Xu et al. 10.1073/pnas.0704579104. |
Fig. 6. Distribution of lineage-specific gene expression. Expression of Ptcra, Il12b, Alb1, Oct4 and Gapdh is measured by quantitative RT-PCR in ESC lines (CCE and J1), primary thymocytes, the thymocyte cell line VL3-3M2, spleen, liver, bone marrow-derived macrophages stimulated with LPS for 4 h (BMDM-LPS4h), and brain. The relative mRNA expression is calculated from primer-specific standard curves using the iCycler Data Analysis Software. Error bars represent mean SD of three independent PCR amplifications. Ptcra expression in hematopoietic stem and progenitor cells measured by quantitative RT-PCR is described in Attema et al. (1).
1. Attema JL, Papathanasiou P, Forsberg EC, Xu J, Smale ST, Weissman IL (2007) Proc Natl Acad Sci, in press.
Fig. 7. Analysis of the reproducibility of bisulfite genomic sequencing. Bisulfite genomic sequencing is arguably the most widely used technique to detect 5-methylcytosine (5-MeC) in genomic DNA. This technique provides a quantitative measurement of DNA methylation at single-molecule resolution in any sequence context (2). The technique involves bisulfite conversion of genomic DNA, whereby unmethylated cytosine is converted to uracil, but 5-methylcytosine remains unchanged. The target sequence is PCR amplified using specific primers to yield fragments in which all uracil (converted from unmethylated cytosine) and thymine residues are amplified as thymine, and only 5-methylcytosine is amplified as cytosine. Following PCR amplification, methylated cytosines can be detected by direct analysis of the PCR products (3, 4, 5) or a more quantitative profile of methylation can be obtained by cloning the PCR fragments and sequencing individual clones, where each clone represents a single molecule in the DNA sample. In this study, we used the later approach to obtain quantitative measurements of CpG methylation in various cell types and tissues. The PCR fragments were TA-cloned into pCR2.1 vector (Invitrogen) and sequenced using M13 forward (-20) or reverse primers. The bisulfite genome sequencing method was first reported by Frommer et al. (6) and Clark et al. (3). Since that time, the validity of the method has been confirmed in a number of studies. However, as with any method, bisufite sequencing is associated with technical difficulties and potential artifacts (2). Incomplete bisulfite conversion, stochastic PCR amplification, PCR bias, and cloning bias are the major causes of artifacts that may complicate interpretation of bisulfite sequencing data. To address the problem of potential artifacts associated with PCR amplification and subsequent cloning of PCR fragments, and to determine whether we should attach any meaning to the variable results observed at a given enhancer when comparing a large number of different cell populations, we have extensively evaluated the reproducibility of bisulfite sequencing method using primers specific to the Ptcra enhancer and promoter regions and the Il12b enhancer region. Four independent bisulfite sequencing experiments were performed at each of these regions (Experiments 1-4) using genomic DNA isolated from ESC CCE cells. Experiments 1 and 2 were performed with independent PCR amplification, subcloning, and sequencing using the same genomic DNA preparation. Experiments 3 and 4 were performed using genomic DNA isolated from ESC CCE cells 10 months later. Similarly, bisulfite sequencing in primary thymocytes (NS, nonstimulated; PI, stimulated with PMA + ionomycin) and bone-marrow-derived macrophages (BMDM) were repeated twice (thymocytes 1 and 2) and three times (BMDM 1 to 3), respectively. In most cases, more than 10 clones were sequenced to estimate DNA methylation level. The combined results from the two, three, or four independent experiments are also shown. As shown in the color-coded table, the analyses from ESC CCE cells yield results with the same general trends, although the quantitative methylation values exhibit considerable variability in some instances. The windows of unmethylated CpG dinucleotides are readily apparent at both the Ptcra enhancer and Il12b enhancer in all four experiments. Nevertheless, specific variations were observed when the same genomic DNA sample was analyzed in two separate experiments (1 versus 2 and 3 versus 4) and also when different genomic DNA samples were analyzed (compare 1 and 2 with 3 and 4). In general, there appears to be greater variability at regions exhibiting intermediate levels of methylation (e.g., Ptcra -4,318, -4,292, and -3,965). The results obtained with thymocytes are more consistent, suggesting that the bisulfite sequencing method exhibits less variability if the CpGs analyzed are homogenously methylated or unmethylated in the population. In this case, the CpG dinucleotides within the amplified regions are nearly completely unmethylated. An example of the other extreme (nearly complete methylation) can be found at the Il12b enhancer regions (-9,512 and -9,420) in ESC CCE cells. The results obtained with bone marrow-derived macrophages support the general principles described above. That is, greater variability was observed at CpG dinucleotides that appeared to possess intermediate levels of methylation, whereas CpG dinucleotides that appeared to be fully methylated or unmethylated exhibited less variability. From these results, we conclude that the bisulfite genome sequencing method used in this study can generate reliable measurement of CpG methylation at multiple DNA regions. However, the results need to be evaluated with some caution. In particular, CpG dinucleotides that exhibit intermediate methylation levels may yield results that are relatively unreliable. For this reason, when modest variability in bisulfite sequencing results is obtained in a comparison of a number of different cell populations, as in this study, one cannot determine with confidence whether the methylation level at the endogenous locus is truly variable from cell type to cell type. Therefore, it is best to focus on general trends and relatively dramatic differences in the methylation profile when different cell samples are compared.
2. Warnecke PM, Stirzaker C, Song J, Grunau C, Melki JR, Clark SJ (2002) Methods 27:101-107.
3. Clark SJ, Harrison J, Paul C, Frommer M (1994) Nucleic Acids Res 22: 2990-2997.
4. Paul CL, Clark SJ (1996) Biotechniques 21:126-133.
5. Gonzalgo ML, Jones PA 91997) Nuclei Acids Res 25:2529-2531.
6. Frommer M, McDonald LE, Millar DS, Collis CM, Watt F, Grigg GW, Molloy PL, Paul CL (1992) Proc Natl Acad Sci 89:1827-1831.
Fig. 8. The mouse Il12b enhancer contains a window of unmethylated CpG dinucleotides. The data shown in B of this figure are the same as in Fig. 3A but include the ratio values of methylated clones to total clones analyzed. (A) A diagram of the Il12b locus is shown. The coding regions and untranslated regions are shown as black and white boxes, respectively. The transcription start site (+1, arrow) was determined according to the published Il12b mRNA sequences (NM_008352). The enhancer (HSS1; 7) and promoter regions of the Il12b gene are marked. Regions selected for bisulfite sequencing are indicated by the two-headed arrows. (B) DNA methylation profiles are shown for the Il12b enhancer and promoter in various embryonic and somatic cells or tissues. Colors indicate methylation levels as described in the legend to Fig. 1.
7. Zhou L, Nazarian AA, Xu J, Tantin D, Corcoran LN, Smale ST (2007) Mol Cell Biol 27:2698-2712.
Fig. 9. The mouse Alb1 enhancer contains narrow windows of unmethylated CpG dinucleotides. The data shown in panels A and B of this figure are the same as in Fig. 3 B and C but include the ratio values of methylated clones to total clones analyzed. (A) DNA methylation profiles are shown for the Alb1 enhancer in various embryonic and somatic cells or tissues. (B) Results from five independent bisulfite sequencing experiments in ESC CCE cells are shown. (C) An aliquot of the cells (ESC CCE, EB, and VL3-3M2) used for bisulfite sequencing was stained with PE-conjugated mouse anti-SSEA-1 antibody and analyzed by flow cytometry. (D and E) Expression of Oct-3/4 and SSEA-1 is high in virtually all undifferentiated ESC and low in differentiated day-6 EBs, as shown by immunofluorescent microscopy.
Fig. 10. DMS genomic footprinting of the Ptcra enhancer region was performed multiple times to determine the validity of the results obtained. Two independent experiments analogous to the experiment in Fig. 4 are shown. Nucleotides protected from DMS modification or exhibiting enhanced modification are labeled as described in the legend to Fig. 4. Footprinting experiments were performed with ESC and EB, as well as with VL3-3M2 thymocytes (a transformed mouse double-positive thymocyte line [see main text]) that were either unstimulated (Ptcra+) or stimulated with PMA + ionomycin (Ptcra-). To identify background bands, the procedure was also performed with ESC but without the addition of piperidine to cleave adjacent to DMS-modified base-pairs.
Fig. 11. Analysis of histone modifications at the Ptcra locus. (A) The Ptcra locus was analyzed using chromatin prepared from non-stimulated (NS) and PMA + ionomycin-stimulated (PI) thymocytes, spleen, and liver cells. Antibodies directed against acetyl histone H3-K9, dimethyl H3-K4, and unmodified histone H3 were used. Precipitated DNA was quantified by real-time PCR using primer pairs specific to the indicated control regions relative to the Ptcra transcription start site. The abundance of each region was plotted relative to the input DNA (% INPUT). A diagram of the Ptcra locus is also shown at the top. (B) The Ptcra locus was analyzed by ChIP using chromatin prepared from the ESC (CCE) line. Antibodies were directed against acetyl H3-K9, dimethyl H3-K4, trimethyl H3-K4, trimethyl H3-K27, unmodified histone H3, and GST (as a negative control). Precipitated DNA was quantified and plotted as described above. The data are representative of experiments from three to five independent chromatin preparations.
Fig. 12. Analysis of histone modifications at the Il12b locus. (A) The Il12b locus was analyzed by ChIP using chromatin from nonstimulated (NS) and LPS-stimulated (LPS-4h) BMDMs. Antibodies directed against acetyl histone H3-K9, dimethyl H3-K4, trimethyl H3-K27, and unmodified histone H3 were used. Precipitated DNA was quantified by real-time PCR and was plotted relative to input DNA (% INPUT) as described in the legend to SI Fig. 11. A diagram of the Il12b locus is shown at the top. (B) The Il12b locus was analyzed by ChIP using chromatin from the ESC (CCE) line. Antibodies were directed against acetyl H3-K9, dimethyl H3-K4, trimethyl H3-K4, trimethyl H3-K27, unmodified histone H3, and GST (negative control). Precipitated DNA was quantified and plotted as described above. The data are representative of experiments from three to five independent chromatin preparations.
Fig. 13. Marks of active and repressed chromatin at promoters for tissue-specific genes and constitutively active genes. As controls for the experiments shown in SI Figs. 11 and 12, ChIP experiments were performed using chromatin prepared from undifferentiated ESC, primary thymocytes (T cells), and mature bone marrow-derived macrophages (BMDM). Antibodies directed against acetyl histone H3-K9 (Ac-H3K9), dimethyl H3-K4 (2Me-H3K4), trimethyl H3-K27 (3Me-H3K27), unmodified histone H3, and GST (as a negative control) were used. Precipitated DNA was quantified by real-time PCR, using primer pairs specific to the promoters of Oct4, Sox2, Math1, Sox1, Nkx2-2, Irx2, HoxA3, Ikaros, and Tnrc5 genes. The abundance of each region was plotted relative to the input DNA (% INPUT). The promoters for Math1, Sox1, Nkx2-2, Irx2, and Ikaros were found to be associated with both H3-K27 and H3-K4 methylation in undifferentiated ESC, as previously reported (8, 9). However, the HoxA3 promoter was associated with H3-K27 methylation, but not H3-K4 methylation. The promoter for the constitutively active gene Tnrc5 was associated with all active histone modifications, but not with H3-K27 methylation.
8. Azuara V, Perry P, Sauer S, Spivakov M, Jorgensen HF, John RM, Gouti M, Casanova M, Warnes G, Merkenschlager M, Fisher AG (2006) Nat Cell Biol 8:532-538.
9. Bernstein BE, Mikkelsen TS, Zie X, Kamal M, Huebert DJ, Cuff J, Fry B, Meissner A, Wernig M, Plath K, et al. (2006) Cell 125:315-326.
Fig. 14. Unmethylated windows emerge at the Ptcra enhancer following stable integration of premethylated constructs into ESC, but not thymocytes. The clone-by-clone data shown in this figure are summarized in Fig. 5. (A) A diagram of the Ptcra enhancer-promoter-GFP plasmid used for stable transfection assays is shown. The 369-bp enhancer and 511-bp promoter fragments from Ptcra were inserted upstream of a destabilized enhanced green fluorescent protein (EGFP) reporter (pd2EGFP; Clontech). A puromycin-resistant cassette was subcloned in the same construct. Two 1.2 kb chicken b-globin insulators were inserted upstream of the Ptcra enhancer and downstream of the EGFP polyA signal. The cloned Ptcra enhancer contains 7 CpG dinucleotides (-4,130 to -3,900). Putative transcription factor binding sites are shown as shaded boxes. Plasmids were linearized with XmnI, methylated with SssI CpG methylase, and electroporated into the ESC CCE line, VL3-3M2 double-positive thymocytes, and EL4 thymocytes, together with pQCIXN (Clontech) containing the neomycin-resistance gene. Unmethylated constructs were also electroporated separately as a control. Cells were selected with neomycin (ESC CCE) or puromycin (VL3-3M2 and EL4). Bisulfite sequencing analyses were performed with the premethylated plasmids and individual neomycin- or puromycin-resistant clones. (B) DNA methylation profiles are shown for five stable clones obtained following transfection of the unmethylated plasmid into ESC CCE cells. Each horizontal line represents an independently sequenced template with methylation at each CpG dinucleotide indicated by a filled circle. (C) Methylation profiles are shown for seven clones obtained following transfection of the premethylated plasmid into ESC CCE cells. (D) The methylation profile of the premethylated plasmid before transfection of ESC CCE cells is shown, confirming that CpG methylation by the SssI methylase was complete. (E and G) Methylation profiles are shown for eight VL3-3M2 clones and four EL4 clones transfected with the unmethylated plasmid. (F and H) Methylation profiles are shown for 20 VL3-3M2 clones and six EL4 clones transfected with the premethylated plasmid.
Fig. 15. GFP expression from unmethylated and premethylated insulator-Ptcra enhancer-promoter-reporter plasmids following stable transfection into ESC and VL3-3M2 cells. (A) GFP expression was monitored by flow cytometry in untransfected VL3-3M2 cells and in representative clones stably transfected with the unmethylated or premethylated insulator-Ptcra enhancer-promoter-reporter plasmid. (B) GFP expression was monitored by flow cytometry in untransfected ESC CCE cells and in representative clones stably transfected with the unmethylated or premethylated insulator-Ptcra enhancer-promoter-reporter plasmid.
Fig. 16. DNA methylation profiles of the Ptcra locus. (A) Diagram of the Ptcra locus and adjacent CpG island-containing genes. The transcription start site (arrows), coding regions (black boxes), untranslated regions (white boxes), and regions selected for bisulfite sequencing (two-headed arrows) are indicated. (B) Ptcra DNA methylation profiles are shown for the ESC CCE line, nonstimulated (NS) and PMA/ionomycin-stimulated (PI) thymocytes, spleen, and liver cells. The percent methylation at each CpG was calculated from the number of methylated clones divided by the total clones sequenced (also shown as a ratio). Methylation levels are represented in a gradation of colors: dark green (0-20%), light green (21-40%), yellow (41-60%), orange (61-80%), and red (81-100%). The data shown in B of this figure are the same as in Fig. 1 but include the ratio values of methylated clones to total clones analyzed.
Fig. 17. The mouse Ptcra enhancer contains windows of unmethylated CpG dinucleotides in sperm, blastocysts, ESC, HSC, hematopoietic progenitors, and non-hematopoietic tissues. The data shown in A of this figure are the same as in Fig. 2B, but include the ratio values of methylated clones to total clones analyzed. (A) DNA methylation profiles are shown for the Ptcra enhancer and promoter, as well as for the flanking CpG islands, in various embryonic and somatic cells or tissues. Only the first four CpG dinucleotides within the Rik23 and Tnrc5 CpG islands are shown (see Fig. 1 and SI Fig. 16). Colors indicate methylation levels as described in the legend to Fig. 1. (B) Methylated CpG dinucleotides within the Ptcra enhancer were randomly distributed among individual clones analyzed by bisulfite sequencing. Bisulfite sequencing results obtained with three different primer pairs (-4.3 kb, -4.1 kb, and -3.9 kb) are shown in a clone-by-clone manner. Results are shown for HSC, MPP, CLP, and CMP. Methylated (filled circles) and unmethylated (open circles) CpG dinucleotides are shown for a number of independently sequenced templates (horizontal lines).
SI Materials and Methods
Bisulfite Sequencing.
Bisulfite treatment of isolated preimplantation embryos was carried out essentially as described (1). Bisulfite treatment of DNA from adult tissues or cultured cells was performed as described (2, 3) with modest modification. Briefly, 1-2 mg of presheared genomic DNA in 50 ml of TE was denatured by adding 5 ml of 3 M NaOH and incubated for 15-30 min at 37°C. For bisulfite treatment, we added 510 ml of 40.5% sodium bisulfite (final concentration 3.3 M) and 30 ml of 10 mM hydroquinone (final concentration 0.5 mM), followed by incubation for 8-16 h at 55°C. Desalting was carried out using QIAquick PCR purification kit (Qiagen), and the eluted DNA (in 50 ml of double distilled H2O) was desulfonated by treatment with 5.5 ml of 3 M NaOH for 15 min at 37°C. DNA was precipitated by the addition of 2 mg of yeast tRNA, 35 ml of 5 M ammonium acetate, pH 7, and 230 ml of ethanol. After centrifugation, the precipitated DNA pellet was resuspended in 50 ml of TE buffer and stored at -20°C until use.
Sequence-specific PCR of the bisulfite-treated DNA was performed by using primers specific to the murine Ptcra locus, Il12b enhancer and promoter, and Alb1 enhancer. PCR was carried out in a volume of 50 ml containing 1´ PCR buffer (Invitrogen), 2.5 mM MgCl2, 1 mM forward and reverse primers, 200 mM dNTPs, and 1 unit of TaqDNA polymerase (Invitrogen). Nested or seminested PCR was performed when the starting DNA was limiting. PCR conditions were: first PCR, 94°C 3 min, 94°C 1 min, 54°C 1 min, 72°C 3 min for 5 cycles, followed by 94°C 30 sec, 54°C 30 sec, 72°C 1 min for additional 25 cycles; second PCR, 94°C 30 sec, 55-56°C 30 sec, 72°C 30 sec for 30 cycles. The PCR fragments were cloned into the pCR2.1 vector (Invitrogen) and transformed into DH5a E. coli cells. Miniprep plasmid DNA was verified by EcoRI restriction analysis and the positive clones were sequenced using M13 forward (-20) or reverse primers. DNA sequencing was performed by the UCLA Sequencing and Genotyping Core Facility and the University of Washington High-Throughput Genomics Unit.
Genomic Footprinting.
Cells (1 ´ 108) were treated with 0.1% DMS (Sigma-Aldrich) in 1 ml of growth media for 1 min at 37°C, followed by immediate rinse with 50 ml of ice-cold PBS. The cells were lysed by the addition of 3 ml of cell lysis buffer (1 mM Tris×HCl, pH7.5/400 mM NaCl/2 mM EDTA/0.2% SDS/0.2 mg/ml proteinase K). This mixture was incubated for 3-5 h at 37°C with periodic mixing. After phenol extraction and chloroform extraction, DNA was precipitated with isopropanol. As a control, DNA extracted from untreated ESC CCE cells was treated in vitro with 0.1% DMS as described (4). Briefly, DNA (100 mg) was treated with 2 ml of 10% DMS in a volume of 202 ml at room temperature for 2 min. The reaction was stopped by the addition of 50 ml of ice-cold DMS stop buffer (1.5 M sodium acetate, pH7.0/1 M 2-mercaptoethanol/100 mg/ml yeast tRNA), followed by the addition of 2.5 volumes of ethanol on dry ice. DNA was precipitated, air-dried, and resuspended in 200 ml of 1 M piperidine in double distilled H2O for 15 min at room temperature.
After DMS treatment, purified genomic DNA from each cell type was cleaved at all methylated guanines by incubation in 200 ml of 1 M piperidine for 30 min at 90°C. The piperidine was removed by lyophilization, and the cleaved DNA pellets were resuspended in 360 ml of TE buffer (10 mM Tris×HCl/1 mM EDTA, pH7.5). Residual piperidine was removed by two successive ethanol precipitations. The resulting DNA pellets were resuspended in double distilled H2O to final concentration of 1 mg/ml.
Chemically modified and cleaved DNA was amplified by LM-PCR as described in ref. 5 with the following modifications: 2 mg of DNA was used for first-strand synthesis, followed by 21-23 cycles of PCR amplification with a profile of 95°C for 1 min, 63°C for 2 min, and 76°C for 5 min plus 15 sec for each additional cycle. The labeling PCR consisted of two rounds of PCR with 95°C for 1 min, 68°C for 2 min, and 76°C for 10 min. The following primers complementary to the coding strand of the Ptcra gene were used: Primer T1, 5'-ATAGTGGGGTGAAGCTGGCATAAGA; Primer T2, 5'- ACTTTCCTGCCCTCTCCTGACCTTG; Primer T3, 5'- CTCTCCTGACCTTGCGCAAGGACCCAGGTT. The footprinting reactions were separated on 6% denaturing polyacrylamide gels and visualized on a Typhoon 9410 PhosphorImager (Amersham Pharmacia Biosciences).
Immunofluorescence and Flow Cytometry.
Cells (2 ´ 105) were attached to coverslips, washed once in PBS, fixed for 10 min in cold 5% paraformaldehyde in PBS (0.05 g/ml), permeabilized and blocked for 20 min in PBS containing 10% FBS/0.1% saponin/0.2% Triton X-100. The cells were then incubated for 1 h with anti-Oct-3/4 monoclonal antibody (BD Transduction Laboratories) diluted 1:200 in block solution, or anti-SSEA-1 monoclonal antibody (Chemicon) diluted 1:50 in block solution. The cells were washed 3 times (5 min each) in PBS containing 0.2% Triton X-100, then incubated for 30 min with FITC-conjugated goat anti-mouse IgG (Jackson Immunoresearch, West Grove, PA) diluted 1:500 in block solution. The cells were washed as above and mounted in Vectashield with 0.1 mg/ml DAPI (Vector Vector Laboratories, Burlingame, CA). Slides were analyzed on a Leica Inverted Confocal Microscope. For flow cytometry, phycoerythrin (PE)-conjugated mouse anti-SSEA-1 (R&D Systems, Minneapolis, MN) was used and the staining protocol provided with the antibody was followed.
ChIP.
ChIP assays were performed as described in ref. 6. For quantitative real-time PCR, the ChIP samples were analyzed in duplicate with the iQTM SYBR Green Supermix (Bio-Rad, Hercules, CA), using iCycler iQTM Real-Time PCR Detection System (Bio-Rad). Primers were designed to amplify sequences every 1 kb through the 20 kb Ptcra locus and every 1-3 kb through the Il12b locus. All primers were tested for PCR efficiency as recommended by the manufacturer (Bio-Rad). A standard curve was prepared for each set of primers using serial titration of the input DNA. The relative amount of precipitated chromatin (% of INPUT) was calculated from primer-specific standard curves using the iCycler Data Analysis Software. The graph was plotted using Microsoft Excel. The following antibodies were used for the ChIP experiments: Ac-H3K9 (Upstate; catalog no. 06-942), 2Me-H3K4 (Upstate; catalog no. 07-030), 3Me-H3K4 (Upstate; catalog no. 07-473), 3Me-H3K27 (Upstate; catalog no. 07-449), histone H3 (Abcam; catalog no. ab1791-100), FoxD3 (Chemicon; catalog no. AB5687) and GST control antibodies (prepared by our laboratory).
1. Millar DS, Warnecke PM, Melki JR, Clark SJ (2002) Methods 27:108-113.
2. Frommer M, McDonald LE, Millar DS, Collis CM, Watt F, Grigg GW, Molloy PL, Paul CL (1992) Proc Natl Acad Sci 89:1827-1831.
3. Clark SJ, Harrison J, Paul C, Frommer M (1994) Nucleic Acids Res 22:2990-2997.
4. Maxam AM, Gilbert W (1980) Methods Enzymol 65:499-560.
5. Weinmann AS, Plevy SE, Smale ST (1999) Immunity 11:665-675.
6. Su RC, Brown KE, Saaber S, Fisher AG Merkenschlager M, Smale ST (2004) Nat Genet 36:502-506.