Isolation and characterization of the zebrafish Danio rerio insulin-like growth factor binding protein-3 promoter region

Abstract

ABSTRACT:  Insulin-like growth factor binding protein-3 (IGFBP-3) is expressed in many different cell types and has a regulatory function in the insulin-like growth factor (IGF) system. To further understand the molecular mechanism, we cloned and sequenced the 5′-promoter region of zebrafish Danio rerio IGFBP-3 and characterized its activity using firefly luciferase transient transfection expression assays. Different fragments of the zebrafish IGFBP-3 5′-flanking region were transfected into HeLa and ZFL cells. In these cell lines, maximum promoter activity was located within 1218 bp of the zebrafish IGFBP-3 flanking region in the HeLa cell line and within 304 bp of the zebrafish IGFBP-3 flanking region in the ZFL cell line. Several putative transcription factors were revealed in the zebrafish IGFBP-3 promoter region, such as organic cation transporter-1, GATA-1 and yin and yang-1. Further study of the in vivo expression of the IGFBP-3 promoter during development was carried out in transgenic zebrafish expressing an IGFBP-3 promoter-driven green fluorescent protein (GFP) encoding the GFP cDNA transgene. The GFP transcripts appeared for the first time in the 32-cell stage. These results indicate that the IGFBP-3 promoter is active in a development-specific manner and suggest that the IGFBP-3 promoter plays an important role in teleost embryo growth. Finally, IGFBP-3 has important IGF-independent effects on cell growth and may involve nuclear localization. By transiently transfecting HeLa cells with various zebrafish IGFBP-3 segments, we identified one nuclear localization signal (NLS): a basic amino acid rich sequence (PSKGRKR) between amino acids 256 and 262 was able to direct enhanced green fluorescence protein predominantly into the nucleus, whereas a deletion of this motif abrogated this nuclear localization property. These data suggest that zebrafish IGFBP-3 contains a NLS around the second NLS sequence, whereas the putative NLS at the first NLS is non-functional.

INTRODUCTION

Insulin-like growth factors (IGF-I and IGF-II) are bound in biological fluids with IGF-binding proteins (IGFBPs). There are six IGFBPs (IGFBP-1, -2, -3, -4, -5 and -6) that bind IGF peptides with high affinities.¹ However, in adult mammals, 80–85% of IGFs circulate as 150-kDa ternary complexes. These ternary complexes are comprised of one molecule each of IGF, IGFBP-3 (the predominant IGFBP in serum) and an acid-labile subunit (ALS).² Expressions of genes encoding the various IGFBPs have been observed in many tissues and are subject to intricate physiological regulation. In fact, IGFBP-3 plays an important role in regulating pharyngeal skeletal and inner ear formation and differentiation in zebrafish Danio rerio.³ IGFBP-3 has growth inhibitory activity that is independent of its IGF-binding properties. Endogenous IGFBP-3 was detected in the nuclei of A549 human lung cancer cells, and the IGFBP-3 existing in nuclei was shown to have nuclear transport ability and to have C-terminal nuclear localization signals (NLSs). This phenomenon indicates that IGFBP-3 has direct effects on gene transcription by interacting with 9-cis retinoic acid receptor-alpha, a nuclear receptor.⁴ As a result of both structural and functional differences among IGFBPs, each IGFBP has its own spectrum of actions and its own unique pattern of expression. The biological activities of these IGFBPs can be modified by post-translational proteolytic cleavage, phosphorylation and glycosylation.⁵

The IGFBP-3 gene is highly conserved among species and is present as a single copy on chromosome 7p14-p12 in humans. However, its promoter activity and gene structure in fish are unknown. Growth hormone (GH) and insulin are two hormones that are important in IGFBP-3 upregulation.⁶ It is interesting to note in our recently published study that both starvation of adult zebrafish and administration of GH significantly increased IGFBP-3 expression.⁷ In addition, we also observed that administration of either IGF-I, IGF-II or insulin resulted in increased IGFBP-3 gene expression in fish. Although the biological significance of IGF-I-induced or IGF-II-induced IGFBP-3 gene expression in zebrafish awaits in-depth investigations, a phenomenon of an IGF-I-induced increase in IGFBP-3 expression was also observed in rhesus monkeys receiving acute or chronic subcutaneous administration of IGF-I.⁷ There is little evidence of the identity of a possible regulator of transcription factor binding in the fish IGFBP-3 promoter region. The human IGFBP-3 gene is located 20 kb downstream of the IGFBP-1 gene, in a tail-to-tail orientation,⁸ but its location in fish is unknown.

We have previously shown that administration of IGF-I, IGF-II or GH leads to significant increases in IGFBP-3 expression,⁶ suggesting that the reported IGF-I, IGF-II and GH stimulation occurs at the level of gene transcription. Many reports have shown IGF-I, IGF-II and GH stimulation in several cell types. Until now, cDNA corresponding to IGFBP-3 mRNA has only been described in zebrafish and tilapia (GenBank accession numbers AJ299410 and AF406954) among teleosts.^6,9 To allow further investigation of the regulation of zebrafish IGFBP-3 at the transcriptional level, we report isolation of the zebrafish IGFBP-3 gene and characterization of its promoter region. Several genomic clones containing the entire gene have been isolated from a zebrafish genomic DNA library using previously isolated zebrafish IGFBP-3 cDNA as a probe for screening. These genomic clones were characterized by sequencing. Approximately 5 kb of the promoter region of the gene was sequenced and several consensus sequences were found that may be important response elements for known hormonal regulators of zebrafish IGFBP-3. We report that the nucleotides between −132 and −1 in the 5′-flanking region are important for transcription of the IGFBP-3 gene in ZFL cells; that nucleotides between −1046 and −1 in the 5′-flanking regions are important to transcription of the IGFBP-3 gene in HeLa cells; and that the IGFBP-3 gene fragment between −1046 and −1 bp is sufficient to drive green fluorescent protein (GFP) gene expression in zebrafish embryos and during various developmental stages. As a part of our investigation to identify NLS-like sequences within zebrafish IGFBP-3 amino acid sequences that are necessary and sufficient for their nuclear import, we constructed different fragments including the first and second NLSs. Thus, nuclear import of IGFBP-3 appears to occur by means of the second NLS, which may play an important role in its nuclear function.

MATERIALS AND METHODS

Isolation of zebrafish IGFBP-3 genomic clones

Approximately one million recombinant bacteriophages from a zebrafish genomic library were screened with ³²P-labeled zebrafish IGFBP-3 cDNA fragments. The zebrafish genomic DNA library was constructed using the pPS vector system and purchased from MoBiTec Molecular Biotechnology (MoBiTec GmbH, Göttingen, Germany). The hybridization buffer used was 50% formamide containing NaDodSO₄ (7 g/100 mL), 0.5 M ethylenediaminetetraacetic acid (EDTA) (pH 8.0) (100 µL/100 mL), 50% PEG8000 (20 mL/100 mL), and 50% formamide (50 mL/100 mL) at 37°C for 16 h; after hybridization, filters were washed four times in 0.1X SSC and 0.1% NaDodSO₄ at 65°C. Positive plaques were used with BNN132 Escherichia coli for purification and restriction mapping of DNA. The zebrafish IGFBP-3 genomic DNA E. coli clones were sent to Genomics Biotechnology (Taipei, Taiwan) for sequencing. Nucleotides and the resulting translation sequences were aligned and compared using the PILEUP and GAP programs (Genetics Computer Group). To compare and align these sequences with sequences that have previously been published, we introduced gaps to maximize identity via procedures described by the genetics computer group gap and molecular evolutionary genetics analysis.¹⁰

Confirmation of the zebrafish IGFBP-3 transcription start point using rapid amplification of 5′cDNA ends

Total RNA was isolated from zebrafish liver using an RNA extraction method following the manufacturer's protocols (ULTRASPEC-II RNA isolation system; Biotecx Laboratories, Invitrogen Life Technologies, Carlsbad, CA, USA). The rapid amplification of 5′cDNA ends (5′RACE) procedure strictly followed the instruction manual (Invitrogen Life Technologies).¹¹ In brief, the sample consisted of approximately 5 μg of total RNA, whereas the primer for IGFBP-3 first-strand cDNA synthesis used the IGFBP-3 GSP1 primer (5′-GGAGCCGAAGGGCTGAACACGC). After the polymerase chain reaction (PCR), the 5′RACE products were analyzed on 1.5% agarose gels and transferred to nylon membranes (Bio-Rad, Hercules, CA, USA). The membranes were probed with [³²P] dCTP-radiolabeled IGFBP-3 cDNA. The hybridization procedure was the same as that used for the isolation and characterization of the zebrafish IGFBP-3 gene. After verifying which fragments hybridized with the probe, restriction endonuclease-digested 5′RACE products were cloned into the pBluescript plasmid vector (Strategene, La Jolla, CA, USA) and then sequenced to confirm the genomic DNA sequence.

Zebrafish IGFBP-3 reporter gene construction, cell culture and transient transfection assay

An approximately 20-kb SacI–KpnI fragment was obtained from a zebrafish IGFBP-3 genomic clone. The fragments for the promoter activity assay were amplified by PCR. After the PCR, the PCR product was digested with KpnI and BglII restriction endonucleases, and the fragment was ligated into the KpnI and BglII sites of the pGL-3 vector (Promega, Madison, WI, USA). The pGL-3 vector is a promoterless plasmid containing multiple cloning sites upstream of the luciferase cDNA. This resulted in the IGFBP-3 promoter activity assay obtaining nine construct plasmids (Fig. 1a). Clones of IGFBP-3 P1→P10, IGFBP-3 P2→P10, IGFBP-3 P3→P10, IGFBP-3 P4→P10, IGFBP-3 P5→P10, IGFBP-3 P6→P10, IGFBP-3 P7→P10, IGFBP-3 P8→P10 and IGFBP-3 P9→P10 were constructed and contained one transcription start site each; the transcription start site was connected immediately upstream of the ATG codon. The IGFBP-3 transcription start site was dependent on the 5′RACE result and assigned at −1 to be near the first nucleotide upstream of the ATG codon, with the definition of −1 referenced to Koval et al.¹²

Figure 1. Identification of promoter activity in the 5′-flanking region of the zebrafish insulin-like growth factor binding protein-3 (IGFBP-3) gene. Different fragments of the 5′-flanking region from −1046 to −27 bp relative to the transcription start site were cloned 5′ to the luciferase reporter gene, pGL3 basic. HeLa and ZFL cells were transfected with the various constructs and luciferase activity was determined. The data, which have been normalized for transfection efficiency against the secreted renilla luciferase activities using the Dual-Glo luciferase assay reagent (Promega), are expressed as a percentage of the activity of the pGL3 control vector. Data are presented as the mean ± standard deviation of at least three experiments. The basic promoter analysis is shown in (a) for non-IGF-treated experiments. Hormonal responsiveness of the zebrafish IGFBP-3 promoter region with luciferase constructs analyzed in ZFL cell lines (b) and HeLa cell lines (c) was for the IGF-treated experiments. After transfection, cells were incubated for 18 h in serum-free medium containing 1 μg/mL IGF-I or 1μg/mL IGF-II. Cell lysates were then assayed for luciferase activity. Data, which have been normalized for transfection efficiency against the secreted renilla luciferase activities of a cotransfected pRL TK vector, are presented as the mean ± standard deviation of at least three separate experiments.

To compare fish IGFBP-3 promoter activity between fish and mammalian systems, we chose fish cell lines and mammalian cell lines. The promoter analysis procedures followed those published previously in fish promoter assay papers.¹³ The ZFL cell line is a liver cell line that may respond to IGF stimuli and that regulates the IGFBP-3 promoter. It was previously reported that in vitro downregulation of growth factors was mediated by IGFBP-3 in cervical cancer. Herein, we chose the HeLa cell line to try to determine whether IGFs stimulate the HeLa cell line to regulate the IGFBP-3 promoter. HeLa, NIH3T3 and ZFL cells (ZFL cell line: ATCC number CRL-2643, zebrafish liver cell line) were grown in Dulbecco's Modified Eagle's medium (DMEM) (4.5 g/L glucose) and L15 medium containing 10% newborn calf serum. HeLa cells were grown in minimum essential medium with 2 mM l-glutamine and Earle's BSS adjusted to contain 90% 1.5 g/L sodium bicarbonate, 0.1 mM non-essential amino acids, 1.0 mM sodium pyruvate and 10% fetal bovine serum at 37°C. Approximately 1 × 10⁴ cells were seeded 12 h before transfection. Lipofectamine transfection was as described previously.¹⁴ Luciferase activity was measured 24 h after transfection and quantified as described previously.¹⁴ The supernatant was used to assay firefly and renilla luciferase activities using the Dual-Glo luciferase assay reagent (Promega). Luciferase activity was measured in a luminometer (Fluoroskan Ascent FL, Thermo Labsystems, Waltham, MA, USA). Plasmids of pEGFP-1 ligated with the IGFBP-3 promoter region were transfected into the same cells. After transfection, cells were treated with IGFs.¹⁵ The procedure was described periously,¹⁶ and the transfection method was the same as that for the luciferase assay. In brief, the IGF concentrations were 1000 ng/mL for all treatments. Values represent the mean ±  standard deviation (SD) of n experiments, where n represents the number of separate experiments for each promoter analysis. Statistical significance was set at P < 0.05 or P < 0.01 and was evaluated using the SAS statistical program (SAS Institute, Cary, NC, USA).

Reporter gene construction, preparation of plasmid DNA, fish egg collection and microinjection

The vector of pEGFP-1 (CLONTECH Laboratories, Mountain View, CA, USA; the entire pEGFP-1 vector sequence can be found in GenBankaccession number U55761) was constructed by inserting IGFBP-3 P1→P10 from the PCR fragment into the pEGFP-1 vector. Plasmid DNA was extracted according to the QIAGEN plasmid handbook for the plasmid maxi kit (QIAGEN GmbH and QIAGEN, Hilden, Germany). After purification, plasmids were precipitated using ethanol and sodium acetate. The DNA pellet was resuspended in ddH₂O and stored at −20°C. The experimental zebrafish were maintained under standard conditions with an artificial photoperiod of 11 h of darkness and 13 h of light to induce and maintain the reproductive cycle. Fertilized eggs were collected and injected with approximately 100 pg of supercoiled plasmid DNA into the cytoplasm of one cell of 2-cell-stage zebrafish embryos. The injection method and GFP assay procedure are described in Chen et al.¹⁶ All experiments were conducted in accordance with guidelines established by the Academia Sinica Committee on the Use, Care and Welfare of Animals at the Institute of Cellular and Organismic Biology, Academia Sinica.

Fluorescence microscopic observation and the reverse transcription polymerase chain reaction

After injection of the plasmid DNA into embryos, eggs were incubated in a 28°C incubator. Embryos were observed and photographed on an Olympus IX70 microscope (Olympus, Center Valley, PA, USA) and LEICA MPS 48/52 dissection microscope (Leica Microsystems, GmbH, Wetzlar, Germany) with a fluorescent filter using a long-wave ultraviolet lamp. The zebrafish embryo RNA extraction method was carried out according to the manufacturer's protocols (ULTRASPEC-II RNA isolation system; Biotecx Laboratories). For each stage, 20 embryos and 10 1-week-old fish were subjected to the IGFBP-3 promoter driven GFP activity assay. These embryos and fish were all injected with the pEGFP-1 vector ligated with the IGFBP-3 promoter region as described above. To investigate whether the expression of the IGFBP-3 promoter is stage-specific or tissue-specific, 1.6 μg of whole embryo or fish total RNA and the primers pEGFP-1 probe primer 5′ primer (5′-ATGGTGAGCAAGGGCGAGGAGC) and pEGFP-1 probe primer 3′ primer (5′-GCCGTCGTCCTTGAAGAAGATGG) were mixed together in a tube and used to amplify internal fragments of GFP first-strand cDNA, and the PCR reaction was run continuously to amplify approximately 317 bp of the GFP coding region according to the manufacturer's instructions (AMPLY RT/PCR KIT, LTK BioLaboratories, Taipei, Taiwan). The amplification program established for the Perkin Elmer 9700 thermocycler was as follows: 1 cycle of 5 min at 60°C and 30 min at 42°C; and 35 cycles of 1 min at 94°C, 2 min at 55°C, and 2 min at 72°C; followed by a final extension of 7 min at 72°C then immediate storage at 4°C. Each reverse transcription polymerase chain reaction (RT-PCR) included control groups: one tube containing only the pEGFP-1 probe primer 5′ primer and pEGFP-1 probe primer 3′ primer, and another tube containing embryo only or fish total RNA only mixtures. After the PCR, the PCR amplification products (5 μL of 20 μL) were analyzed on 1.5% agarose gels and transferred to nylon membranes (Bio-Rad). Membranes were probed with [³²P] dCTP-radiolabeled GFP cDNA fragments. The hybridation procedure was the same as that used for the isolation and characterization of tilapia IGF genes described above.

Enhanced green fluorescence protein reporter analysis of zebrafish IGFBP-3 subcellular distribution

We found two consensus NLSs in the zebrafish IGFBP-3 amino-acid sequences, one located at amino acid number 236 (bipartite: RRFRIPNCDQKGFYKKK) and the other at amino acid number 256 (PSKGRKR). The expression construct encoding different fragments of the N-terminal and C-terminal enhanced green fluorescence protein (EGFP)-tagged zebrafish IGFBP-3 was constructed using PCR cloning steps. The expression construct encoding the fusion protein between EGFP and the zebrafish IGFBP-3 deletion variant is shown in Figure 2a. For cell transfection, 2 × 10⁵ NIH3T3 cells were seeded into each well of a six-well culture plate 1 h before transfection was conducted. Five micrograms of either pEGFP, pEGFP-proPRR, pEGFP-PRR, pEGFP-ATGTGA or pEGFP-PSK was introduced into NIH3T3 cells using lipofectamine 2000 (Life Technologies). Fluorescence microscopy (Olympus) using a fluorescein isothiocyanate (FITC) filter and confocal microscopy were used to visualize the expression of the EGFP fusion proteins.

Figure 2. Enhanced green fluorescence protein (EGFP) reporter analysis of zebrafish insulin-like growth factor binding protein-3 (IGFBP-3) intracellular localization. Fluorescent images were taken directly from the transfected NIH3T3 cell line in culture dishes without prior processing. The figure shows five different constructs for analysis of nuclear localization signal (NLS) function. NIH3T3 cells transfected with the pEGFP-C1 vector alone (e) as a control and with pEGFP-PROPRR (with the pEGFP-N1 vector as a backbone ligated with no IGFBP-3 NLS sequences) (a) displayed cytosolic green fluorescent protein (GFP) fluorescence. NIH3T3 cells transfected with pEGFP-PRR (with the pEGFP-N1 vector as a backbone ligated with the first NLS sequence) displayed cytosolic GFP fluorescence (b), whereas those transfected with pEGFP-ATGTGA (with the pEGFP-N1 vector as a backbone ligated with two NLS sequences) (c) and pEGFP-PSK (with the pEGFP-C1 vector as a backbone ligated with the second NLS sequence) (d) predominantly expressed nuclear GFP fluorescence with very faint cytosolic fluorescence. The left pictures are fluorescent images; the right pictures are EGFP fluorescent and phase-contrast overlapping images; the middle pictures are phase-contrast images. These data show that the first NLS in the zebrafish IGFBP-3 sequence is not functional and that the second NLS sequence is functional.

RESULTS

Sequence analysis of the 5′-flanking region of the zebrafish IGFBP-3 gene

An approximately 5-kb genomic fragment corresponding to the 5′-flanking region of zebrafish IGFBP-3 was obtained after screening the genomic DNA library. The partial nucleotide sequence of the zebrafish IGFBP-3 proximal promoter sequence is given in Figure 3 and shows a sequence of approximately 1.3 kb 5′ upstream of the first methionine sequence. As in the human IGFBP-3 promoter sequence, there are a vitamin D response element,¹⁷ a TATA box and a GC upstream promoter element.¹⁷ In the rat IGFBP-3 gene, the transcription start site is 118 bp upstream of the initiation codon, and a TATA box consensus is located 27 bp 5′ to this CAP site; no CAAT box was present, but a CpG island was identified. Consensus sequences for a number of putative response elements (e.g. activating protein-2, insulin, thyroid stimulating hormone (TSH)/IGF and GH) were present within −700 bp of the CAP site.¹⁸ We compared the 5′-flanking region of the zebrafish IGFBP-3 gene against the TRANSFAC database to identify putative response elements that might mediate basal expression and IGF responsiveness of IGFBP-3 in HeLa or ZFL cell lines. In zebrafish, there is an undefined TATA-like sequence of 41 bp in front of the 5′RACE end. Between −1 and −1309 bp, there are three possible activating protein-1 (AP-1) binding sequences, three possible organic cation transporter-1 (Oct-1) binding sequences, one P300-binding sequence, one myeloid zinc finger-1 (MZF-1) binding sequence, one yin and yang-1 (YY-1) binding sequence, one hepatocyte nuclear factor-3 (HNF-3 beta) binding sequence, and one CCAAT enhancer binding protein (C/EBP beta) binding sequence.

Figure 3. Nucleotide sequences of the 5′-flanking region of the zebrafish insulin-like growth factor binding protein-3 (IGFBP-3) gene. The sequence is numbered relative to the transcription start site (+1). The region of the rapid amplification of 5′cDNA ends is marked by an arrow. Exon sequences and 5′-flanking sequences are shown in uppercase letters, and the intron sequences are shown in lowercase letters. The translation initiation codon begins from the first methonine. Each underlined segment represents putative transcription factor-binding sequences in the zebrafish IGFBP-3 5′-flanking region.

Determination of the zebrafish IGFBP-3 5′-cDNA terminus end

To ascertain the IGFBP-3 5′-cDNA terminus, a 5′ RACE method was carried out by sequencing 10 clones to identify the zebrafish IGFBP-3 5′-cDNA end. The 5′ RACE results identified one 5′-cDNA end in the zebrafish IGFBP-3 gene. The 5′-cDNA end in the zebrafish IGFBP-3 gene was compared to that of the bovine IGFBP-3 gene indicated by an arrow in Figure 3, and they were found to be the same. Bovine IGFBP-3 generated an 87 bp product, indicating that the mRNA cap site is 137 bp 5′ to the translation initiation code.¹⁹ The zebrafish IGFBP-3 5′ cDNA end is 185 bp long from the initiation site to the translation start site.

Demonstration of promoter activity in the zebrafish IGFBP-3 gene

To determine which regulatory regions of the zebrafish IGFBP-3 promoter are involved in regulating IGFBP-3, we constructed sequential 5′ deletions of the proximal promoter region, including the flanking transcription start sites described above (Figs 1 and 3). Constructs were then ligated with the luciferase coding sequence and transfected into HeLa and ZFL cells in triplicate, and the luciferase activity was analyzed (Fig. 1a,b). Data in Figure 1a reveal that the fragment of P7–P10 of IGFBP-3 contains a maximum value of enzymatic activity compared with the other IGFBP-3 promoter region constructs and the promoterless plasmid, pGL3-basic, in ZFL cells, whereas fragment P1–P10 of IGFBP-3 contains a maximum value of enzymatic activity compared with the other IGFBP-3 promoter region constructs and promoterless plasmid, pGL3-basic, in HeLa cells. Deletion of the fragment from P7–P10 to P9–P10 decreased the luciferase activity in both the HeLa and ZFL cell lines. By contrast, minimal luciferase activity was generated with fragments P8–P10 in the ZFL cell line and P9–P10 in HeLa cells, respectively. In a comparison of all fragments, the IGFBP-3 promoter showed that the sequence from P7 to P8 contains a positive regulatory element in ZFL cells. Surprisingly, with the fragment deletion to the number −82 bp (P8–P10) in ZFL cells and deletion to the number −27 bp (P9–P10) in HeLa cells, the luciferase activity represented only basal activity in this fragment, although the fragment from number −1 bp to −132 bp contains five possible transcription factor binding sequences, that is, two AP-1 sites, one MZF-1 site, one P300 site and one TATA site (Figs 1a and 3). To check the hormone responsiveness of the zebrafish IGFBP-3 promoters, expression vectors containing the luciferase reporter gene under the control of different promoter regions of the zebrafish IGFBP-3 gene were treated with IGF-I or IGF-II in the ZFL and HeLa cell lines (Fig. 1b,c). With fragment P4–P10, we observed an approximately 3.8-fold increase by IGF-I in comparison to IGF-II treatment in the ZFL cell line (Fig. 1b). In all six fragments of the promoter region, we observed that promoter activity in all fragments was increased by IGF-II treatment compared to IGF-I treatment in the HeLa cell line (Fig. 1c). In fragment P4–P10, promoter activity was approximately 2.0-fold higher with IGF-II treatment compared with IGF-I treatment (Fig. 1c). In all cases, incubation with IGF-I or IGF-II almost completely abolished the hormone effects in promoter fragment P9–P10 in both the ZFL and HeLa cell lines (Fig. 1b,c). However, the results also showed a 12.4-fold increase in fragment P5–P10 treated with IGF-II compared with fragment P9–P10 transfected into the ZFL cell line (Fig. 1b). The results described above show that the sequence of IGFBP-3 fragments from P1–P10 to P2–P10 contains a positive regulatory element after IGF-II treatment of ZFL and HeLa cells, and the region contains each of one possible binding element of either C/EBP beta or Oct-1. The IGFBP-3 promoter region of P3–P10 to P4–P10 may have an IGF-I negative regulatory element in ZFL cells, whereas that of P4–P10 to P5–P10 may have an IGF-II positive regulatory element in ZFL cells.

Taken together, these results suggest that the region located between nucleotides −809 and −1046 relative to the translation initiation codon contains sequences responsible for IGF-II-dependent activation in the HeLa cell line. A sequence analysis of the promoter region revealed the presence of putative C/EBP beta sites located upstream of the zebrafish IGFBP-3 promoter. Both cell lines showed promoter activity for fragment P5–P10 after IGF-II treatment. However, this region showed two putative AP-1, one putative TATA, one putative P300 and one putative MZF-1 transcription binding sequences. In addition, a number of potential Oct-1, GATA-2 and YY-1 binding sites may contribute to IGFBP-3 gene transcription after IGF-II stimulation in the ZFL cell line (Fig. 1b).

Expression of the IGFBP-3 promoter: GFP in cell lines and zebrafish embryos

To understand the expression timing of IGFBP-3 in fish, we constructed the fragment, IGFBP-3 P1–P10, ligated with the pEGFP-1 vector. This plasmid was transfected into the ZFL cell line and fluorescence in the entire cell was observed (Fig. 4). The data show that the zebrafish promoter can take advantage of the GFP system in cell lines to express the fish promoter. After microinjection of the plasmid with the pEGFP-1 and IGFBP-3 promoter region constructs, patches of fluorescent cells were observed in a zebrafish tail and eye (Fig. 5). The high expression of fluorescence in the tail and eye indicates that the IGFBP-3 promoter had high expression in the zebrafish straightening period. Fluorescence that appeared in the heart and blood cells indicated expression in the segmentation period prior to the straightening period (data not shown).

Figure 4. Transient transfection of ZFL cells with the insulin-like growth factor binding protein-3 (IGFBP-3) promoter fusion green fluorescent protein (GFP) coding region (IGF promoter ligated the pEGFP-1 vector) expressing GFP. The IGFBP-3 promoter used to ligate with pEGFP-1 vector was the fragment P1–P10. The arrow indicates the fluorescent color after the vector was transfected into ZFL cells.

Figure 5. Transient expression analysis of the insulin-like growth factor binding protein-3 (IGFBP-3) promoter in zebrafish embryos. The IGFBP-3 promoter green fluorescent protein (GFP) vector was microinjected into zebrafish 2-cell-stage fertilized eggs. After 24 h of development in fresh water on culture dishes, the GFP was visibly expressed in the tail and eye (indicated by the arrows). The bright-field photograph and fluorescence image are of the same 24-h injected embryo. All photographs were taken with a 45× lens on an Olympus IX70 microscope.

The vector-injected zebrafish embryos were also analyzed by RT-PCR using the primers, pEGFP-1 probe 5′ (5′-ATGGTGAGCAAGGGCGAGGAGC) and pEGFP-1 probe 3′ (5′-GCCGTCGTCCTTGAAGAAGATGG), followed by Southern blotting by a hybridization assay. The data show that after microinjection, the zebrafish IGFBP-3 promoter drove GFP cDNA as early as the 32-cell period, which remained high until 2 weeks, after which at 3 weeks, it was strongly reduced (Fig. 6).

Figure 6. Detection of insulin-like growth factor binding protein-3 (IGFBP-3) driving the green fluorescent protein (GFP) cDNA production of mRNA in each stage of zebrafish embryos and adult fish. Total RNA of whole embryos and juvenile zebrafish was extracted, then reverse transcription polymerase chain reaction (RT-PCR) and Southern blot analyses were conducted. The left picture shows ethidium bromide staining of the RT-PCR products using GFP-specific primers. The right picture shows a Southern blot analysis of the left picture in the gel, and after it was transferred to a membrane, then 32P-labeled GFP cDNA was used as a probe for hybridization. Lane M indicates the 100-bp marker. Lane 1, 32-cell stage; lane 2, 1-K-cell stage; lane 3, 30% epiboly stage; lane 4, gastrula stage; lane 5, 1-week-old fish; lane 6, 2-week-old fish; lane 7, 3-week-old fish. The transcripts were first detected in the 32-cell stage of zebrafish embryos.

Zebrafish IGFBP-3 has one nuclear localization signal

To demonstrate that the two putative NLS motifs in zebrafish IGFBP-3 are functional in cells, we analyzed fluorescently labeled IGFBP-3 containing the NLS motifs in the nucleus and cytoplasm. As described previously,⁶ the PSORT II program predicted two putative NLSs in the zebrafish IGFBP-3 protein sequence (Fig. 2a). As intranuclear localization has not previously been described for fish IGFBP-3 proteins, we were interested in determining whether zebrafish IGFBP-3 is indeed a nuclear protein. Erondu et al. reported the presence of IGFBP-3 in the nucleus in opossum kidney cells.¹⁹ This occurrence may be mediated by the importin-beta nuclear transport factor through a NLS-dependent pathway.⁴ To do so, an expression plasmid and a different zebrafish IGFBP-3 fusion protein were constructed and then introduced into NIH3T3 cells. Herein, we fusion the pre-propeptide cDNA sequences into pEGFP vector to mimic the natural state of IGFBP-3 function. If we only fusion the truncated peptide into pEGFP vector, the protein function may be not be like the original function for IGFBP-3. As shown in Figure 2b, NIH3T3 cells transfected with the pEGFP-C1 vector (pEGFP-C1; GenBank accession number U55763) and pEGFP-PROPRR vector displayed cytosolic EGFP fluorescence (Fig. 2b). The NIH3T3 cells transfected with the pEGFP-ATGTGA and pEGFP-PSK vectors predominantly expressed nuclear EGFP fluorescence with very faint cytosolic fluorescence. Clearly, zebrafish IGFBP-3 has one essential NLS. One NLS sequence was located in the second NLS sequence in the zebrafish IGFBP-3 cDNA sequence.

DISCUSSION

We sequenced and identified the IGFBP-3 promoter region in zebrafish. Promoter analysis revealed that the expression of zebrafish IGFBP-3 appears to be regulated by multiple regulatory elements in the promoter. Sequence analysis of the 5′-flanking region revealed a TATA box 41 bp upstream of the CAP site. No CAAT box was present, but two AP-1 sequence elements were located 5′ to the TATA box. To determine whether the putative promoter element is functional, a 1309-bp 5′-flanking segment with serial deletions was inserted upstream of the luciferase reporter gene and transfected into ZFL and HeLa cells. Results of the promoter assay indicated that these segments have significant serially dependent promoter activity as is expected for eukaryotic promoter elements. Transient transfection experiments with serially deleted constructs revealed that the promoter fragment containing 132 bp upstream of the CAP site showed maximal promoter activity compared to the pGL3 basic vector in ZFL cells, and a fragment extending only 89 bp upstream of the CAP site retained minimal promoter function in ZFL cells. The fragment containing only the TATA sequence showed minimal function in comparison with the pGL3 basic vector promoter. The second shortened sequence element required for promoter function in the zebrafish IGFBP-3 gene is similar to results obtained for bovines and humans.^19,20 However, in HeLa cells, promoter activity was shown to exist 1046 bp upstream of the CAP site, which showed the maximal promoter activity in comparison to the pGL3 basic vector in HeLa cells. Results differed from those of ZFL cells. The fragment of −809 to −1046 bp showed a C/EBP beta binding sequence, and the binding site may be functional in mammalian tumor cells, but was not functional in the zebrafish liver cell line. Comparison of P5–P10, P4–P10 the promoter activity was lower to P6–P10 in ZFL cell lines. The reason for this may be that a repressor protein binding site(s) exists in this region. In general, IGFBP-3 promoter activity in zebrafish liver cell lines is higher than that in mammalian tumor cell lines under no IGF treatment. This may result because different species were used.

From our previous data, we know that elevated levels of IGFBP-3 expression were observed 24 h after injection of IGF-I and 48 h after injection of either IGF-II or insulin. These results suggest that expression of the IGFBP-3 gene might be modulated by IGF-I, IGF-II and insulin.⁶ To understand the molecular mechanism for this transcriptional regulation, it is necessary to assay the promoter activity of zebrafish IGFBP-3 with hormonal treatment. The 5′-flanking region of the zebrafish IGFBP-3 gene does not contain the consensus hormone response element (Fig. 3), and the results are not similar to those published for bovines or humans.^19,20 Of particular interest are the AP-1, C/EBP beta and nuclear factor kappa B elements that are known to regulate IGF transcription activation.²¹ Our question was: can these factors possibly be involved in IGF regulation of IGFBP-3 mRNA transcription? No information is available for fish IGFBP-3 promoter regulation despite the data published for humans and rats, and even the human IGFBP-3 promoter has not successfully been demonstrated to have hormonal responsiveness.²² As a step to understanding the transcription binding elements that mediate IGF's stimulation of zebrafish IGFBP-3 mRNA synthesis in ZFL and HeLa cells, we produced a number of luciferase constructs containing 27–1046 bp of the zebrafish IGFBP-3 5′-flanking region. After transfection, ZFL and HeLa cells were exposed to serum-free media in the absence or presence of IGF-I or IGF-II. The promoter activities of the six constructs were tested by adding tilapia IGFs; this recombinant polypeptide was previously tested and the results have been published.^11,15 The data showed that the deletion fragment of −809 to −1046 bp after treatment with IGF-II had maximal promoter activity in ZFL and HeLa cells, and this implies that Oct-1 and C/EBP beta may be involved in the regulation of IGFBP-3 transcription by IGF-II. However, in transgenic mice with IGF-II, circulating IGF-I has been shown to be negatively correlated and IGFBP-2 to be positively correlated with IGF-II levels, suggesting that IGF-I is displaced from IGFBPs by IGF-II, and that IGF-II is a major regulator of IGFBP-2. Serum levels of IGFBP-3 and IGFBP-4 tended to be higher in phosphoenolpyruvate carboxykinase-IGF-II transgenic mice than in the controls, as evaluated by ligand blot analysis.²³

The data showed that fragment P4–P10 between −426 and −216 bp had higher promoter activities in ZFL cells after IGF-I treatment. After IGF-II treatment, fragment P4–P10 showed increased promoter activity in HeLa cells. In this fragment, we found that two Oct-1, 1 GATA-2 and one YY-1 binding elements existed in this region (Fig. 1b). YY-1 is believed to be a constitutive factor, present in all cell types, which affects the transcription of a large number of genes. Exogenously added purified IGF-I stimulates YY-1 expression, and the rapid responses of YY-1 expression to growth factor deprivation and replacement suggest the possibility that YY-1 may mediate some of the intranuclear responses to autologous stimulation by IGF-I.²⁴ YY-1 may be involved in IGFBP-3 promoter regulation after IGF-I treatment in zebrafish liver cells. The fragment of −625 to −426 bp showed a negative regulatory effect in ZFL cells after IGF-I treatment. It should be noted that the region contains the AP-1 and Nkx-2 binding elements. An AP-1-like motif is key to understanding IGF-I gene regulation.²¹ An AP-1 enhancer mediates phorbol ester 12-O-tetradecanoylphorbol-13-acetate (TPA) -induced transcriptional activation of the chicken IGF-I gene.²⁵ The IGF-I and IGF-II in this paper may have stimulated cells and induced AP-1 production to downregulate IGFBP-3 transcription regulation. These results explain how AP-1 may play an important role in regulation of IGFBP-3 expression. The zinc finger GATA proteins and Nkx-2.5 are coactivators of the GATA-4 protein in growth factors that induce expression of a number of cardiac-specific transcription factors. However, in our experiments, the promoter was tested in ZFL and HeLa cells, and it appeared that Nkx-2 had no function in these two kinds of cells as described previously. We inferred that the functional transcription factor binding element that exists in this region may be AP-1.

To integrate these reports, we can say that the IGFBP-3 promoter is regulated by a complicated pathway. It appears that the regulatory elements in the IGFBP-3 promoter have not been conserved during evolution, suggesting that zebrafish IGFBP-3 expression might be controlled by specific transcription factors. For detailed understanding of the IGFBP-3 promoter function, we used the GFP system as a reporter gene.²⁶ The GFP analyses presented herein indicated that GFP mRNA was detected using primers designed to specifically amplify GFP cDNA, and this was confirmed by the Southern blot analysis. In zebrafish, the early detection of IGFBP-3 mRNA expression in embryos is similar to that in fetal monkeys.^6,27 However, those reports could not confirm whether the IGFBP-3 promoter was present or if it was activated or not. In our GFP system, the IGFBP-3 proximal promoter drove GFP cDNA, which then produced the GFP; the fluorescence could be observed and detected by microscopy or PCR results. In other words, construction of GFP cDNA replaced the IGFBP-3 coding region of genomic DNA sequences. When the promoter began to drive the expression of GFP, the IGFBP-3 mature peptide may have been present in the natural state of the fish. From our GFP results, we know that the IGFBP-3 promoter was detected by PCR from the 32-cell stage. These results coincide with our previous findings.⁶ The green fluorescence was present in zebrafish tail, skin and eye. The results were similar to those using regulation of insulin-like growth factor-binding protein synthesis and secretion in human retinal pigment epithelial cells. Analyses of mRNA and protein localization of the IGF system components in regions with apoptosis during mouse development between 9.5 and 13.5 postcoital days revealed that IGFBP-3 was found in eye muscles, whiskers and somites, and that the expression of IGFBP-3 may stimulate apoptosis by trapping the IGFs.⁴ From our data, the IGFBP-3 promoter-driven GFP in tissues as described above should play an important role in the IGFBP-IGF axis as autocrine/paracrine factors. Determining whether the family of IGFBP-3 modulates IGF biological actions as both negative (inhibitory) and positive (potentiating) modulators will require further evaluation in the next generation of IGFBP-3 promoter–GFP transgenic fish. It is said that the IGFBP-3 proximal promoter may play an important role in controlling IGF expression in teleost embryos as presented by the RT-PCR data.

To identify which residues within the IGFBP-3 sequences are necessary for nuclear accumulation, we constructed different fragments with and without the NLS. The ability of the putative NLS regions within IGFBP-3 to target EGFP to the nucleus was examined by fusing these sequences to EGFP and expressing the resultant fusion protein in NIH3T3 cells. Our results suggest that one possible sequence of IGFBP-3 was imported into the nucleus. Nuclear protein importation directed by bipartite NLSs conventionally requires cytosolic factors, such as importin, Ran and nuclear transport factor 2.⁴ To date, it is known that the nuclear import of IGFBP-3 is mediated by the importin β subunit and GTP hydrolysis.⁴ However, a mechanism of IGF-independent IGFBP-3 action has been identified by cloning the nuclear retinoid X receptor alpha, which is a partner with IGFBP-3, in yeast two-hybrid experiments.⁴ How the promoter function relates to NLS is a question for further study. Therefore, zebrafish IGFBP-3 is indeed a nuclear protein with one NLS. The phenomenon of a dual role, both extracellular and intracellular, of growth factor-binding proteins can be extended not only to some growth factors and their receptors, but also to other substances with complicity in cellular reactions.

In conclusion, we have demonstrated in this study that IGFBP-3 expression in zebrafish is controlled by multiple regulatory elements. The regulatory mechanism controlling IGFBP-3 expression does not appear to have been conserved during evolution because regulatory elements are used in regulating IGFBP-3 expression in fish, but they have not been found in mice or humans. It is hoped that these results will provide important insights into the complex regulation of IGFBP-3 expression in a piscine and will enable us to elucidate the particular elements dictating the differential regulation of IGFBP-3 gene expression. Further study of fish IGFBP-3 may involve this gene-by-gene transfer to promote fish growth and apply the results in fisheries science.

ACKNOWLEDGMENTS

This work was supported in part by a grant from the Marine Research Station, Institute of Cellular and Organismic Biology, Academia Sinica, Ilan, Taiwan.