The Enhancer of split transcription factor Her8a is a novel dimerisation partner for Her3 that controls anterior hindbrain neurogenesis in zebrafish
- 6.6k Downloads
Neurogenesis control and the prevention of premature differentiation in the vertebrate embryo are crucial processes, allowing the formation of late-born cell types and ensuring the correct shape and cytoarchitecture of the brain. Members of the Hairy/Enhancer of Split (Hairy/E(spl)) family of bHLH-Orange transcription factors, such as zebrafish Her3, 5, 9 and 11, are implicated in the local inhibition of neurogenesis to maintain progenitor pools within the early neural plate. To better understand how these factors exert their inhibitory function, we aimed to isolate some of their functional interactors.
We used a yeast two-hybrid screen with Her5 as bait and recovered a novel zebrafish Hairy/E(spl) factor - Her8a. Using phylogenetic and synteny analyses, we demonstrate that her8a evolved from an ancient duplicate of Hes6 that was recently lost in the mammalian lineage. We show that her8a is expressed across the mid- and anterior hindbrain from the start of segmentation. Through knockdown and misexpression experiments, we demonstrate that Her8a is a negative regulator of neurogenesis and plays an essential role in generating progenitor pools within rhombomeres 2 and 4 - a role resembling that of Her3. Her8a co-purifies with Her3, suggesting that Her8a-Her3 heterodimers may be relevant in this domain of the neural plate, where both proteins are co-expressed. Finally, we demonstrate that her8a expression is independent of Notch signaling at the early neural plate stage but that SoxB factors play a role in its expression, linking patterning information to neurogenesis control. Overall, the regulation and function of Her8a differ strikingly from those of its closest relative in other vertebrates - the Hes6-like proteins.
Our results characterize the phylogeny, expression and functional interactions involving a new Her factor, Her8a, and highlight the complex interplay of E(spl) proteins that generates the neurogenesis pattern of the zebrafish early neural plate.
Keywordszebrafish primary neurogenesis midbrain-hindbrain Hairy/E(spl) Her/Hes
Neurogenesis in the early vertebrate neural plate begins at stereotyped loci - termed proneural clusters -, which prefigure the localization of the earliest neuronal groups and the architecture of the primary embryonic neuronal scaffold. These proneural clusters consist of spatially defined progenitor groups engaged in active neurogenesis, within which committed precursors expressing higher levels of proneural genes (such as neurogenin or achaete-scute-like genes, respectively neurog1 and ascl1 in zebrafish) are singled out to differentiate first. An identical scaffold is found in all vertebrate embryos, highlighting the robustness and functional relevance of this organization [1, 2, 3, 4]. Dissecting the regulatory cascades involved in this process is therefore of universal importance.
The control of neurogenesis progression within proneural clusters relies on Hairy/Enhancer of split (E(spl)) factors (Hes in mouse; Her in zebrafish). These transcription factors belong to the basic-helix-loop-helix (bHLH) family, characterized by a DNA-binding basic domain and an HLH domain composed of two alpha helices intervened by a loop of a few amino acids . In addition, Hairy/Enhancer of split (E(spl)) factors contain an Orange domain, which is most probably involved in protein-protein interactions, and a WRPW C-terminal tetrapeptide, which mediates transcriptional repression (reviewed in [6, 7]). During the so-called process of lateral inhibition, the expression of Notch ligands in committed precursors activates Notch signaling in neighbouring progenitors, which in turn induces expression of Hes/Her factors. The latter down-regulate proneural genes, hence maintaining Notch-receiving cells in a progenitor state. Reflecting the intermingled distribution of committed and transiently inhibited progenitors, the proneural and E(spl) genes are expressed in a salt-and-pepper fashion within proneural clusters. E(Spl) factors expressed in proneural clusters in zebrafish include her4.1 [8, 9, 10], hes5/her15, her2 and her12 . In agreement with the lateral inhibition model, her4.1 expression is positively regulated by Notch, and inhibits expression of neurog1 .
Recent work has demonstrated that proneural clusters are delimited negatively, through a process of active neurogenesis suppression taking place in surrounding areas (reviewed in ). These "inhibited" areas, so-called "progenitor pools", are transiently maintained in an refractory state to be recruited in later events of neuronal production, and are organized as tight groups of adjacent cells at stereotyped positions within the neural plate. Major progenitor pools can be found at the presumptive midbrain-hindbrain boundary (MHB) [3, 13] and in longitudinal stripes separating the columns of presumptive moto- and lateral neurons in the hindbrain, or moto-, inter- and sensory neurons in the spinal cord [11, 14, 15]. At the least, the MHB pool is maintained until adulthood in zebrafish, where it participates in the generation of adult-born neurons and oligodendrocytes . Embryonic progenitor pools are characterised by the expression of a specific set of transcription factors, including Zic, BF1/Anf and Rx family members  as well as Hes/Her proteins. In zebrafish, the combinatorial expression of a distinct set of her genes - which to date includes her3, her5, her9 and her11 - characterises all progenitor pools , while in mouse the genes Hes1, Hes3 and Hes5 share sustained expression in adjacent cells of the MHB pool, for example [18, 19, 20]. These her/Hes genes exhibit functional similarities and have been implicated in progenitor pool maintenance: their misexpression inhibits neurogenesis, whereas loss-of-function causes premature expression of proneural genes in at least part of their expression domains [11, 13, 15, 19, 20, 21, 22]. In addition, her3/5/9/11 as well as Hes1 at the mouse MHB all demonstrate an irregular association with Notch: while Hes/her genes in proneural clusters are activated by Notch signaling, the expression of her3/5/9/11 and Hes1 at the MHB is controlled in a Notch-independent manner. The mechanisms accounting for these specific features remain unknown.
The HLH region of Hes/Her factors functions as a dimerisation domain, and the formation of hetero- and homodimers as well as further possible interactions through the Orange domain are key components of the specificity of the actions of these proteins. Heterodimerisation can involve closely related members of the Her/Hes family, or several different transcription factors or transcriptional cofactors . In order to better understand the mechanism of action of Her factors expressed in progenitor pools, and the pathways regulating their activity, we performed a yeast two-hybrid screen using the HLH and Orange domains of Her5 as bait. This led to the recovery of Her8a, a novel Her factor of the Hes6 subfamily expressed in a broad manner at the presumptive midbrain-hindbrain domain of the early zebrafish neural plate. Morpholino-mediated knockdown and misexpression studies establish Her8a as a negative-regulator of neurogenesis playing an essential role in maintaining progenitor pools of rhombomeres (r) 2 and 4. her8a knockdown produces a similar phenotype to that of her3 knockdown, and co-purification demonstrates that Her8a dimerises with Her3. At the MHB however, we show that the predominant activity is exerted by the combination of Her3, 5, 9 and 11. Together, our results identify a new player in progenitor pool formation and highlight the region-specific combinatorial activity of E(spl) factors in this process.
Identification of Her8a as a potential binding partner for Her proteins
To recover binding partners for Her proteins, we used a yeast two-hybrid screen where a 181-amino acid fragment of Her5 (excluding the basic domain and the WRPW motif) was screened against an 18-20 hpf embryo zebrafish library. This screen returned 280 positive clones, from a total of 76.1 million tested interactions. These 280 positive clones represented 75 unique protein-protein interactions. The quality of these interactions was graded using a PBS scoring system - where A is the highest score of confidence, B is very good, C is good, D is low and N/A is used when no score could be assigned (see Materials and Methods). Our screen returned 6 As, 9 Bs, 2 Cs, 49 Ds and 9 N/As (see Additional file 1, Table S1 for a detailed description of all recovered candidates). Gene ontology enrichment analysis of the recovered binding partners revealed an enrichment of proteins involved in protein transport and also heterodimerisation (see Additional file 2, Table S2). Among these candidates, we note the presence of 7 distinct Her factors, in agreement with the postulated capacity for protein heterodimerisation within this family . As further indication of the validity of the assay, Her5 was found to bind with Her11 with a score of B, an interaction that had been shown previously in our laboratory . Her8a, scored A and corresponding to a new E(spl) family member, proved very strongly expressed in the midbrain-hindbrain (MH) area (see below), and we consequently focused on this factor.
Her8a is a new Hes6-like protein, the ortholog of which was lost in the mammalian lineage
her8ais expressed across progenitor pools and proneural clusters in the anterior neural plate
Gain of Her8a function inhibits neurogenesis
Mouse and Xenopus Hes6 proteins are known as positive regulators of differentiation [24, 27]. In this context, the expression of her8a across both pro-neural and non-neurogenic domains was puzzling and prompted us to explore Her8a function.
Her8a is required to maintain the proper neurogenesis pattern in rhombomeres 2 and 4 and acts as binding partner for Her3
To better appreciate the endogenous requirements for Her8a, we next turned to a loss-of-function approach. Embryos at the one-cell stage were injected with morpholinos (MO) directed against the donor splice site of her8a exon 1 (MO1), the acceptor splice site of her8a exon 2 (MO2), or the her8a ATG (MO3), and were analyzed at 3 somites. Reverse transcription PCR was used to reveal strong down-regulation of expression and abnormal splicing of her8a transcripts with both MO1 and MO2, whereas other genes, such as βactin2, remained unaffected (Figure 3G). These observations substantiate that her8aMOs lead to knock-down of her8a expression.
As a final alternative hypothesis, and based on the recovery of Her8a as binding to a Her bHLH domain in yeast cells, we tested whether Her8a and Her3 could act as necessary heterodimerisation partners. As no commercial antibodies are available for Her3 and Her8a, and attempts by our laboratory to have them manufactured failed, we chose a co-purification approach using tagged versions of the full-length zebrafish proteins recombinantly expressed in HEK293T cells. By purifying Strep/Flag-tagged Her3 via its Strep-tag II moiety, Myc-tagged Her8a was successfully co-purified (Figure 4A), demonstrating that both proteins interact with each other. This interaction may be relevant to the maintenance of the progenitor pools within r2 and r4, where both her3 and her8a are strongly expressed, and hereby account for the identical phenotypes of Her3 and Her8a loss of function (summarized in Figure 4B-E).
In the absence of Her3, 5, 9 and 11 activity, endogenous Her8a alone is insufficient to preserve neurog1-free progenitor pools in the midbrain-hindbrain domain
Endogenous her8aexpression in the early neural plate is independent of Notch signaling but requires the expression of SoxB factors
Although many E(spl) transcription factors are downstream effectors of Notch signalling, previous work has shown that zebrafish her genes expressed in progenitor pools, such as her3, 5, 9 and 11 [11, 13, 31] exhibit a non-canonical regulation by Notch: they do not require Notch for their expression, and are insensitive to or transcriptionally inhibited upon ectopic Notch activation. This is in contrast to other family members such as her4.1 that are expressed in neurogenic zones and are activated by Notch signaling .
Her8a is a neurogenesis repressor in the early zebrafish neural plate
Two lines of evidence demonstrate that Her8a can act as a repressor of neurogenesis: firstly, the overexpression of full-length her8a causes a complete loss of neurog1 expression in the early embryo; secondly, we show that morpholino-mediated knockdown of her8a causes ectopic neurog1 expression in rhombomeres 2 and 4. These results are surprising, since the Hes6-like factors studied to date tend to exhibit neurogenesis-promoting activity. When ectopically expressed, Hes6 promotes neurogenesis in the Xenopus embryo , the differentiation of cortical neurons at the expense of astrocytes in the mouse [44, 45], and the differentiation of retinal precursor cells into photoreceptors in mouse retinal explants . These activities at least in part involve functionally antagonizing Hes1, since it was shown that Hes6 alone cannot bind the canonical E(spl) binding site (N box) [24, 45]. Rather, Hes6 dimerises with Hes1 and modifies its DNA binding properties , its capacity for recruiting the co-repressor Groucho or its stability . Interestingly at least some of these properties appear controlled by the loop domain of Hes6, which is five amino acids shorter than that of other E(spl) proteins (see Additional file 3, Figure S1). Indeed, the addition of five amino acid residues into the loop of Hes6 confers Hes1-like repressor activity on the N box, while conversely, the removal of five amino acid residues from the loop of Hes1 completely ceases repression activity and confers Hes6-like activity . We observed that Her8a has an intermediate loop-length compared to Hes6 and Hes1 (Additional file 3, Figure S1). Although the functional significance of this feature remains a matter for investigation, it is possible that it confers specific mechanistic properties to Her8a that distinguish it from Hes6 and bring it closer to the mode of action of Hes1-like proteins. In addition, our phylogenetic and synteny analyses revealed the complex evolution of Hes6-like genes and, contrary to previous belief, that Her8a is not a direct ortholog of Hes6. Rather, Hes6 orthologs comprise zebrafish Her13 and Hes6, both of which have the same loop length as mouse Hes6. We show here that zebrafish her13 is specifically expressed in a pattern coincident with neurogenesis (Additional file 4, Figure S2) reminiscent of Hes6 expression in the developing nervous system of both mouse and Xenopus, highlighting committed progenitors or early neurons [24, 27]. Thus we would predict that zebrafish Her13, rather than Her8a, shares functional properties with mammalian Hes6. The function of Hes6.2 subfamily proteins, to which Her8a belongs, has not been thoroughly tested, largely due to their absence in mammals. In the chicken neural tube however, Hes6.2 exerts a neurogenesis promoting activity , suggesting that Hes6.2 proteins may differ in their activities. We further propose that the splitting of her8a from locus 2 to a distinct genomic location (Figure 1C) permitted the acquisition of a unique expression profile for this gene in zebrafish.
Combined Her activities generate the midbrain-hindbrain progenitor pool pattern through different modalities
Our loss-of-function studies demonstrate that Her8a activity is necessary to prevent neurogenesis within the mediolateral territory of r2 and 4, hence keeping the proneural clusters for moto- and lateral neurons spatially separated within the anterior hindbrain. Combined with the fact that endogenous her8a expression does not depend on Notch at this developmental stage (Figure 6), this function is typical of a "pre-patterning" activity, comparable to that exhibited by the other E(spl) factors Her3, 5, 9 and 11 that delimit the territories competent for neurog1 expression within the neural plate [11, 13, 15]. Despite this functional relevance, the phenotype of her8a morphants appears very restricted compared to the broad expression of her8a, which encompasses the entire mid- and anterior hindbrain (Figure 4B). Functional redundancy and dosage effects have been described for other members of the E(spl) family in the mouse neural tube  and the zebrafish early neural plate [11, 22]. For example, Her5 and Her11 act in an equivalent and dose-dependent manner to block neurog1 expression in the medial and lateral aspects of the presumptive MHB , and Her3 and Her9 also cooperate to inhibit neurogenesis within the longitudinal stripe separating the presumptive moto- and interneuron clusters of the spinal cord . We found that all her genes analyzed here (her3, 5, 9, 11 and 8a) were at least partially co-expressed within the presumptive MH (Figure 5L), strongly suggesting that redundancy may account for the normal development of this area in her8a morphants. The situation in the hindbrain, however, appears different. The only two her genes highlighting progenitor pools in r2-4 are her3 and her8a. While morphant embryos for each of these genes have an identical phenotype, our results argue against a dose-dependent mechanism involving Her3 and Her8a. Indeed, we found that the co-injection of her8aMO and her3MO at active doses did not produce an additional phenotype (Figure 3F) and that, if both morpholinos were injected together in amounts just below their effective concentration, ectopic neurog1 expression was not observed. Given that the two factors do not regulate each other's expression, these experiments suggest that the presence of each factor individually, rather than their overall dose, is relevant to maintain neurogenesis inhibition within r2 and r4. Although Her8a was isolated as a binding partner for Her5 in yeast cells, co-purification shows that the full length Her8a and Her3 proteins heterodimerise (Figure 4A), and the overlapping expression of her8a and her3 makes it possible that this interaction occurs in vivo. Her proteins can dimerise with a variety of partners, as also supported by our yeast two-hybrid results (Additional file 1, Table S1), and Her-Her heterodimers display enhanced stability over homodimers [7, 22]. A parsimonious interpretation of our results is therefore that the heterodimerisation of Her8a and Her3 is required for sufficient activity of these factors in r2 and r4. Alternatively, the individual activities of Her3 and Her8a may control complementary properties necessary to maintain the progenitor pool cell state.
In spite of the high level of her8a expression across the MHB progenitor pool, the results of the present paper also identify that the decisive inhibition of neurog1 expression in this location is played by other factors, namely Her3, 5, 9 and 11. Her5 and 11 were known for their dose-dependent redundant functions, accounting for neurogenesis inhibition across part of this domain . Through knocking-down all four genes, we could achieve for the first time the transformation of most of the MHB into a neurogenic domain (Figure 5), while leaving her8a expression intact. Collectively, our findings show that the progenitor pool pattern of the midbrain and anterior hindbrain is established by the joined activities of five prepatterning E(spl) factors which act in different combination in the MHB and rhombomere domains. They also suggest that distinct mechanisms of action of these factors may be involved in these two domains.
Importantly however, we observed that the massive neurogenic phenotype induced upon blocking Her3/5/9/11 E(spl) activities is only partially followed by neuronal differentiation (Figure 5). In fact, ectopic neurons are restricted to the ventrolateral aspects of the midbrain-hindbrain boundary, like upon blocking Her5 function alone . The corresponding progenitor population may be particularly prone to neuronal differentiation. For all other progenitors, our observations suggest the need for a further commitment event, independent of Her3/5/9/11 activities, to achieve neuronal differentiation following neurog1 induction. Blocking Notch signaling concomitantly to Her3/5/9/11 did not allow further neurogenesis progression (C. Stigloher, unpub.). Persistent her8a expression in this context may contribute to neurogenesis reversion, although it was not possible to evaluate this possibility as embryos blocked for the activities of all five E(spl) factors developed abnormally.
her8aexpression in the early neural plate is controlled by Sox transcription factors but not Notch signaling
Two types of her genes have been recently distinguished based on their Notch response profile: those acting as Notch mediators, depending on Notch signaling for their expression and overexpressed upon Notch activation, and "non-canonical" her genes endogenously independent of Notch and repressed when Notch is experimentally activated (reviewed in ). The former class comprises zebrafish her4.1 and her15, expressed in active neurogenic domains such as proneural clusters of the early neural plate [9, 10, 11]; the latter class is composed of her3, 5, 9 and 11, expressed in progenitor pools [11, 13, 15, 21, 22]. Our results illustrate that her8a is unusual in its expression pattern, which overlaps both proneural clusters and progenitor pools. This property is shared with its ortholog Hes6.2 in chicken . her8a also shows a distinctive response to Notch signaling among hairy/E(spl) genes within the early neural plate, since it endogenously does not depend on Notch signaling but responds positively to the experimental activation of Notch (Figure 6). In agreement with the latter observation, we could identify Su(H) binding sites in the upstream regulatory sequence of her8a. However, this potential appears not to be used within proneural clusters of the early neural plate, demonstrating that Her8a is not a mediator of lateral inhibition. This is also in agreement with its uniform rather than salt-and-pepper expression profile. We found nevertheless that overexpressing her8a abolishes neurog1 expression even in proneural clusters where the two genes are normally co-expressed. Although we cannot ascertain that high overexpression levels mimic endogenous Her8a activity, one hypothesis reconciling this different information is that Her8a function within proneural clusters may generally dampen neurog1 expression, contributing to the function of other Her factors in Notch-inhibited precursors, and ensuring a proper differentiation schedule in committed progenitors. Although not further analyzed in this paper, we noted also that her8a expression becomes dependent on Notch signaling at later developmental stages (K. Webb, unpublished).
Our analyses of her8a expression in morphant contexts for other Her factors did not highlight cross-regulations, although we found several consensus N and E boxes within the 600 bp upstream of the her8a start site. Previous work demonstrated the positive regulation of Xenopus Hes6 by proneural bHLH proteins, in particular Neurogenin . Given the presence of E boxes on the her8a promoter, and the co-expression of her8a and neurog1 in proneural clusters, it will be interesting to test whether her8a expression is also positively controlled by proneural factors in these locations. The Ngn1/Hes6 cascade is positively reinforcing proneural activity in Xenopus , but our functional data would predict an opposite outcome for a Neurog1/Her8a regulation in zebrafish.
Finally, our results show that the levels of her8a expression are under control of a combination of SoxB1 and B2 factors (Sox2/3/19a/19b and Sox21a, respectively) that display intense and partially overlapping expression within the anterior neural plate (Figure 7). In a recent study, Okuda et al.  demonstrated that SoxB1 factors function redundantly to control several successive aspects of zebrafish nervous system development, including neural plate patterning and primary neurogenesis. Although single morphants do not harbor a visible phenotype, her3 expression fails to be induced in quadruple SoxB1 morphants, strongly suggesting that the four factors act redundantly to activate her3 transcription . In a comparable manner, we found that individual SoxB1/B2 morphants display no phenotype, while her8a expression is reduced in the MH domain when all five SoxB1/B2 proteins are abolished. Although we have not tested all possible knock-down combinations, and in particular did not assess the individual relevance of Sox21a in the context of the quadruple knock-out for SoxB1 proteins, these results demonstrate that her8a expression levels are under control of the activity of at least partially redundant SoxB proteins. Expression of these factors is an integral part of the mechanisms patterning the early embryo [40, 41], linking her8a expression with neural plate regionalization. The identification of a Sox2 binding site within the her8a enhancer, and the fact that all SoxB proteins recognize a similar binding motif in vitro, further suggests that part of this control may be direct. In support of this hypothesis, direct binding of SoxB1 factors onto the her3 enhancer has been demonstrated . Like several other SoxB2 proteins, Sox21b was shown to act as a transcriptional inhibitor during dorsoventral patterning of the zebrafish gastrula  and generally promotes neurogenesis . It can however act as an activator in other contexts , and its specific effect on MH neurogenesis and her genes needs to be directly evaluated. It was recently proposed that SoxB1 transcription factors and Notch cooperate through distinct mechanisms in their control of neurogenesis inhibition, including the inhibition of proneural protein activity and the transcriptional upregulation of Hes/her genes, respectively . Our results and those of Okuda et al.  suggest yet another level of regulation, where SoxB proteins directly control the level of expression of some her genes. Whether this is limited to Notch-independent contexts, such as with the regulation of her3 and her8a, remains to be addressed.
In this work, we identify the Hairy/E(spl) transcription factor Her8a as a local inhibitor of neurogenesis in the developing hindbrain. Specifically, we show that Her8a function, like Her3 , is required to generate the non-neurogenic progenitor pools normally separating the presumptive moto- and lateral neurons of r2 and 4. We demonstrate that Her8a is a binding partner for Her3 and we propose that this interaction may be functionally relevant in r2 and r4. We further show that Her8a alone is not sufficient to inhibit neurog1 expression in the presumptive MHB area; this event depends, in contrast, on the combined activities of four other E(spl) factors, Her3, 5, 9 and 11. Unlike canonical E(spl) genes, we demonstrate that her8a does not depend on Notch signaling for its expression at early neural plate stages, and we identify a combination of SoxB factors that together enhance her8a expression. Finally, using phylogenetic analyses, we show that Her8a belongs to a Hes6-like subfamily that was recently lost in the mammalian lineage. This observation provides a context for the strikingly divergent functions of Her8a from Hes6; Hes6, which was previously believed to be the mammalian ortholog of Her8a, displays proneural activity. Together, our results characterize the phylogeny, expression and functional cascades involving a new Her factor, and highlight the complex interplay of E(spl) proteins that generates the neurogenesis pattern of the zebrafish midbrain-hindbrain area.
Yeast Two-Hybrid Analysis
Yeast two-hybrid screening was performed by Hybrigenics, S.A., Paris, France (http://www.hybrigenics.com). The coding sequence for amino acids 20 to 201 of the Danio rerio Her5 protein (GenBank proteic accession number gi: 18858797)
(amino acid sequence DRINQSLETLRMLLLENTNNEKLKNPKVEKAEILESVVHFLRAEQASETDPFQITRVKRARTEES
was PCR-amplified and cloned into pB29 as an N-terminal fusion to LexA (N-Her5-LexA-C). The construct was checked by sequencing the entire insert and used as a bait to screen a random-primed Danio rerio embryo (stages 18-20 hpf) cDNA library constructed into pP6. pB29 and pP6 derive from the original pBTM116  and pGADGH  plasmids, respectively. 76 million clones (7.6 -fold the complexity of the library) were screened using a mating approach with Y187 (mata) and L40DGal4 (mata) yeast strains as previously described . 280 His+ colonies were selected on a medium lacking tryptophan, leucine and histidine, and supplemented with 2 mM 3-aminotriazole to handle bait autoactivation. The prey fragments of the positive clones were amplified by PCR and sequenced at their 5' and 3' junctions. The resulting sequences were used to identify the corresponding interacting proteins in the GenBank database (NCBI) using a fully automated procedure. For each interaction, a Predicted Biological Score (PBS) was computed to assess interaction reliability. This score represents the probability of an interaction being nonspecific. PBS relies on two different levels of analysis; the algorithm and methods used in the calculation are described in detail in Formstecher et al. . Briefly, at first a local score takes into account the redundancy and independency of prey fragments (i.e. the times the interaction was detected with different independent clones and whether it was detected with different or the same fragments), as well as the distribution of reading frames and stop codons in overlapping fragments. Thus, interactions detected with several and different fragments are ranked with a very high confidence score and interactions detected with a single independent fragment are ranked with a moderate confidence score. Secondly, a global score takes into account the interactions found in all the screens performed at Hybrigenics using the same library. This global score represents the probability of an interaction being nonspecific. For practical use, the scores were divided into four categories, from A (highest confidence) to D (lowest confidence). A fifth category (E) specifically flags interactions involving highly connected prey domains previously found several times in screens performed on libraries derived from the same organism. Finally, several of these highly connected domains have been confirmed as false-positives of the technique and are now tagged as F. The PBS scores have been shown to positively correlate with the biological significance of interactions [54, 55].
Gene ontology analysis
Gene ontology enrichment analysis was performed on the recovered yeast-2-hybrid candidates from the categories A, B and C using the AmiGO "Term Enrichment tool"  (available at http://amigo.geneontology.org/cgi-bin/amigo/term_enrichment), using the following settings: ZFIN database as a background set, the maximum p-value set at 0.05 and a minimum number of gene products of two.
Sequence alignment, protein domain identification, phylogenetic and synteny analyses
Protein sequences were retrieved by using a tblastn search  against the non redundant database on NCBI or on Ensembl genomic data (current release of genomes, July 2010). For non-annotated sequences and to further support the expression of the predicted gene, a search for expressed sequence tags was also performed by tblastn on the EST database of the NCBI server. A list of all sequences used for the molecular phylogeny and their genomic locations is provided in Additional file 7, Table S3. Protein sequences were aligned using the ClustalX software. Only unambiguously aligned residues were retained to build the phylogenetic tree: namely the basic, HLH and Orange domains (see Additional file 3, Figure S1). Maximum likelihood phylogeny was constructed using PhyML (substitution model: JTT, number of substitution rate categories: 4; gamma distribution parameter estimated; proportion of invariable sites estimated) . Branch support was estimated by approximate likelihood-ratio test (aLRT, SH-like) . For the synteny analysis, the position and orientation of annotated genes surrounding hes6 family members was retrieved using the Ensembl genome browser.
Adult zebrafish were maintained using standard fish-keeping protocols and in accordance with institute guidelines for animal welfare (defined by the Regierung von Oberbayern and the Services Vétérinaires de l'Essonne). Wildtype (AB) embryos were obtained through natural matings and were staged according to Kimmel et al. .
In situ hybridization
In situ hybridization on embryos was performed as previously described [22, 61] using the following probes: gfp , neurog1 , deltaNP63 , her4.1 , her5 , her9 , her3  and barhl2 . For the her8a and her13 probes, 648bp and 785bp fragments (respectively) were cloned into pCRII-TOPO (Invitrogen) from cDNA from 24hpf AB embryos using the following primers: for her8a forward 5'CACTGCTTGGAAGCAAATGA34, reverse 5'GACTTGGCGTGTGATTGATG3' and for her13 forward 5'TTTCTGTCCAACCCCTTCTG3', reverse 5'GATCCAATCCGATGTTGCTT3' (PCR conditions available on request). The successful clones were verified by sequencing. RNA probes were synthesized following published protocols . For immunohistochemisty the primary antibody was mouse Anti-Hu (diluted 1:1500) (A21271; Molecular Probes) revealed using Cy-2.
GFP of the -8.4neurog1:gfp line  was detected using a chicken primary antibody (GFP-1020, Aves Labs, Inc) followed by a secondary antibody anti-chicken coupled to Alexa-488 (Molecular Probes, Invitrogen).
RNA and morpholino injections
Capped RNAs were synthesized using Ambion mMessage mMachine Kit and embryos were injected at the one-cell stage. For her8a overexpression, full-length her8a was cloned by PCR (forward primer: 5'AATAATGACGGCCTCCAACA3'; reverse primer: 5'GGCTGCATTCATTCACCAG3') and cloned into pXT7. her8a capped mRNA was injected at a concentration of 62.5 ng/μl. NICD overexpression was achieved by injecting capped RNA for nic, which encodes the NICD fragment of zebrafish Notch1 [9, 10]. Morpholinos were purchased from Gene-Tools (Philomath, USA) and gripNAs were purchased from Active Motif (Carlsbad, USA). Two her8a splice morpholinos were used, targeted against the donor and acceptor sites of her8a exons 1 and 2 (respectively her8aMO1 and MO2, see Figure 3G). Both her8aMO1 (ATGTGACATTACCTTTCGCTCCTCT) and MO2 (CGCAGCTAAAATGATAGAAAGCATG) were injected at 1mM. Their efficiency was evaluated by RT-PCR following standard protocols on pools of 25 embryos at the 3-somite stage, in comparison to βactin2 expression, using the following PCR primers (FL: primer designed to amplify the full-length her8 transcript; E2: primer designed to amplify the E1-E2 domain; FW: forward primer; REV: reverse primer): her8a-FL-FW: AATAATGACGGCCTCCAACA; her8a-FL-REV: GGCTGCATTCATTCACCAG; her8a-E2-REV: TCCTCTCTCTGCGTTTCTTCTC; βactin2-FWD: AAGGCCAACAGGGAAAAGAT; βactin2-REV: GTGGTACGACCAGAGGCATAC (expected size 109bp). A third her8a MO, targeted against the her8a ATG (MO3: CATTGCCCATGTTGGAGGCCGTCAT) was also used and injected at 1.5 mM. The her3 morpholino  was injected at 0.5mM. For the combined knockdown of her3, her5, her9 and her11 the following gripNAs were used, each at 0.4mM: her3 gripNA (AGCCATTGTCCTTAAATG) (overlapping the MO sequence published in ), her5 gripNA (GGTTCGCTCATTTTGTGT) , her9 gripNA (TGATTTTTACCTTTCTAT) (overlapping the MO sequence published in ) and her11 gripNA (AGTCGGTGTGCTCTTCAT) . For the combined knock-down of sox genes, we used: sox2/sox3 morpholino (CTCGGTTTCCATACATGTTATACATT) [42, 43] at 1mM (this MO was initially reported to target sox2 only, but we found that its target site is shared between sox2 and sox3), sox19a morpholino (TGCTGTACATGGCTGCCAACAGAAG)  at 1 mM, sox19b morpholino (TAGCCCTTTTCTCAAAACAAACCTG) at 0.25mM, sox21a morpholino (CATGGGCTTTGCCATTTCTTGATAC) (overlapping with the sox21aMO used in  at 1mM.
DAPT treatment was carried out according to Geling et al. . Embryos were placed in 100 μm DAPT (Alexis Biochemicals), 1% DMSO in embryo medium from 50% epiboly to 3 somites. Control embryos received a corresponding treatment with 1% DMSO. After treatment, the embryos were fixed in 4% PFA overnight at 4°C before being processed for in situ hybridisation.
The full length coding sequences of her3 and her8a were subcloned into the N-Strep/Flag (SF) TAP  and myc pCS3+ vectors, respectively. SF-TAP-tagged Her3 and Myc-tagged-Her8a were co-expressed in HEK293T cells. For this purpose 2 μg of pCS3-myc-her8a was co-transfected with either 6 μg pcDNA3.0-SF-TAP-her3 or control vector per 14 cm culture dish of HEK293T cells. After 48 h, cells were lysed in lysis buffer (TBS supplemented with Complete protease inhibitor (Roche), phosphatase inhibitor cocktails I and II (Sigma) and 0.5% NP-40 (Roche)). Beside the solubilization of cytoplasmic proteins, this condition allows the extraction of nuclear proteins which are not tightly bond to DNA. After incubation for 30 min at 4°C cell debris including nuclei were pelleted by centrifugation at 10,000 xg for 10 min, 4°C. The supernatants were filtered through 22-μm syringe filtration units (Millipore). Cleared lysates from two 14 cm culture dishes containing 5-6 mg total protein were incubated with 25 μl Strep-Tactin resin (IBA) for 2 h. After incubation the resin was washed 3 times with lysis buffer. The SF-TAP-tagged Her3 protein was eluted with 100 μl desthiobiotin elution buffer (IBA) and subjected to SDS PAGE. For detection, proteins were blotted onto PVDF membranes (GE Healthcare). The blots were incubated for 1 h with blocking solution (5% dry milk powder in TBST). The SF-TAP-tagged Her3 was detected by incubation of the blots with rabbit polyclonal anti FLAG antibodies (Sigma, 1:2000 in blocking reagent) overnight and secondary anti-rabbit antibodies (Jackson Immuno Research, 1:15,000 in blocking reagent) for 2 h. After each antibody incubation step, the blots were washed 4 times 5 min with TBST. For detection of Myc-Her8a, the blots were stripped using standard protocols and incubated with mouse anti-Myc (Cell signalling, 1:2000) overnight and secondary anti-mouse antibodies (Jackson Immuno Research, 1:15,000 in blocking reagent) for 2 h. Antibody-antigen complexes were visualized using the ECL+ chemiluminescence detection system (GE Healthcare) on Hyperfilms (GE Healthcare).
Acknowledgements and Funding
We thank colleagues from the past and present Bally-Cuif lab for their valuable input into this project, and the fish house teams at the Helmholtz Zentrum München and CNRS. Work in the LBC laboratory at the Helmholtz Zentrum München was funded by the Helmholtz Association, the EU ZF-Models integrated project (contract No. LSHG-CT-2003-503466) and the Center for Protein Science-Munich (CIPSM). Work in the LBC lab at the CNRS is supported by funds from the EU 7th framework integrated projects NeuroXsys and ZF-Health, the Agence Nationale pour la Recherche, the Ecole des Neurosciences de Paris, the Fondation pour la Recherche Médicale, the PIME program and the Schlumberger Association. MC has been recipient of a Marie Curie Intra-European Fellowship and an EMBO Long Term postdoctoral fellowship. Work of MU and CJG is supported by the Helmholtz Alliance for Mental Health in an Aging Society (HelMA, FKZ: HA-215) and the BMBF (neurogenesis from brain and skin cells, FKZ: 01 GN 1009 C), respectively.
- 22.Ninkovic J, Tallafuss A, Leucht C, Topczewski J, Tannhauser B, Solnica-Krezel L, Bally-Cuif L: Inhibition of neurogenesis at the zebrafish midbrain-hindbrain boundary by the combined and dose-dependent activity of a new hairy/E(spl) gene pair. Development. 2005, 132: 75-88.CrossRefPubMedGoogle Scholar
- 25.Jaillon O, Aury JM, Brunet F, Petit JL, Stange-Thomann N, Mauceli E, Bouneau L, Fischer C, Ozouf-Costaz C, Bernot A, Nicaud S, Jaffe D, Fisher S, Lutfalla G, Dossat C, Segurens B, Dasilva C, Salanoubat M, Levy M, Boudet N, Castellano S, Anthouard V, Jubin C, Castelli V, Katinka M, Vacherie B, Biemont C, Skalli Z, Cattolico L, Poulain J, De Berardinis V, Cruaud C, Duprat S, Brottier P, Coutanceau JP, Gouzy J, Parra G, Lardier G, Chapple C, McKernan KJ, McEwan P, Bosak S, Kellis M, Volff JN, Guigo R, Zody MC, Mesirov J, Lindblad-Toh K, Birren B, Nusbaum C, Kahn D, Robinson-Rechavi M, Laudet V, Schachter V, Quetier F, Saurin W, Scarpelli C, Wincker P, Lander ES, Weissenbach J, Roest Crollius H: Genome duplication in the teleost fish Tetraodon nigroviridis reveals the early vertebrate proto-karyotype. Nature. 2004, 431: 946-957. 10.1038/nature03025.CrossRefPubMedGoogle Scholar
- 26.Thisse B, Thisse C: Fast release clones: a high throughput expression analysis. ZFIN Direct Data Submission. 2004Google Scholar
- 28.Bakkers J, Hild M, Kramer C, Furutani-Seiki M, Hammerschmidt M: Zebrafish DeltaNp63 is a direct target of Bmp signaling and encodes a transcriptional repressor blocking neural specification in the ventral ectoderm. Dev Cell. 2002, 2: 617-627. 10.1016/S1534-5807(02)00163-6.CrossRefPubMedGoogle Scholar
- 36.Ambrosetti DC, Basilico C, Dailey L: Synergistic activation of the fibroblast growth factor 4 enhancer by Sox2 and Oct-3 depends on protein-protein interactions facilitated by a specific spatial arrangement of factor binding sites. Mol Cell Biol. 1997, 17: 6321-6329.PubMedCentralCrossRefPubMedGoogle Scholar
- 38.Thisse B, Pflumio S, Fürthauer M, Loppin B, Heyer V, Degrave A, Woehl R, Lux A, Steffan T, Charbonnier XQ, Thisse C: Expression of the zebrafish genome during embryogenesis. ZFIN Direct Data Submission. 2001Google Scholar
- 45.Jhas S, Ciura S, Belanger-Jasmin S, Dong Z, Llamosas E, Theriault FM, Joachim K, Tang Y, Liu L, Liu J, Stifani S: Hes6 inhibits astrocyte differentiation and promotes neurogenesis through different mechanisms. J Neurosci. 2006, 26: 11061-11071. 10.1523/JNEUROSCI.1358-06.2006.CrossRefPubMedGoogle Scholar
- 49.Holmberg J, Hansson E, Malewicz M, Sandberg M, Perlmann T, Lendahl U, Muhr J: SoxB1 transcription factors and Notch signaling use distinct mechanisms to regulate proneural gene function and neural progenitor differentiation. Development. 2008, 135: 1843-1851. 10.1242/dev.020180.CrossRefPubMedGoogle Scholar
- 51.Bartel PL: Unsing the two-hybrid system to detect protein-protein interactions. In Cellular Interactions in Development: A Practical Approach. 1993, 153-179.Google Scholar
- 53.Formstecher E, Aresta S, Collura V, Hamburger A, Meil A, Trehin A, Reverdy C, Betin V, Maire S, Brun C, Jacq B, Arpin M, Bellaiche Y, Bellusci S, Benaroch P, Bornens M, Chanet R, Chavrier P, Delattre O, Doye V, Fehon R, Faye G, Galli T, Girault JA, Goud B, de Gunzburg J, Johannes L, Junier MP, Mirouse V, Mukherjee A, Papadopoulo D, Perez F, Plessis A, Rosse C, Saule S, Stoppa-Lyonnet D, Vincent A, White M, Legrain P, Wojcik J, Camonis J, Daviet L: Protein interaction mapping: a Drosophila case study. Genome Res. 2005, 15: 376-384. 10.1101/gr.2659105.PubMedCentralCrossRefPubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.