Benzo[a]pyrene effects on glycine N-methyltransferase mRNA expression and enzyme activity in Fundulus heteroclitus embryos

Abstract

Benzo[a]pyrene (BaP) is a ubiquitous environmental polycyclic aromatic hydrocarbon (PAH) contaminant that is both a carcinogen and a developmental toxicant. We hypothesize that some of BaP’s developmental toxicity may be mediated by effects on glycine N-methyltransferase (GNMT). GNMT is a mediator in the methionine and folate cycles, and the homotetrameric form enzymatically transfers a methyl group from S-adenosylmethionine (SAM) to glycine forming S-adenosylhomocysteine (SAH) and sarcosine. SAM homeostasis, as regulated by GNMT, is critically involved in regulation of DNA methylation, and altered GNMT expression is associated with liver pathologies. The homodimeric form of GNMT has been suggested as the 4S PAH-binding protein. To further study BaP-GNMT interactions, Fundulus heteroclitus embryos were exposed to waterborne BaP at 10 and 100 μg/L and both GNMT mRNA expression and enzyme activity were determined. Whole mount in situ hybridization showed GNMT mRNA expression was increased by BaP in the liver region of 7, 10 and 14 dpf F. heteroclitus embryos. In contrast to mRNA induction, in vivo BaP exposure decreased GNMT enzyme activity in 4, 10 and 14 dpf embryos. However, in vitro incubations of adult F. heteroclitus liver cytosol with BaP did not cause decreased enzyme activity. In conclusion, BaP exposure altered GNMT expression, which may represent a new target pathway for BaP-mediated embryonic toxicities and DNA methylation changes.

1. Introduction

Benzo[a]pyrene (BaP), a model polycyclic aromatic hydrocarbon (PAH), is a known animal carcinogen and a probable human carcinogen by IARC’s classification. The most widely recognized toxicity associated with PAHs is their mutagenicity and carcinogenicity in mammals, including humans (Mastrangelo et al. 1996) though their embryo and developmental toxicities are becoming more recognized (Billiard et al. 2008; Chen et al. 2005; Pershagen 1989; Tang et al. 2006; Tang et al. 2008)

In the cytoplasm, BaP binds to both 8S and 4S PAH binding proteins (Raha et al. 1990). The 8S PAH binding protein is more commonly known as the aryl hydrocarbon receptor (AhR). The BaP-AhR complex translocates to the nucleus, heterodimerizes with ARNT and then transactivates genes, including CYP1A1, CYP1A2, CYP1B1 and some Phase II enzymes. BaP is bio-activated by the CYP1 family of P450 monooxygenases and/or other enzymes such as prostaglandin H synthase to the toxic metabolite BaP-7, 8-dihydrodiol 9, 10-epoxide (BPDE), which targets the genomic DNA by forming adducts. Besides genotoxic effects, BaP disrupts DNA methylation and thus also has epigenetic mechanisms of toxicity. The binding of BPDE to DNA affects the activities of DNA methyltransferase (Subach et al. 2006), and BaP causes sequence specific hypo- and hypermethylation in multiple breast cancer cell lines (Sadikovic et al. 2004). Currently, the identity and function of the 4S PAH binding protein is still under debate (Ogawa et al. 1997), however, it has been suggested to be glycine N-methyltransferase (GNMT; EC2.1.1.20) (Raha et al. 1994).

GNMT is the enzyme responsible for the transfer of a methyl group from S-adenosylmethionine (SAM) to glycine forming S-adenosylhomocysteine (SAH) and sarcosine, preserving the ratio of SAM/SAH, which is a sensitive indicator of the methylation capacity. GNMT makes up 1% or more of cytosolic proteins in the liver (Luka et al. 2009) and is a critical mediator in the methionine and folate cycles (Mato and Lu 2007). In mammals, the homotetrameric form of GNMT is enzymatically active and is the major folate binding protein in the liver (Cook and Wagner 1984). The homodimeric form has been suggested to be the 4S PAH binding protein (Bhat et al. 1997).

GNMT deficiency in humans is associated with various liver pathologies such as hypermethioninemia and hepatomegaly (Luka and Wagner 2003; Mudd et al. 2001). GNMT has been suggested to be a tumor suppressor gene and GNMT expression was down-regulated in hepatocellular carcinoma (HCC) (Tseng et al. 2003), cholangiocarcinoma (Huang et al. 2008) and prostate cancer tissues (Huang et al. 2007). Additionally, the higher expression of GNMT in females compared to males is proposed to be a possible mechanism explaining the much lower incidence of HCC in females than males (Liao et al. 2009). GNMT knockout mice spontaneously develop chronic hepatitis, glycogen storage disease (Liu et al. 2007), steatosis and HCC (Martinez-Chantar et al. 2008).

Although the controversy of GNMT and 4S PAH binding protein is not resolved, recent studies continue to find interactions between GNMT and BaP. Docking studies suggested a preference for BaP binding to the dimeric form of GNMT (Chen et al. 2004). In the same study, Chen et al. found that BaP exposure of GNMT transfected cells had decreased BPDE adduct formation and CYP1A1 enzyme activity. GNMT may sequester BaP, in turn, diminishing BaP’s effects on the liver detoxification pathways and preventing cytotoxicity in HepG2 cells (Lee et al. 2006b). Recently, a functional xenobiotic response element (XRE) was identified in the 5′ regulatory region of human GNMT and was inducible by BaP treatment in vitro (Lee et al. 2009). Therefore, it was suggested that GNMT activation may be involved in cellular defense mechanisms following BaP exposure.

Waterborne PAH exposure is capable of producing embryo toxicities and HCC in F. heteroclitus (Wang et al. submitted), a teleost that is exposed to environmental contaminants including PAHs and a highly recognized environmental toxicological model organism (Burnett et al. 2007). We have previously reported cloning of F. heteroclitus GNMT cDNA and the mRNA expression profile during embryogenesis measured by quantitative real-time PCR. GNMT mRNA was inducible in F. heteroclitus embryos by BaP exposure in vivo (Fang et al. 2009). In this study, we further characterized the cell/tissue-specific expression of GNMT mRNA in F. heteroclitus developing oocytes and embryos and then further investigated the BaP effects on embryonic GNMT mRNA expression and enzyme activity.

2. Methods

2.1 Fish source, care and handling

A parental population of F. heteroclitus was collected from an uncontaminated site at the Newport River near Beaufort Inlet, NC, USA and raised under the University Institutional Animal Care and Use Committee approved conditions. Sexually mature fish were bred and kept in salt water (20–25 parts per thousand (ppt), Instant Ocean, Cincinnati, OH). The fish were maintained at 20–25°C, 14:10 light-dark cycle. Adult fish were fed twice daily with tropical flake fish food (Tetramin, Tetra Werke, Germany) and live brine shrimp (Salt Creek, Inc., Salt Lake City, UT). First generation offspring, from wild parents, were used for the studies described here.

2.2 Embryo BaP exposure

Eggs and sperm were stripped from the parental F. heteroclitus population and mixed together for in vitro fertilization. Fertilized eggs were randomly sorted into three treatment groups, namely control dimethylsulfoxide (DMSO, 1 μl/mL), 10 and 100 μg/L BaP (Sigma-Aldrich, St. Louis, MO, stock solution 0.1 mg/mL in DMSO; final DMSO concentration was 1 μl/mL in all treatment groups). The pure water solubility of BaP ranges between 2.3 (Agency for Toxic Substances and Disease Registry (ATSDR) 1995) and 4 μg/L (Mackay and Shiu 1977). BaP is highly lipophilic with an octanol-water partition coefficient (K_ow) of 2.2 × 10⁶ (Chiao et al. 1995). Therefore, the higher concentrations of BaP exposure could be considered to mimic either the maternal transfer and/or bioaccumulation of organic chemical to the lipid-rich ova (Kadokami et al. 2004). Ten eggs were pooled randomly and raised in 10 ml of salt water (~21 ppt). Exposures for each experimental sample began at approximately 4.5 hour post fertilization (hpf). Eggs were inspected daily for normal development. Water was changed and eggs were re-dosed every other day. For embryos prepared for in situ hybridization (ISH), 1-phenyl-2-thiourea (PTU, 0.003%, Sigma-Aldrich) was added to the culture water from 5 dpf until the end of the exposure to lessen pigmentation. Ten embryos per pool per time point were collected at 4, 7, 10 and 14 days post fertilization (dpf) (embryos would typically hatch at ~17 dpf). The 4 dpf, F. heteroclitus embryo is under active organodifferentiation, with defined heart chambers and a developing foregut. By 7 dpf, gut and liver are visible and the bilobed urinary bladder is formed. At 10 dpf, the mouth can open, and by 14 dpf, the swimming bladder has inflated (Armstrong and Child 1965).

For ISH, F. heteroclitus embryos were fixed in 4% (w/v) paraformadehyde (PFA) in phosphate buffered saline (PBS, pH 7.4) overnight and stored at −20°C freezer after removal of the chorion membrane by watchmaker’s forceps. For GNMT enzyme activity assays, embryos were collected and frozen at −80°C in 1 ml HEGD buffer (HEPES 10 mM, EDTA 1.5 mM, glycerol 10% v/v, DTT 1 mM, PMSF 0.5 mM and pH 8.0) until further analysis.

2.3 Whole mount in situ hybridization

Whole mount ISH was carried out as described elsewhere (Dong et al. 2002). The fragment of GNMT cDNA (Genbank Accession # FJ607947) used as template for RNA probe synthesis was generated by PCR amplification from F. heteroclitus liver cDNA, using primers: forward: 5′-CTGAAGTACGCGCTGAAG-3′; reverse: 5′-CAAAGGAGACCAGAGCGA-3′. The cDNA fragment contained 188 bp, from 277 to 464 bp (counted from translation start site) in the F. heteroclitus GNMT cDNA sequence, and was highly specific for GNMT according to a BLAST homology search. The fragment spanned across the putative GNMT exons 2, 3 and 4 to avoid genomic DNA binding (exon borders were determined by comparing splicing sites among F. heteroclitus, Danio rerio (#NM_212816.1) and Homo sapiens (#NM_01896) GNMT sequences). The PCR product was purified and inserted to pGEM-T^® Easy Vector (Promega, Madison, MI). Sense and antisense digoxigenin ssRNA probes were synthesized by in vitro transcription using T7 or SP6 RNA polymerase on GNMT plasmid linearized with SpeI or ApaI (New England Biolabs, Ipswich, MA). Following hybridization with probes overnight at 64°C, embryos were washed with 2x saline sodium citrate buffer (SSC, 0.15 M sodium chloride and 15 mM trisodium citrate, pH 7.0) and 0.2x SSC twice for 30 min, respectively. After blocking with 2% blocking reagent (Roche, Indianapolis, IN), embryos were incubated overnight with 4000x diluted anti-DIG antibody conjugated with alkaline phosphatase (Roche) at 4°C. The color reaction was carried out by incubation with BM-purple substrate (Roche). Yolks were carefully removed by watchmaker’s forceps. GNMT expression was recorded with a digital camera (Optronics, Muskogee, OK) and dissecting microscope (Carl Zeiss, Thornwood, NY). Stained regions were outlined as regions of interest (ROIs) and the mean intensity was calculated (=sum intensity divided by ROI pixel area), with the use of KODAK 1D Image Analysis Software. For 7 and 14 dpf ISH, 9–17 embryos were collected for each treatment group; for 10 dpf ISH, 21–28 embryos were collected per group.

2.4 Paraffin embedding, sectioning and H&E staining

Adult F. heteroclitus ovary and 14 dpf BaP-exposed F. heteroclitus embryos were fixed in 4% (w/v) PFA in PBS (pH 7.4) overnight, followed by dehydration in increasing gradients of ethanol (70%, 80%, 90%, 95%, and 100% twice). After cleared in Clearify^™ (American Master Tech Scientific, Lodi, CA), tissues were embedded in molten paraffin (Paraplast embedding media paraplast X-tra, Sigma). Sections of 5 μM thickness were prepared using a microtome (OlympusCut4055, Olympus American, San Jose, CA). For ovary, three adjacent sections were placed on three different slides with an “A” set for hematoxilin and eosin (H&E) staining, a “B” set for GNMT sense section ISH and a “C” set for GNMT anti-sense probe section ISH. For whole mount ISH processed embryos, two adjacent sections were placed on two different slides with an “A” set for eosin contrast staining, a “B” set for observation of the GNMT whole mount ISH signals only.

The sections were cleared with Clearify^™ and a gradient of ethanol (100–70%). The slides were treated as indicated above and covered with cover-slips and observed under a digital camera and light microscope (Olympus America, Center Valley, PA).

2.5 Section in situ hybridization

Section ISH was carried out as described elsewhere (Dong and Willett 2008). The ovary sections were cleared in gradient of ethanol (100–70%) and hybridized with the sense and antisense F. heteroclitus GNMT probes as described above. Following hybridization overnight at 50°C, sections were washed with 2x SSC and 0.2x SSC twice for 20 min, respectively. After blocking with 2% blocking reagent (Roche), embryos were incubated overnight with 4000x diluted anti-DIG antibody conjugated with alkaline phosphatase (Roche) at 4°C. The color reaction was carried out by incubation with BM-purple substrate (Roche). The slides were covered with cover-slips and observed under a digital camera and light microscope (Olympus America, Center Valley, PA).

2.6 GNMT enzyme activity assay

Embryos (10 eggs/pool) or adult livers were homogenized in HEGD buffer. Homogenates were centrifuged at 10,000 × g for 1 hr and the resulted supernatants were the embryo or liver cytosolic preparations. Adult liver cytosol was prepared for method optimization and F. heteroclitus GNMT enzyme kinetics studies. Protein concentrations were measured by Bradford using BSA as a standard. The GNMT enzymatic activity was determined using the method of Cook and Wagner (Cook and Wagner 1984) with some modifications, and performed in duplicate. The GNMT assay mixture (100 μl) contained 0.1 M Tris HCl (pH 7.4), 5 mM DTT, 10 mM glycine (Fisher Scientific, Pittsburgh, PA), 1 mM S-adenosyl-L-[methyl-³H] methionine (0.02 μCi/reaction, PerkinElmer Life Sciences, Waltham, MA) + cold SAM (Sigma-Aldrich), and 180–250 μg sample protein and was incubated at 25°C for 30–60 min. The reaction was terminated with 50 μl of 5% trichloroacetic acid, and 250 μl charcoal (38 mg/ml in 0.1 M acetic acid) was added to absorb the unreacted SAM. The tubes were then incubated at 4°C for 15 min and centrifuged for 3 min (14,000 × g). Supernatant (200 μl) was added to 5 ml of aqueous counting scintillant (Research Products International Corp., Mt. Prospect, IL) and assayed in a Packard Liquid Scintillation Analyzer (GMI, Inc., Ramsey, MN). Assay mixtures without protein served as negative controls and the values obtained were subtracted to give net counts. For method optimization, incubation time was set from 15–90 min and protein amount was loaded from 50–600 μg. To study enzyme kinetic parameters (Vmax, Km), 10 mM glycine was used while SAM concentrations varied from 0.02–3 mM or 1 mM SAM was used while glycine concentrations varied from 0.5–20 mM.

2.7 In vitro BaP effects on GNMT enzyme activity

Samples of adult liver cytosol were incubated with BaP to determine BaP’s ability to inhibit GNMT in vitro. The protein samples (250 μg) were mixed with 1 μL of 1 mM, 5 mM or 10 mM BaP (in DMSO, final concentrations were 10 μM, 50 μM and 100 μM, respectively), 1 μL DMSO or water and incubated at 4°C for 1 hr prior to performing the GNMT activity assay as described above except pH was 9.0. Each treatment group had five samples.

2.8 Statistics

Results of ISH and embryonic enzyme activity were analyzed by GraphPad Prism 5.0 (La Jolla, CA) and presented as mean ± S.E. Statistical differences between treatment groups were determined using a one-way ANOVA followed by Neuman–Keulls post hoc test (p < 0.05). Data of the enzyme kinetic study was analyzed using nonlinear regression analyses (Michaelis-Menten equation).

3. Results

3.1 Expression of GNMT mRNA in F. heteroclitus oocytes

Developing oocytes were divided into stage I (previtellogenic up to 0.6 mm diameter), stage II (vitellogenesis 0.6–1.3 mm), stage III (maturation 1.3–1.9 mm with slight protein accumulation) and stage IV (maturation without protein accumulation) as defined by (Wallace and Selman 1985). We have further divided the stage I follicles into three size based categories: early stage I (stage Ia, 0–0.15 mm), mid stage I (stage Ib, 0.15–0.25 mm), and late stage I (stage Ic, 0.25–0.6 mm). Strongest expression of GNMT mRNA was detected in the ooplasm of stage Ia oocytes (Figure 1C and F). The signal gradually declined in the ooplasm of stage Ib and Ic oocytes (Figure 1I and L) and became barely detectable in later stages (data not shown).

Figure 1.

Expression of GNMT mRNA in adult oocytes. Adult ovaries were sectioned and visualized by H&E staining (A, D, G and J) or in situ hybridization using GNMT sense probe (B, E, H and K) or antisense probe (C, F, I and L). Strong expression of GNMT mRNA was detected in the ooplasm of stage Ia oocytes (primary growth stage, C and F). Signal gradually decreased and became barely detectable by stage Ib and Ic oocytes (previtellogeic follicle stage, I and L, respectively). Oocytes were characterized by the presence of yolk granules in ooplasm (Wallace and Selman 1985) and size.

3.2 Effects of BaP on GNMT mRNA expression in F. heteroclitus embryos

At 7 dpf, strong GNMT expression was detected in the liver of F. heteroclitus embryos by whole mount ISH (Figure 2). In addition to liver, GNMT was also expressed in the urinary bladder in 10 and 14 dpf embryos. The staining intensity in the liver area was quantitated, and BaP increased GNMT mRNA transcription in the liver of 7, 10 and 14 dpf embryos (Figure 3). Compared to controls, BaP at 10 and 100 μg/L induced liver GNMT mRNA expression by 20% and 43% in 7dpf F. heteroclitus embryos, and by 10% and 9% in 10 dpf embryos, and by 20% and 32% in 14 dpf embryos. Paraffin sectioning more clearly showed the distribution of GNMT expression in liver and urinary bladder in 14 dpf BaP-exposed embryos (Figure 4).

Figure 2.

BaP induced GNMT mRNA expression in the liver of developing F. heteroclitus embryos measured by whole mount in situ hybridization. Embryos were exposed to DMSO, or BaP (10 or 100 μg/L, final DMSO concentration was 1 μl/mL in all treatment groups) from 4.5 hour post fertilization (hpf) to 7, 10 or 14 days post fertilization (dpf). From 5 dpf to the end of exposure, PTU was added to reduce pigmentation. Arrows (black) show GNMT expression signals (in purple color) in the livers of 7, 10 and 14 dpf F. heteroclitus. Arrow heads (red) show GNMT signals in the urinary bladders of 10 and 14 dpf F. heteroclitus. Panel A, E and I indicate nonspecific binding in control (DMSO treated) embryos by using GNMT sense probe. Chorion and yolk were removed from the embryo during the experimental procedure.

Figure 3.

Intensity of GNMT expression in the liver region of 7, 10 and 14 dpf F. heteroclitus embryos by whole-mount in situ hybridization (ISH). Expressed regions were analyzed by means of KODAK 1D Image Analysis Software by creating regions of interest. * Indicates significant increase of GNMT expression between the BaP treated and control groups (ANOVA, * p<0.05, ** p<0.01). For 7 and 14 dpf ISH, each bar represents mean ± SE of 9–17 embryos. For 10 dpf ISH, each bar represents mean ± SE of 21–28 embryos.

Figure 4.

Paraffin sections of 14 dpf BaP-exposed embryos stained by GNMT whole mount in situ hybridization (purple). Embryos were exposed to 100 μg/L BaP from 4.5 hpf to 14 days. A is the longitudinal section of the whole fish embryo; B is the eosin contrast staining (red) of the adjacent slide of A; C and D show strong GNMT signals in the liver and urinary bladder, respectively. E and F are magnification of the rectangular region of C and D, respectively.

3.3 F. heteroclitus GNMT enzyme kinetics

F. heteroclitus liver cytosol displayed Michaelis-Menten kinetics for both S-adenosylmethionine (SAM) and glycine substrates when one of the substrates was present in limited concentrations and the other substrate was in excess. The extrapolated Km values for SAM and glycine were 0.18 ± 0.03 and 3.24 ± 0.53 mM, respectively (Figure 5). Estimated Vmax for SAM and glycine were 18.80 ± 0.78 and 22.29 ± 1.17 pmol/mg*min, respectively. As indicated in Figure 5, the curves reached plateau activities with 1 mM SAM and 10 mM glycine, suggesting the saturation concentrations for the enzyme. Therefore, these concentrations were selected for GNMT activity analyses. GNMT activities were linear with reaction time (15 – 90 min) and protein amount (50 – 600 μg) (data not shown).

Figure 5.

Determination of GNMT kinetics with F. heteroclitus liver cytosol. Liver cytosol (250 μg) was used and reactions were performed at 25°C for 30 min. (A) Km (SAM) was 0.18 ± 0.03 mM (R² = 0.92) and Vmax was 18.80 ± 0.78 pmol/mg*min. Reactions were incubated with 10 mM glycine when the SAM concentration was varied from 0.02–3 mM. (B) Km (glycine) was 3.24 ± 0.53 mM (R² is 0.91) and Vmax was 22.29 ± 1.17 pmol/mg*min. Reactions were incubated with 1 mM SAM when the glycine concentration was varied from 0.5–20 mM.

3.4 BaP-mediated inhibition of GNMT enzyme activity

GNMT enzyme activity exists in unfertilized eggs. Constitutive GNMT activity was significantly higher (about two fold) in 4 dpf embryos compared to all other time points tested (Figure 6). Waterborne BaP exposure decreased GNMT enzyme activity in 4, 10 and 14 dpf embryos. BaP exposure of 10 and 100 μg/L inhibited embryonic GNMT enzyme activity by 33% and 36% in 4 dpf F. heteroclitus embryos, by 33% and 55% in 10 dpf embryos, and by 3% and 23% in 14 dpf embryos. However, in vitro incubation with BaP (10–100 μM) had no effect on constitutive GNMT enzyme activity (Figure 7).

Figure 6.

BaP waterborne exposure inhibited GNMT enzyme activity in F. heteroclitus embryos. Embryo cytosolic protein (180 μg) was used and reactions were performed with 1 mM SAM and 10 mM glycine at 25°C for 60 min. Each bar represents n = 5 pools, 10 embryos per pool, * p<0.05, ** p<0.01. Constitutive GNMT enzyme activity was significantly higher at 4 dpf (b).

Figure 7.

In vitro effects of BaP on adult F. heteroclitus liver cytosol GNMT enzyme activity. There were no significant changes in enzyme activity when liver cytosols were incubated with 1 μL of 1 mM, 5 mM or 10 mM BaP (in DMSO, final concentrations were 10 μM, 50 μM and 100 μM, respectively), DMSO solvent, or water before the GNMT assay. Adult liver cytosolic protein (250 μg) was used and reactions were performed with 1 mM SAM and 10 mM glycine (pH 9.0) at 25°C for 30 min. Each reaction was performed five times.

4. Discussion

F. heteroclitus is a estuarine teleost found in contaminated environments along the Atlantic coast of the US with relevant human diseases including cancer and reproductive/developmental deficits (Meyer et al. 2002; Pait 2001; Vogelbein et al. 1990). Because of its small size, external development, and transparent chorion, it is suitable to study developmental toxicities. The F. heteroclitus GNMT deduced protein has 295 amino acids and 74% identity with human GNMT (Fang et al. 2009). As shown in Table 1, the F. heteroclitus liver GNMT Kms compared similarly with previously reported mammalian kinetic values. Therefore, F. heteroclitus is a relevant and useful model organism to further investigate GNMT functions in embryogenesis and BaP-mediated developmental toxicity.

Table 1.

Kinetic parameters of glycine N-methyltransferase of liver cytosol from different species.

Species	Km(SAMa) mM	Km(glycine) mM	Assay Temperature	Reference
Homo sapiens	0.28	12.2	25°C	(Pakhomova et al. 2004)
Homo sapiens	0.27	6.3	30°C	(Ogawa et al. 1993)
Mus musculus	0.18	3.6	25°C	Pakhomova et al. 2004
Rattus norvegicus	0.11	2.0	25°C	Pakhomova et al. 2004
Rattus norvegicus	0.05	0.29	30°C	Ogawa et al. 1993
Oryctolagus cuniculus	0.32	2.9	30°C	Ogawa et al. 1993
Oryctolagus cuniculus	0.2	1	Unspecified	(Kloor et al. 2004)
Oryctolagus cuniculus	0.1	2.2	37°C	Heady and Kerr 1973
Sus scrofa	0.25	11.1	30°C	Ogawa et al. 1993
Fundulus heteroclitus	0.18	3.24	25°C	This study.

^aSAM: S-adenosylmethionine

GNMT protein was first discovered in guinea pig livers (Blumenstein and Williams 1963) and then purified ten years later (Heady and Kerr 1973). Sequences for GNMT have been identified in many other species including human, mouse, rat, rabbit, cow, dog, chicken, zebrafish, medaka and fruitfly, etc. In mammals, GNMT mRNA and protein expression is highest in the liver, pancreas, prostate and kidney (proximal kidney tubules) (Yeo and Wagner 1994). But to date, GNMT expression during oogenesis, embryogenesis and its developmental significance have not been reported in any other organisms. In this study, we found that GNMT mRNA exists in developing oocytes, was expressed throughout development, and was mainly localized in the F. heteroclitus embryonic liver and urinary bladder. GNMT enzyme activity was also detected in unfertilized eggs and developing embryos suggesting its critical role during early development.

In our previous mRNA measurements using quantitative RT-PCR, we found unfertilized F. heteroclitus eggs had the highest expression of GNMT mRNA, about 19-fold higher than 4 dpf embryo expression (Fang et al. 2009), and hypothesized GNMT was possibly a maternally deposited mRNA. In this study, we confirmed that GNMT was actively transcribed and accumulated in growing oocytes and detected GNMT activities in unfertilized eggs. However, the GNMT activity did not peak in unfertilized eggs (as could be predicted based on mRNA levels) but rather later during development. This suggests a large portion of the maternally deposited GNMT mRNA was not translated to GNMT protein before fertilization. Generally, as discussed in (Alizadeh et al. 2005), during meiotic maturation and in embryos shortly after fertilization accumulated maternal mRNA species degrade rapidly for protein synthesis if they encode essential proteins for these specific events. This could account for our previous results that constitutive whole embryo GNMT mRNA expression dropped dramatically post fertilization. In 2 and 3 dpf F. heteroclitus embryos, GNMT expression was still significantly higher than when expression leveled off between 4 and 14 dpf. Possibly resulting from the rapid translation of maternally deposited mRNA, GNMT activity was also significantly higher in early development (4 dpf) than later time-points, suggesting GNMT activity is critical to F. heteroclitus embryogenesis at early stages corresponding to periods of rapid organogenesis (Armstrong and Child 1965). Interestingly, the temporal expression pattern of embryonic GNMT is very similar to DNA methyltransferase 1 (DNMT1). Maternal transcripts of the zebrafish DNMT1 are present at high levels in early embryos and then the mRNA abundance decreases after the blastula stage (Martin et al. 1999; Mhanni and McGowan 2002).

In this study, we found by ISH that the developmental toxicant, BaP, induced GNMT mRNA expression modestly in the embryo liver region in 7, 10, 14 dpf. We also did ISH on 4 dpf embryos, but embryos are undergoing organogenesis and organs like liver and urinary bladder are not easily visualized. It was difficult to distinguish and quantitate the tissue specific expression by ISH at this time point. However, the same trends in GNMT mRNA induction by BaP were found using whole mount ISH and our earlier quantitative RT-PCR results on homogenized whole F. heteroclitus embryo pools where GNMT mRNA induction at 3, 10, and 14 dpf by BaP treatment was detected (Fang et al. 2009). While studies of BaP effects on GNMT are very limited, BaP exposure also induced GNMT mRNA expression in HepG2 cells in vitro (Lee et al. 2006a). Induction of GNMT expression is consistent with the recent recognition that the 5′-regulatory region of the human GNMT contains a BaP inducible XRE (Lee et al. 2009). While the F. heteroclitus GNMT promoter sequence data are not available, our bioinformatics analysis using the Transcription Element Search System (TESS, combined search) on the GNMT promoter of Danio rerio (zebrafish), a more genetically established teleost model, revealed five putative AhR or AhR/ARNT binding sites between -2000 nt and the transcription start site (Table 2). Collectively, these new data suggest that the AhR may have a role in GNMT expression and may explain the BaP-induced mRNA expression we found in F. heteroclitus embryos. The AhR-GNMT relationship is further suggested in AhR null mice where liver constitutive GNMT expression is significantly lower than in wild-type mice (Boutros et al. 2009). Therefore, the transcriptional induction by BaP was possibly mediated through the AhR and XRE in the upstream region of the GNMT gene.

Table 2.

Putative xenobiotic response elements and their locations in the GNMT (zebrafish) promoter regions

Xenobiotic response elements	Locations
AhRa	−1574 to −1558
AhR	−1535 to −1524
ARNTb -AhR	−1185 to −1174
AhR	−765 to −760
AhR	−738 to −727

Locations numbered back from transcription start site (+1). GenBank Accession No. NW_001880090.1.

^aAhR: Aryl hydrocarbon receptor

^bARNT: Ah receptor nuclear translocator

In contrast to mRNA induction, BaP exposure significantly reduced GNMT enzyme activity. To determine if BaP is a direct inhibitor of GNMT, BaP was added to the enzyme reaction system, but it did not alter the GNMT enzyme activity in F. heteroclitus liver cytosol (Figure 7). Our results contrasted to a similar in vitro study (Chen et al. 2004) showing that when BaP was incorporated into a recombinant human GNMT assay system, GNMT activity was reduced by 50% when compared to control DMSO. In either case, the in vitro studies do not represent the in vivo system where BaP metabolites may also be acting as GNMT inhibitors.

Another possible mechanism to explain enzyme inhibition is that the BaP-mediated mRNA increase is reflected in more post translational assembly of the homodimeric BaP-binding form. As a result, the assembly of the enzymatically active homotetrameric form of GNMT could be decreased. In fact, Bhat and coworkers found that BaP altered the post-translational modification of GNMT and resulted in formation of GNMT homodimer, which is the PAH binding but nonenzymatic form (Bhat et al. 1997). In our case, BaP may shift the GNMT translation from tetramer to dimer and thus reduce the GNMT enzyme activity. To specifically address this latter hypothesis a fish-specific GNMT antibody to measure relative homodimer vs. tetramer protein abundances following BaP exposure is necessary. Unfortunately, the commercially available human GNMT antibody does not cross-react with the F. heteroclitus protein (data not shown).

To date, studies suggest GNMT is involved in cellular defense mechanisms in response to environmental stressors and disease states. After BaP exposure, GNMT transfected cells have reduced BPDE formation (Chen et al. 2004), and cell lines expressing constitutively higher GNMT expression have less cytotoxicity than low GNMT expressing cell lines (Lee et al. 2006). Carcinogenic effects of aflatoxin B1 were reduced in GNMT transgenic mice that express human GNMT in the liver and kidney (Yen et al. 2009). GNMT enzyme activity and protein expression were induced after exposure to various xenobiotics, such as ethanol (Villanueva and Halsted 2004), retinoid compounds (Rowling and Schalinske 2001), glucocorticoids (Rowling and Schalinske 2003), excess dietary methionine (Rowling et al. 2002a) and streptozotocin (used to establish diabetic animal models) (Nieman et al. 2006). Thus, GNMT induction is responsive to various environmental stressors and is possibly involved in protective mechanisms. On the other hand, diminished GNMT expression is commonly seen in various cancer states including HCC, prostate cancer and cholangiocarcinoma (Tseng et al. 2003; Huang et al. 2007; Huang et al. 2008). Clearly, there remains much to resolve with respect to the roles of GNMT in disease and the relative utility of message expression versus enzyme activity as a biomarker of effect.

Another concern about the diminished developmental GNMT activities in vivo after BaP exposure is the potential effect on SAM/SAH homeostasis and DNA methylation status. A number of more recent studies have established a relationship between altered GNMT expression and DNA methylation (Lu et al. 2000; Mato et al. 1997; Rowling et al. 2002b). Altered GNMT activity will change the ratio of SAM/SAH and affect the methylation capacity in the body. Compared to control wild-type animals GNMT knockout mice had 7-fold more methionine, 35-fold more SAM and a 100-fold increase in the SAM:SAH ratio (Luka et al. 2006) and changes in global or chromosome specific DNA methylation patterns (Martinez-Chantar et al. 2008). These studies highlight the ability for altered GNMT, SAM and SAH homeostasis to affect critical cellular pathways epigenetically. Our future studies will focus on measuring the embryonic SAM and SAH levels and subsequent global or gene specific methylation changes after BaP exposure in this model organism. If these changes are detected, it will expand our knowledge on the epigenetic mechanisms of BaP toxicity.

PAHs are ubiquitous environmental contaminants which cause various embryo/fetal toxicities and cancer in both fish (Incardona et al. 2004; Wassenberg and Di Giulio 2004) and humans (Choi et al. 2008; Siddiqui et al. 2008; Stjernfeldt et al. 1986). BaP exposure during embryo-fetal development decreases GNMT enzyme activity and may change the pattern of DNA methylation of critical genes at critical times in development, have life-long effects on gene expression and phenotype, then result in diseases later in life (Szyf 2009; Tremblay and Hamet 2008). Recent studies challenge the idea that DNA methylation is irreversible, and some success in epigenetic therapy has been reported (Delcuve et al. 2009). Therefore, better understanding of whether BaP targets methylation homeostasis pathways such as GNMT and epigenetic consequences of that interaction are critical to prevent or potentially treat PAH toxicity.

5. Conclusion

In this study, we found that BaP exposure tissue-specifically induced GNMT mRNA expression measured by whole mount ISH but decreased GNMT enzyme activity in F. heteroclitus embryos throughout development. BaP-mediated changes in GNMT may represent a novel mechanism involved in BaP-induced embryo toxicity and DNA methylation changes.

Abbreviations

BaP

Benzo[a]pyrene

BPDE

BaP-7, 8-dihydrodiol 9, 10-epoxide

dpf

Days post fertilization

PTU

1-phenyl-2-thiourea

PAH

Polycyclic aromatic hydrocarbon

HCC

Hepatocellular carcinoma

hpf

Hours post fertilization

ISH

In situ hybridization

GNMT

Glycine N-methyltransferase

SAM

S-adenosylmethionine

SAH

S-adenosylhomocysteine

XRE

Xenobiotic response element