Promotion of Neurogenesis by Brg1 in Developing Zebrafish
One known function of the Brg1 chromatin-remodeling factor is its interaction with β-catenin and activation of transcription factor TCF/LEF and its downstream target genes (Barker et al.,2001). TCF can also act as a transcriptional repressor of Wnt target genes when it is associated with Groucho (Roose et al.,1998) and CtBP (Brannon et al.,1999). Loss of T cell factor 3 (TCF3), a member of the TCF/LEF family (hdl mutant), leads to complete loss of the eye, forebrain, and part of the midbrain structures (Kim et al.,2000). In addition, Brg1 has been shown to be involved in differentiation of pro-neuronal bHLH transcription factors Ngnr1 and NeuroD in Xenopus (Seo et al.,2005). To determine the role of Brg1 in early zebrafish development, we analyzed the Brg1 mRNA expression pattern using in situ hybridization of one-cell to 24 hr post-fertilization (hpf) stage embryos (Fig. 1). Brg1 transcripts are present at the one-cell stage, suggesting that Brg1 is a maternally expressed gene (Gregg et al.,2003). This was confirmed in Brg1-null yng mutants derived from heterozygote intercrosses, which were found to be less affected due to maternal expression of Brg1 before the onset of zygotic gene activity (see below). Brg1 is expressed ubiquitously until 24 hpf and then becomes more restricted to the anterior region of the embryo. The expression of Brg1 mRNA is strongest in the telencephalon, cerebellum, and hindbrain regions (Fig. 1).
Figure 1. Brg1 is expressed during early embryonic development. A–D: Whole mount in situ hybridization was performed to detect Brg1 expression in wild-type zebrafish embryos collected at the 1-cell stage, 8-cell stage, shield stage (6 hpf), and 24 hpf. Embryos are dorsal; phenotypes were more than 90%; n = 20 for each embryonic stage. Original magnification 40×.
Download figure to PowerPoint
We then analyzed the role of Brg1 during embryogenesis following knockdown of Brg1 by injecting Brg1-specific morpholino oligonucleotides (Brg1-MO) into one- to two-cell-stage embryos. Approximately 80% of the Brg1-MO-injected embryos displayed defects by 24–48 hpf (Fig. 2). Defects in the neural tube were visible as indicated by the absence of the mid/hind-brain boundary (compare Fig. 2C and D, arrows). Evidence of increased cell apoptosis was clearly visible by fluorescence miscroscopy following acridine orange staining, specifically in the anterior portion of the embryos (compare Fig. 2E and F). All embryos that survived to reach 48 to 96 hpf exhibited reduced head size, swelling around the pericardium (Link et al.,2000), and a curved body axis (Fig. 2G–J). The brg1-homozygous (yng) mutants began to show defects at around 28 hpf, and exhibited similar phenotypes (such as a curved body axis) as the Brg1-MO-injected embryos; they died by 6–7 days post-fertilization (dpf) most likely due to cardiovascular defects (Link et al.,2000).
Figure 2. Brg1 expression is critical in neurodevelopment and the proper formation of body axis. A,B: One- to two-cell-stage embryos were injected with control- (A, control) or Brg1-specific MOs (B, Brg1-MO) (10 ng/embryo). Embryos were photographed at 25 hpf. Original magnification 100×. C,D: Dorsal view showing mid/hind-brain boundary under normal conditions (C) or aberrant neural tube development in Brg1-MO-injected embryos at 25 hpf (D). Original magnification 100×. E,F: One- to two-stage embryos were injected with Acridine Orange. Embryos were injected with control-MO (I) or Brg1-MO (J) and photographed at 25 hpf. Morphant phenotypes were more than 90%, n = 30. G–J: Control-MO-injected or Brg1-MO-injected embryos at 48 hpf (E and F) and 4 dpf (G and H). Morphant phenotypes were more than 90%, n = 50. Original magnification 40×.
Download figure to PowerPoint
Reduction in Endogenous Brg1 Activity Leads to Neurogenesis Defects
To examine the scope of the degenerating cell population in embryos expressing reduced levels of Brg1 following Brg1-MO injections, we performed whole mount in situ hybridization experiments on uninjected, control-MO-injected, or Brg1-MO-injected embryos at 10 hpf, 11 hpf (3 somites stage), or 24 hpf. Embryos were fixed, and expression of the following transcription factors was analyzed: six3 (forebrain) (Fig. 3A,B), anf/hesx-1 (forebrain) (data not shown), engrailed 2 (eng2) (mid/hind-brain boundary) (Fig. 3C–F), krox-20 (rhombomeres 3 and 5) (Fig. 3G–J), myoD (somites) (Fig. 3K,L), goosecoid (gsc) (dorsal region of the marginal zone) (Fig. 3M,N), notail (ntl) (notochord) (Fig. 3O–R), and hlx1 (presumptive rostral brain and along the ventral midline of rostral neuroectoderm) (Fig. 3S,T) (Fjose et al.,1994). The severely affected morphants exhibited expansion of the forebrain as detected by six3 and anf expression, but there was a complete absence of expression of the mid/hind-brain specific gene, engrailed 2, and reduction of rhombomere 5 detected by expression of krox-20. The expression of myoD appears unaffected at 12 hpf. Similarly, the expression of goosecoid (gsc) was also unaffected at 10 hpf. However, the zone of expression of notail (ntl) (Schulte-Merker et al.,1992), appeared to be shortened and expanded in Brg1-MO-injected embryos at 10 hpf, and was apparently less confined to the notochord at 24 hpf. Expression of the homeobox gene, hlx1, was severely affected in Brg1-MO-injected embryos. These results demonstrate that Brg1 is critical for the expression of several genes known to be essential for neurogenesis. As such, Brg1-MO-injected embryos exhibit expansion of the forebrain, a less well-defined mid/hind-brain boundary, and reduction of rhombomere 5 in the hindbrain.
Figure 3. Reduction in Brg1 expression leads to aberrant expression of genes involved in neurodevelopment. In situ hybridization analyses showing the expression of various markers in control-MO-injected (10 ng/embryo) or Brg1-MO-injected (10 ng/embryo) embryos at indicated time points. The markers were as follows: six3 (A,B), eng2 (C–F), krox-20 (G–J), myoD (K,L), gsc (M,N), ntl (O–R), and hlx1 (S,T). Morphant phenotypes were more than 90%, n = 40. Embryos are dorsal view (A–P) or lateral view (Q–T). For in situ hybridization analyses, a total of 3,000 embryos was injected and 200 embryos were used per probe. No nonspecific hybridization was observed with sense probes (data not shown). Embryos were photographed using a SPOT camera mounted on a dissecting microscope. Morphant phenotypes were more than 90%, n = 30. Original magnification 40×.
Download figure to PowerPoint
As described above, the yng mutant embryos begin to exhibit defects at 28 hpf, and consistent with this observation, the in situ hybridization experiments performed on yng mutant embryos revealed normal expression of eng2 and krox-20 at stages comparable to those presented for Brg1-MO (data not shown).
It has previously been shown that expression of six3 is critical for the formation of the forebrain through inhibition of the Wnt signal transduction pathway. In addition, over-expression of Wnt1 leads to reduced six3 expression (Lagutin et al.,2003). Since Brg1-MO-injected embryos exhibit expansion of six3 expressing cells in the forebrain, we hypothesized that Brg1 overexpression should decrease six3 expression. Consistent with the previous observation with overexpression of Wnt1 (Lagutin et al.,2003), our results also show a reduction in six3 gene expression in Brg1 mRNA-injected embryos examined at 10 hpf (Fig. 4A and B). The expression of pax6, another forebrain-specific gene, was also significantly reduced (Fig. 4C and D). Also consistent with a role for Brg1 in patterning of the forebrain, overexpression of Brg1 mRNA resulted in reduction of the area defining the forebrain, fusion of the eyes, absence of an eye (Fig. 4E,F), or asymmetrically positioned eyes (data not shown).
Figure 4. Overexpression of Brg1 mRNA leads to reduced forebrain size and rescues the Brg1-MO knockdown phenotype. A–D: Embryos were injected with 500 pg of full-length Brg1 mRNA at the one- to two-cell stage. Embryos were allowed to recover and, at indicated times, were fixed and used for in situ hybridization using six3- or pax6-specific probes. Arrowheads indicate the areas of affected gene expression. E,F: Control-injected or 500 pg Brg1 mRNA–injected embryos at 48 hpf. Note unequal eye/lens and fused eye/lens in Brg1 mRNA–injected zebrafish embryos (F). The phenotypes of the Brg1 mRNA-injected embryos were approximately 30% each in the two panels in F (n = 200). G–I: Five hundred embryos at the one- to two-cell stage were control-MO-injected (10 ng/embryo) or injected with Brg1-MO (10 ng/embryo), or embryos were injected with Brg1-MO (10 ng/embryo) plus 500 pg of full-length Brg1 mRNA. At 24 hpf, embryos were analyzed, quantitated, and photographed. Arrowheads indicate the head of the embryos. Note small head and body size with signs of apoptosis in H (Brg1-MO treated). With the concentration of the Brg1 mRNA that was used, more than 90% of the embryos could be rescued. Original magnification 30×.
Download figure to PowerPoint
To ensure that the defects in neurogenesis observed in Brg1-MO-injected embryos were due to a reduction in Brg1 mRNA levels, we also performed rescue experiments in Brg1-MO-injected embryos. Embryos were injected with 10 ng of control-MO (Fig. 4G), Brg1-MO alone (Fig. 4H), or Brg1-MO plus full-length Brg1 mRNA (Fig. 4I). Embryos rescued with Brg1 mRNA exhibit a normal phenotype compared to Brg1-MO-injected embryos (compare Fig. 4G, H, and I). Approximately 90% of Brg1-MO-injected embryos could be phenotypically rescued by Brg1 mRNA. These developing embryos were observed until 5 dpf and appeared normal.
As noted earlier, yng mutant embryos begin to exhibit abnormalities at 28 hpf and die by 6–7 dpf (Link et al.,2000). Embryos exhibit retinal defects, and there is evidence of lower levels of melanin expression (Fig. 5A,B), and cardiovascular (Fig. 5C,Da,Db) (Link et al.,2000) and brain abnormalities (Fig. 5E–L). As apparent in the data presented in Figure 5E–L, there are severe abnormalities in the order and cell density in the brain. For example, in the sagittal section presented in Figure 5E and F, there is evidence of lower cell density corresponding to the cerebellum (arrowhead). Further analyses of the fore- (Fig. 5G and H), mid- (Fig. 5I and J) and hind- (Fig. 5K and L) brain indicate loss of cells in all the regions (Fig. 5H,J,L, arrowheads) as well as disorganization in the midbrain (compare Fig. 5I and J) regions.
Figure 5. The yng mutant embryos exhibit multiple defects. A,B: yng mutant embryos exhibit lower expression of melanin compared to wild type at 2 dpf. Original magnification 40×. C,Da,Db: yng mutant embryos exhibit cardiovascular defects compared to wild type at 5 dpf. Original magnification 60×. E–L: yng mutant embryos exhibit neurodevelopmental defects compared to wild type at 5 dpf. E and F are sagittal sections. Arrow indicates position of the cerebellum. G–L are coronal sections showing evidence of cell loss in yng mutant embryo (arrowheads). Histological sections were stained with hematoxylin and eosin. hb, hindbrain; P3–P7, pharyngeal arches 3–7; tc, telencephalon; dc, diencephalons; tr, trabeculae; pq, palatoquadrate; ch, ceratohyal; mk, Meckel's cartilage; P5–P7, pharyngeal arches 5–7; pc, parachordal; mo, medulla oblongata; eg, eminentia granularis. Original magnification 200×.
Download figure to PowerPoint
Structurally, the medulla oblongata (mo) appears normal (Fig. 5) (zfin.org). These results indicate that yng mutant embryos exhibit brain abnormalities and, as described below, defects suggestive of neural crest induction and/or differentiation as they display fewer melanocytes, jaw defects (compare Fig. 5C to Da), and cardiovascular defects (Cadigan and Nusse,1997; Link et al.,2000; Gammill and Bronner-Fraser,2003; Meulemans and Bronner-Fraser,2004).
Brg1 Is Required for Neural Crest Cell Induction and Differentiation
The complex abnormalities (such as craniofacial defects, see below) observed in yng mutant embryos at 3–5 dpf, suggested the possibility of defects in neural crest cell induction and/or differentiation. Neural crest progenitor cells are induced at the neural plate border during gastrulation, then separate from other neuronal cell types and begin to express specific genes that define neural crest cells (Barrallo-Gimeno et al.,2004; Raible and Ragland,2005). Genes expressed specifically during neural crest induction and differentiation in vertebrates are the zinc-finger proteins foxd3, tfap2a, and slug/snail2 transcription factors (Meulemans and Bronner-Fraser,2004). To determine the role of Brg1 in neural crest induction, the expressions of foxd3, tfap2a, and snail2 genes were analyzed at 12 hpf in yng mutants or following knockdown of Brg1 using 5 or 10 ng of Brg1-MO in wild type embryos. Expression of foxd3 (Fig. 6–D), tfap2a (Fig. 6E–H), and snail2 (Fig. 6I–L) is dramatically reduced in 10 ng Brg1-MO-injected embryos, but embryos injected with 5 ng Brg1-MO or yng mutant embryos were less affected. In the yng mutant embryos, the expression of the above genes shifted laterally compared to wild type, which is consistent with expansion of the neural plate as described recently in Xenopus with reduced Brg1 expression (Seo et al.,2005). In addition, the reduction of foxd3 expression is clearly visible at the 15-somite stage (16.5 hpf) in yng mutant embryos (Fig. 6M and N).
Figure 6. Brg1-MO-injected embryos exhibit reduced expression of genes critical for neural crest induction. In situ hybridization analyses showing the expression of indicated genes in control-MO-injected (10 ng/embryo) or Brg1-MO-injected (5 or 10 ng/embryo), or yng mutant embryos at 12 hpf or the 15-somite (15s) stages. The expression of foxd3, tfap2a, and snail2 were analyzed. Morphant phenotypes were more than 90%, n = 40. Dorsal view of embryos showing expression of foxd3 (A–D,M,N), tfap2a (E–H), and snail2 (I–L). For the in situ hybridization analyses, a total of 2,000 embryos was injected and 100 embryos were used per probe. No nonspecific hybridization was observed with sense probes (data not shown). Original magnification 40×.
Download figure to PowerPoint
The fact that reduced Brg1 protein levels following injection of Brg1-MO lead to a severe decrease in expression of genes specifying neural crest induction implies a crucial role for Brg1 in this process. Recent reports indicate that mutation in tfap2a (Mont Blanc) leads to the death of neural crest cell derivatives in zebrafish embryos (Barrallo-Gimeno et al.,2004). The tfap2a gene has been shown to be critical for the onset of normal development in trunk neural crest cells and brachial arches 2–7. Deficiency in tfap2a results in defects of thoraco-abdominoschisis (Zhang et al.,1996). Mice deficient in tafp2a exhibit defects in body wall closure and craniofacial skeletal defects (Zhang et al.,1996). The snail2 and closely related slug genes are transcriptional repressors that respond to Wnt signaling and the slug gene promoter has been shown to contain TCF/LEF binding sites (Gammill and Bronner-Fraser,2003). The slug/snail2 genes are involved in the induction and delamination of neural crest cells from the neural tube and migration of neural crest cells (Le Douarin et al.,1994; Meulemans and Bronner-Fraser,2004). Similarly, the foxd3 gene is expressed in pre-migratory neural crest cells, floor plate, somites, and tail bud (Odenthal and Nusslein-Volhard,1998). At later stages of development, foxd3 is expressed in peripheral nervous system glial cells (Kelsh et al.,2000).
Many studies suggest that Wnt/β-catenin signaling is also involved in later stages of development during neural crest cell differentiation after formation of neural crest cells (Lewis et al.,2004). Activation of Wnt/β-catenin signaling in pre-migratory neural crest cells in zebrafish promotes pigment cell formation to the detriment of neurons and glia; in contrast, inhibition of the Wnt/β-catenin signaling pathway promotes neural and glial cell fate (Dorsky et al.,1998; Lewis et al.,2004). At 15 hpf, neural crest cells begin their migration at the midbrain, which progresses in the caudal direction. At 18 hpf, the majority of chondrogenic neural crest cells start descending towards the pharyngeal arches, while the trunk neural crest cells are just beginning to enter the medial migration (Barrallo-Gimeno et al.,2004). Neural crest cell movement continues through 24 hpf. Dlx2 expression was determined in the yng mutant, and was found to be expressed at control levels at 24 and 30 hpf indicating that migratory neural crest cells are unaffected in terms of dlx2 expression (data not shown). However, at 52 hpf, differences in dlx2 expression are visible between yng mutant and wild type embryos. Normally, in wild type embryos, a small mouth appears close to the posterior edges of the eyes (Fig. 7). Later during development, the position of the mouth moves in the anterior direction and the pharyngeal arches extend to make the jaws (zfin.org). However, in yng mutant embryos, the first mandibular arch opens at the midline, and, instead of elongating toward the mouth as normally observed, points out ventrally (Fig. 7, see *). These results indicate that in yng mutant embryos, cranial neural crest migration is affected around 52 hpf.
Figure 7. Expression of dlx2 gene, a specifier for neural crest migration. In situ hybridization analyses showing the expression of dlx2 gene in wild type (Wt) and yng mutant embryos at 52 hpf. Note dlx2 expression in the extending mandibular (m) and the hyoid (h) arches positioned in a ventral-anterior direction in the wild type versus ventral in yng mutant. Morphant phenotypes were more than 90%, n = 30. Original magnification 40×.
Download figure to PowerPoint
In order to further understand the effects of Brg1 deficiency in cranial neural crest cell differentiation, we examined skeletal formation in homozygous yng mutant embryos using Alcian Blue staining. The skeleton of wild-type embryos at 5 dpf consists of a neurocranium and seven pharyngeal arches. The first pharyngeal arch (P1) forms the jaw while the second arch (P2) supports the jaw (Piotrowski et al.,1996; Schilling et al.,1996). The posterior arches (P3–P7) form the pharyngeal gill structures (Fig. 8). The first arch derivatives include Meckel's cartilage (mk), a part of the lower jaw, and the palatoquadrate (pq), a part of the upper jaw. The second (hyoid) arch consists of ventral ceratohyal (ch), basihyal (bh), and dorsal hyosymplectic (hs). The third to seventh pharyngeal arches (P3–P7) consist of ceratobranchials (cb) and hypobranchials (hb) (Piotrowski et al.,1996; Schilling et al.,1996). The neurocranium, a portion of the head skeleton, encloses the brain and the sensory organs. Unlike the pharyngeal skeleton, it is derived from both mesoderm and cranial neural crest cells. In the anterior part of the neurocranium, the trabeculae (tr) fuse across the midline and form the ethmoid plate (ep). In the posterior part of the neurocranium, the trabeculae combine into the parachordals (pc) to form the basal plates (Piotrowski et al.,1996; Schilling et al.,1996). The yng mutant embryos at 5 dpf displayed several clearly visible pharyngeal skeleton defects. Both elements of Meckel's cartilage were ventrally positioned and their shape appeared irregular and fused across the midline and formed two rows of chondrocytes instead of one layer (compare Fig. 8A–D to G–J). The palatoquadrate did not appear triangular in shape and moved posteriorly (compare Fig. 8A,B to G,H). The dorsal process did not articulate with the ethmoid plate (compare Fig. 8A,B to G,H). In the second arch, the ceratohyal was also oriented ventrally (compare Fig. 8A–D to G–J). The basihyal along the midline was greatly reduced or absent (compare Fig. 8C,D to I,J). The posterior pharyngeal arches (P3–P5) were present but were smaller in size. The last 2 arches (P6–P7) were absent or severely reduced (compare Fig. 8A–D to G–J). Examination of the neurocranium from the dorsal view in yng mutant embryos revealed incomplete and aberrant development of parachordal cartilage and the ethmoid plate (compare Fig. 8K,L to E,F). Even though the trabeculae in yng mutants extend normally, they failed to fuse completely into the ethmoid plate, resulting in a smaller, concaved ethmoid plate with a cleft (compare Fig. 8K,L to E,F). The sagittal and cross-sections of 5 dpf embryos of yng mutant also revealed an abnormal pattern of the pharyngeal arches (see Fig. 5E–L). In wild type embryos, the pharyngeal arches (P3–P7) are separated from each other by gill clefts. In contrast, segmentation of the pharyngeal arches in yng mutants is not clearly visible. These results indicate that lack of Brg1 leads to jaw defects, suggesting aberrant differentiation of cranial neural crest cells.
Figure 8. The yng mutant embryos show defects in neural crest cell differentiation. Craniofacial skeletons and schematic drawings of wild type (A,C,E and B,D,F), and yng mutant (G,I,K and H,J,L), embryos following Alcian Blue staining at 5 dpf. The lateral (A and G) and ventral (C and I) views show that all the pharyngeal cartilages form ventrally and the most posterior arches (P6–P7) are reduced or absent in yng mutant embryos. The dorsal processes of the platoquadrate (pq) fail to articulate with the ethmoid plate (ep). The basihyal (bh) along the midline is reduced. The dorsal view of neurocranium (E and K) shows that neural crest–derived ethmoid plate (ep) and mesodermally derived parachordals (pc) are reduced. bh, basihyal; ch, ceratohyal; ep, ethmoid plate; hs, hyosymplectic; mk, Meckel's cartilage; nc, notochord; pc, parachordal; pq, platoquadrate; tr, trabeculae; Wt, wild-type. Note that eyes were removed from the wild type embryo (A) to observe fine details. Original magnification 200×.
Download figure to PowerPoint
Neural crest cells are also critical for the development of the peripheral nervous system (PNS) consisting of sensory neurons and dorsal root ganglia (DRG) (Dorsky et al.,1998; Gammill and Bronner-Fraser,2003; Meulemans and Bronner-Fraser,2004). Similar to neural crest cells, Rohon Beard (RB) sensory neurons form at the border between the neural and non-neural ectoderm, respond to signaling cascades, including the Wnt pathway, and are derived from the same precursors (Hernandez-Lagunas et al.,2005). In Xenopus, Brg1 has been shown to control expression of Ngnr1 (Seo et al.,2005) and reducing expression of Ngnr1 in zebrafish has been shown to lower the number of RB sensory neurons (Cornell and Eisen,2002). To detect the dorsal Rohon Beard sensory neurons as well as dorsal root ganglion (DRG) axons in Brg1-MO-injected embryos, or in yng mutant embryos, we performed immunohistochemistry using antibodies to the pan neuronal marker HuC/D and acetylated tubulin. Our results indicate that in 24-hpf Brg1-MO-injected embryos, there are fewer primary motor nerves at the mid-trunk and fewer Rohon Beard sensory neurons indicated in Figure 9A by arrowheads. At 5 dpf, in yng mutant embryos, there are fewer numbers of dorsal root ganglions (DRGs) in comparison to wild type (Fig. 9B, arrowheads). These results thus indicate that Brg1 is involved in differentiation of DRGs and sensory neurons in the peripheral nervous system originating from trunk neural crest cells.
Figure 9. The Brg1-MO injected and yng mutant embryos exhibit reduced peripheral neurons and axons. A: One- to two-cell stage embryos were injected with control-MO or Brg1-MO, and at 24 hpf embryos were fixed and immunostained with antibodies to acetylated tubulin (acTUB) (lateral) or HuC/D (dorsal). Original magnification 200×. B: The yng mutant embryos were fixed and immunostained at 5 dpf as above. Lateral views are presented. The acTUB and HuC/D were detected using AlexaFluor 594- and AlexaFluor 488- labeled secondary antibodies, respectively. Image analysis was performed with conventional fluorescent or confocal microscopy. In A, note the reduced primary motor nerves and Rohon Beard sensory neurons in Brg1-MO-injected embryos compared to control (arrowheads). In B, note the reduced DRGs in the trunk region of yng mutant embryos compared to wild type indicated by arrowheads. Original magnification 200×.
Download figure to PowerPoint
Brg1 Regulates snail2 Promoter Containing TCF/LEF Binding Site In Vivo
Previous studies in other organisms indicated that following activation of Wnt signaling pathway, Brg1 is recruited to the TCF/LEF binding site and activates downstream target genes such as slug/snail2 (Gammill et al.,2003). This link has not yet been made in zebrafish. In order to determine the expression of various components of the Wnt signal transduction pathway and whether Brg1 can be placed on snail2 promoter regions encoding TCF/LEF binding sites in developing zebrafish, we performed the following experiments: wild-type or Brg1-MO-injected embryos were collected at 6, 12, or 24 hpf and cDNAs were prepared and the expression of Wnt-1, Wnt-4, Wnt-8b, Wnt-10b, Frizzled-2, β-catenin, LEF-1, TCF-1, Engrailed2, Snail2, and Brg1 were analyzed using semi-quantitative RT-PCR. The results indicate that with the exception of Engrailed2, Snail2, and Brg1 whose expression are lower in Brg1-MO-injected embryos, expression of other components of the Wnt signaling were intact (Fig. 10A). To ensure that Brg1 is recruited to the TCF/LEF binding sites of snail2 promoter, we injected plasmids containing human Brg1 cDNA (De La Serna et al.,2000) into one-cell stage embryos and performed a ChIP assay. However, before the assay, the expression of Brg1 was confirmed by immunoblotting in 22-hpf embryos (Fig. 10B). We also confirmed that human Brg1 could rescue the Brg1-MO-injected zebrafish embryos as efficiently as zebrafish Brg1 (Fig. 10C; see Fig. 4G–I). The ChIP assay was performed using 12- and 22-hpf embryos that had been injected with plasmids containing human Brg1 cDNA at one-cell stage. The results indicate that Brg1 immunoprecipitated the snail2 promoter at both 12 and 22 hpf (Fig. 10). Interestingly, the 3-kilo-base pair of the zebrafish snail2 promoter contains three putative TCF/LEF binding sites (CTTTGA/TA/T) (Vallin et al.,2001). In the ChIP assay, we were able to confirm that there was no interaction of Brg1 with a region that contained two putative TCF/LEF binding sites. However, Brg1 did pull down the putative TCF/LEF binding site in region b (Fig. 10D and E). These results confirm that snail2 is a downstream target of Brg1 in this organism.
Figure 10. Brg1 is recruited to the snail2 promoter in developing zebrafish. A: Wnt signaling pathway is intact in Brg1-MO-injected embryos. The cDNAs were prepared from embryos at indicated times and the expression of Wnt1, Wnt4, Wnt8b, Wnt10b, Frizzled2, β-catenin, LEF1, TCF1, Engrailed2, Snail2, Brg1, and β-Actin was determined using specific primers. Note that the expression analyses of the indicated cDNAs in the wild-type are similar to previous reports (Thisse et al.,1995; Dorsky et al.,1999; Bellipanni et al.,2006) and zfin.org. B: Immunoblot analyses of human Brg1 in developing zebrafish. Plasmids containing human Brg1 cDNA (pcDNA3-Brg1) (De La Serna et al.,2000) were injected into one-cell stage embryos. At 22 hpf, lysates were prepared and used to detect Brg1 expression using antibody to human Brg1. Note that we injected plasmids containing human Brg1 because the commercially available Brg1 antibody does not recognize the zebrafish Brg1, although there are high levels of sequence homology between Brg1 cDNA from zebrafish and other organisms. Hela cell lysate was used as a positive control. Lane termed “negative control” represents lysates prepared from zebrafish injected with empty plasmids. β-actin was used as a loading control. C: Human Brg1 efficiently rescues Brg1-MO-injected zebrafish embryos. One-cell stage embryos were left uninjected (top), or injected with Brg1-MO (middle) or Brg1-MO and plasmids containing human Brg1 (50 ng/μl) (bottom). Embryos were photographed at 25 hpf. D: Snail2 promoter regions a and b contain 3 TCF/LEF binding sites (at −1,796 bp and −2,070 bp in a, and at −2,682 bp in b) that were PCR amplified (primers location are indicated by arrows) in ChIP assays. E: Embryos were injected with control plasmids or plasmids containing pcDNA3-human Brg1 at one-cell stage. At 12 and 22 hpf, embryos were collected and prepared for ChIP assays as described in Experimental Procedures. Immunoprecipitated chromatin (or total genomic DNA in the case of input) from each group was PCR amplified for the indicated promoter fragments. Top panels represent where antibody to Brg1 was used to immunoprecipitate chromatin from embryos injected with plasmids containing human Brg1. Middle panels represent where antibody to Brg1 was used to immunoprecipitate chromatin from embryos injected with control empty plasmids. Bottom panels represent PCR products from total genomic DNA input.
Download figure to PowerPoint
In conclusion, we have analyzed zebrafish development in the absence of Brg1. Our results indicate that Brg1 is critical in brain development and neural crest induction and differentiation. These data are consistent with the phenotype of yng mutant embryos at late developmental stages before their death from cardiac defects at 6–7 dpf (Fig. 5) (Link et al.,2000). Our results are also consistent with several recent reports in other vertebrates showing that Brg1 is involved in neurogenesis (Bultman et al.,2000; Matsumoto et al.,2005; Seo et al.,2005). In Xenopus, Brg1 has been shown to interact with the bHLH transcription factors Ngnr1 and NeuroD, and Brg1 deficiency interferes with transcriptional activity of both factors (Seo et al.,2005). More recently, conditional ablation of Brg1 in neural stem cells using nestin-cre transgenic mice crossed with those bearing a floxed-Brg1 gene indicate that although early development of neurons (E10.5) appears normal, lack of Brg1 leads to deficiencies in the differentiation of neural stem cells to astrocytes and oligodendrocytes in favor of generating post-mitotic neurons (Matsumoto et al.,2005). The roles of Brg1 in neural crest cell induction and differentiation is interesting and needs further analysis in terms of Brg1 regulation of specific target genes. As we also show in Figure 10, a critical function for Brg1 in regulation of TCF/LEF target genes such as slug/snail2, genes critical for neural crest delamination and migration, is anticipated (Le Douarin et al.,1994; Gammill et al.,2003; Lewis et al.,2004; Meulemans and Bronner-Fraser,2004). However, the role of Brg1 in severely reducing the expression of other neural crest specifiers such as tfap2a and foxd3 requires additional investigation. Previous reports indicate that down-regulation of Wnt/beta-catenin using a dominant negative form of TCF inhibits foxd3 expression at the 3 somite stage, but not at the later 6 somite stage (Lewis et al.,2004), and this is similar to our observation in which Brg1-MO-injected embryos do not express foxd3 at 12 hpf. Comparable inhibition of TCF did not affect HuC/D expressing dorsal Rohon Beard sensory neurons of the spinal cord detected at the 10-somite stage (Lewis et al.,2004). However, in Brg1-MO-injected embryos at 24 hpf, and in yng mutant embryos at later developmental stages, reduced numbers of such neurons are present, suggesting that in addition to controlling TCF/LEF target genes, Brg1 may have additional functions during neural crest induction. It should also be noted that decreased Brg1 levels during development leads to increased cell death (Fig. 2) and additional experiments are required to determine the specific developmental role for Brg1 in promoting cell survival.