Testicular Dnmt3 expression and global DNA methylation are down-regulated by gonadotropin releasing hormones in the ricefield eel Monopterus albus

In vertebrates, DNA methyltransferase 3 (Dnmt3) homologues are responsible for de novo DNA methylation and play important roles in germ cell development. In the present study, four dnmt3 genes, dnmt3aa, dnmt3ab, dnmt3ba and dnmt3bb.1, were identified in ricefield eels. Real-time quantitative PCR analysis showed that all four dnmt3 mRNAs were detected broadly in tissues examined, with testicular expression at relatively high levels. In the testis, immunostaining for all four Dnmt3 forms was mainly localized to spermatocytes, which also contained highly methylated DNA. All three forms of Gonadotropin-releasing hormone (Gnrh) in the ricefield eel were shown to decrease the expression of dnmt3 genes in the in vitro incubated testicular fragments through cAMP and IP3/Ca2+ pathways. Moreover, in vivo treatment of male fish with three forms of Gnrh decreased significantly the testicular Dnmt3 expression at both mRNA and protein levels, and the global DNA methylation levels. These results suggest that the expression of Dnmt3 and global DNA methylation in the testis of ricefield eels are potentially down-regulated by Gnrh, and reveal a novel regulatory mechanism of testicular Dnmt3 expression in vertebrates.

DNA methylation, a key epigenetic mark, participates in many physiological processes including cellular differentiation 1 , and plays crucial roles in development of vertebrates 2 . DNA methylation occurs mainly at the fifth position of cytosine ( 5 mC) in the dinucleotide CpG 3 , and is catalyzed by a group of enzymes called DNA methyltransferases (Dnmts) including Dnmt1 and Dnmt3. Dnmt1 is involved in the methylation of hemimethylated DNA and thus called maintenance DNA methyltransferase, while Dnmt3 is able to place methylation marks on previously unmethylated CpGs of DNA and thus mainly responsible for the de novo DNA methylation during development 4,5 . It has been demonstrated that cellular DNMT expression is positively correlated with the global DNA methylation level in cell lines 6 .
Multiple Dnmt3 genes exist in vertebrates. In mammals, Dnmt3 subfamily was thought to be composed of three members, namely Dnmt3a, Dnmt3b, and Dnmt3l 7 . Recently, a fourth member, Dnmt3c, was identified 8 . Of these, Dnmt3a, Dnmt3b, and Dnmt3c have been proven to have catalytic activities in vivo 7,8 , whereas Dnmt3l is a catalytically inactive DNA methyltransferase cofactor 7 . In teleost like zebrafish, six different dnmt3 genes are identified 9,10 , which are designated as dnmt3aa, dnmt3ab, dnmt3ba, dnmt3bb.1, dnmt3bb.2, and dnmt3bb.3, respectively, based on the nomenclature in ZFIN (http://zfin.org/). Dnmt3 genes have been shown to be expressed in multiple tissues, including the testis and ovary of mouse 11,12 and zebrafish 9 . During the development of male germ cells in mouse, the expression of Dnmt3a and Dnmt3b exhibits dynamic patterns 13,14 . Dnmt3c was shown to be exclusively expressed in male germ cells in mouse, and its peak expression coincided with male germline de novo DNA methylation 8 . These studies highlight the importance of the tightly regulated expression of Dnmt3a, Dnmt3b, and Dnmt3c during spermatogenesis. However, few studies have explored the transcriptional
Sequence alignment showed that all four forms of ricefield eel Dnmt3 contain the conserved putative functional domains, including one proline-tryptophan-tryptophan-proline (PWWP) motif domain, one plant homeodomain (PHD)-like Zinc finger domain, and one C-terminal catalytic domain (Supplementary Fig. S2 and S3). In contrast to the other three forms, dnmt3ba contains a calponin-homology (CH) domain in the N-terminal region ( Supplementary Fig. S3).
Tissue distribution patterns of dnmt3 mRNAs in male ricefield eels. The expression of dnmt3 genes in tissues of male ricefield eels was analyzed with real-time quantitative PCR (Fig. 1). Results showed distinct tissue patterns for four dnmt3 mRNAs in male ricefield eels, with broad distributions in tissues examined.
The expression of dnmt3aa was detected in all tissues examined, with relatively higher levels in the spleen, eye, cerebellum, and pancreas, and relatively lower levels in the gut, liver, and heart (Fig. 1a). The expression of dnmt3ab was detected at relatively higher levels in the pituitary, olfactory bulb, telencephanlon, cerebellum, mesencephalon, and hypothalamus, relatively lower levels in the medulla oblongata, urinary bladder, spleen, pancreas, eye, testis, blood, kidney, and heart, and much lower levels in the muscle, gut, and liver (Fig. 1b). The expression of dnmt3ba was detected at the highest level in the testis, and relatively lower levels in the spleen, pancreas, medulla oblongata, kidney, urinary bladder, and blood, but hardly detectable in the gut, heart, muscle, and liver (Fig. 1c). The expression of dnmt3bb.1 was also detected at the highest level in the testis, and relatively lower levels in the eye, pituitary, medulla oblongata, hypothalamus, pancreas, and heart, but barely detectable in the muscle, gut, and liver (Fig. 1d). Cellular localization of immunoreactive Dnmt3 and global DNA methylation status in testes of ricefield eels. The specific polyclonal antisera against the ricefield eel Dnmt3aa, Dnmt3ab, Dnmt3ba, and Dnmt3bb.1 were generated, which were shown to specifically recognize the corresponding antigens, respectively ( Supplementary Fig. S4). Western blot analysis indicated that Dnmt3aa and Dnmt3ab proteins were expressed at higher levels in the brain, testis, and spleen, and at lower levels in the kidney and liver (Fig. 2a and c). Dnmt3ba protein was expressed at a higher level in the testis and at lower levels in the brain, spleen, and kidney, but not detectable in the liver (Fig. 2e). Dnmt3bb.1 protein was expressed at higher levels in the brain, testis, spleen, and kidney, and at a lower level in the liver (Fig. 2g). Pre-adsorption of the antisera by excessive corresponding recombinant Dnmt3-N polypeptides abolished the immunoreactive signals in tissue homogenates (Fig. 2b,d,f and h), further confirming the specificities of the antisera generated.
The localization of the four forms of Dnmt3 in testes of ricefield eels was examined by immunohistochemistry (Fig. 3a~d). Immunostaining for all four forms of Dnmt3 was present dominantly in germ cells, particularly spermatocytes, but only weakly in spermatogonia and somatic cells of testes. The pre-adsorption of the antisera with corresponding recombinant Dnmt3-N polypeptides abolished the immunoreactive signals ( Supplementary Fig. S5).
The global DNA methylation status was also examined in testes of ricefield eels by using 5-Methylcytosine (5-MeC) antibody, which could detect 5-methylcytosine but not unmethylated cytosine. Immunostaining for 5-MeC was predominantly present in germ cells, and the immunoreactive signals seem to be strong in spermatocytes but weak in spermatogonia (Fig. 3e).
To examine the possible intracellular signaling pathways involved in the Gnrh-induced down-regulation of dnmt3, Rp-cAMPS (50 μ M, a PKA inhibitor), Go6983 (10 μ M, a PKC inhibitor), U73122 (10 μ M, a PLC inhibitor), or Xestospongin C (1 μ M, an IP 3 R inhibitor) was included in the in vitro incubated testicular fragments of ricefield eels together with Gnrh. Addition of Rp-cAMPS, U73122, or Xestospongin C abolished the inhibitory effects of Gnrh on dnmt3 expression, however, Go6983 showed no or only minor effects (Fig. 5). Rp-cAMPS, Go6983, U73122, or Xestospongin C alone did not show any effects on dnmt3 expression in testicular fragments of ricefield eels ( Supplementary Fig. S7). Moreover, all three forms of Gnrh increased cAMP levels in the in vitro incubated testicular fragments ( Supplementary Fig. S8). These results suggest that Gnrh down-regulates dnmt3 expression in the testis of ricefield eels possibly through both cAMP and IP 3 /Ca 2+ pathways but not the DAG/ PKC pathway.

Down-regulation of dnmt3 induced by Gnrh led to reduction of global DNA methylation levels in testes of ricefield eels in vivo.
The effects of Gnrh on dnmt3 expression in testes of ricefield eels were also examined in vivo via intraperitoneal injection. The treatment with Gnrh significantly down-regulated dnmt3 mRNA as well as Dnmt3 protein levels in testes of ricefield eels ( Fig. 6; Supplementary Fig. S9). Concomitantly, the global DNA methylation levels, as reflected by immunoreactive 5-Methylcytosine levels, were also decreased significantly in testes of ricefield eels after injection of Gnrh (Fig. 7).

Discussion
DNA methyltransferase 3 (Dnmt3) catalyzes the de novo DNA methylation and plays important roles in metabolism and development 5,29 . Multiple forms of Dnmt3 have been shown to exist in mammals 11,30 and teleosts 9,10,31,32 . In the present study, four Dnmt3 homologues were identified in ricefield eels. Similar to those in other vertebrates, ricefield eel Dnmt3 homologues contain the conserved functional domains of DNA methyltransferases, including a catalytic domain at C-terminal, a PWWP motif domain and a plant homeodomain (PHD)-like Zn finger domain at N-terminal. Interestingly, a calponin homology (CH) domain was identified only in the N-terminus of ricefield eel Dnmt3ba, but not the other three paralogues. Similar case also exists in zebrafish 10,33 , but CH domain is not identified in mammalian Dnmt3 homologues 34 . CH domain belongs to a family of actin binding domains which bind to microtubules 35 . These lines of evidence suggest that teleost Dnmt3ba may possess some unique functions.
In zebrafish, RT-PCR analysis showed that two dnmt3a genes were ubiquitously expressed in adult tissues, while four dnmt3b paralogues were differentially expressed 9 . To extend this observation, our present study demonstrated distinct tissue patterns for four dnmt3 mRNAs in male ricefield eels with real-time quantitative PCR analysis, which were further confirmed by Western blot analysis of tissue homogenates from the brain, testis, spleen, kidney, and liver. On the whole, both two dnmt3a genes are expressed more widely and with likely higher abundances (which was also supported by the images of agarose gel electrophoresis of semiquantitative RT-PCR products; data not shown) than dnmt3b genes in tissues of male ricefield eels. In most tissues of bovine, similarly, Dnmt3a mRNA was also detected with more higher levels than Dnmt3b 36 . These lines of evidence suggest that the tissue patterns of dnmt3a and dnmt3b may be conserved across vertebrates, and Dnmt3a homologues may play more important roles for de novo DNA methylation in more tissues of male ricefield eels and other vertebrates as well.
Significant differences in mRNA expression levels of the four dnmt3 genes were observed among tissues in male ricefield eels, with relatively higher levels in the testis especially for dnmt3ba and dnmt3bb.1. Similarly, prominent expression of dnmt3 was also observed in the testis of zebrafish 9 , and dnmt3b in the testis of mammals as well 30 . DNA methylation has important implications for gamete integrity and transmission of epigenetic information to the next generation during spermatogenesis in mammals 29,37 , and DNA methyltransferases are suggested to be involved in these important processes 29,38 . In the mouse, both Dnmt3a and Dnm3b are expressed in the male germ cells at the mRNA and protein levels 14 , and they directly interact and cooperate to establish DNA methylation patterns 38,39 . Consistent with these notions, both strong immunoreactive signals for Dnmt3 and prominent global DNA methylation were detected in male germ cells, particularly in spermatocytes, in the testis of ricefield eels. These results suggest that ricefield eel Dnmt3 may play conserved roles in de novo DNA methylation during spermatogenesis as in other vertebrates.
The expression of Dnmt3 homologues in vertebrates has been shown to be altered by exposure to environmental toxicants, such as bisphenol A 18 , polychlorinated biphenyls 40 , and 2,3,7,8-tetrachlorodibenzo-p-dioxin 19 . However, the information on the regulation of dnmt3 by physiological factors in vertebrates remains scarce. As mRNAs for three forms of Gnrh and two forms of Gnrh receptors were detected in the testis of ricefield eels ( Supplementary Fig. S6), our present study proceeded to examine the possible involvement of Gnrh in the regulation of dnmt3. Both in vitro and in vivo treatments with Gnrh decreased dnmt3 mRNA expression in the testis of ricefield eels. The Gnrh-induced down-regulation of dnmt3 mRNAs was blocked by Rp-cAMPS, U73122, and Xestospongin C, but not by Go6983 in the in vitro incubated testicular fragments, indicating that cAMP and IP 3 / Ca 2+ signaling pathways mediate the down-regulation of dnmt3 by Gnrh in the testis of ricefield eels. Gnrh and/ or their receptors have also been detected in the testis of many other teleosts, such as black porgy 24 and salmon 25 . In the mature testis of rat, Gnrh mRNA was also shown to be present in Sertoli cells 41 and the Leydig cells 23 , and Gnrhr mRNA is mainly expressed in germ cells 23 . Thus, Gnrh signals may also likely regulate Dnmt3 expression in testes of other vertebrates, which warrants further study.
In accordance with changes at mRNA levels, immunoreactive levels of four Dnmt3 homologues were also decreased significantly in testes of ricefield eels after Gnrh treatments in vivo. Notably, the 5-methlcytosine levels in testes of ricefield eels were also significantly decreased after Gnrh treatments, implying a DNA hypomethylation status. Similarly, down-regulation of Dnmt1 and Dnmt3 in TK6 lymphoblastoid cell line induced by hydroquinone resulted in global DNA hypomethylation 42 , while up-regulation of dnmt1 and dnmt3 induced by 2,4-Dichlorophenol is associated with global DNA hypermethylation in the liver of goldfish 43 . The expression of dnmt1 in testes of ricefield eels was not significantly decreased by Gnrh treatments in the present study (data not shown). Taken together, results of present study suggest that Gnrh induced DNA hypomethylation in testes of ricefield eels most likely through the inhibition on the expression of Dnmt3 homologues. Although posttranslational modifications may affect Dnmt3 activities 44,45 , the parallel decreases of immunoreactive Dnmt3, global DNA methylation, and dnmt3 mRNA levels in the testis of ricefield eels after Gnrh treatment in vivo suggest that Gnrh signals regulate Dnmt3 most likely through effects on dnmt3 transcription and/or mRNA stability, rather than posttranslational modifications.
In the testis of vertebrates, apoptosis occurs during normal spermatogenesis 46 , and is thought to be essential for the maintenance of correct ratio of Sertoli cells and gametes 47,48 . Gnrh signals have been shown to induce apoptosis in the testis of immature rats 49 and mature fish [50][51][52] , however, the underlying mechanisms remain largely unknown. In mice, intraperitoneal injection of 5-aza-2′ -deoxycytidine, an inhibitor of DNA methyltransferase, led to DNA hypomethylation in spermatogonia and apoptosis in spermatogonia and spermatocytes, with prominent decreases of Dnmt3a and Dnmt3b immunoreactivities in all germ cells 53 . Decreased expression of Dnmt3a was also observed during apoptosis in primary cultures of rat hepatocytes 54 . Our present study demonstrated that Gnrh down-regulated Dnmt3 expression and reduced global DNA methylation levels in the testis, which may hint one possible mechanism of Gnrh induction of testicular apoptosis. Admittedly, further evidence on apoptosis in the testis of ricefield eels after Gnrh treatments, such as TUNEL staining or Caspase-3 expression, is desperately needed in this respect.
In conclusion, four Dnmt3 homologues, Dnmt3aa, Dnmt3ab, Dnmt3ba, and Dnmt3bb.1, were identified in ricefield eels, with Dnmt3aa being the most predominant form expressed in tissues examined. In the testis of ricefield eels, all four Dnmt3 homologues were shown to be localized predominantly to germ cells, particularly spermatocytes. It was shown for the first time that Gnrh treatment decreased the expression of Dnmt3 at both mRNA and protein levels, possibly through cAMP and IP 3 /Ca 2+ signaling pathways. The results of present study may shed light on the regulation of Dnmt3 genes in the testis of other vertebrates as well.  for the Care and Use of Laboratory Animals. Fishes were sacrificed by decapitation, and tissues were dissected out, frozen immediately in liquid nitrogen, and stored at − 80 °C until RNA extraction or western blot analysis. The testicular tissues for histology and immunohistochemistry were fixed in Bouin's solution for 24 h and stored in 70% ethanol until processing. The phenotypic sex of male ricefield eels was verified by histological examination according to our previous work 55 . The testes of the male ricefield eels employed in the present study were in active spermatogenesis, with the majority of germ cells at the stage of spermatocytes.

Figure 7. Effects of intraperitoneal injections of Gnrh 1, Gnrh 2, or Gnrh 3 on immunoreactive 5-Methylcytosine (5-MeC) levels in testes of ricefield eels.
The male ricefield eels were first examined by biopsy and then received intraperitoneal injections of Gnrh 1, Gnrh 2, or Gnrh 3 (0.1 μ g/g body weight) twice a week for four weeks. The 5-MeC level in the testis was analyzed with the immunohistochemistry as described in Methods, and immunoreactive signals were visualized by DAB chromogen. 5-MeC (1:200) antibody was used as the primary antibody. The secondary antibody was 1:500 diluted horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG H + L (115-035-003, Jackson ImmunoResearch Laboratories, Inc.). The representative testicular sections of the same fish before (a~d) and after (a'~d') treatments were shown, and the insets are the higher magnification of the boxed areas within each image, respectively. The immunoreactive 5-MeC levels are shown in panel e. SPC: spermatocyte; SPG: spermatogonium. Each bar in panel e represents the mean ± SEM (n = 3~4). The experiments were repeated twice and similar results were obtained. *P < 0.05 vs the control; # P < 0.05 for the differences between the indicated groups. Scale bar = 25 μ m.
Scientific RepoRts | 7:43158 | DOI: 10.1038/srep43158 RNA extraction. Total RNA was isolated from frozen tissues using TRIzol (15596-026, Invitrogen MA, USA) and quantified based on the absorbance at 260 nm. The 260/280 nm ratios for all RNA samples were between 1.9 to 2.0. The integrity of RNA was checked with agarose gel electrophoresis.
Cloning of the ricefield eel dnmt3 cDNAs. The ricefield eel testicular cDNA was transcribed from the testis total RNA with the RevertAid H Minus First Strand cDNA Synthesis Kit (K1622, Thermo Scientific, MA, USA) according to the manufacturer's instructions using the adapter primer AP. Two PCR products of 762 and 736 bp were generated with nested PCR, using the primer set eDnmt3-F 1 /R 1 for the first round of amplification, and primer set eDnmt3-F 1 /R 2 or eDnmt3-F 2 /R 1 for the second round of amplification. The sequences of the primers are listed in Supplementary Table S1. Seven clones from each PCR product were sequenced and three different sequences were obtained, which correspond to dnmt3aa, dnmt3ab, and dnmt3ba, respectively. Then the 3′ ends of dnmt3aa, dnmt3ab, and dnmt3ba cDNAs were obtained by the RACE method using nested PCR, and the 5′ ends were extended by nested PCR using gene-specific reverse primers and degenerate forward primers targeted to conserved nucleotide sequences in 5′ utrs of Dnmt3 homologues in other vertebrates. The cDNA sequence of dnmt3bb.1 was initially identified from the ricefield eel pituitary transcriptome database, and then confirmed by PCR cloning from the testicular cDNA. Details are provided in Supplementary Methods.
Real-time quantitative PCR analysis of dnmt3 mRNA expression. Total RNA samples isolated from tissues were first treated with RNase-free DNase I (AM2222, Thermo Scientific) to remove any genomic DNA contamination. The total RNA (1 μ g) was reverse transcribed with random hexamer primers using the RevertAid H Minus First Strand cDNA Synthesis Kit (K1622, Thermo Scientific) according to the manufacturer's instruction. The integrity of all the RNA samples was verified by the successful amplification of actb (actin, beta; AY647143.1). Then 1 μ l of cDNA template was used for the real-time quantitative PCR analysis of dnmt3aa, dnmt3ab, dnmt3ba, and dnmt3bb.1 mRNA levels, respectively. The real-time quantitative PCR and quantification of mRNA expression levels were performed as previously described 56  Western blot analysis. The recombinant proteins or tissue homogenates (300 μ g) were separated on a 8% SDS-PAGE gel and transferred to a methanol-activated polyvinylidene difluoride membrane (ISEQ00010, Merck Millipore, MA, USA) by electroblotting. The membrane was then blocked with 5% nonfat milk powder in 10 mM PBS buffer (137 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4 , 2 nM KH 2 PO 4 ) at 4 °C overnight. The blocked membrane was then incubated sequentially with the primary antiserum and second antibody, and exposed to a chemiluminescence substrate (BeyoECL Plus kit, P0018, Beyotime, Shanghai, China) according to the manufacturer's instructions. Details are provided in Supplementary Methods. Immunohistochemistry. The testicular sections (5 μ m) were deparaffinized, hydrated, and incubated with 3% hydrogen peroxide solution to quench the endogenous peroxidase activity, followed by antigen retrieval in 10 mM citrate buffer (pH = 6.0) at 95 °C for 15 min and blocking in 0.01 M PBS containing 10% normal goat serum for 30 min at room temperature. Then the sections were incubated with the primary mouse anti-Dnmt3aa, anti-Dnmt3ab, anti-Dnmt3ba or anti-Dnmt3bb.1 antiserum (1:200) at 4 °C overnight. After rinsing with PBS for 5 min three times, the sections were exposed to the secondary antibody (HRP-conjugated goat anti-mouse IgG, 1:500 dilution; 115-035-003, Jackson ImmunoResearch Laboratories, Inc., PA, USA) solution. After rinsing with PBS, the sections were developed with 3,3′ -diaminobenzidine (DAB), mounted, examined with a Nikon Eclipse Ni-E microscope (Nikon, Japan), and digitally photographed. To confirm the specificity of the immunostaining, control sections were incubated with the primary antiserum (in its working solution) pre-adsorbed with an excess of corresponding recombinant Dnmt3-N polypeptides. Additional negative controls included replacement of the primary antiserum with PBS or pre-immune serum and the omission of secondary antibody.
The assessment of global DNA methylation in the testis of ricefield eel was performed with immunohistochemistry using the anti-5-methylcytosine antibody (MABE146, Merck Millipore) according to a previous report 58  The immunoreactive levels of 5-Methylcytosine and Dnmt3 in the testes were analyzed with the Image Pro Plus software (Media Cybernetics, Inc., MD, USA) in a way similar to a previous report 59 . All the sections and photoimages for analysis were processed under the same conditions. The measurement for each fish was based on three sections at an interval of about 100 μ m, and at least three fish samples were analyzed for each group. The data are presented as means + SEM (n = 3~4). Details are provided in Supplementary Methods.
In vitro treatment of testicular fragments with Gnrh. The testicular tissues of male ricefield eels were dissected out and chopped into pieces of approximately 1 mm 3 with a scalpel. Approximately 25 mg of testicular minces were placed in each well of a 24-well tissue culture plate (142475, Nunc, Denmark) with 500 μ l of L15 medium (11415064, Gibco, MA, USA) containing 0.1 U/ml penicillin and 0.1 μ g/ml streptomycin (15140163, Gibco), and then incubated at 28 °C in a humidified incubator (SPX-250BSH-II, CIMO, Shanghai, China). After pre-incubation for 18 h, the medium was replaced and the testicular fragments were treated with 10 and 100 nM of ricefiel eel Gnrh 1 (pjGnrh; AAW51121), Gnrh 2 (cGnrh-II; AAW51119), or Gnrh3 (sGnrh; AAW51120), either in the absence or presence of the PKA inhibitor Rp-cAMPS (50 μ M; sc-24010, Santa Cruz, TX, USA), the PKC inhibitor Go6983 (10 μ M; S2911, Selleckchem, TX, USA), the PLC inhibitor U73122 (10 μ M; S8011, Selleckchem,), or the IP 3 R inhibitor Xestospongin C (1 μ M; 1280, TOCRIS, Bristol, UK) for 8 h. Six replicates were performed for each treatment except for Gnrh dose response, where the experiments were repeated three times with three replicates for each treatment. After completion of incubation, tissues were collected and expression of dnmt3aa, dnmt3ab, dnmt3ba, and dnmt3bb.1 was analyzed with real-time quantitative PCR as described above. The experiments were repeated at least three times, and similar results were obtained.
In vivo treatments of male fish with Gnrh. Presumably male ricefield eels (body length 35~45 cm, body weight 50~80 g) were purchased from a local dealer in Guangzhou, Guangdong, P. R. China, kept in 50-litre plastic tanks in laboratory under a natural photoperiod (23.11 °N) and temperature (25.1 ± 1.0 °C) from October to November, 2015, and fed live Tenebrio molitor. The tank water was replaced on alternate days.
After acclimatization for one week, ricefield eels were biopsied to identify their sex in a way similar to a previous report 60 , and only males were included in this study. One week after biopsy, male ricefield eels received intraperitoneal injections of Gnrh1, Gnrh2, and Gnrh3 at 0.1 μ g/g body weight or physiological saline (control) twice a week for four weeks. The dose of Gnrh was determined according to the previous report 60 and the results of the pilot experiments (data not shown). Twenty four hours after the last injection, the testicular tissues of ricefield eels were dissected out, and the expression of Dnmt3 homologues as well as the DNA methylation status were analyzed as above.
Statistical analysis. Data were analyzed by one-way ANOVA followed by the Tukey multiple comparison test using the SPSS 17.0 software (SPSS, Inc., NY, USA). Significance was set at P < 0.05.