Kekkon5 is an extracellular regulator of BMP signaling

Developmental Biology 326 (2009) 36–46

Contents lists available at ScienceDirect

Developmental Biology j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / d e v e l o p m e n t a l b i o l o g y

Kekkon5 is an extracellular regulator of BMP signaling Timothy A. Evans a, Harita Haridas b, Joseph B. Duffy b,⁎ a b

Department of Biology, Indiana University, Jordan Hall, 1001 E 3d Street., Bloomington, IN 47405, USA Department of Biology and Biotechnology, Worcester Polytechnic Institute, Life Sciences and Bioengineering Center, 60 Prescott Street, Worcester, MA 01609, USA

a r t i c l e

i n f o

Article history: Received for publication 14 May 2008 Revised 14 September 2008 Accepted 2 October 2008 Available online 18 October 2008 Keywords: Kek5 LIG BMP/TGF-β Wing Drosophila Crossveins Signal transduction LRR

a b s t r a c t Precise spatial and temporal control of Drosophila Bone Morphogenetic Protein (BMP) signaling is achieved by a host of extracellular factors that modulate ligand distribution and activity. Here we describe Kekkon5 (Kek5), a transmembrane protein containing leucine-rich repeats (LRRs), as a novel regulator of BMP signaling in Drosophila. We find that loss or gain of kek5 disrupts crossvein development and alters the early profile of phosphorylated Mad and dSRF in presumptive crossvein cells. kek5 phenotypic effects closely mimic those observed with Short gastrulation (Sog), but do not completely recapitulate the effects of dominant negative BMP receptors. We further demonstrate that Kek5 is able to antagonize the BMP ligand Glass bottom boat (Gbb) and that the Kek5 LRRs are required for BMP inhibitory activity, while the Ig domain is dispensable in this context. Our identification of Kek5 as a modulator of BMP signaling supports the emerging notion that LIG proteins function as diverse regulators of cellular communication. © 2008 Elsevier Inc. All rights reserved.

Introduction Precise spatial and temporal control of cellular signal transduction is an essential feature of animal development, and is often achieved by accessory proteins that influence the activity, availability, distribution or modification of core signaling components. A classic example of this is the Bone Morphogenetic Protein (BMP) signaling pathway, which regulates the patterning and growth of tissues across animal taxa. In Drosophila, the major BMP-like ligand Decapentaplegic (Dpp) plays major roles in patterning both the early embryo and the adult appendages (Segal and Gelbart, 1985). During larval stages, BMP signaling regulates the growth and patterning of the wing imaginal disc; later during pupal development the BMP pathway cooperates with the EGFR and Notch signaling pathways to control formation of the wing veins. Dpp cooperates with the additional BMP-like ligands Screw (Scw) and Glass bottom boat (Gbb) during embryonic and larval/pupal stages, respectively (Arora et al., 1994; Haerry et al., 1998; Khalsa et al., 1998). Current evidence suggests that heterodimeric ligands represent the primary contributors to in vivo signaling (Shimmi et al., 2005b). BMP signals in the wing are transduced by a serine/threonine kinase receptor complex combining the common type II receptor Punt (Put) with alternate type I receptors Thickveins (Tkv) and Saxophone (Sax) (Brummel et al., 1994; Letsou et al., 1995; Nellen et al., 1994; Penton et al., 1994; Ruberte et al., 1995; Xie et al., ⁎ Corresponding author. Fax: +1 508 831 5936. E-mail address: [email protected] (J.B. Duffy). 0012-1606/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.ydbio.2008.10.002

1994). Receptor activation results in phosphorylation of the cytoplasmic mediator Mothers against Dpp (Mad) (Sekelsky et al., 1995). BMP ligands are subject to extracellular regulation by a number of secreted factors: Short gastrulation (Sog) forms an inhibitory complex with BMP ligands along with Twisted gastrulation (Tsg, in embryos) or Tsg2 (in pupal wings), preventing short-range activation while promoting long-range signaling by facilitating ligand diffusion (Ashe and Levine, 1999; Mason et al., 1994; Ross et al., 2001; Shimmi et al., 2005a). Sog is cleaved by the tissue-specific metalloproteases Tolloid (Tld, in embryos) and Tolloid-related (Tlr, a.k.a. Tolkin, in pupal wings), releasing active ligand dimers (Marques et al., 1997; Serpe et al., 2005). BMP ligands produced in the longitudinal veins diffuse into crossvein-competent regions to initiate crossvein development during pupal stages (Ralston and Blair, 2005), making the crossveins highly sensitive to fluctuations in BMP signaling levels. This is especially true of the posterior crossvein (PCV), which must span a greater distance between BMP-producing longitudinal veins than the smaller anterior crossvein (ACV) (Ralston and Blair, 2005). Numerous recent reports have utilized the PCV to characterize new components of BMP signaling and to further investigate the relationships between existing members (Conley et al., 2000; Ralston and Blair, 2005; Serpe et al., 2005; Shimmi et al., 2005a; Vilmos et al., 2005). For example, molecular characterization of classic crossveinless mutants has revealed that crossveinless (cv) encodes Tsg2, which along with Sog mediates BMP ligand diffusion from the longitudinal veins to the PCV (Bridges, 1920; Shimmi et al., 2005a; Vilmos et al., 2005), while crossveinless-2 (cv-2) encodes a BMP-binding protein with Sog-like

T.A. Evans et al. / Developmental Biology 326 (2009) 36–46

cysteine-rich domains that potentiates BMP signaling in the PCV (Conley et al., 2000). BMP signaling during vein development is further modulated by the PS integrins, which in addition to their canonical role in adhesion influence extracellular Sog distribution in the pupal wing (Araujo et al., 2003). It is therefore becoming increasingly clear that a host of diverse factors cooperate to ensure proper spatiotemporal activation of the BMP pathway. Here we describe a novel regulator of BMP signaling during crossvein formation in Drosophila, Kekkon5 (Kek5). Kek5 is a transmembrane leucine-rich repeat (LRR) and immunoglobulin-like (Ig) domain containing (LIG) protein and one of six members of the Kek family, a suite of related transmembrane LIG proteins in Drosophila (MacLaren et al., 2004). Kek5 is expressed broadly throughout development and like kek1 and kek2 is enriched in the developing embryonic central nervous system (CNS) (Musacchio

37

and Perrimon, 1996). Loss of kek5 leads to defects in crossvein formation due to decreased BMP signaling in presumptive crossvein cells, as indicated by alterations in the profile of phosphorylated Mad (pMad) and dSRF in pupal wings. Similarly, increased kek5 expression disrupts BMP signaling in presumptive crossvein cells and causes loss of crossvein in adult wings. kek5 loss and gain of function phenotypes closely resemble those of sog yet are dissimilar to those of BMP receptors, suggesting that Kek5 acts upstream of the receptors to modulate pathway activity. Consistent with this idea, Kek5 is able to suppress phenotypes caused by increased BMP ligand activity. Finally, we demonstrate that the LRRs of Kek5 are required for its BMP inhibitory activity, perhaps indicating that Kek5 acts similarly to the vertebrate LRR proteins Tsukushi (Tsk) and Biglycan (Bgn) to regulate the activity and/or extracellular distribution of BMP ligands.

Fig. 1. kek5 gene structure and expression pattern. (A) Structure of the kek5 gene and Kek5 protein. Positions of P (P) and piggyBac (pB) transposable elements used to create deletion mutants are denoted by inverted triangles. Extent of deletions in kek5Δ1 and kek5fe148 are indicated below the gene. kek5Δ1 is an 840 bp deletion, beginning 426 bp upstream and ending 58 bp downstream of exon 1. kek5fe148 is a 48 kb deletion, removing the final exon of CG32533 and ending 9 bp upstream of the start of transcription of CG12200, located within the last intron of CG32533. The small predicted gene CG32827 resides within the 70 kb second intron of kek5. (B–J) Kek5 antiserum staining is shown in embryos (B, C, D), third instar wing discs (E, F), and pupal wings (G, I). Kek5 is broadly expressed at these developmental stages in wild type (B, E, G, I, J), while staining is undetectable in the null mutant kek5fe148 (D, F, H). Staining remains within the CNS of kek5Δ1 animals (C) but is undetectable in other tissues (not shown). Inset in (E) highlights membrane localization of Kek5 antisera. Kek5 is broadly expressed in pupal wings. G and I represent different optical sections and J is a magnified view with the vein intervein boundaries indicated (see dashed line and arrow) demonstrating Kek5 expression is present in both vein and intervein regions. For all embryos, anterior is left and dorsal is up. For all wing discs and pupal wings, proximal is left and anterior is up. Black boxes in (A) indicate untranslated regions; white boxes represent protein-coding sequences. N, N-terminal cysteine-rich cap; C, C-terminal cysteinerich cap; LRRs, leucine-rich repeats; Ig, immunoglobulin-like domain; Tm, transmembrane region; and IC, intracellular region. Black boxes numbered 1–6 represent conserved cytoplasmic sequence motifs.

38

T.A. Evans et al. / Developmental Biology 326 (2009) 36–46

Materials and methods Drosophila genetics Flies were raised on standard medium at 27 °C unless otherwise noted. GAL4/UAS stocks used included: P{GawB}apmd544 (ap-GAL4), P {GawB}559.1 (ptc-GAL4), P{en2.4-GAL4}e16E (en-GAL4), P{GawB} BxMS1096 (MS1096-GAL4), and P{GawB}CY2 (CY2-GAL4); P{UAS-sog} and P{UAS-supersog} (Yu et al., 2000), P{UAS-tkvA} (activated Tkv), P {UAS-saxA} (activated Sax), P{UAS-tkv1AΔGS} (dominant negative Tkv), and P{UAS-saxΔI} (dominant negative Sax) (Haerry et al., 1998), and P{UAS-gbb.K} (Khalsa et al., 1998). Kek5 and Kek4 misexpression constructs: full or partial ORFs were amplified via PCR and cloned using Gateway recombination methods (Invitrogen) into a pUAST-GFP destination vector. Transgenic flies were produced using standard germline transformation techniques (Spradling, 1986). Adult wings were dissected in isopropanol and mounted in 80% Canada balsam, and 20% methyl salicylate. Images were captured on a Zeiss Axiophot microscope. Generation of kek5 mutants The kek5Δ1 allele was generated via imprecise excision of the P {GT1}kek5BG02535 P element located within the first intron of kek5 and confirmed by PCR and sequence analysis. kek5fe148 was generated via FLP-mediated mitotic recombination between PBac{WH}f04013 and PBac{RB}CG32533e00904 (Parks et al., 2004). Putative deletions were confirmed by PCR analysis and antibody staining. Immunohistochemistry Embryos and imaginal discs were stained according to standard protocols. Embryos were fixed in PBS containing 10% paraformalde-

hyde and 50 mM EGTA; subsequent steps were carried out in PBS plus 0.5% NP-40 and 0.1% BSA. Imaginal discs were fixed in 1% formaldehyde, 100 mM PIPES, 2 mM MgSO4, 1 mM EDTA, and 0.5% NP-40; subsequent steps were carried out in PBS plus 0.5% NP-40 and 0.1% BSA. White prepupae were collected, aged, dissected in Drosophila Ringer's, and fixed 4–5 h in Brower fix (2% formaldehyde, 100 mM PIPES, 2 mM MgSO4, 1 mM EGTA, and 1% NP-40). Pupal wings were dissected away, blocked with 0.5% BSA for 1–3 h, incubated with primary antibody overnight, washed and incubated with secondary antibody for 1. 5 h, then washed and mounted in 70% glycerol/PBS plus Slowfade (Molecular Probes). All incubation and wash steps were carried out in PBS plus 0.3% Triton X-100 at 4 °C. To generate Kek5 antisera, a unique C-terminal Kek5 peptide (CGSTSQLFDDEGEDGTEV) was conjugated to keyhole limpet hemocyanin and used to immunize rabbits (Cocalico). Polyclonal antisera were pre-adsorbed twice at 1:10 dilution against kek5fe148 null embryos and used at a final dilution of 1:100. Anti-pMad antibody (Persson et al., 1998) and dSRF were used at 1:5000 and 1:1000 (Santa Cruz Biotechnology), respectively. Secondary antibodies used were: goat anti-rabbit Alexa488 and goat anti-rabbit Alexa568 (both 1:400; Molecular Probes). Fluorescence images were captured using a Leica TCS SP confocal microscope or Zeiss Axio Imager equipped with an ApoTome. Results Kek5 is a LIG family member expressed broadly in development The Drosophila proteome includes nine proteins that combine LRRs with Ig domains, termed LIGs (MacLaren et al., 2004). Six of these proteins exhibit significant sequence and structural similarity and represent the Kekkon (Kek) family. Within the Kek family, only Kek1's

Fig. 2. kek5 mutant wings exhibit defects in crossvein patterning and epithelial adhesion. Comparison of wild type (A) with kek5fe148 (B–F) wings. Anterior crossvein (ACV) and posterior crossvein (PCV) are indicated in panel A. kek5 mutant wing defects include truncated or missing PCV (arrow in panel B), ectopic PCV material (arrow in panel C), meandering PCV (arrow in panel D), ectopic crossveins around the ACV (arrow in panel E), and wing blisters (F). (G) Frequency of wing defects in kek5 mutants. Numbers indicate percentage of adult flies with defects in one or both wings. Percent crossvein defect includes all phenotypes shown in panels B–E. Phenotypes shown for kek5fe148 are representative of all genotypes listed.

T.A. Evans et al. / Developmental Biology 326 (2009) 36–46

role has been defined in vivo and its function as an EGFR inhibitor is unique among family members. In order to determine whether other Kek proteins might regulate distinct signaling pathways, we have undertaken an assessment of the additional family members and describe here our analysis of Kek5. The kek5 transcription unit covers 117 kb on the X chromosome and encodes a transmembrane molecule with an extracellular domain combining seven LRRs flanked by cysteine-rich regions with a single C2-type Ig domain, typical of Kek family proteins (Fig. 1A, Fig. S1) (MacLaren et al., 2004; Musacchio and Perrimon, 1996). The Kek5 cytoplasmic tail contains no apparent catalytic domain; however, comparison of orthologs in mosquito, honeybee and flour beetle revealed short, highly conserved sequence motifs spaced throughout this otherwise divergent region (Fig. S1) (MacLaren et al., 2004). This includes the last three amino acids of Kek5 (TEV), which represent a consensus binding sequence for class I PDZ domain proteins, suggesting a role in protein localization (Fig. S1) (Nourry et al., 2003). To investigate the function of Kek5, we initially characterized the Kek5 expression profile. In situ analysis of kek5 mRNA (not shown) is consistent with the protein expression patterns. Although they share a related structure, the overall expression pattern of Kek5 is distinct from that of Kek1 (Fig. 1) (Musacchio and Perrimon, 1996). Throughout embryonic and larval stages, Kek5 expression appears to be ubiquitous in most tissues, and antiserum staining is membrane-localized (Figs. 1B, E, G). Kek5 appears uniformly expressed in imaginal discs (wing, haltere, eye/antenna, and leg), and is widely expressed in both germline and somatic cells of the adult ovary (Fig. 1E and data not shown). This is in marked contrast to kek1, which is expressed in a restricted pattern indicative of EGFR activation within the follicular epithelium and imaginal discs (Musacchio and Perrimon, 1996). In addition, Kek1, Kek2, and Kek5 all exhibit high level expression in the embryonic central nervous system (CNS) (Fig. 1B) (Musacchio and Perrimon, 1996). Thus, Kek5 is a broadly expressed LIG protein that, like other Kek family members, is enriched in the developing embryonic CNS. Kek5 is required for proper crossvein patterning Unlike Kek1, Kek5 does not associate with the EGFR or influence its activity in vivo (Alvarado et al., 2004a). This, along with a more extensive expression pattern than Kek1, suggests a distinct function for Kek5. We therefore set out to determine the role of kek5 in vivo using both loss of function and gain of function approaches. To disrupt the function of kek5, we generated a series of small deletions by excising a transposable element (kek5BG02535) located 59 bp downstream of kek5 exon 1 (Fig. 1A), and a larger deletion through FLPmediated mitotic recombination between two FRT-containing piggyBac elements flanking the kek5 coding region (Parks et al., 2004). One of the small deletions, kek5Δ1, removes the entire first exon of kek5 and 427 bp of upstream sequence and abolishes kek5 expression in most tissues. However, mRNA and protein expression is still detectable within the embryonic and larval CNS of kek5Δ1 animals at comparable levels to wild type (Fig. 1C). In contrast, the larger deletion kek5fe148 lacks all kek5 coding sequence and expression is undetectable via antibody staining (Figs. 1D, F, H) or in situ hybridization (not shown) in kek5fe148 homo- and hemizygous animals, confirming its status as a null allele. Both alleles, kek5Δ1 and kek5fe148, produce fertile adults, similar to kek1 null mutations (Musacchio and Perrimon, 1996). In contrast to kek1, however, flies mutant for kek5 display a number of visible adult phenotypes and exhibit decreased viability. One of the most frequent phenotypes we observed in kek5 mutant flies is aberrant crossvein patterning (Fig. 2; 25–30% penetrance). Most often full or partial truncation of the posterior crossvein (PCV) occurs (Fig. 2B, ∼ 62%); less commonly the PCV appears twigged or forked (Fig. 2C, ∼ 16%), or the PCV ends meander from their correct

39

position (Fig. 2D, ∼ 8%). Occasionally the PCV appears normal and ectopic crossveins appear nearby. The anterior crossvein (ACV) is rarely deleted or truncated in kek5 mutants; more often partial or fulllength ectopic crossveins appear adjacent to the ACV (Fig. 2E, ∼ 14%). The severity and penetrance of these crossvein phenotypes are indistinguishable between kek5Δ1 and kek5fe148, consistent with the fact that both alleles abolish kek5 expression in the developing wing (Fig. 2G). Notably, we were unable to detect any longitudinal vein defects in kek5 mutant wings, indicating that the function of kek5 may be restricted to crossvein development. In both cases, the frequency of defects is enhanced by hemizygosity of our alleles with large deletions, which may simply reflect the notion that large deletions can often exacerbate sensitized phenotypes. We also note a second, apparently distinct, phenotype in kek5 mutants: wing blisters appear at low frequency (less than 5%; Fig. 2F). Although it occurs at low penetrance, this phenotype is specific to loss of kek5 as it does not appear in any of the progenitor strains used to generate kek5 mutants, and appears in both kek5Δ1 and kek5fe148alleles when homo- or hemizygous, or when either is placed over deficiencies for the region (Df(1)JA27 or Df(1)Exel7468). These wing blisters are not simply due to the presence of ectopic vein material, as we observed no correlation between the appearance of blisters and the presence of ectopic crossvein. This phenotype will not be described further here; but consistent with both phenotypes, Kek5 is expressed broadly in the pupal wing in developing vein and intervein regions (Figs. 1G, I, J). Below we focus on the role of kek5 in crossvein patterning. Inhibition of BMP signaling by Kek5 Specification of Drosophila crossveins is controlled by BMP signaling during pupal development. BMP ligands produced in the longitudinal veins diffuse into crossvein-competent regions to initiate crossvein development (Ralston and Blair, 2005), making the crossveins highly sensitive to fluctuations in BMP signaling levels. Loss of crossvein is characteristic of alterations in BMP signaling; in particular, mutations that interfere with BMP ligand distribution (e.g. sog, cv, tlr) produce crossvein-specific phenotypes similar to those we observe for kek5. The PCV is especially sensitive to these alterations (Conley et al., 2000; Ralston and Blair, 2005; Serpe et al., 2005; Shimmi et al., 2005a; Vilmos et al., 2005), raising the possibility that Kek5 may participate in BMP signaling during crossvein development. If this is the case, increased kek5 expression should alter BMP pathway activity and the appearance of vein. If Kek5 acts similarly to Sog, Tsg, and CV-2 to regulate BMP ligand activity or distribution, increased kek5 expression might promote or inhibit BMP signaling. Consistent with the latter possibility, misexpression of kek5 throughout the wing with a variety of GAL4 wing drivers (MS1096-GAL4, enGAL4, and ptc-GAL4) produced phenotypes indicative of decreased BMP signaling, i.e. missing crossveins and decreased wing size (Fig. 3). All misexpression phenotypes were highly penetrant and observed with multiple independent responder lines. In addition, we observed additional phenotypes associated with reductions in BMP signaling, including a strong ectopic scutellar bristle phenotype (Tomoyasu et al., 1998). These effects may also be due to downregulation of BMP signaling by Kek5, as similar phenotypes are produced by misexpression of Sog or dominant negative Tkv with ptc-GAL4 (Figs. 3P–S). While misexpression of Sog or Kek5 leads to local inhibition of crossvein formation, a notable distinction between their effects is that Sog misexpression also triggers the appearance of ectopic crossvein material and Kek5 does not (compare Figs. 3E, G to F, H). This ectopic vein reflects Sog's ability to act at a distance to potentiate BMP signaling, and appears exclusively outside the expression domain of the individual GAL4 drivers (ptc-GAL4 and en-GAL4; no ectopic vein appears with MS1096-GAL4, which is expressed throughout the entire wing blade).

40

T.A. Evans et al. / Developmental Biology 326 (2009) 36–46

Fig. 3. (A–H) Misexpression of Kek5 inhibits crossvein formation and mimics Sog. Kek5 (A) or Sog (B) misexpression throughout the entire wing with MS1096-GAL4 prevents formation of the ACV, a distinct phenotype from dominant negative forms of Sax (C) or Tkv (D), which induce more extensive defects in longitudinal vein formation and wing patterning. Similar Kek5 and Sog effects are produced in combination with ptc-GAL4 or en-GAL4 (E–H). Misexpression with en-GAL4 causes loss of PCV in addition to ACV, likely owing to higher expression levels of this driver. Notably, the ACV appears more sensitive than the PCV to increases in kek5 or sog expression, in contrast to loss of function conditions. (I–L) Kek5 promotes short-range inhibition and restricts long-range activation of BMP signaling by Sog. Removing kek5 from ptc N sog flies causes an increase in ectopic vein outside of the ptc-GAL4 expression domain, reflecting increased long-range BMP activation by Sog (I). In these flies, a full-length ectopic crossvein appears between L4 and L5 in 33% of wings (asterisk in panel I), but was never observed in ptc N sog flies. In contrast, adding UAS-kek5 to this background decreases the amount of ectopic vein (J). Misexpression of a highpotency form of Sog (Supersog) with ptc-GAL4 causes loss of intervein between L4 and L5 (arrow in panel K); this phenotype is suppressed by removal of kek5 (L). Asterisks (⁎ in F, H, I) indicate ectopic vein induced by Sog misexpression, which in all cases occurs outside of the GAL4 expression domain. (M–O) Inhibition of BMP signaling, tkvDN (M) is suppressed by removal of kek5 (N). Misexpression of tkvDN produces large gaps in L4 (class II-moderate defects) in 91% of the wings, which is suppressed ∼4.8 fold by removal of kek5 (O). (P–S) Kek5 can modulate BMP signaling during bristle formation. Wild type flies exhibit four prominent scutellar bristles (P, arrows). Ectopic scutellar bristles are induced in response to misexpression of kek5 (Q), sog (R), or dominant negative tkv (S).

T.A. Evans et al. / Developmental Biology 326 (2009) 36–46

41

strikingly, a full-length ectopic crossvein appears between L4 and L5 in 33% of kek5fe148; ptc N sog wings (asterisk in Fig. 3I); we never observed a full-length ectopic crossvein in this position in ptc N sog wings, although partial crossveins frequently appear there (Fig. 3F). We also observe a corresponding decrease in Sog-induced ectopic vein when Kek5 levels are increased in this background (Fig. 3J), consistent with Kek5 limiting Sog-dependent diffusion of BMP ligands. Providing further evidence for an inhibitory component to Kek5 activity in BMP signaling, loss of kek5 also was able to suppress BMP vein phenotypes induced via misexpression of a dominant negative version of Tkv in the wing (Fig. 3M–O) or a high-potency form of Sog (Figs. 3K, L). To more directly assay the effects of Kek5 on vein patterning and BMP signaling, we examined the profile of pMad and dSRF in pupal wings with altered Kek5 activity. Anti-pMad staining represents a readout of BMP pathway activity and delineates the presumptive longitudinal veins and crossveins of wild type pupal wings at 24 and 30 h APF (Figs. 4A, B). Conversely, dSRF expression delineates intervein, becoming reduced in presumptive vein, while remaining high in intervein regions. Consistent with kek5fe148 adult phenotypes, loss of PCV and ectopic ACV, we detected truncated anti-pMad staining within the presumptive PCV (Fig. 4E), and duplicated loss of dSRF in the ACV region in kek5 mutants (Fig. 5E). To determine if this Fig. 4. BMP pathway activation is altered by changes in Kek5 activity. Pupal wings at 24 and 30 h APF are shown. (A, B) Accumulation of pMad in wild type pupal wings delineates the presumptive longitudinal and cross veins. (C, D) When kek5 is misexpressed in a region overlapping the ACV, BMP signaling is abolished and the ACV fails to form (arrowhead). (E) In kek5 mutant wings, BMP signaling is often decreased in the presumptive PCV, leading to a truncation of this vein (arrowhead). Arrows indicate location of presumptive ACV.

Can Kek5 influence Sog's ability to induce ectopic vein at a distance? If Kek5 acts to restrict BMP ligand diffusion, then reducing its activity should enhance the appearance of ectopic vein associated with Sog misexpression. Consistent with this, loss of kek5 results in an increase in ectopic crossvein material in ptc N sog flies (Fig. 3I). Most

Fig. 5. dSRF expression is altered by changes in Kek5 activity. Pupal wings at 24 and 30 h APF are shown. (A, B) Loss of dSRF in wild type pupal wings delineates the presumptive longitudinal and cross veins. (C, D) When kek5 is misexpressed in a region overlapping the ACV, dSRF expression is not downregulated and the ACV fails to form (arrow). (E) dSRF expression in kek5 mutant wings, indicating the presence of an ectopic ACV. Arrows indicate location of presumptive ACV.

Fig. 6. Kek5 suppresses ligand-dependent increases in BMP signaling. Misexpression of Gbb with ptc-GAL4 induces a large patch of ectopic vein where the ACV would normally form (A). This phenotype is suppressed by addition of Kek5 (B), but not by addition of the related protein Kek4 (D), which has no effect on ACV formation on its own (C). Misexpression of activated forms of the BMP receptors Tkv (E) and Sax (G) induce more severe phenotypes including extensive ectopic vein, fusion of longitudinal veins, and loss of intervein regions. Kek5 is unable to suppress these phenotypes (F, H); instead we observe what appears to be an additive effect where these wings now exhibit an additional decrease in wing size characteristic of Kek5 misexpression (see Figs. 3A, E, G). Flies carrying tkvA were raised at 17 °C to avoid early lethality; all others were raised at 25 °C.

42

T.A. Evans et al. / Developmental Biology 326 (2009) 36–46

represented effects of Kek5 on BMP and not EGFR signaling, which also affects crossvein patterning at 30 h APF, we took advantage of the higher penetrance of the Kek5 misexpression phenotype and assayed the expression of pMAD and dSRF at 24 h APF, a timepoint at which pMAD and dSRF are under the sole control of BMP activity (Ralston and Blair, 2005). In ptc N kek5 wings at 24 and 30 h APF, anti-pMad staining was reduced or absent within the presumptive ACV, but remained high in the longitudinal veins and PCV (Figs. 4C, D). Similarly, dSRF was also affected — failing to be downregulated in the ACV in ptc N kek5 wings at 24 and 30 h APF (Figs. 5C, D). Thus changes in Kek5 activity alters BMP signaling in presumptive crossvein cells upstream of pMAD and dSRF expression. Notably, the effects of altered Kek5 activity on crossvein formation are similar to those observed for the BMP antagonist sog (Figs. 3B, F, H), but distinct from the effects of misexpressing dominant negative type I BMP receptors Thickveins (Tkv) or Saxophone (Sax), which additionally induce strong defects in longitudinal vein specification and wing patterning (Figs. 3C, D). Sog modulates BMP signaling by regulating the extracellular distribution and availability of BMP ligands. Accordingly, Sog misexpression is able to counteract ligand-dependent increases in BMP signaling, but has no effect on phenotypes caused by constitutively active (ligand-independent) forms of BMP receptors

(Yu et al., 2000). In contrast, the vertebrate BMP pseudoreceptor BAMBI acts by inhibiting receptor complexes and is able to antagonize both BMP ligands and constitutively active type I receptors (Onichtchouk et al., 1999). The phenotypic similarities between kek5 and sog, versus dominant negative receptors, suggest that Kek5 may likewise act upstream of BMP receptors. If this is indeed the case, misexpression of Kek5 should be able to suppress increases in pathway activity due to ectopic ligand production, but not misexpression of constitutively active BMP receptors. To test this, we examined the effects of misexpressing Kek5 along with BMP pathway components in the wing. Individually, misexpression of the BMP ligand Gbb with ptc-GAL4 induces an expanded region of ectopic vein where the ACV would normally form, while Kek5 misexpression removes the ACV completely (Figs. 3E, 6A). When Gbb and Kek5 misexpression are combined, ectopic vein is suppressed and the ACV is close to wild type (Fig. 6B). The suppressive effect of Kek5 on Gbb is not simply the result of GAL4 titration, nor is it a common feature of Kek proteins, as we observe no effect on the ptc N gbb phenotype when combined with UAS-kek4 (Figs. 6C, D). In contrast to the effects with Gbb, Kek5 is unable to suppress the misexpression effects of constitutively active forms of Sax or Tkv (Figs. 6E–H). Instead, we observe an additive effect: flies misexpressing both

Fig. 7. Structure-function analysis of Kek5. (A) Schematic of Kek5 deletion variants as described in the text. The ability to disrupt crossvein formation or induce ectopic scutellar bristles is indicated for each variant. Constructs lacking the LRRs (Kek5ΔL), transmembrane and cytodomains (sKek5), or cytodomain alone (Kek5ΔIC) fail to disrupt ACV formation or induce ectopic vein. Adding back the last six residues of the cytoplasmic tail (Kek5ΔIC + P) restores a lower level of activity. Deleting the Ig domain (Kek5ΔI) has no effect. (B–G) The Kek5 cytoplasmic domain contributes to protein localization and stability. GFP fluorescence in wing discs and egg chambers is shown for representative lines in combination with ptc-GAL4 and CY2-GAL4, respectively. Full-length Kek5 is expressed at high levels and appears tightly membrane-localized (B, E, E′). Similar expression levels and localization are observed for Kek5ΔIC + P, which lacks all but the final six residues of the cytoplasmic tail (D, G, G′). In contrast, Kek5ΔIC, in which the entire cytoplasmic domain has been deleted, appears to exhibit decreased stability and localization appears more diffuse (C, F, F′). (E–G) Superficial confocal slice showing lateral follicle cell membranes; (E′–G′) deep slice showing apical, basal, and lateral membranes. Anterior is left.

T.A. Evans et al. / Developmental Biology 326 (2009) 36–46

Kek5 and activated Tkv or Sax exhibit ectopic vein and wing patterning phenotypes comparable to activated receptor alone, plus a decrease in wing size similar to misexpression of Kek5 alone (Figs. 6E–H). Structural analysis implicates the LRRs and cytoplasmic domain in BMP signaling

43

2005; Serpe et al., 2005; Shimmi et al., 2005a; Vilmos et al., 2005). Here we have described a novel regulator of BMP signaling in Drosophila, Kek5, and added to the functions attributed to the LIG family, a group of proteins that have emerged in recent years as a unique class of signaling regulators. How does Kek5 regulate BMP signaling?

If Kek5 acts on extracellular components of BMP signaling, this function should require elements of the Kek5 ectodomain. All Kek proteins exhibit a characteristic modular arrangement of seven LRRs combined with a single Ig domain in their extracellular portions. However, elements of their cytoplasmic domains are also conserved and can contribute to their function, as observed for the KT box of Kek1 in EGFR signaling (Derheimer et al., 2004; Ghiglione et al., 2003). To address the relative contributions of Kek5 sequence elements, we assayed a series of Kek5 GFP-tagged variants for their ability to inhibit BMP signaling (Fig. 7). These include deletion of the LRRs (Kek5ΔL) or Ig domain (Kek5ΔI), a secreted form of Kek5 (sKek5) and complete (Kek5ΔIC) or partial (Kek5ΔIC + P) deletions of the cytoplasmic domain. Kek5ΔIC + P retains only the C-terminal six amino acids, including the conserved PDZ domain-binding site. Deleting the LRRs of Kek5 resulted in a near-complete loss of BMP inhibitory activity. In contrast, a form of Kek5 lacking the Ig domain was comparable to full-length Kek5 in its ability to inhibit crossvein formation. The LRRs alone are not sufficient for this activity, as a secreted form of Kek5 comprising the entire ectodomain exhibited no BMP inhibitory activity (Fig. 7A). Given the requirement for the LRRs, one possibility is that Kek5 might function similarly to the vertebrate LRR proteins Biglycan (Bgn) and Tsukushi (Tsk), which bind to the Sog homolog Chordin and modulate its affinity for BMP ligands (Moreno et al., 2005; Ohta et al., 2004). However, the inability of secreted Kek5 to inhibit BMP signaling suggests that membrane localization may be an integral requirement for Kek5 activity. Previous functional analyses of Kek1 have shown that the cytoplasmic tail, which includes a consensus PDZ-binding motif similar to Kek5, is important for subcellular trafficking and localization (Alvarado et al., 2004a; Ghiglione et al., 2003). Consistent with this, a variant Kek5 transgene lacking the entire cytoplasmic domain (Kek5ΔIC) was unable to reproduce the crossvein inhibition or ectopic bristle phenotypes of full-length Kek5, was not tightly localized to membranes and had significantly lower expression at the protein level than wild type (Figs. 7B–G′, data not shown). Adding back the Cterminal six amino acids (putative PDZ-binding motif — Kek5ΔIC + P), restored membrane localization and expression levels comparable to the full-length protein, but provided only a minimal increase in activity (Fig. 7D). Thus, as in the case of Kek1 the intracellular PDZbinding site contributes to Kek5 localization and/or protein stability. However, while the C-tail of Kek1 suffices as the sole intracellular component for inhibition of the EGFR, there appears to be additional inputs from the cytoplasmic domain of Kek5 necessary to impart full activity in BMP signaling, further distinguishing Kek5 function mechanistically from Kek1.

Our observations suggest that Kek5 may act similarly to Sog to regulate the extracellular distribution or activity of BMP ligands. Our data indicate that Kek5 may influence the ability of Sog to promote BMP ligand diffusion: in the absence of kek5, Sog's ability to promote long-range BMP signaling is augmented, while increased levels of kek5 abrogate this ability. Kek5 is able to antagonize the veinpromoting activity of the BMP ligand Gbb, but is unable to counteract ligand-independent increases in signaling caused by constitutively active forms of the BMP receptors Tkv and Sax, consistent with a model in which Kek5 alters extracellular activity of the pathway. We further demonstrate that the LRRs of Kek5 are required for its influence on BMP signaling, perhaps by mediating interactions between Kek5 and extracellular components of the BMP pathway. Precedence for such a model comes from recent work in vertebrates, in which two members of the Small Leucine-Rich Proteoglycan family (SLRP), Tsukushi (Tsk) and Biglycan (Bgn), were identified as novel extracellular components of BMP signaling (Moreno et al., 2005; Ohta et al., 2004). Both are small, secreted molecules containing 12 and 10 LRRs, respectively. In chick embryos, Tsk is thought to promote the formation of the primitive streak and Hensen's node through inhibition of BMP signaling. Its expression overlaps that of the BMP antagonist Chordin (Chd, the vertebrate ortholog of Sog), and Tsk forms a ternary complex with Chd and BMP capable of inhibiting BMP signaling (Ohta et al., 2004). Likewise, in Xenopus, Moreno et al. (2005) have shown that Bgn inhibits BMP signaling through the formation of a ternary complex including both Chd and BMP-4, where it appears to increase the affinity of Chd for BMP-4. Although it has not yet been demonstrated that the activities of these proteins are mediated by their LRRs, one possibility is that the LRRs of Tsk, Bgn and Kek5 act similarly to bind Chd/Sog and modulate its ligand affinity or diffusibility. The activities of both Tsk and Sog have been placed upstream of BMP receptors by showing that they can antagonize BMP ligands, but not constitutively active receptors (Ohta et al., 2004; Yu et al., 2000); in contrast, the vertebrate transmembrane BMP inhibitor BAMBI acts directly on vertebrate BMP receptors, and is able to suppress the phenotypes caused by constitutively active forms of these receptors (Onichtchouk et al., 1999). With Kek5, we observe an effect more similar to Tsk and Sog, supporting an extracellular role in BMP inhibition. It is important to note however, that our results do not exclude a more complex model invoking interactions with both extracellular and membrane associated components (i.e. ligands and receptors). Indeed, the requirement for membrane tethering of Kek5 and intracellular sequences in addition to the PDZ domain-binding site for full activity provides important support for this interpretation.

Discussion A model for Kek5 function during PCV specification Regulation of extracellular ligand activity and distribution is integral to achieving proper signaling levels in a number of conserved signal transduction pathways. BMP ligands produced in the longitudinal veins diffuse into crossvein-competent regions to specify development of the posterior crossvein (PCV) (Ralston and Blair, 2005), making this structure highly sensitive to fluctuations in BMP signaling levels. Recent studies focusing on BMP signaling during crossvein development have revealed novel regulatory components that provide precise control over BMP pathway activation, and have provided insight into the mechanisms underlying pathway regulation in a developmental context (Conley et al., 2000; Ralston and Blair,

In wild type wings, BMP ligand dimers produced in longitudinal vein cells diffuse into presumptive crossvein territories, where they are released from Sog- and Tsg2-dependent inhibition by the metalloprotease Tlr (Ralston and Blair, 2005; Serpe et al., 2005; Shimmi et al., 2005a; Vilmos et al., 2005). Thus sog, tsg2 (cv), and tlr-1 are required for PCV specification. Loss of kek5 results in both loss and ectopic crossvein, while misexpression triggers loss of crossveins. One simple interpretation is that the absence of kek5 leads to an altered distribution of extracellular components, thereby failing to ensure proper levels of BMP signaling in PCV cells. Perhaps Kek5 helps to

44

T.A. Evans et al. / Developmental Biology 326 (2009) 36–46

direct BMP ligand flux to the presumptive PCV territory through interactions with either the receptors or ligands themselves. In the absence of kek5, BMP ligand flux may be more evenly distributed across PCV and intervein cells, leading to decreased PCV signaling and slightly increased intervein signaling (causing both loss of PCV and ectopic vein as observed in kek5 mutant wings). In contrast, Kek5 misexpression may prevent ligand complexes from exiting the L4 territory where they are produced, thus acting as a potent crossvein inhibitor. Although we have to date been unable to detect specific biochemical interactions between Kek5 and secreted BMP pathway components, it is possible that Kek5 might interact only with a multicomponent complex, perhaps involving Sog, Tsg2, ligand heterodimers, and/or the receptors. Regardless, the phenotypic effects ascribed here to Kek5 suggest the mechanism by which it modulates BMP signaling will likely be distinct from that observed for Kek1's regulation of the EGFR. Kek5 versus Kek1: different pathways Kek1 functions as an EGFR inhibitor in vivo by directly binding the receptor and preventing its activation (Alvarado et al., 2004a; Ghiglione et al., 2003; Ghiglione et al., 1999). kek1 expression is induced by EGFR signaling, thereby allowing feedback attenuation of signaling levels (Ghiglione et al., 1999). We have previously reported that other Kek family members, including Kek5, do not inhibit EGFR signaling, demonstrating that this is not a common feature of the Kek family (Alvarado et al., 2004a). Here we show that Kek5 modulates the activity of a distinct signaling cascade, the Drosophila BMP pathway. This role, as that of Kek1 in EGFR signaling, appears unique among Kek family members as similar effects were not observed upon the loss or gain of the activity of other members of the Kek family (data not shown). Further distinguishing Kek5 from Kek1, Kek5 is broadly expressed throughout development and does not appear to be enriched in tissues or cells where BMP signaling is active, and thus is unlikely to represent a negative feedback loop for BMP signaling. The mechanisms by which Kek1 and Kek5 modulate signaling may also be divergent. Kek1 acts at the receptor level by directly binding and inhibiting EGFR; as such, Kek1 misexpression phenotypes mimic those of dominant negative forms of EGFR. In contrast, Kek5 appears unable to directly recapitulate the phenotypic effects of dominant negative forms of the BMP receptors Tkv or Sax. Although it's possible this reflects quantitative differences in direct receptor inhibition, increased expression of Kek5 does not simply result in phenotypes more akin to those observed with the receptor. Moreover, while Kek1 appears able to inhibit EGFR signaling in all developmental contexts, the same does not appear to hold true for Kek5. For example, while we observe BMP related phenotypes in the wing, we have been unable to detect any alterations in embryonic dorsoventral patterning in kek5 mutant or misexpression embryos (not shown). Together, our results support the idea that Kek5 and Kek1 modulate diverse signaling pathways in mechanistically distinct manners, with Kek1 acting solely via receptor binding and Kek5 likely acting in a more complex manner perhaps altering the distribution of extracellular BMP activity either directly or indirectly via regulation of receptor levels. LIG proteins act as mediators and modulators in diverse signal transduction pathways Kek1 was among the first proteins to be identified that combined LRRs with an Ig domain (Musacchio and Perrimon, 1996), sequence elements that commonly act as protein–protein interaction modules. Many proteins possessing LRRs or Ig domains act as cell adhesion molecules (CAMs), but the functional significance of combining these two motifs in a single protein is unknown. Considering the abundance of LRRs and Ig domains in metazoan proteomes, relatively few

proteins contain both (Pruess et al., 2003), and data addressing the function of these proteins has only recently begun to emerge. One apparent trend is that these LRR and Ig containing (LIG) proteins act as mediators and modulators in diverse signal transduction pathways, by directly transducing signals as in the case of Trk receptors (Teng and Hempstead, 2004) and NGL-1 (Lin et al., 2003), acting positively within a signaling complex as with Lingo-1 (Mi et al., 2004), or inhibiting the activity of receptors as do Kek1 (Alvarado et al., 2004a; Ghiglione et al., 2003; Ghiglione et al., 1999) and LRIG-1 (Gur et al., 2004; Laederich et al., 2004). Kek1 and LRIG-1 both act as modulators of the EGFR/ErbB signaling pathway, in flies and vertebrates respectively. Although parallels between the activities of Kek1 and LRIG-1 have been drawn, vertebrates lack an ortholog of Kek1 (LRIG1 is instead the vertebrate ortholog of the Drosophila protein Lambik) and binding and regulation of receptor activity by these molecules occurs via distinct mechanisms (Alvarado et al., 2004a; Alvarado et al., 2004b; Ghiglione et al., 2003; Ghiglione et al., 1999; Gur et al., 2004; Laederich et al., 2004; MacLaren et al., 2004), suggesting that they each independently acquired the ability to modulate EGFR/ErbB activity. The Drosophila genome encodes nine LIG proteins; of these five share sequence and structural similarity with Kek1 and represent the Kekkon (Kek) family. All six Keks are type I transmembrane proteins containing seven LRRs and one C2-type Ig domain in their extracellular portions (Adams et al., 2000; MacLaren et al., 2004), a common structural arrangement that differentiates them from Peroxidasin (Pxn) (Nelson et al., 1994), the fly LRIG-1 ortholog Lambik (Lbk) (Norga et al., 2003), and the predicted protein CG16974, which represent the three non-Kek Drosophila LIGs. The LRRs of Kek1 are necessary for its interaction with dEGFR (Alvarado et al., 2004a,b; Ghiglione et al., 2003), but no function has yet been ascribed to its Ig domain. We have previously shown that Kek1, Kek2, Kek5, and Kek6 are evolutionarily conserved in the insect lineage, and are able to associate in homo- and heterotypic fashions (MacLaren et al., 2004). Further, the ability of Kek1 to bind and inhibit dEGFR is unique within the family, raising the possibility that other Kek proteins may function as modulators of distinct signaling pathways (Alvarado et al., 2004a). Supporting this idea, we have shown that Kek5 acts in a distinct signaling pathway from Kek1 (BMP rather than EGFR) and likely through a different mechanism (regulation of ligand distribution/activity rather than direct receptor inhibition). The analysis of Kek5 presented here thus diversifies the activities associated with LIG proteins rather than supporting the notion that they function to modulate a single pathway. However, such diversification does not formally exclude the possibility of a common activity among family members, yet to be identified. As such, it will be important to determine whether other LIG molecules, including additional Kek proteins, act as modulators of signal transduction in concert or independently. Of the remaining four Keks, Kek4 and Kek6 are most similar to Kek5 (forming a subgroup within the Kek family), and represent the most likely candidates for similar function (MacLaren et al., 2004). However, preliminary misexpression and mutant analyses of Kek4 and Kek6 have not produced Kek5-like effects, and kek5-kek4 or kek5-kek6 double mutants display no dramatic enhancement of BMP-related kek5 phenotypes, arguing against redundant roles with Kek5 in the context of crossvein specification (T.A.E., unpublished). A number of the conserved Kek5 cytoplasmic motifs are shared with Kek6 (MacLaren et al., 2004), and our phylogenetic analyses further indicate that kek5 and kek6 are represented by a single gene in non-insect arthropods (T.A.E., unpublished), supporting the possibility of a common, as yet unidentified function between these closely related genes. A common thread uniting LIG proteins is expression in the nervous system. Indeed, the primary functions of a number of LIGs (e.g. Trk receptors, NGL-1, LINGO-1) appear to be restricted to the nervous system (Lin et al., 2003; Mi et al., 2004; Teng and Hempstead, 2004). Kek1, Kek2, and Kek5 are all highly enriched in the Drosophila

T.A. Evans et al. / Developmental Biology 326 (2009) 36–46

embryonic CNS, and Kek1 and Kek2 have been implicated in the processes of axon guidance and synapse formation, respectively (Guan et al., 2005; Speicher et al., 1998). It remains to be determined whether the roles of Kek1 and Kek5 in the CNS are related to or independent of their respective activities as modulators of EGFR and BMP signaling. One possibility is that the ancestral functions of Kek family members are as CNS CAMs and that Kek1 and Kek5 (at least) have subsequently acquired additional roles as signaling pathway modulators in other tissues. Acknowledgments We thank Ethan Bier, Kristi Wharton, and the Bloomington Drosophila Stock Center for fly stocks, Peter ten Dijke and Ed Laufer for antibodies, Amy Ralston for helpful discussions and technical advice, Mihaela Serpe and Michael O'Connor for assistance with preliminary Kek5-BMP biochemical interaction studies, Nick Klemen for generating the Kek5 LRR and Ig deletion constructs, and Christina Ernst, as well as other members of the Duffy lab for comments on the manuscript. This work was supported by NIH Genetics training grant fellowship GM07757 to T.A.E. and by NSF grants MCB-0616068 and IBN-0131707 to J.B.D. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.ydbio.2008.10.002. References Adams, M.D., Celniker, S.E., Holt, R.A., Evans, C.A., et al., 2000. The genome sequence of Drosophila melanogaster. Science 287, 2185–2195. Alvarado, D., Rice, A.H., Duffy, J.B., 2004a. Bipartite inhibition of Drosophila epidermal growth factor receptor by the extracellular and transmembrane domains of Kekkon1. Genetics 167, 187–202. Alvarado, D., Rice, A.H., Duffy, J.B., 2004b. Knockouts of Kekkon1 define sequence elements essential for Drosophila epidermal growth factor receptor inhibition. Genetics 166, 201–211. Araujo, H., Negreiros, E., Bier, E., 2003. Integrins modulate Sog activity in the Drosophila wing. Development 130, 3851–3864. Arora, K., Levine, M.S., O, Connor, M.B., 1994. The screw gene encodes a ubiquitously expressed member of the TGF-beta family required for specification of dorsal cell fates in the Drosophila embryo. Genes Dev. 8, 2588–2601. Ashe, H.L., Levine, M., 1999. Local inhibition and long-range enhancement of Dpp signal transduction by Sog. Nature 398, 427–431. Bridges, C.B., 1920. The mutant crossveinless in Drosophila melanogaster. Proc. Natl. Acad. Sci. U. S. A. 6, 660–663. Brummel, T.J., Twombly, V., Marques, G., Wrana, J.L., Newfeld, S.J., Attisano, L., Massague, J., O, Connor, M.B., Gelbart, W.M., 1994. Characterization and relationship of Dpp receptors encoded by the saxophone and thick veins genes in Drosophila. Cell 78, 251–261. Conley, C.A., Silburn, R., Singer, M.A., Ralston, A., Rohwer-Nutter, D., Olson, D.J., Gelbart, W., Blair, S.S., 2000. Crossveinless 2 contains cysteine-rich domains and is required for high levels of BMP-like activity during the formation of the cross veins in Drosophila. Development 127, 3947–3959. Derheimer, F.A., MacLaren, C.M., Weasner, B.P., Alvarado, D., Duffy, J.B., 2004. Conservation of an inhibitor of the epidermal growth factor receptor, Kekkon1, in dipterans. Genetics 166, 213–224. Ghiglione, C., Carraway 3rd, K.L., Amundadottir, L.T., Boswell, R.E., Perrimon, N., Duffy, J. B., 1999. The transmembrane molecule Kekkon 1 acts in a feedback loop to negatively regulate the activity of the Drosophila EGF receptor during oogenesis. Cell 96, 847–856. Ghiglione, C., Amundadottir, L., Andresdottir, M., Bilder, D., Diamonti, J.A., Noselli, S., Perrimon, N., Carraway, I.K., 2003. Mechanism of inhibition of the Drosophila and mammalian EGF receptors by the transmembrane protein Kekkon 1. Development 130, 4483–4493. Guan, Z., Saraswati, S., Adolfsen, B., Littleton, J.T., 2005. Genome-wide transcriptional changes associated with enhanced activity in the Drosophila nervous system. Neuron 48, 91–107. Gur, G., Rubin, C., Katz, M., Amit, I., Citri, A., Nilsson, J., Amariglio, N., Henriksson, R., Rechavi, G., Hedman, H., Wides, R., Yarden, Y., 2004. LRIG1 restricts growth factor signaling by enhancing receptor ubiquitylation and degradation. EMBO J. 23, 3270–3281. Haerry, T.E., Khalsa, O., O, Connor, M.B., Wharton, K.A., 1998. Synergistic signaling by two BMP ligands through the SAX and TKV receptors controls wing growth and patterning in Drosophila. Development 125, 3977–3987. Khalsa, O., Yoon, J.W., Torres-Schumann, S., Wharton, K.A., 1998. TGF-beta/BMP

45

superfamily members, Gbb-60A and Dpp, cooperate to provide pattern information and establish cell identity in the Drosophila wing. Development 125, 2723–2734. Laederich, M.B., Funes-Duran, M., Yen, L., Ingalla, E., Wu, X., Carraway 3rd, K.L., Sweeney, C., 2004. The leucine-rich repeat protein LRIG1 is a negative regulator of ErbB family receptor tyrosine kinases. J. Biol. Chem. 279, 47050–47056. Letsou, A., Arora, K., Wrana, J.L., Simin, K., Twombly, V., Jamal, J., Staehling-Hampton, K., Hoffmann, F.M., Gelbart, W.M., Massague, J., et al., 1995. Drosophila Dpp signaling is mediated by the punt gene product: a dual ligand-binding type II receptor of the TGF beta receptor family. Cell 80, 899–908. Lin, J.C., Ho, W.H., Gurney, A., Rosenthal, A., 2003. The netrin-G1 ligand NGL-1 promotes the outgrowth of thalamocortical axons. Nat. Neurosci. 6, 1270–1276. MacLaren, C.M., Evans, T.A., Alvarado, D., Duffy, J.B., 2004. Comparative analysis of the Kekkon molecules, related members of the LIG superfamily. Dev. Genes Evol. 214, 360–366. Marques, G., Musacchio, M., Shimell, M.J., Wunnenberg-Stapleton, K., Cho, K.W., O, Connor, M.B., 1997. Production of a DPP activity gradient in the early Drosophila embryo through the opposing actions of the SOG and TLD proteins. Cell 91, 417–426. Mason, E.D., Konrad, K.D., Webb, C.D., Marsh, J.L., 1994. Dorsal midline fate in Drosophila embryos requires twisted gastrulation, a gene encoding a secreted protein related to human connective tissue growth factor. Genes Dev. 8, 1489–1501. Mi, S., Lee, X., Shao, Z., Thill, G., Ji, B., Relton, J., Levesque, M., Allaire, N., Perrin, S., Sands, B., Crowell, T., Cate, R.L., McCoy, J.M., Pepinsky, R.B., 2004. LINGO-1 is a component of the Nogo-66 receptor/p75 signaling complex. Nat. Neurosci. 7, 221–228. Moreno, M., Munoz, R., Aroca, F., Labarca, M., Brandan, E., Larrain, J., 2005. Biglycan is a new extracellular component of the Chordin-BMP4 signaling pathway. EMBO J. 24, 1397–1405. Musacchio, M., Perrimon, N., 1996. The Drosophila kekkon genes: novel members of both the leucine-rich repeat and immunoglobulin superfamilies expressed in the CNS. Dev. Biol. 178, 63–76. Nellen, D., Affolter, M., Basler, K., 1994. Receptor serine/threonine kinases implicated in the control of Drosophila body pattern by decapentaplegic. Cell 78, 225–237. Nelson, R.E., Fessler, L.I., Takagi, Y., Blumberg, B., Keene, D.R., Olson, P.F., Parker, C.G., Fessler, J.H., 1994. Peroxidasin: a novel enzyme-matrix protein of Drosophila development. EMBO J. 13, 3438–3447. Norga, K.K., Patel, P.H., Huang, Y., Pan, H.L., Mackay, T.F.C., Bellen, H.J., 2003. Lambik, a novel putative LRR/immunoglobulin transmembrane protein, is a bristle number QTL that modulates Notch signaling. A. Dros. Res. Conf. 44, p. 768C. Nourry, C., Grant, S.G., Borg, J.P., 2003. PDZ domain proteins: plug and play. Sci. STKE 2003, RE7. Ohta, K., Lupo, G., Kuriyama, S., Keynes, R., Holt, C.E., Harris, W.A., Tanaka, H., Ohnuma, S., 2004. Tsukushi functions as an organizer inducer by inhibition of BMP activity in cooperation with chordin. Dev. Cell 7, 347–358. Onichtchouk, D., Chen, Y.G., Dosch, R., Gawantka, V., Delius, H., Massague, J., Niehrs, C., 1999. Silencing of TGF-beta signalling by the pseudoreceptor BAMBI. Nature 401, 480–485. Parks, A.L., Cook, K.R., Belvin, M., Dompe, N.A., Fawcett, R., Huppert, K., Tan, L.R., Winter, C.G., Bogart, K.P., Deal, J.E., Deal-Herr, M.E., Grant, D., Marcinko, M., Miyazaki, W.Y., Robertson, S., Shaw, K.J., Tabios, M., Vysotskaia, V., Zhao, L., Andrade, R.S., Edgar, K.A., Howie, E., Killpack, K., Milash, B., Norton, A., Thao, D., Whittaker, K., Winner, M.A., Friedman, L., Margolis, J., Singer, M.A., Kopczynski, C., Curtis, D., Kaufman, T.C., Plowman, G.D., Duyk, G., Francis-Lang, H.L., 2004. Systematic generation of highresolution deletion coverage of the Drosophila melanogaster genome. Nat. Genet. 36, 288–292. Penton, A., Chen, Y., Staehling-Hampton, K., Wrana, J.L., Attisano, L., Szidonya, J., Cassill, J.A., Massague, J., Hoffmann, F.M., 1994. Identification of two bone morphogenetic protein type I receptors in Drosophila and evidence that Brk25D is a decapentaplegic receptor. Cell 78, 239–250. Persson, U., Izumi, H., Souchelnytskyi, S., Itoh, S., Grimsby, S., Engstrom, U., Heldin, C.H., Funa, K., ten Dijke, P., 1998. The L45 loop in type I receptors for TGF-beta family members is a critical determinant in specifying Smad isoform activation. FEBS Lett. 434, 83–87. Pruess, M., Fleischmann, W., Kanapin, A., Karavidopoulou, Y., Kersey, P., Kriventseva, E., Mittard, V., Mulder, N., Phan, I., Servant, F., Apweiler, R., 2003. The proteome analysis database: a tool for the in silico analysis of whole proteomes. Nucleic Acids Res. 31, 414–417. Ralston, A., Blair, S.S., 2005. Long-range Dpp signaling is regulated to restrict BMP signaling to a crossvein competent zone. Dev. Biol. 280, 187–200. Ross, J.J., Shimmi, O., Vilmos, P., Petryk, A., Kim, H., Gaudenz, K., Hermanson, S., Ekker, S.C., O, Connor, M.B., Marsh, J.L., 2001. Twisted gastrulation is a conserved extracellular BMP antagonist. Nature 410, 479–483. Ruberte, E., Marty, T., Nellen, D., Affolter, M., Basler, K., 1995. An absolute requirement for both the type II and type I receptors, punt and thick veins, for dpp signaling in vivo. Cell 80, 889–897. Segal, D., Gelbart, W.M., 1985. Shortvein, a new component of the decapentaplegic gene complex in Drosophila melanogaster. Genetics 109, 119–143. Sekelsky, J.J., Newfeld, S.J., Raftery, L.A., Chartoff, E.H., Gelbart, W.M., 1995. Genetic characterization and cloning of mothers against dpp, a gene required for decapentaplegic function in Drosophila melanogaster. Genetics 139, 1347–1358. Serpe, M., Ralston, A., Blair, S.S., O, Connor, M.B., 2005. Matching catalytic activity to developmental function: tolloid-related processes Sog in order to help specify the posterior crossvein in the Drosophila wing. Development 132, 2645–2656. Shimmi, O., Ralston, A., Blair, S.S., O, Connor, M.B., 2005a. The crossveinless gene encodes a new member of the Twisted gastrulation family of BMP-binding proteins which, with short gastrulation, promotes BMP signaling in the crossveins of the Drosophila wing. Dev. Biol. 282, 70–83.

46

T.A. Evans et al. / Developmental Biology 326 (2009) 36–46

Shimmi, O., Umulis, D., Othmer, H., O, Connor, M.B., 2005b. Facilitated transport of a Dpp/Scw heterodimer by Sog/Tsg leads to robust patterning of the Drosophila blastoderm embryo. Cell 120, 873–886. Speicher, S., Garcia-Alonso, L., Carmena, A., Martin-Bermudo, M.D., de la Escalera, S., Jimenez, F., 1998. Neurotactin functions in concert with other identified CAMs in growth cone guidance in Drosophila. Neuron 20, 221–233. Spradling, A.C., 1986. P element-mediated transformation. In: Roberts, D.B. (Ed.), Drosophila: A Practical Approach. IRL Press, Oxford, pp. 175–197. Teng, K.K., Hempstead, B.L., 2004. Neurotrophins and their receptors: signaling trios in complex biological systems. Cell. Mol. Life Sci. 61, 35–48. Tomoyasu, Y., Nakamura, M., Ueno, N., 1998. Role of dpp signalling in prepattern

formation of the dorsocentral mechanosensory organ in Drosophila melanogaster. Development 125, 4215–4224. Vilmos, P., Sousa-Neves, R., Lukacsovich, T., Marsh, J.L., 2005. Crossveinless defines a new family of Twisted-gastrulation-like modulators of bone morphogenetic protein signalling. EMBO Rep. 6, 262–267. Xie, T., Finelli, A.L., Padgett, R.W., 1994. The Drosophila saxophone gene: a serine-threonine kinase receptor of the TGF-beta superfamily. Science 263, 1756–1759. Yu, K., Srinivasan, S., Shimmi, O., Biehs, B., Rashka, K.E., Kimelman, D., O, Connor, M.B., Bier, E., 2000. Processing of the Drosophila Sog protein creates a novel BMP inhibitory activity. Development 127, 2143–2154.