Dog Steroidogenic Factor-1: Molecular cloning and analysis of epigenetic regulation.

The authors declare that they have no conflict of interest.

To identify the true full-length sequence of dog SF-1 mRNA, we extracted total RNA from dog adrenal grand and performed 5′- and 3′-RACE ( Fig. 2a Fig. 2. (a) Representative electrophoresis image. All polymerase chain reaction (PCR) fragments (arrows) were sub-cloned and sequenced. Red arrows indicate the fragments including dog Steroidogenic factor 1 ( SF-1 ) mRNA sequence. Black arrows represent non-specific products. M: Marker. (b) The full-length cDNA sequence of dog SF-1 identified in this study. Numbers on the left represent nucleotide positions. The 5′ and 3′ untranslated regions (UTRs) are indicated in lower case. The open reading frame is indicated in upper case. The start codon and the stop codon are underlined. Sequences in red represent the region newly identified in this study. Sequences in blue represent the polyadenylation signal. (c) Comparison of exon-intron structure of ENSCAFT00000032206 and the full-length dog SF-1 . Exons are shown as boxes and are numbered. The coding region newly identified in this study are filled with red. ). In 5′-RACE PCR, five PCR products were obtained and sequenced. One of the five PCR fragments contained the dog SF-1 sequence. The other fragments were non-specific products. In 3′-RACE PCR, four PCR products were obtained and sequenced. One of the four PCR fragments contained dog SF-1 sequence. The other fragments were non-specific products. Combined the results of 5′- and 3′-RACE, a true full-length dog SF-1 cDNA sequence was determined. No other mRNA variant was detected. The full-length dog SF-1 cDNA was consisted of 3,016 base pairs (bp) including a 162 bp 5′ untranslated region (UTR), 1,386 bp ORF, and 1,468 bp 3′ UTR ( Fig. 2b ). The sequence was submitted to the DNA Data Bank of Japan (DDBJ) and assigned an accession number (ID: LC494495 ). (a) Representative electrophoresis image. All polymerase chain reaction (PCR) fragments (arrows) were sub-cloned and sequenced. Red arrows indicate the fragments including dog Steroidogenic factor 1 ( SF-1 ) mRNA sequence. Black arrows represent non-specific products. M: Marker. (b) The full-length cDNA sequence of dog SF-1 identified in this study. Numbers on the left represent nucleotide positions. The 5′ and 3′ untranslated regions (UTRs) are indicated in lower case. The open reading frame is indicated in upper case. The start codon and the stop codon are underlined. Sequences in red represent the region newly identified in this study. Sequences in blue represent the polyadenylation signal. (c) Comparison of exon-intron structure of ENSCAFT00000032206 and the full-length dog SF-1 . Exons are shown as boxes and are numbered. The coding region newly identified in this study are filled with red. Since the full-length sequence of dog SF-1 cDNA was clarified, we attempted to identify the dog SF-1 position in the Canis lupus familiaris genome database (CanFam 3.1) using BLAT. However, most of the dog SF-1 sequence identified in this study was not present in the database, indicating that the genomic sequence immediately upstream of the transcription start site (TSS) of ENSCAFT00000032206 is not registered. Therefore, we performed genome walking using Splinkerette PCR. Sequencing analysis of Splinkerette PCR products determined that the dog SF-1 is located on chromosome 9 (estimated position: 58,460,020-58,484,999) and contains seven exons ( Fig. 2c ). The identified genomic sequence was submitted to DDBJ and assigned an accession number (ID: LC494496 ). We compared the nucleic acid sequence between the full-length dog SF-1 determined in this study and ENSCAFT00000032206 ( Fig. 2c ). The full-length dog SF-1 cDNA was 300 bp long upstream of the predicted start codon of ENSCAFT00000032206. The 3′ UTR of the full-length dog SF-1 cDNA was 1,716 bp, which was shorter than that of ENSCAFT00000032206. Other parts of the full-length dog SF-1 cDNA sequence almost matched to those of ENSCAFT00000032206, except for some single nucleotide polymorphisms. Based on the cDNA sequence, the dog SF-1 protein is composed of 461 amino acids, with an estimated molecular weight of 51.6 kDa and theoretical isoelectric point of 7.66. When the amino acid sequence of the full-length dog SF-1 was aligned to the SF-1 of several mammals, the sequence in DBD completely matched, demonstrating that SF-1 is also highly conserved in dogs ( Fig. 3 Fig. 3. Multiple sequence alignments of deduced amino acid (aa) sequence of full-length dog Steroidogenic factor 1 (SF-1) with other species. Alignments were performed with clustalW. Conserved aa sequences are indicated by a dark background. Highly similar aa sequences are indicated by a dark grey background. Weakly similar aa sequences are indicated by a light grey background. Numbers on the right show the position of the aa sequence. DNA binding domain is boxed in red. ). The protein sequence of the full-length dog SF-1 shared 95.0 and 94.8% identity to the sequence of human and mouse, respectively. These data indicated that the full-length cDNA identified in this study is probably an original and normal sequence of dog SF-1 . Multiple sequence alignments of deduced amino acid (aa) sequence of full-length dog Steroidogenic factor 1 (SF-1) with other species. Alignments were performed with clustalW. Conserved aa sequences are indicated by a dark background. Highly similar aa sequences are indicated by a dark grey background. Weakly similar aa sequences are indicated by a light grey background. Numbers on the right show the position of the aa sequence. DNA binding domain is boxed in red. Clarifying the TSS of dog SF-1 identified the genome position and sequence of the basal promoter. The DNA sequence of the dog SF-1 basal promoter (from −120 bp to +120 bp) shared 88.9% and 81.8% identity to that of human and mouse, respectively. Similar to human and mouse SF-1 , the basal promoter of dog SF-1 contains regulatory elements including the SOX9 binding site, E box, CCAAT box, and Sp1/Sp3 site ( Fig. 4a Fig. 4. (a) Overview of the CpG sites in the Steroidogenic factor 1 ( SF-1 ) basal promoter (−120 pb ~ +120 bp) in dogs, humans, and mice. Lollipops indicate the position of individual CpG sites. Horizontal lines indicate transcription factor binding sites. Sx, a binding site for SRY-box9 (SOX9); E, an E box; C, a CCAAT box; Sp, a binding site for Sp1 or Sp3 transcription factors. (b, c) Effect of a demethylating reagent (5-aza-dC) on the expression of dog SF-1 gene in adipose tissue-derived mesenchymal stem cells (AD-MSCs). (b) Expression of the dog SF-1 and dog glyceraldehyde-3-phosphate dehydrogenase ( GAPDH ) genes by reverse transcription-polymerase chain reaction (RT-PCR). Expression of the dog SF-1 increased depending on 5-aza-dC concentration. GAPDH was used as an internal control. (c) Relative SF-1 gene expression levels measured by quantitative RT-PCR, with normalization to GAPDH expression using the Pfaffl method. The data shown represent the mean ± standard error (SE) (n=3). * P <0.05, ** P <0.01. N.D.: not detected. ). DNA methylation at the basal promoter in the SF-1 gene has been reported to regulate tissue-specific expression in humans and mice. Dog SF-1 also has 15 CpG sites around exon 1 ( Fig. 4a ). Thus, dog SF-1 may be also under the control of DNA methylation. (a) Overview of the CpG sites in the Steroidogenic factor 1 ( SF-1 ) basal promoter (−120 pb ~ +120 bp) in dogs, humans, and mice. Lollipops indicate the position of individual CpG sites. Horizontal lines indicate transcription factor binding sites. Sx, a binding site for SRY-box9 (SOX9); E, an E box; C, a CCAAT box; Sp, a binding site for Sp1 or Sp3 transcription factors. (b, c) Effect of a demethylating reagent (5-aza-dC) on the expression of dog SF-1 gene in adipose tissue-derived mesenchymal stem cells (AD-MSCs). (b) Expression of the dog SF-1 and dog glyceraldehyde-3-phosphate dehydrogenase ( GAPDH ) genes by reverse transcription-polymerase chain reaction (RT-PCR). Expression of the dog SF-1 increased depending on 5-aza-dC concentration. GAPDH was used as an internal control. (c) Relative SF-1 gene expression levels measured by quantitative RT-PCR, with normalization to GAPDH expression using the Pfaffl method. The data shown represent the mean ± standard error (SE) (n=3). * P <0.05, ** P <0.01. N.D.: not detected. To investigate whether DNA methylation is involved in the regulation of dog SF- 1 expression, we performed demethylation assay with the demethylating reagent 5-aza-dC using dog AD-MSCs. Dog SF-1 mRNA was not detected in the AD-MSCs. However, 5-aza-dC treatment induced the dose-dependent expression of dog SF-1 ( Fig. 4b and 4c ). This result suggested that DNA methylation affects the expression of dog SF-1 . We next analyzed the expression of dog SF-1 gene and DNA methylation around exon 1 in steroidogenic tissues, including adrenal, ovary, and testis. The gene expression was detected in adrenal, ovary, and testis, but the expression level was tissue-dependent ( Fig. 5a and 5b Fig. 5. Analysis of dog Steroidogenic factor 1 ( SF-1 ) expression and DNA methylation in the promoter. (a) Expression of the dog SF-1 gene in adrenal gland, ovary, testis and adipose tissue-derived mesenchymal stem cells (AD-MSCs) by reverse transcription-polymerase chain reaction (RT-PCR). The upper and lower panels show dog SF-1 and dog glyceraldehyde-3-phosphate dehydrogenase ( GAPDH ) expression, respectively. (b) Relative dog SF-1 gene expression levels measured by quantitative RT-PCR with normalization to GAPDH expression using the Pfaffl method. The data shown represent the mean ± SE (n=3). (c) DNA methylation level measured by combined bisulfite restriction analysis (COBRA) at the dog SF-1 promoter region in tissues and cultured cells. Numbers on the left indicate the position of CpG site from TSS. The data shown represent the mean ± SE (n=3). (d) Scatter plot of dog SF-1 gene expression and DNA methylation levels in the promoter, defined by quantitative RT-PCR and COBRA. Red line indicates linear regression line. (e) Bisulfite sequencing analysis around exon 1 of dog SF-1 gene. (Top) Diagram of the dog SF-1 gene. Exon 1 is shown as a white box. Vertical lines indicate the position of individual CpG sites. Black arrowheads represent the position of the CpG sites measured in COBRA (d). (Bottom) The open and closed circles indicate the unmethylated and methylated states of each CpG site, respectively. ). The DNA methylation rates of three CpG sites around the dog SF-1 promoter, −82 bp, + 286 bp, and + 426 bp from TSS, was analyzed using COBRA. DNA methylation rates were 0–3% in ovary, 30–39% in adrenal, 63–69% in testis, and 74–100% in AD-MSCs ( Fig. 5c ). There was a clear inverse correlation between the gene expression and the DNA methylation rates among samples ( Fig. 5d ). Bisulfite sequencing analysis was performed to investigate the DNA methylation state of individual CpG sites around the dog SF-1 promoter from positions −93 bp to + 426 bp, which contains 30 CpG sites ( Fig. 5e ). As expected, based on the COBRA results, methylation levels were low throughout the promoter in ovary, while AD-MSCs were highly methylated. In addition, the inverse correlation of methylation pattern to expression extended downstream of exon 1. These results indicated that the promoter activity of dog SF-1 is under the control of DNA methylation. Analysis of dog Steroidogenic factor 1 ( SF-1 ) expression and DNA methylation in the promoter. (a) Expression of the dog SF-1 gene in adrenal gland, ovary, testis and adipose tissue-derived mesenchymal stem cells (AD-MSCs) by reverse transcription-polymerase chain reaction (RT-PCR). The upper and lower panels show dog SF-1 and dog glyceraldehyde-3-phosphate dehydrogenase ( GAPDH ) expression, respectively. (b) Relative dog SF-1 gene expression levels measured by quantitative RT-PCR with normalization to GAPDH expression using the Pfaffl method. The data shown represent the mean ± SE (n=3). (c) DNA methylation level measured by combined bisulfite restriction analysis (COBRA) at the dog SF-1 promoter region in tissues and cultured cells. Numbers on the left indicate the position of CpG site from TSS. The data shown represent the mean ± SE (n=3). (d) Scatter plot of dog SF-1 gene expression and DNA methylation levels in the promoter, defined by quantitative RT-PCR and COBRA. Red line indicates linear regression line. (e) Bisulfite sequencing analysis around exon 1 of dog SF-1 gene. (Top) Diagram of the dog SF-1 gene. Exon 1 is shown as a white box. Vertical lines indicate the position of individual CpG sites. Black arrowheads represent the position of the CpG sites measured in COBRA (d). (Bottom) The open and closed circles indicate the unmethylated and methylated states of each CpG site, respectively.

The dog SF-1 mRNA recorded in the database (CanFam3.1) lacks the 5′-end sequence coding for the DBD, which is conserved in other mammals. In this study, we identified the true full-length cDNA of dog SF-1 . It possessed the 5′-end sequence coding DBD and shared high similarities with sequences in human, mouse, bovine, pig, and cat. In addition, determination of the 5′-end sequence of dog SF-1 mRNA enabled us to identify the genomic location and the genomic sequence of the promoter. There are many unclarified genomic sequences in the dog genome database [ 13 ]. Most of the RefSeq genes in dog have been computationally predicted based on CanFam3.1. In fact, 3,181 RefSeq genes in CanFam3.1 have unclarified regions within 5,000 upstream from TSS, suggesting that the correct 5′-end sequences of those genes have been veiled. Many previous studies have reported that point mutations of the human SF-1 gene cause adrenal insufficiency and disorders of sex development (DSD). In particular, mutations in DBD, such as p.G35E or p.R92Q, led to severe phenotype of those diseases [ 8 ]. The 46, XY DSD, which is the most common SF-1 -related disease featuring a DBD mutation, includes clitoral enlargement, small inguinal testes, and absent or rudimentary Müllerian structures as the typical phenotype. In the field of veterinary medicine, dog 78, XY DSD exhibits symptoms including testicular hypoplasia with clitoral enlargement, persistent Müllerian duct syndrome, cryptorchidism, and hypospadias, similar to human 46, XY DSD [ 17 , 23 ]. However, dog SF-1 gene mutations have not been detected in the dog 78, XY DSD. One of the reasons is that the genome sequence of dog SF-1 DBD have not been identified. Our results provide the sequence of dog SF-1 DBD and information that will be useful in veterinary medicine diagnosis and research. In this study, the basal promoter sequence of dog SF-1 gene was identified. Epigenetic analyses of the dog SF-1 promoter revealed that the expression of dog SF-1 gene is under the control of DNA methylation. These results indicate that epigenetic mutation influences gene expression of SF-1 in dog. Aberrant hypomethylation at the basal promoter of the SF-1 gene has been reported to induce ectopic gene expression in human endometriosis [ 33 ]. In dogs, ectopic-endometrium and endometrioma have been reported, which are homologous diseases to human endometriosis [ 1 , 2 , 7 , 21 ]. However, the relationship between those dog diseases and SF-1 has not been clarified. Based on our findings, it is possible that ectopic SF-1 gene is overexpressed in dog endometriosis by epigenetic mutation in the promoter. In conclusion, the complete sequences of mRNA and the promoter region of dog SF-1 were identified. Expression of dog SF-1 is under the control of DNA methylation at the promoter. Our results provide a molecular biological basis for a better understanding of developmental and metabolic mechanisms for dogs.

An adrenal gland from a male mixed breed dog, an ovary from a female beagle dog, a testis from a male beagle dog, and adipose tissues from a male chihuahua dog were collected during surgery at the University of Miyazaki Veterinary Teaching Hospital, Miyazaki, Japan, with the signed informed consent from dog owners and the ethical approval of the animal ethics committee of Faculty of Agriculture, University of Miyazaki, and the university’s research committee. All samples were grossly normal. The tissues used in this study were as follows; a part of the adrenal gland including capsule, cortex, and medulla; a part of the ovary including germinal epithelium, tunica albuginea, cortex, and medulla; and a part of the testis including tunica albuginea, vascular layer, parenchyma, and interstitium. The adrenal gland, ovary, and testis were immediately frozen in liquid nitrogen and then stored at −80°C until use. Intra-abdominal adipose tissues were aseptically collected from a 9-year-old male chihuahua dog. Adipose tissues were cut into pieces ≤0.2 mm 3 and digested at 37°C with 0.1% (w/v) Trypsin-EDTA (FUJIFILM Wako Pure Chemical Corp., Tokyo, Japan). After digestion, the cell suspension was centrifuged at 1,000 rpm for 3 min to collect the cells. The cells were cultured in Dulbecco’s modified Eagle’s medium-low glucose (DMEM-LG; Sigma-Aldrich, St. Louis, MO, USA) supplemented with 10% (v/v) fetal bovine serum (Thermo Fisher Scientific, Waltham, MA, USA), 2 mM GlutaMAX TM Supplement (Thermo Fisher Scientific), 100 U/m l Penicillin-Streptomycin (Thermo Fisher Scientific), and 0.1% 2-mercaptoethanol (Thermo Fisher Scientific). After 48 hr, the medium was replaced, and non-adherent cells were removed. The medium was changed every 3 days. When the growth of the AD-MSCs was 70–80% confluent, the cells were detached by incubation in 0.05% trypsin-EDTA for 5 min at 37°C. For demethylation assays, AD-MSCs were cultured for 96 hr in medium containing 0, 1, 5, or 10 µ M 5-aza-2′-deoxycytidine (5-aza-dC; Merck Millipore, Billerica, MA, USA). Total RNA was extracted from tissues and cells using ISOGEN II (FUJIFILM Wako Pure Chemical Corp.) following the manufacturer’s instructions. Quality and concentration were measured using a NanoDrop ® 2000C spectrophotometer (Thermo Fisher Scientific). Rapid amplification of cDNA ends (RACE) reactions were performed using the GeneRacer TM Kit with SuperScript TM III RT (Thermo Fisher Scientific) and template cDNA from total RNA obtained from adrenal gland. The GeneRacer TM Kit provides a method to the obtain full-length 5′ and 3′ ends of cDNA by removing the mRNA cap structure, ligating the GeneRacer TM RNA Oligo to the mRNA, and reverse transcribing the mRNA with oligo dT primer. Specific amplification products were obtained through polymerase chain reaction (PCR) performed under the following thermocycling conditions: 30 cycles of 98°C for 10 sec, 60°C for 5 sec, and 72°C for 3 min. The primers used are summarized in Table 1 Table 1. Primers used in this study Primers Sequence (5′-3′) Application 5′_dSF1-F CGACTGGAGCACGAGGACACTGA RACE 5′_dSF1-R GTCCACGATGGAGATGAAGG RACE 5′_Nested_dSF1-F GGACACTGACATGGACTGAAGGAGTA RACE 5′_Nested_dSF1-R GCTCTGGGTACTCAGACTTGATG RACE 3′_dSF1-F TCCAGAAGTGCCTGACAGTG RACE 3′_dSF1-R GCTGTCAACGATACGCTACGTAACG RACE 3′_Nested_dSF1-F AGCATCTGGGCAACGAGATG RACE 3′_Nested_dSF1-R CGCTACGTAACGGCATGACAGTG RACE CDS_dSF1-F ATGGACTATTCGTACGACGAG RACE CDS_dSF1-R TCAAGTCTGCTTGGCTTGCA RACE Sp_adaptor CGAAGAGTAACCGTTGCTAGGAGAGACC Splinkerette PCR Sp_Nested_adaptor GTGGCTGAATGAGACTGGTGTCGAC Splinkerette PCR Sp_dSF1-F CATGGACTATTCGTACGACGAGGACCTG Splinkerette PCR Sp_Nested_dSF1-F GCTACCACTACGGACTGCTCACG Splinkerette PCR Sp_dSF1-R ACCTTGCAGCTCTCGCACGTG Splinkerette PCR Sp_Nested_dSF1-R ACGTGAGCAGTCCGTAGTGGTAGC Splinkerette PCR M13-F TGTAAAACGACGGCCAGT Sequence M13-R CAGGAAACAGCTATGACCATG Sequence Seq_dSF1-F ACCGCACGCGCTGATATAG Sequence Seq_dSF1-R ACGACAAAACCCCGATTCTGAG Sequence GAPDH-F AATGCCTCCTGCACCACCAAC qPCR GAPDH-R GAAGGCCATGCCAGTGAGCTTC qPCR dSF1-F TCCAGAAGTGCCTGACAGTG qPCR dSF1-R TGAAGCCATTGGCTCGAATCTG qPCR Bis_dSF1-F GATTTAAATGAAGAGAAATATTAATAAAGAAGG Bisulfite PCR Bis_dSF1-R ACCATAAACACATTCACAAACTAC Bisulfite PCR . All PCR products were extracted with Wizard ® SV Gel and PCR Clean-Up System (Promega, Madison, WI, USA) and were ligated into pBluescript II SK (−) by In-Fusion (TaKaRa Bio Inc., Kusatsu, Japan). Ligated PCR products were sub-cloned and sequenced. Genomic DNA was isolated from tissues and cells by phenol and chloroform separation, and ethanol precipitation. This DNA was suspended in TE buffer (nacalai tesque, Kyoto, Japan). Quality and concentration were measured using the aforementioned NanoDrop ® 2000C spectrophotometer. Genomic DNA isolated from adrenal grand was amplified using Splinkerette PCR [ 26 ] with specific primers and sequenced to determine the genomic locations of dog SF-1 . All Splinkerette PCR experiments were performed under the following thermocycling conditions: 30 cycles of 98°C for 10 sec, 60°C for 5 sec, and 72°C for 3 min. The primers used are summarized in Table 1 . All PCR products were sub-cloned and sequenced described above. Ensemble IDs of transcriptions used in this study were as follows: Homo sapiens (ENST00000373588.8), Mus musculus (ENSMUST00000112883.7), Rattus norvegicus (ENSRNOT00000017651.3), Bos taurus (ENSBTAT00000011869.3), Sus scrofa (ENSSSCT00000034748.2), Felis catus (ENSFCAT00000026158.3), and Canis lupus familiaris (ENSCAFT00000032206.3). Multiple alignments of SF-1 protein sequences were analyzed using clustalW (https://clustalw.ddbj.nig.ac.jp). The open reading frame (ORF) of dog SF-1 was identified with ORF finder (https://www.ncbi.nlm.nih.gov/orffinder/). The amino acid sequence deduced from the cDNA sequence was obtained with the EMBOSS Transeq (https://www.ebi.ac.uk/Tools/st/emboss_transeq). Calculated molecular weights and predicted isoelectric points were obtained with EMBOSS Pepstats (https://www.ebi.ac.uk/Tools/seqstats/emboss_pepstats). Sequence identity of cDNA and basal promoter was analyzed using EMBOSS Stretcher (https://www.ebi.ac.uk/Tools/psa/emboss_stretcher) and EMBOSS Water (https://www.ebi.ac.uk/Tools/psa/emboss_water), respectively. TFBIND (http://tfbind.hgc.jp) was used to search for transcription factor binding sites in the dog SF-1 basal promoter. Total RNA was extracted from tissues and cells using ISOGEN II (FUJIFILM Wako Pure Chemical Corp.) following the manufacturer’s instructions. For the reverse transcription-polymerase chain reaction (RT-PCR), first-strand cDNA was synthesized using total RNA (1 µ g) with random hexamers and ReverTra Ace reverse transcriptase (TOYOBO Co., Ltd., Osaka, Japan). The cDNA template was amplified using BIOTAQ TM HS DNA Polymerase (Bioline Ltd.; London, UK) and specific primers for dog SF-1 and dog glyceraldehyde-3-phosphate dehydrogenase ( GAPDH ). All PCR experiments were performed under the following thermocycling conditions: 95°C for 10 min; 30 cycles of 95°C for 30 sec, 60°C for 30 sec, and 72°C for 1 min, with a final extension at 72°C for 10 min. Quantitative real-time PCR (qPCR) was performed using SYBR ® Green PCR master mix (Applied Biosystems, Woburn, MA, USA). Data were normalized to GAPDH expression. Gene expression levels are presented as the fold-change in expression, which was calculated using the Pfaffl method [ 22 ]. The sequences of the primers used in this study are summarized in Table 1 . Sodium bisulfite treatment of genomic DNA was performed using the EZ DNA Methylation-Gold TM Kit (Zymo Research, Irvine, CA, USA). PCR amplification was performed using BIOTAQ TM HS DNA Polymerase (Bioline Ltd.) and specific primers for dog SF-1 promoter. The sequences of primers used in this study are summarized in Table 1 . All PCR experiments were performed under the following thermocycling conditions: 95°C for 10 min; 35 cycles of 95°C for 30 sec, 55°C for 30 sec, and 72°C for 1 min, with a final extension at 72°C for 10 min. For COBRA [ 32 ], the PCR product was treated by HpyCH4IV (New England Biolabs Inc., Ipswich, MA, USA) or TaqI (New England Biolabs Inc.). Concentration of the treated PCR products was measured using MultiNA (Shimadzu, Kyoto, Japan). To determine the methylation states of individual CpG sites at the dog SF-1 promoter, the PCR product was gel-extracted, sub-cloned into the pGEM-T Easy vector (Promega), and sequenced. Methylation sites were visualized and quality control was performed using the QUMA web-based tool (http://quma.cdb.riken.jp/) [ 14 ]. Differences between two independent samples were evaluated by performing two-tailed Student’s t -test. All error bars represent the standard error of the mean. Linear regression and Pearson product-moment correlation coefficient were used to analyze correlations between gene expression and DNA methylation.