The Opitz Syndrome Gene MID1 Is Essential for Establishing Asymmetric Gene Expression in Hensen's Node

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

provided by Elsevier - Publisher Connector Available online at www.sciencedirect.com

Developmental Biology 258 (2003) 397–405 www.elsevier.com/locate/ydbio

The Opitz syndrome gene MID1 is essential for establishing asymmetric gene expression in Hensen’s node

Alessandra Granata and Nandita A. Quaderi*

MRC Centre for Developmental Neurobiology, King’s College London, 4th Floor New Hunt’s House, Guy’s Hospital Campus, London, SE1 1UL, UK Received for publication 13 January 2003, revised 21 February 2003, accepted 25 February 2003

Abstract Patterning the avian left–right (L/R) body axis involves the establishment of asymmetric molecular signals on the left and right sides of Hensen’s node. We have examined the role of the chick Midline 1 gene, cMid1, in generating asymmetric gene expression in the node. cMid1 is initially expressed bilaterally, but its expression is then conﬁned to the right side of the node. We show that this restriction of cMid1 expression is a result of repression by Shh on the left side of the node. Misexpression of cMid1 on the left side of the node results in bilateral Bmp4 expression and a loss of Shh expression. Correspondingly, downstream left pathway genes are repressed while right pathway genes are ectopically activated. Conversely, knocking down endogenous right-sided cMid1 results in a loss of Bmp4 expression and bilateral Shh expression. This results in an absence of right pathway genes and the ectopic activation of the left pathway on the right. Here, we present a revised model for the establishment of asymmetric gene expression in Hensen’s node based on the epistatic interactions observed between Shh, cMid1, and Bmp4.

Keywords: L/R asymmetry; Hensen’s node; MID1; SHH; BMP4; Opitz syndrome; Chick; Antisense morpholinos

Introduction expression pattern. In this report, we show that cMid1 is expressed on the right side of Hensen’s node and that it We have previously shown that mutations in MID1 cause plays a key role during the early stages of left–right (L/R) X-linked Opitz G/BBB syndrome (Quaderi et al., 1997) determination. (OS; OMIM 300000 and 145410), a rare congenital defect The internal organs of vertebrates show invariant asym- affecting midline structures. Hypertelorism and hypospa- metries along the L/R axis, both in terms of placement and dias are the two principal features of OS, but tracheo- morphology. Incorrect patterning of the L/R axis results in oesophagal fistulas, cardiac malformations, and CNS de- laterality defects in which the internal organs are abnor- fects resulting in mental retardation are also common. mally positioned and formed (Capdevila et al., 2000). L/R Recently, progress has been made in understanding MID1 determination in chick involves the establishment of asym- cellular function: MID1 is a microtubule-associated ubiq- metric gene expression in Hensen’s node. Once molecular uitin ligase that targets the catalytic subunit of protein phos- asymmetry has been established, distinct signalling path- phatase 2A (PP2A) for degradation (Liu et al., 2001; Trock- ways on the left and right sides of the node transmit L/R enbacher et al., 2001). However, this knowledge does not information to the lateral plate mesoderm (LPM), and then significantly aid our understanding of MID1 biological on to the organ primordia, resulting in asymmetric morpho- function as PP2A has many targets and is involved in a genesis (Capdevila et al., 2000; Wright, 2001). A number of multitude of processes (Sontag, 2001). To investigate the genes have now been identified that are asymmetrically roles that MID1 might play during development, we have expressed in the chick node (Boettger et al., 1999; Garcia- cloned the chick MID1 orthologue (cMid1) and analysed its Castro et al., 2000; Kawakami and Nakanishi, 2001; Levin et al., 1995; Monsoro-Burq and Le Douarin, 2001; Rodri- * Corresponding author. Fax: ϩ44-(020)-7848-6550. guez-Esteban et al., 2001), and considerable progress has E-mail address: [email protected] (N.A. Quaderi). been made in characterising the asymmetric signalling cas-

Fig. 1. Expression of cMid1 in Hensen’s node. (A) Weak, bilateral cMid1 expression in the node is first seen at stage 4. Ectodermal staining is also seen in an anterior horse shoe-shaped domain. (B) Symmetrical expression of cMid1 in the ectoderm of the node is seen at stage 4ϩ. (C) cMid1 expression is confined to the right side of the node by stage 5. (D) cMid1 expression is stronger in the mesendoderm of the right side of the node at stage 7. Expression of cMid1 in the mesoderm lateral to the node is also seen. (E) By stage 8, cMid1 shows weak, symmetrical expression in the node. (A–D) Transverse sections through the node, at the level of the white line, are also shown. (F, G) Activin-soaked beads were implanted to the left of the node in stage 4 embryos. (G) cMid1 expression is upregulated on the left side of the node. (H) Control beads soaked in PBS do not affect cMid1 expression in the node. High-magnification inserts of the node show the small changes in gene expression. Red arrows denote the extent on gene expression on the manipulated side, and black arrows denote the extent on gene expression on the nonmanipulated side. All embryos are shown dorsal side upward. L, left; R, right.

cade that patterns the L/R axis (Capdevila et al., 2000; Bmp4 and propose a revised model for the establishment of Wright, 2001). asymmetric gene expression in Hensen’s node. The current model for the establishment of asymmetric gene expression in the node involves a mutually repressive interaction between Shh on the left and Bmp4 on the right Materials and methods side of the node (Monsoro-Burq and Le Douarin, 2001). Here, we present data showing that cMid1 acts upstream of Cloning cMid1 Bmp4 and plays a pivotal role in mediating the mutually antagonistic interactions between Shh and Bmp4. We de- The following degenerate primer pair designed from scribe the regulatory relationships between Shh, cMid1, and coding exon 9 of the human MID1 sequence (Gaudenz et

Table 1 Key L/R genes used in this study

Gene Asymmetrical expression in Hensen’s node and LPM Reference cAct-RIIa First detected in the right side of the node at stage 4 Levin et al., 1995 Shh Symmetrical expression detected in the node at stage 4Ϫ Levin et al., 1995 Expression is restricted to the left side of the node at stage 5Ϫ cMid1 Symmetrical expression detected in the node at stage 4 This report Expression is restricted to the right side of the node at stage 5Ϫ Bmp4 First detected in the right side of the node at stage 5Ϫ Monsoro-Burq and Le Douarin, 2001 Fgf8 First detected in the right side of the node at stage 6 Boettger et al., 1999 Nodal Expressed in the left perinodal region and LPM from stage 7 Levin et al., 1995 Snail Expressed in the right side LPM from stage 8ϩ Isaac et al., 1997 A. Granata, N.A. Quaderi / Developmental Biology 258 (2003) 397–405 399

Fig. 2. Regulatory interactions between cMid1 and the right pathway. (A, D, F, H) A cMid1 expression vector, which also expresses GFP, was electroporated into the left side of Hensen’s node in stage 4 embryos. (B) The typical extent of electroporation is shown by the region of GFP fluorescence 2 h after electroporation. (C) By stage 7, the region of fluorescence has become more diffuse as a result of cell movement. (D) Ectopic Bmp4 expression on the left side of the node. (F) Ectopic Fgf8 expression on the left side of the node. (H) Ectopic Snail expression in the left LPM. (E, G, I) Ectopic left-sided expression of Bmp4, Fgf8,orSnail is not seen in embryos electroporated with a control GFP vector. (J, M, O, Q) A fluorescein-tagged antisense morpholino directed against cMid1 was electroporated into the right side of Hensen’s node in stage 4 embryos. (K) The typical extent of electroporation is shown by the region of fluorescein fluorescence. (L) By stage 7, the region of fluorescence has become more diffuse as a result of cell movement. (M) Bmp4 expression is downregulated in the right side of the node. (O) Fgf8 expression is downregulated in the right side of the node. (Q) Snail expression is downregulated in the right LPM. (N,P,R) Downregulation of right-sided gene expression is not seen in embryos electroporated with a control morpholino. Inserts, arrows and orientation as in Fig. 1. al., 1998; Quaderi et al., 1997) was used to amplify chick by using standard techniques. Ten independent clones were genomic DNA: 5Ј-GCNCCNAARCAYGARTGG-3Ј and isolated, and a subset was sequenced and used to generate the 5Ј-ATNGTNARRCAYTTRTTCCA-3Ј. The expected 306-bp consensus cMid1 sequence. The cMid1 sequence has been product was used to screen a chick embryonic cDNA library deposited in GenBank under Accession No. AF374463.

Fig. 3. Regulation of cMid1 by BMP4 and FGF8. (A, B) BMP-soaked beads were implanted to the left of the node in stage 4 embryos. (B) cMid1 expression is upregulated on the left side of the node. (C) Control beads soaked in PBS do not affect cMid1 expression in the node. (D, E) FGF8-soaked beads were implanted to the left of the node in stage 4 embryos. Neither FGF8 (E) nor PBS-soaked control (F) beads affected cMid1 expression in the node. (G, H, J, L) Chordin-soaked beads were implanted to the right of the node in stage 4 embryos. Neither chordin- (H) nor PBS-soaked control beads (I) affected cMid1 expression in the node. (J) Fgf8 expression is downregulated in the right side of the node. (L) Shh expression is upregulated on the right side of the node. (K, M) Control beads soaked in PBS did not affect Fgf8 or Shh expression in the node. Inserts, arrows and orientation as in Fig. 1. 400 A. Granata, N.A. Quaderi / Developmental Biology 258 (2003) 397–405

In situ hybridisation, embryo culture, and bead quence differs slightly from that reported by Richman et al. implantations (2002); however, these nucleotide polymorphisms do not alter the predicted amino acid sequence apart from at posi- In situ hybridisation was performed using standard tech- tion 106 where there is a conserved glu/asp variation. The niques (Myat et al., 1996). A 478-bp cMid1 subclone was predicted protein sequences of human (Quaderi et al., 1997) used as a template for riboprobe generation. Embryos were and chick MID1 share 95% identity, indicating that its explanted ventral side up in EC culture (Chapman et al., function is conserved between the two species. We have 2001). Beads were implanted into a small incision adjacent used whole-mount in situ techniques to determine the cMid1 to the node. For clarity, when the position of implanted expression pattern during development. cMid1 is first seen beads is referred to in the main text, it is with respect to a at stage 3 in an anterior crescent-shaped domain (data not dorsal view of the embryo. Affigel-Blue beads (Bio-Rad) shown). This expression is stronger at stage 4, when weak were soaked in SHH (R&D Systems; 0.27 mg/ml in 1% cMid1 expression is also seen in Hensen’s node (Fig. 1A). BSA/PBS), or a monoclonal anti-SHH blocking antibody The bilaterally symmetric expression in the ectoderm of the (0.5 mg/ml (Ericson et al., 1996)), or activin A (R&D node is stronger by stage 4ϩ (Fig. 1B); however, by early Systems; 0.5 mg/ml in PBS), or BMP4 (R&D Systems; 1 stage 5, cMid1 expression is downregulated on the left side ␮g/ml in PBS) or chordin (R&D Systems; 2 mg/ml in PBS). of Hensen’s node (Fig. 1C). This asymmetry is still evident Heparin acrylic beads (Sigma) were soaked in FGF8 (R&D at stage 7, but at this stage cMid1 is seen in the mesendo- Systems; 1 mg/ml in PBS). Control bead implantations were derm rather than the ectoderm of the node (Fig. 1D). After performed by using beads soaked in 1% BSA/PBS (for the one-somite stage, cMid1 expression in the node is down- SHH) or PBS. regulated, and by stage 8, a weak, symmetrical expression pattern is observed (Fig. 1E). Expression constructs, antisense oligonucleotides, and The asymmetric pattern of cMid1 expression is highly electroporations suggestive of regulatory relationships with other molecules involved in L/R determination. The first molecular marker The cMid1 coding region was amplified by PCR and of L/R asymmetry in chick is the right-sided expression of cloned into the pCAB expression vector between the ␤-actin activin receptor IIA (cAct-RIIa) in the node at stage 4 promoter and IRES-GFP sequence. A fluorescein-tagged (Levin et al., 1995). Right-sided signalling through cAct- antisense morpholino oligonucleotide directed against the RIIa induces Bmp4 (Monsoro-Burq and Le Douarin, 2001), cMid1 translational start sequence (cMid1.mo: 5Ј-TCAGTTC- which induces Fgf8 (Boettger et al., 1999), which in turn CGACTCCAGTGTTTCCAT-3Ј) and a fluorescein tagged induces Snail (Isaac et al., 1997) in the LPM. The asym- standard control morpholino (control.mo: 5Ј-CCTCTTAC- metric expression profiles of the key L/R genes used in this CTCAGTTACAATTTATA-3Ј) were purchased from Gene study are described in Table 1. Activin is able to induce the Tools, LLC (Corvallis, OR). A modified EC culture tech- expression of right pathway genes, suggesting that an en- nique was used to transfer stage 4 embryos, dorsal side up dogenous activin-like molecule initiates the right-sided cas- with the vitelline membrane removed, into a small electro- cade (Levin et al., 1995). To examine whether cMid1 is also poration chamber containing a positive platinum electrode regulated by activin, beads soaked in activin A were im- in its base. Embryos were positioned such that this electrode planted on the left of Hensen’s node in stage 4 embryos was on the left (for pCAB.cMid1) or on the right (for (Fig. 1F), which were then cultured to stage 6/7. Activin- cMid1.mo) of Hensen’s node. pCAB.cMid1 (2 ␮g/␮lin treated embryos showed bilateral cMid1 expression in the 0.1% Fast Green for visualisation) or cMid1.mo (1 mM in node (n ϭ 25/31; 81%) (Fig. 1G). Control beads soaked in 0.1% Fast Green for visualisation) was applied to the left or PBS had no effect on cMid1 (n ϭ 16/16) (Fig. 1H). Thus, right side of the node, respectively. Electroporations were activin is able to positively regulate cMid1 expression. performed by placing a negative tungsten electrode over the We then performed a series of experiments to determine DNA/morpholino solution and applying four to five pulses where cMid1 acts in the known right-sided signalling path- (5 V, 50 ms) using an Intracept TSS10 dual pulse isolated way. The cMid1 expression vector pCAB.cMid1, which stimulator (Intracel, UK). Control electroporations were also expresses GFP, was focally electroporated into the left performed by using pCAB.GFP or a control morpholino. side of the node in stage 4 embryos (Fig. 2A). The typical Embryos were then grown as EC cultures until the appro- extent of pCAB.cMid1 expression is shown by the region of priate stage. GFP, or fluorescein, fluorescence was used to GFP fluorescence approximately 2 h after electroporation monitor the position and efficiency of electroporation. (Fig. 2B). By stage 7, a more diffuse region of GFP fluorescence is observed as a result of extensive cell movements around the node during gastrulation (Fig. 2C). Embryos Results were cultured to stage 7 and assayed for Bmp4 and Fgf8 expression, or cultured to stage 8ϩ/9Ϫ and assayed for We have cloned the chick Midline1 orthologue, cMid1, Snail expression. cMid1 electroporations resulted in ectopic and determined its consensus cDNA sequence. Our se- Bmp4 (n ϭ 7/7; 100%; Fig. 2D) and Fgf8 (n ϭ 11/15; 73%; A. Granata, N.A. Quaderi / Developmental Biology 258 (2003) 397–405 401

Fig. 2F) on the left side of the node, and ectopic Snail (n ϭ blocking BMP signalling using noggin results in bilateral 18/23; 78%; Fig. 2H) in the left LPM. It is noteworthy that Shh expression in the node (Monsoro-Burq and Le Douarin, ectopic Bmp4 and Fgf8 expression is confined to the node, 2001); our chordin-treated embryos also showed bilateral to give a symmetrical expression pattern, and not seen in Shh expression (n ϭ 7/8; 88%; Fig. 3L). PBS-treated con- surrounding tissue (Fig. 2D and F). Control GFP electro- trols showed no change in Shh expression (n ϭ 4/4; Fig. porations had no effect on Bmp4 (n ϭ 5/5; Fig. 2E), Fgf8 (n 3M). The observation that blocking BMP4 signalling does ϭ 7/7; Fig. 2G), or Snail (n ϭ 11/11; Fig. 2I). These results not affect cMid1 expression confirms that Bmp4 is down- indicate that cMid1 can ectopically induce Bmp4, and other stream of cMid1 in the right pathway. However, BMP4 can downstream members of the right pathway, on the left side upregulate left-sided cMid1 expression (see above), and it is of the embryo. known that BMP4 can inhibit endogenous Shh expression To investigate whether cMid1 is necessary for progres- on the left side of Hensen’s node (Monsoro-Burq and Le sion of the right pathway, we examined the effect of knock- Douarin, 2001). Taken together, these results suggest that ing-down endogenous MID1. A fluorescein-tagged anti- the upregulation of left-sided cMid1 in BMP4 treated em- sense morpholino oligonucleotide directed against cMid1 bryos may be mediated by a downregulation of left-sided (cMid1.mo) was focally electroporated into the right side of Shh expression. Hensen’s node at stage 4 (Fig. 2J). The typical extent of Shh is initially symmetrically expressed, but is then re- cMid1.mo electroporation is shown by the region of fluo- stricted to the left side of the node by stage 5Ϫ (Levin et al., rescein fluorescence (Fig. 2K). By stage 7, a larger region of 1995). In contrast, expression of cMid1 is restricted to the fluorescence is observed as a result of extensive cell move- right side at this stage. Given the mutually exclusive ex- ments around the node (Fig. 2L). Embryos were cultured to pression patterns of Shh and cMid1 in Hensen’s node be- stage 7 and assayed for Bmp4 or Fgf8, or cultured to stage tween stage 5Ϫ to stage 7, we investigated whether an 8ϩ/9Ϫ and assayed for Snail. Embryos electroporated with antagonistic relationship exists between them. Beads soaked cMID1.mo showed an absence of Bmp4 (n ϭ 10/11; 91%; in SHH protein were implanted to the right of the node in Fig. 2M) and Fgf8 (n ϭ 13/15; 87%; Fig. 2O) in the node, stage 4 embryos (Fig. 4A), which were then cultured to and an absence of Snail in the LPM (n ϭ 10/16; 63%; Fig. stage 6/7. SHH beads implanted at this stage were able to 2Q). The downregulation of right-sided Bmp4 and Fgf8 was abolish cMid1 expression in the node (n ϭ 35/47; 74%; Fig. limited to the node, and did not extend into the primitive 4B), while beads implanted slightly later (stage 5Ϫ) did not streak (Fig. 2M and O). Embryos electroporated with a affect cMid1 (data not shown). Control beads soaked in 1% control morpholino showed no change in Bmp4 (n ϭ 6/6; BSA/PBS had no effect on cMid1 (n ϭ 27/27; Fig. 4C). Our Fig. 2N), Fgf8 (n ϭ 7/7; Fig. 2P), or Snail (n ϭ 11/11; interpretation of these results is that ectopic SHH can pre- Fig. 2R). vent the induction of cMid1 expression in the node, but Although cMid1 expression is first detected before that cannot downregulate right-sided cMid1 expression once es- of Bmp4orFgf8 (see Table 1), there is some temporal tablished. This interpretation is supported by our observa- overlap in their expression with cMid1 on the right side of tion that blocking BMP signalling using chordin results in the node, allowing for the possibility that they positively bilateral Shh expression without a concomitant downregu- regulate cMid1. Beads soaked in BMP4 or FGF8 were lation of right-sided cMid1 expression (see above). To fur- implanted on the left of Hensen’s node in stage 4 embryos ther investigate the regulatory relationship between Shh and (Fig. 3A and D), which were then harvested at stage 6/7 cMid1, we analysed the effect of blocking endogenous when left-sided cMid1 expression has normally been down- SHH. Beads soaked in SHH-blocking antibody were ap- regulated. BMP4-treated embryos showed bilateral cMid1 plied to the left of Hensen’s node at stage 4 (Fig. 4D). expression in the node (n ϭ 15/15; 100%; Fig. 3B), while Embryos were harvested at stage 6/7 and bilateral cMid1 control beads soaked in PBS had no effect on cMid1 (n ϭ expression was observed (n ϭ 25/29; 86%; Fig. 4E). Con- 10/10; Fig. 3C). Neither FGF8 (n ϭ 0/48; 0%; Fig. 3E) nor trol PBS-soaked beads had no effect on cMid1 (n ϭ 9/9; PBS control (n ϭ 20/20; Fig. 3F) beads had any effect on Fig. 4F). These results demonstrate that the restriction of cMid1. FGF8 activity was verified by its ability to induce cMid1 to the right side of the node is a result of repression Snail (data not shown; Boettger et al., 1997). by SHH on the left side. We then analysed the effect of blocking endogenous To test whether cMid1 is able to regulate Shh expression, BMP4 signalling by applying the BMP antagonist chordin pCAB.cMid1 was focally electroporated into the left of the to the right side of Hensen’s node in stage 4 embryos (Fig. node at stage 4 (Fig. 5A, and see Fig. 2B), and the embryos 3G), which were then cultured to stage 6. Neither chordin (n cultured to stage 5ϩ/6. cMid1 electroporations specifically ϭ 0/18; 0%; Fig. 3H) nor PBS control (n ϭ 7/7; Fig. 3I) abolished Shh expression in the node, without affecting Shh beads had any effect on cMid1 expression. However, chor- expression in the head process (n ϭ 12/16; 75%; Fig. 5B). din-treated embryos showed an absence of Fgf8 expression Control GFP electroporations had no effect on Shh (n ϭ in the node (n ϭ 7/9; 78%; Fig. 3J), confirming its activity. 7/7; Fig. 5C). To investigate whether the repression of Shh PBS-treated embryos showed no change in Fgf8 expression by cMid1 affects the downstream left pathway, we tested the (n ϭ 4/4; Fig. 3K). It has previously been demonstrated that effect of cMid1 on Nodal (Levin et al., 1995). pCAB.cMid1 402 A. Granata, N.A. Quaderi / Developmental Biology 258 (2003) 397–405

Fig. 4. Regulation of cMid1 by SHH. (A, B) SHH-soaked beads were implanted to the right of the node in stage 4 embryos. (B) cMid1 expression is downregulated in the right side of the node. (C) Control beads soaked in 1% BSA/PBS do not affect cMid1 expression in the node. (D, E) ␣SHH blocking antibody-soaked beads were implanted to the left of the node in stage 4 embryos. (E) cMid1 expression is upregulated on the left side of the node. (F) Control beads soaked in PBS do not affect cMid1 expression in the node. Inserts, arrows and orientation as in Fig. 1. Fig. 5. Regulation of the left pathway by cMid1. (A, B, D) A cMid1 expression vector was electroporated into the left side of Hensen’s node in stage 4 embryos. (B) Shh expression is downregulated on the left side of Hensen’s node. (D) Nodal expression is downregulated in the left perinodal region and LPM. (C, E) Left-sided downregulation of Shh or Nodal is not seen in embryos electroporated with a control GFP vector. (F, G, I) An antisense morpholino directed against cMid1 was electroporated into the right side of Hensen’s node in stage 4 embryos. (G) Ectopic Shh expression on the right side of the node. (I) Ectopic Nodal expression in the right perinodal region and LPM. (H, J) Ectopic right-sided Shh or Nodal expression is not seen in embryos electroporated with a control morpholino. Inserts, arrows and orientation as in Fig. 1. was electroporated into the left side of the node in stage 4 MID1 on Shh and downstream left pathway genes, the embryos (Fig. 5A), which were then harvested at stage 7. antisense morpholino cMid1.mo was focally electroporated cMid1 electroporations repressed Nodal expression in both into the right side of the node in stage 4 embryos (Fig. 5F, the perinodal region and LPM (n ϭ 12/14; 86%; Fig. 5D). and see Fig. 2K). These embryos were then cultured to stage Control GFP electroporations had no effect on Nodal (n ϭ 5ϩ/6 and assayed for Shh expression, or cultured to stage 7 8/8; Fig. 5E). and assayed for Nodal expression. Embryos electroporated To study the effects of knocking down endogenous with cMID1.mo showed ectopic Shh on the right side of the A. Granata, N.A. Quaderi / Developmental Biology 258 (2003) 397–405 403

Fig. 6. Establishment of molecular asymmetry, and L/R identity, in Hensen’s node. Regulatory interactions involving cMid1 have been integrated into the known asymmetric signalling cascade to provide a revised model for early L/R patterning in chick. (A) The first molecular sign of L/R asymmetry is seen in Hensen’s node at stage 4. However, Shh and cMid1 are expressed symmetrically at this stage, and Hensen’s node does not yet have a fixed L/R identity. (B) The L/R identity of Hensen’s node is fixed by stage 5. This coincides with the point at which Shh and cMid1 show a laterally restricted, mutually exclusive, expression pattern. A series of regulatory interactions involving Shh, cMid1, and Bmp4 actively maintains asymmetric gene expression in Hensen’s node.

node (n ϭ 10/10; 100%; Fig. 5G) and ectopic Nodal in the Discussion right perinodal region and LPM (n ϭ 8/10; 80%; Fig. 5I). Ectopic Shh expression is restricted to the right side of the We investigated the regulatory relationships between node. Electroporations using a control morpholino had no cMid1 and other molecules involved in L/R determination effect on Shh (n ϭ 5/5; Fig. 5H) or Nodal (n ϭ 9/9; Fig. 5J). (Capdevila et al., 2000; Wright, 2001) and have integrated These results indicate that the presence of MID1 on the right cMid1 into the known signalling cascade to generate a side of the node represses right-sided Shh, thus confining revised model for the establishment of molecular asymme- Shh, and the subsequent induction of the left pathway, to the try in Hensen’s node (Fig. 6). Both Shh and cMid1 are left side of Hensen’s node. As cMid1 induces Bmp4, and initially expressed bilaterally in the node until stage 5Ϫ. BMP4 can repress Shh (Monsoro-Burq and Le Douarin, Subsequently, an activin-like signal, possibily Activin ␤B 2001), it is likely that that the repression of Shh by cMid1 on (Levin et al., 1997), signalling through the cAct-RIIa recep- the right side of Hensen’s node is mediated by BMP4. tor, activates the right pathway by stabilising right-sided 404 A. Granata, N.A. Quaderi / Developmental Biology 258 (2003) 397–405 cMid1 expression. cMid1 then induces the expression of human L/R axis. However, the lack of OS-associated later- Bmp4 on the right side of the node. BMP4 represses right- ality defects does not necessarily exclude MID1 involve- sided Shh expression (Monsoro-Burq and Le Douarin, ment in human L/R determination: functional redundancy 2001), thus restricting Shh to the left side of the node. between MID1 and its closely related homologue MID2 Left-sided SHH represses cMid1 expression on the left, (Buchner et al., 1999) in this context may allow L/R deter- restricting cMid1, and the subsequent induction of Bmp4,to mination to proceed normally in the absence of MID1 while the right side of the node. Thus, the regulatory relationships development of midline structures is compromised. involving Shh, cMid1, and Bmp4 actively maintain asym- The negative regulatory relationship between cMid1 and metric gene expression in the node by restricting Shh to the Shh might not be confined to Hensen’s node, and may play left, and cMid1 and Bmp4 to the right side of Hensen’s a part in patterning other organs during embryogenesis. In node. particular, antagonism between cMid1 (see Fig. 1D) and The first molecular sign of asymmetry in the node, the Shh in the axial mesendoderm may be involved in position- right-sided expression of cAct-RIIa, is detected at stage 4 ing the eye fields (Roessler and Muenke, 2001). Shh signal- (Levin et al., 1995). However, the L/R identity of the node ling from the prechordal plate to the overlying neural plate is still labile at this stage (Pagan-Westphal and Tabin, splits the single, central eye field into bilateral eye fields 1998), and Shh and cMid1 are still symmetrically expressed. (Roessler and Muenke, 2001). Patients with SHH mutations However, by stage 5, Hensen’s node gains an intrinsic L/R have hypotelorism, or in extreme cases, have a single cen- identity and is no longer under the influence of peripheral tral eye (Belloni et al., 1996; Roessler et al., 1996). Con- tissues (Pagan-Westphal and Tabin, 1998). Notably, this versely, mutations in GLI3 (Vortkamp et al., 1991) and PTC coincides with the time at which the symmetrical expression (Hahn et al., 1996; Johnson et al., 1996), both of which of Shh and cMid1 in the node becomes laterally restricted normally limit SHH signalling, result in hypertelorism. OS into two mutually exclusive left and right-sided domains patients also suffer from hypertelorism, supporting our hy- (Fig. 6). pothesis that MID1 normally acts to negatively regulate Ectopic expression of cMid1 on the left side of the node SHH activity. We propose that a loss of MID1 function (Cox is sufficient to induce left-sided Bmp4 and repress endoge- et al., 2000) in OS patients results in SHH “over-activity”, nous Shh, so blocking the progression of the left pathway, causing an expansion of the ventral midline. Preliminary and ectopically activating the right pathway on the left side results suggest that manipulating cMid1 expression in the of the embryo. Conversely, knocking-down endogenous head process can effect Shh expression in the axial mesen- MID1 results in an absence of right pathway genes and doderm (data not shown), and further experiments are in ectopic expression of Shh, and the subsequent activation of progress. the left pathway, on the right side of the embryo. It has previously been noted that regulation of L/R gene expression in Hensen’s node by the application of exogenous Acknowledgments factors using beads or cell pellets only works over very short distances (Monsoro-Burq and Le Douarin, 2001). We We thank Claudio Stern and Andrew Lumsden for help- have found that manipulating gene expression by electro- ful discussions and critical appraisal of this work, Andrea poration of pCAB.cMid1, or cMid1.mo, is also only effec- Streit for technical advice, and Dawn Savery for technical tive when performed at close range: even slightly lateral assistance. We would also like to thank the following for electroporations had no effect on L/R gene expression. Fur- probes: Cliff Tabin for Shh and Nodal; Philippa Francis- thermore, the changes in Shh, Bmp4, and Fgf8 expression West for Bmp4; Ivor Mason for Fgf8; David Wilkinson for were confined to the node and did not extend to adjacent Snail and Claudio Stern for the pCAB vector. This work tissues. was supported by a Wellcome Trust Career Development Factors involved in the very early stages of L/R deter- Fellowship and a Royal Society grant (to N.A.Q.). mination in chick have been described (Essner et al., 2002; Levin and Mercola, 1999; Levin et al., 2002), but it is still unclear why the repression of cMid1 by Shh on the left side References of the node does not begin until stage 5Ϫ. One possibility is that this antagonistic interaction is mediated by a left-sided Belloni, E., et al., 1996. Identification of Sonic hedgehog as a candidate gene responsible for holoprosencephaly. Nat. Genet. 14, 353–356. molecule that is not expressed, or active, at earlier stages. Boettger, T., et al., 1999. FGF8 functions in the specification of the right The identity of this molecule remains to be revealed. body side of the chick. Curr. Biol. 9, 277–280. There is often an association between midline and L/R Buchner, G., et al., 1999. MID2, a homologue of the Opitz syndrome gene defects (Roessler and Muenke, 2001), but laterality defects MID1: similarities in subcellular localization and differences in expres- are not typical of OS. This may reflect the fact that the sion during development. Hum. Mol. Genet. 8, 1397–1407. Capdevila, J., et al., 2000. Mechanisms of left–right determination in mechanisms of early L/R determination are poorly con- vertebrates. Cell 101, 9–21. served between species (Capdevila et al., 2000; Meyers and Chapman, S.C., et al., 2001. Improved method for chick whole-embryo Martin, 1999) and MID1 is not required for patterning the culture using a filter paper carrier. Dev. Dyn. 220, 284–289. A. Granata, N.A. Quaderi / Developmental Biology 258 (2003) 397–405 405

Cox, T.C., et al., 2000. New mutations in MID1 provide support for loss of via binding to the regulatory subunit alpha 4. Proc. Natl. Acad. Sci. function as the cause of X-linked Opitz syndrome. Hum. Mol. Genet. USA 98, 6650–6655. 9, 2553–2562. Meyers, E.N., Martin, G.R., 1999. Differences in left–right axis pathways Ericson, J., et al., 1996. Two critical periods of Sonic Hedgehog signaling in mouse and chick: functions of FGF8 and SHH. Science 285, 403– required for the specification of motor neuron identity. Cell 87, 661– 406. 673. Monsoro-Burq, A., Le Douarin, N.M., 2001. BMP4 plays a key role in Essner, J.J., et al., 2002. Conserved function for embryonic nodal cilia. left–right patterning in chick embryos by maintaining Sonic Hedgehog Nature 418, 37–38. asymmetry. Mol. Cell 7, 789–799. Garcia-Castro, M.I., et al., 2000. N-Cadherin, a cell adhesion molecule Myat, A., et al., 1996. A chick homologue of Serrate and its relationship involved in establishment of embryonic left–right asymmetry. Science with Notch and Delta homologues during central neurogenesis. Dev. 288, 1047–1051. Biol. 174, 233–247. Gaudenz, K., et al., 1998. Opitz G/BBB syndrome in Xp22: mutations in Pagan-Westphal, S.M., Tabin, C.J., 1998. The transfer of left–right posi- the MID1 gene cluster in the carboxy-terminal domain. Am. J. Hum. tional information during chick embryogenesis. Cell 93, 25–35. Genet. 63, 703–710. Quaderi, N.A., et al., 1997. Opitz G/BBB syndrome, a defect of midline Hahn, H., et al., 1996. Mutations of the human homolog of Drosophila patched development, is due to mutations in a new RING finger gene on Xp22. in the nevoid basal cell carcinoma syndrome. Cell 85, 841–851. Nat. Genet. 17, 285–291. Isaac, A., et al., 1997. Control of vertebrate left–right asymmetry by a Richman, J.M., et al., 2002. Isolation and characterisation of the chick snail-related zinc finger gene. Science 275, 1301–1304. orthologue of the Opitz syndrome gene, Mid1, supports a conserved Johnson, R.L., et al., 1996. Human homolog of patched, a candidate gene role in vertebrate development. Int. J. Dev. Biol. 46, 441–448. for the basal cell nevus syndrome. Science 272, 1668–1671. Rodriguez-Esteban, C., et al., 2001. Wnt signaling and PKA control Nodal Kawakami, M., Nakanishi, N., 2001. The role of an endogenous PKA expression and left-right determination in the chick embryo. Develop- inhibitor, PKIalpha, in organizing left–right axis formation. Develop- ment 128, 2509–2515. ment 128, 3189–3195. Levin, M., Mercola, M., 1999. Gap junction-mediated transfer of left-right Roessler, E., et al., 1996. Mutations in the human Sonic Hedgehog gene patterning signals in the early chick blastoderm is upstream of Shh cause holoprosencephaly. Nat. Genet. 14, 357–360. asymmetry in the node. Development 126, 4703–4714. Roessler, E., Muenke, M., 2001. Midline and laterality defects: left and Levin, M., et al., 1995. A molecular pathway determining left–right asym- right meet in the middle. Bioessays 23, 888–900. metry in chick embryogenesis. Cell 82, 803–814. Sontag, E., 2001. Protein phosphatase 2A: the Trojan Horse of cellular Levin, M., et al., 1997. Left/right patterning signals and the independent signaling. Cell Signal. 13, 7–16. regulation of different aspects of situs in the chick embryo. Dev. Biol. Trockenbacher, A., et al., 2001. MID1, mutated in Opitz syndrome, en- 189, 57–67. codes an ubiquitin ligase that targets phosphatase 2A for degradation. Levin, M., et al., 2002. Asymmetries in H(ϩ)/K(ϩ)-ATPase and cell Nat. Genet. 29, 287–294. membrane potentials comprise a very early step in left–right patterning. Vortkamp, A., et al., 1991. GLI3 zinc-finger gene interrupted by translo- Cell 111, 77. cations in Greig syndrome families. Nature 352, 539–540. Liu, J., et al., 2001. Phosphorylation and microtubule association of the Wright, C.V., 2001. Mechanisms of left–right asymmetry: what’s right and Opitz syndrome protein mid-1 is regulated by protein phosphatase 2A what’s left? Dev. Cell 1, 179–186.