Mesendoderm gene regulatory networks in zebrafish

GLOBAL JOURNAL OF BIOCHEMISTRY

Mesendoderm gene regulatory networks in zebrafish Chung-Hao Chaoa,b,#, Wen-Fang Tsenga,#, Tzu-Min Chana, Hua-Ling Chena, Chang-Ben Huanga, Te-Hsuan Janga, Hsu Hong-Yia, Horng-Dar Wangb, Chiou-Hwa Yuha,c,d,* a

Division of Molecular and Genomic Medicine, National Health Research Institutes, 35 Keyan Road, Zhunan Town, Miaoli County 350, Taiwan, R. O. C b College of Life Science and Institute of Biotechnology, National Tsing-Hua University, Hsin-Chu, Taiwan, R. O. C c College of Life Science and Institute of Bioinformatics and Structural Biology, National Tsing-Hua University, HsinChu, Taiwan, R. O. C d Department of the Biological Science & Technology, National Chiao Tung University, Hsin-Chu, Taiwan, R. O. C # These authors contributed equally * Author for correspondence: Chiou-Hwa Yuh, email: [email protected] Received 14 Jan 2011; Accepted 28 Jan 2011, Available Online 28 Jan 2011

Abstract The genomic program for development operates mainly by the regulated expression of genes encoding transcription factors and the signaling pathways. Complex genetic regulatory networks control developmental fate. In this study we built the mesendoderm GRNs in zebrafish by integration of the spatial and temporal expression pattern with the interaction relationship from literatures searching and perturbation analysis. We established the subcircuits between specific transcription factors using morpholinos against them, and measured certain gene expression profiles by real time quantitative RT-PCR (Q-PCR) and validation by in situ hybridization. From our experiments, we confirmed some interactions identified from literature, and we also identified new target genes downstream of those candidates, which were not found in the literatures search. Some of the interaction nodes were identified through computational searching of the conserved non-coding region, and validated by functional analysis. This is the systematic approach towards transcriptional regulation in zebrafish development and it gives new insight to evolution of deuterostome. Keywords: Transcription regulation; Gene regulatory network; Zebrafish endoderm development; In situ hybridization

1. Introduction Gene regulatory networks (GRNs) operate by the coordinated expression of genes involving transcription factors and the components of signaling pathways. Deciphering GRNs is essential to understanding the developmental process, cancer formation, pancreatic and liver specification, stem cell networks and other biological phenomena [1,2]. With the genomic information available for many species, systems biology has become an increasingly important means to understand the secret of life in the post-genomic era [3]. One of the best examples of GRNs came from sea urchin and the study of endoderm specific gene, Endo16. The regulation of the Endo16 is controlled by a series of transcription factors in a logical manner, and the interplay of those transcription factors forms a network [4,5]. Multiple transcription factors regulate each other and form precise networks in development [6]. GRNs are dedicated systems that integrate many

specific motifs to adjust the spatial and temporal expression patterns of genes according to the inputs and the supplied outputs [7]. Zebrafish (Danio rerio) is an excellent model organism for studying development [810], carcinogenesis [11-13], organogenesis [14,15], vasculogenesis [16], and many other biological processes [17-19]. Coordinated interplay of maternal factors and downstream genes is important for axis formation [20], endoderm formation [21,10], and muscle development [22]. Nodal, an important molecule for endoderm formation [23,24], is a novel member of the transforming growth factor-beta (TGFbeta) superfamily [25]. Nodal is first expressed in primitive streak-stage embryos at the time of mesoderm formation. It then becomes highly localized in the node at the anterior of the primitive streak. This region is analogous to Hansen’s node in the chicken (Gallus gallus), and dorsal lip in both Xenopus and zebrafish [26]. The Nodal pathway is conserved among

Global Journal of Biochemistry | Volume 2 | Issue 4 | 2011 www.simplex-academic-publishers.com © 2011 Simplex Academic Publishers. All rights reserved.

215

GLOBAL JOURNAL OF BIOCHEMISTRY zebrafish, Xenopus, and human in endoderm and mesoderm differentiation, and is associated with vertebrate gastrulation and axial patterning [27]. Nodal-related protein 1 (Ndr1) and Nodalrelated protein 2 (Ndr2) have important roles in organizer development and in the formation of mesoderm and endoderm in zebrafish [28]. By cell transplantation, the activation of Nodal signaling is sufficient to commit cells to both endodermal fate and behavior in zebrafish [29]. The sustained Nodal signaling is required to ensure endoderm formation, but transient Nodal signaling is sufficient for mesoderm formation [30]. Nodal and the receptor interaction initiate phosphorylation and activation of Smad2/3 to bind to Smad4. The Smad2/3/4 complex in combination with other specific partners like FoxH1 and Bon allows the transcriptional activation of some specific target genes [31,32]. Nodal signaling operated several endoderm-specific transcription factors. Early endoderm formation in zebrafish requires transcription factors downstream of Nodal signaling, including bon, gata5, og9x, pou5f1, sox32 and sox17 [24]. Mutation of those genes caused an endodermal tissue defect in zebrafish [33], bon (Bon/Clyde), gata5 (gata5/faust, fau), and sox32 (sox32/Casanova, cas) were identified downstream of Nodal signaling but upstream of the early endodermal marker sox17. Embryos deficient in sox32 lack endoderm and develop cardiac bifida [34]. SRY-box-containing gene 32 (sox32), expressed in endoderm, was cloned in a subtractive screen for Nodal-responsive genes [35,36]. The knockdown of sox32 using morpholino antisense oligonucelotide results in a lack of sox17 expression in endodermal precursor cells during gastrulation. sox32 induces endodermal markers and represses mesodermal markers during gastrulation. The over-expression of sox32 restores endoderm markers in the absence of Nodal signaling [30]. In addition, Bon is required for the expression of sox32. This revealed that sox32 is located downstream of Nodal signaling and upstream of sox17. Previously, we the used perturbation functional analysis of the sox17 transcription unit and found the Sox32 and Pou5f1 input are indeed encoded in the DNA sequence of the sox17 cis-regulatory element [37]. The activation of sox32 depends upon gata5, bon, and eomesodermin [38]. Genes og9x, bon and sox32 are immediate early targets of Nodal signaling to specify the endoderm lineage by activating downstream genes such as sox17 and foxa2 [33], gata5 is expressed in the endodermal progenitors from the late blastula stages. The gata5 mutant expresses lower levels of sox17 and foxa2 than wild type. Using a

complementary mutant and over-expression analysis, it was shown that gata5 and bon are required for sox32 activation in endoderm formation [38]. Eomesodermin, a maternal Tbox protein, acts with gata5 and bon to regulate the expression of sox32 via an element located 1476 bp upstream of the translational initiation site of sox32 [39]. The regulation of endoderm formation by eomesodermin does not require Nodal signaling or og9x [39], which is a novel paired-like homeobox protein that depends on a functional Nodal signaling pathway. It was demonstrated that og9x, bon, and sox32 are all immediate early targets of Nodal signaling, while sox17 requires protein synthesis in order to be induced by expressing a constitutively active Nodal receptor in the presence of translation inhibitors [33]. These results highlight the complexity of the transcriptional network operating during endoderm formation. As mentioned above, foxa2 is one of the downstream target genes of Nodal signaling. In situ hybridization shows that foxa2 is first expressed at the dorsal side of hypoblast just before gastrulation [40]. Later on, foxa2 is expressed in endoderm and axial mesoderm at 8 hours post fertilization (hpf). In adult fish, foxa2 expresses in multiple territories including gut, liver and pancreas of endoderm and floor plate of ectoderm [41]. Otx genes, cognates of the Drosophila head gap gene Orthodenticle, are highly conserved in all vertebrates. The Otx gene family plays the crucial role in head development. In mouse, Otx2 plays multiple roles in each site of development including epiblast, anterior visceral endoderm, anterior mesendoderm, anterior neuroectoderm, forebrain/midbrain, and cephalic neural crest cells [42,43,44]. Gbx2 is a well-known transcription factor which regulates anterior hindbrain development to antagonize Otx2 which regulates midbrain development, and this mechanism is similar in several species (mouse, chicken and zebrafish). The interface between Gbx2 and Otx2 creates a signal center to control the local brain patterning. Gbx1 is expressed earlier than Gbx2 at a region adjacent to the posterior part of Otx2 at gastrulation, and Gbx2 will be activated later and lying near Otx2. Gbx1 probably acts with ectoderm or endomesoderm genes at gastrulation to establish the isthmus organizer and then activates Gbx2 to maintain MHB. However, it is still unclear how Otx2 and Gbx2 boundary are settled and how these genes regulate the downstream genes to promote the brain segmentation. Global Journal of Biochemistry | Volume 2 | Issue 4 | 2011 www.simplex-academic-publishers.com

216

© 2011 Simplex Academic Publishers. All rights reserved.

GLOBAL JOURNAL OF BIOCHEMISTRY Although there have been many studies of zebrafish development, there are very few integrated studies regarding the gene regulation networks in zebrafish. Nevertheless, wide varieties of zebrafish maternal and zygotic regulatory factors and signaling pathways have been discovered. We have recently built the zebrafish GRNs [45], and integrated the evidence of interaction in the literature and the spatial/temporal expression patterns extracted from the Zebrafish Information Network (ZFIN) [46]. Earlier work on zebrafish development produced an enormous amount of data, but the systematic study of the endoderm transcription factors is lacking and is the objective of this study. We established the network architecture on the basis of the results of morpholino perturbation coupled with expression profiles in Q-PCR and whole-mount in situ hybridization. This is systematic approach towards transcriptional regulation in zebrafish development. 2. Experimental 2.1. Zebrafish strains Embryos were maintained and staged, and adult zebrafish were maintained at 28.5°C essentially as described [46a], but with slight modification. Natural spawning provided the embryos used in this study. The AB strain zebrafish were purchased from Zebrafish International Resource Center (ZIRC) at the University of Oregon, Eugene, OR. 2.2. Morpholino injections The morpholino oligonucleotides were injected at different dosages according to the titration result. For sox32, we injected 4.6 ng/embryo, 9.2 ng/embryo for sox17, 5 ng/embryo for gata6 and 15 ng/embryo for gata5. We injected the morpholino into the yolk at the one-cell or two-cell stage. Embryos were injected at one-cell stage with MO by Nanoject Ⅱ (Drummond Scientific Co., Broomall, PA). After injection, embryos were observed under Leica DMIRB Inverted Fluorescence Microscope coupling with CoolSNAP TM Cooled CCD Camera (Roper Scientific, Trenton, NJ). The sequences of the morpholinos used are: sox32-MO, 5' CGGTCGAGATACATGCTGTTTTGCG3'; sox17-MO, 5' CGCATCGGGACTGCTCATCTCAAAC3';

gata5-MO, 5' ATCCAGTGAATAAGCTAGATTTCGA3'; gata6-MO, 5' CATGTCGGTAGGGTCCACAATGAGG3'; otx2-MO, 5' TTGTTTGCGTCTTCAGCGGTGGAGG3'; foxa2-MO, 5' ATTTTGACAGCACCGAGCATCCTGG3'; gbx2-MO, 5' ACGGTGTGCTGAAAGCTGCACTCAT3'; gbx1-MO, 5' ACCGCTCGGTCTCTGCATCGTGAAC3'; six3b-MO, 5' AGTCGCCGGACAGAGACAAAACGAC3'; 2.3. RNA extraction and Q-PCR Total RNA from staged embryos was extracted with an RNeasy Mini kit (catalog #74106, QIAGEN Co. Valencia, CA). Twenty embryos were collected at 5 hpf, 8 hpf, 11 hpf, 16 hpf and 24 hpf. For RT-PCR, cDNA synthesis was performed using MultiScribeTM Reverse Transcriptase (Applied Biosystems, Foster City, CA). In a 50 l reaction mixture, we added 900 ng of total RNA together with the following reagents from the kit: 10 TaqMan RT buffer (5 l), 25mM MgCl2 (11 l), DeoxyNTP mixture (10 l), Random Hexamer (2.5 l), RNase Inhibitor (1 l) and MultiScribe Transcriptase (50 U/ l) 1.25 l). The cycling parameters were: initial incubation at 25oC for 10 minutes, extension at 48 oC for 30 minutes, denaturation at 95 oC for 5 minutes, and stop at 4oC. Real time reverse transcriptase quantitative PCR (Q-PCR) was carried out with Sybr Green (Applied Biosystems) in an ABI PRISM 7900 sequence detection system (Applied Biosystems). In 10 l of reaction mixture, we added 3.8 l of 20-fold diluted cDNA, which is equivalent to 0.0095 embryos (see calculation below) with the following reagents: 1.25 pmol/l of primer pairs (1.2 l), 2 SybrGreen Mixture (5 l). The cycling parameters were: 50 oC for 2 minutes and 95oC for 10 minutes, and then 40 cycles of denaturation at 95 oC for 15 seconds, and annealing and extension at 60 oC for 1 minute. Twenty embryos were used to generate the RNA, and about 2.5 embryos were used in 50 l volumes to generate the cDNA, and diluted 20-fold for Q-PCR experiment. We thus have the cDNA of about 0.0025 embryo /l. We used 3.8 l of diluted cDNA for each Q-PCR reaction, equivalent to 0.0095 embryo ((2.5 embryos/50)/20 = 0.0025 embryo/μl; 3.8 μl 0.0025 = 0.0095 embryo). We selected 36 transcription factors and designed corresponding


217

GLOBAL JOURNAL OF BIOCHEMISTRY Table 1. The sequences of the Q-PCR primers of the candidate genes. Forward Reverse

ID

Gene

1

smad2

CTCCTGCAGAACTGTCACCA

GAATTGGAGGGGTCTGTGAA

2

foxH1

CCGTCAGGATGAAACCATCT

CCTGAGGCTATGCAGGAGAG

3

og9x

ACCCGGAGATTCGTCTTTCT

GCAGAGCTTCCTCCAAACTG

4

bon

ACCCAAAGGAACACGAACAG

AACAGGCGATGTGTTTAGGG

5

gata5

CGACAACACTGTGGAGGAGA

GTAGCGCCAGACACTGTTGA

6

sox32

GGTGTCTGTCGGGTCTGAAT

TCTGCTTCTTGCATGTACGG

7

sox17

CCGCTCTCAGACTCCAAATC

TTACTCAGCTCCGCATTGTG

8

foxa2

ATGTCCGGAATGAGTTCGAC

AGCGTTCATGTTGCTGTACG

9

otx2

ACGTTGTAGACCTGCCATCC

ATACGCATTCACGACCCTTC

10

gbx1

AAGCCGGAAATGTCAACAAC

GCTGTGCTTCATGCATCTGT

11

pax2a

GCTGATGCCTTCACTCAACA

AATCTGAGCTGACGCTGGTT

12

pou5f1

GCTTAAACACAAGCGCATCA

ATCCTGAGGGTTCTCGGAGT

13

gbx2

CTGCTCCCAGACTTCAGGAC

GAAGCACCACTGACCGGTAT

14

cdx4

CAACACGCTGAAAGGACAGA

TTCAAGAGCCTCCAGCATTT

15

drl

GGGAAGGGTTTCCCTTATGA

GAAACTCATCCCGCACTGAT

16

etv6

ACACTACCGCAACCACATCA

ATGCGGAAGACTTTGGTGTC

17

hey1

CGTCACTTTCAAGCCTGTCA

GCGTTCGTGTTGAGTTCAGA

18

irx3a

CAGGACGAACTGCATTCTGA

CTGTTGTGGTGGAGGTGATG

19

mybbp1a

AGAGCTGCACAGACCCTGAT

TCTACTGCTGCCTCCTGGTT

20

otx1

GTGCGTCATACCAACATTGC

CACTCCGTCCTTCTCCTCTG

21

prdm1

ATCGTGGCCTGAACCACTAC

GAGGAGGTACCAAGCTGCTG

22a

six3a

TGCCACTGGAAGCGACCATA

TTATCCCTCTCATTTCTTCCTGCCA

22b

six3b

CTCAAGTGGGCAATTGGTTT

ATACTGGAGACGCTGGTCGT

23a

tbx16

TGGGAACTCATCCGCTTTAC

GCTCACCAGCACGAGTATGA

23b

tbx24

GACTGCGAGGTCCTTCTCAC

TCCCGGTGTCTTCTATGGAG CTGTTTGAAGCCCTGAGGAG

24

etv5

CACCCTTTGTAGTGCCCTGT

25

foxA1

CTGCTTCGTCAAAGTGTCCA

CGCTGGACTGCTCTCTCTTT

26

gata6

ATGCTGCACTCACTGCAATC

AGACATGCGCTTCTGAGGTT

27

trh3

TACTTCTTCAACGGGCTGCT

CGTTCAGGACTGGTCCATTT

28

sox4b

CAGCGGCCACATCAAGAGA

CGGATGAACGGGATCTTGTC

29

brn1.2

CTCCATCGAGGTGAGTGTCA

GTCATCCGCTTCTCCTTCTG

30a

pou1

CTTCAACTGGAGCGGTTTG

TGGCGCTTGAAGAAGTTGTA

30b

pou12

AGACCACCATTTGCAGGTTC

CTTTCGTTTCCTGCCTTGAG

31

gata4

TGCAGAAGGAGAGCCAGTCT

ATGCTGGAAACGCAGATACC

32

otx5

ACTGCAGCTCGTACCTGTCC

TCCAGGAATTCTGGTCCTTG

33

foxa3

CCGCCACTCTCTATCCTTCA

CTGCTGCCATCCTGAGACTT

C1

actin

CTCCATCATGAAGTGCGACGT

CAGACGGAGTATTTGCGCTCA

C2

18s rRNA

CCGCAGCTAGGAATAATGGA

CATCGTTTACGGTCGGAACT

Thirty-six transcription factors that are expressed on mesendoderm at early stages of zebrafish embryo development were selected for Q-PCR. Actin and 18 S rRNA were used as internal controls for comparison between different batches. The primers were designed to correspond to the selection from the Primer3 program.

primers based on the Primer3 program (v.0.4.0) (http://frodo.wi.mit.edu/primer3/).The nucleotide sequence for each pair of Q-PCR primers used in this study is given in Table 1. The calculation of relative fold change was performed as follows:

The rate of PCR amplification in control versus specific morphants (after normalizing against housekeeping genes actin or 18 S rRNA) reflects differences in expression of the gene in the morpholino knockdown embryos. For Global Journal of Biochemistry | Volume 2 | Issue 4 | 2011 www.simplex-academic-publishers.com

218



Table 2. The sequences of the primers for the candidate genes used for in situ hybridization. Gene

Forward

Reverse

foxh1

AGAGGGGTGACAGTTGTTGG

AAAGGGAAGGTCCGTTCTGT

og9x

AGCACATGCTGTCCTCTCCT

GAATCTCCGGGTGTTTCAGA

bon

GCTTGCCAGAATCGAGAATT

AACAGGCGATGTGTTTAGGG

gata5

TACCCACCACTCGTGTACCA

CTCCTTCCCGGTAGAGTTCC

sox32

GCTATATGTTTCAAGGTCAAGC

AAGCAGCAGTCATCATTCAAGT

sox17

CGTCTACGAGTCAAACATATGCA

CTGCTAACACAATGTGACAACTC

foxa2

ACGGCCAGTCGAACATAAAC

CGACTCGGATCCTTGCTTAG

otx2

AGCACAATGGCAGTCGTGCAC

CCTCAAAGACTAAATGAAGCCAT

gbx1

GCCCTGTCTCATTCGTCATT

CGCCTTTCAGTTTGCCTTT

pax2a

AAAGTTCAGCAGCCTTTCCA

TGCTCTGGCTTGATGTGTTC

drl

GCAGAACAGAGCATCACAAT

ACTTCCAGTGCGGCAAT

etv6

GCAGCAGCTCATTATTAAGGAGA

ATGAAGGTGGTGATGATGATGA

six3a

TCTTCCTGCCAAACTTTGCT

CCAGTTCCCTTTTCTTGCTG

six3b

CCTGGAAACCCACAAGTTCA

CGCCTCGTTCATTCATACTG

tbx16

GAGAAGGCTAAATAATGCAGGCT

GCTCACCAGCACGAGTATGAG

tbx24

ACCCAAACAGCCTTTCCTTT

GGTGGATTGAATTTGGGATG

etv5

ATGGACGGAT TTTATGACCA

TTCCCTTCAAGTCTCTCTGG

foxA1

CATGGACTCCAGCTCCATCAC

ACAGGCCTGGAATACACACC

gata6

GTTATCAACACAGCCGAGCA

GGCGTCATCCTCAAGAGAAG

sox4b

CGAGGAAGAAGGTGAAGTCG

GAGGAGAACTGCGATG

pou1

CTTCAACTGGAGCGGTTTG

ACCGTCCTCCAGTCTGTGTT

gata4

GGAGCTGCGTCTTACGAGTC

CTGTGCAGTACGGAGCTGTC

The gene was selected according to the Q-PCR result and the primers were designed to correspond to the selection from the Primer3 program. The Sp6, T3, or T7 promoter sequence was added to the 5’-end of each primer.

calculation of the absolute molecules of RNA, we use the known DNA as our standard to calculate the molecules from the cycle number. Each Ct (cycle number at threshold) can be converted into molecules according to the standard curve, and then converted into molecules per embryo by dividing with the actual embryo number. 2.4. Whole-mount in situ hybridization The DNA templates for in situ hybridization probes were generated by PCR amplification. The primers used for generation of the DNA template used for the in situ probes are given in Table 2. Sp6, T3, or T7 promoter sequence was added to the 5’-end of each primer. The DNA sequences of those promoters are:

SP6: 5’ATTTAGGTGACACTATAG3’ T3: 5’AATTAACCCTCACTAAAGGGAGA3’ T7: 5’TAATACGACTCACTATAGGG3’ To prepare the DIG-labeled antisense RNA probes, we used either T7 or T3 RNA polymerase (catalog #600123 and #600111, Stratagene, La Jolla, CA) or Sp6 MEGAscript Kit (catalog #1330, Ambion, Austin, TX). DIG RNA-labeling mixture was purchased from Roche (catalog #11277073910, Penzberg, Germany). Embryos were collected at the desired time-points. Embryos were fixed with 4% paraformaldehyde overnight, and then dehydrated in methanol at –20oC. The protocol for whole-mount in situ hybridization was adopted from Thisse B and Thisse C at ZFIN (http://zfin.org) with some modification. Embryos were rehydrated gradually with PBST


219

GLOBAL JOURNAL OF BIOCHEMISTRY (PBS, 0.1% Tween 20). For embryos older than 24 hours, 10 g/ml proteinase K (P-2308, Sigma-Aldrich, St. Louis, MO) in PBST was used for digestion for 12 minutes. It was stopped by incubation in 4% paraformaldehyde in PBS for 20 minutes, and then rinsed 5 times for 5 minutes in PBST. The embryos were incubated at 65 oC overnight in HYB+ (50% formamide, 5 SSC, 0.1% Tween 20, 50 g/ml heparin and 500 g/ml wheatgerm tRNA (R-7876, SigmaAldrich)), which contained 0.5 ng/μl of DIGlabeled antisense RNA probe. The embryos were washed for 15 minutes at 65 oC with: 75% HYB, 25% 2 SSC, 50% HYB–, 50% 2 SSC, 25% HYB–, 75% 2 SSC, and 2 SSC. HYB– is 50% formamide, 5 SSC, 0.1% Tween 20. The embryos were then washed twice with 0.2 SSC for 30 minutes at 65 oC, followed by gradual transfer to 2 blocking buffer (100 mM maleic acid, 150 mM NaCl, 0.1% Tween 20, 2% Blocking Reagent (catalog # 1096176, Roche) for at least 1 hour. The embryos were incubated overnight at 4oC with 1:5000 diluted antidigoxigenin-AP (catalog # 11093274910, Roche) in 2 blocking buffer. After 6 washes with 100 mM maleic acid, 150 mM NaCl, 0.1% Tween 20 each for 15 min and 3 washes with alkaline Tris buffer (100 mM Tris–HCl (pH 9.5), 50 mM MgCl2, 100 mM NaCl, 0.1 % Tween 20) for 5 min, bound antibody was detected with NBT/BCIP solution (catalog # 11681451001, Roche) in alkaline Tris buffer. After staining, the labeled embryos were mounted in 80% glycerol. We observed the embryos under a stereomicroscope (SZX-ILLD100, Olympus, Tokyo, Japan) with an Olympus DP70 Digital Microscope Camera to capture the images. 2.5. Establish gene regulatory network by BioTapestry interactive network viewer We used BioTapestry Editor (version 5.0.1) to build the gene regulatory network. The software is available at http://www.bio tapestry.org. Our BioTapestry model uses the framework described earlier [47], which is an interactive tool for building, visualizing, and simulating genetic regulatory networks. 3. Results 3.1. Gene expression profiles of 35 transcription factors in zebrafish development To understand the gene regulatory networks control the mesendoderm specification, we searched the zebrafish information networks (ZFIN, http://zfin.org) for the transcription factors expressed in mesendoderm between zygote and gastrula stage, and identified 35 transcription factors. We first examined the

temporal expression profiles of those selected genes by Q-PCR with RNAs isolated from embryos at 5, 8, 11, 16, and 24 hours postfertilization (hpf). There are smad2, etv6 and trh3 expression appears earliest of all six groups (Figure 1), we assigned them to group I. Group II contains genes express high at 5hpf, and then gradually decreased. The genes in group II are pou5f1, drl, bon, og9x, sox32, tbx16, foxa3 and otx1. Many of them have been known to be downstream of Nodal and are responsible for mesendoderm specification, such as pou5f1, drl, bon, og9x and sox32. Group III contains three transcription factors: gbx1, gata6 and sox17 which expressions peak between 8 hpf and 11 hpf, and almost vanish after 16 hpf. Group IV contains ten transcription factors which expressions peak at 11 hpf and then gradually declined. Group V contains seven genes which starts to express at 11 hpf and lasts until 24 hpf, sox4b belongs to this group. This suggests both sox32 and sox17 function in early zebrafish embryonic development, whereas sox4 has a late function. Group VI contains genes which has significant drop at 8 hpf, and expressions peaks at 11 or 16 hpf, and stay high until 24 hpf. The temporal expression profile of those transcription factors corresponds to their biological function. 3.2. Different developmental defects of Sox32 and Sox17 morphants by morpholino antisense oligonucleotides (MO) in zebrafish To establish the subcircuits of those transcription factors, we used morpholino antisense oligonucleotides (MO) to block the translation of those mRNA in zebrafish embryos. Titration experiment of sox32 MO and sox17 MO were shown as an example (Figure 2) and all of the morpholinos used in this study were titrated with the same manner. Different amounts of sox32 MO (2.3, 4.6, 6.9, and 9.2 ng per embryo) were injected into one cell stage embryo. Three typical phenotypes of the sox32 morphants are shown in Figure 2A. Class I, the embryo is slightly different from the control morphants on ventralized body. Class II, the yolk sac almost disappears, pericardial edema is obvious and the midbrain-hindbrain boundary is not clear. Class III, the most severe phenotype changes with more expanded swollen pericardial sacs and shortened body trunk. The titration results of sox32 MO are given as a bar graph in Figure 2B. At 2.3 ng of sox32 MO, about 7% of sox32 morphants look normal, 83% show the class I phenotype and 10% have the class II

Global Journal of Biochemistry | Volume 2 | Issue 4 | 2011 www.simplex-academic-publishers.com

220



Figure 1. Expression profiles of 35 transcription factor. The X axis represents hours after fertilization, and the Y axis represents RNA molecules/embryo defined by Q-PCR. The 35 transcription factors are expressed at different stages in zebrafish embryos. RNA was isolated from embryos at 5 hpf, 8 hpf, 11 hpf, 16 hpf, and 24 hpf, and their absolute molecular number per embryo was verified. The data came from 32 different batches of experiments and the standard deviation is shown as bar extending to both sides of the mean.


221


Figure 1. (continued)

222


Figure 2. Morphology of sox32 and sox17 MO injected embryos. Dose-dependent sox32 knockdown was done in different titrated injections (2.3, 4.6, 6.9, and 9.2 ng) of sox32 MO/embryo. Dose-dependent sox17 knockdown was done in the same titrated injections of sox17 MO/embryo. According to the severity of the defect, there are three classes for sox32 morphants. (A) Selected images of sox32 morphants, from normal to Class I, Class II, and Class III. (B) The bar graph shows the percentage of each class of phenotype at different doses of sox32. The dark blue bar indicates the normal type; red represents Class I, green represents Class II, and purple represents Class III. The light blue bar indicates the hatching rate for each amount of sox32 morpholino injected. Increased amount of morpholino injected was associated with increased severity of the defect. We chose 4.6 ng of sox32MO for the rest of the experiment. (C) Selected images of sox17 morphants, from normal to Class I, Class II, and Class III. (D) The bar graph shows the percentage of each class of phenotype at different doses of sox17: dark blue, normal type; red, Class I; green, Class II; purple, Class III. The light blue bar indicates the hatching rate for each amount of sox17 morpholino injected. Increased amount of morpholino injected was associated with increased severity of the defect. We choose 9.2 ng of sox17 MO for the rest of the experiment.

phenotype. With 4.6 ng of sox32 MO, no morphant looks normal, 53% have the class I phenotype, 45% have the class II phenotype and 2% show the class III phenotype. When the dosage of the sox32 MO is increased to 6.9 ng or 9.2 ng, the severity of the morphants becomes more obvious, as shown in Figure 2B. Together, the knockdown of sox32 by MO produces different levels of developmental defects in a dosage-dependent manner. Thus, we chose 4.6 ng per embryo as the dose for sox32 MO injection. The injection of 4.6 ng of sox32 MO into zebrafish embryos resulted in several specific developmental defects at different stages, compared to control morpholino-injected embryos, which develop normally. The earliest time of the development defect starts at 24 hpf and the sox32 morphants appear incomplete in the head formation with no clear midbrainhindbrain (MHB) boundary. At 36 hpf, they have a thickened yolk sac and lack endodermal

derivative tissue such as anus, intestine and liver. The heart appears to have pericardial edema and the brain defect remains. This phenotype resembles those of the mutant sox32 reported earlier [48]. After 48 hpf, the defects become much more severe and the embryos cannot hatch normally. sox17 identified in many species is expressed during gastrulation exclusively in the endoderm region [49]. By sox32/Casanova mutant analysis and sox17 cis-element assay, sox17 has been shown to be a direct target gene of sox32 [35]. We also identified a cis-element on sox17 regulatory region for the direct binding of Sox32 and activating sox17 transcription [37]. We next examined the effects of the knockdown sox17 mRNA translation by the injection of sox17 MO. We found injection of sox17 MO to embryos generates three classes of different morphology types, similar to those of sox32 knockdown (Figure 2C, classes I, II and III). Titrated doses of sox17 MO from 2.3, 4.6,


223

GLOBAL JOURNAL OF BIOCHEMISTRY and 6.9 to 9.2 ng per embryo were injected (Figure 2D). With 2.3 ng of sox17MO, 36% of sox17 morphants were normal and 64% showed the class I phenotype. With 4.6 ng of Sox17 MO, 27% of morphants had the class I phenotype, 53% showed the class II phenotype, and 20% of embryos displayed the class III phenotype. When the sox17 MO was increased to 6.9 ng and 9.2 ng per embryo, the morphants become much more severe, as shown in Figure 2D. Besides this, the class III morphants had some morphology changes typical of the sox32 morphants, including a thickened yolk sac, and a lack of intestine and endodermal tissue formation. The morphology defects conserved between the sox32 and sox17 morphants were similar in head, yolk sac, pericardial edema, lacking of intestine, and endodermal tissue. This implies that the two transcription factors may coordinate each other or sox17 has a downstream sox32. Other morpholino effect on oligonucleotides were injected at different dosages according to the titration methods described above. 3.3. The effects of sox32 knockdown on gene expression involved in three germ layer specifications sox32 is expressed in the yolk syncytial layer and mesendoderm at the early stages, identified by in situ hybridization [36]. Several genes have been identified as the sox32 downstream target genes, including ntl, sox17, bmp2b, gata2, gata5, tbx6, gsc, foxa1, foxa2, foxa3 and shha [39,35,33,36]. To establish the subcircuit for sox32, we selected 35 candidate transcription factors, and examined their expression upon sox32 knockdown. After injection of sox32 MO into one-cell stage embryos, we collected control and sox32 morphants at 5, 8, 11, 16, and 24 hpf for RNA isolation and Q-PCR. Six genes were upregulated and eight were down-regulated, as shown by Q-PCR after knockdown of sox32 (Table 3). The up-regulated genes sox17, otx2, foxa2, pax2a, pou1 and sox32 itself, are important transcription factors for early embryo development. We also used an in situ hybridization assay to verify the Q-PCR result by comparison of the expression pattern of control and sox32 MO. As shown in Figure 3, sox17 expression was down-regulated five-fold in the embryos with the knockdown of sox32 (Figure 3 A1) at 8 hpf. It is verified that there is a down-regulation of sox17 in the sox32 morphants at at 8 hpf (Figure 3 A3) in the endodermal lineage, as determined by in situ hybridization. We found sox32 activates itself at 5 hpf, and sox32

expression is lower in the sox32 morphants at the dorsal lip, where sox32 is initiated (Figure 3B). The expression of otx2 at 8 hpf was decreased in the sox32 morphant (Figure 3C). We detected a slight decrease of foxa2 gene expression in the endoderm of foxa2 but there was no effect on ectodermal expression in the 8 hpf embryos of the sox32 morphants (data not shown). foxa2 was expressed exclusively in the floor plate and the neuroectoderm, which is not affected by sox32 morphants at 11 hpf (data not shown) in later development. sox17 and foxa2 are essential for endoderm formation and are key components of Nodal signaling [21], otx2 is a downstream factor of the beta-catenin, Fgf, and Nodal signal pathways [35,50]. Its role in the development of mesendoderm is further investigated in our research in the conserved kernel consisting of otx2, gata5, and gata6 (submitted for publication elsewhere). Spatially, the expressions of sox17 and foxa2 in the presumptive endoderm region are much less at 8 hpf in embryos of sox32 morphants. Temporally, the peak of sox17 and foxa2 expression is after that of sox32. Thus, it is possible that sox32 initiates the expression of those two genes in the very early stage of endoderm formation. In summary, sox32 activates itself at 5 hpf at presumptive in endoderm and induces sox17 and foxa2 at 8 hpf endoderm lineage. sox32 activates sox17 expression in the endoderm continuously at 11 hpf. The expression of og9x at 8 hpf on the mesendodermal lineage was decreased in the sox17 morphant (Figure 3D). In addition, sox32 activates pax2a (Figure 3E), which is involved in MHB formation, and sox32 activates pou1 (Figure 3F), which is involved in formation of the central nervous system. Normally, Sox32 expression becomes very low after 11 hpf, and is located in the endoderm. pax2a is expressed in the midbrain and hindbrain boundary [51], and pou1 is expressed in the central nervous system [51], which is different from where sox32 is expressed. Therefore, the activation of pax2a and pou1 by sox32 may be indirect through other signaling pathways. 3.4. The genes repressed by sox32 at later stages sox32 negatively regulated many important transcription factors, including drl, og9x, bon, gbx1, pou5f1, gata6, and sox32 itself, at later stages (Table 3). Some of the genes were expressed in the endoderm area, and others were expressed in different areas during embryonic development. og9x, bon, and sox32, early zygotic expression transcription factors, have Global Journal of Biochemistry | Volume 2 | Issue 4 | 2011 www.simplex-academic-publishers.com

224



Table 3. Downstream target genes of sox17, sox32, foxa2, gata6, gata5, otx2, six3b, gbx2, and gbx1.

5hrs

5hrs sox32 (+) sox17 (+)

8 hrs og9x (-)

sox17 downstream genes 8 hrs 11 hrs 16 hrs og9x (+) sox32 (+) sox17 (-) foxa1 (-) sox17 (-)

downstream genes 11 hrs 16 hrs sox17 (+) sox32 (-) pax2a (+) pou1 (+) gbx1 (-) pou5f1 (-) drl (-) gata6 (-) bon (-) sox32 (-) foxa2 downstream genes 11 hrs 16 hrs 24 hrs sox4b (+) foxa2 (-) gbx1 (+) foxa2 (-) brn1.2 (+) pou1 (+) pou12 (+) shhb (+) nk x2.2b (+) ntn1b (+) gata4 (-) foxa2 (-) sox32 8 hrs sox17 (+) otx2 (+) foxa2 (+) drl (-) og9x (-) bon (-)

24 hrs drl (-) etv6 (-) foxh1 (-) six3a (-) six3b (-) etv5 (-) foxa1 (-) gata6 (-) trh3 (-) gata4 (-) gata5 (-) sox17 (-) 24 hrs drl (-) sox32 (-)

36 hrs gbx1 (+) brn1.2 (+) pou1 (+) pou12 (+) foxa (+) shha (+) shhb (+) nk x2.2a (+) nk x2.2b (+) ntn1b (+) gata4 (-) foxa2 (-)


225


5hrs tbx16 (+) foxa3 (+) gata5 (+) foxa2 (+) sox32 (+) og9x (+)

5hrs irx3a (+) gata6 (+) sox32 (+) trh3 (-) 5hrs gata6 (+) foxa3 (+) gata5 (+) sox17 (+) foxa2 (+)

gata6 downstream genes 8 hrs 11 hrs 16 hrs (+) pou1 gbx1 (+) gbx1 (+) cdx4 (+) pou5f1 (-) pou1 (+) (-) (+) foxh1 pou12 prdm1a (+) sox17 (+) gata6 (-) pou5f1 (-) irx3a (-) bon (-) foxh1 (-) gata6 (-) sox32 (-) sox4b (-) drl (-) gata4 (-) foxa3 (-) gata5 (-) sox32 (-) sox17 (-) gata5 downstream genes 8 hrs 11 hrs 16 hrs gata4 (+) sox17 (+) gata5 (-) sox17 (+) gata5 (-) gata5 (-) otx2 downstream genes 8 hrs 11 hrs 16 hrs bon (-) brn1.2 (+) pou1 (+) otx2 (-) pou1 (+) pou12 (+) pou12 (+) pax2a (-) bon (-) gbx2 (-) otx2 (-) foxh1 (-) prdm1a (-) six3a (-) six3b (-) tbx16 (-) tbx24 (-) gata6 (-) trh3 (-) sox4b (-) gata4 (-) foxa3 (-) bon (-) sox32 (-) sox17 (-) otx2 (-)

24 hrs gbx1 (+) drl (-) mybbp1a (-) foxh1 (-) tbx16 (-) gata6 (-) sox4b (-) gata4 (-) foxa3 (-) gata5 (-) sox17 (-)

24 hrs gata5 (-)

24 hrs gbx1 (+) brn1.2 (+) pou12 (+) mybbp1a (-) foxh1 (-) tbx16 (-) tbx24 (-) otx2 (-)


226



six3b downstream genes 5hrs 8 hrs 11 hrs 16 hrs otx1 (+) mybbp1a (+) gbx1 (+) gbx1 (+) mybbp1a (-) foxh1 (+) pax2a (+) pax2a (+) otx2 (+) gbx2 (+) otx1 (+) etv5 (-) otx1 (+) pou1 (+) og9x (+) prdm1a (-) otx2 (+) six3b (-) tbx16 (-) gata6 (-) brn1.2 (-) foxa3 (-) bon (-) gbx2 downstream genes 5hrs 8 hrs 11 hrs 16 hrs bon (-) gbx2 (+) gbx2 (+) six3b (+) pou1 (+) pou12 (+) foxh1 (-) fgf8 (+) sox32 (-) eng2a (+) otx2 (-) sox32 (-) sox17 (-) gbx1 downstream genes 5hrs 8 hrs 11 hrs 16 hrs cdx4 (+) og9x (+) foxa (+) gbx1 (-) six3a (+) foxa (+) fgf8 (+) shha (+) gbx1 (-) twhh (+) fgf8 (+) sox32 (+) gbx1 (-) foxa1 (-)

24 hrs pax2a (+) hey1 (-) irx3a (-) six3b (-) tbx16 (-) tbx24 (-) sox4b (-) foxa3 (-)

24 hrs gbx2 (+) mybbp1a (-) six3a (-) tbx24 (-) sox4b (-)

24 hrs gbx1 (-) foxh1 (-)

We measured the expression profile change in the morphants at the five time-points labeled in gray. More than three batches were used for the Q-PCR experiments, and the fold change compared to the specific MO and control MO was calculated. We selected more than twofold changes, and the statistical significance of the quantitative data was determined by Student’s t-test. ★★ indicated p < 0.01, ★ indicated 0.05 < p < 0.01, ☆ indicated p > 0.05. Genes that are repressed by transcription factors are colored green; and those activated by transcription factors are colored red.

important roles for endoderm formation downstream of the Nodal signaling pathway [33,36]. The expression of og9x at 8 hpf is increased 3-fold after knockdown of sox32 (Figure 4H), and the enhanced expression is located in the endodermal lineage (data not shown). The expression of bon is also increased after knockdown of sox32 at both 8 hpf and 11 hpf (Figure 4I), and the increased expression also happened in endodermal cells (data not

shown). The expression of sox32 itself was increased in the sox32 morphants at 11 hpf, 16 hpf and 24 hpf (Figure 4A). It was shown that bon and og9x activated sox32 expression in the presumptive endoderm, and their expressions were turned off at a later stage. Sox32 is a zygotic endoderm transcription factor with peak expression between 8 hpf and 11 hpf, and it has to be turned off at a later stage. From our data, sox32 negatively regulates bon, og9x and sox32


227


Figure 3. Example of sox32 and sox17 activated endoderm-specific transcription factors by Q-PCR and in-situ hybridization. For each gene, the Q-PCR result is shown in 1, in situ hybridization of the control morphant is shown in 2, and in situ hybridization of the same gene in the morphant is shown in 3. In the Q-PCR, when sox32 is knocked down, the gene is decreased, indicating it is positively regulated by sox32 in wild type condition. If the decreasing is more than twofold, it is colored red. When sox32 is knocked down, the gene is increased indicated it is negatively regulated by sox32 in wild type condition. If the increasing is more than twofold, it is colored green. However, if the changes are less than twofold or the gene’s expression level almost vanishes, it is colored blue. The result came from more than three different batches and the standard deviation is indicated as a line extended from the mean. The statistical significance of the quantitative data was determined by Student’s t-test: ★★ indicated p＜0.01, ★ indicated 0.05＜p＜0.01, ☆ indicated p＞0.05. The arrow indicates the in situ signals are repressed due to the knockdown of sox32. (A) sox17 is repressed at 8 hpf in sox32 morphants. (B) sox32 is repressed at 5 hpf in sox32 morphants. (C) otx2 is repressed at 8 hpf in sox32 morphants. (D) og9x is repressed at 8 hpf in sox17 morphants. (D) pax2a is repressed at 11 hpf in sox32 morphants. (H) pou1 is repressed at 11 hpf in sox17 morphants.

228


Figure 4. Complete set of downstream target genes of sox32 by Q-PCR. When the sox32 is knockdown, the gene is down-regulated indicated it is positively regulated by sox32 in wild type conditions. If the down-regulation is more than twofold, it is colored red. When sox32 is knocked down, the gene is up-regulated, indicating it is negatively regulated by sox32 in wild type condition. If the up-regulation is more than twofold, it is colored green. If the up-regulation is less than twofold or the gene’s expression level almost vanishes, it is colored blue. The result came from more than three different batches and the standard deviation is indicated as a line extending from the mean. The statistical significance of quantitative data was determined by Student’s t-test. ★★ indicated p＜0.01, ★ indicated 0.05＜p＜0.01, ☆ indicated p＞0.05. The arrow indicates the in situ signals are repressed due to the knockdown of sox32. Global Journal of Biochemistry | Volume 2 | Issue 4 | 2011 www.simplex-academic-publishers.com © 2011 Simplex Academic Publishers. All rights reserved.

229

GLOBAL JOURNAL OF BIOCHEMISTRY itself at a later stage. The purpose of the repression is to limit upstream factors and turn off the expression of an activator when it is not needed. Sox32 also negatively regulated some genes, including drl, gbx1, and gata6, which are expressed in other territories [52,53], drl and gata6 are transcription factors that are expressed at the mesoderm during development, and function in mesoderm specification [52]. sox32 represses both of them at beginning of the somite stage (Figure 4 G and L). gbx1, a neuroectoderm transcription factor [53], possibly responsible for brain specification at 11 hpf, is also repressed by sox32 (Figure 4F). sox32 expression is very low at 11 hpf, and it is expressed in the endoderm, while drl and gata6 are expressed in the mesoderm, and gbx1 is expressed in the ectoderm, and they are all expressed in nonoverlapping areas. Thus, this kind of repression is possible through an indirect signaling pathway. We detected increased tbx16 expression after sox32 knockdown by Q-PCR (Figure 4 K). According to the data reported above, endodermal transcription factor sox32 may be able to influence the specification of other tissues. These surprising results may account for the different morphology defects seen after sox32 MO injection. The purpose of the repression is to repress neighboring areas between mesoderm versus endoderm, and ectoderm versus endoderm, and to form a sharp boundary. 3.5. The effects of sox17 knockdown on gene expression involved in three germ layer specifications sox17 is an HMG box transcription factor that is essential for endoderm formation [24] and is a downstream target of sox32. An earlier study had identified several sox17 upstream genes, including pou5f1, acvr1b, sox32, ndr1, og9x, bon, fgf8, fgf17b and fgf24 [21,35,54,50]. However, the molecular events controlled by sox17 are largely unknown. In an attempt to decipher these events, we injected sox17 MO to knockdown sox17 expression in zebrafish embryos and established the subcircuits as described for sox32 knockdown. We analyzed the sox17 MO-injected embryos by Q-PCR at five different developmental stages as described earlier. The results showed that there were two genes activated and thirteen genes repressed by sox17 (Table 3). 3.6. The gene activated by sox17 at early stages In our Q-PCR experiments, og9x expression was down-regulated more than twofold in the embryos with knockdown of

sox17 (Figure 5A). The down-regulation of og9x in the sox17 morphants is verified in the endodermal lineage by in situ hybridization (data not shown). sox32 expression was downregulated in the endodermal lineage more than twofold in the embryos with knockdown of sox17 (Figure 5 B). Our data showed that og9x activates sox32 expression before 8 hpf. In addition, we showed that sox32 activates itself at 5 hpf, and it activates sox17 at 5, 8, and 11 hpf. Here, we showed that sox17 activates og9x at 8 hpf and activates sox32 at 11 hpf. The autopositive feedback loop forms a stable reenhancement among those three important transcription factors. 3.7. The gene repressed by sox17 at later stages Although it is known that sox17expressing cells differentiate into presumptive endodermal cells, the role for sox17 in endoderm specification and its downstream target genes are not clear. By using sox17 knockdown and examination of the expression profiles of some genes known to have a role in endoderm differentiation, we found several new genes downstream of sox17 that are negatively regulated by sox17. First, we found sox17 represses itself in the endoderm area from 11 hpf to 24 hpf (Figure. sox17 represses other endoderm 5F). transcription factor, such as foxa1 and foxa3 (Figure 5C and P). This kind of repression is important and is found to turn off expression of some transcription factors in many other developmental systems. In addition to those endoderm markers, sox17 represses several genes expressed in other areas based on Q-PCR and in situ hybridization studies. sox17 repressed six neuroectoderm-specific genes, six3a, six3b, etv5, etv6, trh3 and sox4b at 24 hpf (Figure 5I, J, K, H, M, and N). Six3a and Six3b are Six3 family transcription factors, displayed similar expression patterns in early forebrain development, and were involved in the growth of the brain and the head [55]. Endodermal specific transcription factor sox17 represses ectodermal expressed genes, six3a and six3b at 24 hpf (Figure 5I and J). Two Ets variants, Etv5 and Etv6, are Ets-related molecules expressed in the brain, eye and central nervous system at late somite stages [56]. sox17 represses etv5 and etv6 at 24 hpf, but not at early stages (Figure 5K and H). Another neuroectoderm transcription factor, trh3, expressed in the brain region was enhanced in sox17 morphants at 24 hpf (Figure 5M). The expression of transcription factor sox4b (SRY (sex-determining region Y)-box 4b) [56], for central nervous system, were enhanced 2.5-fold Global Journal of Biochemistry | Volume 2 | Issue 4 | 2011 www.simplex-academic-publishers.com

230



Figure 5. Complete set of downstream target genes of sox17 by Q-PCR. When the sox17 is knockdown, the gene is down-regulated indicated it is positively regulated by sox17 in wild type conditions. If the down-regulation is more than twofold, it is colored red. When sox17 is knocked down, the gene is up-regulated, indicating it is negatively regulated by sox17 in wild type condition. If the up-regulation is more than twofold, it is colored green. If the up-regulation is less than twofold or the gene’s expression level almost vanishes, it is colored blue. The result came from more than three different batches and the standard deviation is indicated as a line extending from the mean. The statistical significance of quantitative data was determined by Student’s t-test. ★★ indicated p＜0.01, ★ indicated 0.05＜p＜0.01, ☆ indicated p＞0.05. Global Journal of Biochemistry | Volume 2 | Issue 4 | 2011 www.simplex-academic-publishers.com © 2011 Simplex Academic Publishers. All rights reserved.

231

GLOBAL JOURNAL OF BIOCHEMISTRY in sox17 morphants (Figure 5N). It has been suggested that endoderm transcription factor sox17 might contribute only to endoderm formation. However, according to our new discovery, sox17 participates in regulation of ectoderm formation and specification. Furthermore, sox17 represses the mesoderm-specific genes drl, foxh1, gata4, gata5, and gata6 at 24 hpf. drl is a critical transcription factor expressed at an intermediate cell mass of mesoderm and affects formation of the hematopoietic system [56a]. The expression of drl is increased in the intermediate cell mass of sox17 morphants (data not shown). foxh1 (Forkhead box H1), expressed on diencephalons and ventral mesoderm at late somite stages, can negatively regulate vascular formation in zebrafish [32]. After injection of sox17 MO, the expression of foxh1 becomes enhanced in the same area as the control morphants (Figure 5D). Mesoderm-specific transcription factors gata4/gata5/gata6 (gata-4/5/6) are all expressed in overlapping patterns between the endoderm and mesoderm at an early stage [57]. At 24 hpf, gata-4/5/6 expression is increased about 2.5-fold in sox17 morphants (Figure 5O, E and L). The cross-boundary inhibition is important to restrict the expression domain and ensure the boundary between neighboring domains. 3.8. The upstream inputs for sox32 and sox17 The sox32 and Sox17 MO knockdown experiments have identified many downstream target genes of sox32 and sox17. The upstream inputs of two endodermal transcription factors were also identified using a similar approach. Table 5 summarizes all of the upstream inputs from the Q-PCR data. gata5 activates sox32 at 5 hpf, gbx1 activates sox32 at 8 hpf, and sox17 activates sox32 at 11 hpf. Other than those positive inputs, sox32 and gbx2 repress sox32 at 11 hpf, gbx2, otx2, and sox32 repressed sox32 at 16 hpf. For the upstream input to sox17, we found sox32, gata5 and otx2 can activate sox17 expression from 5 hpf (otx2) to 8 hpf (gata5 and sox32) until 11 hpf (gata5 and sox32). On the other hand, other genes repress sox17 at 11 hpf, those genes include sox17 itself, gbx2 and otx2. 3.9. The effects of otx2, gata6, gata5, foxa2, gbx2, gbx1 and six3b knockdown on gene expression involved in three germ layer specifications The downstream target genes of other six endodermal transcription factors were also identified. All of knockdown experiments were carried out using a similar approach. Table 3 listed the downstream targets and Table 4 listed the upstream inputs. Here, we summarized the

important discovery from each knockdown experiment. 3.10. foxa2 repressed itself and activates genes involved in endoderm formation and CNS development From Q-PCR results, foxa2 began to express at 5 hpf, increasing from 8 hpf and peak at 11 hpf, then decreased at 16 hpf and maintained to express until 36 hpf (Figure. 1Group IV-D). After foxa2 knockdown, foxa2 RNA expression was up-regulated 2- to 3- fold from 11 hpf to 36 hpf (Figure 6-H). The phenomenon was validated by in situ hybridization. At 11 hpf, foxa2 expression was enhanced not only in endodermal cells but also in axis (Figure 7A vs. B). At 16 hpf, the midline expression was stronger and extended to notochord (Figure 7C vs. D). It indicates that foxa2 can regulate its mRNA expression through negative feedback. In addition, in foxa2 morphants, og9x expression was enhanced at 8 hpf (Figure 6-A and Figure 7E vs. F) and sox4b was down-regulated at 11 hpf (Figure 6-C and Figure 7G vs. H). In addition to genes involved in foxa2 knockdown endoderm formation, repressed several genes which are expressed in central nervous system (CNS). In foxa2 morphants, gbx1 was repressed at 24 and 36 hpf (Figure 6-K), brn1.2 and pou1 was downregulated from 11 to 36 hpf (Figure 6-D and 6E) and pou12 was repressed from 16 to 36 hpf (Figure 6-J). The localization and expression of each gene was examined at 24 hpf. gbx1, which are expressed in hindbrain and spinal cord [53], was inhibited in foxa2 morphants (Figure 7I vs. J). brn1.2, pou1 and pou12, which belong to class III POU gene subfamily [58], expressed weaker in brain and spinal cord in foxa2 morphants (Figure 7K vs. L, Figure 7M vs. N, Figure 7O vs. P). 3.11. foxa2 expression was affected in gata6, otx2 and sox32 perturbation experiments At 5 hpf, foxa2 was first detected on the dorsal side of hypoblast and the expression was decreased in otx2 and gata6 morphants at 5 hpf (Figutr 8C and 9E). foxa2 expression was downregulated in sox32 morphants at 8 hpf. Compared with control (Figure 7-Q) and sox32 morphants (Figure 7-R), foxa2 expression in endoderm disappeared in sox32 morphants. In addition, the width of the axis was broadened in sox32 morphants. These indicated that foxa2 was positively regulated by gata6, otx2 and sox32.


232


GLOBAL JOURNAL OF BIOCHEMISTRY Table 4. Upstream inputs of foxa2, gata6, gata5, otx2, six3a, six3b, sox32, sox17, gbx2, and gbx1.

5hrs gata6 (+) otx2 (+) 5hrs otx2 (+) gata5 (+)

5hrs gata6 (+) otx2 (+)

5hrs

5hrs gbx1 (+)

foxa2 gene expression affected by others-all 8 hrs 11 hrs 16 hrs sox32 (+) foxa2 (-) foxa2 (-) gata6 gene expression affected by others-all 8 hrs 11 hrs 16 hrs sox32 (-) otx2 (-) gata6 (-) gata6 (-) six3b (-) gata5 gene expression affected by others-all 8 hrs 11 hrs 16 hrs gata5 (-) gata5 (-) gata5 (-) gata6 (-) otx2 gene expression affected by others-all 8 hrs 11 hrs 16 hrs six3b (+) six3b (+) otx2 (-) sox32 (+) otx2 (-) gbx2 (-) otx2 (-) six3a gene expression affected by others 8 hrs 11 hrs 16 hrs otx2 (-)

six3b gene expression affected by others 8 hrs 11 hrs 16 hrs gbx2 (+) otx2 (-) six3b (-) sox32 gene expression affected by others-all 5hrs 8 hrs 11 hrs 16 hrs gata5 (+) gbx1 (+) sox17 (+) otx2 (-) sox32 (+) gbx2 (-) gbx2 (-) gata6 (+) sox32 (-) sox32 (-) gata6 (-) gata6 (-) sox17 gene expression affected by others-all 5hrs 8 hrs 11 hrs 16 hrs otx2 (+) gata5 (+) gata5 (+) otx2 (-) sox32 (+) sox32 (+) sox32 (+) sox17 (-) gata6 (+) gbx2 (-) gata6 (-) sox17 (-) gbx2 gene expression affected by others-all 5hrs 8 hrs 11 hrs 16 hrs gbx2 (+) gbx2 (+) six3b (+) otx2 (-) gbx1 gene expression affected by others-all 5hrs 8 hrs 11 hrs 16 hrs gata6 (+) six3b (+) six3b (+) gbx1 (-) gbx1 (-) gata6 (+) sox32 (-) gbx1 (-) 5hrs

24 hrs foxa2 (-)

24 hrs sox17 (-) gata6 (-)

24 hrs gata5 (-) gata6 (-) sox17 (-) 24 hrs otx2 (-)

24 hrs sox17 (-) gbx2 (-) 24 hrs sox17 (-) six3b (-) 24 hrs sox32 (-)

24 hrs sox17 (-) gata6 (-)

24 hrs gbx2 (+)

24 hrs otx2 (+) foxa2 (+) gata6 (+) gbx1 (-)


233


Figure 6. Complete set of downstream target genes of foxa2 by Q-PCR. When the foxa2 is knockdown, the gene is down-regulated indicated it is positively regulated by foxa2 in wild type conditions. If the down-regulation is more than twofold, it is colored red. When foxa2 is knocked down, the gene is up-regulated, indicating it is negatively regulated by foxa2 in wild type condition. If the up-regulation is more than twofold, it is colored green. If the up-regulation is less than twofold or the gene’s expression level almost vanishes, it is colored blue. The result came from more than three different batches and the standard deviation is indicated as a line extending from the mean. The statistical significance of quantitative data was determined by Student’s t-test. ★★ indicated p＜0.01, ★ indicated 0.05＜p＜0.01, ☆ indicated p＞0.05. 234


Figure 7. The expression of foxa2 downstream genes and foxa2 upstream input in control and morphants (A – B) 11 hpf stage, dorsal view, head to the top. The expression of foxa2 is activated in foxa2 morphants. (C-D) 16 hpf stage, later view with dorsal to the right. The foxa2 positively regulates itself. (E-F) 8 hpf stage, lateral view, animal to the top. The expression of og9x is positively regulated by foxa2. (G-H) 11 hpf, dorsal view with head to the top. sox4b is down-regulated in foxa2 morphants. (I-P) 24 hpf, lateral view with anterior to the left and dorsal to the top. The expression of gbx1, brn1.2, pou1 and pou12 are down-regulated in foxa2 morphants. (Q-R) 8 hpf, the expression of foxa2 in endoderm is decreased in sox32 morphant comparing to control morphants.


235

GLOBAL JOURNAL OF BIOCHEMISTRY 3.12. Evolutional conserved subcircuits and its divergences We have done research on the interactions among otx2, gata5 and gata6, and found those evolutional conserved transcription factors form a subcircuits and determine the initial specification of the zebrafish endoderm (submitted and under review). We found otx2 can activate both gata5 and gata6 however a feedback loop from Gata to otx2 is missing. In this study, we observed the down regulation of gata5, sox17, foxa2, foxa3 and gata6 in the otx2 morphants at 5 hpf (Figure 8 A~E). This indicates the importance of otx2 on those endoderm specific transcription factors at the earliest time. Otx2 is expressed in an overlapping time frame and spatial domain with gata5, foxa2, gata6, foxa3 and sox17 at 5 hour’s stage, and knockdown of Otx2 function depresses the activity of those genes, the interaction of Otx2 on those genes is very possible to be direct. We have used chromatinimmunoprecipitation by Otx2 antibody to prove the direct binding of Otx2 on the specific regulatory modules of gata5 and gata6, and the binding indeed contribute to the positive regulatory function (submitted). In addition, gata6, foxa3 and sox17 were negatively regulated by otx2 (Figure 8E, D and B) later in development at 16 hpf. 3.13. The transcription factors positively regulated by otx2 on later neuroectoderm During the somite stage, Otx2 express in presumptive brain, forebrain and midbrain. In the otx2 morphants expression of brn1.2, pou1, pou12 and gbx1 were decreased and it indicated otx2 positively regulated these four genes, and knockdown of otx2 function depressed their expression level (Figure 8 H, I, J and X). Those genes expressed in the neuroectoderm at this stage. brn1.2, pou1and pou12 are ClassⅢ POU genes that may contribute to early regionalization of the zebrafish embryonic brain [58,59]. They are expressed in the presumptive brain, forebrain and midbrain, temporal-spatial overlap with otx2, the interaction of otx2 on those genes is very possible to be direct. Nevertheless, gbx1 expressed in the hindbrain, temporal-spatial not overlap with otx2 and the interaction of otx2 is probably indirect. 3.14. The transcription factors negatively regulated by otx2 on neuroectoderm, endoderm or mesoderm otx2 also negatively regulated many important transcription factors that expression on several territory and had different development functions. For example, otx2 represses some

early expressed genes: otx2, foxh1, tbx16, sox32 and bon at later stage (Figure 8G, M, Q, V, and F). From the expression profile of otx2, foxh1, tbx16, sox32 and bon, we noticed that those genes were expressed earlier in development and then went down. The down regulation of those early genes is very important in the normal development. The purpose of such feedback loop is to have a temporal peak of expression, when it is not needed, the expression should be turned off. foxh1 is a maternal transcription factor which would be turn off at later stage when it is not needed. bon is a maternal transcription factor that expresses in the presumptive endoderm that would be negative regulated by otx2 at 8 to 16 hours post fertilization (Figure 8F). It also expresses in axial mesoderm, which can regulate neural patterning. It has been reported that tissue specific knockdown bon in the axial mesoderm would reduce otx2 expression [60]. Therefore, they found bon expressed in the axial mesoderm can positively regulate otx2. In our study, we found otx2 can negatively regulate bon. The interaction between otx2 and bon is like a negative feedback loop, first Bon activates otx2 and otx2 repressed bon at later. otx2, foxh1, gbx2, pax2a, six3a, six3b, mybbp1a, and sox4 are transcription factors that would express in the brain and have functions in brain specification or maintenance. Otx2 down regulated them at later somite stages (16, 24 hpf) (Figure 8), at this stage Otx2 expressed in the diencephalons and midbrain. The function of Otx2 prevented transcription factors expressed in nearly brain regions. six3a and six3b are transcription factors that expressed in forebrain, gbx2 is transcription factors that expressed in hindbrain, pax2a is transcription factors that expressed in midbrain hindbrain boundary, sox4 is transcription factors that expressed telencephalon, diencephalons, and midbrain hindbrain boundary. When otx2 expresses normally it would keep these transcription factors form midbrain, knockdown otx2 would increase these transcription factors expression. mybbp1a was expressed co localized with otx2 in the midbrain and is down regulated by otx2 at 24 hours post-fertilization. tbx16, sox32, gata4, prdm1, tbx24 and trh3 were transcription factors that express in the tissue of endoderm or mesoderm. otx2 negative regulated these transcription factors at later stage (Figure 8), the interaction of otx2 on these transcription factors is possible to be indirect. Other genes that regulated by otx2 may regulate its expression in the early hypoblast stage.


236



Figure 8. Complete set of downstream target genes of otx2 by Q-PCR. When the otx2 is knockdown, the gene is down-regulated indicated it is positively regulated by otx2 in wild type conditions. If the down-regulation is more than twofold, it is colored red. When otx2 is knocked down, the gene is up-regulated, indicating it is negatively regulated by otx2 in wild type condition. If the up-regulation is more than twofold, it is colored green. If the up-regulation is less than twofold or the gene’s expression level almost vanishes, it is colored blue. The result came from more than three different batches and the standard deviation is indicated as a line extending from the mean. The statistical significance of quantitative data was determined by Student’s t-test. ★★ indicated p＜0.01, ★ indicated 0.05＜p＜0.01, ☆ indicated p＞0.05. Global Journal of Biochemistry | Volume 2 | Issue 4 | 2011 www.simplex-academic-publishers.com © 2011 Simplex Academic Publishers. All rights reserved.

237

GLOBAL JOURNAL OF BIOCHEMISTRY 3.15. gata5 and gata6 positively regulate the mesendoderm transcription factors The positive feedback loop between gata5 and gata6 locks on to the specification of endoderm to ensure the expression of endodermspecific transcription factors, and the direct binding site were identified in the functionally important regulatory module (submitted). Figure 9 shows the positive regulatory role of gata6 on tbx16, foxa3, gata5, sox32, foxa2 and og9x at 5 hpf, on gbx1, cdx4, prdm1a, sox17 at 8 hpf. Figure 10 shows the positive regulatory function of gata5 on irx3a, gata6 and sox32 at 5hpf, and on gbx1, brn1.2, gata4, sox17 and otx2 at 8hpf. Similar to other early endoderm transcription factors mentioned earlier, gata5 and gata6 will repress the expression of endoderm transcription factors later in development. 3.16. gbx1 and gbx2 involved in the gene regulatory networks Gbx1 has some regulatory function on three germ layers at initial stage. The endoderm genes, og9x, foxa and sox32, were activated in 8 hpf (Figure. 11C, D and H). From our Q-PCR result, we found most of the genes regulated by gbx1 were different from gbx2, and expect fgf8 which is related to MHB. However gbx1 had effect on the genes specifically located on spinal cord or ectoderm at very early stage. It seemed that gbx1 had different task on the posterior part of hindbrain development. Unlike gbx1, gbx2 only positive regulates itself, six3b, pou12, fgf8, eng2a and pou1 (Figure 12B, C, D, E, and I). Later in development, gbx1 represses the expression of foxa1, gbx1 and foxh1 (Figure 11I, J and K), and gbx2 represses many genes’ expression including sox32, sox17, foxh1, otx2, mybbp1a, six3a, tbx24 and sox4 (Figure 12G~O). When we looked at the upstream input of gbx1 and gbx2, we found that gata6, gata5, sox32, six3b, otx2, gbx1 and foxa2 could positively regulate gbx1, and gata5, six3b, otx2 and gbx2 could regulate gbx2 (Table 4). 4. Discussion 4.1. The role of Nodal, sox32, and sox17 in endoderm formation The molecular genetics of axis formation in zebrafish has been reviewed extensively [20], revealing that in five major signaling pathways: BMP, Nodal, Fgf, canonical Wnt and retinoic acid signals generate positional information that leads to the restricted expression of transcription factors [20]. The endoderm formation is specified by the combinatorial effect among Nodal, Fgf, and Bmp signaling pathways in zebrafish. Nodal and Bmp

belong to the transforming growth factor- family. Two members of the Nodal family, Ndr1 and Ndr2, are essential inducers of both mesoderm and endoderm [28]. Bmps negatively regulate the formation of the endoderm precursor on the ventral side, whereas Fgf signaling restricts the endoderm precursor cells on the dorsal side [50]. The Nodal signaling activates Sox32 and the downstream endoderm transcription factors in zebrafish development. Nodal-related transcription Many factors including Sox32, Sox17, Bon, Gata5, Pou5f1 and Og9x are needed for endoderm formation. The thyroid growth and differentiation at much later stages require three transcription factors (Hhex, Nk2.1a, and Pax2.1) downstream of the Nodal pathway [61]. The induction of the zebrafish ventral brain and floor plate depends upon Ndr1/Ndr2 signaling, probably through the Foxa2 transcription factor [62]. Our subcircuits of sox32 and sox17 provide an insight into how the Nodal signal transmitted from the initial stage activates many transcription factors. Og9x, Bon, eomesodermin, and Gata5 have been shown to activate the expression of sox32. pou5f1 and sox32 together are necessary and sufficient to activate endoderm development and stimulate sox17 expression [63]. Gata5 activates the expression of sox17 and foxa2 because mutation of gata5 appears to express lower levels of sox17 and foxa2 than the wild type. In this study, we knocked down the expression of sox32 and sox17 and searched for the interrelationships of sox32 and sox17 to all those endodermal-specific transcription factors. We also examined the effects of sox32 and sox17 knockdown on the expression of other lineagespecific transcription factors. 4.2. The gene regulatory networks based on Sox32 and Sox17 perturbation We have combined all of the upstream and downstream information from the knockdown of sox32 and sox17, as well as other related transcription factors. The sox32 and sox17 subcircuits were established and are shown in Figure 13 and Figure 14. Five different time-points in the first 24 hours of zebrafish development that represent late blastula (5 hpf), middle gastrula (8 hpf) and segmentation (11 hpf, 16 hpf and 24 hpf) were chosen to build the subcircuits. Previously, we built the zebrafish GRNs with the spatial and temporal expression patterns by using the interaction evidence in the literature. We have updated our GRNs according to our experimental data. At 5 hpf, we found that gata5 positively regulates the expression of sox32, and sox32 Global Journal of Biochemistry | Volume 2 | Issue 4 | 2011 www.simplex-academic-publishers.com

238



Figure 9. Complete set of downstream target genes of gata6 by Q-PCR. When the gata6 is knockdown, the gene is down-regulated indicated it is positively regulated by gata6 in wild type conditions. If the down-regulation is more than twofold, it is colored red. When gata6 is knocked down, the gene is up-regulated, indicating it is negatively regulated by gata6 in wild type condition. If the up-regulation is more than twofold, it is colored green. If the up-regulation is less than twofold or the gene’s expression level almost vanishes, it is colored blue. The result came from more than three different batches and the standard deviation is indicated as a line extending from the mean. The statistical significance of quantitative data was determined by Student’s t-test. ★★ indicated p＜0.01, ★ indicated 0.05＜p＜0.01, ☆ indicated p＞0.05. Global Journal of Biochemistry | Volume 2 | Issue 4 | 2011 www.simplex-academic-publishers.com © 2011 Simplex Academic Publishers. All rights reserved.

239


Figure 10. Complete set of downstream target genes of gata5 by Q-PCR. When the gata5 is knockdown, the gene is down-regulated indicated it is positively regulated by gata5 in wild type conditions. If the down-regulation is more than twofold, it is colored red. When gata5 is knocked down, the gene is up-regulated, indicating it is negatively regulated by gata5 in wild type condition. If the up-regulation is more than twofold, it is colored green. If the up-regulation is less than twofold or the gene’s expression level almost vanishes, it is colored blue. The result came from more than three different batches and the standard deviation is indicated as a line extending from the mean. The statistical significance of quantitative data was determined by Student’s t-test. ★★ indicated p＜0.01, ★ indicated 0.05＜p＜0.01, ☆ indicated p＞0.05.

240


Figure 11. Complete set of downstream target genes of gbx1 by Q-PCR. When the gbx1 is knockdown, the gene is down-regulated indicated it is positively regulated by gbx1 in wild type conditions. If the down-regulation is more than twofold, it is colored red. When gbx1 is knocked down, the gene is up-regulated, indicating it is negatively regulated by gbx1 in wild type condition. If the up-regulation is more than twofold, it is colored green. If the up-regulation is less than twofold or the gene’s expression level almost vanishes, it is colored blue. The result came from more than three different batches and the standard deviation is indicated as a line extending from the mean. The statistical significance of quantitative data was determined by Student’s t-test. ★★ indicated p＜0.01, ★ indicated 0.05＜p＜0.01, ☆ indicated p＞0.05.


241


Figure 12. Complete set of downstream target genes of gbx2 by Q-PCR. When the gbx2 is knockdown, the gene is down-regulated indicated it is positively regulated by gbx2 in wild type conditions. If the down-regulation is more than twofold, it is colored red. When gbx2 is knocked down, the gene is up-regulated, indicating it is negatively regulated by gbx2 in wild type condition. If the up-regulation is more than twofold, it is colored green. If the up-regulation is less than twofold or the gene’s expression level almost vanishes, it is colored blue. The result came from more than three different batches and the standard deviation is indicated as a line extending from the mean. The statistical significance of quantitative data was determined by Student’s t-test. ★★ indicated p＜0.01, ★ indicated 0.05＜p＜0.01, ☆ indicated p＞0.05.

242


Figure 13. Complete set of downstream target genes of six3b by Q-PCR. When the six3b is knockdown, the gene is down-regulated indicated it is positively regulated by six3b in wild type conditions. If the down-regulation is more than twofold, it is colored red. When six3b is knocked down, the gene is up-regulated, indicating it is negatively regulated by six3b in wild type condition. If the up-regulation is more than twofold, it is colored green. If the up-regulation is less than twofold or the gene’s expression level almost vanishes, it is colored blue. The result came from more than three different batches and the standard deviation is indicated as a line extending from the mean. The statistical significance of quantitative data was determined by Student’s t-test. ★★ indicated p＜0.01, ★ indicated 0.05＜p＜0.01, ☆ indicated p＞0.05.


243


Figure 14. The gene regulatory networks control the developmental process at five different stages. (A) Overview of the GRNs at 5hpf. Otx2 activated gata5, gata6 sox17, foa2, and foxa3. This stimulation was soon amplified by the autoregulatory lock-on system in gata5 and gata6 to each other and sox32 to itself. This stable output then further activated other downstream genes including irx3a, sox32, sox17, foxa2, foxa3, and tbx16. (B) Overview of the GRNs at 8hpf. As development progressed, more downstream genes were up-regulated such as gata4 by gata5, prdm1a by gata6, shha, twhh, fgf8, foxa by gbx1. Some gene no longer needed were down-regulated such as bon by otx2, sox32, and gbx2, otx2 and gata5 by themselves, og9x and drl by sox32 and gata5, and foxa1 by sox17. (C) Overview of the GRNs at 11hpf. More genes were downregulated, such as gata5, gata6, and otx2 shout down by themselves, bon, drl, pou5f1, foxh1, and sox32, by gata6. More ectoderm expression genes were up-regulated; gbx2 and six3b form a autoregulatory lock-on system and drive their downstream genes such as otx2, eng2a, fgf8, pou12, pou1, brn1.2, pax2a and otx1. (D) Overview of the GRNs at 16hpf. gbx2 keep on lock-on itself , but six3b started to down-regulate itself. Except for those ectoderm expression genes driven by gbx2, otx2, six3b, most of the genes in the mesendoderm GRNs were shut down. (E) Overview of the GRNs at 24hpf. Although most of the early genes were shut down, foxa2 starts to take over part of the activating roles of six3b, gbx2 and otx2 on gbx1, pou1, pou12; it also activate other neural specific genes such as shhb, nkx2.2b, and ntn1b. Global Journal of Biochemistry | Volume 2 | Issue 4 | 2011 www.simplex-academic-publishers.com

244


GLOBAL JOURNAL OF BIOCHEMISTRY activates sox17 in accordance with the literature. In addition, we discovered otx2 activates sox17 expression, and sox32 activates sox32 itself at this early stage. The role of otx2 that determines the endoderm specification has been demonstrated in sea urchin [64]. In vertebrates, otx2 is expressed in the anterior endoderm. A double mutant lacking both otx2 and cripto (a critical component in the Nodal signaling pathway) fails to form endoderm, and instead generates an excess of mesoderm. Our results indicate that in zebrafish otx2 is important in the early endoderm formation by positively regulating sox17 and the other four endodermspecific transcription factors. sox32 is required for the expression of gata5 in endoderm formation [38]. In our subcircuits, we found gata5 activates sox32 expression. The positive feedback loop between gata5 and sox32 is one of the examples of a lock-on system that is important to ensure the endodermal cell fate. 4.3. Role of foxa2 in endoderm formation The importance of foxa2 in endoderm formation is well-known that is under regulation by sox32 in Nodal signaling pathway. Nodal signal induces og9x [33], bon and gata5, and initiates the expression of sox32 which then activates sox17 and foxa2 expression [65]. From our sox32 knockdown experiment, foxa2 was down-regulated in sox32 morphants. In situ pattern showed that the endodermal expression of foxa2 disappeared in the morphants. This is the same as that showed previously [35]. Besides this change in endoderm, we also found that the axial expression of foxa2 was broadened more than control morphants. og9x, a maternal pairedlike homeobox protein, expressed transiently in mesoderm precursor before shield stage, was turned off after 70% epiboly [33]. However, og9x expression was activated 3 fold in foxa2 morphants at 8 hpf. This implied that foxa2 has a negative feedback to the upstream regulator og9x to shut down the expression of og9x at later stage. sox4b is suggested as a key regulator of pancreatic alpha cell differentiation in zebrafish [66]. In foxa2 morphants sox4b was repressed at 11 hpf. This suggests that foxa2 plays an important role in pancreas formation partly by regulating sox4b. In addition to sox32, gata6 and otx2 also activate foxa2. gata6 is expressed at hypoblast at 5 hpf (data not shown). At 12 hpf it is expressed in lateral plate mesoderm [67]. Later it is expressed in cardiac progenitor at 18 hpf and heart tube and gut at 24 hpf. gata6 is important for cardiovascular and gut and liver development, but little study is found in early embryonic stage. One of the most related gata

family, gata5, is well known to be a upstream regulator of foxa2 for endoderm formation in nodal signaling pathway [39,38]. Because the expression pattern and sequence homology is very similar, we suggested that gata6 plays a role as gata5. That places gata6 upstream of foxa2 and it can induce foxa2 transcription at early stage. 4.4. Role of gbx1 and gbx2 in gene regulatory network of zebrafish embryonic development From our GRNs, gbx family not only interacts with endoderm in early stage but also play different roles in neuronal development. gbx1 regulates several genes in posterior hindbrain and spinal cord, and gbx2 influences the other transcriptional factors in the diencephalons and anterior hindbrain. Interestingly, we also found some feedback between transcription factors. First, gbx2 and otx2 repress each other to build the sharp boundary in MHB region. Besides, gbx2 positively regulates fgf8 and from literature, fgf8 positively regulates gbx2, such positive feedback loop promoting gbx2 and fgf8 expressed in the anterior hindbrain (rhombomere1). fgf8 and wnt1 thus establish the isthmus signal center to regulate adjacent neuron tissue development. Second, gbx2 indirectly regulates gbx1 through other genes. otx2 activates gbx1, and otx2 is repressed by gbx2. Such interrelationship probably causes that gbx1 cannot express in the gbx2 region which are located on the opposite side of otx2 and always keep some space from otx2 [53]. 4.5. Interesting functional motifs reveals the robustness of the endoderm GRNs in zebrafish In these subcircuits, we identified several interesting functional motifs that are important building blocks in zebrafish developmental GRNs. First, sox32 and sox17 serve as an autoregulatory lock on the endoderm fate. sox32 activates sox17 and sox32 itself at 5 hpf, sox17 activates sox32 at 11 hpf. This positive enhancement ensures the endoderm areas accumulate enough sox32 and sox17 in a short period of time. Secondly, the function of an early activation switch to late repression can be found in several genes. sox32 activates sox32 itself at 5 hpf, and then represses itself from 11 hpf until 24 hpf. otx2 activates sox17 at 5 hpf and then represses sox17 at 16 hpf. Thirdly, late transcription factors that repress early activator can be identified. gata5 activates sox17 at 5 hpf, and then sox17 represses gata5 at 24 hpf. gbx1 activates sox32 at 8 hpf, and then sox32 represses gbx1 at 11 hpf. Fourthly, we found many early transcription factors can be repressed


245

GLOBAL JOURNAL OF BIOCHEMISTRY by both sox32 and sox17 at later stages. sox32 and sox17 repress drl1, pou5f1, foxh1, and bon. Furthermore, sox17 represses itself at later stages. The repressive systems have been identified in many other developmental GRNs as an important functional motif that controls the temporal expression peak and restricts the spatial expression pattern of specific transcription factors [68,69]. The architecture of the GRNs for endoderm formation is robust due to those important blocks. First, sox32 and sox17 positively regulate each other at the autoregulatory lock. gata5 and sox32 also positively regulate each other. After the Nodal signal is transduced from the receptors, and the activated sox32 and gata5 can establish a domain expressing the endodermal transcription factors in a short period of time through this lock-down system. To ensure the endoderm can further specify, those early transcription factors are turned off at a later stage. We found many repression effects in our GRNs. From an early activator to a late repressor, one gene can have different roles for the same target at different stages of development. As late transcription factors repress early activators, a gene can turn off the expression of its own activator. Furthermore, sox32 and sox17 appear to shut down a series of early transcription factors. Precisely coordinated gene regulation can be achieved with these comprehensive positive auto-regulatory networks and the negative feedback loops. 4.6. The establishment of more complete subcircuits of sox32 and sox17 A genome-wide search for sox32 downstream target genes and other endodermspecific transcription factors are important to complete the gene networks. We have performed microarray analysis for sox32 and sox17 morphants in order to identify the whole spectrum of the genes under the control of sox32 and sox17. Our results have revealed that expression was either reduced or increased for many genes upon the sox32 knockdown by injection of an anti-sense oligonucelotide into embryos. To investigate the genome-wide Sox32 binding site, we are in the process of using chromatin immunoprecipitation coupled with microarray analysis (ChIP-on-chip). The discovery of Sox32 binding sites may help identify the target genes and reveal possible functions during animal development. This analysis will give us an insight into the in vivo target sites for the important transcription factors, and thus provide another important basis

for the mechanism of Sox32-regulated endoderm formation. Direct interaction is vital for validation of the GRNs structure. In development, the cisregulatory modules function as information processors. The autonomous regulatory functions generated by each cis-regulatory module are the direct consequence of its design. To test our GRNs model, we performed a comparative genomic search for sox17. The candidate modules were linked to a GFP reporter gene and the functional analysis was performed, and we report that finding elsewhere. Combined with the ChIP-chip experimental data, we will generate a model of how the cis-regulatory elements logically regulate the differential gene expression in three germ layers of zebrafish development. Acknowledgements We would like to thank the ZeTH (Zebrafish Core Facility of NTHU and NHRI) funded by NSC (National Science Council, Taiwan, NSC 99-2321-B-400-001) for providing fish/material for this study. We are grateful for help from the confocal microscopy facility in the core facilities of the National Health Research Institutes, especially to Ms. Su-Feng Hu. Funding support from National Health Research Institute grants MG-094-PP-14, MG-095-PP-08, MG-096-PP-05, and MG-097-PP-07 and NSC 97-3112-B-400-008 from National Science Council to Dr Yuh Chiou-Hwa is gratefully acknowledged. References 1.

2.

3.

4.

5.

Levine, M. and Davidson, E. H., Gene regulatory networks for development, Proc. Natl. Acad. Sci. USA 102 (2005) 49364942. Oliveri, P. and Davidson, E. H., Development. Built to run, not fail, Science 315 (2007) 1510-1511. Oliveri, P. and Davidson, E. H., Gene regulatory network controlling embryonic specification in the sea urchin, Curr. Opin. Genet. Dev. 14 (2004) 351-360. Yuh, C. H., Bolouri, H. and Davidson, E. H., Cis-regulatory logic in the endo16 gene: switching from a specification to a differentiation mode of control, Development 128 (2001) 617-629. Yuh, C. H. and Davidson, E. H., Modular cis-regulatory organization of Endo16, a gut-specific gene of the sea urchin embryo, Development 122 (1996) 1069-1082.


246


GLOBAL JOURNAL OF BIOCHEMISTRY 6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16. 17.

18.

19.

Davidson, E. H., Rast, J. P., Oliveri, P., Ransick, A., Calestani, C., Yuh, C. H., Minokawa, T., Amore, G., Hinman, V., Arenas-Mena, C. et al., A genomic regulatory network for development, Science 295 (2002) 1669-1678. Ben-Tabou de-Leon, S. and Davidson, E. H., Gene regulation: gene control network in development, Annu. Rev. Biophys. Biomol. Struct. 36 (2007) 191. Schoenebeck, J. J. and Yelon, D., Illuminating cardiac development: Advances in imaging add new dimensions to the utility of zebrafish genetics, Semin. Cell Dev. Biol. 18 (2007) 27-35. Zacchigna, S., Ruiz de Almodovar, C. and Carmeliet, P., Similarities between angiogenesis and neural development: what small animal models can tell us, Curr. Top. Dev. Biol. 80 (2008) 1-55. Zorn, A. M. and Wells, J. M., Molecular basis of vertebrate endoderm development, Int. Rev. Cytol. 259 (2007) 49-111. Cha, Y. I., Solnica-Krezel, L. and DuBois, R. N., Fishing for prostanoids: deciphering the developmental functions of cyclooxygenase-derived prostaglandins, Dev. Biol. 289 (2006) 263-272. Cohen, M. M., Jr., Molecular dimensions of gastrointestinal tumors: some thoughts for digestion, Am. J. Med. Genet. A 122 (2003) 303-314. Stern, H. M. and Zon, L. I., Cancer genetics and drug discovery in the zebrafish, Nat. Rev. Cancer 3 (2003) 533-539. Mari-Beffa, M., Santamaria, J. A., Murciano, C., Santos-Ruiz, L., Andrades, J. A., Guerado, E. and Becerra, J., Zebrafish fins as a model system for skeletal human studies, Sci. World J. 7 (2007) 1114-1127. Perlin, J. R. and Talbot, W. S., Signals on the move: chemokine receptors and organogenesis in zebrafish, Sci. STKE, (2007) 45. Lamont, R. E. and Childs, S., MAPping out arteries and veins, Sci. STKE (2006) 39. Kari, G., Rodeck, U. and Dicker, A. P., Zebrafish: an emerging model system for human disease and drug discovery, Clin. Pharmacol. Ther. 82 (2007) 70-80. Lieschke, G. J. and Currie, P. D., Animal models of human disease: zebrafish swim into view, Nat. Rev. Genet. 8 (2007) 353367. Pyati, U. J., Look, A. T. and Hammerschmidt, M., Zebrafish as a powerful vertebrate model system for in vivo studies of cell death, Semin. Cancer Biol. 17 (2007) 154-165.

20. Schier, A. F. and Talbot, W. S., Molecular genetics of axis formation in zebrafish, Annu. Rev. Genet 39 (2005) 561-613. 21. Alexander, J., and Stainier, D. Y., A molecular pathway leading to endoderm formation in zebrafish, Curr. Biol. 9 (1999) 1147-1157. 22. Ochi, H. and Westerfield, M., Signaling networks that regulate muscle development: lessons from zebrafish, Dev. Growth Differ. 49 (2007) 1-11. 23. Le Good, J. A., Joubin, K., Giraldez, A. J., Ben-Haim, N., Beck, S., Chen, Y., Schier, A. F. and Constam, D. B., Nodal stability determines signaling range, Curr. Biol. 15 (2005) 31-36. 24. Schier, A. F. and Talbot, W. S., Nodal signaling and the zebrafish organizer, Int. J. Dev. Biol. 45 (2001) 289-297. 25. Zhang, J., A new trick to tune down TGFbeta signal, Cell. Res. 14 (2004) 439-440. 26. Zhou, X., Sasaki, H., Lowe, L., Hogan, B. L. and Kuehn, M. R., Nodal is a novel TGFbeta-like gene expressed in the mouse node during gastrulation, Nature 361 (1993) 543547. 27. Jones, C. M., Kuehn, M. R., Hogan, B. L., Smith, J. C. and Wright, C. V., Nodalrelated signals induce axial mesoderm and dorsalize mesoderm during gastrulation, Development 121 (1995) 3651-3662. 28. Feldman, B., Gates, M. A., Egan, E. S., Dougan, S. T., Rennebeck, G., Sirotkin, H. I., Schier, A. F. and Talbot, W. S., Zebrafish organizer development and germ-layer formation require nodal-related signals, Nature 395 (1998) 181-185. 29. David, N. B. and Rosa, F. M., Cell autonomous commitment to an endodermal fate and behaviour by activation of Nodal signaling, Development 128 (2001) 39373947. 30. Aoki, T. O., David, N. B., Minchiotti, G., Saint-Etienne, L., Dickmeis, T., Persico, G. M., Strahle, U., Mourrain, P. and Rosa, F. M., Molecular integration of casanova in the Nodal signalling pathway controlling endoderm formation, Development 129 (2002) 275-286. 31. Attisano, L., Silvestri, C., Izzi, L. and Labbe, E., The transcriptional role of Smads and FAST (FoxH1) in TGFbeta and activin signaling, Mol. Cell Endocrinol. 180 (2001) 3-11. 32. Kunwar, P. S., Zimmerman, S., Bennett, J. T., Chen, Y., Whitman, M. and Schier, A. F., Development 130 (2003) 5589. 33. Poulain, M. and Lepage, T., Mezzo, a paired-like homeobox protein is an


247


34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

immediate target of Nodal signalling and regulates endoderm specification in zebrafish, Development 129 (2002) 49014914. Kikuchi, Y., Agathon, A., Alexander, J., Thisse, C., Waldron, S., Yelon, D., Thisse, B. and Stainier, D. Y., casanova encodes a novel Sox-related protein necessary and sufficient for early endoderm formation in zebrafish, Genes Dev. 15 (2001) 1493-1505. Dickmeis, T., Mourrain, P., Saint-Etienne, L., Fischer, N., Aanstad, P., Clark, M., Strahle, U. and Rosa, F., A crucial component of the endoderm formation pathway, CASANOVA, is encoded by a novel sox-related gene, Genes Dev. 15 (2001) 1487-1492. Sakaguchi, T., Kuroiwa, A. and Takeda, H., A novel sox gene, 226D7, acts downstream of Nodal signaling to specify endoderm precursors in zebrafish, Mech. Dev. 107 (2001) 25-38. Chan, T. M., Chao, C. H., Wang, H. D., Yu, Y. J. and Yuh, C. H., Functional analysis of the evolutionarily conserved cis-regulatory elements on the sox17 gene in zebrafish, Dev. Biol. 326 (2009) 456-470. Reiter, J. F., Kikuchi, Y. and Stainier, D. Y., Multiple roles for Gata5 in zebrafish endoderm formation, Development 128 (2001) 125-135. Bjornson, C. R., Griffin, K. J., Farr, G. H., 3rd, Terashima, A., Himeda, C., Kikuchi, Y. and Kimelman, D., Eomesodermin is a localized maternal determinant required for endoderm induction in zebrafish, Dev. Cell 9 (2005) 523-533. Strahle, U., Blader, P. and Ingham, P. W., Expression of axial and sonic hedgehog in wildtype and midline defective zebrafish embryos, Int. J. Dev. Biol. 40 (1996) 929940. Norton, W. H., Mangoli, M., Lele, Z., Pogoda, H. M., Diamond, B., Mercurio, S., Russell, C., Teraoka, H., Stickney, H. L., Rauch, G. J. et al., Monorail/Foxa2 regulates floorplate differentiation and specification of oligodendrocytes, serotonergic raphe neurones and cranial motoneurones, Development 132 (2005) 645-658. Kurokawa, D., Kiyonari, H., Nakayama, R., Kimura-Yoshida, C., Matsuo, I. and Aizawa, S., Regulation of Otx2 expression and its functions in mouse forebrain and midbrain, Development 131 (2004) 33193331. Kurokawa, D., Sakurai, Y., Inoue, A., Nakayama, R., Takasaki, N., Suda, Y.,

44.

45.

46.

47.

48.

49.

50.

51.

Miyake, T., Amemiya, C. T. and Aizawa, S., Evolutionary constraint on Otx2 neuroectoderm enhancers-deep conservation from skate to mouse and unique divergence in teleost, Proc. Natl. Acad. Sci. USA 103 (2006) 19350-19355. Kurokawa, D., Takasaki, N., Kiyonari, H., Nakayama, R., Kimura-Yoshida, C., Matsuo, I. and Aizawa, S., Regulation of Otx2 expression and its functions in mouse epiblast and anterior neuroectoderm, Development 131 (2004) 3307-3317. Chan, T. M., Longabaugh, W., Bolouri, H., Chen, H. L., Tseng, W. F., Chao, C. H., Jang, T. H., Lin, Y. I., Hung, S. C., Wang, H. D. et al., Developmental gene regulatory networks in the zebrafish embryo, Biochim. Biophys. Acta 1789 (2009) 279-298. Sprague, J., Bayraktaroglu, L., Bradford, Y., Conlin, T., Dunn, N., Fashena, D., Frazer, K., Haendel, M., Howe, D. G., Knight, J. et al., The Zebrafish Information Network: the zebrafish model organism database provides expanded support for genotypes and phenotypes, Nucleic Acids Res. 36 (2008) D768-D772; 46a. Westerfield, M., et al., An on-line database for zebrafish development and genetics research, Semin. Cell Dev. Biol. 8 (1997) 477-488 Longabaugh, W. J., Davidson, E. H. and Bolouri, H., Computational representation of developmental genetic regulatory networks, Computational representation of developmental genetic regulatory networks, Dev. Biol. 283 (2005) 1-16. Driever, W., Solnica-Krezel, L., Schier, A. F., Neuhauss, S. C., Malicki, J., Stemple, D. L., Stainier, D. Y., Zwartkruis, F., Abdelilah, S., Rangini, Z. et al., A genetic screen for mutations affecting embryogenesis in zebrafish, Development 123 (1996) 37-46. Grapin-Botton, A. and Constam, D., Evolution of the mechanisms and molecular control of endoderm formation, Mech. Dev. 124 (2007) 253-278. Poulain, M., Furthauer, M., Thisse, B., Thisse, C. and Lepage, T., Zebrafish endoderm formation is regulated by combinatorial Nodal, FGF and BMP signaling, Development 133 (2006) 21892200. Muller, F., Albert, S., Blader, P., Fischer, N., Hallonet, M. and Strahle, U., Direct action of the nodal-related signal cyclops in induction of sonic hedgehog in the ventral midline of the CNS, Development 127 (2000) 3889-3897.


248


GLOBAL JOURNAL OF BIOCHEMISTRY 52. Peterkin, T., Gibson, A. and Patient, R., GATA-6 maintains BMP-4 and Nkx2 expression during cardiomyocyte precursor maturation, EMBO J 22 (2003) 4260-4273. 53. Rhinn, M., Lun, K., Amores, A., Yan, Y. L., Postlethwait, J. H. and Brand, M., Cloning, expression and relationship of zebrafish gbx1 and gbx2 genes to Fgf signaling, Mech. Dev. 120 (2003) 919-936. 54. Dougan, S. T., Warga, R. M., Kane, D. A., Schier, A. F. and Talbot, W. S., The role of the zebrafish nodal-related genes squint and cyclops in patterning of mesendoderm, Development 130 (2003) 1837-1851. 55. Kikuta, H., Kanai, M., Ito, Y. and Yamasu, K., gbx2 Homeobox gene is required for the maintenance of the isthmic region in the zebrafish embryonic brain, Dev. Dyn. 228 (2003) 433-450. 56. Thisse, B., Heyer, V., Lux, A., Alunni, V., Degrave, A., Seiliez, I., Kirchner, J., Parkhill, J. P. and Thisse, C., Spatial and temporal expression of the zebrafish genome by large-scale in situ hybridization screening, Methods Cell Biol. 77 (2004) 505-519; 56a. Patterson, L. J., Scl is required for dorsal aorta as well as blood formation in zebrafish embryos, Blood 105 (2005) 3502-3511. 57. Reiter, J. F., Alexander, J., Rodaway, A., Yelon, D., Patient, R., Holder, N. and Stainier, D. Y., Gata5 is required for the development of the heart and endoderm in zebrafish, Genes Dev. 13 (1999) 2983-2995. G. and Gerster, T., 58. Hauptmann, Combinatorial expression of zebrafish Brn1- and Brn-2-related POU genes in the embryonic brain, pronephric primordium, and pharyngeal arches, Dev. Dyn. 218 (2000) 345-358. 59. Spaniol, P., Bornmann, C., Hauptmann, G. and Gerster, T., Class III POU genes of zebrafish are predominantly expressed in the central nervous system, Nucleic Acids Res. 24 (1996) 4874-4881. 60. Trinh, L. A., Meyer, D. and Stainier, D. Y., The Mix family homeodomain gene bonnie and clyde functions with other components of the Nodal signaling pathway to regulate neural patterning in zebrafish, Development 130 (2003) 4989-4998. 61. Elsalini, O. A., von Gartzen, J., Cramer, M. and Rohr, K. B., Zebrafish hhex, nk2.1a, and pax2.1 regulate thyroid growth and differentiation downstream of Nodaldependent transcription factors, Dev. Biol. 263 (2003) 67-80. 62. Dufort, D., Schwartz, L., Harpal, K. and Rossant, J., The transcription factor

63.

64.

65.

66.

67.

68.

69.

HNF3beta is required in visceral endoderm for normal primitive streak morphogenesis, Development 125 (1998) 3015-3025. Reim, G., Mizoguchi, T., Stainier, D. Y., Kikuchi, Y. and Brand, M., The POU domain protein spg (pou2/Oct4) is essential for endoderm formation in cooperation with the HMG domain protein casanova, Dev. Cell. 6 (2004) 91-101. Yuh, C. H., Dorman, E. R., Howard, M. L. and Davidson, E. H., An otx cis-regulatory module: a key node in the sea urchin endomesoderm gene regulatory network, Dev. Biol. 269 (2004) 536-551. Maduro, M. F., Endomesoderm specification in Caenorhabditis elegans and other nematodes, Bioessays 28 (2006) 10101022. Mavropoulos, A., Devos, N., Biemar, F., Zecchin, E., Argenton, F., Edlund, H., Motte, P., Martial, J. A. and Peers, B., sox4b is a key player of pancreatic alpha cell differentiation in zebrafish, Dev. Biol. 285 (2005) 211-223. Heicklen-Klein, A. and Evans, T., T-box binding sites are required for activity of a cardiac GATA-4 enhancer, Dev. Biol. 267 (2004) 490-504. Davidson, E. H. and Erwin, D. H., Gene regulatory networks and the evolution of animal body plans, Science 311 (2006) 796800. Oliveri, P., Walton, K. D., Davidson, E. H. and McClay, D. R., Repression of mesodermal fate by foxa, a key endoderm regulator of the sea urchin embryo, Development 133 (2006) 4173-4181.


249