- 1. Technical NotePhyNexuswww.phynexus.comPerformance Enhancement and Automationof Chromatin Immunoprecipitation (ChIP)Procedures with PhyTip Column TechnologyIntroductionA central element of genome function is the expression of the individual genome elements and the factors that regulate thoseevents. Transcription factors, histone proteins, and covalent modifications of various proteins that interact with DNA arecritical players within the regulation of gene expression events. Due to their far-reaching functional implications, there arecurrently many efforts to study these protein-DNA interactions and their associated covalent modifications in a detailed andcomprehensive manner 1,2. To study these events many investigators perform chromatin immunoprecipitation (ChIP) analysesdirected to specific genomic loci, whereby locus-specific primer sets are used for detection of the sequence of interest byPCR3,4, which has recently been adapted to provide quantitative results by real-time PCR5,6. In addition, DNA microarraytechnology is applied for genome-wide identification of those genomic loci that interact with specific interaction partners7orinteraction partners that possess specific covalent modifications 8. It is clear from the current breadth and scope of effortssuch as these that functional events governing gene regulation will continue to be the subject of intense study for many yearsto come.An essential component to successful ChIP analyses is the ability to conduct high-performance immunoaffinity separations ofcrosslinked protein-DNA complexes, and key to that success is the quality of antibodies applied. For example, the antibodiesmust possess high affinity to achieve maximum sensitivity for the targeted protein-DNA complex and thus minimize falsenegatives, but must also possess very high specificity so as to minimize false positives due to background noise contributed bynon-specific binding events. The optimization process for ChIP antibodies must insure that rigorous wash procedures areapplied so that non-specific binding does not occur with the antibody or the antibody-support resin, yet at the same time thiswashing must generate minimal losses of the target complex. With proper optimization, it has been shown ingenome-wide ChIP studies that false positive error rates can range from 4%7 to 27%9, with more typically quoted false positiverates ranging between 6-10%10,11. False negative error rates though obviously not directly measurable are often quoted onthe basis of retrospective statistical analyses as being ~33%10. In addition, bioinformatics-based approaches have beendescribed for reducing the false positive error rate by identifying those regulatory sequence motifs most likely to be valid12.These models have also been boosted with training data of validated positive and negative sequences13. While this boostingwas shown to provide 40% improvement over the non-boosted models, there were still invariably unaccounted-for falsepositives.
2. Performance Enhancement and Automation of Chromatin Immunoprecipitation (ChIP)Page 2 Procedures with PhyTip Column TechnologyWhile thorough ChIP antibody optimization combined with rigorous bioinformatics can lead to reductions in false positives,there will always be questions as to which positive interactions are true and which are false, thus requiring validation andconfirmation of reported protein-DNA interactions. This situation is eloquently summarized by authors9 that haveperformed extensive work in the area of ChIP analyses:However, despite ever increasingly sophisticated statistical analysis of the genome-wide ChIP-chip data currently beingperformed in yeast10, independent and stringent empirical examination oferror rates will continue to be required tovalidate the global conclusions presented of genome-wide DNA-binding factor targets due to experimental error. These issueswill become more pronounced as genome-wide analysis moves increasingly toward human studies where there are threeorders of magnitude more DNA than in yeast.In view of the current state of ChIP analyses, one area that has not been rigorously optimized is the ChIP separation format.An optimized separation format for ChIP analyses would insure that: The antibody-protein-DNA complex is extracted to the Protein A or Protein G resin with maximum efficiency sothat the highest possible target signal can be obtained. This insures the lowest possible rate of false negatives. The trapped complex is washed in a manner that allows for efficient and complete removal of any non-specificbinders. This should be achieved without excessive washing, transfer steps or the need for repeated ChIP rounds.This insures the lowest possible rate of false positives. The trapped complex is eluted from the resin with extremely high yields and is eluted into a small volume thatmaximizes final target complex concentration. This insures maximum analytical rigor and sensitivity in determiningenrichment factors. Once the complex is eluted from the resin and is released from its crosslinking, the sample is in a form that lendsitself to simple and quantitative DNA cleanup procedures. This further insures the lowest possible rate of false negatives.All of the above steps are easily automated and each sample is subjected to a fresh separation column. This insures maximumprecision and zero carryover from sample-to-sample.PhyNexus PhyTip column technology provides an optimized ChIP separation format that meets the specificationsdescribed above. As shown in the photo below at left, the PhyNexus PhyTip 1000+ (top) and 200+ (bottom) columnproducts each possess a microcolumn of agarose-based resin at the opening of a pipet tip. As shown in the drawing belowat right, the design of the PhyTip column is such that highly efficient capture of the target protein is achieved bymaximizing transport to the agarose beads, which is enhanced through repeated back-and-forth sample exposure cycles.This design also insures exceptional purity for the target protein by reducing the amount of unwashed volume to virtuallyzero, thus minimizing any opportunities for non-specific binding events. Lastly, this same design provides for high-yieldelutions to occur into exceptionally small volumes. All of these features work together to provide dramatically increasedtarget protein concentrations at high levels of purity, all from starting samples as small as a few hundred microliters.PhyTip Column TechnologyPhyTip column technology is a simple and straightforward means of performing fully optimized ChIP separations prior toconventional ChIP, ChIP-chip or validation of individual ChIP-chip interactions all of which can also be easily automated forhigh-throughput procedures. The benefits described above as being provided by the PhyTip column technology areillustrated below for different colon cancer cell lines and different histone modifications associated with the hMLH1 gene. 3. Performance Enhancement and Automation of Chromatin Immunoprecipitation (ChIP) Page 3 Procedures with PhyTip Column TechnologyMaterials and MethodsAntibodies and cell linesHistone H3 antibodies, SDS Lysis Buffer, ChIP Dilution Buffer, and salmon-sperm DNA/Protein A and Protein G agarosewere obtained from Upstate Biotechnology (Lake Placid, NY). RKO and SW480 cell lines were obtained from the AmericanType Culture Collection. Dulbeccos modified growth medium, L-15 medium and fetal bovine serum were obtained fromInvitrogen (Carlsbad, CA). Protein G PhyTip Columns were from PhyNexus, Inc. (San Jose, CA).Cell growth, crosslinking and chromatin preparationRKO and SW480 colorectal cancer cell lines were obtained from American Type Culture Collection. RKO cells weregrown in high-glucose Dulbeccos modified Eagles medium supplemented with 10% fetal bovine serum and 1%penicillin/streptomycin. SW480 cells were grown in L-15 medium supplemented with 10% fetal bovine serum and 1%penicillin/streptomycin in a humidified atmosphere containing 5% CO2 at 37C.EZChIP Assay Kit from Upstate Biotechnology was used for chromatin preparation with some modifications. 1x106 cellswere harvested from each cell line. Proteins were cross-linked to DNA by addition of formaldehyde directly to the culturemedium to a final concentration of 1% for 10 min at room temperature. The cross-linking reaction was quenched by adding0.125 M glycine for 5 min at room temperature. The medium was then removed and cells were washed with 1X PBScontaining a combination of protease inhibitors. The PBS was removed and 0.2X trypsin was added to the cells. After a5-min incubation at 37C, ice-cold 1X PBS containing 10% FBS was added to stop trypsinization. The cells were pelleted, andwashed twice with 1X PBS plus protease inhibitors as above and resuspended in 200 l 1% SDS, 10mM EDTA, 50mM Tris,pH 8.1. The lysate from each cell line was diluted to a final volume of 2 mL in 0.01% SDS, 1.1% TritonX-100, 1.2mM EDTA,16.7mM Tris-HCl, pH 8.1, 167mM NaCl.DNA was sheared by sonication on ice using a Sonifier 250 ultrasonic cell disruptor (Branson Ultrasonics), 40% duty cyclefor 2 minutes at low power. Insoluble material was removed by centrifugation at 12,000 x g for 10 minutes at 4. 20 l ofeach sample was retained as DNA input control for PCR analysis. The sonicated samples were pre-cleared with 80 l ofsalmon sperm DNA/Protein A and Protein G agarose beads (3:1 ratio of Protein A to Protein G) for 1 h at 4C withagitation followed by centrifugation at 4000 x g for 1 minute to pellet agarose beads.The resulting supernatant from each cell line was divided equally into five 400 l fractions. 2.5 l of each antibody wasadded to separate fractions and incubated overnight at 4C. Control samples containing no antibody were also prepared forthe two cell lines. Following the overnight incubation, each sample was divided into two 200 l fractions for analysis usingconventional immunoprecipitation and PhyTip column enrichment.Standard Chromatin ImmunoprecipitationTen microliters of the 3:1 Protein A and Protein G agarose were added to each sample for one hour at 4C with agitationfollowed by centrifugation at 4000 x g for 1 minute to pellet agarose beads. The beads were washed by resuspending in 1mlice-cold Low-Salt Wash Buffer (0.1% SDS, 1% Triton X-100, 2mM EDTA, 20mM Tris-HCl, pH 8.1, 150mM NaCl) followedby centrifugation at 4000 x g for 1 minute. The wash was repeated once with High Salt Wash Buffer (0.1% SDS, 1% TritonX-100, 2mM EDTA, 20mM Tris-HCl, pH 8.1, 500mM NaCl), once with LiCl Wash Buffer (0.25M LiCl, 1% IGEPAL-CA630,1% deoxycholic acid (sodium salt), 1mM EDTA, 10mM Tris, pH 8.1) and two washes with TE Wash Buffer (10mM Tris-HCl,1mM EDTA, pH 8.0). Following the final wash step, the immune complex was extracted from the beads by adding 120 lElution Buffer (1% SDS. 0.1M NaHCO3) at room temperature for 15 minutes. 4. Performance Enhancement and Automation of Chromatin Immunoprecipitation (ChIP)Page 4Procedures with PhyTip Column TechnologyChromatin Immunoprecipitation with PhyTip ColumnsProtein G PhyTip columns (5 l bed volume) were blocked with 30 g/mL salmon sperm DNA in PBS using a singleintake/expel cycle of 190 l at a flow rate of 50 l/min. The samples, wash, and elution buffers were placed in individual wellsof a microtiter plate (200 l per well). Liquid handling steps were performed using the PhyNexus ME-200 computer-controlled multichannel flow controller. All of the actions associated with the ChIP protocol can be completely automatedthrough an MEA Personal Purification System, as shown in Figure 1. Figure 1. PhyNexus MEA Personal Purification SystemCapture of the immune complex was achieved by passing the samples back and forth through the column using an intake/expel volume of 190 l for nine cycles at a flow rate of 150 l/min. The columns were washed with the same series ofbuffers (Low Salt Wash, High Salt Wash, LiCl Wash, TE Wash) as described in the preceding section by passing 190 l ofeach wash solution through the columns for three cycles at a flow rate of 200 l/min. The elution step was performed bypassing 120 l of elution buffer back and forth through the column for five cycles at a flow rate of 50 l/min. This process isillustrated in Figure 2.Figure 2. PhyTip ChIP extraction process 5. Performance Enhancement and Automation of Chromatin Immunoprecipitation (ChIP)Page 5 Procedures with PhyTip Column TechnologyPCR AmplificationAfter elution from the beads and columns, each sample was heated at 65C for four hours to reverse the DNA-chromatincrosslinking, followed by Proteinase K digestion, phenol:chloroform extraction and ethanol precipitation to recover the DNA.Alternatively the DNA can undergo high-yield quantitative purifications in an entirely automated and high-throughput mannerby applying reverse-phase PhyTip columns. A 266 bp sequence of the 5 upstream region of the hMLH1 gene (AB017806)was amplified via PCR. The sequence of the primers used were: Forward: 5 GGCTCCACCAC TAAATAACGCTG, Reverse:5 GCCTCTGCTGAGGTGATCTGG. Amplification was performed using Taq DNA Polymerase (Roche) using the followingprogram: 95C for 5 min (95C for 45 s, 60C for 45 s, 72C for 45 s for 35 cycles), 72C for 10 min, heated lid 105C. PCRproducts were visualized using agarose gel electrophoresis and EtBr staining of amplicons.ResultsAcetylation and methylation patterns of histone H3 surrounding the promoter region of the mismatch repair gene hMLH1 incolorectal cancer cell lines RKO (hMLH1 promoter silenced) and SW480 (hMLH1 promoter active) have been describedpreviously3, and were used for the purpose of comparing results obtained by standard ChIP procedures to those obtained byPhyTip ChIP procedures.Acetylation of Histone H3The first histone modification studied was acetylation of Histone H3, with the gel image shown below in Figure 3.The symbols above the individual lanes signify the absence (-) or presence (+) of antibody. Therefore, all lanes withoutantibody should show no PCR product any PCR product that does show in these lanes would be the result of inadequatewashing of non-specific DNA binding to the resin itself or any other potentially contaminating surfaces. In addition, forthose cases where antibody is present for this particular histone modification, there should be a clear PCR product presentin the case of SW480 cells and no PCR product in the case of RKO cells. Any PCR product that does show in the RKO cellline-plus-antibody lane would be a false positive.The PhyTip ChIP column procedure generates a strong PCR product for the SW480 cell line, with no discernible PCRproduct in any of the other lanes indicating that the wash procedure was sufficient for both cell lines, and that there wasnot a false positive for the RKO cell line. The standard ChIP procedure generates a clearly discernible signal for the SW480cell line, with no discernible signal in the no antibody lane for this same cell line. However, there are faint PCR productsin both lanes of the RKO cell line, indicating that the wash procedure for this cell line was not sufficient and runs the risk ofcontributing to false positive signals, as shown in the with antibody lane for the RKO cell line.Figure 3. Anti-acetyl-Histone H3 ChIP results, standard vs PhyTip processing 6. Performance Enhancement and Automation of Chromatin Immunoprecipitation (ChIP) Page 6 Procedures with PhyTip Column TechnologyDimethylation of Histone H3-K9The second histone modification studied was dimethylation of Histone H3-K9 with the gel image shown below in Figure 4.Just as with the previous histone modification, all lanes without antibody should show no PCR product. Any PCR productthat does show in these lanes would be the result of inadequate washing of non-specific DNA binding to the resin itself orany other potentially contaminating surfaces. In addition, for those cases where antibody is present for this particularhistone modification, there should be a clear PCR product present in the case of RKO cell lines and no PCR product in thecase of SW480 cell lines. Any PCR product that does show in the SW480 cell line-plus-antibody lane would be a falsepositive.The PhyTip ChIP column procedure generates a clearly discernible PCR product for the RKO cell line, with no discerniblePCR product in any of the other lanes, indicating that the wash procedure was sufficient for both cell lines, and that therewas not a false positive for the SW480 cell line. The standard ChIP procedure generates a very faint PCR product for theRKO cell line (and is close to being undetectable, and hence a false negative), with no discernible signal in the no antibodylane for this same cell line. However, there is also a faint PCR product in the no antibody lane for the SW480 cell line(and contrary to what is already known about this histone modification, is actually more discernable than the withantibody lane for the RKO cell line). This indicates that the wash procedure for this cell line was not sufficient and runs therisk of contributing to false positives. Figure 4. Anti-H3-K9-Me2 ChIP results, standard vs PhyTip processingDiscussionAs demonstrated for a known gene (hMLH1) oppositely regulated by two different H3 modifications in two different celllines, PhyTip column technology can provide significant improvements in signal-to-noise for ChIP analyses. This translates toimprovements in false positive error rates by providing more specific signals, as well as improvements in false negative errorrates by providing stronger signals.These improvements in performance have implications in a variety of ChIP-related contexts. For one, conventional ChIPanalyses whereby a set or sets of loci-specific primers will benefit from having an overall higher degree of reliability andautomation. This would be specifically beneficial if a large number of experiments are performed for a particular event, orfor validation of individual genome-wide interactions reported as suggested by previous authors. In addition, theimprovements can potentially lead to overall improvements in false negatives and false positives in ChIP-chip experiments.continued... 7. Performance Enhancement and Automation of Chromatin Immunoprecipitation (ChIP) Page 7 Procedures with PhyTip Column Technology Another advantage of the PhyTip column approach to ChIP analyses is that the improvements described above are obtained with significantly reduced starting sample quantities. In many standard ChIP procedures, the starting number of cells is typically 1x106, which are then lysed and commonly brought to a volume of 2 mL for standard ChIP processing. In the case of PhyTip processing for ChIP analyses, 0.2 mL of lysate volume representing the equivalent of 1x105 cells provided the improvements reported over processing this same volume by standard ChIP procedures. The improved results for smaller amounts of starting material mean that PhyTip ChIP approaches are well-suited to those situations where the starting cells are very precious and limited in quantity, such as primary tumor cell lines. In addition, this advantage also means that larger numbers of different experiments can be performed upon the same sample, e.g. examining a range of different histone modifications within a given sample. In addition to these substantial benefits for ChIP analyses, the PhyTip column approach to ChIP analyses lends itself to very simple automation of the separation process. This creates not only greater precision in ChIP analyses, but also dramatically reduces hands-on time. In addition, the PhyNexus MEA Personal Purification System has accessory items such as on-board heating that can automate the process of reversing protein-DNA crosslinking with 65C incubations, and the ability to also perform standard pipetting operations on the MEA allows for addition of Proteinase K to these de-crosslinked samples. Furthermore, reverse-phase DNA purification can be performed in a high-yield and automated manner on the MEA Personal Purification System, thus eliminating the need for phenol-chloroform extractions requiring labor-intensive manipulation of the samples resulting in losses of DNA. Acknowledgments Drs. Jeffrey Sklar and Hui Li of Yale Medical School are gratefully acknowledged for providing the experimental work and data for this Technical Note. References 1.B.D. Strahl and C.D. Allis, The language of covalent histone modifications, Nature, 2000, 403, 41-45. 2.T. Jenuwein and C.D. Allis, Translating the histone code, Science, 2001, 293, 1074-1080. 3.J.A. Fahrner, S. Eguchi, J.G. Herman and S.B. Baylin, Dependence of histone modifications and gene expression on DNA hypermethylation in cancer, Cancer Res., 2002, 62, 7213-7218. 4.S. Chuikov, J.K. Kurash, J.R. Wilson, B. Xiao, N. Justin, G.S. Ivanov, K. McKinney, P. Tempst, C. Prives, S.J. Gamblin, N.A. Barlev and D. Reinberg, Regulation of p53 activity through lysine methylation, Nature, 2004, 432, 353-360. 5.J.C. Wang, M.K. Derynck, D.F. Nonaka, D.B. Khodabakhsh, C. Haqq and K.R. Yamamoto, Chromatin immunopreciptation (ChIP) scanning identifies primary glucocorticoid receptor target genes, Proc. Natl. Acad. Sci. USA, 2004, 101, 15603-15608. 6.J.V. Geisberg and K. Struhl, Quantitative sequential chromatinimmunoprecipitation, a methodforanalyzing cooccupancy of proteins at genomic regions in vivo, Nucleic Acids Res., 2004, 32, e151. 7.T.H. Kim, L.O. Barrera, M. Zheng, C. Qu, M.A. Singer, T.A. Richmond,Y. Wu, R.D. Green andB. Ren, A high-resolution map of active promoters in the human genome, Nature, 2005, 436, 876-880. 8.B.E. Bernstein, E.L. Humphrey, R.L. Erlich, R. Schneider, P. Bouman, J.S. Liu, T. Kouzarides and S.L. Schreiber, Methylation of histone H3 Lys 4 in coding regions of active genes, Proc. Natl. Acad. Sci USA, 2002, 99, 8695-8700. 9.M.J. Oberley, D.R. Inman and P.J. Farnham, E2F6 negatively regulates BRCA1 in human cancer cells without methylation of histone H3 on lysine 9, J. Biol. Chem., 2003, 278, 42466-42476. 10. T.I. Lee, et al, Transcriptional regulatory networks in Saccharomyces cerevisiae, Science, 2002, 298, 799-804. 11. Z. Li, S. Van Calcar, C. Qu, W.K. Cavenee, M.Q. Zhang and B. Ren, A global transcriptional regulatory role for c-Myc in Burkitts lymphoma cells, Proc. Natl. Acad. Sci. USA, 2003, 100, 8164-8169. 12. E.M. Conlon, X.S. Liu, J.D. Lieb and J.S. Liu,Integratingregulatorymotif discovery and genome-wide expression analysis, Proc. Natl. Acad. Sci. USA, 2003, 100, 3339-3344. 13. P. Hong, X.S. Liu, Q. Zhou, X. Lu, J.S. Liu and W.H. Wong, A boosting approach for motif modeling using ChIP-chip data, Bioinformatics, 2005, 21, 2636-2643.Copyright 2005, PhyNexus, Inc., All Rights ReservedPSL # 90-10-03 September 2005www.phynexus.com