Complexity of cis-regulatory organization of six3aduring forebrain and eye development in zebrafish
© Chao et al; licensee BioMed Central Ltd. 2010
Received: 13 November 2009
Accepted: 26 March 2010
Published: 26 March 2010
Six3a belongs to the SIX family of homeodomain proteins and is expressed in the most anterior neural plate at the beginning of neurogenesis in various species. Though the function of Six3a as a crucial regulator of eye and forebrain development has been thoroughly investigated, the transcriptional regulation of six3a is not well understood.
To elucidate the transcriptional regulation of six3a, we performed an in vivo reporter assay. Alignment of the 21-kb region surrounding the zebrafish six3a gene with the analogous region from different species identified several conserved non-coding modules. Transgenesis in zebrafish identified two enhancer elements and one suppressor. The D module drives the GFP reporter in the forebrain and eyes at an early stage, while the A module is responsible for the later expression. The A module also works as a repressor suppressing ectopic expression from the D module. Mutational analysis further minimized the A module to four highly conserved elements and the D module to three elements. Using electrophoresis mobility shift assays, we also provided evidence for the presence of DNA-binding proteins in embryonic nuclear extracts. The transcription factors that may occupy those highly conserved elements were also predicted.
This study provides a comprehensive view of six3a transcription regulation during brain and eye development and offers an opportunity to establish the gene regulatory networks underlying neurogenesis in zebrafish.
Zebrafish (Danio rerio) has long been an excellent vertebrate model organism for developmental biology [1–3] and is used to study the mechanisms of axis formation , endoderm differentiation [3, 5] and muscle development . It was also recently suggested to be an excellent model system for eye genetics . Development is controlled by the hierarchical regulation between signaling pathways and transcription factors. The basic operating principle for specifying a territory or regulatory state is controlled coordinately by transcription factors and signal transduction machinery in the cis-regulatory code. Gene regulatory networks (GRNs) for early embryogenesis have been established in sea urchin mesendoderm , Xenopus endoderm , Drosophila dorsal/ventral polarity  and zebrafish mesendoderm . It is essential to decode the cis-regulatory operation for the key transcription regulators contributing to the GRNs to elucidate the development of tissues and organs.
The Sine Oculis Homeobox (SIX) proteins share two evolutionarily conserved functional motifs in which the 115 amino acid SIX protein-protein interaction domain is located just upstream of the homeobox DNA binding domain [12, 13]. Disrupting the SIX domain or the homeodomain abolishes the ability of Six3a to induce rostral forebrain enlargement in zebrafish, implying that these domains are essential for six3 gene function . However, it also has been suggested that part of the biochemical and functional specificity between members of the SIX protein family is due to their non-conserved C-terminal segments .
Six3 is a member of the SIX family and is expressed in the most rostral portion of the brain in many animals. The first member of the SIX family, sine oculis (so), was identified in Drosophila , and later, six3 was discovered in many other species, including mouse , chick , zebrafish , medaka  and Xenopus . Together with the products of other homeobox genes, such as Otx  and Emx , Six3 plays a central role in the patterning of forebrain and eye development [22–25]. The function of Six3a in the development of the forebrain and eyes has been demonstrated in many species. Overexpression of six3 induces rostral forebrain enlargement in zebrafish and promotes ectopic lens formation  and ectopic retinal primordia formation in medaka  and Xenopus . The telencephalon of six3a and six3b double morphant embryos is markedly reduced in size, owing to impaired cell proliferation . In humans, study of severe malformation of the brain identified that mutations in the homeodomain of the Six3 gene may relate to holoprosencephaly [29–31]. Therefore, previous studies suggest a conserved function for Six3 in eye and forebrain development in metazoans.
The transcriptional regulation of six3a has been investigated previously (e.g., olSix3.2 in the developing medaka forebrain  and six3.1 in zebrafish retina and forebrain ). However, the minimal binding elements and the transcription factors were not identified in those studies. The negative regulator for forebrain specification, LOM4b, has also been found , but how it regulates six3a in terms of binding sites is not known. Several eye-field transcription factors (EFTFs) are expressed in a dynamic, overlapping pattern in the presumptive eye and forebrain. A genetic network regulating vertebrate eye field specification has been proposed in Xenopus using a combination of subsets of (EFTFs) and functional (inductive) analysis .
To delineate the network of trans-acting factors that control the evolutionarily conserved activity of six3a during forebrain development, we studied the function of the conserved non-coding regions in zebrafish six3a. Functionally important regions in the genome usually evolve more slowly than non-functional regions due to selective pressure. Alignment of conserved non-coding DNA sequences among different species using bioinformatics tools (e.g., the UCSC genome browser) provides a useful method for identifying the important regulatory elements of a gene (Chen and Blanchette, 2007; Werner et al., 2007). To facilitate the identification of functionally important elements in six3a, we applied the same strategy as for conserved non-coding regions. We took advantage of the power of computational analysis, the availability of the zebrafish genome sequence, and the efficiency of embryo transgenesis to analyze the regulatory control of one of the two six3 homologs in zebrafish, six3a. Elucidating the regulation of six3a and identifying the transcription factors responsible for correct expression of six3a helps to show how the gene is regulated and reveals the comprehensive gene regulatory networks (GRNs) for forebrain and eye specification across all vertebrates.
Identification of the cis-regulatory modules for six3aexpression in zebrafish
Both medaka and zebrafish have Six3 duplications. We aligned 14 protein sequences for the Six3 homologs (Additional file 1) and performed phylogenetic analysis (Additional file 2). From the protein sequence comparison, we found that zebrafish Six3a is more similar to medaka Six3.2 (85% identity) than to Six3.1 (75% identity). From the phylogenetic analysis, medaka Six3.2 and zebrafish Six3a were the closest homologs, while medaka Six3.1 was more closely related to zebrafish Six3b.
Brn3 and Pax6.1 binding sites on modules B and C contain no enhancer function in the forebrain or eyes up to 24 hpf
We first performed deletion analysis to search for the important elements. The genomic sequence spanning from -3782 to +110, containing modules A-D and Bp, was fused with GFP (Fig. 2-I). Transgenesis analysis showed that this 4-kb fragment (3087-Bp) expressed GFP accurately at all stages (Fig. 2-II-A). This is similar to the results by Bovolenta in medaka six3.2, in which the 4.5-kb fragment was responsible for six3.2 expression. Serial deletion mutants demonstrated that the genomic sequence that started from module D (1060-Bp, Fig. 3-II-C) exhibited a strong expression pattern similar to the 4-kb fragment (3087-Bp, Fig. 3-II-A). However, the sequence containing modules B and C (B:C-Bp) had no enhancer function in the forebrain or eyes up to 24 hpf (Fig. 2). The six3a promoter analysis by Wargelius et al. indicates that the retinal enhancer at 48 hpf is located in the 1.2-kb fragment between -3302 and -1409, which contains both Brn3 and Pax6.1 binding sites. Our B:C module (-2932 to ~ -1692) overlapped with this 1.2-kb fragment. It is possible that module B:C-Bp functions late only in the retina.
The 1060△42-Bp and 1060-Bp constructs (Fig. 3-II-C) displayed the same activity, indicating that the 42 bp in module D can be excised without change of function (the sequence is shown in Fig. 5 black box). Although 1060△42-Bp exhibits very strong activity, 898-bp shows much lower expression both early and late (Fig. 2), indicating that the sequence differences between these constructs (-1755 to -1594) contributed dramatically to the transcription of six3a. This sequence contains two DNA fragments that were shown to have no enhancement function in other constructs, part of the B:C module (-1755 to -1693) and the 42 bp in module D (-1646 to -1605). Thus, the most important elements in module D were minimized to the 46-bp region (-1692 to -1647). From electrophoresis mobility shift assays (EMSAs), we also found that the 46-bp region interacted with proteins present in the 24 hpf nuclear extracts (data shown in the EMSA section).
From our serial deletion mutants, the sequence between 681-Bp (Fig. 3-II-F) and 448-Bp (Fig. 3-II-G) seemed to contribute, to some degree, to GFP expression at 24 hpf. Half of the embryos bearing the 681-Bp construct expressed GFP at 24 hpf in the forebrain, but 448-Bp is very similar to Bp, and only 4% of embryos expressed GFP. This region actually contains the Box E discovered by Bovolenta's lab, which is a late neural and retinal enhancer in medaka six3.2 . Consistent with their discovery, our results suggested that a booster enhancer is located between 681-Bp and 448-Bp.
Identification of modules D and A, together with the basal promoter, as early and late enhancers
From our cross-species comparison, we identified ten conserved non-coding modules. We then tested the function of individual modules by transgenesis. First, we tried to identify the basal promoter. From our transgenesis results, we found that the sequence from -694 to +110 drove weak GFP expression restricted to 1-3 cells in the forebrain, which was only observed at 24 hpf. The expression rate of the Bp was about 10% (Fig. 3-III-A), suggesting that this region only has weak promoter activity. Thus, we named it "basal promoter", which contains the transcriptional start site (Fig. 3-I, black arrow). Wargelius et al. found that the zebrafish six3a sequence between 805 and 236 bp upstream of the ATG is required for early expression. Our results showed that a similar sequence located in the basal promoter (695 bp from the first exon) has a basal level of transcription activation up to 24 hpf, consistent with Wargelius's discovery. Conte et al. found that the sequence upstream of the translation start site in medaka six3.2 (Boxes I and L) is responsible for brain expression at later stages (stages 24 to 40) . Because the stages of our study were early compared to Wargelius's , we cannot conclude that the basal promoter does not have an enhancer function at later stages as discovered in medaka.
Embryos injected with modules 1, 3, B, C, 4, 5 or 6 individually showed expression patterns similar to that of the basal promoter. Only 5-22% of embryos carrying these constructs expressed low levels of GFP, as shown in Fig. 3-III-B for module 1. In spite of the sequence conservation, modules 1, 3, 4, 5 and 6 did not appear to contribute to the spatio-temporal control of six3a in zebrafish, at least up to 24 hpf. Interestingly, neither module B nor C was functional. Previously, a possible Pax6.1 binding site in module B and a putative Brn3b binding site in module C were shown to be important for regulating six3a in zebrafish, but this was not found in our study. Again, these results were consistent with the deletion analysis described earlier.
Module 2 of six3a increased the expression of GFP in embryos to 53% compared to 5% with Bp alone (Fig. 2-III-C); however, the results from eight batches of embryos were inconsistent (Additional file 5). Module A and D enhanced Bp expression dramatically, and the results from eight batches of experiments were consistent (Additional file 5). Fig. 3-III-B shows the strongest expression pattern of module 2. Half of the time, module 2 resulted in no expression. On the other hand, the expression pattern from modules A and D were always "strongly expressed in a large number of cells", but the expression from modules 1, 3, B, C, 4, 5 and 6 were always "expressed in only a few cells". Embryos injected with module A showed GFP expression in the forebrain and eyes from 14 hpf (data not shown). At 24 hpf, GFP was strongly expressed in the forebrain and retina (Fig. 3-III-D).
Because zebrafish six3a is expressed from 6 hpf and constitutively expressed in the neuroectoderm throughout anterior neuroectoderm formation , Bp, module 2 and A were not sufficient to control six3a expression. This suggests that an extra module is responsible for zebrafish six3a expression during early stages. Bovolenta and coworkers showed that a 4.5-kb region upstream of the medaka olsix3.2 transcription initiation site is responsible for olsix3.2 expression. They identified an early neural enhancer, Box D. We aligned the medaka and zebrafish genomes and found that Box D of Conte et al. was similar to the six3a genomic sequence -1754 to -1571 bp in zebrafish; we named it module D. Module D significantly activated transcription (Fig. 3-III-E, F), and it was the only module that drove expression in early stages (Fig. 3-III-G for 8 hpf, 3III-H for 11 hpf). Our result showed that the early neural enhancer in medaka olsix3.2 also appears in a similar position and exhibits an enhancer function for six3a expression in the early zebrafish neuroectoderm. At 24 hpf, modules D and A displayed an enhancer effect when linked to the basal promoter, except the expression domain from D-Bp (Fig. 3-III- F, I), which seemed to extend to a more ventral region than A-Bp (Fig. 3-III-D, L). This suggests that a suppressor function is missing from D-Bp. More related data is shown in the "silencer function of module A on the suppression of the ectopic expression from module D" section.
Deletion analysis of module A identified several activator binding sites
To understand the transcriptional regulation of six3a in detail, we dissected module A by using deletions and mutations to identify the minimal sequence and the bound transcription factors required for enhancer function. Comparison of module A sequences among zebrafish, Xenopus, opossum, Tetraodon, human and mouse revealed four highly conserved elements with 100% similarity among those species (Fig. 4, bottom). We first generated five deletion constructs by PCR (Fig. 4-I). The A462-Bp construct was 462 bp long and lacked 311 bp from the 5'-end of module A. In the A433-Bp construct, 29 bp were deleted, including the first highly conserved element "TCATTAA". In the A415-Bp construct, the second highly conserved element "CGCTAACAA" was deleted. Both of the "AAATGC" and "CAGCT" elements were deleted in the D392-Bp construct. A 22-bp, highly conserved element was deleted in the A315-Bp construct. These constructs were separately microinjected into zebrafish embryos at the one-cell stage, and the zebrafish were analyzed for GFP expression. A462-Bp mutants showed GFP expression in the forebrain and retina at 24 hpf (Fig. 4-III-A). Furthermore, deletion of the first element (A433-Bp) slightly decreased GFP expression in the forebrain (Fig. 4-III-B). The deletion of the second highly conserved element (A415-Bp) caused decreased GFP expression in the eye and forebrain (Fig. 4-III-C). A construct containing none of the highly conserved elements (A392) showed very low expression in both the forebrain and eye, suggesting the importance of those elements (Fig. 4-III-D).
We further deleted the 3'-end of module A to find the minimal active elements. The 3' deletion mutant A△ 372-Bp, which contained all four highly conserved elements, had very low GFP expression and was expressed in only 30% of embryos. It was similar to the 5' deletion mutant A392-Bp. Although A392-Bp and A△372-Bp each contained half of module A, neither could drive GFP expression, indicating the possibility of cooperation of the transcriptional factors bound to each of these two DNA fragments. We therefore examined the functions of the four highly conserved elements plus the downstream sequences up to A△298 (Fig. 4).
The importance of the four highly conserved elements was demonstrated by the serial deletion mutants, in which we successively removed more sequence from the A△298-Bp construct. Comparison of the results between A462△298-Bp and A433△298-Bp (Fig. 4-III-E, F) confirmed the importance of the enhancer function in the "TCATTAA" sequence. Analysis of the results between A462△298-Bp and A462△298-Bp-M (Fig. 4-III-F, G) corroborated the importance of "CGCTAACAA" for forebrain and eye development. The GFP expression pattern between A462△298-Bp-M and A392△298-Bp (Fig. 4-III-G, H) supported the activation function of two highly conserved elements "AAATGC" and "CAGCT". Duplication of the 22-bp, highly conserved element showed no activation function at all. This again indicated that multiple elements must work together, and implied that transcription factor occupancy of these cis-elements has a synergistic effect.
Deletion analysis in module D elucidated multiple activator binding elements
As mentioned previously, module D is the only module that drives expression in early stages of zebrafish development. Therefore, we next dissected module D to find the minimal regulatory elements. Alignment of this module among Tetraodon, fugu and medaka revealed clusters of conserved regions (Fig. 5-I). The comparison of the results between D141-Bp and D105-Bp showed decreased GFP expression in the eyes and forebrain in D105-Bp-injected embryos (Fig. 5-III-B, C). Further removal of nucleotides between D105-Bp and D69-Bp decreased GFP expression at both 11 and 24 hpf (Fig. 5-III-C, D). The deletion to a shorter construct, D30-Bp, which contained a highly conserved 28-bp region, showed GFP expression in the eyes and forebrain (Fig. 5-III-E). When we duplicated this element (D28(2)-Bp), it drove GFP not only at 24 hpf (Fig. 5-III-F) but also at 8 and 11 hpf. Our results indicated that the highly conserved 28-bp sequence might be bound by some early transcription factor(s). Mutation of six nucleotides in this 28-bp element from "ctaatt" to "AGCCGG" abolished the enhancer activity, indicating that the occupancy of transcription factor(s) at this core sequence was responsible for activation. We then used the ctaatt sequence to search the transcription factor binding site database PROMO http://alggen.lsi.upc.es/cgi-bin/promo_v3/promo/promoinit.cgi?dirDB=TF_8.3 and found that POU1F1a binds to "CTAAT" and "ATTAC" in the 28-bp highly conserved element. Both sequences were destroyed in our D28M(2)-Bp construct.
In contrast to module A, the multiple elements in module D seem to function independently. Constructs D184△74-Bp (Fig. 5-III-G) and D69 (Fig. 5-III-E) represent the 5' and 3' halves of module D; both drove expression in the neuroectoderm, although at a lower level compared to the full-length construct. However, comparing the three deletion mutants D184△30-Bp (Fig. 5-III-H), D105△30-Bp (Fig. 5-III-K) and D69△30-Bp (Fig. 5-III-O), we found that in the absence of the 28-bp region (removed in △30), the sequence between D184 and D105 was essential for expression. Previous deletion mutants indicated that the most important elements in module D included the 46-bp region between -1692 and ~ -1647, which co-localizes exactly with the sequence between D184 and D105.
Silencer function of module A for the suppression of the ectopic expression of module D
EMSA for modules D and A
We found multiple elements in module A during our functional study. However, synergy between those elements is required for activation, and this may be due to cooperativity in DNA binding. We used EMSAs to test transcription factor binding to module A (Fig. 7-II) and found that a large complex was formed with the probe A462 298 in a dose-dependent manner (Fig. 7-II). The complex was reduced during competition with unlabelled DNA (Fig. 7-II). However, no band shift was detected when we used smaller probes (data not shown). It is likely that those elements bind transcription factors in the 24 h nuclear extract with low affinity. Thus, multiple elements are required to achieve a higher level of occupancy on the DNA and to activate six3a transcription.
Two evolutionarily conserved modules and many minimal functional elements work together to achieve correct six3aexpression
Six3a is an important transcription factor expressed in the presumptive brain and eye territory and continuously expressed in the forebrain and retina. Forebrain development is evolutionarily conserved among vertebrates and so is the underlying mechanism that operates the GRNs. By analyzing the transcriptional regulation of six3a, we not only uncovered the evolutionary cis-regulatory code responsible for the expression of six3a, but also drew a preliminary picture of how those transcription factors specify forebrain development.
We found that the early enhancer module D activated six3a expression as early as 8 hpf. However, expression extended to the ventral area. Module A was necessary to restrict this expression. Bovolenta's lab found that Box D (similar to our module D) is an early enhancer, and Box A (similar to our module A) is a silencer . The interactions between these two Boxes are necessary for correct six3.2 expression in medaka. They also found that Box E is a late neural and retinal enhancer of medaka six3.2. Our results are consistent with their findings and suggest that a booster enhancer is located between 681-Bp and 448-Bp (similar to Box E). However, it is a very weak enhancer compared to modules A and D.
There are also some inconsistencies between our data and previous reports on six3a. We found that module A acted as an enhancer at 24 hpf, and the basal promoter had no enhancer function, as suggested by the Bovolenta  and Fjose  groups. Because the developmental stages of our study are early compared to the aforementioned studies, we cannot conclude that the basal promoter does not have an enhancer function at later stages, as they discovered. However, we have injected the Bp-GFP construct and observed expression at later stages (48 and 72 hpf), but we still do not detect GFP expression (data not shown). This preliminary result indicated that this region (the basal promoter of zsix3a and Box I and L of olsix3.2) does not have enhancer activity. Also, module B:C had no enhancer function, as proposed by Wargelius et al. However, we suggest that POU and Pax6.1 activate six3a expression by binding to modules A and D.
We not only demonstrated the functionality of these evolutionarily conserved elements using transgenesis in zebrafish, but we also made many deletion and mutation constructs to identify the minimal elements necessary for regulation in order to predict the transcription factors responsible for six3a regulation. EMSAs, using nuclear extracts from zebrafish embryos, revealed binding proteins that occupied these functional elements. The predictions of transcription factors are given in Fig. 9-I, and the working model for six3a regulation is shown in Fig. 9-II.
The POU domain transcription factor is important for activating six3aexpression
We searched for transcription factor binding sites on the evolutionarily conserved and functional minimal elements that we identified. The 28-bp, conserved element has two HNF1Bs (GCTAATTA and TAATTACT), two POU domain transcription factor (CTAAT and ATTAC), one C/EBP beta (TTGCTTA) and one PR-alpha (AGAACACAA) binding sites. Mutational analysis showed that the POU domain transcription factor binding site was responsible for the enhancer function; mutation of the six-nucleotide sequence containing this site eliminated the enhancer function. In module A, four highly conserved elements work synergistically as an enhancer. The first element (TCATTAA) has a binding site for the POU domain transcription factor (ATTAA).
Two POU domain transcription factors are expressed in zebrafish brain and eyes: pou3f3a (POU class 3 homeobox 3a; previous names:zp12, wu:fc33a11, pou12 and brn1.1), which is predominantly expressed in the central nervous system ; and pou3f3b (POU class 3 homeobox 3b; previous names:zp23pou, wu:fb92g06( 1 ), zp23, fb92g06, zfpou1( 1 ), pou1 and pou23), which is expressed in developing neural tissues . They are expressed at the right time and place for the activator to bind to six3a modules A and D. It is likely that Pou3f3a and Pou3f3b are important in regulating six3a expression by binding to modules A and D. Although we did not find that the proposed Brn3 (which is a POU domain transcription factor) binding site on module B had any activation function, as described previously , we suggest that Pou3f3a and Pou3f3b activate six3a by binding to the 28-bp element in module D and the first highly conserved element in module A.
Pax6a binds to six3amodule A to activate expression in the eyes and brain
In module A, the second element (GCGTAACAA) has a binding site for a paired box protein (GCGTAAC), and the third (AAATGC) also has a binding site for a paired box protein (AAATGC). In module D, in addition to the 28-bp element, two highly conserved elements show enhancer function and are bound by proteins in the 24 h nuclear extract. The EMSA probe #4 (CGCTCCCTGCTGATT) has a binding site for a paired box protein (CGCTCCC).
Many paired box proteins in zebrafish are likely important in regulating six3a expression. Among these, pax6a (paired box gene 6a), which is expressed in the anterior neural plate at the bud stage  and in the brain and immature eye at the prim5 stage, is the most likely activator of module D expression. The regulatory mechanisms governed by pax6 are evolutionarily conserved among teleosts and mammals . Although we did not find that the proposed Pax6 binding site on module C had any activation function, as proposed previously , we suggest that Pax6a activates six3a by binding to the 2nd and 3rd highly conserved elements in module A.
Other possible transcription factors regulating six3aexpression
The 28-bp conserved element of module D has two homeobox protein binding sites (TAATTA), and the fourth element of module A (GACAGCT) has a binding site for HoxA3 (GACAG). It is possible that a homeobox protein, Gbx1 (gastrulation brain homeobox 1), activates six3a expression as early as 5 hpf. Our unpublished real-time RT-PCR data indicated that knockdown of Gbx1 expression by morpholino injection decreased six3a expression at 5 hpf. Gbx1 is expressed in the presumptive brain at the 75% epiboly stage . However, if Gbx1 expression extends to the hindbrain neural plate, neural plate and ventral mesoderm , early expression will result in an ectopic pattern unless there is a repression system. Another homeobox transcription factor candidate for regulating six3 expression is the LIM homeobox gene, LMO4b, which has been proposed as a negative regulator of forebrain growth acting via the restriction of six3 expression during early segmentation stages . It is possible that LMO4b negatively regulates six3a by binding to modules A and D together.
We found that the second conserved element (CGTAACA) of module A shows similarity to the forkhead transcription factor binding site. The fifth EMSA probe of module D (CATAGAACAGAGCAGTGAAAGCTAG) has a binding site for the forkhead transcription factor (TAGAACA). Thus, the forkhead transcription factor may also activate module D expression in the presumptive brain. Foxa2, also called hepatocyte nuclear factor 3β (hnf-3β), is a member of the forkhead box transcription factor protein family. In situ hybridization shows that Foxa2 is first expressed on the dorsal side of the hypoblast just before gastrulation . Later, at 8 hpf, it is expressed in the endoderm and axial mesoderm. In adult fish, Foxa2 is expressed in multiple territories, including the gut, liver and pancreas of the endoderm and the floor plate of the ectoderm . Thus, during zebrafish development, Foxa2 is expressed in all three germ layers. Its role in the ventral central nervous system was proposed by Norton et al. Nodal signaling from the notochord induces Foxa2 expression in the medial floor plate (MFP), and Foxa2 activates downstream genes (e.g., Ntn1b, Shha and Shhb) that are required for MFP maintenance and differentiation.
One of the forkhead transcription factors, zfoxl1, is strongly expressed in neural tissues, such as the midbrain, hindbrain and the otic vesicle at the early embryonic stage. It is a novel regulator of neural development that acts by suppressing shh expression . Our results and previous data from Conte et al.  indicated that the suppressor function mediated by module A eliminates the ectopic expression carried by module D. It is possible that zfoxl1 binds to module A via the 2nd conserved element and to module G through the 28-bp element to execute its repressor function over module D.
In this study, we identified two enhancers in the regulation of zebrafish six3a. Those modules are evolutionarily conserved across all vertebrates. Module D was also identified as an enhancer in medaka. Module A not only exhibited an enhancer function but also had a suppressor effect, eliminating the ectopic expression driven by module D. Further analysis of the minimal binding elements in both modules demonstrated that the multiple elements in module A work synergistically in binding to DNA, while the 45-bp and 28-bp elements in module D have strong DNA-protein interaction activity and are functionally important. Possible activators that bind to modules D and A are Pou3f3a, Pou3f3b, Pax6a, Gbx1 and Foxa2. Possible repressors binding to module D and A are LMO4 and zfoxl1. The discovery of these elements and transcription factor binding sites provides a new insightful view of the entire interplay of transcription factors in the GRNs for forebrain development.
Zebrafish husbandry, experimentation and care/welfare
AB strain D. rerio fish were purchased from the Zebrafish International Resource Center (ZIRC), Oregon. Fish were maintained at 28°C in our zebrafish facility, a continuous flow-through system, with a 14 h light/10 h dark cycle. The developmental stages were as described previously . Embryos from naturally spawning AB strain zebrafish were used in this study. To generate embryos for injection, male and female fish were placed in a 1 L fish tank with inner mesh and a divider the night before injection. Zebrafish embryos were obtained from naturally spawning adults stimulated by light and by removing the divider. The embryos were stored at 28.5 before and after microinjection.
All experiments involving zebrafish were conducted according to the guidelines of Institutional Animal Care and Use Committee (IACUC) of the National Health Research Institutes (NHRI). The animal protocol involving zebrafish was approved by IACUC of NHRI; the approved protocol numbers are NHRI-IACUC-095050-A, 096037-A, 098017-A and 098087 under the name of Dr. Yuh, Chiou-Hwa, who is the corresponding author.
DNA constructs and site-directed mutagenesis
The zebrafish six3a DNA was obtained from zebrafish genomic DNA or a six3a BAC clone (DKEY-254J21) purchased from the BACPAC Resource Center (BPRC) at the Children's Hospital, Oakland Research Institute, Oakland, CA. The genomic sequence was obtained from the Ensembl Genome Browser Database http://www.ensembl.org/Danio_rerio/index.html. Cis-regulatory elements in the zebrafish six3a genome, which are conserved among several species, were identified using the UCSC Genome Browser Database http://genome.ucsc.edu/cgi-bin/hgGateway. In order to obtain the cis-regulatory elements for our constructs, we used PCR to amplify the BAC DNA by specific forward and reverse primers containing restriction enzyme sites for ligation into the pEGFP-N1 vector (BD Biosciences, San Jose, CA). After digesting with restriction enzymes and purifying from the gel, the DNA was inserted into an EGFP-N1 vector. The primer sequences used to amplify the DNA are shown in [additional data 1], and the restriction enzyme sites are underlined.
Site-directed mutagenesis was performed using a QuickChange Site-Directed Mutagenesis kit (Stratagene). This method utilizes PfuTurbo DNA polymerase, which replicates both plasmid strands with high fidelity and without displacing the mutant oligonucleotide primers. A 50-ng sample of template DNA was used in each reaction. After temperature cycling (95°C for 30 s, then 18 cycles of 95°C for 30 s, 55°C for 1 min and 68°C for 12 min), the product was treated with DpnI. The nicked vector DNA containing the desired mutations was then transformed into XL2-Blue ultracompetent cells. The oligonucleotides used to generate the mutations are given in Additional file 3. The underlined sequences are the 5' and 3' parts of the primer sequences. The internal sequences were deleted after site-directed mutagenesis.
Microinjection and microscopic photography
Embryos were injected using either DNA from the construct, which was PCR-amplified with the specific forward and reverse primers (EGFP-N1-r-poly(A) containing the SV40 poly(A) signal), or the linearlized constructs formed by digestion with XhoI. For microinjection, the morpholinos or DNAs were prepared in PBS with 0.05% (w/v) phenol red. Embryos were injected at the one-cell stage with 2.3 nl of 25 ng/μl six3a-green fluorescent protein (GFP) constructs by Nanoject (Drummond Scientific Co., Broomall, PA). Embryos were collected at different stages for GFP visualization and photography using a Nikon Optiphot-2 upright microscope with the episcopic-fluorescence attachment EFD-3 (Nikon Inc., Melville, NY) coupled with a DXM1200 Nikon Digital Camera.
Radioactively labeled DNA probes were designed based on the functional assay results. The G184 EMSA probe was generated by PCR using the D184-F and D184-R primers and was digested with XhoI to produce 5' protruding ends for labeling. The D184△74 probe was generated by PCR using the D184-F and D184△74-R and was digested with XhoI to produce 5' protruding ends for labeling. The D69 probe was generated by PCR using the D69-F and D184-R primers and was digested with XhoI and SacI to produce 5' protruding ends for labeling. The #1-#5 and D28 probes in module D were synthesized oligonucleotides and annealed to form double strands using the method described below. The module A EMSA probe was generated by PCR using the A462-F and AΔ298-R primers and was digested with XhoI and KpnI to produce 5' protruding ends for labeling.
The oligonucleotides used for EMSA and PCR primers for generating the DNA fragments for EMSA are listed in Additional file 1. The 5' protruding sequence on the annealed double-stranded oligonucleotide probes, 5'ATCG, that generated by restriction enzyme digestion, was used for Klenow labeling by filling in with [α32P] dCTP. The double-stranded DNA was annealed in 0.1 M NaCl buffered with 10 mM Tris (pH 8.0) at a 100 μM final concentration. The double-stranded DNA was labeled with Klenow polymerase, and the free un-incorporated nucleotides were removed with a G-25 Sephadex spin column. Reactions were prepared per tube as follows: to 2 μl nuclear extract or buffer C , 1 μl specific competitor (100 ng/Al) or H2O was added, plus 1 μl labeled probe (1 × 105cpm/μl), 16 μl pre-mix binding buffer (containing nonspecific competitors [polyd(I)-d(C)/polyd(I)/d(C), 10 μg]), and 1× binding buffer [20 mM HEPES-KOH (pH 7.9), 75 mM KCl, 0.5 mM DTT]. The tubes were incubated on ice for 15 min before loading onto a 6% native polyacrylamide/TBE gel. The gel was run at 150 V for 2.5 h, placed on 3 MM paper, wrapped with plastic wrap, dried for 60 min at 80°C in a vacuum gel dryer and subjected to phosphoimager exposure.
We thank Mr. Chen, Wen-Chuan for maintaining the zebrafish stocks. Funding support from National Health Research Institute (MG-094-PP-14, MG-095-PP-08, MG-096-PP-05, MG-097-PP-07, MG-098-PP-06, and MG-099-PP-06) and NRPGM grant (NSC 97-3112-B-400-008) to Dr Yuh Chiou-Hwa is gratefully acknowledged.
- Schoenebeck JJ, Yelon D: Illuminating cardiac development: Advances in imaging add new dimensions to the utility of zebrafish genetics. Semin Cell Dev Biol. 2007, 18: 27-35. 10.1016/j.semcdb.2006.12.010.PubMed CentralView ArticlePubMedGoogle Scholar
- Zacchigna S, Ruiz de Almodovar C, Carmeliet P: Similarities between angiogenesis and neural development: what small animal models can tell us. Curr Top Dev Biol. 2008, 80: 1-55. full_text.View ArticlePubMedGoogle Scholar
- Zorn AM, Wells JM: Molecular basis of vertebrate endoderm development. Int Rev Cytol. 2007, 259: 49-111. full_text.View ArticlePubMedGoogle Scholar
- Schier AF, Talbot WS: Molecular genetics of axis formation in zebrafish. Annu Rev Genet. 2005, 39: 561-613. 10.1146/annurev.genet.37.110801.143752.View ArticlePubMedGoogle Scholar
- Alexander J, Stainier DY: A molecular pathway leading to endoderm formation in zebrafish. Curr Biol. 1999, 9: 1147-1157. 10.1016/S0960-9822(00)80016-0.View ArticlePubMedGoogle Scholar
- Ochi H, Westerfield M: Signaling networks that regulate muscle development: lessons from zebrafish. Dev Growth Differ. 2007, 49: 1-11.View ArticlePubMedGoogle Scholar
- Fadool JM, Dowling JE: Zebrafish: a model system for the study of eye genetics. Prog Retin Eye Res. 2008, 27: 89-110. 10.1016/j.preteyeres.2007.08.002.PubMed CentralView ArticlePubMedGoogle Scholar
- Davidson EH, Rast JP, Oliveri P, Ransick A, Calestani C, et al: A genomic regulatory network for development. Science. 2002, 295: 1669-1678. 10.1126/science.1069883.View ArticlePubMedGoogle Scholar
- Loose M, Patient R: A genetic regulatory network for Xenopus mesendoderm formation. Dev Biol. 2004, 271: 467-478. 10.1016/j.ydbio.2004.04.014.View ArticlePubMedGoogle Scholar
- Stathopoulos A, Van Drenth M, Erives A, Markstein M, Levine M: Whole-genome analysis of dorsal-ventral patterning in the Drosophila embryo. Cell. 2002, 111: 687-701. 10.1016/S0092-8674(02)01087-5.View ArticlePubMedGoogle Scholar
- Chan TM, Longabaugh W, Bolouri H, Chen HL, Tseng WF, et al: Developmental gene regulatory networks in the zebrafish embryo. Biochim Biophys Acta. 2009, 1789: 279-298.View ArticlePubMedGoogle Scholar
- Kawakami K, Ohto H, Takizawa T, Saito T: Identification and expression of six family genes in mouse retina. FEBS Lett. 1996, 393: 259-263. 10.1016/0014-5793(96)00899-X.View ArticlePubMedGoogle Scholar
- Tessmar K, Loosli F, Wittbrodt J: A screen for co-factors of Six3. Mech Dev. 2002, 117: 103-113. 10.1016/S0925-4773(02)00185-5.View ArticlePubMedGoogle Scholar
- Kobayashi M, Toyama R, Takeda H, Dawid IB, Kawakami K: Overexpression of the forebrain-specific homeobox gene six3 induces rostral forebrain enlargement in zebrafish. Development. 1998, 125: 2973-2982.PubMedGoogle Scholar
- Weasner BP, Kumar JP: The non-conserved C-terminal segments of Sine Oculis Homeobox (SIX) proteins confer functional specificity. Genesis. 2009, 47: 514-523. 10.1002/dvg.20517.PubMed CentralView ArticlePubMedGoogle Scholar
- Oliver G, Mailhos A, Wehr R, Copeland NG, Jenkins NA, et al: Six3, a murine homologue of the sine oculis gene, demarcates the most anterior border of the developing neural plate and is expressed during eye development. Development. 1995, 121: 4045-4055.PubMedGoogle Scholar
- Bovolenta P, Mallamaci A, Puelles L, Boncinelli E: Expression pattern of cSix3, a member of the Six/sine oculis family of transcription factors. Mech Dev. 1998, 70: 201-203. 10.1016/S0925-4773(97)00183-4.View ArticlePubMedGoogle Scholar
- Loosli F, Koster RW, Carl M, Krone A, Wittbrodt J: Six3, a medaka homologue of the Drosophila homeobox gene sine oculis is expressed in the anterior embryonic shield and the developing eye. Mech Dev. 1998, 74: 159-164. 10.1016/S0925-4773(98)00055-0.View ArticlePubMedGoogle Scholar
- Zhou X, Hollemann T, Pieler T, Gruss P: Cloning and expression of xSix3, the Xenopus homologue of murine Six3. Mech Dev. 2000, 91: 327-330. 10.1016/S0925-4773(99)00270-1.View ArticlePubMedGoogle Scholar
- Suda Y, Nakabayashi J, Matsuo I, Aizawa S: Functional equivalency between Otx2 and Otx1 in development of the rostral head. Development. 1999, 126: 743-757.PubMedGoogle Scholar
- Patarnello T, Bargelloni L, Boncinelli E, Spada F, Pannese M, et al: Evolution of Emx genes and brain development in vertebrates. Proc Biol Sci. 1997, 264: 1763-1766. 10.1098/rspb.1997.0244.PubMed CentralView ArticlePubMedGoogle Scholar
- Seo HC, Drivenes , Ellingsen S, Fjose A: Expression of two zebrafish homologues of the murine Six3 gene demarcates the initial eye primordia. Mech Dev. 1998, 73: 45-57. 10.1016/S0925-4773(98)00028-8.View ArticlePubMedGoogle Scholar
- Davis RJ, Shen W, Heanue TA, Mardon G: Mouse Dach, a homologue of Drosophila dachshund, is expressed in the developing retina, brain and limbs. Dev Genes Evol. 1999, 209: 526-536. 10.1007/s004270050285.View ArticlePubMedGoogle Scholar
- Gehring WJ: New perspectives on eye development and the evolution of eyes and photoreceptors. J Hered. 2005, 96: 171-184. 10.1093/jhered/esi027.View ArticlePubMedGoogle Scholar
- McCollum CW, Amin SR, Pauerstein P, Lane ME: A zebrafish LMO4 ortholog limits the size of the forebrain and eyes through negative regulation of six3b and rx3. Dev Biol. 2007, 309: 373-385. 10.1016/j.ydbio.2007.07.004.PubMed CentralView ArticlePubMedGoogle Scholar
- Loosli F, Winkler S, Wittbrodt J: Six3 overexpression initiates the formation of ectopic retina. Genes Dev. 1999, 13: 649-654. 10.1101/gad.13.6.649.PubMed CentralView ArticlePubMedGoogle Scholar
- Bernier G, Panitz F, Zhou X, Hollemann T, Gruss P, et al: Expanded retina territory by midbrain transformation upon overexpression of Six6 (Optx2) in Xenopus embryos. Mech Dev. 2000, 93: 59-69. 10.1016/S0925-4773(00)00271-9.View ArticlePubMedGoogle Scholar
- Ando H, Kobayashi M, Tsubokawa T, Uyemura K, Furuta T, et al: Lhx2 mediates the activity of Six3 in zebrafish forebrain growth. Dev Biol. 2005, 287: 456-468. 10.1016/j.ydbio.2005.09.023.View ArticlePubMedGoogle Scholar
- Wallis DE, Roessler E, Hehr U, Nanni L, Wiltshire T, et al: Mutations in the homeodomain of the human SIX3 gene cause holoprosencephaly. Nat Genet. 1999, 22: 196-198. 10.1038/9718.View ArticlePubMedGoogle Scholar
- Domene S, Roessler E, El-Jaick KB, Snir M, Brown JL, et al: Mutations in the human SIX3 gene in holoprosencephaly are loss of function. Hum Mol Genet. 2008, 17: 3919-3928. 10.1093/hmg/ddn294.PubMed CentralView ArticlePubMedGoogle Scholar
- Lacbawan F, Solomon BD, Roessler E, El-Jaick K, Domene S, et al: Clinical spectrum of SIX3-associated mutations in holoprosencephaly: correlation between genotype, phenotype and function. J Med Genet. 2009, 46: 389-398. 10.1136/jmg.2008.063818.PubMed CentralView ArticlePubMedGoogle Scholar
- Conte I, Bovolenta P: Comprehensive characterization of the cis-regulatory code responsible for the spatio-temporal expression of olSix3.2 in the developing medaka forebrain. Genome Biol. 2007, 8: R137-10.1186/gb-2007-8-7-r137.PubMed CentralView ArticlePubMedGoogle Scholar
- Wargelius A, Seo HC, Austbo L, Fjose A: Retinal expression of zebrafish six3.1 and its regulation by Pax6. Biochem Biophys Res Commun. 2003, 309: 475-481. 10.1016/j.bbrc.2003.07.010.View ArticlePubMedGoogle Scholar
- Zuber ME, Gestri G, Viczian AS, Barsacchi G, Harris WA: Specification of the vertebrate eye by a network of eye field transcription factors. Development. 2003, 130: 5155-5167. 10.1242/dev.00723.View ArticlePubMedGoogle Scholar
- Spaniol P, Bornmann C, Hauptmann G, Gerster T: Class III POU genes of zebrafish are predominantly expressed in the central nervous system. Nucleic Acids Res. 1996, 24: 4874-4881. 10.1093/nar/24.24.4874.PubMed CentralView ArticlePubMedGoogle Scholar
- Matsuzaki T, Amanuma H, Takeda H: A POU-domain gene of zebrafish, ZFPOU1, specifically expressed in the developing neural tissues. Biochem Biophys Res Commun. 1992, 187: 1446-1453. 10.1016/0006-291X(92)90464-V.View ArticlePubMedGoogle Scholar
- Zou S, Kamei H, Modi Z, Duan C: Zebrafish IGF genes: gene duplication, conservation and divergence, and novel roles in midline and notochord development. PLoS One. 2009, 4: e7026-10.1371/journal.pone.0007026.PubMed CentralView ArticlePubMedGoogle Scholar
- Lakowski J, Majumder A, Lauderdale JD: Mechanisms controlling Pax6 isoform expression in the retina have been conserved between teleosts and mammals. Dev Biol. 2007, 307: 498-520. 10.1016/j.ydbio.2007.04.015.View ArticlePubMedGoogle Scholar
- Rhinn M, Lun K, Amores A, Yan YL, Postlethwait JH, et al: Cloning, expression and relationship of zebrafish gbx1 and gbx2 genes to Fgf signaling. Mech Dev. 2003, 120: 919-936. 10.1016/S0925-4773(03)00135-7.View ArticlePubMedGoogle Scholar
- Sprague J, Clements D, Conlin T, Edwards P, Frazer K, et al: The Zebrafish Information Network (ZFIN): the zebrafish model organism database. Nucleic Acids Res. 2003, 31: 241-243. 10.1093/nar/gkg027.PubMed CentralView ArticlePubMedGoogle Scholar
- Strahle U, Blader P, Ingham PW: Expression of axial and sonic hedgehog in wildtype and midline defective zebrafish embryos. Int J Dev Biol. 1996, 40: 929-940.PubMedGoogle Scholar
- Norton WH, Mangoli M, Lele Z, Pogoda HM, Diamond B, et al: Monorail/Foxa2 regulates floorplate differentiation and specification of oligodendrocytes, serotonergic raphe neurones and cranial motoneurones. Development. 2005, 132: 645-658. 10.1242/dev.01611.PubMed CentralView ArticlePubMedGoogle Scholar
- Nakada C, Satoh S, Tabata Y, Arai K, Watanabe S: Transcriptional repressor foxl1 regulates central nervous system development by suppressing shh expression in zebra fish. Mol Cell Biol. 2006, 26: 7246-7257. 10.1128/MCB.00429-06.PubMed CentralView ArticlePubMedGoogle Scholar
- Kimmel CB, Ballard WW, Kimmel SR, Ullmann B, Schilling TF: Stages of embryonic development of the zebrafish. Dev Dyn. 1995, 203: 253-310.View ArticlePubMedGoogle Scholar
- Calzone FJ, Theze N, Thiebaud P, Hill RL, Britten RJ, et al: Developmental appearance of factors that bind specifically to cis-regulatory sequences of a gene expressed in the sea urchin embryo. Genes Dev. 1988, 2: 1074-1088. 10.1101/gad.2.9.1074.View ArticlePubMedGoogle Scholar