Chromatin Immunoprecipitation (ChIP) on Unfixed Chromatin from Cells and Tissues to Analyze Histone Modifications

Reagents

100-bp DNA size ladder (Promega)

Agarose

Antisera (affinity-purified)

These should be raised against histone peptides with mono-, di-, or trimethylation or acetylation at a specific lysine/arginine residue of interest. In addition, include a control precipitation with an (IgG) antiserum that is not directed against chromatin proteins.

Cell culture from which nuclei are to be extracted (1 x 10⁷ to 1 x 10⁸ cells are required; see Steps 9-16)

Cell culture medium, appropriate for cells of interest (see Step 10)

ChIP elution buffer

ChIP incubation buffer

Dialysis-lysis buffer

DNA loading buffer (6X)

Ethanol (70%, v/v)

Ethidium bromide solution (20 mg/ml in H₂ O)

Glycogen solution (20 mg/ml) (Roche)

Isopropanol

Liquid nitrogen (for frozen tissue only; see Step 1)

Micrococcal nuclease (MNase; 10 units/µl in 50% [v/v] glycerol) (Amersham Bioscience)

Store in 10- to 20-µl aliquots at �20°C. Each aliquot should be used only once to ensure equal enzyme activity in different chromatin preparations.

MNase digestion buffer

MNase stop solution

NaCl (5 M)

Nuclei preparation buffer I, prechilled on ice

Nuclei preparation buffer II, prechilled on ice

Nuclei preparation buffer III, prechilled on ice

Phenol:chloroform:isoamyl alcohol 25:24:1 (v:v:v)

For extraction of genomic DNA, the phenol should be saturated beforehand with 100 mM Tris-Cl (pH 7.5) and stored at 4°C under 10 mM Tris-Cl (pH 7.5).

Phosphate-buffered saline (PBS), cold

Protein A (e.g., CL-4B Sepharose from Amersham Bioscience) or protein G Sepharose

Protein A and protein G are bacterial cell wall proteins that bind to the Fc region of antibodies. These proteins are covalently coupled to Sepharose. The choice between protein A or protein G Sepharose depends on the nature of the antibody used for ChIP. In general, protein A works best for rabbit polyclonal antisera and for mouse monoclonal antibodies from the IgG2a, IgG2b, and IgG3 subclasses. Protein G Sepharose is preferred for mouse IgG1 monoclonal antibodies and for polyclonal antisera from mouse, rat, sheep, and goat.

Proteinase K (10 mg/ml) (optional; see Step 70)

Sodium butyrate (optional; for analyzing histone acetylation only)

To analyze histone acetylation, we recommend adding sodium butyrate (to a final concentration of 5 mM) to the solutions used for the purification of nuclei and for the preparation of input chromatin. Sodium butyrate prevents loss of histone acetylation via the nonspecific action of endogenous histone deacetylases.

Sodium dodecyl sulfate (SDS; 10 %, w/v)

TBE buffer (1X)

TE buffer (1X, pH 7.5)

Tissue samples (fresh or frozen) from which nuclei are to be extracted (see Steps 1-8)

Trypsin solution (0.05% [w/v]) (Sigma)

Tubing preparation solution I

Tubing preparation solution II

Washing buffer A

Washing buffer B

Washing buffer C

Equipment

Centrifuges:

Bench-top centrifuge with cooling system for 1.5-ml microcentrifuge tubes
 

Centrifuge with a swing-out bucket rotor for 15-ml polypropylene tubes
 

High-speed centrifuge with cooling system and a swing-out bucket rotor for 14-ml polypropylene tubes

^{Dialysis tubing (0.5-mm thick, 10-kDa pore width) (VWR international)}

^{Homogenizer, prechilled on ice}

^{We use a tissue grinder/homogenizer (from BDH) that has a glass mortar (tube) and a pestle with a hard plastic head. The clearance between pestle and mortar is 0.15-0.25 mm.}

^{Horizontal gel electrophoresis tank for agarose gels}

^Ice

^{Magnetic stirrer (see Step 32)}

^{Microcentrifuge tubes (1.5 ml and 2.0 ml)}

^{Chromatin immunoprecipitations and incubations with Protein A (G) Sepharose (Steps 44-69) are performed in microcentrifuge tubes. These tubes may be siliconized beforehand (e.g., with a 2% [v/v] dichloromethylsilane solution) in order to prevent nonspecific association of chromatin and antibodies to the inner walls of the tubes. In our laboratory, we have obtained comparable results with nonsiliconized and siliconized microcentrifuge tubes.}

^{Microscope, inverted light (optional; see Step 8)}

^{Mortar and pestle, prechilled in liquid nitrogen (for frozen tissue only; see Step 1)}

^{Muslin cheesecloth}

^{Prepare the cheesecloth by rinsing with H₂ O and then autoclaving.}

^Parafilm

^{Pasteur pipettes}

^{Polypropylene tubes, 14 ml (e.g., 17 x 100-mm Falcon tubes) and 15 ml (e.g., 17 x 120-mm Falcon conical tubes)}

^{Rotating wheels at 4°C and room temperature}

^{Spectrophotometer}

^{Tray for staining gels (see Step 41)}

^{Universal tubing clamps (5 mm) (Spectrum Laboratories)}

^{UV lamp}

^{Vortex mixer}

^{Water bath set at 37°C}

^METHOD

^{Nuclei Preparation from Tissues and Cells}

^{Steps 1-8 describe the purification of nuclei from tissue, while Steps 9-16 describe the purification of nuclei from cultured cells. To prevent chromatin degradation, all steps of the nuclei purification procedure should be performed on ice, or at 4°C (e.g., precool the centrifuge rotors). In addition, one set of micropipettes should be dedicated only to the preparation of nuclei, chromatin, and ChIP analysis, to avoid DNA contamination. Wear gloves throughout all procedures, and respect the safety rules, especially when handling phenol.}

^{Purification of Nuclei from Tissue (2 h)}

^{1. Dissect fresh tissue (maximum 0.2 g in total) and rinse it in cold PBS.
See Troubleshooting. For many tissue types, frozen tissue (snap-frozen in liquid nitrogen) can be used as well. This tissue should first be crunched into powder in a mortar filled with liquid nitrogen; this powder is used for Step 2. The mortar should be prechilled with liquid nitrogen and the tissue kept constantly under liquid nitrogen.}

^{2. Homogenize the tissue in a prechilled glass homogenizer with 5-10 ml of ice-cold nuclei preparation buffer I, until no clumps of cells persist (~10-20 strokes). Filter the suspension through four layers of muslin cheesecloth moistened beforehand with nuclei preparation buffer I.}

^{3. Transfer the cell suspension to a 14-ml polypropylene tube, and centrifuge the samples in a swing-out rotor (3000g , 5 min, 4°C).}

^{4. Pour off the supernatant, and resuspend the cells in 2 ml of ice-cold nuclei preparation buffer I. Add 2 ml of ice-cold nuclei preparation buffer II, mix gently, and place the tubes on ice a maximum of 5 minutes.
See Troubleshooting.}

^{5. Prepare two new 14-ml polypropylene tubes, each containing 8 ml of ice-cold nuclei preparation buffer III. Carefully layer 2 ml of each cell suspension (from Step 4) onto each 8-ml sucrose cushion. Cover each tube with a piece of Parafilm.}

^{6. Centrifuge the tubes in a prechilled swing-out rotor (10,000g , 20 min, 4°C).
The nuclei will form a pellet at the bottom of the tube, whereas the cytoplasmic components will remain in the top layer. At this step, the nuclear pellet should be white.}

^{7. Carefully take off the supernatant with a Pasteur pipette.
This is a critical step, as the top solution (which contains the detergent IGEPAL CA-630) should not come into contact with the nuclear pellet at the bottom of the tube. One way to achieve this is to remove the supernatant in about three steps, changing the Pasteur pipette each time.
See Troubleshooting.}

^{8. Resuspend the nuclear pellet in 1 ml of MNase digestion buffer, and keep the samples on ice. If possible, MNase digestion (Step 23) should be started immediately. Nuclei can be stored for up to 1 day at 4°C.
At this point, the nuclei can be counted using a microscope slide for counting cells. The number of nuclei obtained per gram of tissue varies according to tissue type. For liver, for example, this protocol yields ~2 x 109 nuclei/g tissue.}

^{Nuclei Preparation from Cultured Cells (2 h)}

^{9. Culture 1 x 107 to 1 x 108 cells. Ensure that the cells are not grown beyond semiconfluency.}

^{10. Rinse the cells in PBS, add 2 ml of trypsin solution (for adhering cells only), and incubate them at 37°C. When trypsination is complete, stop the reaction by adding 5 ml of culture medium to the cells.}

^{11. Divide the cell suspension between two 14-ml polypropylene tubes, and centrifuge the samples in a swing-out rotor (4000g , 5 min, 4°C).}

^{12. Pour off the supernatant, and resuspend the cells in 2 ml of ice-cold nuclei preparation buffer I. Add 2 ml of ice-cold nuclei preparation buffer II, mix gently, and place the tubes on ice a maximum of 5 minutes.
See Troubleshooting.}

^{13. Prepare two new 14-ml polypropylene tubes, each containing 8 ml of ice-cold nuclei preparation buffer III. Carefully layer 2 ml of each cell suspension (from Step 4) onto each 8-ml sucrose cushion. Cover each tube with a piece of Parafilm.}

^{14. Centrifuge the tubes in a prechilled swing-out rotor (10,000g , 20 min, 4°C).
The nuclei will form a pellet at the bottom of the tube, whereas the cytoplasmic components will remain in the top layer. At this step, the nuclear pellet should be white.}

^{15. Carefully take off the supernatant with a Pasteur pipette.
This is a critical step, as the top solution (which contains the detergent IGEPAL CA-630) should not come into contact with the nuclear pellet at the bottom of the tube. One way to achieve this is to remove the supernatant in about three steps, changing the Pasteur pipette each time.
See Troubleshooting.}

^{16. Resuspend the nuclear pellet in 1 ml of MNase digestion buffer, and keep the samples on ice. If possible, MNase digestion (Step 23) should be started immediately. Nuclei can be stored for up to 1 day at 4°C.
At this point, the nuclei can be counted using a microscope slide for counting cells. The number of nuclei obtained per gram of tissue varies according to cell type.}

^{Micrococcal Nuclease (MNase) Fractionation and Purification of Chromatin}

^{Preparation of Dialysis Tubing (2 h including cooling time)}

^{17. Cut the tubing into pieces of convenient length (10-20 cm).}

^{18. Boil the tubes for 10 minutes in 0.5 liter of tubing preparation solution I.}

^{19. Rinse the tubes twice in distilled H₂ O.}

^{20. Boil the tubes for 10 minutes in 0.5 liter of tubing preparation solution II.}

^{21. Allow the tubes to cool down, and store them in tubing preparation solution II at 4°C. Ensure that the tubes are entirely submerged.}

^{22. Before use, wash the tubing twice, inside and out, with H₂ O.
Several batches of dialysis tubing can be prepared and stored at 4°C for several weeks.}

^{MNase Fractionation (30 min)}

^{23. Aliquot the resuspended nuclei (from Step 8 or 16) into two 1.5-ml microcentrifuge tubes (500 µl in each tube).}

^{24. Add 1 µl of MNase enzyme (10 units/µl in 50% [v/v] glycerol) to each tube, and mix gently.}

^{25. Incubate the two tubes in a 37° water bath. One tube should be incubated for 2 minutes; the other tube should be incubated for 5 minutes.
Keep these two digestions separate in subsequent steps, until fractions are chosen for combining (Step 43).}

^{26. Add 20 µl of MNase stop solution to each tube.}

^{27. Chill the samples on ice.}

^{Recovery of Soluble Chromatin Fractions (16 h)}

^{28. Centrifuge the 1.5-ml tubes with the MNase-digested nuclei (from Step 27) to pellet the nuclei (10,000 rpm, 10 min, 4°C).}

^{29. Transfer the supernatant into another 1.5-ml tube. Store it for up to 1 day at 4°C.
This supernatant contains the first soluble fraction of chromatin, S1, which comprises small fragments only. Do not discard the pellet.}

^{30. Carefully resuspend the pellet in 500 µl of dialysis-lysis buffer.
At this stage, we normally proceed to Step 31; however, a more expedient lysis/dialysis procedure used in our laboratory has yielded chromatin fragments of comparable quality. It replaces Steps 31-34 as follows:}

^{i. After resuspending the pellet in 500 µl of dialysis-lysis buffer, place the samples for 1 hour at 4°C.}

^{ii. Centrifuge the samples (10,000 rpm, 10 min, 4°C) in a microcentrifuge.}

^{iii. Proceed with Steps 35-36.}

^{31. Close one side of the washed dialysis tubing (from Step 22) with a universal closure clamp. Transfer the 500 µl of resuspended nuclei (from Step 30) into the dialysis tube, and close the second side with another clamp.}

^{32. Submerge the tube in 1-2 liters of dialysis-lysis buffer. Perform dialysis for 12-16 hours at 4°C with constant mild stirring using a magnetic stirrer.}

^{33. Transfer the dialyzed nuclei into a 1.5-ml microcentrifuge tube.}

^{34. Centrifuge the nuclei (10,000 rpm, 10 min, 4°C) in a microcentrifuge.}

^{35. Transfer the supernatant in a new 1.5-ml microcentrifuge tube. Store it for up to 1 day at 4°C.
This is the second soluble chromatin fraction, S2, comprising the larger fragments of chromatin that were removed from the nuclei during lysis/dialysis (Step 32).}

^{36. Resuspend the pellet in 50 µl of dialysis-lysis buffer. Store it for up to 1 day at 4°C.
This is chromatin fraction P.}

^{Quality Control of Chromatin (3 h)}

^{Chromatin Immunoprecipitation}

^{Incubation of Chromatin with Antiserum (16 h)}

^{44. Mix comparable amounts of the S1 and S2 fractions in a 1.5-ml microcentrifuge tube (4-10 µg in total).
It is important to mix similar amounts of chromatin from the S1 and S2 fractions. This ensures that chromosomal regions that are both less accessible and highly accessible to MNase are present in the input chromatin that will be used for ChIP.}

^{45. Bring the volume to 1 ml with ChIP incubation buffer.}

^{46. Add 5-10 µg of the antibody of choice.
In order to control for nonspecific background signal, it is important to include a control precipitation with an (IgG) antiserum that is not directed against chromatin proteins.}

^{47. Close the tubes, and seal the lids with Parafilm.}

^{48. Rotate the tubes at 20-30 rpm for 12-16 hours at 4°C.
During this incubation time, the antibodies will bind to their specific epitopes.
See Troubleshooting.}

^{Preparation of Protein A (G) Sepharose (45 min)}

^{49. Weigh 0.25 g of protein A (or G) Sepharose beads into a 14-ml polypropylene tube.}

^{50. Add 10 ml of H₂ O, and mix.}

^{51. Centrifuge the tubes for 3 minutes at 1500g in a swing-out rotor, and discard the supernatant.}

^{52. Repeat Steps 50 and 51 four times.}

^{53. Add 1 ml of H₂ O, and resuspend the beads.}

^{54. Distribute 100-µl aliquots into 10 1.5-ml microcentrifuge tubes. Store these aliquots at 4°C.
These aliquots are used for the extraction of antibody-bound chromatin from ChIP experiments (starting in Step 55).}

^{Extraction of Immunoprecipitated Chromatin with Protein A (G) Sepharose (6-7 h)}

^{55. Add 50 µl of protein A (or G) Sepharose (from Step 54) to each tube (from Step 48).}

^{56. Rotate the tubes at 20-30 rpm for 4 hours at 4°C.}

^{57. Centrifuge the tubes at 1500g in a swing-out rotor for 3 minutes.}

^{58. Transfer the supernatant to a 2-ml microcentrifuge tube. Store it at 4°C.
This fraction contains the chromatin that did not link to the antibody (i.e., the "unbound fraction").}

^{59. Resuspend the Sepharose beads in 1 ml of washing buffer A.}

^{60. Transfer the resuspended beads to a 15-ml Falcon tube.}

^{61. Bring the total volume to 10 ml with washing buffer A. Mix briefly.}

^{62. Centrifuge for 3 minutes at 1500g in a swing-out rotor at 4°C. Carefully discard the supernatant.}

^{63. Resuspend the beads in 10 ml of washing buffer B. Briefly mix.}

^{64. Centrifuge for 3 minutes at 1500g in a swing-out rotor at 4°C. Carefully discard the supernatant.}

^{65. Resuspend the Sepharose beads in 10 ml of washing buffer C.}

^{66. Centrifuge for 3 minutes at 1500g in a swing-out rotor at 4°C. Carefully discard the supernatant.}

^{67. To elute the chromatin, resuspend the Sepharose beads in 500 µl of ChIP elution buffer, and transfer the sample to a 1.5-ml microcentrifuge tube.}

^{68. Incubate the samples for ~30 minutes at room temperature on a rotating wheel at 20-30 rpm. After this incubation, centrifuge for 3 minutes at 1500g in a microcentrifuge at room temperature.}

^{69. Carefully transfer the supernatant into a 2-ml microcentrifuge tube. Store it at 4°C.
This tube contains the chromatin eluted from the Sepharose beads (i.e., the "bound fraction").
See Troubleshooting.}

^{DNA Extraction from Precipitated Chromatin (3 h)}

^{70. Add 500 µl of phenol:chloroform:isoamyl alcohol (25:24:1, v:v:v) to the bound (Step 69) and unbound (Step 58) fractions.
The DNA extractions from the immunoprecipitated chromatin fractions might be slightly enhanced by including a proteinase K (PK) digestion step just before phenol:chloroform:isoamyl alcohol extraction. For PK digestion, add PK to each sample (to a final concentration of 100 µg/ml), and incubate the samples for 1 hour at 50°C.}

^{71. Vortex the samples for 30 seconds.}

^{72. Centrifuge the samples at 13,000 rpm (~15,000g ) for 15 minutes in a microcentrifuge.}

^{73. Carefully transfer the upper (aqueous) phase to another 2-ml microcentrifuge tube.}

^{74. Add NaCl to a final concentration of 250 mM.}

^{75. Add 10-20 µg of glycogen, and mix.
Because the DNA concentration in the bound fraction is usually low, we recommend the use of glycogen as a coprecipitator.}

^{76. Add 1 volume of isopropanol, and mix.}

^{77. Keep the samples for at least 2 hours at -80°C.}

^{78. Centrifuge the samples at 13,000 rpm in a microcentrifuge for 30 minutes. Carefully discard the supernatant.}

^{79. Rinse each pellet with 1 ml of 70% (v/v) ethanol.}

^{80. Centrifuge the samples at 13,000 rpm for 5 minutes in a microcentrifuge. Carefully discard the supernatant.}

^{81. Dry the pellets for 5-10 minutes at room temperature, and resuspend each pellet in 10-50 µl of 1X TE buffer. The DNA samples can be stored at 4°C.}

^{Assessment of Precipitated Chromatin}

^{82. Measure the OD₂₆₀ of each sample (from Step 81) in order to calculate how much DNA to use as a template in the subsequent PCR amplification (see PCR-Based Analysis of Immunoprecipitated Chromatin ).
The ratio of the DNA in the bound fraction versus the total starting material (corresponding to the bound and unbound fractions together; this value was obtained at Step 37) indicates the efficiency of the ChIP assay, as it represents the percentage of immunoprecipitated chromatin. In a standard analysis of histone modifications, no more than 15% of the input native chromatin should be precipitated. However, this depends on the nature and the abundance of the histone modifications, and on the characteristics and concentrations of the antibodies used (see Discussion).}

^{TROUBLESHOOTING}

^{Problem: How much tissue (and which kind) must be used to purify enough nuclei for a ChIP experiment?}

^{[Step 1]}

^{Solution: This protocol works well on tissues, such as liver, brain, lung, and placenta, and also on early mammalian embryos. In our laboratory, for instance, we have performed studies on 8.5-9.5 d.p.c. mouse embryos and placentas (Umlauf et al. 2004 ). Some 60 dissected embryos were used for each ChIP experiment. No more than ~0.2 g of tissue should be used for the volumes and tube sizes indicated in the protocol.}

^{Problem: How can the yield of intact nuclei be maximized?}

^{[Steps 4 and 12]}

^{Solution: At Step 4 (or 12) of the nuclei purification procedure, it is critical not to extend the incubation in nuclei preparation buffer II for >10 minutes. For that reason, Step 5 should be initiated after 5 minutes of incubation in order to commence centrifugation (Step 6) at exactly 10 minutes after the addition of nuclei preparation buffer II in Step 4. Longer incubations can greatly reduce the yield of intact nuclei. For many tissues (liver, kidney, placenta), a final concentration of 0.2% of the nonionic detergent IGEPAL CA-630 (in nuclei preparation buffer II) will be enough to lyse the cellular membranes during the 10-minute incubation. However, we recommend testing 0.4% IGEPAL CA-630 for other tissues. For instance, this higher concentration of detergent slightly improves the yield of nuclei from brain and muscle tissues.}

^{Problem: Is it a problem if the top layer (containing the IGEPAL CA-630) comes into contact with the nuclei pellet?}

^{[Steps 7 and 15]}

^{Solution: At Step 7 (or 15), it is essential that no traces of the top layer (containing the IGEPAL CA-630) come into contact with the nuclei pellet (even small traces of IGEPAL CA-630 may aberrantly affect the subsequent digestion of chromatin by MNase). Usually, we remove the top layer and the sucrose cushion from the tube by using Pasteur pipettes. This is done by aspirating from the surface of the solution, while changing the Pasteur pipette very often. If the top layer nevertheless comes in contact with the nuclei pellet, the pellet should be gently rinsed once with 1 ml of nuclei preparation buffer III before proceeding with Step 8 (or 16).}

^{Problem: What can be done if, after the MNase digestion, the chromatin appears to be digested too much or too little?}

^{[Step 43]}

^{Solution: Fractionation of chromatin depends on the batch of MNase used, the concentration of the nuclei in the tube, the time of incubation, and the tissue type from which the nuclei were purified. If one observes too much digestion of the chromatin (i.e., almost all chromatin is digested to mono- and dinucleosome fragments), a lower concentration of MNase should be used. Inversely, in case little material is obtained in the S1 fraction, the amount of enzyme should be increased.}

^{Problem: What can be done to improve ChIP when using chicken antisera that do not bind well to either protein A or protein G?}

^{[Step 47]}

^{Solution: We recommend adding 5 µg of a rabbit anti-chicken antiserum directly after Step 47 for a second precipitation of 3-4 hours, before proceeding with the extraction of the antibody-bound chromatin.}

^{Problem: When an antiserum is used for the first time, how does one verify that the histone modification it is directed against has become enriched in the antibody-bound fraction?}

^{[Step 69]}

^{Solution: This can be done by purifying the histone proteins from the antibody-bound fraction (from Step 69), followed by electrophoresis through acid-urea-Triton gels. After electrophoresis, proteins are Western-blotted to nylon filters, which are immunostained with the antiserum following standard procedures (Gregory et al. 2001 ).}

^DISCUSSION

^{There are different ways to obtain input chromatin. Several groups in the field prepare "cross-linked chromatin," for example, by chemically cross-linking proteins and DNA with specific substances such as formaldehyde. However, usually only a small fraction of the chromatin is precipitated, and this method relies on random shearing after cross-linking, which does not always produce small-enough chromatin fragments at the regions of interest. For this reason, and to be able to conduct experiments on fresh and frozen tissues, we and others have preferred to make use of "native chromatin." In our protocol (Fig. 1 ), the chromatin is fractionated by incubation of purified nuclei with micrococcal nuclease (MNase), an enzyme that cleaves preferentially at the linker DNA between the nucleosomes. By performing partial digestions with MNase, it is possible to obtain native chromatin fragments of, on average, one to five nucleosomes in length (Fig. 2 ). These oligo-nucleosome fragments are purified from the nuclei and are then used to perform ChIP. The choice of native chromatin as the input material for ChIP is advantageous because the epitopes, recognized by the antibody, remain intact during the chromatin preparation. As a consequence, native chromatin tends to give higher levels of precipitation for a specific histone modification than formaldehyde cross-linked chromatin. Because fractionation occurs between the nucleosomes, rather than randomly, precipitated native oligo-nucleosome fragments are also particularly suitable to be used for "ChIP on chip." In a recent ChIP on chip study (Bernstein et al. 2006 ), both native and formaldehyde-cross-linked chromatin were precipitated with antisera against histone modifications. DNA samples extracted from the precipitated chromatin were used as probes to hybridize DNA tiling arrays covering many large chromosomal regio, ns. In this large-throughput study, results obtained with native chromatin were very similar to those obtained with cross-linked chromatin. For locus-specific, smaller-scale studies, amplification by the polymerase chain reaction (PCR) remains the method of choice. Different PCR-based approaches can be used to determine how much DNA is precipitated at a site of interest (see PCR-based Analysis of Immunoprecipitated Chromatin ).}

^{Although ChIP is presently the best methodology to analyze histone modifications at specific chromosomal loci, it has several limitations. First, unlike DNA methylation studies, ChIP does not allow analysis of histone modifications in individual cells or on individual chromosomes. ChIP studies are always performed on populations of (cultured) cells or on tissue samples comprising many cells. Moreover, although sequential precipitations with different antisera can be done (Bernstein et al. 2006 ), or antisera against combinations of different histone modifications can be used, it is not easy to determine whether there are specific combinations of covalent modifications on individual histones at a given locus. Again, this is because many cells are used for chromatin purification and ChIP, and chromatin is usually fractionated into fragments that comprise multiple nucleosomes. Last, it should be noted that quantification of the levels of histone modifications at specific chromosomal loci is difficult to obtain by ChIP, because levels of precipitation do not depend solely on the local abundance of the modification studied. They also depend on the quality of the prepared chromatin, on the specificity and concentration of the antiserum used, and on the global abundance of the histone modification that is being studied. Factors that can influence the outcome of the experiment are (1) the distribution of the histone modifications on the chromosomes, (2) the amount of antiserum used, and (3) the "strength" of the antibodies (i.e., the affinity for their epitope). On the other hand, the efficiency of precipitation of modified histones at a locus of interest greatly depends on whether the modification is common or rare in the genome. For instance, a rare modification (e.g., H3-K4 methylation) gives usually good precipitation at the site where it is present. This can be explained by the fact that, in the ChIP, the quantity of antibody added to the tube is high enough to precipitate all the chromatin that carries that specific modification. However, for a modification that is abundant in the genome, the indicated amount of antibody (5-10 µg) sometimes does not precipitate all the chromatin that has the modification. These different factors should be taken into account when comparing different chromatin immunoprecipitation experiments.}

^{ACKNOWLEDGMENTS}

^{We thank Richard Gregory and David Umlauf for design of methodologies, and Bryan M. Turner and Laura P. O’Neill (Birmingham, UK) for introducing us to ChIP on unfixed chromatin. The CNRS, the Association pour la Recherche sur le Cancer (ARC), and the ESF EuroCORES Programme EuroSTELLS are acknowledged for grant support.}