Def Functions as a Cell Autonomous Factor in Organogenesis of Digestive Organs in Zebrafish

Digestive organs originate from the endoderm. Morphogenesis of the digestive system is precisely controlled by multiple factors that dictate the cell fate and behavior so that the specific digestive organs are timely formed in the right place and develop into right size and structure. We showed previously that digestive organ expansion factor (def) is a gene whose expression is enriched in the liver, pancreas and intestine. Loss-of-function of def in the defhi429 mutant confers hypoplastic digestive organs partly due to alteration of expression of genes related to the p53 pathway. However, the molecular mechanism for the involvement of Def in the organogenesis of digestive organs is still largely unknown. For example, it is not known whether Def regulates specific pathways in a specific organ. To address this question, we generated four independent Tg(fabp10a:def) transgenic fish lines which over-expressed Def specifically in the liver. We characterized Tg-I, one of the transgenic lines, in detail with genetic, molecular and histological approaches. We found that Tg-I restored the liver but not exocrine pancreas and intestine development in the defhi429 mutant. However, Tg-I adult fish in the wild type (WT) background exhibits reduced liver-to-body ratio and all four transgenic lines conferred abnormal intrahepatic structure. Microarray data analysis showed that certain specific functional pathways were affected in the liver of Tg-I. These results demonstrate that Def functions in a cell autonomous manner during early liver development and aberrant Def protein expression might lead to disruption of the structural integrity of a normal adult liver.


Introduction
Digestive system is formed by digestive tract and accessory organs such as liver and exocrine pancreas [1]. Zebrafish has been proven to be an excellent model system for the study of molecular control of organogenesis of digestive organs [2]. As in other vertebrate, zebrafish digestive system originates from the endoderm, one of the three germ layers defined during gastrulation [3]. Organogenesis of zebrafish digestive organs is controlled by spatially and temporally expressed endodermal factors and mesodermal signaling molecules [4][5][6][7][8][9][10]. We previously reported that Def was a pan-endodermal enriched factor that is essential for the growth of digestive organs in zebrafish [11]. Def is an evolutionally conserved molecule across eukaryotes from human to yeast [11]. Recent studies showed Def homologs in Saccharomyces cerevisiae (Utp25p) [12,13] and Arabidopsis thaliana (NOF1) [14] are nucleolus-localized proteins and regulate pre-rRNA processing as the component of small subunit (SSU) processome. The nucleolus is a subnuclear structure that exhibits dynamic morphological changes during cell cycle. The nucleolus serves as the site for rRNA biosynthesis, processing and maturation, and also as the site for assembly of ribosome large and small subunit [15]. Therefore, disruption of the nucleolus function is normally detrimental to a cell [15]. Recently, evidence has shown that some nucleolar factors are also essential for organogenesis during embryogenesis. For example, loss-of-function of wrd36 confers small eyes and hypoplastic digestive organs [16] and bap28 mutation leads to neurodegeneration [17] in zebrafish.
In this work, we sought to address the question how Def, as a nucleolar factor, regulates organogenesis of digestive organs in zebrafish. We focused on a specific question: can Def's function in the liver be uncoupled from that in the exocrine pancreas and intestine? Does Def regulate specific functional pathways in a specific organ? To address this question, we generated four independent Def transgenic lines in that def expression was under the control of liver-specific promoter fabp10a [18]. We crossed Tg-I, one of the transgenic lines with def hi429 mutant and found that only the liver but not intestine and exocrine pancreas development in the mutant was rescued to normal by Tg-I. Through comparing the global gene expression profiles we found that Def overexpression in the adult liver of Tg-I altered the expression of genes in specific functional pathways. Histology analysis revealed that the adult liver in all four transgenic lines suffered from disorganized intrahepatic structure. Our result shed lights on understanding how Def regulates organogenesis of digestive organs in zebrafish.

Ethics statement
This study did not involve non-human primates. All experiments described in this study were performed in full accordance with the guidelines for animal experiments released by the National Institute of Animal Health. This study is approved by the Animal Ethic Committee at Zhejiang University (ETHICS CODE Permit NO. ZJU2011-1-11-009Y).

Zebrafish lines and maintenance
Zebrafish (Danio rerio) were raised and maintained in the standard Zebrafish Unit (produced by Aisheng Zebrafish Facility Manufacturer Company, Beijing, China) at Zhejiang University under a constant 14 hour on/10 hour off light cycle at 28uC. Zebrafish wild type AB strain was used in this study. The def hi429 mutant line [19] was provided by Professor Nancy Hopkins at Massachusetts Institute of Technology (Cambridge, USA). The two pairs of primers derived from lacZ and def were used to genotype the def hi429 mutant [11]. Tg(fabp10a:def) transgenic fish lines were generated through injection of the fabp10a:def plasmid DNA into zebrafish embryos. Primer pair def_Ex4_Fw336 (59- Figure 3. The def transgene in Tg-I rescues the liver but not intestine and exocrine pancreas development in def hi429 . (A, B) WISH was performed to analyze the digestive organs in embryos at 4dpf produced by Tg-I +/2 def hi429/+ 6 def hi429/+ using the liver marker fabp10a, intestinal marker fabp2 and exocrine pancreas marker trypsin (A). Embryos from (A) were then individually genotyped and matched to their phenotype. Number of embryos identified in each category was as shown in (B). (C) Progenies from Tg-I +/2 def hi429/+ 6 def hi429/+ at 3.5dpf were also examined simultaneously by Fast Red-labeled fabp10a and trypsin probes BCIP/NBT-labeled fabp2 and insulin probes. Number of embryos in each category was obtained as shown after genotyping a total of 51 embryos from the cross. en: endocrine pancreas, ex: exocrine pancreas, in: intestine, lv: liver. doi:10.1371/journal.pone.0058858.g003 GGACACCGATGAAGAGGACGGGA-39) and def_Ex6_Rv730 (59-AGGCCTGTCCGATGGCTGGA-39) that spans intron IV and V of the def gene was used to genotype the Tg(fabp10a:def) transgenic lines.
DNA constructs and microinjection of plasmid DNA def full length cDNA was amplified by primer pair NheI-def (59-TTCGCTAGCATGGGCAAAAGAAGGCGAGGAAAA-39) and def-BamHI ( 5 9-CGCGGATCCT-CATGTGCTTTTCTCCTCCCCCGT-39) according to the sequence information reported previously [11]. For construction of the fabp10a:def fusion DNA construct, zebrafish def full length cDNA tailed with SV40 polyadenylation (pA) signal was cloned downstream of a 2.8-kb 59-flanking sequence of zebrafish fabp10a gene in the pEGFP-C1 vector [18], in such way the fabp10a promoter will drive def expression specifically in hepatocytes in the transgenic fish. The fabp10a-def DNA plasmid DNA was linearized by NotI and SfiI, purified and adjusted to 500 ng/ml in ddH 2 O containing phenol red dye. Approximately 200 pl of the DNA solution was injected into the zebrafish fertilized eggs at one-cell stage.
For generation of transgenic fish using the miniTol2 transposon system, the fabp10a:def fusion DNA fragment was cloned into the pDB739 vector between the left and right recognition boarders for Tol2 transposase. Tol2 mRNA was obtained by in vitro transcription of pT3TS/Tol2 plasmid linearized by SmalI. One nanolitre of the plasmid DNA (50 ng/ml) and Tol2 mRNA (50 ng/ ml) mix was injected into the fertilized eggs at one-cell stage.

DNA extraction and southern blotting analysis
Zebrafish genomic DNA was extracted from embryos with Genomic DNA Cell & Tissue Kit (Aidlab) using the protocol recommended by the manufacturer. 30 mg genomic DNA was digested with EcoRI, separated by electrophoresis in 0.6% agarose gel. Probe was Digoxigenin (DIG)-labeled and southern blot hybridization was performed according to the manufacturer's instructions (Roche Diagnostics). def probe was amplified by primers def probe_Fw301 (59-GCTGAAGTTGAAGGTGA-TAGTGAA-39) and def probe_Rv575 (59-TCTTCATCCTGTTCCTGTTGTGCT-39) using the def plasmid as the template.
Whole-mount RNA in situ hybridization (WISH) WISH was performed as described [11]. def full-length, fabp10a, fabp2, trypsin and insulin RNA probes were labeled with DIG and their sequences information was described previously [11]. Photos were taken under a Nikon AZ100 microscope.

RNA extraction, northern blotting analysis and qPCR
Total RNA from different samples was extracted using TRIzol (Invitrogen) according to manufacturer's instructions. Probes were DIG-labeled and northern blot hybridization was performed according to the manufacturer's instructions (Roche Diagnostics). The 59-ETS, ITS1 and ITS2 probes were as described previously [17]. For real-time quantitative PCR (qPCR), total RNA was treated with DNase I prior to reverse transcription and purified with RNeasy mini kit (Qiagen). First strand cDNA was synthesized using M-MLV Reverse Transcriptase (Invitrogen). The qPCR was  Table S1.

Microarray hybridization and data analysis
We monitored the gene expression in samples using Gene-ChipH Zebrafish Genome Array (Affymetrix). The Affymetrix GeneChipH Zebrafish Genome Array can be used to study gene expression of over 14,900 Danio rerio genes.
Total RNA was first treated with DNase I and then purified through Qiagen RNeasy mini kit (Qiagen). RNA was quantified on the Nanodrop ND-1000 and RNA integrity was checked with Agilent 2100 Bioanalyzer. Samples were then amplified and labeled using the 39 IVT Express Kit (Affymetrix) and hybridized with the GeneChipH Zebrafish Genome Array for 16 h. After hybridization, the arrays were then washed and stained with the Affymetrix fluidics station 450, followed by scanned with the Affymetrix 3000 7G plus scanner. Dat and Cel files were obtained by the AGCC software, and the CHP files were generated with MAS5.0 method by Affymetrix Expression Console. Raw data were imported into the Agilent GeneSpring GX 11.0 software for data analysis.

Protein analysis and antibodies preparation
Zebrafish embryos were deyolked in ice-cold PBS by passing through a 21 gauge needle and then lysed in Tris lysis buffer (PH8.0, EDTA free) containing 16Complete Protease Inhibitor Cocktail (EDTA-free, Roche) and 25 mM MG132 (Sigma) by passing through a 26 gauge needle. Liver lobes were dissected from the adult fish after anaesthesia, and lysed directly in the same Tris lysis buffer. Protein samples were stored at 220uC after denaturation.
Proteins were separated on SDS-PAGE gel and then transferred to PVDF membrane (Millipore). Membranes were subjected to western blotting analysis with appropriate antibodies. Chemiluminescent signals were detected with ECL solutions according to manufacturer's instructions (Pierce). A zf-Def rabbit polyclonal antibody was used as described [11]. A mouse monoclonal (38F3) antibody against human Fibrillarin (ab4566) was purchased from Abcam.

Cryo-sectioning and immunofluorescence staining
Cryo-sectioning was performed as described [11]. For immunofluorescence staining of cryo-sections, the slides were immersed in 0.01 M citrate (pH 6.0) and boiled in a microwave oven for three times, each for 5 min with the purpose to expose the antigens. After overnight cooling at 4uC in the retrieval buffer, the slides were briefly washed in PBST and incubated with proper primary antibody, followed by Alexa Fluor conjugated secondary antibody (Invitrogen). The slides were then mounted with a mount medium containing DAPI (Vector), and the images were taken under a Leica TCS SP5 confocal microscope.

LBR measurement and H&E (Hematoxylin and Eosin) staining
LBR was measured according to the method described previously [20]. H&E staining was performed as described [21]. Cell activity and structure 30q14Q Enzymes and proteinase inhibitor 13q14Q

Ion binding 7q2Q
Carriers and transporters 5q10Q

DNA binding and biosynthesis 1q2Q
Protein degradation 3q3Q Muscle function related 1q2Q Neural function related 2q1Q Immune response 2q1Q Unclassified 4q5Q Unknown and unnamed proteins 29q27Q *Refer to

Results
Generation of Tg (fabp10a:def) transgenic zebrafish The fabp10a:def plasmid DNA was linearized by NotI and SfiI ( Figure 1A) and was then injected into zebrafish fertilized eggs at one-cell stage. Three months later, these founders (F0 fish) were crossed with wild type (WT) fish. The progenies of each F0 fish were collected and pooled. Genomic DNA was extracted from the pooled embryos and used in PCR for screening for the germ line integration founders. Progenies (F1) from positive founders were allowed to grow to adulthood. The scales of each individual F1 fish were collected and lysed. The obtained DNA samples were used in PCR to identify the individuals harboring the transgene. WT fish is expected to produce a PCR product of a single band of 681 bp while a transgenic fish would yield a 395 bp band corresponding to the transgene that lacks the intron IV and V of the def gene ( Figure 1B and C, lanes for WT and plasmid) in addition to the 681 bp band. Out of 37 founder fish, one line was identified to be a transgenic fish and was designated as Tg-I ( Figure 1C, lane for Tg-I). We also obtained other three transgenic lines, including the line Tg(fabp10a:def)-II (Tg-2) ( Figure 1C, lane for Tg-II), Tg(fabp10a:def)-III (Tg-III) and Tg(fabp10a:def)-IV (Tg-IV) (data not shown), through using the miniTol2 transposon system.
We found that the progenies of Tg-I +/2 6 WT and of cross of male and female Tg-I +/2 showed a 1:1 and 1:3 Mendelian segregation ratios (data not shown), respectively, suggesting that there is a single insertion in Tg-I. Southern blotting was performed to analyze the transgene in Tg-I and Tg-II. The result confirmed that Tg-I line harbored with a single insertion while Tg-II line with multiple insertions ( Figure 1D).
Def is over-expressed in the liver in Tg-I fabp10a is a liver specific gene whose expression is detectable in the hepatocytes from 2 dpf to adulthood in zebrafish [18,22]. It is a relatively strong promoter which has been used to drive the reporter genes EGFP or RFP in previous studies [18,23]. Since the fabp10a promoter is the most well characterized liver specific promoter in zebrafish [18] we used it to drive the expression of def in the hepatocytes. WISH was performed on the progenies derived from Tg-I with the DIG-labeled def full-length probe. We found that only those embryos harbored the transgene showed strong def expression specifically in the liver in a pattern similar to that did the fabp10a gene at 4.5 dpf [18], while the WT siblings exhibited weak endogenous def expression in the entire digestive organs (Figure 2A) as reported previously [11]. Western blot analysis with a Def polyclonal antibody showed that Def protein level in Tg-I was approximately 9-fold higher than that in the wild type control ( Figure 2B). These evidences demonstrated that Def is overexpressed in Tg-I. Def protein is localized in the nucleolus Def homologs in Saccharomyces cerevisiae (Utp25p) [12,13] and Arabidopsis thaliana (NOF1) [14] have been shown to be localized in the nucleolus. We performed co-immunostaining of sections from 4.5 dpf old Tg-I fish with a Def antibody and an Fibrillarin (Fib, a nucleolar marker) antibody. The result showed that the overexpressed Def in the liver of Tg-I is co-localized with Fib within or at the periphery of the dense fibrillar component (DFC) regions of the nucleoli ( Figure 2C).
The def transgene in Tg-I rescues the liver but not intestine and exocrine pancreas development in def hi429 We showed in our previous report that def expression is enriched in the liver, exocrine pancreas and intestine but not in the endocrine pancreas. This expression pattern coincides with the hypoplastic phenotype of the liver, exocrine pancreas and intestine but with a normal endocrine pancreas exhibited by the def hi429 mutant [11], suggesting that Def likely functions in a cell autonomous manner. In addition, our result showed that the development of the entire digestive system including liver, intestine and exocrine pancreas in the def hi429 mutant could be restored to normal by def mRNA injection [11]. Apparently, the function of Def in a specific organ cannot be assessed by def mRNA injection into the def hi429 mutant embryos since all cells will have equal chance to get the injected def mRNA. On the other hand, the Tg-I transgenic fish provided us with a unique chance because the def transgene is specifically expressed in the hepatocytes. We reckoned that if Def works in a cell autonomous manner the transgene will only be able to rescue the liver development but not the intestine and exocrine pancreas development in def hi429 . To test this hypothesis, the Tg-I fish was first mated with def hi429 heterozygous (def hi429/+ ) fish to get the Tg-I +/2 def hi429/+ F1 fish. The progenies of Tg-I +/2 def hi429/+ 6 def hi429/+ were examined by WISH using the liver marker fabp10a, intestinal marker fabp2, and exocrine pancreatic marker trypsin ( Figure 3A). All embryos were then genotyped individually and classified into three groups, namely siblings (including Tg-I +/2 , Tg-I +/2 def hi429/+ , def hi429/+ , and wildtype), Tg-I +/2 def hi429 double and def hi429 single lines ( Figure 3B). Data analysis revealed that the liver development in all Tg-I +/2 def hi429 embryos was rescued by the def transgene ( Figure 3B). In contrast, the Tg-I def hi429 embryos still exhibited hypoplastic intestine and exocrine pancreas, same as that observed in def hi429 ( Figure 3B). We confirmed the above result by performing WISH to examine simultaneously the expression patterns of fabp10a, fabp2, trypsin and the endocrine marker insulin in the progenies of Tg-I +/2 def hi429/+ 6def hi429/+ ( Figure 3C). These genetic data demonstrate that 1) the def transgene in Tg-I is functional equivalent to the endogenous def gene, and 2) although Def expression is enriched in the entire digestive system it works in a cell autonomous manner.
The def transgene continues to express at a high level in the adult liver We followed expression of the def transgene in Tg-I to the adulthood. We first compared the transcriptional expression of def in the liver samples from three months old Tg-I and WT fish via qPCR. The result showed that def was expressed 268-folds higher in Tg-I than that in the WT fish. Western blot of Def in the liver samples showed that the level of Def protein was 3.3-fold higher in Tg-I than that in the WT fish ( Figure 4A).

The liver of Tg-I accumulates 18S rRNA intermediate precursor
Def is a nucleolar protein that has been found to regulate 18S rRNA processing in yeast and Arabidopsis [12][13][14]. Northern blot hybridization was performed to compare rRNA processing in the adult liver using probes detecting different rRNA precursors as indicated in Figure 4B (right panel). The result revealed that 18S rRNA precursor ( Figure 4B, form iv as indicated, detected by 59-ETS and ITS1 probes) but not 28S rRNA precursors ( Figure 4B, form iii as indicated, detected by ITS2 probe) dramatically accumulated in the adult liver in Tg-I.

The liver of Tg-I has a reduced LBR
We have shown that Def is essential for the expansion growth of digestive organs during the embryonic stage. We were intrigued to know whether Def over-expression in Tg-I will produce a larger liver. Zebrafish LBR is a sex but not age-dependent constant which has been used to evaluate liver growth by other researchers [20]. We measured the LBRs of 6-month-old WT and Tg-I male fish, respectively. To our surprise, the LBR in Tg-I is approximately 35.4% lower (p,0.01) than that in the WT fish ( Figure 4C).
Gene expression profiles were altered in the liver of Tg-I In view of LBR reduction in Tg-I we decided to perform microarray analysis with the purpose to find out whether Def overexpression will affect global gene expression profiles that might lead to LBR reduction. Total RNA from the adult liver of three independent batches of 3-month-old Tg-I and WT (AB strain) male fish was extracted using TRIzol (Invitrogen), each batch with the liver samples from three fish mixed together. Before subjecting to microarray hybridization, the RNA samples were confirmed to express dramatically high level of def by qPCR (data not shown). After data validation according to the procedure instructed by Affymetrix, the genes were filtered on flags that at least 3 out of 6 samples had values Present [P] or Marginal [M]. The logarithm base 2 of the signal ratio of Tg-I to WT ,21 or .1, as well as the unpaired t-test p value #0.05 were used as the cut-off value to identified candidate genes differently expressed in the liver in Tg-I. A total of 208 genes including def were identified (Table S2). After removing the same genes detected by redundant probe sets, a total of 202 genes were identified as differently expressed genes, of which 110 genes were up-regulated (Table S2) and 92 genes were down-regulated (Table S2). qPCR was performed to confirm the microarray data. Among 16 genes selected (selected from the list of genes with expression at least 2.5-fold higher or lower in Tg-I when compared to that in the WT control), expression of 12 genes (accounting for 75%) was confirmed to be up-or down-regulated in Tg-I ( Figure 5A) by qPCR analysis.
Blast search revealed that 137 genes could be assigned a biochemical function while 65 belonged to unclassified, unknown or unnamed genes (Table 1 and Table S2). Gene ontology analysis classified the 137 genes into 11 categories. Among these 137 genes, 12 genes encode for transcription related proteins (including transcription factors, co-transactivators and transinhibitor), 12 for RNA binding and processing proteins (including RNA binding and splicing factors, RNA processing and ribosome biogenesis factors and tRNA synthetases), 44 for cell activity and structure related proteins (including cell cycle, proliferation, apoptosis, adhesion and migration proteins, energy metabolism, endocytosis and chromatin remodeling proteins, receptors, chaperones, cytokine, cytoskeleton related proteins and membrane proteins), 27 for various enzymes and proteinase inhibitor (including synthetase, synthase, hydratase, hydroxylase, dehydrogenases, decarboxylase, transferases, transaldolase, lyases, hydrolase, lysozyme, oxidase, reductase, protein kinases, protein phosphatase, pyrophosphatase, phospholipase, peptidase, mutase, metalloproteinase and proteinase inhibitor), 9 for ion binding proteins, 15 for carrier and transporter proteins, 3 for DNA binding and biosynthesis proteins, 6 for proteins involved in protein degradation, 9 for proteins involved in muscle function, neural function and immune response (Table S2).
Genes in specific functional pathways were down-or upregulated in a coordinate way in the liver of Tg-I We then summarized genes up-or down-regulated in each of the 11 categories with the purpose to find out the pathway(s) coordinately affected by Def over-expression in the liver in Tg-I. Data analysis revealed that only eight transcription factor genes were expressed differently in the liver between Tg-I and the WT fish. Interestingly, six out of the eight liver transcription factor genes, namely cebpa, rogdi, znf598, arid3b, hnf4a and znhit1 were down-regulated ( Figure 6A). hnf4a and cebpa were decreased by 2.4-fold and 3.6-fold in Tg-I, respectively in our microarray analysis, and their expression reduction was further validated by qPCR analysis (Figure 5A).
Hnf4a and C/EBPa are two well-known liver key transcription factors that control the expression of large amount of liver genes [24,25]. Therefore, down-regulation of Hnf4a and C/EBPa is expected to cause down-or up-regulation of their target genes which are normally up-or down-regulated by them. Bioinformatic analysis of the 60 down-regulated functional known genes in Tg-I revealed that, in addition to the 6 transcription factor genes mentioned in the above, many of the metabolic enzyme genes (15 genes) and solute carrier/transport genes (10 genes) ( Figure 6B) were collectively down-regulated (Table S2). On the other hand, among the 77 up-regulated genes, genes in the categories of RNA binding and processing (9 genes) ( Figure 6C), cell cycle and proliferation (7 genes) ( Figure 6D), energy metabolism (7 genes), chaperones (5 genes), cytoskeleton related (6 genes), non-catabolic enzymes (12 genes) and calcium ion binding (6 genes) were collectively up-regulated in Tg-I (Table S2). Considering the fact that liver is a relatively homogenous tissue, this observation strongly suggests that Def regulates specific functional pathways in the liver.

Def over-expression causes disorganization of intrahepatic structure
Up-regulation of genes involved in cell cycle and proliferation (mainly related to cell cycle arrest) ( Figure 6D) and genes related to chaperone and cytoskeleton and others suggest that the proliferation and interconnection of hepatocyes might be affected in Tg-I. Indeed, analysis of H&E stained liver sections from 3-monthold WT and Tg-I fish showed that cells in the Tg-I liver loosened in cell-cell junction and failed to form regular patterns as that observed in the WT fish ( Figure 7A and B). Furthermore, the liver in Tg-I had increased number and size of gap regions within when compared to the liver in a wild type fish ( Figure 7B).
In order to exclude the possibility that this abnormal adult liver phenotype in Tg-I could be due to positional effect (e.g disrupt a gene or affect the expression of adjacent genes by the insertion), we analyzed H&E stained liver sections from the remaining three independent transgenic lines Tg-II +/2 , Tg-III +/2 and Tg-IV +/2 . We first performed the qPCR analysis and found that def expression was approximately 145, 27, 18 and 26 folds higher in Tg-I +/+ , Tg-II +/2 , Tg-III +/2 and Tg-IV +/2 , respectively ( Figure 8A and B). Interestingly, western blot analysis showed that the protein levels in these four independent transgenic lines did not differ in a corresponding scale to that of their transcripts and showed 6.4, 2.8, 2.3 and 4.9 folds higher in Tg-I +/+ , Tg-II +/2 , Tg-III +/2 and Tg-IV +/2 , respectively, than its expression in the WT liver ( Figure 8C and D). Surprisngly, qPCR analysis of the same 16 genes listed in Figure 5A showed that the expression of these genes in Tg-II +/2 , Tg-III +/2 and Tg-IV +/2 at different age or the same age did not follow the trend or even opposite changes when compared with that observed in Tg-I ( Figure 5B-D, Figure 9). We also noticed that even in different batches of Tg-I at the same age the expressions of these 16 genes were also different ( Figure 9). Despite the discrepancy in gene expression among different Def over-expression transgenic lines, histological analysis showed the adult liver of Tg-II +/2 , Tg-III +/2 and Tg-IV +/2 exhibited similar phenotype ( Figure 7C-E) as that observed in Tg-I ( Figure 7B).

Discussion
The alimentary tract comprises functional distinct organs each contains specific cell types. From the view of developmental biology, the key question is how specification of different organs and interconnection of these organs to form a complete digestive system is achieved during the process of organogenesis. Three groups of factors are believed to contribute to this process: organ specific factors that define the identity of a specific organ, panendodermal factors that coordinate the development of different organs to form a functional digestive system, and signaling molecules from surrounding mesoderm cells that mediate the coordination. Def expression is enriched in the digestive organs, raising a question as to how Def performs its function in different organs. To address this question, it is ideal to express Def in a specific digestive organ in the background of def hi429 . In this way we could examine 1) whether the specific organ development can be restored to normal, and 2) whether it will alter the cell fate in the organs expressing Def. To address this question we generated the transgenic fish Tg(fabp10a:def) in which def is specifically expressed in the liver. Our result showed that the def transgene in Tg-I only rescued the liver but not intestine and exocrine pancreas development in def hi429 . Meanwhile, the liver of Tg-I expressed the liver marker fabp10a but not the intestinal marker fabp2 and exocrine pancreatic marker trypsin. These data demonstrated that Def functions as a cell autonomous factor to regulate the development of digestive organs although it appears not to be a key cell fate determinant.
Previous reports showed that Def homologs in yeast and Arabiopsis are nucleolar proteins and are involved in the processing of 18S but not 28S rRNA intermediate precursors [12][13][14]. We wondered whether continuous over-expression of Def in Tg-I will affect the processing of 18S rRNA. To our surprise, we found that processing of 18S rRNA precursor in the adult liver of Tg-I was affected as that observed in the yeast mutant Utp25 2/2 and Arabidopsis mutant nof1, suggesting that the level of Def must be kept in a narrow range because too little or too much both will affect the processing of rRNA precursors mediated by Def. Lossof-function of Def leads to hypoplastic digestive organs, it is of our great interest to find out whether over-expression of Def would promote liver development in our future studies.
By comparing global gene expression profiles we previously reported that 141 genes were down-regulated and 23 genes upregulated in the def hi429 mutant [11]. However, that experiment was performed using total RNA extracted from whole embryos at 5 dpf, therefore the data cannot help us to identify specific pathways in a specific organ based on down-regulated genes because of heterogeneity of origin of the total RNA used. The liver is considered to be a relatively homogenous organ because it contains limited types of cells, with majority (.60%) being hepatocytes. Therefore, over-expressing Def in the liver will not only allow us to study Def's role in the liver development but also serves as a model to define Def's role in regulating gene expression. We compared the global gene expression profiles in the adult liver between Tg-I and WT. We found that the def transgene in Tg-I altered a fraction of genes that can be clearly classified into a few distinct pathways in the liver. Among these pathways, we noticed that nine genes related to rRNA binding and processing (namely elavl1, cars, yars, snrpg, sfrs5a, cars, ddx27, bop1, qars) were upregulated. This indicates that Def might not only directly regulate Figure 9. Summary of the expression changes of the 16 genes in the adult livers of four transgenic lines with respect to that in the WT liver. The summary was based on the qPCR data shown in Figure 7. Notably, the data showed that the expression of these 16 genes in Tg-II +/2 , Tg-III +/2 and Tg-IV +/2 did not follow the trend or even opposite changes when compared with that observed in Tg-I. Surprisingly, the expression of these 16 genes in different batches of Tg-I at the same age were also different. doi:10.1371/journal.pone.0058858.g009 rRNA processing as a component of the SSU processome but also affect rRNA homeostasis through other genes and pathways.
It is well-known that the expression of liver-specific and/orenriched genes are under the control of a network formed by transcription factors including hepatic nuclear factors HNF1, HNF3, HNF4a, HNF6, and C/EBPa (CCAAT/enhancer binding protein) etc [26][27][28][29][30][31][32]. Therefore, down-regulation of hnf4a and C/EBPa provides a logic explanation why majority of genes in the categories of catabolic enzyme and solute carrier/transporter were down-regulated in the liver of Tg-I.
Analysis of the adult liver revealed that the LBR is reduced significantly in Tg-I. Histology study showed that hepatocytes appeared to be loosely connected and there were apparent distended gap regions inside the liver mass in all four independent transgenic lines as well. These results suggest that over-expression of Def in these four transgenic lines not only affects the processing of 18S rRNA precursors but also changes the structural integrity of the adult liver. The structural abnormality of the adult liver in Tg(fabp10a:def) might be related to the up-regulation of genes involved in the pathways of cell cycle and proliferation, cytoskeleton related, chaperones and calcium ion binding, and to down-regulation of genes in the pathways of cell adhesion and migration, solute carrier/transporter and DNA remodeling factors. However, one has to be cautious in linking the gene expression profiles with the abnormal liver structure in this specific case since we found that the up-or down-regulated genes in one transgenic line was not the case or even opposite in another transgenic line. This discrepancy might be due to the timing of RNA sampling since the liver structure in all four transgenic lines has already accumulated pathologically identifiable disruptions at 3 months which, conceivably, would result in highly variable gene expression profiles in different transgenic lines at this time point due to different degrees of damages to the hepatocytes. We hypothesize that Def over-expression initially causes changes in expression of certain key genes that are essential to make a perfect liver in all transgenic lines. Changes in the expression of these key genes will ultimately disrupt the liver structure within a time period which, depending on the level of Def expressed, is different in duration for different transgenic lines. Apparently, a key task in the future is to determine the time point when the liver structure just displays abnormality for each independent transgenic lines and then to use this time point or prior to this time point for RNA sampling.
In conclusion, the generation of the Tg(fabp10a:def) transgenic fish has allowed us to study Def's function both in the embryonic and adult liver. We found that while Def is essential for the organogenesis of the liver at the embryonic stage its continuous high expression in the adult liver is harmful, leading to structural abnormality of the liver. Therefore, Def protein level must be tuned to an appropriate level to maintain the normal function of the liver. However, we currently do not know whether alteration of gene expression profiles in the Tg-I liver is due to aberrant processing of 18S rRNA precursors or to Def's other function. Neither we know whether Def works by its own or team up with other factors to perform its function. Future efforts on addressing these questions will no doubt help us to understand more about the role of the nucleolus in organogenesis of digestive organs.