A ryanodine receptor-dependent Ca_i²⁺ asymmetry at Hensen's node mediates avian lateral identity

Establishment of the vertebrate left-right (LR) body plan initiates proper positioning of the internal organs. Early differences in small molecules and signals across the midline organizer are translated into an asymmetric genetic cascade that enforces LR identity. Effectors of these early events include cilia motility (Essner et al.,2005; McGrath et al.,2003; Tabin and Vogan,2003), cell coupling (Levin and Mercola, 1999), ion pumps(Levin et al., 2002; Adams et al., 2006; Shu et al., 2007), inositol polyphosphates (Sarmah et al.,2005), serotonin (Fukumoto et al., 2005a; Fukumoto et al.,2005b) and retinoic acid (Chauzad et al., 1999; Tsukui et al., 1999; Zile et al., 2000). Although genetic effectors, including sonic hedgehog (Shh), Nodal, lefty and Pitx2, are largely conserved across vertebrates, there may be significant differences as to which upstream participants are used. Some differences might reflect evolutionary divergence and structural changes in the embryo. The precise relationships among these participants are also unclear.

One poorly understood early asymmetry signal is calcium. In mouse and zebrafish, intracellular calcium (Ca_i²⁺) is enriched along the left margin of the node and Kupffer's vesicle, respectively(McGrath et al., 2003; Sarmah et al., 2005). In mouse, a loss-of-function mutation in left-right dynein (iv/iv) or polycystin 2 collapses calcium asymmetry at the node and randomizes LR identity (McGrath et al.,2003), prompting suggestions that it is cilia-generated flow across the ventral node that generates the calcium asymmetry and subsequent left-sided gene expression. A role for calcium asymmetry in zebrafish is less clear, as Ca_i²⁺ fluxes are also linked to KV formation and subsequent laterality establishment(Sarmah et al., 2005; Schneider et al., 2008).

Whether and how Ca_i²⁺ asymmetry across the node might contribute to lateral identity in other vertebrates is unknown. Although the node architecture of the avian embryo might preclude nodal flow (Manner et al., 2001), its node monocilia are positioned toward the ventral endoderm(Essner et al., 2002) and it expresses several ciliary-related genes, including left-right dynein,kinesin 3B and polycystin 2(Qiu et al., 2005). Thus, at least some elements of this laterality mechanism might be conserved in chick. Ca_i²⁺ levels have not been examined in chick, although,at neurulation, a left-sided external calcium pool affects laterality through the activation of notch signaling(Raya et al., 2004).

We report here that the avian embryo uses an asymmetric enrichment of Ca_i²⁺ at the node to mediate laterality. This Ca_i²⁺ enrichment emerges at early gastrulation and originates from the activity of ryanodine receptors and calcium-induced calcium release (CICR). Several previously identified laterality mediators,specifically serotonin, H⁺V-ATPase and H⁺K⁺ATPase, regulate LR identity by directly affecting node Ca_i²⁺ concentrations. Our data significantly advance the understanding of how the avian embryo achieves its LR identity,and show that, despite differences in spatial architecture,Ca_i²⁺ asymmetry across the node might be a conserved mechanism of laterality in amniotes.

Chick eggs (Hyline W98) were obtained from the UW Poultry Research Laboratory, Madison, WI. Normal and retinoid-deficient Japanese quail eggs were generated as described previously(Zile et al., 2000). Embryos were staged according to Hamburger and Hamilton (HH)(Hamburger and Hamilton,1951).

HH3+ to HH6 embryos were incubated (60 minutes, 37°C) in a slide chamber with Tyrode's buffer (TWC) containing 25 μM Fura-2-AM and 0.1%Pluronic F-127 (Molecular Probes), rinsed and then incubated (20 minutes) with one of the following pharmacological agents: Bapta-AM (1 mM), EGTA (1.0-2.5 mM), xestospongin C (1 μM), U73122 (10 μM), 9,21-didehydroryanodine (2μM), ryanodine isomers (10 μM), dantrolene (2 μM), fluoxetine (10μM), 2-methyl-5HT (25 μM), ondansetron (25 μM), ML10302 (25 μM),GR125487 (25 μM), concanamycin (100 nM) or lansoprazole (7 μM)(concentrations were determined experimentally). Embryos were rinsed and immediately imaged. Studies of CICR (ryanodine plus EGTA) used calcium-free TWC.

Ca²⁺ imaging used an inverted microscope (Nikon TE2000-U) with a Fluor-S ×10 lens, Xenon lamp, and Sutter filter wheel with Fura-2 filters (Chroma Technology, Brattleboro, VT); HH4/HH5 analysis was carried out at a magnification of ×20. Using MetaFluor imaging software (Universal Imaging, West Chester, PA), the left and right regions of interest were defined and images collected at 1 second intervals for 3 minutes using a CCD digital camera (Cool Snap-ES, Photometrics, Tucson, AZ). Fluorescent signal emission at 510 nm, following dual excitation at 340 and 380 nm, was calculated for each region and timepoint. Mean response per region and per embryo were calculated from 10 consecutive images. At the end of the assay,ionomycin challenge (0.1 mM) affirmed embryo viability and determined R_max; non-responding embryos were discarded. Subsequent treatment with 2 M MnCl₂ quenched Ca²⁺ -bound Fura-2 and defined the background fluorescence (R_min).

Data were analyzed using two approaches. The line-scan function in MetaFluor quantified the mean Fura-2 signal within the left and right side at a fixed distance from the node; results were expressed as the mean Fura-2 signal of six to 12 embryos per treatment. To calculate the fold change in left versus right regions, we used the equation Ca_i²⁺=K_d×[(R-R_min)/(R_max-R)]×(F_max³⁸⁰/F_min³⁸⁰)(Grynkiewicz et al., 1985),where the constant terms K_d and F_max³⁸⁰/F_min³⁸⁰ were cancelled because a ratio was calculated. R_min is the emission ratio after MnCl₂ treatment, R_max is the emission ratio after ionomycin treatment, and R represents the emission ratio of experimental interest. R_min and R_max were calculated individually for each region. Results were expressed as the mean left-right ratio for five to nine embryos per treatment.

A microbead soaked in the agent of interest was implanted to the left or right of Hensen's node of in ovo HH3++/HH4 embryos: Bapta-AM (30 mM),ionomycin (1 mM), U73122 (10 mM), mixed ryanodine isomers (2.5 mM), EGTA (5 mM), calmidizolium (5 μM), fluoxetine (100 μM), 2-CH3-5HT (5 mM),ondansetron (5 mM), ML10302 (5 mM), GR125487 (5 mM), concanamycin (100 μM),lansoprazole (7 mM) or DMSO only. Ryanodine and EGTA were applied as small agarose plugs. Beads were removed 4 hours later. At HH10/HH11, embryos with significant cranial or midline defects were discarded. The remainder were scored for cardiac laterality as left or right loop, by two treatment-blinded individuals. We analyzed eight to 25 embryos per treatment.

cDNA encoding chick Nodal (cNR-1)(Levin et al., 1995) was provided by C. Tabin. HH4 embryos were implanted to the left or right of Hensen's node with DMSO, Bapta-AM, ionomycin, ryanodine-EGTA or calmidizolium-soaked microbeads. At HH8, embryos were processed for whole-mount in situ hybridization as described previously(Smith et al., 1997). For whole embryo immunostaining, antibody directed against the carboxy-terminus of all three RyR isoforms (C-18, 1/500, Santa Cruz) was reacted with HH3+ to HH6 embryos as described (Smith et al.,1997). Signal was visualized using alkaline phosphatase-conjugated secondary antibody (Southern Biotech) and BM-Purple (Roche). Embryos were processed batch-wise to ensure consistent treatment. We analyzed eight to 16 embryos per treatment.

Binary data (e.g. heart laterality) were analyzed using χ²analysis (SAS 9.1, SAS Institute, Cary, NC). Normally distributed data were subjected to an unpaired, two-tailed t-test employing the appropriate variance parameter using SigmaStat v.2.0 (Systat Software, Point Richmond,CA). Data not normally distributed were examined using the Mann-Whitney U-test. P<0.05 was set as the level of significance.

We used Fura-2-based ratiometric imaging to evaluate Ca_i²⁺ concentrations at the chick Hensen's node. At HH3++/HH4, a modest and bilaterally symmetric Ca_i²⁺elevation was first detected along the anterior margin of the node(Fig. 1A,D). A modest left-side enrichment was occasionally observed (Fig. 1D). At HH4+/HH5-, the calcium signal extended posteriorly and the emerging prechordal plate split this signal into bilateral fields, with levels being lower at the midline (Fig. 1B,E). By HH5-, left Ca_i²⁺ levels were consistently elevated when compared with the right side(Fig. 1C,F); anterior node levels were reduced. By HH6, the Ca_i²⁺ signal extended more posteriorly and it was significantly enriched along the left side (LR ratio 2.85±0.44; Table 1, Fig. 2B). The left Ca_i²⁺ field appeared to overlap with a left-sided extracellular calcium enrichment previously described(Raya et al., 2004). Treatment with the intracellular Ca_i²⁺ chelator Bapta-AM reduced Ca_i²⁺ levels throughout, and ablated the left-right differential (Table 1, Fig. 2C), affirming that this signal represented Ca_i²⁺.

We manipulated Ca_i²⁺ levels to the left or right side of Hensen's node and analyzed the effects on asymmetry. An early asymmetry indicator in chick is the right loop of the heart tube at HH10. DMSO treatment had little effect and most embryos had normal, right-looped hearts(Fig. 3, Fig. 4A). A small percentage had midline-positioned heart tubes that failed to loop or fully fuse; their incidence was not treatment dependent and they were excluded from the analysis. Left but not right Bapta-AM treatment at HH4 significantly increased the incidence of left-looped hearts (40% situs inversus; Fig. 3, Fig. 4B). To test whether Ca_i²⁺ enrichment alone was sufficient to influence laterality, we applied the potent Ca_i²⁺ ionophore ionomycin. Right but not left ionomycin treatment increased the frequency of left-looped hearts (31%; Fig. 3, Fig. 4C).

Nodal is an early laterality determination gene that is restricted to the left-side of Hensen's node and the lateral plate mesoderm during early somitogenesis (Levin et al.,1995). Although DMSO did not affect Nodal expression(left treatment, 10/10 normal; right treatment, 8/8 normal; Fig. 5A,B), left-sided Bapta-AM treatment at HH4 caused bilateral Nodal expression in the lateral plate mesoderm in 60% of embryos (n=10; Fig. 5C) and normal expression in the remainder. Right-sided Bapta-AM did not affect Nodal (8/8 normal; Fig. 5D). Conversely,right-sided ionomycin reduced Nodal expression (4/6 embryos, 67%; Fig. 5E), but left-sided treatment did not (1/7 lacked Nodal). Both calcium effectors also affected cranial morphology consistent with the known effects of calcium on neural fold elevation. Although we cannot rule out that such cranial changes may have disrupted midline formation and thus affected LR identity(Schneider et al., 2008), this seems unlikely because, while left and right treatment equally affected cranial morphology, only left Bapta and right ionomycin significantly affected Nodal, suggesting that the effects on Nodal and the neural folds were separable events. This suggested that Ca_i²⁺acted upstream of Nodal to affect laterality.

Both extracellular (Raya et al.,2004) and intracellular calcium(McGrath et al., 2003; Sarmah et al., 2005; Shu et al., 2007) affect LR laterality and could originate the Ca_i²⁺ enrichment. The left Ca_i²⁺ enrichment overlaps at later stages (HH6)with a left-side enriched extracellular calcium pool that has previously been demonstrated to govern chick laterality, heart looping and Nodalexpression (Raya et al.,2004). EGTA treatment ablated the Ca_i²⁺asymmetry and attenuated LR Ca_i²⁺ levels(Fig. 2D, Table 1), suggesting that extracellular calcium contributed to the Ca_i²⁺enrichment. EGTA also reversed heart looping when applied to the left side(25%, P=0.006) but not the right (0%, Fig. 3).

Intracellular calcium could originate from either phosphoinositide or RyR activity. Phosphoinositides are implicated in zebrafish laterality through roles in Kupffer's vesicle formation(Schneider et al., 2008) and ciliary beating (Sarmah et al.,2005; Sarmah et al.,2007). However, neither xestospongin C, which inhibits IP₃-mediated Ca_i²⁺ release, nor the phosphoinositidyl-phospholipase C antagonist U73122, affected absolute Ca_i²⁺ levels (Fig. 2E,F) or the LR Ca_i²⁺ ratio(Table 1), nor did U73122 affect asymmetry (Fig. 3). Thus the Ca_i²⁺ did not originate from IP₃-mediated sources.

RyRs mediate calcium mobilization from sarco/endoplasmic stores following stimulation by smaller calcium quantities that often originate from extracellular sources, a process known as calcium-induced calcium release(CICR) (Zucchi and Roncha-Testoni,1997). RyRs have not been previously implicated in laterality. Treatment with the RyR antagonist 9,21-dehydroryanodine ablated the LR Ca_i²⁺ asymmetry and significantly lowered Ca_i²⁺ levels (Fig. 2G, Table 1), as did a ryanodine isomer mixture and a distinct RyR antagonist, dantrolene(Table 1). Ca_i²⁺ levels were not as low as those obtained with Bapta-AM, probably because the CICR component of RyR activity was still present. Accordingly, the ryanodine-EGTA combination was more potent than was each separately in reducing Ca_i²⁺(Table 1), implicating CICR as the RyR stimulus. Ryanodine/EGTA also abolished the Ca_i²⁺ enrichment in the HH3++/4 node(Table 2), suggesting that CICR/RyR originated this signal. Similarly, left-side ryanodine treatment had little effect on heart looping (12.5%), whereas the ryanodine/EGTA combination significantly reversed heart looping (36%; Fig. 3, Fig. 4E). Left-side ryanodine/EGTA treatment at HH3++/4 repressed Nodal expression in five out of nine embryos (Fig. 5F).

An important Ca_i²⁺ effector is the Ca_i²⁺ sensor protein calmodulin, which interacts with and activates numerous signaling proteins, including RyRs(Berridge et al., 2000; Zalk et al., 2007). Treatment with the calmodulin inhibitor calmidizolium significantly reduced Ca_i²⁺ levels at both HH3+/4(Table 2) and HH6(Table 1), actions that are consistent with the known ability of calmodulin to directly regulate RyR activity (Zalk et al., 2007). Left but not right calmidizolium treatment also caused a significant incidence of cardiac situs inversus (36%; Fig. 3, Fig. 4D). Left calmidizolium treatment abolished Nodal expression in 59% of embryos(10/17, Fig. 5G).

Avians, like mammals, have three RyR isoforms with differing expression and regulatory control. Their contributions to early development are unknown. Immunostaining against all RyR isoforms at HH4 revealed low but discernable levels along the anterior node margin (Fig. 6B,D), overlapping the Ca_i²⁺ Fura-2 signal. Expression expanded along the primitive streak at HH5(Fig. 6E) and the neural plate at HH6 (Fig. 6F). Its expression was symmetric, indicating that its asymmetric activity was regulated post-translationally, a mechanism that commonly governs RyR activity(Zucchi and Ronca-Testoni,1997). Thus, RyRs are present at the right time and location to mediate the Ca_i²⁺ enrichment.

The relationship between node Ca_i²⁺ asymmetry and other laterality effectors is unknown. The initial symmetry of the Ca_i²⁺ signal, followed by its left-side enrichment,suggested the existence of asymmetric regulators. The laterality effector serotonin affects both Nodal expression and heart looping in chick(Fukumoto et al., 2005a; Fukumoto et al., 2005b). Treatment at HH3++/4 with the serotonin re-uptake inhibitor fluoxetine, which prolongs serotonin signaling, repressed Ca_i²⁺ at the H3++/4 anterior node (Table 2). At HH6, fluoxetine ablated the LR Ca_i²⁺ differential(1.38±0.13; Table 3) by reducing left-side Ca_i²⁺ levels(Fig. 7B; Table 3), suggesting that serotonin might normally suppress right-side Ca_i²⁺. Consistent with this, left but not right fluoxetine treatment significantly randomized heart looping (36%; Fig. 8).

Serotonin affects laterality through its 5-HT₃ and 5-HT₄ receptors (Fukumoto et al., 2005b). 5-HT₃ (2-methyl-5-HT) and 5-HT₄(ML10302) agonists rapidly reduced left-side Ca_i²⁺levels and ablated Ca_i²⁺ asymmetry(Fig. 7C,E; Table 3), whereas their respective antagonists (ondansetron, GR125487) elevated right Ca_i²⁺ and sustained left Ca_i²⁺levels (Fig. 7D,F; Table 3), indicating that serotonin acted to keep right Ca_i²⁺ low. Similarly,right but not left application of 5-HT₃ and 5-HT₄antagonists significantly randomized heart-looping direction (ondansetron,32%; GR125487, 30%), whereas their agonists reversed heart-loop direction only when applied to the left (2-methyl-5HT, 31%; ML10302, 32%; Fig. 8). These findings endorse the model of Fukumoto et al. (Fukumoto et al., 2005a; Fukumoto et al.,2005b) that serotonin operates on the right side of the chick to affect LR identity. The repression of right-sided Ca_i²⁺by serotonin is a novel mechanism for this laterality effector.

Also implicated in avian laterality are the H⁺-V-ATPase(Adams et al., 2006) and H⁺K⁺-ATPase proton pumps(Levin et al., 2002; Raya et al., 2004); their inhibition randomizes heart laterality and alters asymmetric gene expression. The H⁺-V-ATPase antagonist concanamycin prevented the node Ca_i²⁺ enrichment at HH3++/4(Table 2) and at HH6(Table 3; Fig. 9B). Left but not right concanamycin treatment also increased the incidence of cardiac situs inversus(58%; Fig. 9D). Thus,H⁺V-ATPase may affect laterality through its ability to initiate and sustain node Ca_i²⁺ asymmetry.

H⁺K⁺-ATPase affects avian laterality by preventing a transient depolarization to the right of Hensen's node(Levin et al., 2002); its inhibition flattened the LR asymmetry of extracellular calcium(Raya et al., 2004). It affected Ca_i²⁺ in a complex manner. At HH4, the H⁺/K⁺-ATPase antagonist lansoprazole significantly reduced the node Ca_i²⁺ enrichment(Table 2). However, at HH6, its action opposed H⁺V-ATPase, and it abolished Ca_i²⁺ asymmetry (1.03±0.19; Table 3) by elevating right Ca_i²⁺ (Fig. 9C), an action consistent with suggestions that it acts upon the right-side of the embryo (Levin et al.,2002). Interestingly, right-side inhibition of H⁺K⁺-ATPase did not affect heart looping(Fig. 9D), a finding that is at odds with its ability to elevate right Ca_i²⁺. Previous laterality studies of H⁺K⁺-ATPase used bilateral inhibitor treatment (Levin et al.,2002; Raya et al.,2004). These complex actions of lansoprazole may indicate shifting roles for the ATPase in establishing asymmetry. Nonetheless, both H⁺V-ATPase and H⁺K⁺-ATPase contribute to laterality through their regulation of node Ca_i²⁺.

Vitamin A-deficient (VAD) quail embryos have a marked tendency (72%) to form left-sided hearts (Zile et al.,2000); the expression of laterality genes, such as Nodaland Pitx2, is also altered (Zile et al., 2000). Vitamin A-sufficient (VAS) HH6 quail had a modest LR Ca_i²⁺ enrichment compared with Gallus galluschicks of equivalent stage (LR ratio 1.44±0.40, Fig. 10A), although the mean Fura-2 signal of the quails did not differ appreciably from that of the chick(right, 168.4±13.7; left, 187.0±14.9; L versus R, P=0.014). Interestingly, HH6 VAD quail had an inverted Ca_i²⁺ asymmetry, and, in a majority (5/7),Ca_i²⁺ levels were significantly higher on the right of the node versus the left (Fig. 10B): the mean LR Ca_i²⁺ ratio for those five embryos was 0.60±0.21. This was not due to increased right Ca_i²⁺ but to significantly reduced left Ca_i²⁺ levels (right, 164.1±9.9; left,161.6±17.0; n=7; left Fura-2 of VAD versus VAS, P=0.016). Although unexpected, this inverted LR Ca_i²⁺ ratio in the majority of VAD quail embryos is consistent with their high incidence of reversed cardiac laterality(Zile et al., 2000).

Here, we show that, like mouse, the chick embryo uses node Ca_i²⁺ asymmetry to mediate laterality. Ca_i²⁺ is first enriched along the anterior margin of the node at HH3++/4, and overlaps with RyR expression. Its induction and left-side enrichment requires RyR and extracellular calcium, implicating CICR as a novel origin for this signal. H⁺K⁺-ATPase and H⁺V-ATPase activity are also contributory, although by HH6 their activities are distinct and opposing. 5-HT₃ and 5-HT₄keep right-side Ca_i²⁺ levels low, representing a novel mechanism for the laterality effects of serotonin. Although the avian node structure may preclude nodal flow, its use of Ca_i²⁺asymmetry supports the hypothesis that aspects of this laterality mechanism are conserved among amniotes.

Figure 11 summarizes our findings and details underlying this model are discussed below. At HH2 and HH3(not shown), the elongating primitive streak already expresses much of the machinery that governs CICR/RyR activity, including serotonin and its receptors and transporters (Fukumoto et al., 2005a; Fukumoto et al.,2005b), H⁺V-ATPase(Adams et al., 2006) and H⁺K⁺-ATPase (Levin et al., 2002); the surrounding blastoderm expresses Cx43(Levin and Mercola, 1999). At HH3+, the first asymmetric gene, cAct-RIIa, appears(Levin et al., 1995). At this same time, Ca_i²⁺ enrichment emerges along the anterior margin of Hensen's node commensurate with RyR expression; its ablation by EGTA/ryanodine and calmidizolium suggests that it originates from CIRC/RyR. Ca_i²⁺ emergence coincides with a transient depolarization of cells along the left side of the node(Levin et al., 2002), and could enhance CICR/RyR activity. Left-sided depolarization, the instigator of which is unknown, may thus be a crucial early step to create asymmetry. By HH4+/HH5, Ca_i²⁺ levels are asymmetric and enriched along the left side of the node. The previously symmetric Shh expression also becomes restricted to the left side of the node(Levin et al., 1995), and,importantly, the sodium-calcium transporter NCX1 is enriched along the right node (Linask et al., 2001). This right-side restricted NCX1, which uses Na⁺ influx to drive Ca_i²⁺ export, may steepen further the LR Ca_i²⁺ asymmetry. By neurulation (HH6), the Ca_i²⁺ asymmetry has increased and extends posteriorly. H⁺V-ATPase sustains the left-side Ca_i²⁺elevation, whereas H⁺K⁺-ATPase keeps right Ca_i²⁺ levels low. Although the mechanism underlying Ca_i²⁺ repression by serotonin is unclear, the physiological properties of the right side of the node suggest that 5HT₃ may function as a Na⁺K⁺ exchanger,perhaps in coordination with NCX1 and/or H⁺K⁺-ATPase, to maintain low Ca_i²⁺. Gap junctions in the blastoderm(Levin and Mercola, 1999)could serve to coordinate and stabilize the Ca_i²⁺ and ATPase signals across the anterior node. Bifurcation of the anterior Ca_i²⁺ field by the prechordal plate, along with the proposed barrier activity of the midline(Danos and Yost, 1996), could then allow LR Ca_i²⁺ levels to be regulated autonomously.

Although Ca_i²⁺ asymmetry across the node is necessary for laterality in mouse (McGrath et al.,2003) and now chick, its origins and relationship to other laterality mediators has been unclear. The identification of CICR/RyR as the origin of Ca_i²⁺ enrichment and asymmetry at Hensen's node is a novel function for RyRs in the early embryo. The requirement for CICR/RyR offers a second mechanism for left-side enriched extracellular calcium to govern avian LR identity, in addition to its regulation of Notch signals (Raya et al., 2004);any interaction between Ca_i²⁺ and Notch remains to be determined. Because RyR proteins are expressed symmetrically along the anterior node and the primitive streak, their activity is likely to be regulated post-translationally. One important mediator is membrane potential and pH, through direct effects upon the protein and indirectly by controlling the activity of calcium exchangers and channels that mediate calcium influx. The complex abilities of the proton ATPases to regulate Ca_i²⁺ levels may reflect a possible involvement in Ca²⁺ release from CICR/RyR-gated stores and this is discussed in greater detail below.

In mouse, movement by monocilia within the ventral node is proposed to generate a fluid current that creates the Ca_i²⁺asymmetry across the node; mutation of left-right dynein heavy chain(Lrd; Dnahc11 - Mouse Genome Informatics) leads to immotile cilia, the collapse of Ca_i²⁺ asymmetry and the randomization of lateral identity (McGrath et al., 2003). Although the avian node architecture is quite different from that of mouse and it may lack a `ventral node'(Manner, 2001), HH4-chick node possesses monocilia that project ventrally from the epiblast and towards the ventral endoderm (Essner et al.,2002). Unfortunately, it was not possible to determine which cell layer at the node was Ca_i²⁺ enriched. However, the existence of Ca_i²⁺ enrichment at the avian node shows that at least some aspects of the mechanisms used to establish LR laterality are conserved among amniotes. In this light, it may be worth noting that polycystin-2, a calcium channel linked to LR identity in mouse and zebrafish(Pennekamp et al., 2002; Bisgrove et al., 2005; Obara et al., 2006; Schottenfeld et al., 2007) can initiate CIRC/RyR-mediated calcium release(Nauli et al., 2003) in a protonation-dependent manner(Gonzalez-Perrett et al.,2002). The contribution of polycystins to avian lateral identity remains unexplored.

How serotonin and the proton ion channels affect lateral identity is unclear. Our data suggest that serotonin operates specifically on the right side of the node to reduce Ca_i²⁺ and maintain asymmetry;5-HT₃ and 5-HT₄ inhibition rapidly elevated right-side Ca_i²⁺ and randomized heart looping. Because serotonin and its 5-HT₃ and 5-HT₄ receptors are bilaterally expressed at these stages (Fukumoto et al., 2005b), there must exist signals that preclude left-sided serotonin action. These could include the serotonin transporters, two of which facilitate serotonin activity and have right-biased expression(Fukumoto et al., 2005a), and monoamine oxidase, which is expressed along the right node margin(Fukumoto et al., 2005b) and which could create a degradative barrier to prevent LR serotonin communication. Thus, several components of serotonin signaling might be good candidates to reduce right-side calcium levels and steepen the Ca_i²⁺ asymmetry.

How could serotonin rapidly decrease Ca_i²⁺? The 5-HT₄ is a G protein-coupled receptor that signals through Gαs and protein kinase A, and thus could regulate diverse elements of calcium homeostasis, including RyR itself(Zucchi and Roncha-Testoni,1997). The 5-HT₃ is a relatively non-selective,ligand-gated cation channel that acts either as a Na⁺K⁺exchanger or to stimulate inward calcium movement(Derkach et al., 1989; Hargreaves et al., 1994); that the 5-HT₃ agonist decreased Ca_i²⁺ implicates the former mechanism. Insights may come from the cation ATPase contributions to lateral identity (Levin et al.,2002; Adams et al.,2006; Raya et al.,2004; Ellertsdottir et al.,2006; Shu et al.,2007). In zebrafish, NCX4a and Na,K-ATPase α2 affect laterality by controlling blastomere Ca_i²⁺ levels(Shu et al., 2007); morphants had elevated Ca_i²⁺, a situation that parallels the avian loss of 5-HT₃ and H+K+-ATPase. Because NCX4a and its related transporters are unique to fish, it was unclear whether similar mechanisms governed calcium asymmetry in other vertebrates. Our data suggest that the H⁺K⁺-ATPase and Na⁺K⁺ exchange activity of 5-HT₃ could have analogous roles in chick to keep right Ca_i²⁺ low.

The H⁺K⁺-ATPase prevents cell depolarization on the right side of the chick embryo (Levin et al., 2002). As depolarization would favor calcium influx and stimulate CICR/RyR, this may be a key H⁺K⁺-ATPase role following midline establishment. That H⁺K⁺-ATPase inhibitors reduce external calcium (Raya et al., 2004) and increase Ca_i²⁺ is consistent with this mechanism. Conversely, H⁺V-ATPase activity(Adams et al., 2006) was essential to induce and sustain the left-side Ca_i²⁺elevation. How it regulates Ca_i²⁺ is unknown. Protonation is a potent inhibitor of both RyR(Zucchi and Roncha-Testoni,1997) and membrane calcium channels, such as polycystin-2(Gonzalez-Perrett et al.,2002), and thus proton export could control CICR/RyR activity directly, or indirectly via voltage-gated calcium channels. Vacuolar ATPases also govern vesicular trafficking and, thus, could regulate membrane calcium channel levels (Jarvis and Zamponi,2007). Taken together, our findings affirm the importance of proton ATPases to LR asymmetry through their effects on Ca_i²⁺. Their disparate effects upon Ca_i²⁺ induction and maintenance indicate that they serve multiple roles. It is unknown whether their activity might be coupled with that of the Na⁺K⁺-ATPases, which, in zebrafish, govern laterality in a calcium-dependent manner(Ellertsdottir et al., 2006; Shu et al., 2007). Additional studies should clarify these issues.

Retinoids mediate LR identity: retinoid receptor antagonists randomize laterality (Tsukui et al.,1999), whereas VAD quail frequently have left-sided hearts(∼70%) (Zile et al.,2000). Retinoids also affect the bilateral symmetry of the somites(Kawakami et al., 2005; Sirbu and Duester, 2006), and act in parallel with Shh and upstream of Nodal, Lefty1 and Pitx2 to affect left identity(Zile et al., 2000; Tsukui et al., 1999). However,their precise contributions remain elusive. We found that a majority of VAD quail had reversed LR Ca_i²⁺ asymmetry, an outcome consistent with their high frequency of reversed heart laterality. Right Ca_i²⁺ levels were not enriched; rather, left Ca_i²⁺ enrichment was absent, which suggests that the latter process requires retinoids. An obvious explanation is a transcriptional role for retinoids in the gene(s) that governs Ca_i²⁺elevation, such as the proton ATPases or a serotonergic repressor. Alternatively, retinoids may sustain the midline boundary that enables left and right to function as separate compartments; for example, through induction of the Nodal inhibitor Lefty(Tsukui et al., 1999; Tabin, 2006). RA also regulates Cxn43 expression(Clairmont and Sies, 1997), and thus could affect gap junctions that might transmit and amplify a laterality signal, such as Ca_i²⁺ or an ion balance(Levin and Mercola, 1999). These mechanisms could be explored in future studies.

Supported by NIH MERIT Award R37 AA11085 to S.M.S. M.H.Z. is supported by NIH award R01 HL61982, USDA NRI 2005-35200-15257 and the Michigan Agricultural Experiment Station. We thank HyLine International for the generous donation of our chickens and John Fallon for helpful discussions.