| Edited by Toshihiko Shiroishi. Yongsu Jeong: Corresponding author. E-mail: yongsu@khu.ac.kr |
The Six genes are vertebrate homologs of the Drosophila sine oculis (So), which was initially identified as a component of the retinal determination cascade. So acts as a transcription factor in concert with Eya and Dach to regulate compound eye organogenesis (Pignoni et al., 1997; Chen et al., 1997). One of the murine Six genes, Six1, is expressed in various tissues during development and plays crucial roles in differentiation, morphogenesis and organogenesis (Christensen et al., 2008; Kumar, 2009). In humans, SIX1 is associated with the autosomal developmental disorder BOR (branchio-oto-renal) syndrome (Ruf et al., 2004), and has been implicated in cancer and tumorigenesis (Kumar, 2009). Previous studies using Six1-deficient mice have suggested its critical role in forming the nervous system and sensory structures. The mouse olfactory epithelium emerges from the olfactory placodes, which invaginate toward the forebrain to form the olfactory pits. Six1–/– mutants showed lack of pioneer neurons in the olfactory epithelium as well as defects in axonal outgrowth and neuronal migration to the olfactory bulb (Ikeda et al., 2007). In addition to the olfactory placode, other cranial placodes such as otic, trigeminal and epibranchial placodes generate cranial sensory neurons of the trigeminal (V), facial (VII), acoustic (VIII), glossopharyngeal (IX) and vagus ganglia (X) (Schlosser, 2006). Six1 has been shown to be critical for early differentiation and survival of these placodally- derived cranial sensory neurons (Christophorou et al., 2009; Konishi et al., 2006; Pandur and Moody, 2000; Schlosser et al., 2008; Zou et al., 2004). In the trunk, Six1 is also expressed abundantly in the dorsal root ganglia, where some developmental abnormalities were observed in Six1–/– / Six4–/– mutant mice (Ando et al., 2005).
Members of Six proteins are characterized by the presence of 2 evolutionarily conserved domains, homeobox nucleic acid recognition domain (HD) and Six domain (SD), and can be subdivided into 3 subclasses, Six1/2, Six4/5 and Six3/6 (Kawakami et al., 2000). Previous studies have refined the consensus binding sites for the Six proteins. The site bound by Six3/6 contains the traditional ATTA homeodomain core recognition sequence, while Six1/2 and Six4/5 bind to a different consensus site, TCAGGTT (Kumar, 2009). Six1 and Six4 are expressed in the developing muscle and control early steps of myogenic development, as indicated by Six1–/– / Six4–/– mutant mice showing general muscle hypoplasia (Grifone et al., 2005; Niro et al., 2010). Six1 and Six4 have been shown to activate Myogenin promoter through the unique consensus site (Spitz et al., 1998). Interestingly, Six proteins do not appear to function just as transcriptional activators. Six1 which promotes placodal fate affects the expression of ectodermal genes via both transcriptional activation and repression depending upon the presence of cofactors (Bricaud and Collazo, 2006; Brugmann et al., 2004). While progress has been made, molecular mechanisms of how Six1 influences critical developmental decisions at the level of gene regulation remain largely unknown.
In the present study, efforts to determine the direct transcriptional targets of Six1 led to the identification of a regulatory sequence that is bound by Six1 homeoprotein. We employed Chromatin Immunopreciptation (ChIP)-Display to locate the regions occupied by Six1. Despite being relatively low- throughput compared to ChIP-chip and ChIP-Seq, ChIP-Display provided an opportunity to discover novel targets of Six1. We showed that one of the Six1-bound regions, which is evolutionarily conserved in vertebrates, is a functional noncoding element engaged in controlling gene expression at specific sites of the developing nervous and sensory structures.
ChIP Display was performed essentially as previously described (Barski and Frenkel, 2004). Pooled embryos at E10.5 were fixed with 1% formaldehyde for 20 min with shaking. After a 10 min incubation with 10 mM glycine, trunks were dissected and disrupted in lysis buffer (20 mM Tris-HCl, pH 7.5, 1% Triton X-100, and 1 mM EDTA) and protease inhibitor cocktail (Sigma) by passing through 30G needles. Chromatin was sonicated and diluted with 20 mM Tris-HCl, pH 7.5, 140 mM NaCl, 0.5% Triton X-100, and protease inhibitor cocktail. After preclearing with protein A agarose beads (Upstate), the chromatin was incubated overnight with anti-Six1 (Abcam) or IgG (Santacruz), followed by incubation with protein A agarose beads, and washed with 20 mM Tris-HCl, pH 7.5, 140 mM NaCl and 0.5% Triton X-100. After elution with 20 mM Tris-HCl, pH7.5, 10 mM EDTA, and 1% SDS and decrosslinking, DNA was purified and subjected to Display. The immunoprecipitated DNA was treated with SAP (Roche), digested with Ava II, and ligated to linkers (sense, 5’-TTCGCGGCCGCAC-3’; antisense, 5’-GWCGTGCGGCCGCGAA-3’). The DNA was purified with MinElute kit (Qiagen) and subjected to PCR amplification using nested primers (5’-CGGCCGCACGWCCN-3’). After the resulting PCR products were resolved on 6% PAGE, DNA from each band was purified with SpinX columns (Corning) and subjected to reamplification by PCR. The resulting DNAs were digested with Hinf I and resolved on 3.5% agarose gel electrophoresis. Bands of interest were excised and the DNA was purified with GeneClean III kit (BIO101) for sequencing.
To confirm Six1 binding in vivo, chromatin DNAs were prepared from each of the trunks and dorsal brains and subjected to quantitative PCR (QPCR) (Jeong and Epstein, 2005). Each reaction contained 0.6 μl of the reference dye (diluted 1:1600), 2 μl of chromatin DNA, 1 μl of 2 μM forward primer, 1 μl of 2 μM reverse primer and 10 μl of 2x Brilliant SYBR Green QPCR Master Mix (Stratagene) in a final reaction volume of 20 μl. PCRs were performed as follows: 1 cycle at 95°C for 10 min and 40 cycles at 95°C for 30 s, 55°C for 1 min, and 72°C for 30 s. PCRs of 3 independent replicates were each performed in triplicate. The primer sequences for SRE1 were 5’-TTTAATACGGGACAAACAGTAGAG-3’ (forward) and 5’-ATTATGGATGCGTTTTTAGAGCAC-3’ (reverse). The primer sequences for other targets are available upon request.
For whole-mount immunostaining, embryos were fixed in 4% paraformaldehyde in phosphate buffered saline (PBS) at 4°C for 2 h and were incubated in PBS containing 0.4% Triton X-100, 5% sheep serum and 3% M.O.M. blocking reagent (Vector Labs) at 4°C for 6 h. Then, the embryos were incubated overnight with an anti-neurofilament antibody (2H3, Hybridoma Bank). After washing 6 times (1 h each) in PBS containing 0.4% Triton X-100, embryos were incubated overnight with Cy3 anti-mouse IgG antibody and washed 3 times (1 h each). For section immunostaining, 20 μm- embryo sections were incubated with anti-Six1 antibody (Abcam) followed by alkaline phosphatase-conjugated goat anti-mouse secondary antibody (Millipore).
All evolutionarily conserved regions (ECR) sequences assayed were cloned into the Not I restriction site of a reporter vector containing the beta-globin minimal promoter, lacZ cDNA, and SV40 large T antigen poly(A) site. The ECR sequences from mouse, human, dog, rat, chicken and frog were generated by PCR amplification of their respective genomic DNA using the following primers: mouse, 5’-ATTAGCGGCCGCCTTGAGGTTCCACTCACTTAAGC-3’ and 5’-ATTAGCGGCCGCCTTAATGTCCTTGGAAACATCCTG-3’; human, 5’-ATTAGCGGCCGCCACAGCAATGAGAAGTGGGGGCTGAG-3’ and 5’-ATTAGCGGCCGCAAAGGAAGAAGTGTCGAGTCTC-3’; dog, 5’-ATTAGCGGCCGCTTAAGAGAAGGTTTATTTCCTG-3’ and 5’-ATTAGCGGCCGCAAGACTCCAGGCTTGGTGCCCTG-3’; rat, 5’-ATTAGCGGCCGCAGAAAGGGAAAAACTGAATC-3’ and 5’-ATTAGCGGCCGCCTTGAGGTTCTGCTCACTTAAG-3’; chicken, 5’-ATTAGCGGCCGCGAGATGAGGAAAAAGCAAGGGTG-3’ and 5’-ATTAGCGGCCGCATATTCTAGTTATTGAAGGGTTAAAG-3’; frog, 5’-ATTAGCGGCCGCTAAATATATTTTTAATACGTGAC-3’ and 5’-ATTAGCGGCCGCATAAATAAATGATGAATGC-3’. A construct, harboring a small deletion of the Six1/2/4/5 core motif (TCAGGT) from mouse SRE1, was generated by ligating 2 PCR products flanking the region of interest that were amplified using 2 primer pairs: 5’-ATTAGCGGCCGCCTTGAGGTTCCACTCACTTAAGC-3’ and 5’-CATTAAAGCCAGCGACTCACC-3’; 5’-AGCCGTGTGATAGTCTCTG-3’ and 5’-ATTAGCGGCCGCCTTAATGTCCTTGGAAACATCCTG-3’.
Protocols for animal use (KHUASP-10-003 and KHUASP-10-004) were reviewed and approved by the Kyung Hee University Institutional Animal Care and Use Committee. Mice were housed in groups of 2–5 per cage at a temperature of 21°C, with a 12-h/12-h light/dark cycle. Transient transgenic embryos were generated by pronuclear injection into fertilized eggs derived from FVBN (Charles River) strains. Transgenes were prepared for microinjection as described (Jeong and Epstein, 2005). The genotyping of embryos carrying reporter constructs was performed by PCR using Proteinase K-digested yolk sacs as DNA templates.
To explore the binding targets of Six1, we dissected E10.5 mouse embryos and performed ChIP-Display (Barski and Frenkel, 2004) using a Six1 antibody (Fig. 1A, B). Immunoprecipitated DNA fragments were dephosphorylated, digested and ligated to linkers, followed by PCR amplification. After the resulting PCR products were resolved by PAGE, only bands reproduced in 2 independent Six1 ChIP lanes and in none of the IgG control lanes were excised, eluted from the gel and reamplified for identification (Fig. 1C). Sequences obtained in this manner were aligned with the mouse genome using UCSC (genome.ucsc.edu) and Ensemble (www.ensembl.org) genome browser and mapped to 8 distinct regions. To ascertain Six1 occupancy at each of these regions, we performed ChIP-quantitative PCR using target-specific primers in 3 independent experiments. Robust enrichments were obtained in Six1 ChIPs as compared to paired IgG control ChIPs, confirming in vivo Six1 binding to each target sequence (Fig. 1D and data not shown). Five of the confirmed Six1-bound regions were found within or 5’ proximal to known genes, which include Myogenin, Cyclin D, SallI, Slc12a2 and Gdnf (Table 1). These genes were originally identified as the target genes of Six1 (Ando et al., 2005; Chai et al., 2006; Kobayashi et al., 2007; Li et al., 2003; Spitz et al., 1998). We also identified 3 novel Six1-bound regions far distant from the nearest annotated transcription units Tbx3, FoxD3 and Shh (Table 1).
 View Details | Fig. 1 ChIP-Display identifies a novel Six1-occupied genomic region. (A) Chromatin DNAs were isolated from trunks (caudal to the dashed line) of E10.5 mouse embryos and immunoprecipitated with anti-Six1 antibody. The immunoprecipitated DNAs were treated with shrimp alkaline phosphatase to prevent ligation of linkers to the DNA ends generated by sonication. The DNAs were then digested with Ava II to scatter non-specifically precipitated fragments, ligated to linkers and subjected to PCR amplification using nested primers. After the resulting PCR products were resolved on PAGE, only bands reproduced in 2 independent ChIPs were excised and the DNA from each band was PCR-reamplified. After digestion with Hinf I to scatter any non-specific remnants, the amplified products were resolved on 3.5% agarose gel electrophoresis and sequenced for identification. (B) Six1 immunostaining. Six1 expression was detected in the dorsal root ganglia and somites. (C) An example of ChIP-Display results. PCR amplification products from two independent Six1 ChIPs and IgG controls were resolved in 6% polyacrylamide gel. A band (arrowheads), which was reproduced in Six1 ChIP lanes, but not in IgG control lanes was excised. (D) ChIP from embryos using antibody to Six1 or IgG. QPCR results from three independent experiments reveal a significant enrichment of SRE1 DNA in Six1- versus IgG-bound chromatin isolated from trunk but not dorsal brain of E10.5 mouse embryos (*P < 0.005, Student’s t-test). |
 View Details | Table 1 Six1-bound regions identified by ChiP-Display |
To test the functional relevance of the novel Six1-occupied regions in vivo, we cloned each sequence into a reporter expression construct containing beta-globin minimal promoter, lacZ reporter and SV40 polyA, and then analyzed reporter expression in transgenic mouse embryos. Embryos carrying the Six1-bound sequences of which the nearest genes are Tbx3 or FoxD3 showed no specific X-gal staining in the developing mouse tissues. This does not discount the possibility of acting as a regulatory element since Six1 has been shown to function as a context-dependent activator or repressor of target gene expression (Bricaud and Collazo, 2006; Brugmann et al., 2004). Embryos carrying the target sequence residing ~380 kb away from Shh locus directed lacZ expression to the developing nervous system (Fig. 2A). X-gal staining was detected in the cranial ganglia such as the trigeminal (V), facial (VII), acoustic (VIII), glossopharyngeal (IX) and vagus nerves (X) (Fig. 2A–C). In the trunk, lacZ expression was present in the neural tube, DRG, sensory and motor nerves (Fig. 2D, E). This pattern of X-gal staining appeared similar to that of neurofilaments (compare Fig. 2A–E and F–J). This sequence was designated SRE1, Six1-bound Regulatory Element 1.
 View Details | Fig. 2 Isolation of a regulatory sequence controlling reporter activity in the cranial and spinal nerves. (A–E) X-gal staining in the cranial and spinal nerves of transgenic embryo carrying mouse SRE1 sequence. (B–D) Enlargements of panel (A). lacZ expression was present in the trigeminal (V), facial (VII), acoustic (VIII), glossopharyngeal (IX) and vagus nerves (X). At the trunk level (D, E), X-gal staining was detected in the neural tubes, dorsal root ganglia (DRG), and sensory and motor nerves. (F) Schematic illustrating the pattern of neurofilament staining. (G–J) Whole-mount immunostaining using 2H3 antibody revealed the cranial and spinal ganglia and peripheral nerves of E10.5 wild-type embryo. fl, forelimb; ov, optic vesicle; ot, otic vesicle; S, sensory nerve; M, motor nerve. |
Given that vertebrate species show similar expression patterns and functions of Six1 in the developing embryo (Kumar, 2009; Pandur and Moody, 2000), we explored whether the mouse SRE1 may be preserved across phyla. Comparative sequence analysis of the 784-bp fragment of the SRE1 was undertaken using ECR browser (ecrbrowser.dcode.org), and identified evolutionarily conserved regions (ECR) from a number of vertebrate species (Fig. 3A). Comparisons between mouse and individual species such as human, monkey, dog or rat showed a high degree of overall homology of 84%, 84%, 80% and 93%, respectively (Fig. 3A). Conservation of chicken and frog sequences was found on average to be lower when compared to that of mouse, with homologies of 25% and 15%, respectively. On closer inspection, however, alignment of short stretches of chicken and frog sequences revealed higher homology scores, 75% and 73%, respectively, compared to the overall average. SRE1 sequences were not found in zebrafish or lower vertebrates. To determine the extent to which the conservation of sequence reflected the conservation of function, each of the homologous regions was assayed for its reporter activity in transgenic mouse embryos. DNA fragments were amplified by PCR from human, dog, rat, chicken and frog genomic DNAs, and cloned into the reporter expression construct. Transgenic embryos carrying mammalian SRE1 sequences including human, dog and rat showed strong X-gal staining in a manner comparable with the mouse SRE1 (Fig. 3C–E). X-gal staining was detected in the cranial ganglia, DRG, as well as in the neural tube, in keeping with the significant preservation of nucleotide identity across much of this element. SRE1 sequences from chicken and frog were also able to activate reporter expression in the developing mouse embryos despite the fact that the degrees of conservation of sequences were reduced (Fig. 3F, G). These results suggest that SRE1 functions as a conserved regulatory element. Since Shh is not expressed in the cranial and spinal ganglia, it is likely that SRE1 is linked to transcriptional regulation of other gene, which remains to be determined.
 View Details | Fig. 3 Conservation of SRE1 sequence and function. (A) Vista plot comparing the alignment of human SRE1 sequences with macaque, dog, mouse, rat, chicken, frog, fugu and zebrafish. (B) Nucleotide position and the number of transgenic embryos exhibiting reproducible reporter activity for each SRE1 sequences. (C–G) SRE1 reporter activity derived from (C) human, (D) dog, (E) rat, (F) chicken, and (G) frog sequences. |
To understand molecular mechanisms governing the regulatory activity of SRE1, we surveyed the SRE1 sequences for known transcription factor-binding sites. A binding site matching the consensus for Six1/2/4/5 (TCAGGTT) was identified in the sequence of SRE1 in a region that was highly conserved between the vertebrate species (Fig. 4). To address whether this site is required for the regulatory activity of SRE1, a reporter construct that lacked the Six1/2/4/5-binding site in the mouse SRE1 sequence was generated and tested for its ability to drive lacZ expression in transgenic embryos. Deletion of the Six1/2/4/5-binding site in the mouse SRE1 resulted in loss of reporter expression in the cranial nerves and DRG (Fig. 4B), suggesting that SRE1 is directly regulated by Six1. X-gal staining was also abrogated in other regions such as the neural tube and motor nerves where Six1 is not expressed. Six4, which is capable of binding to the same Six1/2/4/5 consensus site, has been shown to be expressed in the neural tube (Ohto et al., 1998). Whether Six4 is a candidate regulator of SRE1 remains to be determined.
 View Details | Fig. 4 Requirement of conserved Six1/2/4/5-binding site in SRE1. Alignment of SRE sequences from human, macaque, dog, rat, mouse, chicken and frog is shown at the top. Conserved sequences are shaded in yellow. DNA recognition sequences matching the consensus for Six1/2/4/5 are shown. (A, B) X-gal staining of transgenic embryos at E10.5 carrying reporter constructs with wild-type (A) or deletion of DNA binding site (B). Underneath each figure is a schematic of the reporter construct. The table at the bottom indicates the total number of transgenic embryos generated for the mutant construct. P, beta-globin minimal promoter. |
Identifying functional regulatory elements that control the spatial and temporal expression of genes is an unresolved task in the decoding of human and other mammalian genomes. To uncover remote-acting transcriptional enhancers is, in particular, challenging because they are scattered among the huge noncoding portion of the genome. Recent studies for locating non-coding regulatory elements and transcription factor- binding sites utilize the coupling of ChIP with high-throughput analysis such as ChIP-PET and ChIP-Seq. While these approaches provide genome-wide maps of protein-DNA binding sites, they require vast amounts of starting material as well as tremendous financial resources. Instead, we employed ChIP-Display, which has been previously shown to be amenable and efficient for small-scale experiments (Barski and Frenkel, 2004; Jariwala et al., 2007; Patrick et al., 2009). In addition to Myogenin, Cyclin D, SallI, Slc12a2 and Gdnf, previous studies have also identified other binding targets of Six1, which include Atp2a1, Eno3, Ifitm3, Kcne1, Mybph, Myeov2, Myl1, Tnnc1, Tnnc2, Tnnt3 (Niro et al., 2010), Myf5 (Giordani et al., 2007), Ezrin (Yu et al., 2006), cyclinA1 (Coletta et al., 2004). Our ChIP-Display was validated by the identification of ~25% of genes that have been reported as Six1-bound targets. Furthermore, our study demonstrated that the combination of ChIP-Display, in vivo transgenic reporter assay and comparative genomics is a successful strategy to identify novel functional regulatory elements conserved among species. Future experiments using the same strategy will lead to the discovery of additional novel noncoding elements regulated by Six1.
Results from our study indicated that SRE1 is an evolutionarily conserved transcriptional enhancer, but did not reveal its associated target gene. The gene nearest to SRE1 is Shh, which has been known to play critical roles in diverse developmental processes including the neural tube patterning. Shh is expressed in the ventral region of the neural tube along the entire CNS axis (Echelard et al., 1993), but its overall pattern of expression is different from SRE1-regulated reporter expression, ruling out Shh as a candidate target. We surveyed ~1 Mb upstream or downstream of SRE1 sequence, but did not find any known transcription units showing relevant expression pattern, suggesting that SRE1 may operate over a greater distance. Alternatively, there are emerging evidences to show that regulatory elements might act in trans to control genes on other chromosome (Spilianakis et al., 2005; Williams et al., 2010). On the basis of 2H3 immunostaining, neurofilaments, type IV intermediate filament heteropolymers composed of light, medium, and heavy chains, are expressed in a manner similar to SRE1-regulated reporter expression. Their coding sequences reside on chromosomes 11 and 14, while SRE1 is located on chromosome 5. Whether there are interchromosomal associations between the neurofilament genes and SRE1 remains to be determined. Identification of SRE1-associated gene and elucidation of its function should reveal novel regulatory mechanisms of the Six1 transcription factor.
This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (2010-0015499).
|
Index