Itch Is Required for Lateral Line Development in Zebrafish

The zebrafish posterior lateral line is formed during early development by the deposition of neuromasts from a migrating primordium. The molecular mechanisms regulating the regional organization and migration of the primordium involve interactions between Fgf and Wnt/-catenin signaling and the establishment of specific cxcr4b and cxcr7b cytokine receptor expression domains. Itch has been identified as a regulator in several different signaling pathways, including Wnt and Cxcr4 signaling. We identified two homologous itch genes in zebrafish, itcha and itchb, with generalized expression patterns. By reducing itchb expression in particular upon morpholino knockdown, we demonstrated the importance of Itch in regulating lateral line development by perturbing the patterns of cxcr4b and cxcr7b expression. Itch knockdown results in a failure to down-regulate Wnt signaling and overexpression of cxcr4b in the primordium, slowing migration of the posterior lateral line primordium and resulting in abnormal development of the lateral line.


Introduction
The zebrafish lateral line is an insightful system for studies of cellular development as it displays evolutionarily-conserved developmental mechanisms ranging from progenitor migration, neural differentiation and planar cell polarity to sensory transduction [1]. Furthermore, the lateral line is thought to have evolved into sensory structures of the cochlea and inner ear in drier vertebrates, making it an organ of broad interest for developmental neurobiology [2,3]. The posterior lateral line (pLL) is a mechanosensory organ running along the body and tail of fish and amphibians. It is built during early development through coordinated cell migration, proliferation, epithelial morphogenesis and differentiation of a group of about one hundred cells forming the pLL primordium. The pLL primordium arises from placodal cells that undergo partial epithelial-mesenchymal transition and acquire migratory properties. As the primordium migrates towards the tail along the myoseptum, cells in the trailing zone of the primordium become organized into rosette-like epithelial structures that mature into proneuromasts, which are reiteratively formed and deposited every 3-4 hours. These cells differentiate as the accessory and hair cells of the 6-7 mature neuromasts of the primary pLL. When the primordium reaches the end of the tail, it fragments into a few terminal neuromasts [4][5][6]. Thus, the timing of neuromast deposition and the underlying molecular mechanisms of its regulation are critical for the development of this organ.
The migration of the primordium and the formation of the neuromasts is coordinated by Wnt and Fgf signaling. Through a feedback mechanism, Wnt/b-catenin signaling is restricted to the leading zone of the primordium and Fgf signaling occurs in the trailing zone [7,8]. These localized activities maintain the polarized activation of two chemokine receptors: cxcr4b is expressed in the leading zone while cxcr7b is restricted to the trailing zone. The differential expression of cxcr4b and cxcr7b is essential for directed collective migration of the primordium cells [1,9]. However, the mechanisms downstream of these receptors that convey their actions are unclear.
The ubiquitin ligase ITCH has been shown to influence signaling downstream of several important receptors. In particular, ITCH recognizes and down-regulates several SH3-domain proteins, which have been shown to limit epidermal growth factor receptor internalization and signaling [10]. Although no direct link has been established between ITCH and FGF signaling, ITCH targets proteins involved in receptor tyrosine kinase internalization like CBL and SH3GL2 (endophilin) [10][11][12].
ITCH directly interacts with ligand-activated CXCR4 and promotes its ubiquitylation at the plasma membrane [13,14], which is important for the regulation of CXCR4 trafficking and signaling [14,15]. In human cell lines, ITCH depletion significantly attenuates CXCR4-induced ERK-1/2 activation and modestly increases CXCR4 surface levels [16].
ITCH also regulates Wnt signaling through its interaction with Disheveled (Dvl) [17]. Dvl is a central mediator of Wnt signaling where it functions as a scaffold protein bridging the receptors and downstream signaling components [18,19]. In HEK-293 cells, knockdown of ITCH significantly increased Wnt-induced TOPflash activity and the accumulation of free b-catenin induced by Wnt3a. Wnt3a-mediated induction of Wnt target genes AXIN2 and NKD1 was also potentiated, suggesting that ITCH negatively regulates the canonical Wnt pathway [17].
Given this implication of Itch in the two major signaling pathways, Cxcr4 and Wnt, involved in pLL primordium migration, we investigated the effects of Itch depletion in lateral line formation in zebrafish embryos. Our study presents the first direct demonstration of the implication of the ubiquitin-ligase Itch in the regulation of signal transduction in a living organism.

Itch in Danio rerio
To investigate the impact of itch loss of function in early vertebrate development, we used the well-established model that is the zebrafish [20]. We identified the ITCH orthologues in zebrafish by querying the NCBI database Danio rerio sequences with the amino acid sequence of the human Itch HECT domain (NP_113671.3) using the tBlastn algorithm [21]. This search retrieved 11 complete coding sequences with a maximum nucleotide sequence identity of 43-88%. All the sequences identified encoded predicted proteins with the typical domain architecture of the Nedd4 sub-family of ubiquitin ligases, consisting of a C2 domain, two to four WW domains and the catalytic HECT domain (CWH ligases).
We aligned these cds sequences together with the human sequences for all nine members of the CWH sub-family (Table S1), and constructed a phylogenetic tree using the Phylip package SeqDist algorithm (Fig. 1) [22]. All zebrafish sequences clustered with their human counterparts, giving weight to the hypothesis that they represent orthologues. Two genes grouped with human ITCH. We designate these sequences itcha and itchb, corresponding respectively to NCBI gene IDs 100331274 on chromosome 6 and 100330031 on chromosome 23. Interestingly, the itchb sequence is absent from the ZFIN database but was however confirmed in the latest zebrafish genome sequencing project [23]. Except for itch and nedl2, only one copy of each of the other CWHs was found in the zebrafish genome.
We next examined the expression pattern of itcha and itchb in developing embryos. We preformed reverse-transcription of RNA extracted from embryos at 6, 24, and 48 hours post-fertilization (hpf) followed by PCR amplification with primers designed to specifically amplify itcha or itchb. Both genes were successfully amplified from 6 hpf embryos and their expression was maintained at 48 hpf (Fig. 2 A) and later on (not shown). In situ hybridization experiments were conducted on embryos at 6 and 24 hpf. A generalized pattern of expression was discerned in these experiments, suggesting that both genes are expressed in every tissue ( Fig. 2 B-D). This is consistent with reports ITCH gene in mammals where RNA is detected in all analyzed tissues [24]. Faint staining could be observed in the pLL primordium with both itcha and itchb specific probes, distributed throughout the structure, whereas no staining appeared when sense probes were used (Fig. 2 E-G).
Knockdown of itcha or itchb affects embryo's growth We next examined the impact of loss of expression of itch genes in early development. We designed antisense splice junction morpholino oligonucleotides (MOs) targeting itcha and itchb (Table S2). We first assessed the efficacy of the MOs by performing RT-PCR on RNA extracted from control or injected embryos using oligonucleotide primers flanking the target exons of each MO. There was a marked and specific reduction of itcha mRNA in embryos injected with MOs against itcha but not itchb, and likewise for itchb with the itchb but not the itcha MO, with amplification of the lower, exon-skipped band readily visible in the middle panel (Fig. 3 A). Both MOs induced considerable cell death, visualized by acridine orange staining [25] (Fig. S1). Injection of morpholino has been shown to induce cell death by activating the proapoptotic protein p53, an unspecific side effect efficiently counteracted by downregulation of p53 expression [26]. Therefore, p53 MO [27] was consistently co-injected to reduce the general toxicity effect of MOs.
We then assessed the morphology of injected embryos. Knockdown of either gene caused a general growth delay, the effect of itchb knockdown being generally more severe (Fig. 3 B-I and K-N). itchb knockdown embryos were about two hours late in their development at 24 hpf, judging from the number of somites and pLL primordium migration. They remained smaller throughout the first five days of development, the discrepancy in size tending to increase with time (320:21+2:71mm at 72 hpf compared to 380:15+2:62mm; pv0:01) (Fig. 3 J,O). Other  Table S1. doi:10.1371/journal.pone.0111799.g001 Itch in Zebrafish Lateral Line PLOS ONE | www.plosone.org general features included small head, deformed hindbrain, small eyes, immature fin buds at 72 hpf, pericardium oedema and low survival after 96 hpf. They were able to swim but were generally not responsive to touch and tended to rest on the bottom of the dish when undisturbed (not shown). In comparison, itcha knockdown was less severe, but knockdown embryos were still significantly smaller than control siblings (344:06+1:36mm; pv0:01) (Fig. 3 L-O). The coinjection of both MOs resulted in slightly smaller embryos (318:57+3:21mm) and high mortality (Fig. 3 N,O).
In an attempt to rescue and assess the specificity of the more severe itchb phenotype, we used in vitro-transcribed human ITCH mRNA (that is not targeted by the itchb MOs), injected alone as a control, or in itchb knockdown embryos. Co-injection of ITCH mRNA partially rescued growth defects, with embryos reaching a significantly larger size than upon itchb knockdown (342:02+2:13mm), and was comparable to itcha knockdown (Fig. 3 P).

itchb is involved in pLL development
In mammalian cells, Itch has been identified as an important regulator of CXCR4 signaling [14,16]. This cytokine receptor is well described in Zebrafish for its role in pLL primordium migration [28]. We therefore examined the consequences of itch knockdown on pLL development. We used DiAsp, a vital marker of neuromast hair cells to label the pLL of WT embryos [29]. We also used cldnb:gfp embryos that express GFP in the developing pLL primordium as well as the neuromasts [30]. Both methods yielded identical results and showed that the pLL primordium did not reach the tail tip upon itchb knockdown, usually stalling around the end of the digestive track at 48-52 hpf (Fig. 4 D). In itcha knockdown, the primordium migration was slightly delayed at 48 hpf( Fig. 4 C), but had reached the tail by 52 hpf (not shown). The number of neuromasts formed at 72 hpf was drastically reduced in itchb knockdown, but not in itcha knockdown (Fig. 4 C, E and N).
The primordium maintained its overall morphology, and the characteristic rosette-like organization of the neuromast progenitors was readily identifiable under all conditions, but itchb morphants were often more disorganized and seemed to collapse in the trailing end ( Fig. 4 J-M). They also often exhibited less cohesion in the border cells, as can be seen in Movie S3.
To determine if primordium migration was affected by ablation of itchb, we performed time-lapse microscopy experiments in cldnb:gfp embryos. We recorded primordium migration starting just before deposition of the first trunk neuromast in control, morphant or rescued embryos (   Movies S1-S2). Moreover, cells from the primordium leading edge in morphant embryos were disorganized and moved around and back instead of straight toward the tail as in control primordia. This cell behavior suggests that the slowing of pLL primordium migration is not merely the consequence of general growth failure, but due to genuine interference with directional cell migration. The primordium migration defect was rescued by human ITCH mRNA injection ( itchb knockdown perturbs signaling events required for pLL primordium migration Migration of the primordium along the myoseptum is directed by the cytokine receptor Cxcr4b, expressed in the leading edge of the primordium. Directionality is ensured by expression of another cytokine receptor, Cxcr7b, in the trailing end of the migrating primordium, whose expression is restricted by Cxcr4b and acts as a sink for the ligand Sdf-1, thereby preventing Cxcr4b signaling in the trailing end and allowing directionality of migration [31]. Given the slow migration of the pLL primordium in itchb morphants, we examined cxcr4b and cxcr7b expression in the primordium. We visually assessed in cldnb:gfp embryos that the primordium had reached somite 10 before fixation and in situ hybridization. This occurred at about 28 hpf for control embryos and 30 hpf for itchb morphants. Comparing cxcr4b expression in these stage-matched embryos, while cxcr4b expression was restricted to the leading two-third of the primordium in control embryos (Fig. 5 A), it was clear that cxcr4b was overexpressed in itchb morphants and that its distribution encompassed the entire primordium, extending to the deposited cells behind the migrating primordium (

Wnt signaling is increased in itchb morphants
Wnt signaling also plays an important role in pLL primordium migration and is thought to act upstream of cxcr4 [7]. Wei et al. [17] identified Disheveled, an important intermediate in the Wnt signaling pathway as a target of Itch in mammalian cells. To determine if Wnt signaling was modified by itchb knockdown, we examined the expression of the Wnt target gene lef1, whose expression is normally limited to the leading cells of the migrating primordium [7], as seen in control embryos (Fig. 5 E). In itchb morphants lef1 expression appeared to be increased, though it was still restricted to the leading two thirds of the structure (Fig. 5 F).
To get a sense of the general level of lef1 expression in itchb morphants, we performed a semi-quantitative RT-PCR experiment with RNA extracted from stage-matched control (WT) or itchb morphant (MoB) embryos. After reverse transcription, we performed serial dilutions of the cDNA reactions and proceeded to PCR amplification with specific primers for lef1 and actin (Table  S3). Amplification of actin was comparable between the two sets of cDNAs ( Fig. 5 G), showing comparable extraction and reversetranscription efficacy. In contrast, lef1 was amplified from much lower amounts of cDNAs from itchb morphants than from control embryos, indicating higher expression of lef1 in this group (Fig. 5 G, H). These data are consistent with general Wnt signaling activation in itchb morphants.

Discussion
itchb is required to maintain primordium migration This study aimed at exploring how early vertebrate development was affected by knockdown of the ubiquitin ligase itch gene using the vertebrate model organism Danio rerio. We found two genes coding for proteins with similar sequence identity with human ITCH, itcha and itchb. RT-PCR experiments demonstrate that both genes were expressed in zebrafish larvae, and a generalized pattern of expression was observed using in situ hybridization. Duplicated genes are common in zebrafish and in these cases paralogues have been found to present partial genetic redundancy, although they also often diverge in either their function or their expression pattern [32]. Though knockdown of either or both genes affected overall growth, itcha and itchb indeed appeared to differ in their function, the most noticeable distinction being in the effect of the knockdown of itchb but not itcha on perturbation of pLL development.
The process governing pLL primordium migration is well described and involves Wnt, Fgf and cytokine signaling. The principal determinant of pLL primordium migration is the establishment of distinct expression domains of the cytokine receptors cxcr4b and cxcr7b. As the primordium migrates along the myoseptum, cxcr4b is restricted to the leading half of the cycles. Marked reduction of itcha and itchb was obtained after injection of the appropriate but not the other MO. (B-I) Morphology of embryos injected with MOs targeting itchb (MoB) showed general growth failure, already visible at 24 hpf (B,C) and persisting through to 96 hpf (H,I) as compared to control embryos (CTRL). Scale bar: 500mm. (J) Total length of control embryos or embryos injected with MOs against itchb (MoB) was measured from the tip of the head to the tip of the tail muscle. Two-factor analysis of variance indicates no significant interaction of injection with age. Pairwise comparison was performed between control and MoB injected embryos at each time point. The number of embryos measured is indicated at the bottom of the graph. ÃÃ represents statistical significance at pv0:01. (K-N) Morphology of control embryos (CTRL), itcha knockdown embryos (MoA), itchb knockdown (MoB) or itcha and itchb knockdown embryos (MoAzB) at 48 hpf. itcha knockdown embryos were slightly delayed, but were morphologically intact as compared to itchb or itcha plus itchb knockdown embryos. (O) Global growth of injected embryos was assessed by total length measurements at 72 hpf. itcha and itchb knockdown embryos were significantly smaller than control embryos, and itchb knockdown were smaller than itcha as assessed by Kruskal-Wallis one-way ANOVA combined with Dunn's method of group comparison with a significant threshold fixed at pv0:01 (ÃÃ). itcha plus itchb knockdown embryos were not significantly smaller than itchb knockdown embryos (n.s.). (P) Growth defect of itchb knockdown was partially rescued with injection of in vitro transcribed human ITCH mRNA. Kruskal-Wallis one-way ANOVA combined with Dunn's method of comparison established no-significant differences between control embryos (CTRL) and embryos injected with ITCH mRNA alone (mRNA), whereas the total length was significantly reduced in both itchb MO-injected embryos (MoB) and embryos injected with a combination of itchb MO and ITCH mRNA (Rescue). Total length was significantly larger in the Rescue group than in the MoB group, indicating partial rescue. The number of embryos measured is indicated at the bottom of the graphs. Each graph summarizes at least four different experiments. ÃÃ indicates pv0:01.  primordium and inhibits cxcr7b, which consequently is present only in the trailing end [1,7,33]. This asymmetric distribution directs primordium migration along the myoseptum in response to Sdf1 secretion and disturbing this equilibrium results in slowed primordium migration speed or stalling of the primordium [1,7,[33][34][35]. Wnt signaling occurs mainly in the leading region of the pLL primordium and activates Fgf signaling in the medial and trailing region [7]. Fgf signaling organizes the primordium precursor cells in rosettes that will become the neuromasts and restricts Wnt signaling to the leading cells [36,37].
The expression of cxcr4b is regulated at the transcriptional level by Wnt and oestrogen signaling, as response elements for Lef1 and Esr1 are present in the upstream control region of the cxcr4b gene [34,35]. lef1 morphants and mutants demonstrate truncated pLL, as the migrating primordium collapses before it reaches the end of the tail caused by decreased cell proliferation and lack of progenitors [4,35,38]. Lef1 depletion alone has no effect on cxcr4b or cxcr7b expression, and pLL primordium migration and differentiation appears to be normal. However, increasing Wnt signaling, as occurs in apc mutants, strongly inhibits primordium migration while increasing cxcr4b expression domain and excluding cxcr7b from the primordium [7].
itchb morphants similarly exhibited displacement of the cxcr4b and cxcr7b expression domains, consistent with increased cxcr4b and Wnt signaling. In cultured mammalian cell lines, Itch has been shown to increase both Wnt and Cxcr4 signaling. Itch can regulate the Wnt signaling pathway by recognizing and ubiquitylating phosphorylated Dvl [17]. Dvl is recruited to the activated Wnt receptor complex and activates both the canonical and noncanonical signaling pathways [19]. Upon activation by Wnt, Dvl become hyperphosphorylated, and this phosphorylation is essential to fully activate b-catenin stabilization [39,40]. Inactivation of Itch stabilizes phosphorylated Dvl, increasing Wnt signaling [17]. In zebrafish, Dvl degradation has been shown to be implicated in Wnt signaling regulation [41]. On the other hand, Dvl expression has also been shown to be stable in zebrafish embryos during primordium migration [42]. It must be stressed that Itch specifically targets phosphorylated Dvl and promotes its proteasomal degradation [17]. Consequently, Itch depletion could increase Wnt signaling and expression of Wnt signaling pathway target genes, without affecting the overall level of Dvl protein [17].
CXCR4 is a direct target of ITCH ubiquitin ligase activity in mammalian cells [14,15]. ITCH is known to regulate CXCR4 internalization and signaling in conjunction with b-arestin and STAM [16,43]. CXCR4 ubiquitylation is required for its rapid ligand-induced lysosomal degradation [15]. ITCH interacts directly with CXCR4 and HRS to direct CXCR4 degradation [14,44]. Moreover, ITCH is necessary for CXCR4-mediated activation of the ERK/MAPK pathway [16]. It has not been possible to directly measure the effect of Itch depletion on Cxcr4b protein, as antibodies against CXCR4 do not discriminate Cxcr4a and Cxcr4b in Western blot analysis and did not yield a reliable signal in immunofluorescence. Nevertheless, cxcr4b is clearly overexpressed in the primordium at the mRNA level, which could be a direct effect of its increased signaling. It is known that Sdf1/ Cxcr4 signaling exerts a positive feedback on cxcr4b expression in the primordium [1,34]. Therefore, increased cxcr4b expression is consistent with Cxcr4b protein stabilization. Cxcr4b also exerts transcriptional control over cxcr7b expression, effectively excluding Cxcr7b form the Cxcr4b expression domain [1]. This is consistent with the reduced cxcr7b signal in itchb morphants.
Defects in pLL primordium migration in itchb morphants can thus be attributed to increased Wnt and Cxcr4b signaling in vivo.
Since the phenotypic manifestation of perturbed Cxcr4b signaling and Wnt overactivation are confounded, it is not possible to discriminate which pathway is most affected by itchb depletion, though most likely itchb affects both.

Neuromast deposition is altered in zebrafish depleted of itchb
Although neuromast deposition is clearly a consequence of and influenced by pLL primordium migration, manipulations specifically altering neuromast deposition have no effect on pLL primordium migration [45]. The processes are thus independently regulated by overlapping signals. Examining the pLL of itchb deficient embryos, it was clear that both the number of neuromasts and the rate of primordium migration were reduced by half. Reduced neuromast deposition is observed when the proliferation rate is decreased, for example by treating the embryos with DNA replication inhibitors such as hydroxyurea [45]. Increased cell death also leads to fewer neuromast deposition, as seen in bap28 homozygous mutants and tcf7 ATG morphants. Importantly, the reduction in proneuromast number in tcf7 MO injected embryos is non-specific and not due to loss of tcf7 function, and was completely rescued by the coinjection of p53 MOs [45]. Both itcha and itchb MOs indeed induced significant cell death when injected alone, but this effect was reversed by addition of p53 MO. Moreover, only itchb morphants exhibited defects in neuromast deposition, and this effect was rescued by injection of human Itch mRNA. p53-mediated cell-death therefore is unlikely to be responsible for the neuromast deposition defects. Nevertheless, reduction of Itch has been related to increased cell death or reduced proliferation in a number of cases. First, p63 and p73, two isoforms of p53, are direct targets of Itch, and their stabilization through Itch depletion is known to activate apoptotic pathways [46][47][48]. Although it has not been reported specifically in the primordium, p63-mediated cell death occurs in zebrafish embryos and larvae [49]. In mammalian cells, ITCH depletion increased LATS1, a serine/threonine kinase in the Hippo pathway, enhancing FAS-induced apoptosis and reducing proliferation, survival, and migration [50]. This is consistent with results showing that in HEK-293T cells, ITCH depletion decreased cell survival and enhanced TRAIL-induced cell-death [51]. Increased cell death is thus a plausible explanation for loss of neuromasts in itchb morphants, but its assessment is difficult due to the MOs against itchb and in vitro-transcribed human ITCH mRNA (Rescue). Whereas the number of neuromasts in the pLL was markedly reduced after injection of MOs against itchb, injection of in vitro transcribed ITCH mRNA alone or of MOs against itcha had no effect. The MOs phenotype was rescued by co-injection of ITCH mRNA in itchb knockdown. The number of embryos for each group is indicated at the bottom of the graph. ÃÃ represents statistical significance at pv0:01. n.s. = non-significant. (J-M) Morphology of the migrating primordium in 26 hpf embryos (CTRL), itcha knockdown embryos (MoA) and itchb knockdown embryos (MoB). Cells at the tip of the primordium of MoB embryos were slightly disorganized, but the typical rosette formation occurred normally. (O) Migration of the pLL primordium was quantified by measuring the distance between the primordium leading edge in the first frame (t~0) compared to the last frame (t~60) as shown in F-I (doted line). The number of embryos counted or imaged is indicated at the bottom of the graph. Each graph summarizes four different experiments. ÃÃrepresents statistical significance at pv0:01. n.s. = non-significant. EGFP fluorescence is presented as reversed black and white to facilitate visualization in all micrographs. Scale bars: (A-E) 34mm,  recognized caveat in the use of MOs to examine apoptotic pathways [26].
It must be noted that Itch depletion has also been shown to induce cell proliferation through a number of signaling pathways. In hematopoietic stem cells, Itch deficiency increases proliferation by stabilizing Notch1 signaling [52]. In the zebrafish pLL, Notch signaling determines differentiation of pLL ganglion neurons and, later on, of hair cells [53,54]. Although it was shown that preventing hair cell differentiation leads to fewer neuromasts in the pLL [55], the effect of promoting Notch signaling on this system is unknown. Moreover, since DiAsp was successfully accumulated by those neuromasts that were formed, we conclude that itchb morphant hair cells were functional. In the context of Hedgehog signaling, Itch depletion promoted tumorigenicity, preventing the formation of a degradation complex between Itch, Numb and Gli1. This had the effect of stabilizing Gli1 and promoting medulloblastoma growth [56]. There is thus a multitude of pathways through which Itch could influence both cell-death and cell proliferation in the pLL primordium that could result in the formation of fewer neuromasts. A transcriptomic or proteomic approach could yield knowledge about how Itchb affects neuromast formation.
In the zebrafish pLL primordium, proliferation is regulated mainly by Wnt and Fgf signaling, and seems to be largely independent of cell differentiation [45]. Increasing Wnt signaling alone leads to increased proliferation, but the process is dependent on Fgf signaling [45,57]. In short, Fgf signaling promotes proliferation whereas Wnt signaling limits the proliferation zone to the trailing end of the primordium, where rosette formation occurs, while there is little proliferation in the leading zone that instead directs migration [45]. The domain of cxcr7b expression seems to delimit the zone where the depositing cells reside, although it is not clear how Cxcr7b activity mediates this process [45]. cxcr7b expression is limited by Wnt signaling and lef1 expression that occurs in the leading zone of the primordium [7]. lef1 is a target gene of Wnt signaling and its expression increases in apc mutants [7]. We show here that Itch depletion in zebrafish resulted in increased Wnt signaling, as was shown by higher lef1 expression, consistent with the identification of Dvl as a target of Itch ubiquitylation in mammalian cells [17].
Increased Wnt signaling in zebrafish is associated with slower migration of the primordium, but is not sufficient to induce deposition of fewer neuromasts. Apart from increased apoptosis, the reduction in neuromast number in itchb morphants could occur through altered Fgf signaling [45,58]. No direct impact of Itch on Fgf signaling has been reported so far in the literature. Fgfr1 is a direct target of the closely related Nedd4-1 ubiquitin ligase and the inhibition of this interaction leads to an important increase in Fgf signaling that alters anterior development in zebrafish [59]. Ubiquitylation is important to regulate Fgfr internalization and signaling, both directly and indirectly through the regulation of Sprouty 2 [59,60]. Tyrosine kinase receptor internalization is influenced by ubiquitylation of the endocytic machinery, and targets of Itch have been implicated in this process [10][11][12][61][62][63][64][65]. However, should Itch be directly involved in Fgf regulation, one would expect that itchb depletion would lead to increased Fgf signaling. Increasing Fgf signaling does not lead to reduced proliferation, but instead leads to the formation of supplementary rosettes in the migrating primordium [37]. It does not seem likely then that the Fgf pathway is perturbed independently of Wnt and Cxcr4b in itchb morphants.
The HECT-domain ubiquitin ligase Itch is an important negative regulator of signaling. It has been mainly linked to immunological responses, but is widely expressed and likely involved in many developmental and regulatory signaling events.
In mice, Itch deficiency results in spontaneous development of late onset and progressively lethal systemic autoimmune-like disease, attributable to biased differentiation of CD4 z cells into T H 2 cells and chronic activation [66]. The inflammatory response is also attributable to expansion of the B1b lymphocytes leading to IgM elevation and IgE production [67]. These immunological defects are mainly attributable to accumulation of the transcription factor JunB, in the absence of Itch [66,67].
In human, in addition to multisystem autoimmune diseases akin to the Itchy mice phenotype, patients with ITCH mutations displayed morphologic and developmental abnormalities [68]. Together, the results described above confirm that Itch constitutes an important signaling hub in the cell, maintaining the balance in several important signaling pathways. Different vertebrate models, including the zebrafish introduced here, are likely to unveil the molecular defects underlying these Itch-related pathologies.

Ethics Statement and Transgenic Animals
A colony of wild-type Longfin zebrafish (Danio rerio) was bred and maintained according to standard procedures in our animal facility [69]. The transgenic line Tg(cldnb:lynEGFP) expressing membrane-tethered EGFP (enhanced GFP) under the claudinb promoter was used as it labels the migrating lateral line primordial, the neuromast organs as well as the chain of interneuromast cells deposited during migration [30]. All experiments were performed in compliance with the guidelines of the Canadian Council for Animal Care and approved by the Comité de déontologie de l'expérimentation sur les animaux (CDEA) of the University of Montreal. Embryos were anesthetized in 0.02% tricain (MS-222, Sigma) in Embryo medium prior to all experiments.

Antisense Morpholino Oligonucleotides and RNA Injections
To knockdown the zebrafish itcha and itchb genes, we designed splice-junction blocking MOs specific to the donor and acceptor splice-sites of itcha and itchb exons 12 and 13 (Gene Tools, Philomath, OR). MOs and mRNAs were diluted in nuclease-free water with 0.2% FastGreen vital dye to judge of injection volume. To avoid toxicity effects, p53 MO was coinjected [27]. All MOs are listed in the file (Table S2).

Reverse Transcription-PCR
Total RNA was extracted from pools of approximately 50 embryos, treated as stated, using TRIzol Reagent (Invitrogen, Carlsbad, CA). cDNA was synthesized from 1 mg total RNA using the SuperScript VILO cDNA Synthesis Kit (Invitrogen, Carlsbad, CA). All amplifications were carried on using Phusion High Fidelity DNA polymerase (New England Biolab, Ipswich, MA).
Semi-quantitative PCR reactions were setup using serial dilutions of cDNAs followed by 30 cycles of amplification. All primers are listed in the file (Table S3).

Lateral Line Staining
The lateral line of control or injected embryos was labeled using the vital dye 4-(4-diethylaminostyryl)-N-methylpyridinium iodine (4-di-2-ASP, Invitrogen) diluted to 0.5 mM in embryo medium according to an established procedure [71]. Embryos were dechorionated manually and staged at 72 hpf then incubated in the solution for 30 minutes at 28.5uC. They were then washed three times for 10 minutes in fresh embryo medium and anesthetized in tricain before imaging on an epifluorescence dissection microscope (Olympus) equipped with a Flea2 CCD Camera (IEEE 1394, Point Grey Research Inc. Richmond, BC, Canada). This protocol allows for visualization of full neuromasts that were counted for each fish.

Acridine Orange Staining
Zebrafish were incubated in 1 mg/ml acridine orange for 30 min then repeatedly washed in embryo medium. Larvae were anesthetized in tricain and mounted in low melting point agarose before being visualized under a 10X water immersion lens mounted on a Quorum Technologies spinning disk confocal microscope as bellow.

Confocal Microscopy and Time-Lapse Recordings
Embryos were anesthetized in 0.02% tricain (MS-222) in embryo medium and embedded in 1% low melting point agarose. Imaging was performed on a Quorum Technologies spinning-disk confocal microscope (Quorum WaveX Technology Inc Guelph, On, Canada) mounted on an upright Olympus BX61W1 fluorescence microscope with water-immersion lenses. The setup was fitted with a Hamamatsu ORCA-ER camera and image acquisition was done with the Volocity software (Perkin-Elmer) and analyzed with the ImageJ software (NIH). Stacks were acquired at 1 mm thickness and assembled in ImageJ. When necessary, adjacent frames were aligned and stitched together using Photoshop CS 5.1 (Adobe Systems Incorporated) Auto-Blend Layers function. For time-lapse microscopy, images were captured every three minutes for 60 minutes. To quantify the primordium migration, the distance between the tip of the primordium at t~0 and the tip of the primordium at t~60 minutes was measured in ImageJ after superposition of the images.

Whole-Mount In Situ Hybridization
In situ hybridization was performed using sense and antisense probes designed against the zebrafish orthologs of itch (Table S4) to view endogenous localization of itch mRNA. Embryos of 6 hpf and 24 hpf were processed for in situ hybridization as previously described [72]. To measure the impact of itch down-regulation on signaling in the migrating primordium, probes against cxcr4b, cxcr7b and lef1 (Table S4), were synthesized and used in an identical procedure on MO and control injected cldnb:gfp embryos age-staged at 26 hpf according to primordium migration. Figure S1 Cell death in the pLL primordium of itcha and itchb morphants. (A-C), acridine orange staining in the primordium region of control cldnb:gfp embryos (CTRL), itcha knockdown (MoA), and itchb knockdown (MoB). p53 MO was omitted in this experiment. In these conditions, cell death occurred in the primordium cells, predominantly in the trailing end of the migrating primordium in both itcha (B) and itchb (C) morphants. Acridine orange staining is visible as brighter dots representing the nucleus of apoptotic cells over the dimmer EGFP signal in the cell membrane of the transgenic primordium. Scale bar: 45mm.

(TIF)
Table S1 Accession number of sequences used for alignment. (PDF) Movie S1 CTRL. Example of a pLL primordium migration at 28 hpf in a control embryo. Stacks were acquired at 1mm thickness and assembled in ImageJ. Movie length: 60 min. (AVI) 28 hpf in a human ITCH-injected embryo (a second control). Stacks were acquired at 1mm thickness and assembled in ImageJ. Movie length: 60 min. (AVI) Movie S3 MoB. Example of a pLL primordium failing to migrate at 30 hpf in a itchb-morpholino-injected embryo. Stacks were acquired at 1mm thickness and assembled in ImageJ. Movie length: 60 min. (AVI) Movie S4 Rescue. Example of a pLL primordium migration at 28 hpf in a itchb-morpholino-injected embryo after rescue with human ITCH mRNA. Stacks were acquired at 1mm thickness and assembled in ImageJ. Movie length: 60 min. (AVI)