D-Six4 Plays a Key Role in Patterning Cell Identities Deriving from the Drosophila Mesoderm

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

Developmental Biology 294 (2006) 220–231 www.elsevier.com/locate/ydbio

D-six4 plays a key role in patterning cell identities deriving from the Drosophila mesoderm

Ivan B.N. Clark a,b, Joanna Boyd a, Graham Hamilton c, ⁎ David J. Finnegan b, , Andrew P. Jarman a

a Centres for Integrative Physiology and Neuroscience Research, University of Edinburgh, George Square, Edinburgh EH8 9XD, UK b School of Biological Sciences, Kings Buildings, University of Edinburgh, Edinburgh EH9 3JR, UK c Division of Molecular Genetics, Institute of Biomedical and Life Sciences, University of Glasgow, Anderson College Complex, 56 Dumbarton Road, Glasgow G11 6NU, UK Received for publication 6 June 2005; revised 24 February 2006; accepted 27 February 2006 Available online 3 April 2006

Abstract

Patterning of the Drosophila embryonic mesoderm requires the regulation of cell type-specific factors in response to dorsoventral and anteroposterior axis information. For the dorsoventral axis, the homeodomain gene, tinman, is a key patterning mediator for dorsal mesodermal fates like the heart. However, equivalent mediators for more ventral fates are unknown. We show that D-six4, which encodes a Six family transcription factor, is required for the appropriate development of most cell types deriving from the non-dorsal mesoderm — the fat body, somatic cells of the gonad, and a specific subset of somatic muscles. Misexpression analysis suggests that D-Six4 and its likely cofactor, Eyes absent, are sufficient to impose these fates on other mesodermal cells. At stage 10, the mesodermal expression patterns of D-six4 and tin are complementary, being restricted to the dorsal and non-dorsal regions respectively. Our data suggest that D-six4 is a key mesodermal patterning mediator at this stage that regulates a variety of cell-type-specific factors and hence plays an equivalent role to tin. At stage 9, however, D-six4 and tin are both expressed pan-mesodermally. At this stage, tin function is required for full D-six4 expression. This may explain the known requirement for tin in some non-dorsal cell types. © 2006 Elsevier Inc. All rights reserved.

Keywords: SIX genes; Transcriptional regulation; Patterning; Mesoderm; Drosophila

Introduction the identity and function of the upstream regulatory mediators themselves. A fundamental question in developmental biology concerns The specification of the mesoderm in Drosophila melano- the means by which uncommitted cells become specified to gaster provides a tractable model system in which to study how form a diversity of tissues according to their spatial location. the expression of cell type regulators is patterned within a large In general, it is clear that a relatively small number of group of cells that are initially identical. A diverse range of signaling and transcription factors are expressed in response to organs derives from the Drosophila mesoderm, including the positional information, and in turn, these act combinatorially heart, the somatic and visceral muscles, the fat body and the to regulate the expression of more specialized cell type somatic component of the gonad. For parasegments 4–12, an regulatory factors. There is much interest in understanding the approximate fate map can be constructed outlining the meso- combinatorial regulation of cell type factors, particularly dermal regions that give rise to these organs (Fig. 1). Trans- through computational analysis of their cis-regulatory regions. plantation experiments show that fate determination is This is hampered, however, by an incomplete understanding of dependent on cell position, and therefore, patterning the mesoderm requires positional information (Beer et al., 1987). This is provided in part by inductive signaling from the overlying ⁎ Corresponding author. ectoderm, which results in the establishment of specific E-mail address: [email protected] (D.J. Finnegan). expression patterns for mesodermal transcription factors.

visceral mesoderm is formed in the dorsal eve domain through the activation of bagpipe by Tin and Dpp and its repression by wg/slp (Azpiazu and Frasch, 1993; Azpiazu et al., 1996; Lee and Frasch, 2005). Apart from its dorsal function, tin also has a poorly understood role in the development of more ventral mesodermal fates (Azpiazu and Frasch, 1993). Outside the dorsal domain, it has been suggested that the non-dorsal mesoderm is divided into ventral and dorsolateral domains (Riechmann et al., 1998). This was based on the response of fat body cells to Wg signaling, although it is not clear whether this distinction has a genetic basis. The dorsolateral domain contains cells with dual fat body/SGP competence, although normally only those cells in parasegments 10–12 take on an SGP fate. Apart from the role of Tin dorsally, the patterning of mesodermal fates in the dorsoventral Fig. 1. A schematic fate map of one parasegment of the Drosophila trunk axis is poorly understood. A major unanswered question is mesoderm. The dashed boxes indicate the domains in which the precursors of different larval tissues are patterned (representing parasegments 4–12 only). whether there is a factor in the non-dorsal mesoderm that Each parasegment is divided into two regions along the anteroposterior axis, performs functions complementary to those of Tin in the dorsal termed the eve and slp domains (Riechmann et al., 1997). Along the region. A candidate for such a factor is the Six family dorsoventral axis, three functional domains have be proposed: dorsal, homeodomain protein, D-Six4. In mouse, Six1, Six4, and Six5 dorsolateral, and ventral. The dorsolateral and ventral domains are distinguished genes are coexpressed during myogenesis, while Six1 and Six4 by the opposite effects of wg as opposed to en and hh signaling on fat body development (Riechmann et al., 1998). Approximate positions of the stripes of at least are required during mesoderm development (Grifone et en, hh, and wg expression in the overlying ectoderm are indicated. Ectodermal al., 2005). In human, reduction of SIX5 expression may underlie dpp signaling maintains expression of tin in the dorsal mesoderm from stage 10. some of the abnormalities associated with Type 1 Myotonic The dorsolateral eve domain gives rise to the somatic gonadal precursors (SGPs) Dystrophy (DM1) (Fillipova et al., 2001; Klesert et al., 1997; in parasegments 10–12 and to the primary clusters of fat body precursors in Thornton et al., 1997). In Drosophila, the sole Six4/Six5 more anterior parasegments. This figure is based on Fig. 1 from Riechmann et al. (1998), and the caveats mentioned therein apply. homologue, D-Six4, is the only Six homeoprotein expressed in the early mesoderm, and its mutation disrupts gonad and muscle development (Seo et al., 1999; Kirby et al., 2001). Along the anteroposterior axis, the parasegmental mesoderm Here, we present evidence that D-Six4 is a key mesodermal is divided into two domains that correspond to the action of the patterning factor and is necessary for the correct development of pair rule genes, even skipped (eve) and sloppy paired (slp) various cell types deriving from the non-dorsal mesoderm, (Riechmann et al., 1997). The eve domain includes the cells including fat body, SGP, and somatic muscles. Correspondingly, underlying ectodermal stripes of hedgehog (hh) and engrailed at stages 10/11, D-Six4 is expressed in non-dorsal mesoderm in (en) expression, and these genes participate in the development a complementary pattern to tin. Moreover, with its cofactor of the tissues that derive from this region (Azpiazu et al., 1996; Eyes absent (Eya), D-Six4 is sufficient to drive the specification Riechmann et al., 1998). The action of hh is antagonized by that of certain non-dorsal fates. In addition, our results clarify the of wingless (wg), which signals to cells of the slp domain function of tin in ventral mesodermal cells: we propose that leading to body wall muscle and heart development (Azpiazu et earlier in development (at stages 8/9), part of tin's function al., 1996; Baylies et al., 1995; Park et al., 1996). In the slp ventrally is to initiate expression of D-Six4. domain, twist (twi) is expressed at a high level and contributes to the development of the somatic muscles (Baylies and Bate, 1996), while Notch signaling modulates twi to low levels in the Materials and methods eve domain (Tapanes-Castillo and Baylies, 2004). Histochemistry In the dorsoventral axis, the homeodomain transcription factor, Tinman (Tin), plays a central role in establishing dorsal Antibody staining of whole-mount embryos was performed using standard mesoderm fates (Azpiazu and Frasch, 1993; Bodmer, 1993; methods and detected using secondary antibodies conjugated to Alexa 488 or Lockwood and Bodmer, 2002; Yin and Frasch, 1998). In the 568 fluorochromes (Molecular Probes). Primary antibodies used were Ladybird β dorsal region, ectodermal Decapentaplegic (Dpp) signaling (mouse, 1/1) (Jagla et al., 1997), Srp (rabbit, 1/500) (Hayes et al., 2001), - galactosidase (mouse, 1/1000, Promega), Eya (mouse 1/100, Developmental maintains the expression of tin, which is lost from the Studies Hybridoma Bank, developed by N. Bonini), myosin heavy chain (rabbit, remainder of the mesoderm following gastrulation (Frasch, provided by B. Patterson), GFP (mouse, 1/1000, Molecular Probes), D-Mef2 1995; Staehling-Hampton et al., 1994). Tin and Dpp combine (rabbit, 1/1000, provided by K. Jagla), Zfh-1 (1/500, mouse, provided by Z. with factors involved in anteroposterior patterning to establish Lai), Tin (rabbit, 1/500, provided by K. Jagla), cleaved human caspase-3 (Asp the primordia of the various dorsal mesodermal organs. For 715, 1/100, Cell signaling Technology), FasIII (mouse, 1/100, from DSHB), and Vasa (1/1000, provided by R. Lehmann). For fluorescence in situ hybridization a example, in the dorsal slp domain, Tin cooperates with Wg to tyramide amplification kit (Molecular Probes) was used with Alexa 488. activate specific sets of target genes, leading to heart and Digoxygenin-labeled D-six4 RNA probe consisting of nucleotides 472 to 1616 dorsal muscle development (Furlong, 2004). Conversely, the of the cDNA sequence (AF099185) was transcribed using Roche reagents. For 222 I.B.N. Clark et al. / Developmental Biology 294 (2006) 220–231 double labeling, RNA in situ hybridization was performed first followed by strain by standard methods. The tin zfh-1 double mutant was a gift of R. Lehmann immunofluorescence. (Broihier et al., 1998). A P[UAS-D-Six4] construct was made by PCR amplification and cloning of the D-Six4 ORF into pUAST. Transformant flies Microscopy were made by microinjection by standard methods. twi-Gal4 and P[UAS-Eya] stocks were obtained from the Bloomington stock centre; UAS-dpp 94B/Cy;UAS- Fluorescently labeled embryos were visualized by laser scanning confocal dpp 42B was obtained from Helen Skaer. microscopy on either an SP (Leica) or a Pascal (Zeiss) microscope system. For analysis of recessive mutant phenotypes (D-Six4289 and zfh-1 tin), homozygous Results embryos were recognized by the absence of β-galactosidase staining in the progeny of parents carrying a P[Ubx-lacZ] marked balancer. Images were D-six4 expression pattern suggests it is a candidate factor for processed and arranged using Adobe Photoshop software. Images are depicted in colorblind friendly format (green and magenta). patterning the lateral and ventral mesoderm

Fly stocks We asked whether the expression of D-six4 might indicate a specific requirement in mesoderm patterning. D-six4 mRNA is Wild-type flies were of the Oregon R stock. The D-Six4289 allele is described expressed in the trunk mesoderm from stage 9; there is no by Kirby et al. (2001). Other mutants were Df(3R)GC14 for tin and zfh-12 expression in the cephalic mesoderm (Kirby et al., 2001; Seo et (provided by Z.-C. Lai). For the D-six4-III-GFP enhancer construct, the primers al., 1999). Expression is initially in a broad domain that TCTAGACAGCAAAGACCGTGAGTTG and GGATCCGAATGGATTGC- CATCCAGTTG were used to amplify the D-Six4 third intron sequence from coincides with D-Mef2, a protein that is expressed throughout wild-type genomic DNA, and the fragment was inserted into the pH Stinger the mesoderm at this stage (Fig. 2A). By stage 10, D-six4 vector (Barolo et al., 2000). The resulting plasmid was used to transform the w1118 mRNA becomes restricted to ventral and lateral mesoderm

Fig. 2. Pan-mesodermal D-six4 expression becomes refined to the lateral and ventral mesoderm. (A, B) D-six4 mRNA expression pattern in wild-type embryos, compared to the pan-mesodermal marker, D-Mef2. (A) Stage 9, showing widespread mesodermal D-six4 expression. (B), Stage 10, showing loss of D-six4 RNA from the dorsal mesoderm. (C–F) Embryos carrying a GFP reporter transgene driven by the D-six4 third intron enhancer (D-six4-III-GFP). (C) Stage 10. Relative to D-Mef2, stronger ventral/lateral GFP represents the restriction of D-six4 expression. Remaining dorsal expression represents GFP perdurance. (C) (D) At stage 10, D-six4-III-GFP does not overlap with expression of dorsal Tin (magenta). (E) Stage 10 embryo at higher magnification stained for GFP and Srp. GFP is coexpressed with Srp in the dorsolateral eve domain but is expressed to a higher level in the dorsolateral and ventral slp domain (areas between the Srp clusters). (F) twi-Gal4, UAS-dpp embryo. Misexpression of dpp results in reduction of mesodermal D-six4 expression (but no effect on ectodermal head expression). Panels C and E are single confocal sections to minimize artifactual colocalization. In all micrographs anterior is to the left, dorsal at the top. I.B.N. Clark et al. / Developmental Biology 294 (2006) 220–231 223

(Fig. 2B), before being rapidly lost from all mesodermal cells overlapping spatial domains of gene expression in the except the SGPs. mesoderm: a dorsal domain expressing tin and a ventral and Using a GFP reporter gene construct (referred to as D-six4- lateral domain in which D-six4 is expressed. D-six4 is therefore III-GFP), we identified an enhancer within the D-six4 third a candidate for the counterpart of tin in patterning more ventral intron that activates GFP in a pattern corresponding closely to mesodermal fates. the mesodermal expression of D-six4 RNA. At stage 9, D-six4- III-GFP is coexpressed with D-Mef2 in a broad mesodermal D-six4 is required for fat body development domain. Subsequently, by stage 10, GFP expression becomes largely restricted ventrally, although some perduring protein Since D-six4 has functions in both gonad and muscle remains in the dorsal region (Fig. 2C). At this point, lateral/ formation (Kirby et al., 2001), we investigated whether it also ventral D-six4-III-GFP expression is complementary to the has a role in development of the fat body, the other major organ dorsal expression of Tin (Fig. 2D) (although the two are arising from the non-dorsal mesoderm. Fat body precursors are coexpressed earlier). Once restricted, the dorsal limit of D-six4- marked in the embryo by expression of Srp (a GATA tran- III-GFP expression coincides with that of serpent (srp) protein, scription factor), which is required for their specification and which marks the dorsolateral fat body cells (Abel et al., 1993). differentiation (Abel et al., 1993; Rehorn et al., 1996; Sam et al., The levels of D-six4 mRNA and D-Six4-III-GFP expression are 1996). At stage 14 in wild-type embryos, Srp-staining cells form modulated in the anteroposterior axis of the segment, being a continuous layer in the lateral mesoderm (Fig. 3A). In embryos stronger in the slp domain (between the dorsolateral Srp clusters homozygous for the D-six4289 allele, which is probably a of the eve domain) (Fig. 2E). This anteroposterior modulation of functional null (Kirby et al., 2001), there are very few Srp- D-Six4 expression resembles that of twi, raising the possibility positive cells in the trunk at this stage (Fig. 3B), demonstrating that different levels of protein have different functional that D-six4 is essential for fat body development. From their consequences (Baylies and Bate, 1996; Tapanes-Castillo and distribution, it is likely that the remaining Srp-positive cells are Baylies, 2004). hemocytes, which derive from the cephalic mesoderm. At stage 10, inductive dpp signaling from the dorsal The layer of fat body precursors visible at stage 14 is derived ectoderm acts to maintain tin expression, thereby driving the from cells originating in three distinct clusters in each of the dorsal restriction of Tin (Staehling-Hampton et al., 1994). embryonic parasegments 4–9(Fig. 1). At stage 10, a primary Conversely, the ventral restriction of D-six4 may depend on an cluster appears dorsolaterally in the eve domain. Later, two inhibitory effect of dpp signaling. Consistent with this, secondary clusters appear, one immediately posterior to the misexpression of dpp throughout the mesoderm reduces original cluster and the other in a more ventral and anterior expression of D-six4 RNA to a low level (Figs. 2F and G). position (Figs. 3C and E) (Moore et al., 1998; Riechmann et al., Thus, we suggest that dpp signaling acts to establish two, non- 1998). The secondary clusters are present in parasegments

Fig. 3. D-six4 is required for fat body development. Expression of Srp in wild-type (A, C, E) and homozygous D-Six4289 (B, D, F) embryos. (A, B) Stage 14, showing the loss of most Srp expression in the mutant. (C, D) Stage 13, showing the loss of dorsolateral Srp clusters (brackets) but the initial establishment of ventral secondary clusters (arrows). (E, F) Stage 12, showing that the loss of dorsolateral Srp clusters occurs early. The ventral secondary clusters are present but not completely represented in the optical cross sections in panels C–F. 224 I.B.N. Clark et al. / Developmental Biology 294 (2006) 220–231

10–12 as well as in the more anterior regions. In the absence positive cells at the earliest stage at which they can be of D-six4 function, few Srp-staining cells are visible in recognized, implying that there is indeed a partial loss of initial dorsolateral positions at stage 12, suggesting a defect in initial SGP specification (Fig. 4B), although we cannot rule out a very specification of the primary fat body precursors (Fig. 3D and early onset of cell death in the mutant. F). On the other hand, the secondary fat body precursors are Many of the SGPs that initially appear in D-six4289 mutants initially present ventrally (Figs. 3D and F), although Srp appear to die subsequently. The number of cells expressing Eya expression in these precursors is not maintained by stage 14. decreases markedly during stage 12, while a greater number of These findings are consistent with previous observations apoptotic cells is observed in the region normally occupied by indicating that the ventral and dorsolateral clusters develop by SGPs (Figs. 4C and D). Some of these apoptotic cells also distinct mechanisms (Riechmann et al., 1998). Although we express Eya, implying that programmed cell death accounts for cannot rule out the possibility that primary precursors are the reduction in SGP number at this time. Thus, like its function specified but fail to express Srp, it appears likely that D-six4 in fat body development, D-six4 appears to be required in SGPs has functions in initial specification and subsequent mainte- both for initial specification and for survival or maintenance of nance of different populations of fat body precursors. fate. In summary, D-six4 is required for two cell types (fat body Defect in SGP specification in D-Six4 mutants and SGPs) arising from the dorsolateral mesoderm of the eve domain. In contrast, it is not required for the visceral muscles, In parasegments 10–12, the dorsolateral mesoderm of the which arise from the adjacent dorsal region of the eve domain eve domain gives rise not to fat body cells but to the SGPs (Fig. (Fig. 1; data not shown). 1). These attract and coalesce with migrating germ cells to form the primitive gonads in wild-type embryos (reviewed in (Santos A specific subset of muscles is disrupted in D-Six4 mutants and Lehmann, 2004)). In D-Six4 mutants, the germ cells fail to coalesce with SGPs, becoming scattered throughout the Given the specific requirement for D-six4 in the eve domain posterior half of the embryo (Kirby et al., 2001). It is not of the mesoderm, we then examined the role of D-six4 in clear from previous studies whether this phenotype results from patterning cell fates for tissues that arise from the slp domain. a failure in SGP specification or from changes in their These include cells of the heart dorsally and somatic muscles, functional properties. We examined the initial appearance of which arise from cells throughout the entire dorsoventral extent SGPs using expression of Eyes absent (Eya) protein as a marker. of this domain (Borkowski et al., 1995). D-six4 mutant embryos In wild-type embryos, Eya expression becomes refined to three exhibit a characteristic pattern of muscle defects. The bilateral clusters of SGPs at stage 11 (Fig. 4A) (Boyle et al., cardioblasts appear normal, as do the dorsal somatic muscles 1997). In D-six4289 mutants, these clusters consist of fewer Eya- (Figs. 5A and B). Both these tissues originate in the dorsal

Fig. 4. D-six4 is required for the specification and maintenance of SGPs. (A, B) Stage 11 embryos stained for Eya protein. (A) Wild type, with SGPs seen in three bilateral clusters in parasegments 10–12 (arrows). (B) D-six4289 mutant, showing clusters containing fewer SGPs. The cluster size appears larger than expected from the reduced cell number because Eya expression is not exclusively nuclear in the mutant. (C, D) Stage 12 embryos stained for Eya and cleaved caspase 3 to mark apoptotic cells. (C) Wild type. (D) D-six4289 mutant, showing fewer Eya expressing cells and more apoptotic cells. Some Eya-staining SGPs are apoptotic (arrow). I.B.N. Clark et al. / Developmental Biology 294 (2006) 220–231 225

Fig. 5. D-six4 is required for the development of specific ventral and lateral muscles. Wild-type (A, C, E) and D-six4289 homozygous (B, D, F) late stage embryos stained for myosin heavy chain to show the body wall muscles. (A, B) Dorsal views, showing the cardioblasts (CB) and dorsal somatic muscles (DM) appear normal in D-six4289 mutants. (C, D) Lateral views, showing severe disruption in the D-six4289 mutant of muscles including the lateral transverse muscles (LT) and the segment border muscle (SBM). The unaffected dorsal muscles (DM) are visible. (E, F) Ventrolateral views, showing disruption in the mutant of specific ventral muscles, including ventral acute 3 (VA3); other muscles, including the ventral oblique muscles VO1-3 (VO) are affected to a lesser degree. (G, H) Stage 11 embryos stained for Ladybird protein. (G) Wild type, with arrows indicating the Ladybird expression associated with the SBM founder cell; other expression is associated with the nervous system. (H) D-Six4289 mutant, with most segments showing loss of Ladybird expression in the area of the SBM founder cell, while neural expression is unaffected. (I) Ventrolateral view of tin zfh-1 double mutant embryo stained for myosin heavy chain, showing severe disorganization of muscle pattern. 226 I.B.N. Clark et al. / Developmental Biology 294 (2006) 220–231 region of persistent Tin expression. In contrast, lateral and sufficient for the tin-independent somatic muscles that require ventral somatic muscles are disrupted (Figs. 5C–F). The lateral D-six4. body wall muscles are severely affected, with many muscles The tin fat body phenotype is much more severe in missing (Figs. 5C and D). Although it is difficult to identify all embryos that also lack the zinc finger/homeodomain protein the remaining muscles in embryos stained for myosin, a number encoded by zfh-1 (Broihier et al., 1998; Moore et al., 1998), of consistent effects can be discerned. In the ventral region, suggesting that these two transcription factors have somewhat certain muscles are consistently affected to a greater extent than redundant functions in fat body patterning. Strikingly, others (Figs. 5E and F). For example, the ventral acute muscle, although D-six4-III-GFP expression is affected only slightly VA3, and segment border muscle (SBM) are absent in most in zfh-1 mutant embryos (Fig. 6C), expression is abolished in segments, while some ventral oblique (VO) muscles and most double tin zfh-1 mutant embryos (with the exception of a of the ventral longitudinal (VL) muscles are usually present small number of cells in posterior parasegments) (Fig. 6D). (muscle nomenclature according to Bate (1993). D-six4 is We suggest therefore that the roles of tin and zfh-1 in fat body therefore required for the development of specific muscles that and SGP development are at least partly mediated by their arise from the dorsolateral and ventral regions. This requirement regulation of D-six4. Moreover, the extensive loss of D-six4- suggests that D-six4 regulates specific muscle identity genes in III-GFP expression suggests that together tin and zfh-1 must founder myoblasts. These genes encode transcription factors also influence the development of the D-six4-dependent that govern the eventual size, shape, position, and innervation of somatic muscles. Consistent with this possibility, muscle the individual muscles (Bate, 1993). Consistent with this patterning in tin zfh-1 double mutant embryos is much more proposal, the expression of ladybird, which is a muscle identity disrupted than reported for the single mutants, with little gene for the SBM founder myoblast, is absent in D-Six4 mutant apparent organization of muscles in the ventral and lateral embryos (Figs. 5G and H). Conversely, the expression of dorsal regions (Fig. 5I). muscle identity genes, collier and eve, is unaffected (data not shown). D-Six4 and eyes absent cooperate to specify mesodermal organs D-Six4 may mediate early functions of tinman and zfh-1 in fat body and SGP patterning D-Six4 expression may be passively required during the patterning of ventral and dorsolateral mesodermal fates (e.g., to In at least some aspects of somatic muscle patterning, the maintain fates), or it may actively confer information needed for roles of tin and D-six4 appear complementary. However, such patterning. To address this, we examined the effect of embryos lacking tin function also show loss of SGPs and misexpression on mesodermal patterning. Using the binary mild defects in fat body development (Boyle et al., 1997; Moore Gal4/UAS system, misexpression of UAS-D-Six4 alone et al., 1998), suggesting that the transient pan-mesodermal throughout the mesoderm using a twi-Gal4 driver failed to expression of tin has a role to play in these cells as well as in affect cell fate significantly (data not shown), which is some ventral and lateral somatic muscles. We investigated the consistent with the first possibility. However, an alternative possibility that regulation of D-six4 may account for these tin explanation is that an additional tissue-restricted factor(s) is phenotypes. Strikingly, in tin mutant embryos, there is a marked required for D-Six4 function. Eya is a likely candidate, since it reduction in D-Six4-III-GFP expression in most parasegments is a cofactor of the related Six protein, Sine oculis, in (Figs. 6A and B). This loss of expression may therefore at least Drosophila eye development (Pignoni et al., 1997), and partly underlie the SGP and fat body defects in tin mutants, homologous proteins synergize with other Six family members although clearly the remaining D-six4 expression must be to activate transcription in other organisms (Heanue et al., 1999;

Fig. 6. Effect of tin and zfh-1 on D-six4 expression. (A–D) GFP in D-six4-III-GFP embryos. (A) Wild type. (B) tin mutant, showing marked reduction of GFP expression. (C) zfh-1 mutant, showing a lesser reduction in GFP expression. (D) tin zfh-1 double mutant, showing loss of almost all the mesodermal D-six4 pattern. I.B.N. Clark et al. / Developmental Biology 294 (2006) 220–231 227

Laclef et al., 2003; Li et al., 2003; Ozaki et al., 2004; Xu et al., required for Eya's nuclear translocation, as has been reported 2003; Xu et al., 2002). Consistent with this, eya is coexpressed for their vertebrate homologues in cell culture experiments with D-Six4 RNA in the mesoderm (unpublished and Kirby et (Ohto et al., 1999). al., 2001). Like D-six4, eya expression becomes restricted With these observations in mind, we asked whether first to the dorsolateral and ventral mesoderm and later to the restricted eya expression limits the ability of D-Six4 to SGPs (Boyle et al., 1997, and data not shown). Unlike D- perturb mesoderm development when misexpressed. Mis- six4, however, eya expression is not modulated in the expression of eya alone has little or no effect on mesodermal anteroposterior axis. Mesodermal phenotypes of eya mutants organs (data not shown and Boyle et al. (1997)). In contrast closely resemble those of D-Six4 for fat body, SGP, and comisexpression of eya with D-six4 in the twi domain (twi- muscle development (unpublished observations and Boyle et Gal4 UAS-eya UAS-D-six4) strongly affects the mesoderm. al., 1997; Moore et al., 1998). In gonad development, it was Several cell types derived from the ventral and dorsolateral proposed that eya has a role in the maintenance of SGP cell mesoderm appear expanded. Greatly increased numbers of fate (Boyle et al., 1997; Warrior, 1994), but like D-six4,it SGPs result in an enlarged gonad (Figs. 7A–C). There are may also be required for initial specification since there also more fat body precursors, and these now extend into the appear to be fewer SGPs even at the time they are first dorsal region of the mesoderm, suggesting that dorsal cells specified (Fig. 5A in Boyle et al., 1997). In D-six4 mutant have taken on more ventral fates (Figs. 7A and B). In embryos, the exclusive nuclear localization of Eya protein is contrast, derivatives of the dorsal mesoderm are disrupted. In lost (Fig. 4B and data not shown). This is consistent with a stage 17 embryos, no myosin-expressing cardioblasts are direct protein interaction between D-Six4 and Eya being visible, while there are gaps in the pattern of dorsal muscles

Fig. 7. Misexpression of D-Six4 and eya specifically disrupts mesoderm patterning. (A, B) Lateral views of stage 13 embryos stained for Srp (green) to mark the fat body (FB) and Zfh-1 (magenta) to mark the SGPs/gonad. (A) twi-Gal4 control, showing the location of fat body and gonad relative to the dorsal mesoderm (DoMe). (B) twi-Gal4 UAS-D-six4/UAS-eya embryo, showing enlargement of gonad and dorsal expansion of fat body when D-six4 and eya are misexpressed. (C) twi-Gal4 UAS-D-six4/UAS-eya embryo showing coalescence of large numbers of SGPs (marked in magenta by zfh-1) around the germ cells (marked in green by Vasa). Note that correct migration and coalescence of germ cells seems to occur despite the enlarged gonad. (D–H) Stage 17 embryos stained for myosin heavy chain, showing the effect of misexpression on somatic muscles and cardioblasts. (D) twi-Gal4 UAS-D-Six4/UAS-Eya embryo, dorsal view. The cardioblasts are absent (cf., Fig. 5A) and several dorsal somatic muscles (DM) are missing. (E) twi-Gal4 UAS-D-Six4/UAS-Eya embryo, lateral view. Lateral body wall muscles are expanded relative to twi- Gal4 control embryo (F). The ventral acute 3 muscle (VA3) appears to be duplicated in some segments. (G, H) Higher magnification views of embryos similar to those shown in E and F, in which myosin stained muscles are visualized after volume rendering (Amira software). Extra muscle tissue in the lateral (l) region is clearly seen after misexpression. (I, J) Stage 12 embryos stained for Fas III, showing a disruption of expression in the visceral mesoderm upon misexpression (I) compared with the control embryo (J). 228 I.B.N. Clark et al. / Developmental Biology 294 (2006) 220–231

(Fig. 7D, cf., Fig. 5A). Similarly, there are gaps in the Defective D-six4 function results partly in failure of cell fate visceral mesoderm at stage 12, as revealed by fasciclin III maintenance and/or cell survival. This is a common mutant expression (Figs. 7I and J). phenotype of members of the Six and Eya gene families. The effects of misexpression on the lateral and ventral Strikingly, however, expression of D-Six4 with its cofactor, somatic musculature are more difficult to analyse. Consistently, Eya, throughout the mesoderm causes the expansion of D-six4- however, in the ventral region, there are duplications of the VA3 dependent cell types (fat body and SGPs) with the concomitant muscle, while the pattern of VO muscles is disrupted (Figs. 7D– disruption of other mesodermal derivatives, including the F). In general, this suggests that at least some of the muscles that cardioblasts and visceral mesoderm. This supports an active are disrupted in the D-six4 mutant appear duplicated upon role for D-six4 in initial patterning of cell fates. It is possible, misexpression, whereas muscles that are little affected in the therefore, that maintenance/survival phenotypes are a secondary mutant (dorsal and some ventral) are disrupted by misexpres- effect of defects in the initial establishment of cell identity. sion. The lateral somatic musculature is harder to analyze but Specific aspects of muscle cell identity are affected in D-six4 appears expanded upon misexpression (Figs. 7G and H). These mutant embryos. The phenotype is variable, but the external results support the idea of distinct developmental pathways for lateral and some ventral muscles are consistently disrupted. muscle subsets, one of which requires D-six4. Significantly, When D-Six4 and Eya are misexpressed/overexpressed in the other pathways (some requiring Tin) are susceptible to D-six4/ mesoderm, an aberrant but regular muscle pattern is formed, eya function. suggesting that they have a patterning role, as opposed to a function in differentiation or myoblast fusion as we have Discussion previously suggested (Kirby et al., 2001). It is likely that D-six4 participates in the activation of certain muscle identity genes in We have established that D-six4 is a key factor for the founder myoblasts (Bate, 1993). Expression of the SBM development of a variety of tissues that originate from the non identity gene, ladybird, requires D-Six4, while misexpression dorsal mesoderm (Fig. 8). It is required for the SGPs, fat body of D-six4 and eya specifically in founder cells (using a precursors and specific lateral and ventral muscles and is likely dumbfounded-Gal4 driver) results in a muscle phenotype to be a competence factor or patterning mediator, acting to indistinguishable from that of embryos misexpressing these regulate a variety of key tissue and cell identity genes, such as genes throughout the mesoderm (data not shown). srp for the fat body and ladybird for the segment border muscle The relationship between D-six4 and tin is complex, partly founder cells. Different target genes would be regulated in because it changes over time (summarized in Fig. 8) and also different locations by the combinatorial action of D-six4 and because tin has functions in the ventral and lateral mesoderm other factors involved in dorsoventral and anteroposterior axis that have remained obscure. The best characterized functions of patterning. D-six4 may play additional roles later in gonad tin concern the dorsal mesoderm (Lee and Frasch, 2005), development, since its expression is maintained in SGPs reflected in its restricted dorsal expression at stage 10/11. At this throughout embryogenesis, whereas it is expressed transiently time, D-six4 expression is complementary to that of tin, and in most of the mesoderm. there are no discernable effects on dorsal mesoderm structures

Fig. 8. Schematic summary. (A) Schematic summary of muscles affected in mutant embryos of tin (data taken from Azpiazu and Frasch, 1993) and D-six4 (this work). The muscles assigned represent a conservative estimate of those that are clearly affected in one or other mutant. It is possible that the remaining muscles (grey–blue) also require either tin or D-six4, but the degree of muscle disorganization precludes definitive identification. (B) Summary of D-six4′s place in mesoderm patterning. ‘?’ Represents a possible third patterning factor, reflecting the uncertainty as to whether Tin's function in non-dorsal founder cell specification is direct or indirect. The nature of the redundancy between tin and zfh-1 is unclear. See text for details. I.B.N. Clark et al. / Developmental Biology 294 (2006) 220–231 229 in D-six4 mutants. We propose that these two genes play The role of D-six4 in mesoderm patterning appears to be complementary roles in their respective domains, promoting the conserved in other organisms. Expression of human SIX5 is development of specific cell types in conjunction with reduced in Type I Myotonic Dystrophy (Fillipova et al., 2001; additional patterning factors. Despite their complementary Klesert et al., 1997; Thornton et al., 1997), which may suggest a expression patterns at this stage, there is no evidence that tin role in myogenesis since the most severe forms of this condition and D-six4 are mutually antagonistic: although Tin can act as a display muscle developmental defects (Harper, 1989). The repressor as well as an activator (Choi et al., 1999), there is no murine orthologues, Six4 and Six5, are both expressed during significant expansion of D-six4 expression along the dorsoven- myogenesis (Oliver et al., 1995; Ozaki et al., 2001), although tral axis in a tin mutant and vice versa (unpublished data). It is their precise roles are not yet established as single gene knock- more likely that, like tin, D-six4 is directly regulated by dpp out models have no clear muscle defects, perhaps owing to signaling. compensatory interactions (Klesert et al., 2000; Ozaki et al., In addition to dorsal mesoderm defects, tin mutant embryos 2001; Sarkar et al., 2000). Six4 mutation, however, strongly show SGP, fat body and specific lateral and ventral muscle exacerbates the muscle loss of mice mutant for the more defects that presumably depend on its early pan-mesodermal divergent homologue, Six1 (Grifone et al., 2005). It is striking in expression (Azpiazu and Frasch, 1993; Boyle et al., 1997; particular that hypaxial progenitors (which contribute to limb Moore et al., 1998). At least one tin function appears to depend muscles) lose their identity in Six1 Six4 double mutant mice on its regulation of D-six4 in the early mesoderm before their (Grifone et al., 2005). These muscle progenitors require the mutually exclusive refinement of expression. Like D-six4, tin is function of an lb homologue, Lbx1 (Alvares et al., 2003), and required for correct SGP development: a reduced number of there is evidence that Lbx1 may be a target of Six/Six4 (Grifone SGPs appear at stage 11, and the number diminishes further et al., 2005). Thus, it appears that the function of Six4/5 genes until stage 13 when germ cell migration defects become might be conserved to a high degree. apparent (Broihier et al., 1998). However, the ventral expression The requirement for D-six4 in diverse cell types, linked by of tin is lost before the SGPs are apparent, suggesting that their location of origin during mesoderm patterning, may another factor mediates its function in SGP development. D- represent a primordial state. The C. elegans homologue (unc- six4 may be this factor, since initially the two genes are 39) is also required for a number of mesodermal cell types transiently coexpressed broadly in the mesoderm, and D-six4 (Yankowitz et al., 2004). Although knowledge of Six4 and Six5 expression is partly dependent on tin function. At this stage, Tin function is incomplete, it is notable that Six1 is required for the could be a direct transcriptional activator of D-six4, since there development of diverse organs such as muscle, kidney, and otic are a number of sequences in the third intron matching the vesicle (Laclef et al., 2003; Ozaki et al., 2004; Xu et al., 2003). canonical Tin binding site (ACAAGTGG). It is interesting to note that Lbx1 regulation may be achieved by The pattern of lateral and ventral muscle defects in embryos the combinatorial action of Six1/4 and Hox genes (Alvares et al., lacking Tin is different from that of D-six4 mutants (Azpiazu 2003), which would thus behave as patterning factors in a and Frasch, 1993, Fig. 8A). Muscles affected by tin include similar way to D-Six4. Our studies suggest that diverse roles of LL1, LO1, VL3, VL4, and VT1 (Azpiazu and Frasch, 1993), SIX genes in vertebrate organogenesis as apparent cell- or which do not require D-Six4 or Eya (this study and Boyle et al., tissue-type regulators may have their evolutionary origins in a 1997). Conversely, muscles that are severely affected by D-six4 general primordial developmental patterning mechanism, part mutation appear normal in tin mutants, including VA3, the of which may be preserved more clearly in the role of D-six4 in SBM, and the external lateral muscles LT1, LT2, LT3, and LT4 mesoderm development in Drosophila. (Figs. 7B and C in Azpiazu and Frasch, 1993). Based on these findings, we propose that muscles fall into at least three Acknowledgments categories. The visceral, cardiac, and dorsal somatic muscles all require tin function directly through persistent dorsal tin We thank R. Lehmann for reagents and useful discussion. expression (Furlong, 2004). A second group, comprising a Also H. Skaer, Z.-C. Lai, and K. Jagla for reagents and the latter subset of ventral and lateral muscles, requires tin function via its also for sharing his knowledge of muscle development. This transient pan-mesodermal expression, either directly or perhaps work was supported by a grant from the Muscular Dystrophy through unknown patterning mediators. A third group, a Campaign. A.P.J. is a Senior Fellow of the Wellcome Trust different subset of lateral and ventral muscles, is dependent (042182). on D-six4/eya function and is not affected in tin single mutants, presumably because the reduced D-six4 expression in these References embryos (Fig. 6B) is sufficient for their patterning. Muscles in this last category resemble the fat body precursors in their Abel, T., Michelson, A.M., Maniatis, T., 1993. A Drosophila GATA family functional requirements, being dependent on an early, partially member that binds to Adh regulatory sequences is expressed in the redundant function of tin and zfh-1, which is necessary to developing fat body. Development 119, 623–633. initiate D-six4 expression in most parasegments. Confirmation Alvares, L.E., Schubert, F.R., Thorpe, C., Mootoosamy, R.C., Cheng, L., Parkyn, G., Lumsden, A., Dietrich, S., 2003. Intrinsic, Hox-dependent of this model awaits a comprehensive characterization of cues determine the fate of skeletal muscle progenitors. Dev. Cell 5, muscle identity gene expression in founder cells in tin and D- 379–390. six4 mutant embryos. Azpiazu, N., Frasch, M., 1993. tinman and bagpipe: two homeo box genes that 230 I.B.N. Clark et al. / Developmental Biology 294 (2006) 220–231

determine cell fates in the dorsal mesoderm of Drosophila. Genes Dev. 7, Laclef, C., Hamard, G., Demignon, J., Souil, E., Houbron, C., Maire, P., 1325–1340. 2003. Altered myogenesis in Six1-deficient mice. Development 130, Azpiazu, N., Lawrence, P.A., Vincent, J.P., Frasch, M., 1996. Segmentation and 2239–2252. specification of the Drosophila mesoderm. Genes Dev. 10, 3183–3194. Lee, H.H., Frasch, M., 2005. Nuclear integration of positive Dpp signals, Barolo, S., Carver, L.A., Posakony, J.W., 2000. GFP and beta-galactosidase antagonistic Wg inputs and mesodermal competence factors during transformation vectors for promoter/enhancer analysis in Drosophila. Drosophila visceral mesoderm induction. Development 132, BioTechniques 29, 726–732. 1429–1442. Bate, M., 1993. The mesoderm and its derivatives. In: Bate, M., Martinez Arias, Li, X., Oghi, K.A., Zhang, J., Krones, A., Bush, K.T., Glass, C.K., Nigam, S.K., A. (Eds.), The development of Drosophila melanogaster, vol. 2. Cold Spring Aggarwal, A.K., Maas, R., Rose, D.W., Rosenfeld, M.G., 2003. Eya protein Harbour Laboratory Press, pp. 1013–1090. phosphatase activity regulates Six1-Dach-Eya transcriptional effects in Baylies, M.K., Bate, M., 1996. twist: a myogenic switch in Drosophila. Science mammalian organogenesis. Nature 426, 247–254. 272, 1481–1484. Lockwood, W.K., Bodmer, R., 2002. The patterns of wingless, decapentaplegic, Baylies, M.K., Martinez Arias, A., Bate, M., 1995. Wingless is required for the and tinman position the Drosophila heart. Mech. Dev. 114, 13–26. formation of a subset of muscle founder cells during Drosophila Moore, L.A., Broihier, H.T., Van Doren, M., Lehmann, R., 1998. Gonadal embryogenesis. Development 121, 3829–3837. mesoderm and fat body initially follow a common developmental path in Beer, J., Technau, G.M., Campos-Ortega, J.A., 1987. Lineage analysis of Drosophila. Development 125, 837–844. transplanted individual cells in embryos of Drosophila melanogaster:IV. Ohto, H., Kamada, S., Tago, K., Tominaga, S.I., Ozaki, H., Sato, S., Kawakami, Commitment and proliferative capabilities of mesodermal cells. Roux's K., 1999. Cooperation of six and eya in activation of their target genes Arch. Dev. Biol. 196, 222–230. through nuclear translocation of Eya. Mol. Cell. Biol. 19, 6815–6824. Bodmer, R., 1993. The gene tinman is required for specification of the heart and Oliver, G., Wehr, R., Jenkins, N.A., Copeland, N.G., Cheyette, B.N., visceral muscles in Drosophila. Development 118, 719–729. Hartenstein, V., Zipursky, S.L., Gruss, P., 1995. Homeobox genes and Borkowski, O.M., Brown, N.H., Bate, M., 1995. Anterior–posterior subdivision connective tissue patterning. Development 121, 693–705. and the diversification of the mesoderm in Drosophila. Development 121, Ozaki, H., Watanabe, Y., Takahashi, K., Kitamura, K., Tanaka, A., Urase, K., 4183–4193. Momoi, T., Sudo, K., Sakagami, J., Asano, M., Iwakura, Y., Kawakami, K., Boyle, M., Bonini, N., DiNardo, S., 1997. Expression and function of clift in the 2001. Six4, a putative myogenin gene regulator, is not essential for mouse development of somatic gonadal precursors within the Drosophila embryonal development. Mol. Cell. Biol. 21, 3343–3350. mesoderm. Development 124, 971–982. Ozaki, H., Nakamura, K., Funahashi, J., Ikeda, K., Yamada, G., Tokano, H., Broihier, H.T., Moore, L.A., Van Doren, M., Newman, S., Lehmann, R., 1998. Okamura, H.O., Kitamura, K., Muto, S., Kotaki, H., Sudo, K., Horai, R., zfh-1 is required for germ cell migration and gonadal mesoderm Iwakura, Y., Kawakami, K., 2004. Six1 controls patterning of the mouse otic development in Drosophila. Development 125, 655–666. vesicle. Development 131, 551–562. Choi, C.Y., Lee, Y.M., Kim, Y.H., Park, T., Jeon, B.H., Schulz, R.A., Kim, Y., Park, M., Wu, X., Golden, K., Axelrod, J.D., Bodmer, R., 1996. The wingless 1999. The homeodomain transcription factor NK-4 acts as either a signaling pathway is directly involved in Drosophila heart development. transcriptional activator or repressor and interacts with the p300 coactivator Dev. Biol. 177, 104–116. and the Groucho corepressor. J. Biol. Chem. 274, 31543–31552. Pignoni, F., Hu, B., Zavitz, K.H., Xiao, J., Garrity, P.A., Zipursky, S.L., 1997. Fillipova, G.N., Thienes, C.P., Penn, B.H., Cho, D.H., Hu, Y.J., Moore, J.M., The eye-specification proteins So and Eya form a complex and regulate Klesert, T.R., Lobanenkov, V.V., Tapscott, S.J., 2001. CTCF-binding sites multiple steps in Drosophila eye development. Cell 91, 881–891. flank CTG/CAG repeats and form a methylation-sensitive insulator at the Rehorn, K.P., Thelen, H., Michelson, A.M., Reuter, R., 1996. A molecular DM1 locus. Nat. Genet. 28, 335–343. aspect of hematopoiesis and endoderm development common to vertebrates Frasch, M., 1995. Induction of visceral and cardiac mesoderm by ectodermal and Drosophila. Development 122, 4023–4031. Dpp in the early Drosophila embryo. Nature 374, 464–467. Riechmann, V., Irion, U., Wilson, R., Grosskortenhaus, R., Leptin, M., 1997. Furlong, E.E., 2004. Integrating transcriptional and signalling networks during Control of cell fates and segmentation in the Drosophila mesoderm. muscle development. Curr. Opin. Genet. Dev. 14, 343–350. Development 124, 2915–2922. Grifone, R., Demignon, J., Houbron, C., Souil, E., Niro, C., Seller, M.J., Riechmann, V., Rehorn, K.P., Reuter, R., Leptin, M., 1998. The genetic control Hamard, G., Maire, P., 2005. Six1 and Six4 homeoproteins are required for of the distinction between fat body and gonadal mesoderm in Drosophila. Pax3 and Mrf expression during myogenesis in the mouse embryo. Development 125, 713–723. Development 132, 2235–2249. Sam, S., Leise, W., Hoshizaki, D.K., 1996. The serpent gene is necessary for Harper, P.S., 1989. Myotonic Dystrophy. W.B. Saunders Company. progression through the early stages of fat-body development. Mech. Dev. Hayes, S.A., Miller, J.M., Hoshizaki, D.K., 2001. serpent, a GATA-like 60, 197–205. transcription factor gene, induces fat-cell development in Drosophila Santos, A.C., Lehmann, R., 2004. Germ cell specification and migration in melanogaster. Development 128, 1193–1200. Drosophila and beyond. Curr. Biol. 14, R578–R589. Heanue, T.A., Reshef, R., Davis, R.J., Mardon, G., Oliver, G., Tomarev, S., Sarkar, P.S., Appukuttan, B., Han, J., Ito, Y., Ai, C., Tsai, W., Chai, Y., Stout, Lassar, A.B., Tabin, C.J., 1999. Synergistic regulation of vertebrate muscle J.T., Reddy, S., 2000. Heterozygous loss of Six5 in mice is sufficient to development by Dach2, Eya2, and Six1, homologs of genes required for cause ocular cataracts. Nat. Genet. 25, 110–114. Drosophila eye formation. Genes Dev. 13, 3231–3243. Seo, H.C., Curtiss, J., Mlodzik, M., Fjose, A., 1999. Six class homeobox genes Jagla, K., Jagla, T., Heitzler, P., Dretzen, G., Bellard, F., Bellard, M., 1997. in Drosophila belong to three distinct families and are involved in head ladybird, a tandem of homeobox genes that maintain late wingless development. Mech. Dev. 83, 127–139. expression in terminal and dorsal epidermis of the Drosophila embryo. Staehling-Hampton, K., Hoffmann, F.M., Baylies, M.K., Rushton, E., Bate, M., Development 124, 91–100. 1994. dpp induces mesodermal gene expression in Drosophila. Nature 372, Kirby, R.J., Hamilton, G.M., Finnegan, D.J., Johnson, K.J., Jarman, A.P., 2001. 783–786. Drosophila homolog of the myotonic dystrophy-associated gene, SIX5, is Tapanes-Castillo, A., Baylies, M.K., 2004. Notch signaling patterns Drosophila required for muscle and gonad development. Curr. Biol. 11, 1044–1049. mesodermal segments by regulating the bHLH transcription factor twist. Klesert, T.R., Otten, A.D., Bird, T.D., Tapscott, S.J., 1997. Trinucleotide repeat Development 131, 2359–2372. expansion at the myotonic dystrophy locus reduces expression of DMAHP. Thornton, C.A., Wymer, J.P., Simmons, Z., McClain, C., Moxley III, R.T., 1997. Nat. Genet. 16, 402–406. Expansion of the myotonic dystrophy CTG repeat reduces expression of the Klesert, T.R., Cho, D.H., Clark, J.I., Maylie, J., Adelman, J., Snider, L., flanking DMAHP gene. Nat. Genet. 16, 407–409. Yuen, E.C., Soriano, P., Tapscott, S.J., 2000. Mice deficient in Six5 Warrior, R., 1994. Primordial germ cell migration and the assembly of the develop cataracts: implications for myotonic dystrophy. Nat. Genet. 25, Drosophila embryonic gonad. Dev. Biol. 166, 180–194. 105–109. Xu, P.X., Zheng, W., Laclef, C., Maire, P., Maas, R.L., Peters, H., Xu, X., 2002. I.B.N. Clark et al. / Developmental Biology 294 (2006) 220–231 231

Eya1 is required for the morphogenesis of mammalian thymus, parathyroid E.A., 2004. UNC-39, the C. elegans homolog of the human myotonic and thyroid. Development 129, 3033–3044. dystrophy-associated homeodomain protein Six5, regulates cell motility and Xu, P.X., Zheng, W., Huang, L., Maire, P., Laclef, C., Silvius, D., 2003. Six1 is differentiation. Mech. Dev. 272, 389–402. required for the early organogenesis of mammalian kidney. Development Yin, Z., Frasch, M., 1998. Regulation and function of tinman during dorsal 130, 3085–3094. mesoderm induction and heart specification in Drosophila. Dev. Genet. 22, Yankowitz, J.L., Shakir, M.A., Hedgecock, E., Hutter, H., Fire, A.Z., Lundquist, 187–200.