Analysis of her1 and her7 Mutants Reveals a Spatio Temporal Separation of the Somite Clock Module

Somitogenesis is controlled by a genetic network consisting of an oscillator (clock) and a gradient (wavefront). The “hairy and Enhancer of Split”- related (her) genes act downstream of the Delta/Notch (D/N) signaling pathway, and are crucial components of the segmentation clock. Due to genome duplication events, the zebrafish genome, possesses two gene copies of the mouse Hes7 homologue: her1 and her7. To better understand the functional consequences of this gene duplication, and to determine possible independent roles for these two genes during segmentation, two zebrafish mutants her1hu2124 and her7hu2526 were analyzed. In the course of embryonic development, her1hu2124 mutants exhibit disruption of the three anterior-most somite borders, whereas her7hu2526 mutants display somite border defects restricted to somites 8 (+/−3) to 17 (+/−3) along the anterior-posterior axis. Analysis of the molecular defects in her1hu2124 mutants reveals a her1 auto regulatory feedback loop during early somitogenesis that is crucial for correct patterning and independent of her7 oscillation. This feedback loop appears to be restricted to early segmentation, as cyclic her1 expression is restored in her1hu2124 embryos at later stages of development. Moreover, only the anterior deltaC expression pattern is disrupted in the presomitic mesoderm of her1hu2124 mutants, while the posterior expression pattern of deltaC remains unaltered. Together, this data indicates the existence of an independent and genetically separable anterior and posterior deltaC clock modules in the presomitic mesdorm (PSM).


Introduction
Somitogenesis is an essential and complex process during early vertebrate development. As the body axis elongates, transient metameric structures, called somites, bud off from the PSM at the tail bud adjacent to both sides of the notochord. This complex process requires the carefully coordinated activation and inhibition of gene transcription and is controlled by a molecular oscillator [1][2][3][4]. Extensive studies have been carried out to elucidate the mechanisms that control cyclic gene expression, revealing important roles for signaling pathways such as D/N-, Wnt-and FGF-signaling. However, the genetic network and interplay between these pathways is not fully understood yet. Typically, loss of function of one component in this network does not lead to breakdown of the whole process. Instead, only partial somitic defects occur at distinct positions along the body axis. Thus, it seems likely that the system possesses the ability to compensate for the loss of individual signal inputs found in loss of function situations [5][6][7][8]. Alternatively, it suggests that during embryonic development multiple mechanisms exist to control segmentation over time.
The process of somitogenesis commences when the first anlagen of the somites are generated and involves three steps that are essential for somite formation. First, the unsegmented PSM is prepatterned, followed by the establishment of rostro-caudal (r/c) polarity and finally by the formation of somitic borders [9,10]. However, it remains elucidated whether these three steps are functionally linked or are driven by independent mechanisms. One of the major pathways involved in the process of pre patterning is the D/N-signaling pathway. The components of the D/N pathway, together with their target genes from the hairy and enhancer of Split (hes) family, constitute a genetic feed-back loop [11,12] which ultimately results in cyclic gene expression. Morpholino oligonucleotide (MO) mediated knock down studies in zebrafish have shown that loss of Her function disrupts the cyclic expression of D/N components, suggesting an important role for Her transcription factors in the D/N-mediated oscillation mechanism [13,14].
Her genes encode basic Helix-Loop-Helix (bHLH) transcription factors, which act in a protein complex with the co-repressor Groucho [15]. Due to a gene duplication, zebrafish possess two homologues of murine Hes7 [16], annotated as her1 and her7. Both genes have been reported to play important and separate roles during pre patterning of the unsegmented PSM. MO mediated knock down studies indicate an essential requirement for her1 in the formation of the first three somites [14], whereas her7 was shown to play a role in segmentation posterior to the ninth somite [6]. Moreover, loss of function of both her genes, either in the b567 mutant or through MO mediated knock down [14], results in disruption of all somites. These findings suggest non-redundant roles or temporally separate roles for both her genes during specific stages of segmentation.
In this study, we present novel zebrafish her1 and her7 mutants, and analyse the role of both her genes in pre-patterning of the PSM during early embryonic development. Furthermore, we analyse PSM pre-patterning in double-mutant fish lacking both DeltaC and Her1 function. Expression analysis of the clock genes in double-mutant embryos revealed a critical role for Her7 dependent posterior PSM oscillations in the synchronization of gene expression in adjacent cells. In contrast, we found that Her1 drives the pre-patterning of the first three somites in the anterior PSM. Together, our study demonstrates distinct spatio-temporal requirements for her1 and her7 during somite formation.

Results and Discussion
Characterisation of the her1 and her7 Mutant Alleles ENU-induced point mutations were identified in the her1 and her7 genes by 59-end sequencing of the relevant genomic DNA derived coding sequences amplified from mutagenized fish.
One allele with a single base pair transition was identified for each gene. The her1 hu2124 allele (acc no X97329) contains a C.A transition at position 185, resulting in a premature stop codon (TCG(S).TAG/(stop)). The her7 hu2526 allele (acc no AF240772, [17] contains an A.T transition at position 208, also resulting in a premature stop codon (AAA(K).TAA/(stop) (Fig. 1A,B). In both mutants the stop-codon is located upstream of the basic domain. her1 hu2124 is truncated within the loop located at the end of exon 2, and her7 hu2526 is truncated within HelixI of the HLH-domain located in exon 2. Thus, both mutant proteins lack a full HLHdomain, and are hypothesized to lack dimerization function.

Her1 is required for Patterning the Anterior-most Somites
Previous studies have demonstrated that her1-morphant embryos show a spectrum of phenotypes ranging from mild morphological defects in the anterior 1 to 3 somites, to more severe defects observed along the entire axis [6,13,14]. This variability in phenotype may be attributed to incomplete knock down using MO, and therefore can make it difficult to determine precisely the requirement for Her1 during segmentation. Therefore, to better understand the function of Her1 during early somitogenesis, we compared segmentation events between her1 hu2124 homozygous mutant embryos and wild type siblings. Whereas wild type siblings showed normal somite formation (Fig. 1C), her1 hu2124 homozygous mutant embryos exhibit defects in the borders of the first (anterior) somites (Fig. 1D). Consistently, analysis of myogenic differentiation 1 expression (myoD, [18]) reveals a diffuse pattern within the misshapen somites of her1 hu2124 mutant embryos when compared to wild type embryos or to more posterior somites in the mutant (Fig. 1E, F).
To determine the requirement for Her1 in establishing r/c polarity, the expression pattern of mesoderm posterior (mesp) [19] was compared in wild type and her1 hu2124 mutant embryos. mespb expression in her1 hu2124 mutant embryos was disrupted during the pre-patterning of somites 1 to 3. While wild type embryos display a stripe expression pattern of mespb (Fig. 1K, L), a ''salt and pepper''like expression pattern was observed in the her1 hu2124 mutant (Fig. 1M, N). During later stages of segmentation, when border formation is unaffected in the her1 hu2124 mutant, wild type-like expression of mesp is restored (Fig. 1G-J). This indicates that the maintenance of r/c polarity in the anterior-most somites is regulated through Her1 activity. To understand the relationship between the morphological somite defects observed in the her1 hu2124 mutant and the molecular oscillation clock, the expression patterns of deltaC, her1 and her7 were examined between 90% eiboly and bud stage, when the first 3 somites are prepatterned (Fig. 2). While wild type embryos display cyclic deltaC expression ( Fig. 2A), her1 hu2124 mutants exhibit disruption of the cyclic deltaC expression in the anterior PSM (Fig. 2D). Only one deltaC expression domain is detectable in the Her1 loss of function situation. Importantly, oscillating deltaC expression in the posterior PSM was detected in both wild type ( Fig. 2A) and her1 hu2124 mutant embryos (Fig. 2D), indicating that cyclic deltaC expression in the posterior PSM is independent of Her1 function. To further confirm both Her1-dependent and -independent deltaC oscillations, deltaC expression was analyzed at the 10-12 somite stage, when somite border defects are no longer observed in her1 hu2124 mutants ( Fig. 3). At this stage, her1 hu2124 mutants express only a single stripe of deltaC in the anterior PSM, in contrast to the 1-2 stripes of expression observed in wild type embryos, indicating that cyclic deltaC expression in the anterior PSM is indeed dependent on Her1 activity ( Fig. 2A, D). In contrast, different phases of oscillation in the posterior PSM were detected in both, wild type embryos ( Fig. 2A) and in her1 hu2124 mutant embryos (Fig. 2D), indicating that deltaC expression oscillates in the absence of functional Her1 in the posterior PSM. Thus, the absence of Her1 leads to impaired deltaC expression in the anterior PSM, whereas cyclic gene expression in the posterior PSM is not affected. These findings support the conclusion that cyclic deltaC expression in the posterior part of the PSM occurs independent of Her1. Furthermore, our investigation suggests that two deltaC clock modules exist, in which the posterior and anterior deltaC expression waves are driven separately. Although loss of Her1 activity results in disruption to both anterior deltaC expression and formation of anterior somite borders, later during segmentation these somite borders are restored while deltaC expression remains disrupted in the her1 hu2124 mutant. It is therefore unlikely that the morphological somite defects in her1 hu2124 mutant embryos are caused by disrupted deltaC expression.
Next, the expression pattern of her genes in the her1 hu2124 mutant was analyzed. Cyclic expression of her1 is disrupted in her1 hu2124 homozygous mutants between 90% epiboly and bud stage (Fig. 2B, E). In contrast, oscillation of her7 is not affected at this stage in the her1 hu2124 mutant (Fig. 2C, F), suggesting that Her1 negatively regulates its own expression, but is not required for her7 expression during early segmentation. Interestingly, during later segmentation stages oscillating her1 expression patterns are observed ( Fig. 2E-H), demonstrating that her1 resumes oscillation over the course of development, even in the absence of Her1. However, the domain of cyclic expression of both her1 and her7 in the posterior PSM appears expanded anteriorly, and with a simultaneous lack of an expression wave (Fig. 3E-L). Nevertheless, defects in somite formation are not observed in later stages, indicating that altered her1 and her7 expression does not affect somite boundary formation. Thus, Her1 acts in a temporally restricted manner and contributes to the segmentation clock independent of the DeltaC-Her7 feedback loop during early development.
her7 and deltaC Oscillation are Regulated Through Her1 During Early Development bea/deltaC mutant embryos exhibit segmentation defects along their antero-posterior axis, beginning between the third and fifth somite. In addition to these morphological defects, expression analysis revealed that expression of segmentation clock genes is perturbed (Fig. 4A, B; [20,21]). Examination of her1 hu2124 mutant embryos revealed a complementary pattern of somite disruption, whereby only the first three somite borders are disrupted (Fig. 1D). To better understand the relationship between DeltaC and Her1, homozygous double mutant embryos for her1 and deltaC were created and somite border defects were analyzed and compared between double mutants, single mutants and wild type embryos ( Fig. 4A-C). her1 hu2124 /bea tm98 homozygous double mutants show disruption of somitic borders along the entire axis (Fig. 4C). In addition, half segmental ( Fig 4D) expression of myoD is disrupted in all somites (Fig. 4F), compared with the restricted anterior perturbation in her1 hu2124 mutants (Fig. 1F) and the defects observed in bea tm98 mutants starting from somites three to five (Fig. 4E). The same segmentation defect was observed by analyzing expression of a segment border marker. In wild type embryos at prim-6 stage eplin is expressed along the segment borders in a characteristic v-shape (Fig. 4G, [22]. This pattern is disrupted in the three anterior-most somites in her1 hu2124 mutants (Fig. 4H). In bea tm98 mutants expression of eplin in all somites but the first three or four are disrupted (Fig. 4I). Double her1 hu2124 /bea tm98 mutants display disrupted eplin expression along the whole axis (Fig. 4J).
To investigate the influence of the loss of both Her1 and DeltaC on the segmentation clock, the expression of deltaC, her1 and her7 was examined in embryos between 90% epiboly and bud stage and compared to the expression patterns observed in single mutants and wild type embryos. Analysis of bea tm98 mutants revealed that the expression pattern of all three genes oscillates normally prior to the three somite stage, although expression is slightly diffuse compared to the wild type embryos (Fig. 2G, H, I, respectively, [20,21]. In her1 hu2124 mutants, as described above, expression of her1 is perturbed (Fig. 2E) and deltaC oscillation is only disrupted in the anterior PSM (Fig. 2D), whereas her7 Figure 1. Her1 mutants exhibit defects in somitogenesis. Electropherogram of her1 (A) and her7 (B) amplicons in wild type (top), homozygous her1 hu2124 (bottom, A) and her7 hu2526 (bottom, B) mutant fish. Schematics above the sequences depict the exon-and intron-organization and the protein domains encoded by the exons. Point mutations are indicated by asterisks. (C, D) Brightfield pictures of wild type and her1 mutant embryos, lateral view, anterior to left. Compared to wild type embryos (C, asterisks), the first 3 somitic borders in the her1 hu2124 mutant appear diffuse and partly disrupted (D, bracket). In situ analysis of myoD expression in wild types indicates characteristic half-segmental expression within the somites (E, asterisks indicate somites 1-3). myoD expression is diffuse in the first 3 somites of her1 hu2124 embryos (F, bracket). (G, H) and (K, L) show halfsegmental respectively r/c polarity wild type expression pattern of mespb at 10-12 somite stage and between 90% epiboly and bud stage, respectively. (I, J) and (M, N) represents mespb expression in the her1 hu2124 mutant at 10-12 somite stage and between 90% epiboly and bud stage, respectively. Expression of mespb is disturbed in the her1 hu2124 mutant between 90% epiboly and bud stage (M, N), when the anlagen of the first somites are pre-patterned. Compared to one or two stripes in the wild-type (K, L), mespb is expressed in a salt and pattern (M, N). mespb expression is unperturbed at 10-12 somite stage in her1 hu2124 mutants (I, J). Dorsal views, anterior to top. doi:10.1371/journal.pone.0039073.g001 Figure 2. Expression analysis of segmentation genes in her1 hu212 and bea tm98 mutants. In situ hybridisation analysis of segmentation clock genes deltaC, her1, and her7 in wild type (A-C) her1 hu2124 mutants (D-F), bea tm98 mutants (G-I) and her1 hu2124 /bea tm98 double mutants (J-L) at 90% epiboly. Cyclic deltaC expression is disrupted in the anterior PSM of her1 hu2124 mutants. Instead of one or two expression stripes as in the wild type (A, arrowheads) only one stripe of expression is observed (D, arrowhead). Expression domains in the posterior PSM display different sizes indicating unperturbed oscillation of deltaC in the tail bud of her1 hu2124 mutants (D, bars). Cyclic expression of her1 is fully disrupted in the her1 hu2124 mutant (E) when compared to wild type (B), whereas her7 expression remains oscillatory (compare C and F). Cyclic expression of all three genes is observed in bea tm98 although some slight initial perturbation is observed (G-I). In her1 hu2124 /bea tm98 double mutants, all three clock genes show fully disrupted expression patterns at 90% epiboly. Dorsal views, anterior to the top, number in each panel indicate cycling phases. doi:10.1371/journal.pone.0039073.g002 Expression of her1 and her7 oscillates in the her1 hu2124 mutant but on average one expression stripe is lacking (see asterisks in G, H and K, L, respectively) compared to the respective wild type expression domains (asterisks in E, F and I, J). Further, the patterns in the PSM of mutants appear stretched towards the anterior compared to wild type (see bars in A-D) suggesting that one expression wave is lacking. Dorsal view, anterior to the top. doi:10.1371/journal.pone.0039073.g003 expression appears cyclic (Fig. 2F). In contrast, cyclic her7 expression in her1 hu2124 /bea tm98 double mutant embryos is completely disrupted, with her7 expressed in a gradient with declining expression from posterior to anterior (Fig. 2L). In addition, posterior deltaC oscillation is disrupted in the double mutant (Fig. 2J) when compared with the her1 hu2124 mutant (Fig. 2D). Instead of two different expression phases, which were observed in the posterior PSM in the her1 hu2124 mutant embryos (Fig. 2D), an invariant posterior expression pattern of deltaC was observed in her1 hu2124 /bea tm98 double mutants (Fig. 2J). Thus, cyclic expression of all three analyzed clock genes is completely disrupted in her1 hu2124 /bea tm98 double mutant embryos from the time point of initiation of segmentation. This indicates that cyclic her7 expression and posterior deltaC oscillation are regulated in a combinatorial manner through both a Her1 auto regulatory feedback loop and a D/N signaling module.

Mutant Embryos
In light of the phenotypic variability observed in her1 morphants, we re-analyzed the expression of the clock genes her1, her7 and deltaC during somitogenesis in her7 hu2526 mutants. Cyclic expression of deltaC is disrupted in her7 hu2526 mutant embryos (Fig. 5 A-C), similar to those phenotypes observed in her7 morphants, or in D/N mutants [6,13,23] at the 10-12 somite stage. Expression of her1 and her7 is disrupted in her7 hu2526 homozygous mutants in a similar manner to that observed in the her7 morphant (Fig. 5D-I). Thus, her7 morphants and her7 hu2526 mutants show similar disruption of the segmentation clock genes at the 10-12 somite stage. Furthermore, we found that expression of all examined clock genes is unperturbed during early somitogenesis (Fig. 5J-O). D/N mutants, such as bea, des, aei or mib, or MO mediated knock down of deltaC, notch1a, deltaD and E3 ligase, display somitic border defects from the 3 rd , 7th, 8th and 9th somite onwards, respectively. In line with the observed border defects cyclic gene expression of deltaC, her1 and her7 are disrupted [5,6,21,24]. In a similar fashion, cyclic gene expression of deltaC, her1 and her7 in her7 hu2526 mutants are disrupted in conjunction with somitic border malformation.

Her7 Plays an Essential Role During Pre-patterning
To determine the temporal onset of somite defects in her7 hu2526 mutant embryos, myoD expression was examined at 12-14 somite stage. The anterior limit of somitic boundary defects (ALD) in the her7 hu2526 mutant was observed around the level of the 8 th somite (Fig. 6A, B). The myoD expression pattern was disturbed at the same axial level (Fig. 6C, D [6]). To examine the posterior extent of somitic defects, eplin expression was analysed in the mutants after completion of somitogenesis, permitting visualization of the somite borders. In her7 hu2526 mutant embryos eplin expression is disrupted with high penetrance between somite 8(+/23) to somite 17(+/23) (n = 56, Fig. 6F and graph in Fig. 6G). Somitic borders posterior to this region appear unaffected indicating that a posterior limit of defects (PLD) exists in upon Her7 loss-offunction. In line with this finding, disrupted mesp expression was observed during, but not prior to, this time interval (Fig. 6H-K,). Thus, Her7 has a non-redundant role in somite border formation between the ,8 th and ,17 th somite.
In summary, molecular and morphological analysis of her1 and her7 mutants indicate a non-redundant requirement for both these genes in the correct segmentation of distinct somite regions in the zebrafish. Our data resolves previous seemingly contradictory data arising from her1 morphant analysis [23] and in vitro studies with  In situ hybridisation analysis of deltaC, her1 and her7 in wild type (A, B and D, E and G, H, respectively) and her7 hu2526 mutants (C, F, I, respectively) at 10-12 somite stage and between 90% epiboly and bud stage (J, L, N for wild type expression patterns and K, M, O for respective expression patterns in the mutant embryos). Expression patterns of deltaC, her1 and her7 at 10-12 somite stage are disrupted in the mutant appear unperturbed between 90% epiboly and bud stage. Expression patterns of mespa and mespb are not affected in the her7 hu2526 mutant between 90% epiboly and bud stage (Q and S, respectively) and similar to the wild type (P and R, respectively). doi:10.1371/journal.pone.0039073.g005 her1 promoter constructs [25]. Observations in the latter study strongly supported a her1 negative auto regulation mechanism; however in vivo analysis of the her1 morphant did not provide any supporting evidence for this conclusion. Our her1 hu2124 mutant analysis now suggests that the regulatory requirement of Her1 decreases during the course of segmentation. During early somitogenesis, Her1 activity constitutes a negative auto regulatory feedback loop, in agreement with the findings of Kawamura et al., 2005 [25], while later during development the auto regulatory potential of Her1 is considerably reduced or absent, as suggested by the residual expression in the inter stripe regions of her1 hu2124 mutants. Furthermore, Her1 does not negatively feed back on her7 in a direct manner, at either early or late somitogenesis, as our own study previously has suggested [23]. However, Her1 is required to regulate the rhythm of her gene oscillation, as shown by the altered expression patterns that suggest an increase in wavelength towards the anterior. This effect is most probably caused indirectly by loss of the repressive activity of Her1 on delta gene expression. Nevertheless, the increase in wavelength towards the anterior is not associated with changes in somite size (data not shown). Furthermore, analysis of clock genes in her1 hu2124 /bea tm98 double mutants revealed that cyclic her7 expression and posterior deltaC oscillation in the PSM are governed by a Her1 auto regulatory feedback loop. Morphologically, her1 hu2124 /bea tm98 double mutant exhibit a cumulative phenotype, strongly supporting distinct roles for Her1 and DeltaC during somitogenesis. Future studies should seek to identify the D/N independent Her1 targets that control anterior somite formation. Moreover, her7 hu2526 mutant analysis confirmed the role previously suggested for her7 in somitogenesis during the 1-12 somite stage. In addition, the observation of a PLD in the her7 hu2526 mutant further suggests a temporally restricted role for Her7 during somitogenesis.
In summary, the comparison between single her1 hu2124 and her7 hu2526 mutants and her1 hu2124 /bea tm98 double mutants suggests independent roles for both her genes in regulation of distinct phases of the segmentation clock. There subsequently remains an open question about the direct downstream targets of DeltaC, which together with Her1 are able to initiate cyclic her7 expression.

Ethic Statement
Adult zebrafish were handled according to relevant national and international guidelines and was approved by the German environment and customer protection office Cologne (1 11 Abs. 1 No. 1 for animal protection law (BGBL.I.S. 1005-1120). Only embryos up to 32 hpf were used for these experiments, which do not require approval of the animal experiments committee according to national and European law.

Genotyping and Used Mutant Fish
Fish were maintained at 28.5uC on a 14-h light/10-h dark cycle. Embryos were collected by natural spawning and staged according to Kimmel et al., 1995 [26].
her1 and her7 heterozygous mutants were identified by screening the ENU-mutagenised Tilling Library at the Hubrecht Institute, Utrecht. To identify her1 and her7 homozygous carriers, the 59 end of the relevant gene was amplified from genomic DNA from fin clips and analyzed by sequencing. The her1 hu2124 or her7 hu2526 alleles, respectively, were genotyped by PCR using the following primers: her1F 59-GAG AAG AAA CGG AGA GAC CGG-39 and her1R 59-CTT TAC ATA CGT GTA GAC AGG-39; her7F 59-GAT GAA AAT CCT GGC ACA GAC T-39 and her7R 59-TCT GAA TGC AGC TCT GCT CG-39. The amplicons were purified using AcroPrep TM 96 plates (PALL) and sequenced.
The bea tm98 mutant was used in this study [27].
Automated in situ hybridization was carried out following the protocol of Leve et al., 2001 [17] using a programmable liquid handling system (InsituPro, Intavis) described by Plickert et al., 1997 with a hybridization temperature of 65uC. Digoxygeninlabeled RNA probes were prepared using RNA labeling kits (Roche). Staining was performed with BM purple (Roche). Wholemount embryos were observed under a stereomicroscope (Leica) and digitally photographed with Leica DFC 480. Flat mounted embryos were analyzed with an Axioplan2 microscope connected to an Axiocam system (Zeiss).