Hypoxia induces telomerase reverse transcriptase (TERT) gene expression in non-tumor fish tissues in vivo: the marine medaka (Oryzias melastigma) model

Background Current understanding on the relationships between hypoxia, hypoxia-inducible factor-1 (HIF-1) and telomerase reverse transcriptase (TERT) gene expression are largely based on in vitro studies in human cancer cells. Although several reports demonstrated HIF-1- mediated upregulation of the human TERT gene under hypoxia, conflicting findings have also been reported. Thus far, it remains uncertain whether these findings can be directly extrapolated to non-tumor tissues in other whole animal systems in vivo. While fish often encounter environmental hypoxia, the in vivo regulation of TERT by hypoxia in non-neoplastic tissues of fish remains virtually unknown. Results The adult marine medaka (Oryzias melastigma) was employed as a model fish in this study. We have cloned and characterized a 3261-bp full-length TERT cDNA, omTERT, which encodes a protein of 1086 amino acids. It contains all of the functional motifs that are conserved in other vertebrate TERTs. Motif E is the most highly conserved showing 90.9–100% overall identity among the fish TERTs and 63.6% overall identity among vertebrates. Analysis of the 5'-flanking sequence of the omTERT gene identified two HRE (hypoxia-responsive element; nt. – 283 and – 892) cores. Overexpression of the HIF-1α induced omTERT promoter activity as demonstrated using transient transfection assays. The omTERT gene is ubiquitously expressed in fish under normoxia, albeit at varying levels, where highest expression was observed in gonads and the lowest in liver. In vivo expression of omTERT was significantly upregulated in testis and liver in response to hypoxia (at 96 h and 48 h, respectively), where concomitant induction of the omHIF-1α and erythropoietin (omEpo) genes was also observed. In situ hybridization analysis showed that hypoxic induction of omTERT mRNA was clearly evident in hepatocytes in the caudal region of liver and in spermatogonia-containing cysts in testis. Conclusion This study demonstrates for the first time, hypoxic regulation of TERT expression in vivo in a whole fish system. Our findings support the notion that hypoxia upregulates omTERT expression via omHIF-1 in non-neoplastic fish liver and testis in vivo. Overall, the structure and regulation of the TERT gene is highly conserved in vertebrates from fish to human.


Background
The enzyme telomerase, comprising a telomerase reverse transcriptase (TERT) subunit and a RNA subunit (TR), is ubiquitous from unicellular protozoa to mammals [1,2]. In humans, transcriptional regulation of hTERT is the major mechanism of telomerase activation. Hypoxia responsive elements (HREs), to which the hypoxia-inducible factor 1 (HIF-1) binds to regulate transcription of HIF-responsive genes, have been detected in the hTERT promoter [3,4]. Current understanding on the mechanistic relationship between hypoxia, hypoxia-inducible factor-1 (HIF-1) and hTERT gene expression is largely based on in vitro studies in human cancer cells. For instance, HIF-1 has been shown to transactivate hTERT expression in human choriocarcinoma cells [4], cervical cancer cells [3] and ovarian carcinoma cells [5]. Hypoxia induction of hTERT expression and subsequent telomerase activity have been reported for solid tumor cells (i.e. ovarian carcinoma and colon adenocarnoma cells) [6]. The HIF-1α protein has been shown to correlate well with hTERT expression in human placentas [4]. Conversely, Koshiji et al. [7] demonstrated that HIF-1α downregulates hTERT expression in human colon cancer cells, presumably via the displacement of Myc binding from the hTERT promoter. These conflicting findings suggest multiple mechanistic relationships between hypoxia, HIF-1α and hTERT expression in human cancer cells. Yet, it remains virtually unknown whether these findings can be directly extrapolated to non-tumor tissues in other whole animal systems in vivo.
Fish often encounter hypoxia in the aquatic environment, which is a pressing environmental problem worldwide [8][9][10]. Gracey et al. [11] used cDNA microarrays to study the hypoxia-induced gene expression profiles of the euryoxic fish, Gillichthys mirabilis, and found that genes encoding glycolytic enzymes and those involved in protein synthesis are downregulated, which support the notion that metabolic depression is an important strategy for survival under hypoxia [12][13][14]. Interestingly, induction of genes involved in the suppression of cell proliferation and growth has also been observed which could serve as bioindicators of lowered growth during hypoxia. For example, insulin-like growth factor binding protein-1 (IGFBP-1) which regulates the availability of insulin-like growth factors in the circulation is involved in growth regulation. The role of HIF-1 in modulation of the IGFBP-1 gene and its involvement in hypoxia-induced growth and developmental retardation have recently been reported in zebrafish [15,16]. Although a detailed pattern of response to hypoxia in fish is emerging, it is not clear at what level of hypoxia these responses are initiated and how the various responses are related to each other. So far, we know of no reports on the effect of HIF-1 on TERT transcription in any piscine system. Given the important roles that TERT plays in apoptosis, DNA repair, cell proliferation and growth [17][18][19][20], a better understanding of the relationship between hypoxia and fish TERT expression will shed light on the in vivo regulation of fish TERT and its role in adaptive and toxicological responses in fish to hypoxic stress.
To investigate hypoxic regulation of fish TERT expression in vivo, we isolated and cloned the TERT gene (omTERT) from the marine medaka, Oryzias melastigma, and examined its in vivo organ expression patterns in whole fish under normoxia and in response to hypoxic stress. Analysis of omTERT expression by quantitative RT-PCR and in situ hybridization showed significant in vivo induction of omTERT expression in fish livers and testes. Gene transactivation studies using luciferase reporter constructs showed that ectopic expression of omHIF-1α enhanced omTERT promoter activity in vitro. Taken together, our findings support the notion that hypoxia upregulates omTERT expression via omHIF-1 in non-tumor fish tissues in vivo. The open reading frame (ORF) of omTERT encodes a polypeptide of 1086 amino acids, with a molecular weight (MW) of 124.5 kDa and a pI of 9.60. The physical properties of omTERT are very similar to that of the Fugu TERT (ORF = 1074 a.a., MW = 122 kDa, pI = 9.65; [21] Multiple sequence alignment indicated that omTERT contains all of the functional motifs that are highly conserved in TERTs from other vertebrate species ( Fig. 1; [22][23][24][25][26]). Moreover, sequence comparison of homologus domains from different TERTs indicated that motifs C (58.8%) and E (63.6%) are most highly conserved (Table 1). These two motifs in omTERT contain signature residues that distinguish telomerase RT from other RTs -the aromatic residue following the two aspartic acid residues in motif C and the tryptophan-X-glycine-X-serine/leucine residues in motif E ( Fig. 1; [27]). These signature sequences along with the TERT-specific T motif indicate that omTERT is a vertebrate TERT. Additionally, conserved amino acid residues known to be indispensable for telomerase activity [22,23,28] are also found in motifs GQ, 1, 2, B', C and E in omTERT (highlighted in bold in Fig. 1). Interestingly, the telomerase activity-determining aspartic acid in motif A of non-teleost TERTs is replaced by glutamic acid in all the fish TERTs. While negligible homology is present among the non-functional flexible linker region (the sequence stretch between regions v-I and v-II; [23]) in the TERTs examined, the length of this linker region in fish TERTs (O. melastigma contains 144 residues; grouper 153 residues; and Fugu 129 residues) are considerably shorter than those of other vertebrate TERTs (for example, frog contains 201 residues, chicken 301 residues, human 155 residues and mouse 160 residues).

In vivo expression and response pattern of omTERT to hypoxia
Real-time PCR revealed ubiquitous expression of omTERT mRNA in all of the 11 organs examined in both male and female adult fish (Fig. 2), where highest expression was observed in the fish gonads. Moderate omTERT expression was detected in gill, brain, kidney, gut and spleen followed by eye, heart, skin, muscle. omTERT was expressed only very weakly in liver.
In situ hybridization analysis showed that under normoxia, omTERT mRNA was only weakly expressed in the liver parenchyma (Fig. 4A). At 96 h hypoxia, omTERT expression was induced significantly in the caudal region of liver (Fig. 4B). In testis, omTERT expression was local- Multiple alignment of the deduced amino acid sequence of omTERT with selected vertebrate TERT homologues Similar to the TERT promoters of Fugu, chicken and human [21,26,29], the omTERT 5'-flanking sequence lacks a canonical TATA box but contains three putative CCAAT boxes near the transcription start site (TSS; at nt positions -10, -41 and -91; Fig. 5). In contrast to hTERT and cTERT, no CpG island is found in the omTERT promoter, which is notably AT-rich (60%). Comparison of the omTERT and fTERT 5'-flanking sequences using the Pro-moterWise program (EMBL-EBI) [30] failed to detect any homologous region of significant sequence identity. Using BLAST, the O. latipes TERT (olTERT) promoter sequence was identified in the medaka genome database [31] (scaffold 3719 at the NIG DNA Sequencing Center, Japan), and a 177-bp region that is conserved in both the omTERT and the O. latipes TERT (olTERT) promoters (ca. 75% identity) upstream of the TSS was identified. Signifi-cant sequence homology (69%) was also identified in the 5'-UTRs of the omTERT and olTERT genes.
Hypoxic induction of omTERT, omHIF-1α and omEpo expres-sions in liver and testis of marine medaka in vivo Figure 3 Hypoxic induction of omTERT, omHIF-1α and omEpo expressions in liver and testis of marine medaka in vivo. Male fish (n = 5) were exposed to either normoxic (6.4 mg O 2 L -1 ) or hypoxic (1.8 mg O 2 L -1 ) conditions for 24, 48 and 96 h. Expression of (A) omTERT, (B) omHIF-1α and (C) omEpo in the livers and testes was quantified using real-time RT-PCR and normalized with 18S rRNA. Data are presented as mean relative fold change ± SEM with respect to the gene expression level in the normoxic tissue sample at 24 h (arbitrarily set to 1). Asterisk (*) indicates significant difference from the other treatment group (p < 0.05).
In vivo expression of omTERT mRNA To identify potential transcription factor binding sites (TFBSs) in the omTERT promoter, the first 1000-bp sequence upstream of the TSS was analyzed using a filter string-based search supported by the TESS program [32]. Two putative HREs (hypoxia-responsive elements; at nt. -283 and -892) were identified which contain the exact HRE sequence (ACGTG) as the functional HRE described in the zfIGFBP-1 gene promoter of zebrafish [16]. However, a HIF-1 ancillary sequence (HAS) adjacent to the HRE, that affects hypoxia and HIF-1 responsiveness of the zfIGFBP-1 gene, is not present in the omTERT promoter. Moreover, a single putative HRE core (GCGTG) was also identified in the Fugu TERT promoter [Genbank: AY861384].
Other putative TFBSs of particular interest were also found: AP1 (nt.

Induction of omTERT promoter by omHIF-1α
To test the hypothesis that HIF-1α regulates omTERT transcription, the effect of omHIF-1α overexpression on luciferase activity driven by 3.5kb of omTERT promoter sequence was evaluated in HIF-1α-deficient CHO Ka13 cells. When cells were co-transfected with both pCMV-omHIF-1α and pTERT-luc, luciferase activity was significantly increased by ca. 2-fold when compared with cells transfected with pTERT-luc alone (p < 0.05) (Fig. 6).
In situ localization of omTERT mRNA in livers (A-C) and testes (D-F) of marine medaka

Discussion
The marine medaka, O. melastigma, is a hypoxia-tolerant species and can survive in a hypoxic environment (containing 0.8-1.8 mg O 2 /l) for 1-12 weeks with <10 % mortality (Au et al., unpublished observation). The hypoxia tolerance of O. melastigma was found to be comparable to that of other hypoxia-tolerant fish species such as the mudsucker, Gillichthys mirabilis [11] and the common carp, Cyprinus carpio [35]. In this study, we have cloned and characterized the expression pattern of the full-length telomerase reverse transcriptase (omTERT) cDNA from the marine medaka, O. melastigma. The deduced omTERT protein contains all of the functional motifs that are conserved in other vertebrate TERTs (Table 1), and shares particularly high sequence identity with the TERT of Fugu (61%) and moderate sequence identity (29-35%) with TERTs from other vertebrate species. Among the different motifs in TERT, motif E is the most highly conserved showing 90.9-100% overall identity among the fish TERTs and 63.6% overall identity among vertebrates (Table 1). Banik et al. [25] showed that the sequence spanning motif E to motif E-I is required for telomerase activity and curiously, analysis of the DNA sequences of the fish TERTs in this study showed that the lengths of motifs 1 and 2 (within the RT domain) are shorter by 3 and 8-11 residues, respectively, than other vertebrate TERTs. However, it is unclear what effect these changes may have on the RT activity of TERT. Overall, the findings of this study indicate that the structure and regulation of TERT is highly conserved in vertebrates from fish to human.
Real-time PCR analysis showed that omTERT is expressed ubiquitously, albeit at varying levels, in different organs of the adult marine medaka (Fig. 2). In particular, major expression of omTERT was detected in ovary and testis where germ cells are actively dividing and differentiating.
In situ hybridization analysis of testes demonstrated that high omTERT expression is restricted to highly proliferating sperm cells (i.e. spermatogonia and spermatocytes) and differentiating spermatids but not in mature spermatozoa (Fig. 4D). This is in agreement with the findings of Achi et al. [36], which indicated a progressive decline in telomerase activity in the rat testes during sperm cell development. In this context, it was not surprising that omTERT was also found to be highly expressed in regenerative tissues such as gill, gut and kidney. A similarly high level of omTERT expression that was observed in the medaka brain (Fig. 2) may be attributed to the presence of undifferentiated neuronal precursor cells as was demonstrated in the Xenopus brain [37]. Recent evidence indicate that TERT expression in the brain enhances cell survival of developing neurons [17,18]. In this study, a low level of omTERT expression was also detected in liver and muscle tissues of the marine medaka. This is in agreement with the presence of telomerase activity in the muscle and liver tissues of adult zebrafish [38], rainbow trout [39] and marine medaka (Au et al., unpublished data). Marine medaka exposed to prolonged hypoxia showed significant induction of omTERT expression in the testis (96 h) and liver (48 h), as determined by both quantitative RT-PCR and in situ hybridization techniques, that positively correlated with elevated expression levels of the omHIF-1α and omEpo genes (p < 0.05; Figs. 3B and 3C, respectively). Because of the unavailability of omHIF-1αspecific antibodies, detection of omHIF-1α protein by either Western blotting or immunohistochemistry was not possible, and expression of omHIF-1α and the HIF-1regulated omEpo gene was therefore used as an indicator of hypoxic induction of omHIF-1 in this study. To determine whether omTERT is responsive to the HIF transcription factor, two putative HRE core sequences (ACGTG) were identified in the omTERT promoter by computer analysis (Fig. 5), and gene transfection experiments in CHO cells indicated that the omTERT promoter is upregulated by omHIF-1α (Fig. 6). The results strongly suggest that omHIF-1α has a role in the regulation of omTERT expression. And based on the molecular responses of omTERT to hypoxia in the medaka liver and testis (Figs. 3 and 4), it is highly likely that upregulated expression of omTERT Effect of omHIF-1α overexpression on omTERT promoter activity under hypoxia is mediated (at least in part) via omHIF-1 in whole fish in vivo. The mechanistic relationships between omHIF-1α and omTERT (expression at the mRNA and protein levels), and the associated changes in omTERT telomerase activity in various fish organs under normoxia and hypoxia (and upon normoxic recovery) are currently being investigated in our laboratory to better understand the role(s) of these molecules in cell proliferation and growth.
The differential responses of hepatocytes (in caudal region of the liver) and testicular cells (in spermatogonia) in the marine medaka to hypoxic stress indicate that hypoxic induction of TERT is cell-type specific. Anti-apoptotic and protective functions of TERT against hypoxia-induced apoptosis and DNA damage have been demonstrated in a variety of mammalian cells (see review by Chung et al. [20]). It is tempting therefore to speculate that the increase in omTERT expression in medaka liver and testis under hypoxia could be a protective response to enhance survival of specific cell types. For instance, the role of TERT in maintaining the genomic stability of male germ cells is particularly critical during spermatogenesis to ensure the transfer of intact chromosomes to the offspring [36]. Additionally, the presence of a putative E-box (to which the c-Myc/Max heterodimer could potentially bind) and c-Myb binding sites in the omTERT promoter suggests that omTERT activation may aid in promoting cell proliferation and differentiation [40,41] by counteracting the potentially adverse effects of hypoxia on fish tissues. The molecular mechanisms operating through TERT in regulating cellular homeostasis and gametogenesis in fish certainly warrant further investigation. Functional studies and exposure experiments are currently underway in our laboratory to investigate the significance of the estrogen-(EREs) and dioxin-responsive (DREs) elements in the omTERT promoter on in vivo gene expression in the marine medaka. Conceivably, if omTERT expression is responsive to challenges by various estrogenic and/or dioxin-related pollutants, and the responses could be linked to Darwinian fitness traits (such as cell proliferation, growth and reproduction) in the marine medaka, this will open the possibility of TERT being used as a novel and meaningful molecular biomarker of effects for monitoring environmental stress and pollution in the marine environment [42].

Conclusion
We report here the isolation and characterization of the omTERT gene from O. melastigma, a potential marine model fish for molecular ecotoxicology. Our findings suggest that hypoxic induction of omTERT expression in nontumor liver and testis in whole fish system may be mediated, in part via the omHIF-1 transcription factor. Data from the marine medaka model suggest that the structure and regulation of TERT gene is conserved in vertebrates from fish to human.

Marine medaka
The Dissolved oxygen was monitored continuously using dissolved oxygen meters and polarographic probes (Cole-Parmer 5643-00, Illinois, USA). After the exposure period, adult fish (n = 5 for each sex) were sampled and tissues were immediately dissected out, snap-frozen in liquid nitrogen and stored at -80°C until ready to be processed.

Real-time PCR
Real-time PCR was performed on the iCycler iQ Real-time PCR System (BioRad) using the SYBR Green I dye-based detection method. Similar protocols were also employed to quantify mRNA expression for two hypoxia marker genes, omHIF-1α and the erythropoietin gene (omEpo). First-strand cDNA was diluted to 1/10 and 5 μL were used for each real-time PCR reaction. Triplicates were run for PCR reaction. To normalize the target gene expression for differences in cDNA input, cDNA was diluted 1:2500 for measuring 18S rRNA levels. Diluted cDNAs (5 μL aliquots) were added to the wells of a 96-well thin-wall PCR plate, and to each was added 20 μL PCR master mix containing 12.5 μL 2 × iQ SYBR Green Supermix (BioRad), 0.5 μL of each target gene primer (10 μM) and 6.5 μL water or 1 μL of each 18S rRNA primer (10 μM) and 5.5 μL water. PCR primers were designed using the Primer3 program [44] and their sequences are shown in Table 2. The PCR plate was heated at 95°C for 2 min followed by 40 cycles of 95°C for 20 s, 60°C for 30 s and 72°C for 30 s. For quantification of PCR results, C T (the cycle at which the fluorescence signal is significantly different from background) was determined for each reaction. From a standard curve showing the relationship between the quantity of starting material (a dilution series of fish testis cDNA) and C T , PCR efficiency of the omTERT expression assay was determined to be 98.6% (R 2 = 0.999). Melting-curve analysis was performed by running a heat-dissociation protocol after the PCR to differentiate between the desired target amplicons and any primer-dimers or DNA contaminants. The size and identity of each gene amplicon were verified by agarose gel-electrophoresis and DNA sequencing, respectively. The C T values for both reference and target genes, and the efficiencies (E = 10-1/slope ) of both assays were substituted into Equation (1) to yield the mean normalized expression (MNE) value.
To calculate the relative fold-change in target gene expression, the MNE of the experimental sample was divided by the MNE of the control (Equation 2).

Gene
Sense primer Antisense primer Amplicon size (bp) Genbank accession number diluted adaptor-ligated genomic library (for primary PCR) or 1:50 diluted primary PCR product (for secondary PCR) and 1 μL 50 × BD Advantage 2 polymerase mix (BD Biosciences). The amplification profile used was the same as that in RACE-PCR.