebrate Nkd proteins, mNkd1, blot, although similar to mouse nkd2, it may correspond to is also capable of antagonizing Wnt activity. alternatively spliced 3Ј UTRs. Mouse nkd2 clones contain- Wnt signals converge on DNA sequence elements that ing three classes of 3Ј UTR generated via alternative splic- recognize TCF/Lef family proteins. One assay for Wnt ing were retrieved from EST and phage screens, consistent response employs a multimerized version of a TCF/Lef with the three bands seen on the Northern blot (Fig. 1C). binding site, cloned adjacent to a weak promoter driving a The longest contiguous nkd2 cDNA sequence is 3386 bp reporter gene in tissue culture. Incubation of reporter and might correspond to the 3.8-kb band seen on the construct-transfected Cos7 cells with cells expressing Wnt1 Northern blot. Shorter cDNA contigs, including the 5Ј leads to activation of the reporter gene (Fig. 6A). To test the coding region (see Materials and Methods), were 1917 and effect of mNkd1 on Wnt-dependent transcription, we co- 2718 bp, consistent with the 1.9- and 2.6-kb bands seen on transfected increasing amounts of mouse nkd1-expressing the Northern blot (Fig. 1C). plasmid and observed dose-dependent inhibition of the In 8.5 days postcoitum (d.p.c.) mouse embryos, nkd1 is Wnt-reporter construct (Fig. 6A). By contrast, transfection expressed in the neural folds along the entire rostrocaudal of comparable amounts of the mNkd1 plasmid did not axis (not shown). By 9.5 d.p.c., nkd1 and nkd2 transcripts affect ␤-catenin-induced activation of the Wnt reporter (Fig. accumulate throughout the forelimb buds, and at the ante- 6B), indicating that mNkd1, similar to fly Nkd (Rousset et rior and posterior of each somite boundary (Figs. 4C and al., 2001), acts upstream of ␤-catenin in the Wnt pathway. 4D). Branchial arch expression is evident by 9.5 d.p.c. and is We then asked whether mNkd1 can alter Wnt signaling well developed for both genes by 10.5 d.p.c. (Fig. 5B and not in Drosophila. We misexpressed mNkd1 using the binary shown). In later embryos, nkd1 is transcribed in the endo- Gal4/UAS system (Brand and Perrimon, 1993) and observed

Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved. 102 Wharton et al.

FIG. 7. Misexpression of mNkd1 in Drosophila results in wg-like adult phenotypes. (A) Ventral abdomen of B119-Gal4; UAS-mNkd1 adult female showing segmentally repeated sternite bristle clusters. One segment shows loss of a hemisternite bristle cluster (yellow arrow) and lateral displacement and disorientation of the contralateral bristle cluster (green arrow). (B) Ventral abdomen of typical wg, B119-Gal4; UAS-mNkd1 adult female showing consistent loss of some bristles in every segment with laterally displaced (green arrow) and missing (yellow arrows) hemisternite bristle clusters. (C) Wild-type leg. (D) Representative legs from wg, E105-Gal4; UAS-mNkd1 adults are variably truncated.

adults that were missing hemisternite bristle clusters (Fig. during metamorphosis (Shirras and Couso, 1996) and is also 7A) or had laterally displaced sternite bristles (not shown). seen in otherwise wild-type adults that have been exposed This phenotype is indicative of compromised wg signaling to low-level overexpression of fly Nkd (Zeng et al., 2000).

Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved. Vertebrate Naked Cuticle-Related Proteins 103

The effect of overproduced fly Nkd is enhanced by loss of nkd homologs in the completed Caenorhabditis elegans a wild-type copy of a positively acting component of Wg genome despite there being five Wnts and at least three signaling such as wg, dsh, or arm (Zeng et al., 2000). We Dsh-related proteins (see Wnt homepage). In the current therefore asked whether mNkd1 would be more potent in view of metazoan evolution, the common nematode/fly flies compromised for Wg signaling. We used two Gal4 ancestor is thought to be less ancient than the common driver lines, B119 and E105, that are effective lines for fly/vertebrate ancestor, implying that C. elegans must have revealing low-level antagonism of Wg activity by mutant discarded its nkd gene (de Rosa et al., 1999). Five of the forms of fly Nkd (K. A. Wharton, unpublished data). Pro- common Wnt signaling components were recently de- duction of mNkd1 using B119-Gal4 in heterozygous wg/ϩ scribed in a primitive metazoan, the Cnidarian Hydra animals consistently resulted in loss of hemisternite bristle (Hobmayer et al., 2000), so it will be interesting to see clusters (Fig. 7B). Some adults completely lacked sternite whether they too have a nkd gene. One possible explana- bristles (not shown), a phenotype similar to that seen when tion for the apparent absence of nkd from C. elegans is that fly Nkd is produced in wild-type animals using the same nkd may have evolved to control Wnt signaling over many Gal4 line (Zeng et al., 2000). When fly Nkd was produced more cell diameters than are usually employed by Wnt using E105-Gal4, most adults had severe leg truncations, signals in C. elegans. For example, mom-2 is required in the another phenotype associated with loss of wg (Zeng et al., P2 blastomere for spindle orientation and cell fate determi- 2000). Production of mNkd1 using E105-Gal4 produced rare nation in the immediately adjacent EMS blastomere (Han, adults with leg truncations (Ͻ1%). In wg/ϩ animals, E105- 1997). However, lin-44 controls polarized cell divisions over driven mNkd1 production resulted in up to 12% of adults many cell diameters (Herman et al., 1995), and egl-20 (n ϭ 158 adults) with at least one truncated leg (Fig. 7D). controls cell migrations occurring over half of the length of The mNkd1-induced phenotypes were also enhanced in the animal (Whangbo and Kenyon, 1999), so other mecha- adult Drosophila heterozygous for a dsh mutation (not nisms may limit Wnt signaling over many cell diameters in shown). Based upon our analysis of the transformations C. elegans. No additional genes closely related to nkd are observed with multiple Gal4 lines, the mNkd1-induced apparent in the completed D. melanogaster genome. There- phenotypes were qualitatively similar to, but less penetrant fore, redundancy between nkd and another related gene and expressive than, those induced by fly Nkd (not shown). cannot account for the absence of cell-autonomous pheno- Thus, mNkd1, similar to fly Nkd, can antagonize Wnt type in nkdϪ/Ϫ clones in many adult Drosophila tissues signaling in both vertebrate and invertebrate systems. (Zeng et al., 2000). One possible explanation for these fly and worm mysteries is that a protein with sequence distinct from Nkd might play similar regulatory role. DISCUSSION The relationship between wg and nkd in the fly may provide some clues to understand how the Nkd-related Our previous studies have shown that Drosophila nkd proteins might regulate Wnt activity vertebrates. In Dro- acts as an inducible antagonist for Wg signaling during early sophila embryos, Wg controls the transcription of the target segmentation (Zeng et al., 2000). nkd encodes a novel genes en and hh in cells adjacent to the cells that produce cytoplasmic protein that acts through the Wnt signaling Wg (Lee et al., 1992; Martinez Arias et al., 1988; Tabata et component Dsh (Zeng et al., 2000; Rousset et al., 2001). al., 1992). In nkd mutants, both en and hh are transcribed in The mammalian proteins we identify here appear to be broader stripes than in wild type, at a time when there is no Drosophila nkd orthologs by three criteria. First, the verte- apparent abnormality in Wg protein distribution (Martinez brate Nkd proteins share a conserved region with fly Nkd, Arias et al., 1988; Moline et al., 1999). One interpretation of the EFX domain, that is necessary and sufficient for the the abnormalities in en and hh expression in the nkd interaction between Nkd and the basic/PDZ region of Dsh embryo is that mutant cells are hypersensitive to Wg (Zeng (Rousset et al., 2001; Rousset et al., manuscript in prepara- et al., 2000). By analogy to the fly, a mammal with com- tion). Second, just as fly nkd is inducible by Wg signals promised nkd activity might be hypersensitive to certain (Zeng et al., 2000), the vertebrate nkd genes are expressed in Wnt signals. Mammalian Nkds may cooperate with or act gradient patterns reminiscent of Wnt signaling gradients. redundantly with other Wnt pathway antagonists such as Third, production of one of the mouse Nkd proteins, Axin and APC. Mouse axin mutants display both recessive mNkd1, in cultured cells or flies antagonizes Wnt- lethal and dominant visible phenotypes that suggest excess dependent readouts. The data are consistent with the idea Wnt signaling: dominant mutants display partial axis du- that these proteins, similar to fly Nkd, are inducible antago- plications, and recessive mutants display ectopic axes nists for vertebrate Wnt activity. (Perry et al., 1995). Homozygous mouse mutants harboring We show that the EFX domain conserved between fly and an APC mutation die during early gastrulation (Moser et al., mouse Nkd proteins corresponds to a minimal domain that 1995), possibly due to abnormal Wnt signaling. It will be of interacts with Dsh and the three mouse Dsh homologs, the interest to determine whether mice with mutations in Dvl proteins. The evolutionary conservation of the binding nkd1 and/or nkd2 exhibit similar phenotypes. domain supports the importance of the Nkd/Dsh interac- In some Drosophila tissues Wg is distributed in a gradient tion in Wnt signaling. Nonetheless, there are no apparent (Strigini and Cohen, 2000) and can act as a , i.e.,

Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved. 104 Wharton et al. in a concentration-dependent fashion, to regulate tissue- ing endogenous nkd levels in vertebrate animals results in specific target genes (Strigini and Cohen, 1999). Wg appears abnormal development or disease attributable to Wnt hy- to induce nkd transcription in many tissues and at different persensitivity. stages in embryos and imaginal discs (Zeng et al., 2000). In nkd several tissues is transcribed where the Wg signal has ACKNOWLEDGMENTS been received, indicating that nkd may be a common target of Wg signaling in the fly. The expression patterns of the We thank Ljiljana Milenkovic, Janine Hartford, Matt Fish, and mouse nkd genes suggest that the same may be true in Mei Zhang for expert technical assistance; Chris Karlovich and mammals. However, not all fly nkd transcription is depen- Rick Myers for radiation hybrid mapping of human nkd2; Ryan dent upon an inducing Wg signal. The initial transcription Rountree, Cathy Thut, David Kingsley, and Tony Oro for in situ of nkd in the embryo, for example, probably occurs in hybridizations; Lara Collier, Bo Yu, and Roel Nusse for expression response to early segmentation gene activities (Zeng et al., plasmids and cells; Brigid Hogan for the 8.5 d.p.c. mouse embryo 2000). Similarly only some of the mouse nkd transcription library; Peter Klein, Tony Oro, Elizabeth Hick, and Clifford Tabin patterns may be dependent on Wnt signals. for helpful comments on the manuscript, and the rest of the Scott Epistasis experiments employ double mutants to reveal lab for continued advice and support. K.W. was supported by a HHMI Postdoctoral Fellowship for Physicians and by the NIH (K08 relationships between genes with distinct phenotypes that HD 01164-04); G.Z. was supported by HHMI; R.R. was supported may act in a common biological pathway. Epistatic rela- by the Association pour la Recherche sur le Cancer and by the tionships are relatively straightforward to determine in the Human Frontier Science Program; M.P.S. is an Investigator of the fly and are often applicable to conserved vertebrate path- Howard Hughes Medical Institute. HHMI supported this research. ways. In wg mutants, in contrast to nkd mutants, en and hh transcription initiates normally but fades during the time when en and hh normally require Wg for maintenance of REFERENCES their expression (Bejsovec and Martinez Arias, 1991; Lee et Aberle, H., Bauer, A., Stappert, J., Kispert, A., and Kemler, R. (1997). al., 1992; Martinez Arias et al., 1988; Tabata et al., 1992). ␤-catenin is a target for the ubiquitin-proteasome pathway. Transcription of en in wg nkd double-mutant embryos EMBO J. 16, 3797–3804. fades as in wg single mutants (Bejsovec and Wieschaus, Behrens, J., von Kries, J. P., Kuhl, M., Bruhn, L., Wedlich, D., 1993), indicating that nkd requires intact Wg signaling to Grosschedl, R., and Birchmeier, W. (1996). Functional interac- reveal its loss of function phenotype. tion of ␤-catenin with the transcription factor LEF-1. 382, One interpretation of the similarity between the wg and 638–642. wg nkd double-mutant phenotypes is that removal of nkd Bejsovec, A., and Martinez Arias, A. (1991). Roles of wingless in has little effect if wg has already been inactivated, because patterning the larval epidermis of Drosophila. Development 113, 471–485. nkd is itself induced by the transduced Wg signal (Rousset Bejsovec, A., and Wieschaus, E. (1993). Segment polarity gene et al., et al., 2001; Zeng 2000). The epistatic relationship interactions modulate epidermal patterning in Drosophila em- between Axin and Wg in Drosophila has not yet been bryos. Development 119, 501–517. reported. However, embryos harboring a mutation in Dro- Bhanot, P., Brink, M., Samos, C. H., Hsieh, J. C., Wang, Y., Macke, sophila APC (D-APC2) undergo expansion of en expression J. P., Andrew, D., Nathans, J., and Nusse, R. (1996). A new that is indistinguishable from that seen in nkd mutants member of the frizzled family from Drosophila functions as a (McCartney et al., 1999). The phenotype of a D-APC2 Wingless receptor. Nature 382, 225–230. mutant embryo, similar to the nkd phenotype, requires Brand, A. H., and Perrimon, N. (1993). Targeted gene expression as intact Wg signaling: a D-APC2 wg mutant resembles a wg a means of altering cell fates and generating dominant pheno- mutant (McCartney et al., 1999). Because nkd and D-APC2 types. Development 118, 401–415. Brannon, M., Gomperts, M., Sumoy, L., Moon, R. T., and Kimel- have similar Wg-dependent phenotypes in Drosophila, ver- man, D. (1997). A ␤-catenin/XTcf-3 complex binds to the siamois nkd tebrate genes may be acting in some tissues in an promoter to regulate dorsal axis specification in Xenopus. Genes analogous fashion to vertebrate APC, a key tumor suppres- Dev. 11, 2359–2370. sor gene for colon cancer (Polakis, 1997). Brunner, E., Peter, O., Schweizer, L., and Basler, K. (1997). pangolin We recently proposed that Drosophila Nkd might control encodes a Lef-1 homologue that acts downstream of Armadillo to the sensitivity of Wg-responsive cells by regulating the transduce the Wingless signal in Drosophila. Nature 385, 829– ability of Dsh to transmit these signals from the Fz recep- 833. tors to the ␤-catenin “destruction complex” (Rousset et al., Burgoyne, R. D., and Weiss, J. L. (2001). The neuronal calcium 2ϩ 2001). Our demonstration that the vertebrate Nkd proteins sensor family of Ca -binding proteins. Biochem. J. 353, 1–12. can interact with the three vertebrate Dsh-related proteins Cavallo, R. A., Cox, R. T., Moline, M. M., Roose, J., Polevoy, G. A., Clevers, H., Peifer, M., and Bejsovec, A. (1998). Drosophila Tcf and, in the case of mNkd1, antagonize Wnt signals makes and Groucho interact to repress Wingless signalling activity. these nkd-related genes potentially pivotal regulators of Nature 395, 604–608. Wnt signals in vertebrates. Future experiments will seek to de Rosa, R., Grenier, J. K., Andreeva, T., Cook, C. E., Adoutte, A., understand the biochemical consequences of the Nkd/Dsh Akam, M., Carroll, S. B., and Balavoine, G. (1999). Hox genes in interaction during Wnt signaling, whether the nkd genes brachiopods and priapulids and protostome evolution. Nature are inducible by vertebrate Wnt signals, and whether alter- 399, 772–776.

Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved. Vertebrate Naked Cuticle-Related Proteins 105

DiNardo, S., Heemskerk, J., Dougan, S., and O’Farrell, P. H. (1994). McMahon, A. P., and Bradley, A. (1990). The Wnt-1 (int-1) proto- The making of a maggot: Patterning the Drosophila embryonic oncogene is required for development of a large region of the epidermis. Curr. Opin. Genet. Dev. 4, 529–534. mouse brain. Cell 62, 1073–1085. Freeman, M. (2000). Feedback control of intercellular signalling in Mochizuki, T., Wu, G., Hayashi, T., Xenophontos, S. L., development. Nature 408, 313–319. Veldhuisen, B., Saris, J. J., Reynolds, D. M., Cai, Y., Gabow, P. A., Goodrich, L. V., Johnson, R. L., Milenkovic, L., McMahon, J. A., Pierides, A., Kimberling, W. J., Breuning, M. H., Deltas, C. C., and Scott, M. P. (1996). Conservation of the hedgehog/patched Peters, D. J., and Somlo, S. (1996). PKD2, a gene for polycystic signaling pathway from flies to mice: Induction of a mouse kidney disease that encodes an integral membrane protein. patched gene by Hedgehog. Genes Dev. 10, 301–312. Science 272, 1339–1342. Goodrich, L. V., and Scott, M. P. (1998). Hedgehog and patched in Molenaar, M., van de Wetering, M., Oosterwegel, M., Peterson- neural development and disease. Neuron 21, 1243–1257. Maduro, J., Godsave, S., Korinek, V., Roose, J., Destree, O., and Han, M. (1997). Gut reaction to Wnt signaling in worms. Cell 90, Clevers, H. (1996). XTcf-3 transcription factor mediates 581–584. ␤-catenin-induced axis formation in Xenopus embryos. Cell 86, Heisenberg, C. P., Tada, M., Rauch, G. J., Saude, L., Concha, M. L., 391–399. Geisler, R., Stemple, D. L., Smith, J. C., and Wilson, S. W. (2000). Moline, M. M., Southern, C., and Bejsovec, A. (1999). Directional- Silberblick/Wnt11 mediates convergent extension movements ity of wingless protein transport influences epidermal patterning during zebrafish gastrulation. Nature 405, 76–81. in the Drosophila embryo. Development 126, 4375–4384. ␤ Herman, M. A., Vassilieva, L. L., Horvitz, H. R., Shaw, J. E., and Morin, P. J. (1999). -catenin signaling and cancer. Bioessays 21, Herman, R. K. (1995). The C. elegans gene lin-44, which controls 1021–1030. the polarity of certain asymmetric cell divisions, encodes a Wnt Moser, A. R., Shoemaker, A. R., Connelly, C. S., Clipson, L., Gould, protein and acts cell nonautonomously. Cell 83, 101–110. K. A., Luongo, C., Dove, W. F., Siggers, P. H., and Gardner, R. L. Hobmayer, B., Rentzsch, F., Kuhn, K., Happel, C. M., von Laue, (1995). Homozygosity for the Min allele of Apc results in disrup- C. C., Snyder, P., Rothbacher, U., and Holstein, T. W. (2000). tion of mouse development prior to gastrulation. Dev. Dyn. 203, WNT signalling molecules act in axis formation in the diploblas- 422–433. Noordermeer, J., Klingensmith, J., Perrimon, N., and Nusse, R. tic metazoan Hydra. Nature 407, 186–189. (1994). dishevelled and armadillo act in the wingless signalling Ikeya, M., Lee, S. M., Johnson, J. E., McMahon, A. P., and Takada, pathway in Drosophila. Nature 367, 80–83. S. (1997). Wnt signalling required for expansion of neural crest Nusse, R., van Ooyen, A., Cox, D., Fung, Y. K., and Varmus, H. and CNS progenitors. Nature 389, 966–970. (1984). Mode of proviral activation of a putative mammary Ju¨rgens, G., Wieschaus, E., Nu¨sslein-Volhard, C., and Kluding, H. oncogene (int-1) on mouse chromosome 15. Nature 307, 131– (1984). Mutations affecting the pattern of the larval cuticle in 136. Drosophila melanogaster. II. Zygotic loci on the third chromo- Ozawa, M., Baribault, H., and Kemler, R. (1989). The cytoplasmic some. Wilhelm Roux’s Arch. Dev. Biol. 193, 283–295. domain of the cell adhesion molecule uvomorulin associates Kadowaki, T., Wilder, E., Klingensmith, J., Zachary, K., and Perri- with three independent proteins structurally related in different mon, N. (1996). The segment polarity gene porcupine encodes a species. EMBO J. 8, 1711–1717. putative multitransmembrane protein involved in Wingless pro- Parr, B. A., and McMahon, A. P. (1995). Dorsalizing signal Wnt-7a cessing. Genes Dev. 10, 3116–3128. required for normal polarity of D-V and A-P axes of mouse limb. Kawasaki, H., Nakayama, S., and Kretsinger, R. H. (1998). Classi- Nature 374, 350–353. Biometals fication and evolution of EF-hand proteins. 11, 277– Parr, B. A., Shea, M. J., Vassileva, G., and McMahon, A. P. (1993). 295. Mouse Wnt genes exhibit discrete domains of expression in the Kengaku, M., Capdevila, J., Rodriguez-Esteban, C., De La Pena, J., early embryonic CNS and limb buds. Development 119, 247– Johnson, R. L., Belmonte, J. C. I., and Tabin, C. J. (1998). Distinct 261. WNT pathways regulating AER formation and dorsoventral Peifer, M., McCrea, P. D., Green, K. J., Wieschaus, E., and Gum- polarity in the chick limb bud. Science 280, 1274–1277. biner, B. M. (1992). The vertebrate adhesive junction proteins Klingensmith, J., Yang, Y., Axelrod, J. D., Beier, D. R., Perrimon, ␤-catenin and plakoglobin and the Drosophila segment polarity N., and Sussman, D. J. (1996). Conservation of dishevelled gene armadillo form a multigene family with similar properties. structure and function between flies and mice: isolation and J. Cell. Biol. 118, 681–691. characterization of Dvl2. Mech. Dev. 58, 15–26. Peifer, M., Rauskolb, C., Williams, M., Riggleman, B., and Wie- Lee, J. J., von Kessler, D. P., Parks, S., and Beachy, P. A. (1992). schaus, E. (1991). The segment polarity gene armadillo interacts Secretion and localized transcription suggest a role in positional with the wingless signaling pathway in both embryonic and signaling for products of the segmentation gene hedgehog. Cell adult pattern formation. Development 111, 1029–1043. 71, 33–50. Peifer, M., Sweeton, D., Casey, M., and Wieschaus, E. (1994). Liu, P., Wakamiya, M., Shea, M. J., Albrecht, U., Behringer, R. R., wingless signal and Zeste-white 3 kinase trigger opposing and Bradley, A. (1999). Requirement for Wnt3 in vertebrate axis changes in the intracellular distribution of Armadillo. Develop- formation. Nat. Genet. 22, 361–365. ment 120, 369–380. Martinez Arias, A., Baker, N. E., and Ingham, P. W. (1988). Role of Perrimon, N., and McMahon, A. P. (1999). Negative feedback segment polarity genes in the definition and maintenance of cell mechanisms and their roles during pattern formation. Cell 97, states in the Drosophila embryo. Development 103, 157–170. 13–16. McCartney, B. M., Dierick, H. A., Kirkpatrick, C., Moline, M. M., Perry, W. L., Vasicek, T. J., Lee, J. J., Rossi, J. M., Zeng, L., Zhang, Baas, A., Peifer, M., and Bejsovec, A. (1999). Drosophila APC2 is T., Tilghman, S. M., and Costantini, F. (1995). Phenotypic and a cytoskeletally-associated protein that regulates wingless sig- molecular analysis of a transgenic insertional allele of the mouse naling in the embryonic epidermis. J. Cell. Biol. 146, 1303–1318. Fused . Genetics 141, 321–332.

Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved. 106 Wharton et al.

Pfeiffer, S., and Vincent, J. P. (1999). Signalling at a distance: Siegfried, E., Wilder, E. L., and Perrimon, N. (1994). Components of transport of Wingless in the embryonic epidermis of Drosophila. wingless signalling in Drosophila. Nature 367, 76–80. Semin. Cell Dev. Biol. 10, 303–309. Slusarski, D. C., Yang-Snyder, J., Busa, W. B., and Moon, R. T. Polakis, P. (1997). The adenomatous polyposis coli (APC) tumor (1997). Modulation of embryonic intracellular Ca2ϩ signaling by suppressor. Biochim. Biophys. Acta 1332, F127–F147. Wnt-5A. Dev. Biol. 182, 114–120. Pongs, O., Lindemeier, J., Zhu, X. R., Theil, T., Engelkamp, D., Stark, K., Vainio, S., Vassileva, G., and McMahon, A. P. (1994). Krah-Jentgens, I., Lambrecht, H. G., Koch, K. W., Schwemer, J., Epithelial transformation of metanephric mesenchyme in the Rivosecchi, R., et al. (1993). Frequenin—A novel calcium- developing kidney regulated by Wnt-4. Nature 372, 679–683. binding protein that modulates synaptic efficacy in the Drosoph- Strigini, M., and Cohen, S. M. (1999). Formation of morphogen ila nervous system. Neuron 11, 15–28. gradients in the Drosophila wing. Semin. Cell Dev. Biol. 10, Riddle, R. D., Ensini, M., Nelson, C., Tsuchida, T., Jessell, T. M., 335–344. and Tabin, C. (1995). Induction of the LIM gene Lmx1 Strigini, M., and Cohen, S. M. (2000). Wingless gradient formation by WNT7a establishes dorsoventral pattern in the vertebrate in the Drosophila wing. Curr. Biol. 10, 293–300. limb. Cell 83, 631–640. Tabata, T., Eaton, S., and Kornberg, T. B. (1992). The Drosophila Riese, J., Yu, X., Munnerlyn, A., Eresh, S., Hsu, S. C., Grosschedl, hedgehog gene is expressed specifically in posterior compart- R., and Bienz, M. (1997). LEF-1, a nuclear factor coordinating ment cells and is a target of engrailed regulation. Genes Dev. 6, signaling inputs from wingless and decapentaplegic. Cell 88, 2635–2645. 777–787. Takada, S., Stark, K. L., Shea, M. J., Vassileva, G., McMahon, J. A., Rousset, R., Mack, J. A., Wharton, K. A., Jr., Axelrod, J. D., Cadigan, and McMahon, A. P. (1994). Wnt-3a regulates somite and tailbud K., Fish, M. P., Nusse, R., and Scott, M. P. (2001). naked cuticle formation in the mouse embryo. Genes Dev. 8, 174–189. targets dishevelled to antagonize Wnt signal transduction. Genes Wehrli, M., Dougan, S. T., Caldwell, K., O’Keefe, L., Schwartz, S., Dev. 15, 658–671. Vaizel-Ohayon, D., Schejter, E., Tomlinson, A., and DiNardo, S. Rubinfeld, B., Albert, I., Porfiri, E., Munemitsu, S., and Polakis, P. (2000). arrow encodes an LDL-receptor-related protein essential (1997). Loss of ␤-catenin regulation by the APC tumor suppressor for Wingless signalling. Nature 407, 527–530. protein correlates with loss of structure due to common somatic Whangbo, J., and Kenyon, C. (1999). A Wnt signaling system that mutations of the gene. Cancer Res. 57, 4624–4630. specifies two patterns of cell migration in C. elegans. Mol. Cell. Satoh, S., Daigo, Y., Furukawa, Y., Kato, T., Miwa, N., Nishiwaki, 4, 851–858. T., Kawasoe, T., Ishiguro, H., Fujita, M., Tokino, T., Sasaki, Y., Willert, K., Brink, M., Wodarz, A., Varmus, H., and Nusse, R. Imaoka, S., Murata, M., Shimano, T., Yamaoka, Y., and Naka- (1997). Casein kinase 2 associates with and phosphorylates mura, Y. (2000). AXIN1 mutations in hepatocellular carcinomas, dishevelled. EMBO J. 16, 3089–3096. and growth suppression in cancer cells by virus-mediated trans- Zeng, W., Wharton, K. A., Jr., Mack, J. A., Wang, K., Gadbaw, M., fer of AXIN1. Nat. Genet. 24, 245–250. Suyama, K., Klein, P. S., and Scott, M. P. (2000). naked cuticle Schneider, S., Steinbeisser, H., Warga, R. M., and Hausen, P. (1996). encodes an inducible antagonist of Wnt signalling. Nature 403, ␤-catenin translocation into nuclei demarcates the dorsalizing 789–795. centers in frog and fish embryos. Mech. Dev. 57, 191–198. Shirras, A. D., and Couso, J. P. (1996). Cell fates in the adult Received for publication February 22, 2001 abdomen of Drosophila are determined by wingless during pupal Accepted February 23, 2001 development. Dev. Biol. 175, 24–36. Published online April 17, 2001

Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved.