HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 20, 17765-17774, May 17, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/20/17765    most recent M200954200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Xie, W. Articles by DeJong, J. Search for Related Content PubMed PubMed Citation Articles by Xie, W. Articles by DeJong, J. Social Bookmarking                      What's this?

Regulation of ALF Gene Expression in Somatic and Male Germ Line Tissues Involves Partial and Site-specific Patterns of Methylation^*

Wensheng Xie, SangYoon Han, Mohammed Khan, and Jeff DeJong

From the Department of Molecular and Cell Biology, University of Texas at Dallas, Richardson, Texas 75080

Received for publication, January 29, 2002, and in revised form, March 7, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

ALF (TFIIA/-like factor) is a germ cell-specific counterpart of the large (/) subunit of general transcription factor TFIIA. Here we isolated homologous GC-rich promoters from the mouse and human ALF genes and used promoter deletion analysis to identify sequences active in COS-7 and 293 cells. Further, bisulfite sequence analysis of the mouse ALF promoter showed that all 21 CpG dinucleotides between 179 and +207 were partially methylated in five somatic tissues, brain, heart, liver, lung, and muscle, and in epididymal spermatozoa from adult mice. In contrast, DNA from prepubertal mouse testis and from purified spermatocytes were unmethylated except at C⁺¹⁹G and C⁺¹⁷⁰G. We also found that ALF expression correlates with a strong promoter-proximal DNase I-hypersensitive site present in nuclei from testis but not from liver. Finally we show that in vitro methylation of the ALF promoter inhibits activity and that 5-aza-2'-deoxycytidine treatment reactivates the endogenous ALF gene in a panel of seven different mouse and human somatic cell lines. Overall the results show that silencing in somatic cells is methylation-dependent and reversible and that a unique CpG-specific methylation pattern at the ALF promoter precedes expression in pachytene spermatocytes. This pattern is transient as remethylation of the ALF promoter in haploid germ cell DNA has occurred by the time spermatozoa are present in the epididymis.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The ALF¹ gene (TFIIA/-like factor; TFIIA) encodes a 478-amino acid protein related to the large (/) subunit of general transcription factor TFIIA (1, 2). ALF, together with the small () subunit of TFIIA, can stabilize TATA-binding protein (TBP)-TATA box interactions and can restore RNA polymerase II-dependent activity to TFIIA-depleted HeLa cell extracts. In contrast to the ubiquitously expressed TFIIA/ gene, ALF is only expressed in germ cells (3).

The genes for several other RNA polymerase II-associated factors are also selectively transcribed in reproductive tissues. For instance, the Drosophila TAF_II80-related cannonball gene is expressed only in testis (4), the mouse TRF2/TLF gene is preferentially expressed in testis (5, 6), and the mouse TAF_II105 gene is preferentially expressed in testis and ovary (7). These factors, as well as ALF, may be present in germ cell-specific preinitiation or coactivator complexes necessary for gametogenic or developmental patterns of gene expression (8). In fact, mutations in the cannonball and TRF2/TLF genes cause defects in spermatogenesis (4-6), inactivation of the TAF_II105 gene impairs oogenesis in mice (7), and Drosophila TAF_II60 mutants show defects in both male and female germ cell development (9).

Expression of the RNA polymerase II machinery in testis is tightly controlled during mouse development. In particular, genes for testis-specific factors such as ALF and TRF2/TLF and for non-tissue-specific factors such as TBP, TFIIA/, TFIIA, and TFIIB genes are turned on or up-regulated, respectively, in prepubertal mice at postnatal day 14 (3, 10, 11). The timing of expression corresponds with the appearance of primary spermatocytes in the pachytene stage of meiotic prophase (12). Interestingly meiotic gene expression often involves initiation at promoters that are distinct from those utilized in somatic tissues. For instance, the human TRF2/TLF gene initiates from a testis-specific promoter that is separate from its weak somatic cell promoter (13). Likewise the mouse TBP gene initiates from five unique testis-specific promoters in addition to its normal somatic cell promoter (14). In contrast, the tissue-specific ALF gene is only transcribed from a germ cell-specific promoter. At present, the mechanisms that control the tissue-specific expression and collective up-regulation of general transcription factor genes in meiotic prophase are not known.

Previous studies have shown some germ cell-specific genes to be regulated by methylation at CpG dinucleotides, an epigenetic modification catalyzed by a family of CpG-specific DNA methyltransferases (15-17). For example, the promoters of the germ cell-specific cyclin A1, Ldh-C, MAGE, Pdha-2, Pgk-2, tH2B, and Tnp1 genes contain CpG sites that are methylated in somatic tissues and unmethylated in testis (18-24). The methylated promoter would presumably be packaged into an inactive chromatin configuration through the action of methyl-CpG-binding domain proteins and associated histone deacetylases (16, 25), whereas the unmethylated promoter present in germ cells would be accessible to the transcription machinery. However, while a role for methylation in controlling X-linked and imprinted genes has been established (26, 27), the correlation between methylation and tissue-specific or developmentally regulated gene expression is not absolute (28-30). Still these studies raise the possibility that the differential expression of genes encoding the RNA polymerase II machinery in somatic and germ line tissues might involve DNA methylation.

Here we examine mechanisms that control expression of the germ cell-specific transcription factor ALF. We isolated homologous GC-rich promoters of the mouse and human ALF genes, characterized core promoter elements required for expression in vivo, and found that the endogenous ALF promoter is accessible to DNase I in testis but not liver. Together with an analysis of CpG-specific methylation patterns, the results suggest that methylation and chromatin inaccessibility are associated with silencing in somatic cells and show that this effect can be reversed by 5-aza-2'-deoxycytidine (azaC). The data also reveal unique CpG- and cell-specific patterns of methylation at the ALF promoter and demonstrate that changes in those patterns occur during germ cell differentiation.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Mouse and Human ALF Genes-- The mouse ALF gene was isolated in a PCR-based screen of a bacterial artificial chromosome (BAC) library (Incyte Genomics) using mALF-10 (5'-GCTGGCATGGCCTTCATCAACCTGGTG-3') and mALF-11 (5'-CCCGCACGCCCTCGATGACATCTTCAA-3'). One clone (26090) was full-length, and a 7.0-kb BglII fragment that contained Exons 1-3 (GenBank^TM accession no. AF452125) was isolated and subcloned into pRSET (Invitrogen). A human BAC 96012D that contains the human ALF gene was purchased from Research Genetics (GenBank^TM accession no. AC073082).

Cell Lines-- Human testicular embryonal carcinoma NTERA-2 cl.D1 and mouse spermatogonia GC-1 spg cell lines were purchased from American Type Culture Collection (ATCC). The human embryonal kidney 293 cell line was a gift from Dr. Santosh D'Mello. The human lung cancer cell line HCC38 was a gift from Dr. John Minna at Southwestern Medical Center. The human neuroblastoma SH-SY5Y and mouse neuroblastoma NIE/115 cell lines were gifts from Dr. Gail Breen. All cells except HCC38 were grown at 37 °C in 5-10% CO₂ in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, streptomycin (100 µg/ml), and penicillin (100 units/ml). HCC38 was grown in RPMI 1640 medium with 5% fetal bovine serum.

RNase Protection Assays-- To make hybrid genomic DNA-cDNA RNase protection assay (RPA) constructs, the mouse ALF promoter and Exon 1 was amplified from BAC 26090 with either mALFrp-1 (5'-CCGGAATTCAAGCGAGCCTGGCGGCTCTC-3') and mALF-6 (5'-CACCAGGTTGATGAAGGCCATGCCAG-3') to give a 93-bp product called PCR I or with mALFrp-2 (5'-CCGGAATTCTCGCGGTTGCGCAGCAACG-3') and mALF-6 to generate a 126-bp product called PCR II. Exons 1 and 2 were amplified from the mouse ALF cDNA using ALFrp-3 (5'-ATACGAGCTCCTTTCAACACCTGCTCCTCGAT-3') and mALF-10 (5'-GCTGGCATGGCCTTCATCAACCTGGTG-3') to generate a 129-bp product called PCR III. PCR products were digested with HaeIII and ligated to make short (PCR I-PCR III) and long (PCR II-PCR III) fragments. These were digested with EcoRI and SacI and subcloned into pGEM T-Easy (Promega). A mouse angiotensin-converting enzyme (ACE) gene construct was prepared as described previously (31). RPA was performed using total RNA from the testis of 40-day-old mice hybridized with 1 × 10⁴ cpm/µg [-³²P]UTP-labeled sense or antisense RNA using the RPAIII kit (Ambion).

Transfection and Luciferase Activity Assays-- The mouse and human ALF promoters were amplified by PCR from the human and mouse BAC clones. For the human gene, reactions contained hALFpa-2 (5'-GACAGCACCTCCAGCACCTG-3') together with one of eight primers: hALFpa-0 (5'-AGCCTGGGCACCATTGAGCA-3'), hALFpa-3 (5'-GTGATCATGCCACTGCACTGCA-3'), hALFpa-4 (5'-GATGCTGCTGTACCACGCTG-3'), hALFpa-5 (5'-CTAGACCCAACCTAACCATCCG-3'), hALFpa-7 (5'-CGACCGCCTCTCCGCCTTGACC-3'), hALFpa-8 (5'-CCGCTCCATCTATTAACGTTCTC-3'), hALFpa-9 (5'-CGTTCTCCGTGGTTGCGCACCT-3'), or hALFpa-6 (5'-CGTTCAAAACGTGCCCAGTG-3'). For the mouse gene, reactions contained mALFpa-1 (5'-GCCAGCCGCTCTGTGCCTAACC-3') together with one of four primers: mALFpa-2 (5'-ACTTCTGCCTGAGGACTGGGGA-3'), mALFpa-3 (5'-CCTACGGAGAACAGAGGACAGC-3'), mALFpa-4 (5'-GAGGTTGCCCCATCGACCTGAC-3'), or mALF-pa-5 (5'-AGCAACGAGGACCCACGGTTCA-3'). Each 3'-primer contains an XhoI site, while all but one 5'-primer contains a KpnI site (hALFpa-0 contains a SacI site). The amplified fragments were inserted upstream of the firefly luciferase reporter gene in the pGL3 Enhancer plasmid (Promega).

Mutagenesis was carried out using mALFpa-7m (5'-GAGGACCCACGGTGTGCTCGCGAGCCTGG-3') in which the TTCAAAA sequence was replaced with TGTGCTC. PCR with mALFpa-7m and mALFpa-1 was performed with pmALF216 as the template. The amplified fragment was used as a primer with mALFpa-4 in a second PCR that was cloned into the pGL3 Enhancer plasmid.

COS-7 and 293 cells in a 24-well plate were grown to 50-80% confluence and transfected with 0.2 µg of DNA using the FuGENE 6 reagent (Roche Molecular Biochemicals). Transfections also contained 2.5 ng of pCMV Sport--gal DNA (ATCC) to monitor efficiency. Two days later whole cell extracts from triplicate transfections were assayed using the Luciferase Assay System (Promega) and a Turner TD-20e luminometer. Results are expressed as luciferase activity/-galactosidase activity.

DNase I Hypersensitivity Site Analysis-- Mouse liver and testis nuclei were prepared as described previously (32) with several modifications. Tissues were homogenized in 10 mM HEPES, pH 7.6, 15 mM KCl, 0.15 mM spermine, 0.5 mM spermidine, 1 mM EDTA, 2.4 M sucrose, 1% low fat powdered milk, 0.1% Triton X-100, 0.5 mM dithiothreitol, and 0.5 mM phenylmethylsulfonyl fluoride. The homogenate was transferred to an SW27 rotor tube that contained a cushion of 10 ml of homogenization buffer without milk. Nuclei were pelleted by centrifugation at 75,000 × g for 60 min at 4 °C and resuspended in 20 mM HEPES, pH 7.5, 50 mM NaCl, 1 mM EDTA, 1 mM dithiothreitol, and 1 mM phenylmethylsulfonyl fluoride at a concentration of 10 A₂₆₀ units/ml. Aliquots of 200 µl were added to 100 µl of the same buffer supplemented with 15 mM MgCl₂ and 4 mM CaCl₂, and 0.5-5 units of DNase I (Promega) were added for 10 min at 37 °C. Reactions were stopped by addition of EDTA to 40 mM and SDS to 1% (v/v), and DNA was purified after overnight incubation at 55 °C with 400 µg/ml proteinase K. PvuII-digested DNA was analyzed by Southern analysis using a random-primed PCR fragment prepared with mALFpa-3 and mALFpa-9 (5'-AGATGTTATTCCTGTTTTTACC-3').

Sodium Bisulfite Sequence Analysis-- Genomic DNA was isolated from liver, brain, heart, lung, muscle, and adult and prepubertal testis of male CD-1 mice. Sodium bisulfite treatment of genomic DNA (33) was performed as follows. Approximately 2 µg of genomic DNA was denatured in 0.3 M NaOH for 15 min at 37 °C. Freshly prepared sodium bisulfite and hydroquinone (Sigma) were added to a final concentration of 3.1 M and 0.5 mM, respectively, in a final volume of 1 ml. After 8-10 h of incubation at 50 °C, DNA was purified using miniprep columns (Promega) and dissolved in 50 µl of H₂O. NaOH was added to 0.3 M for 5 min at room temperature and neutralized with 6 M ammonium acetate, and DNA was precipitated.

PCR amplifications were performed with Elongase (Invitrogen) using bisulfite-treated DNA. Primer methy-4 (5'-TGTTTTGAAATTTGGGTGATTTTA-3') and methy-3 (5'-AAATAAAATTACCCCATCAACCTAA-3') were designed to amplify the bottom strand of DNA between 179 and +207 of the mouse ALF promoter. Primers for the mouse TBP gene were methytbp-1 (5'-CTCTACTTAAAAACCTTAATAAAAA-3') and methytbp-2 (5'-TTAGTTTGATTTTTAGGTTTTTGG-3'). Each primer also contained an EcoRI site to facilitate cloning. Cycling conditions were 2 min at 94 °C, 5 cycles of 1 min at 94 °C, 1 min at 50 °C, 1.5 min at 70 °C, 25 cycles of 30 s at 94 °C, 1 min at 50 °C, 1.5 min at 70 °C, and finishing at 70 °C for 6 min. PCR products were purified by the QIAquick Gel Extraction kit (Qiagen) and sequenced with the Radiolabeled Terminator Cycle Sequencing kit (United States Biochemical Corp.). Liver PCR products were digested with EcoRI and subcloned into pBlueScript II (Stratagene).

In Vitro Methylation Analysis-- Human or mouse ALF promoter-reporter constructs (20 µg) were treated with 20 units of SssI methylase (New England BioLabs) for 3 h with 160 µM S-adenosylmethionine (New England BioLabs). Methylation status was verified by digestion with HpaII or MspI. Methylated constructs were transfected into COS-7 or 293 cells and assayed for luciferase activity.

5-Aza-2'-deoxycytidine Treatment and RT-PCR-- Cells were seeded in 60-mm plates and grown to 80% confluence before the addition of 5-aza-2'-deoxycytidine (Sigma) at a final concentration of 3 µM (34). After 48 h with or without azaC, cells were washed twice with phosphate-buffered saline and harvested. RNA was extracted using TRIzol reagent (Invitrogen). Reverse transcription reactions were performed using the Advantage RT-for-PCR kit (CLONTECH) using an oligo(dT)₁₈ primer. PCR amplification to detect the presence of first strand cDNA for human ALF mRNA used 2a2-17 (5'-GGTGCTGTCATGGCCTGCCTCAACCCGG-3') and 2a2-8 (5'-ATGCTAGCTGAACCACTG-3'). Reactions to detect first strand cDNA for mouse ALF mRNA used mALF-3 (5'-GTTTTACGCCGGAAGACCTGA-3') and mALF-5 (5'-GTCCTCGTTGTCGCTGCTA-3'). Detection of glyceraldehyde-3-phosphate dehydrogenase was performed with primers provided in the kit.

Other Techniques-- Unit gravity sedimentation of spermatocyte populations was performed as described previously to obtain populations of pachytene, round, and elongating/elongated spermatocytes that were 70-90% pure (35). Mature spermatozoa were isolated from mouse epididymis.

Mobility shift assays were performed as described previously using proteins expressed and purified from Escherichia coli by Ni²⁺ affinity chromatography (1). The probes were mALF (5'-GAGGACCCACGGTTCAAAAGCGAGCCTGGC-3'), AdML (5'-AAGGGGGGCTATAAAAGGGGGTGGG-3'), and AdML mutant (5'-AAGGGGGGCTAGAGAAGGGGGTGGG-3').

Genomic DNA blotting was performed with liver or testis DNA digested with PstI, PstI and HpaII, or PstI and MspI. Probe A is a 318-bp fragment produced by PstI-XbaI digestion of pmALF866. Probe B is a 147-bp PCR product made with mALFrp-1 (5'-CTCGCGGTTGCGCAGCAACG-3') and mALFrp-6 (5'-TGGTCCAACTACTTCCGGTACGGTC-3'). Probe Mito is a 524-bp EcoRI fragment produced from a pBluescript clone that contains an 856-bp PCR product made with two mitochondrial DNA-specific primers, mMito-1 (5'-GGAATTCCGAGCTTGGTGATAGCTGGT-3') and mMito-2 (5'-GGAATTCTATTCTCCGAGGTCGC-3').

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The Mouse and Human ALF Promoters Are GC-rich-- The human ALF gene, located on chromosome 2, is downstream of the stoned B-like factor (SBLF) gene and upstream of the luteinizing hormone receptor gene (Fig. 1A) (3). We isolated the mouse ALF gene using a PCR-based screen of a mouse BAC library and mapped the promoter (GenBank^TM accession no. AF452125) by RPA. In RPA experiments, total adult mouse testis RNA was annealed to ³²P-labeled antisense strand RNAs transcribed from the mALFrp1-3 and mALFrp2-3 constructs. Digestion with RNaseA/RNaseT1 produced seven fragments, indicating that transcription initiates from one of seven sites within a 44-nucleotide window (Fig. 1B, lanes 5 and 7). The major site of initiation (site 3) is designated as position +1 (Fig. 1C), and the initiating ATG is located 27 nucleotides downstream. The mouse ACE gene was used as a control for RPA (Fig. 1B, lane 9) (31).

View larger version (56K): [in this window] [in a new window]   Fig. 1.   Organization and mapping of the mouse and human ALF gene promoters. A, the genomic organization of the human ALF, SBLF, and partial luteinizing hormone receptor (LHR) genes are depicted along with splicing patterns that generate SBLF, ALF, and the chimeric SALF (SBLF-ALF) mRNAs. B, RNase protection assays were performed with RNA probes (lanes 11-16) from the mALFrp1-3, mALFrp2-3, and mACErp1-2 constructs. Seven protected fragments were observed in reactions with antisense probes (lanes 5 and 7), but not with sense probes (lanes 6 and 8). Control RPA reactions (mACE) are shown in lanes 9 and 10. An mALF promoter sequencing reaction is shown in lanes 1-4. C, alignment of the human and mouse ALF proximal promoters. Open triangles beneath the mouse sequence show seven initiation sites between 19 to +25 (labeled 1-7). Dark arrows show the lengths of selected deletion constructs. Identical residues, including CpG dinucleotides, are shaded. A GC box and TTCAAA sequences are boxed, and Exon 1 is indicated by brackets. Sequences are numbered at left. D, distribution of CpG dinucleotides in human ALF, mouse ALF, and mouse TBP promoters. S and T refer to somatic and testis exons, and the dark bars show regions analyzed for methylation in this study. Other features are MER5, ALU elements, and simple (GA, CT, or T) repeats. nt, nucleotides.

An alignment of promoter-proximal sequences from the mouse and human ALF genes is shown in Fig. 1C. The sequences are 42% identical between 200 and 101 (numbering refers to the mouse promoter), 68% identical between 100 and 1, and 79% identical in Exon 1. Sequences outside these regions were only weakly similar. Both promoters are GC-rich (for mouse: 51% between 200 and 101, 71% between 100 and 1, 57% in Exon 1, and 58% between +48 and +148) and contain many CpG sites in their promoter regions (Fig. 1D). A GC box (GGGCGG in mouse and GGGCGT in human) at 90 is predicted to bind the Sp1 transcription factor. In addition, a conserved TATA-like element (TTCAAA) is located 35 nucleotides upstream of initiation site 3, and a similar element is present in the TRF2/TLF gene (GenBank^TM accession no. NM004865) (13). Computer analysis also revealed putative binding sites for cAMP-response element-binding protein, AP1, upstream stimulatory factor, CCTC-binding factor, and other factors. These sites were not analyzed further as they were not conserved.

The ALF Promoter Is Active in COS-7 and 293 Cells-- To test whether human ALF promoter constructs were active in somatic cells, we fused a series of truncated promoters to a luciferase reporter (phALF2693, phALF1029, phALF489, phALF219, phALF168, phALF130, phALF84, and phALF53) (Fig. 2A) and transfected them into COS-7 and 293 cells. Although the endogenous ALF gene is silent in somatic tissues, these deletion constructs were expressed to varying levels (Fig. 2B). The highest activity (~50% of the level seen with pGL3 Control, a vector that contains the SV40 promoter and enhancer) was observed with phALF219 in 293 cells, while the shortest construct, phALF53, was the least active. Mouse ALF promoter constructs (pmALF866, pmALF507, pmALF216, and pmALF57) (Fig. 2A) were also active in COS-7 and 293 cells (Fig. 2B). The highest activity (~95% of the level seen with the pGL3 Control) was seen with pmALF216 in 293 cells, while the shortest construct, pmALF57, was the least active. The background level of luciferase activity was determined using the promoter-less pGL3 Enhancer vector.

View larger version (23K): [in this window] [in a new window]   Fig. 2.   The mouse and human ALF promoters are active in 293 and COS-7 somatic cell lines. A, diagram of ALF promoter-luciferase constructs. B, luciferase reporter activity was determined for human constructs (phALF53, phALF84, phALF130, phALF168, phALF219, phALF489, phALF1029, and phALF2693) or for mouse constructs (pmALF57, pmALF216, pmALF507, pmALF866, and pmALF216mut) in COS-7 cells or 293 cells. The pGL3 Control and pGL3 Enhancer plasmids are positive and negative controls, respectively. pmALF216mut is identical to pmALF216 except that the TATA-like TTCAAA sequence has been altered (see text). C, bandshift assay reactions were performed with TBP and either ALF (lanes 1-4) or TFIIA (lanes 5-8) using the mALF TTCAAA element. Complex formation is abolished by the addition of mALF TTCAAA (10 pmol, lanes 2 and 6) or AdML TATA competitors (10 pmol, lanes 3 and 7) but not by a mutant (mut.) AdML TATA competitor (10 pmol, lanes 4 and 8).

We next evaluated properties of the conserved TATA-like TTCAAA element. Bandshift reactions were performed using the AdML TATA or mouse ALF TTCAAA boxes in the presence of TBP and either TFIIA or ALF (Fig. 2C). TBP-dependent complexes were formed on the mALF TTCAAA element (lanes 1 and 5), and they were competed by unlabeled mALF TTCAAA or AdML TATA oligonucleotides (lanes 2, 3, 6, and 7) but not by an AdML mutant (lanes 4 and 8). To test the functional importance of the TTCAAA element, we changed it from TTCAAAA to TGTGCTC (pmALF216mut) so that it was no longer recognized by TBP (data not shown). Surprisingly transfection analysis showed the pmALF216mut construct was as active as wild type pmALF216 (Fig. 2B). Taken together the results of these assays suggest ALF is controlled by a short, GC-rich, and TATA-less promoter (see "Discussion").

DNase I Hypersensitivity at the ALF Promoter-- We wished to know whether the ALF promoter was accessible in testis where it is expressed and inaccessible in a somatic tissue where it is silent. To address this question, we digested nuclear chromatin from testis and liver with DNase I. Following digestion, the restriction enzyme PvuII was used to generate a 2.5-kb fragment that spanned Exon 1 of the ALF promoter (Fig. 3). In the presence of DNase I, a fragment of 0.65 kb was produced in chromatin from both liver and testis (lanes 2, 3, 5, and 6). This band reflects a constitutive hypersensitive site located ~100 nucleotides downstream of Exon 1. In addition, a small and somewhat broad band of ~0.45 kb was observed only in testis (lanes 5 and 6). This tissue-specific hypersensitive site maps within the GC-rich ALF promoter. The results show that ALF expression is associated with an accessible chromatin configuration, presumably to facilitate binding of the RNA polymerase II machinery.

View larger version (42K): [in this window] [in a new window]   Fig. 3.   DNase I-hypersensitive site analysis. Nuclei from mouse liver and testis were digested with either no DNase I (lanes 1 and 4), 0.5 units of DNase I (lanes 2 and 5), or 5 units of DNase I (lanes 3 and 6) followed by digestion with PvuII. ALF promoter fragments were visualized by Southern blot analysis with a probe located at the 5'-end of the 2.5-kb PvuII fragment. The diagram below shows the positions of hypersensitive sites (designated HS-1 and HS-2).

Differential Methylation of ALF in Liver and Testis-- The observation that ALF promoter constructs were active in somatic cells whereas the endogenous gene is silent suggests regulation by an epigenetic mechanism. In particular, the presence of multiple CpG dinucleotides in the ALF promoter points to a possible role for cytosine methylation. To test this possibility, genomic DNA from mouse liver or testis was digested with the methylation-sensitive HpaII and methylation-insensitive MspI restriction enzymes. We focused the analysis on four sites (CCGG) located in a 2.1-kb PstI fragment spanning the promoter, Exon 1, and part of Intron 1 (Fig. 4A). Digestion of liver DNA with PstI/HpaII and hybridization with probe A revealed a strong 0.7-kb band and a weaker undigested band of 2.1 kb (Fig. 4B, lane 3). Digestion of liver DNA with PstI/HpaII and hybridization with probe B showed a weak 0.55-kb band and two partial digestion products of 2.1 and 1.4 kb (lane 10). Based on the intensities of the partially digested fragments, we estimate that ~5% of genomic DNAs are methylated at the HpaII site located just upstream of the transcription start site, while the three downstream sites are methylated to near completion. In contrast, results with testis genomic DNA revealed complete digestion to either the 0.7-kb band (lane 6) or the 0.55-kb band (lane 13), indicating that these sites are not methylated in this tissue.

View larger version (58K): [in this window] [in a new window]   Fig. 4.   Digestion of mouse liver and testis genomic DNA with HpaII and MspI reveals tissue-specific methylation of the ALF promoter. A, schematic of the mouse ALF promoter and the locations of PstI (P), HpaII (H), and MspI (M) sites. B, genomic DNA blots using PstI-, PstI-MspI-, and PstI-HpaII-digested DNA from liver and testis were hybridized with probe A (lanes 1-6) and probe B (lanes 7-13). PstI-HpaII reactions show incomplete digestion of liver DNA (lanes 3 and 10) and complete digestion with testis DNA (lanes 5 and 13). Arrows show the positions of partially digested products. C, a mitochondrial DNA probe shows the unmethylated mitochondrial genome is completely digested with HpaII or MspI (lanes 2, 3, 5, and 6).

To show that the results in Fig. 4B were not due to incomplete digestion, DNA from both tissues was hybridized with a mitochondrial DNA probe (probe mito). This genome is unmethylated and should be fully digested with either HpaII or MspI. The results show the expected mitochondrial DNA digestion products of 1.7 and 0.65 kb (Fig. 4C, lanes 2, 3, 5, and 6), indicating that digestion was complete.

Bisulfite Sequence Analysis of ALF and TBP Promoter DNA-- To assess whether other promoter-proximal CpGs were methylated we used a sodium bisulfite sequencing assay (33). Bisulfite treatment converts cytosine to uracil but does not affect methylated cytosines. Thus, sequencing reactions will show an adenosine (A) band at the position of an unmethylated cytosine, and a guanosine (G) band at the position of a methylated cytosine. The region tested was the bottom strand of the mouse ALF promoter between 179 and +207. In reactions with liver DNA, we observed 22 bands in the "G" lane; 21 of these corresponded to the position of a CpG dinucleotide in the region tested, while one at the end was due to incorporation of a G residue opposite a cytosine in the PCR primer (Fig. 5A). Interestingly bands in the G lane comigrated with those in the "A" lane, indicating that some but not all DNAs were methylated. Additional experiments were performed with DNA from brain, heart, lung, and muscle, and a representative PhosphorImager trace is shown in Fig. 6A. While other studies have shown some genes to be methylated differently among somatic tissues (29), ALF was always methylated. In contrast, DNA from adult mouse testis was essentially unmethylated except at C⁺¹⁹G and C⁺¹⁷⁰G (Figs. 5A and 6A).

View larger version (56K): [in this window] [in a new window]   Fig. 5.   Methylation of the mouse ALF and TBP genes in liver and testis. A, bisulfite sequence analysis of the ALF promoter (179 to +207, see Fig. 1D) in liver and testis. Bands in the G lane indicate sites of methylation in the original genomic DNA. B, bisulfite sequence analysis of the TBP promoter (+220 to +471, see Fig. 1D) in liver and testis. The TBP gene is numbered with the major somatic initiation site as +1. The maps to the left side show the locations of the testis-specific exon (shaded box) and CpGs (triangles).

View larger version (36K): [in this window] [in a new window]   Fig. 6.   Partial and site-specific methylation in somatic and germ line tissues. A, the top line shows a representative PhosphorImager intensity trace of the G lane of bisulfite sequencing reactions from brain (Br), heart (He), liver (Li), lung (Lu), and muscle (Mu) tissues. CpG sites are numbered above the trace. Aligned beneath are PhosphorImager traces from bisulfite sequencing experiments using whole testis (adult), prepubertal whole testis (8 days old (8 d.o.)), pachytene, round, elongating/elongated spermatocytes, and epididymal spermatozoa. DNA from prepubertal testis and isolated spermatocytes is demethylated at all sites other than C+19G and C+170G. DNA from epididymal spermatozoa is remethylated in a pattern similar to that seen in somatic tissues. The diagram to the left shows the progression of spermatogenesis, and the status of ALF expression in each tissue is shown to the right. B, bisulfite-treated liver DNAs (clones 1-13) were subcloned into pBluescript and sequenced. Black triangles show the position of methylated CpGs. The percentage of methylation at each site ranged from 0 to 77% (%modified). Exon 1 and the C+19G and C+170G dinucleotides are indicated. n.d., no data.

We also examined methylation of the TBP gene. The region tested begins 220 nucleotides downstream of the ubiquitously expressed "somatic cell" promoter and ends 471 nucleotides downstream (Fig. 1D) (14). This region is transcribed in all tissues but also contains a promoter active only in testis. No G bands were seen in reactions with either liver or testis DNA, indicating that none of the 26 CpGs in this region are methylated (Fig. 5B). Thus, unlike ALF, the mouse TBP gene is unmethylated and transcribed in both tissues. This result shows that methylation may play a role at some but not all general transcription factor genes up-regulated in meiotic prophase.

Partial Methylation of the ALF Promoter on Individual DNAs-- The observation that the ALF promoter was partially methylated implies that the patterns present on any individual DNA would vary. To show the extent of this variation, individual PCR products from bisulfite-treated liver DNA were cloned and sequenced. Of 13 isolates, 12 different patterns were observed (Fig. 6B). Methylation at any given CpG ranged from 0 to 77% (0/13 at positions C⁸³, C⁶⁶, and C⁹ to 10/13 at position C⁺¹⁷⁰), and the number of methylated CpGs per clone ranged from 0 to 71% (0/21 for clones 5 and 13 and 15/21 for clone 10). These data are consistent with the degree of methylation estimated by genomic Southern and bisulfite sequencing analyses. For instance, 1 of 13 subclones (7.7%) was methylated at the proximal CCGG site (C⁹⁹), and ~5% of liver genomic DNA was undigestable by HpaII at this site (Fig. 4B). The results imply that there might not be a particular CpG critical for silencing.

C⁺¹⁹G- and C⁺¹⁷⁰G-specific Methylation in Male Germ Cells-- We have previously shown that mouse ALF mRNA is present in germ cells beginning at the pachytene stage of meiosis and continuing through the appearance of elongating/elongated haploid spermatids (3). We therefore wanted to determine how ALF was methylated during germ cell development. To test this point, germ cells were separated at unit gravity over a 1-4% gradient of bovine serum albumin to obtain populations of pachytene spermatocytes, round spermatids, and elongating/elongated spermatids. Bisulfite sequencing showed that DNA from these cells was unmethylated except at two positions, C⁺¹⁹G and C⁺¹⁷⁰G (Fig. 6A). These results are similar to those using DNA from whole testis (Figs. 5A and 6A) but display a lower "background," presumably because somatic or premeiotic cells of the testis are absent. Bisulfite sequencing of DNA from prepubertal testis (8 days old) also showed a hypomethylated, C⁺¹⁹G- and C⁺¹⁷⁰G-specific pattern (Fig. 6A). This was a bit surprising since at this stage germ cells have not yet entered the first cycle of meiosis, and ALF is not yet expressed. Thus, the results show that the ALF promoter is hypomethylated prior to its expression (see "Discussion"). In contrast, DNA from mature epididymal spermatozoa was methylated in a manner indistinguishable from that observed in somatic tissues (Fig. 6A).

The results of these experiments reveal dynamic cell-specific changes in ALF promoter methylation during germ cell differentiation. These changes involve formation of a characteristic C⁺¹⁹G- and C⁺¹⁷⁰G-specific methylation profile early in germ cell development (at least by day 8 postpartum), a continuation of this pattern throughout spermatogenesis, during which time expression of the mouse ALF gene occurs, and reversal to a partially methylated state in mature spermatozoa.

In Vitro Methylation Represses ALF Promoter Activity-- The results of the experiments in Figs. 4, 5, and 6 show that a methylated ALF promoter was present in tissues where the gene is not expressed. To obtain more direct evidence that this modification affects expression, the phALF130 and phALF489 deletion constructs were modified in vitro with the SssI methylase (Fig. 7A) and tested for activity. To show that the constructs were methylated, they were digested with either HpaII or MspI. As shown in Fig. 7A, methylation prevented digestion of pGL3 Control and phALF489 by HpaII (lanes 5 and 11), whereas unmethylated constructs were digested by both HpaII and MspI (lanes 2, 3, 8, and 9). When these constructs were introduced into COS-7 cells, expression from the methylated pGL3 Control vector was 2.5-fold lower than its unmethylated counterpart, while the methylated phALF489 construct was reduced 68-fold (Fig. 7B). Similarly methylation of phALF489 and phALF130 also diminished activity when introduced into 293 cells (~100-fold). We conclude that methylation of the ALF promoter exerts a strong negative effect on expression in these cells and suggest that methylation might inactivate the endogenous ALF promoter in somatic cells.

View larger version (53K): [in this window] [in a new window]   Fig. 7.   Effect of in vitro methylation and azaC treatment on ALF gene activity. A, pGL3 Control and phALF489 were methylated in vitro with SssI methylase. The extent of methylation was assessed by comparing digestion patterns of unmethylated (lanes 1-3 and 7-9) and methylated (lanes 4-6 and 10-12) constructs with HpaII (lanes 2, 5, 8, and 11) or MspI (lanes 3, 6, 9, and 12). B, methylated pGL3 Control, phALF130, and phALF489 constructs were transfected into COS-7 or 293 cells and assayed for activity. The results are shown as the ratio of the activity between the unmethylated and methylated constructs. C, human somatic cell lines (HCC38, 293, NTERA-2, SY5Y, and HeLa) treated with 5-aza-2'-deoxycytidine showed an ALF-specific RT-PCR product in all cell lines except SY5Y (lanes 6-8 and 10). D, mouse somatic cell lines NIH/3T3, NIE/115, and GC-1 also showed an ALF-specific RT-PCR product after azaC treatment (lanes 2, 4, 6, and 8). Amplification of glyceraldehyde-3-phosphate dehydrogenase (G3PDH) was used to normalize template concentrations.

5-Aza-2'-deoxycytidine Activates ALF Expression in Somatic Cell Lines-- If methylation of the endogenous ALF gene promoter causes the gene to be inactive in somatic cells, treatment with azaC should reverse inhibition. AzaC is incorporated into DNA as a non-methylatable cytosine analog and interferes with DNA methyltransferase function (34, 36), and it is presumed that these effects are responsible for its ability to activate gene expression. Five human cell lines, HCC38, 293, NTERA-2, SY5Y, and HeLa, were tested by RT-PCR for the presence of ALF mRNA before and after azaC treatment. The 5'-end primer was located in Exon 1 to avoid detection of the chimeric SALF transcript (data not shown). Prior to treatment, no expression was observed in any of these cells (Fig. 7C, lanes 1-5). However, following 48 h of treatment, an ALF-specific RT-PCR product was detected in four of five lines (HCC38, 293, NTERA-2, and HeLa) (lanes 6-8 and 10), indicating the gene was now active.

Since our methylation analyses used DNA from mouse tissues, we also tested whether azaC treatment would reactivate ALF expression in mouse cell lines. Prior to azaC treatment the mouse NIH/3T3, NIE/115, and GC-1 cell lines did not contain ALF mRNA (Fig. 7D, lanes 1, 3, and 5). Following treatment, however, ALF-specific RT-PCR products were present (lanes 2, 4, and 6). The results of these experiments show that the endogenous ALF gene is silenced by a mechanism that depends on DNA methylation and that this effect can be reversed by treatment with azaC.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Characteristics of the ALF Promoter-- ALF is a germ cell-specific counterpart of the large (/) subunit of general transcription factor TFIIA. The studies here explore mechanisms of differential gene expression in somatic and germ line tissues using the ALF gene as a model. The results show that ALF promoters from two species, mouse and human, display homology in a GC-rich region that extends ~100 base pairs upstream of the start site and that constructs that lack this region (pmALF57 and phALF53) are relatively inactive (Fig. 2B). Furthermore a mutation in a conserved TATA-like TTCAAA element (pmALF216mut) had no effect on promoter activity (Fig. 2B) even though this sequence was stably bound by TFIIA·TBP or ALF·TBP complexes in vitro (Fig. 2C). The experiments demonstrate that the homologous GC-rich sequences required for ALF promoter activity in COS-7 and 293 cells coincide with the short region of homology between mouse and human and that transcription factors available in somatic cell lines such as COS-7 or 293 are sufficient for expression even though the endogenous gene is silent.

The GC-rich ALF promoter resembles those found in housekeeping genes as these tend to be TATA-independent and initiate at multiple sites (37). In addition, testis-specific genes including cyclin A1, Ldh-C, MAGE, Pgk-2, and tH2B also contain promoters that are GC-rich (Refs. 18-23 and 38; for review, see Ref. 39). In fact, some genes that use TATA-dependent promoters in somatic cells switch to GC-rich promoters in testis (39), and very short promoter-proximal sequences are often sufficient to specify testis-specific expression in transgenic mice (e.g. Pdha-2 (187 bp), ACE (91 bp), Ldh-C (100 bp), proenkephalin (116 bp), and Prm1 (113 bp)) (40-44). These studies suggest that a class of GC-rich, TATA-less promoters are uniquely active only in germ cells.

ALF Promoter Methylation and Silencing in Somatic Cells-- Although there is still debate whether methylation controls developmental and tissue-specific patterns of gene expression, our results on the ALF promoter favor the idea that methylation is involved in somatic cell silencing. This is strongly supported by the inactivation of the ALF promoter by in vitro methylation (Fig. 7B) and the ability to activate expression of the endogenous gene with azaC (Fig. 7, C and D). In addition, HpaII sites (CCGG) located in the promoter and Intron 1 are methylated in liver but not testis (Fig. 4B), and all 21 promoter-proximal CpGs tested are partially methylated in brain, heart, liver, lung, and muscle (Figs. 5A and 6A), whereas only two of these sites are methylated in testis.

The process by which particular CpGs are selected for methylation and how this affects expression are interesting issues. Since DNA methyltransferases are not sequence-specific (15), one possibility is that variations might be affected by heterogeneous nucleosomal positioning on chromosomes from different cells or even within the same cell. In support of this idea, mutations in the Arabidopsis DDM and human ATRX genes, both of which encode SWI/SNF-like chromatin-remodeling factors, alter patterns of genomic methylation (45, 46). In an earlier study it was shown that a Rous sarcoma virus long terminal repeat-containing plasmid methylated at levels as low as 7% reduced expression by 67-90% (47). Likewise the partially methylated ALF promoter may recruit methyl-CpG-binding domain proteins and associated histone deacetylases that form an inaccessible nucleosomal configuration that spreads over the entire promoter (16, 48). Although the exact mechanism remains to be elucidated, DNase I hypersensitivity analysis does show the ALF promoter to be less accessible in liver nuclei than it is in testis (Fig. 3).

Two additional observations relate to the effects of methylation. First, azaC did not activate ALF expression in human neuroblastoma SY5Y cells, perhaps because they lack a factor(s) required for expression or because they are insensitive to azaC treatment. Second, a region of the TBP gene downstream of the ubiquitously expressed somatic cell promoter (Fig. 1D) (14) was not methylated at any of the 26 CpG sites examined (Fig. 5B). Since this region contains a testis-specific promoter that is off in somatic cells (Fig. 1D), this result suggests that a regulatory role for methylation may be restricted to genes like ALF that are germ cell-specific, and genes like TBP that are up-regulated in meiotic prophase but also ubiquitously expressed may utilize distinct mechanisms.

ALF Promoter Methylation and Expression during Male Germ Cell Differentiation-- The patterns of ALF methylation in male germ cells highlights several interesting characteristics (Fig. 6A). First, there is the surprising ability to distinguish C⁺¹⁹G and C⁺¹⁷⁰G from other CpG sites. As the distance between these sites (151 nucleotides) is the approximate length of one nucleosomal unit, we speculate that a specific chromatin configuration could be involved. Second, the C⁺¹⁹G- and C⁺¹⁷⁰G-specific pattern of methylation is transient since DNA from epididymal spermatozoa is remethylated at all CpGs. Reprogramming has also been documented at the Pgk-2 gene (49) and, together with our data, suggest a de novo methylation system that acts late in germ cell development. Remarkably remethylation generates a heterogeneous pattern similar to that seen in somatic tissues, implying that each haploid paternal genome will be methylated in a relatively distinct pattern. Finally, germ cell-specific genes such as ApoA1, Pgk-2, and Pdha-2 are selectively demethylated sometime between birth and postnatal day 10 (38, 48-50), and this may also occur at the ALF promoter.

Hypomethylation suggests an association with gene activation (for reviews, see Refs. 17 and 39). Interestingly ALF is hypomethylated in prepubertal testis at postnatal day 8, preceding the time of expression by at least 6 days (Fig. 6A). This may indicate that demethylation results in a chromatin configuration that allows transcription to occur once the appropriate trans-acting factors are available (51). In fact, a DNase I-hypersensitive site indicative of an open chromatin configuration was observed in testis (Fig. 3), and a similar site appears at the demethylated Pgk-2 promoter just before it is expressed (52). It is also possible that demethylation is a consequence of transcription factor binding as several reports have shown that NF-B, Sp1, EBNA-1, and VP16 can prevent methylation and facilitate demethylation (53-57). In either case, the delay in expression until the pachytene stage of meiosis reveals additional constraints that prevent the transcription apparatus from functioning earlier.

Conclusion-- The results of this study reveal dynamic changes in methylation status at the ALF promoter during the course of male germ cell differentiation in which a CpG-specific hypomethylated state precedes gene activation and is reversed prior to fertilization. In addition, the data show that methylation-associated silencing of the ALF gene in somatic cells is reversible by 5-aza-2'-deoxycytidine. Overall the work sets the stage for using ALF as a model to study relationships between DNA methylation, chromatin packaging, and preinitiation complex assembly in male germ cells.

	ACKNOWLEDGEMENTS

We thank Drs. Gail Breen, Santosh D'Mello, and John Minna for cell lines. We thank Ashok Upadhyaya for recombinant proteins for bandshift experiments and Dr. Boning Gao for advice on bisulfite sequencing. We thank JingHong Mu and Ashok Upadhyaya for helpful comments on the manuscript.

	FOOTNOTES

^* This work was supported by grants from the American Cancer Society and The Welch Foundation (to J. D.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The nucleotide sequence(s) reported in this paper has been submitted to the GenBank^TM/EBI Data Bank with accession number(s) AF452125.

To whom correspondence should be addressed: Dept. of Molecular and Cell Biology, University of Texas at Dallas, 2601 N. Floyd Rd., Richardson, TX 75080. Tel.: 972-883-6882; Fax: 972-883-2409; E-mail: dejong@utdallas.edu.

Published, JBC Papers in Press, March 11, 2002, DOI 10.1074/jbc.M200954200

	ABBREVIATIONS

The abbreviations used are: ALF, TFIIA/-like factor; mALF, mouse ALF; hALF, human ALF; TF, transcription factor; TBP, TATA-binding protein; TLF, TBP-like factor; TAF, TBP-associated factor; MAGE, melanoma antigen; azaC, 5-aza-2'-deoxycytidine; BAC, bacterial artificial chromosome; RPA, RNase protection assay; ACE, angiotensin-converting enzyme; RT, reverse transcription; SBLF, stoned B-like factor.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

CiteULike