Analyzing Chromatin Structure and Transcription Factor Binding in Yeast

METHODS: A Companion to Methods in Enzymology 15, 295–302 (1998) Article No. ME980633

Analyzing Chromatin Structure and Transcription Factor Binding in Yeast Philip D. Gregory, Slobodan Barbaric,1 and Wolfram Ho¨rz2 Institut fu¨r Physiologische Chemie, Universita¨t Mu¨nchen, Schillerstrasse 44, 80336 Munich, Germany

The study of chromatin, once thought to be a purely structural matrix serving to compact the DNA of the genome into the nucleus, is of increasing value for our understanding of how DNA functions in the cell. This article provides two basic procedures for the study of chromatin in vivo. The first is a DNase I-based method for the treatment of isolated nuclei to resolve the chromatin structure of a particular region; the second employs dimethyl sulfate footprinting of whole cells in vivo to determine the binding of factors to cis elements in the locus of interest. Specific examples illustrating the techniques described are given from our work on the regulation of the yeast PHO8 gene, but have also been successfully and reliably applied to the study of many other yeast loci. These procedures make it possible to correlate the binding of a transactivator with an altered or perturbed chromatin organization at a specific locus. © 1998 Academic Press

The genetic processes of the eukaryotic cell must function with a dynamic matrix of DNA, histone, and nonhistone proteins termed chromatin (1). Perhaps the best-studied examples of chromatinmediated effects come from the study of gene expression. The regulation of transcription in vivo is provided an additional dimension by the chromatin organization of the regulatory unit (2, 3). A transactivator recognition element positioned within a nucleosome, for example, can determine both the ability of the factor to bind to this element (4) and its requirement for machines able to bring about chro1 Present address: Laboratory of Biochemistry, Faculty of Food Technology and Biotechnology, University of Zagreb, Pierottijeva 6, 10000 Zagreb, Croatia. 2 To whom correspondence should be addressed. Telefax: 49 89 –5996 440. E-mail: [email protected].

1046-2023/98 $25.00 Copyright © 1998 by Academic Press All rights of reproduction in any form reserved.

matin remodeling to assist in transcriptional activation (5). Furthermore, recent work has identified a plethora of transcriptional coactivators with intrinsic histone acetylation and deacetylation activities formally connecting the covalent modification of chromatin with the proper regulation of transcription (6, 7). Determining the specific chromatin structure of a given promoter or enhancer is therefore of importance in the understanding of the factors and mechanisms that regulate its expression. Here we describe our methods for the isolation of nuclei and the use of DNase I to map the positions of nucleosomes and nucleosome-free (DNase I-hypersensitive) sites that we have successfully employed to study the chromatin structures of the yeast PHO5 (8, 9), PHO8 (10), and TDH3 (11) promoters. To provide a complementary approach, we also describe a protocol for in vivo footprinting using dimethyl sulfate (DMS), which we have employed to determine the binding of the Pho4 transactivator to its sites within a chromatin context in vivo (4, 12, 13). Since the nucleosome and transacting factors compete for occupancy of the DNA this technique can demonstrate binding to sites in vivo and potentially, as has been possible with Pho4, allow the correlation of an altered nucleosomal structure with the binding of the transactivator (4).

METHODS Isolation of Yeast Nuclei Several alternative methods are available for the preparation of nuclei. The method provided here aims to balance speed and purity and is based on the 295

296

¨ RZ GREGORY, BARBARIC, AND HO

original method of Wintersberger et al. (14). There are methods that generate nuclei of appreciably higher purity, which are, however, significantly more time consuming (15, 16), and also faster methods that treat nuclei in crude lysates (17–20). The technique described in detail below provides nuclei of sufficient purity to allow a high degree of reproducibility between experiments. Protocol Grow a 1-liter yeast culture to early logarithmic phase (2– 4 3 10 cells/ml) and collect the cells by centrifugation (3000g/10 min). This is approximately an optical density at 600 nm of 2– 4, which should provide in the order of 0.2 mg of DNA. The measurement of cell density may vary between spectrophotometers; thus, calibration of a particular spectrophotometer is advisable. Cells should be kept cold and processed immediately after removal from the incubator as excessive storage of the cells on ice may lead to difficulties in achieving a suitable degree of cell lysis. Wash the cells by resuspension in 50 ml ice-cold water, transfer to preweighed 50-ml centrifuge tubes, and centrifuge at 3000g for 5 min at 4°C to sediment the cells. Discard the wash water and determine the wet weight of cells. Suspend the pellet in 2 vol (relative to wet weight of cells) of preincubation solution (0.7 M b-mercaptoethanol, 2.8 mM EDTA) and incubate at 28°C with shaking for 30 min. This step facilitates digestion of the cell wall with Zymolyase. It may be necessary to maintain induction or repression by adding the appropriate inducer or repressor or adjusting the incubation temperature for temperature-sensitive strains (21). Collect the cells by centrifugation at 3000g for 5 min at 4°C and wash in 50 ml cold 1 M sorbitol. Pellet the cells again (3000g, 5 min, 4°C) and resuspend in 5 vol (relative to wet weight of cells) lysis solution (1.0 M sorbitol, 5 mM b-mercaptoethanol). Determine the optical density at 600 nm of the resulting cell suspension by diluting 10-ml aliquots 200-fold in water. Absorbance should be in the range between 0.5 and 1. Add 1/50 vol of a freshly prepared Zymolyase solution [20 mg Zymolyase 100T (ICN) dissolved in 1 ml water] and incubate with gentle agitation at 28°C. After 15 and 30 min determine the optical density at 600 nm as before. As a relative measure of lysis values should drop to 5–20% of the original measurement. Lysis is, however, strain and growth stage dependent; for example, stationary cells are more difficult to lyse than logarithmically growing cells. We have successfully used nuclei ob-

tained from cells that gave 60% of the starting absorbance at the end of the Zymolyase treatment. In such cases it is advisable to monitor a constitutively accessible restriction site in chromatin as a control (22). After Zymolyase treatment centrifuge at 2000g for 5 min at 5°C, wash the cell pellet in 50 ml 1 M sorbitol, and collect the cells again by centrifugation at 3000g for 10 min at 5°C. Resuspend the cells in 7 ml Ficoll solution [18% (w/v) Ficoll, 20 mM KH2PO4 (pH 6.8), 1 mM MgCl2, 0.25 mM EGTA, 0.25 mM EDTA] per 1 g cells (original wet weight). Cells lyse at this stage, but nuclei are stabilized by the Ficoll. Distribute aliquots equivalent to 0.5 or 1 g wet wt of cells (suitable for subsequent digestion experiments) into 10-ml polypropylene centrifuge tubes and collect nuclei at 30,000g for 30 min at 5°C. Discard the supernatant, freeze the nuclear pellet in liquid nitrogen or a dry ice/ethanol bath, and store at 270°C. DNase I Analysis of Chromatin Structure in Isolated Nuclei The wrapping of the DNA around the histone octamer to form a nucleosome is sufficient to inhibit DNase I cleavage of the DNA over the core particle and bias the digestion to the linker regions between nucleosomes (1). In the method described below only double-stranded cuts are scored. Alternative methods exist for nucleosome mapping that employ a primer extension protocol and score single-stranded cuts. In these cases, however, the optimum extent of DNase I digestion is lower than that used here. We have employed the procedure described here to observe chromatin transitions at the PHO5 (8), PHO8 (10), and TDH3 (11) promoters in yeast. DNase I analysis identifies regions of hypersensitivity that are indicative of the active state, exemplified by the active PHO8 promoter shown in Fig. 2, and may also demonstrate the presence of positioned nucleosomes across a region, e.g., the PHO5 promoter under repressing conditions (8). We have also successfully employed micrococcal nuclease as an alternative enzyme for probing chromatin structure (8). However, this nuclease exhibits strong sequence specificity (23) and necessitates the careful and extensive use of free-DNA controls throughout any investigation. We therefore consider DNase I the enzyme of choice. A schematic of the protocol is given in Fig. 1, and an example of a typical result of this assay with respect to the chromatin transition at the PHO8 promoter is shown in Fig. 2.

CHROMATIN ANALYSIS AND FACTOR BINDING IN YEAST

Protocol Suspend a nuclear pellet from approximately 500 mg cells (wet wt) in 3 ml digestion buffer [15 mM Tris–HCl (pH 7.5), 75 mM NaCl, 3 mM MgCl2, 0.05 mM CaCl2, 1 mM b-mercaptoethanol] by vortexing. Recover the nuclei by centrifugation at 2000g for 5 min at 5°C, and resuspend in digestion buffer in a total volume of 1.2 ml. Transfer aliquots of 200 ml to six microfuge tubes. Dilute DNase I in DNase I dilution buffer [10 mM Tris– HCl (pH 7.4), 0.1 mg/ml bovine serum albumin (BSA)], and add four different concentrations in the range 0.5 to 20 U/ml final concentration to the nuclear suspensions and incubate for 20 min at 37°C. Keep one sample on ice and one at 37°C without nuclease to control for the extent of digestion due to endogenous nucleases. Terminate digestion by adding 10 ml of 1 M Tris–HCl (pH 8.8), 4 ml 0.2 M EDTA (pH 8.0), 5 ml 20% SDS, and 20

297

ml proteinase K solution [10 mg/ml proteinase K dissolved in 10 mM Tris–HCl (pH 8.0)] and incubate the samples for 30 min at 37°C. Purify total DNA by the addition of 1/5 vol 5 M NaClO4 and phenol/chloroform/isoamyl alcohol extraction followed by chloroform extraction. Remove the supernatant and recover the nucleic acids by the addition of 2.5 vol ethanol. After centrifugation at maximum speed for 15 min at 4°C, resuspend the pellet in 125 ml TE, add 10 ml DNase-free RNase solution [5 mg/ml ribonuclease A in 5 mM Tris– HCl (pH 7.5)], and incubate for 1 h at 37°C. Precipitate the nucleic acids by the addition of 5 ml 5 M NaCl and 0.6 vol isopropanol, and recover immediately by centrifugation at maximum speed for 10 min at room temperature. Wash the DNA pellet with cold 70% ethanol, lyophilize briefly or air-dry, and dissolve in 80 ml TE. To determine the extent of DNase I digestion, analyze 5-ml aliquots by PAGE in 1% agarose gels and stain with ethidium bromide. Optimal extents of digestion may differ according to the chromatin structure of the region of interest and as a product of how distant the region of interest is in relation to the secondary digest and probe position. In general, a range of digests are taken for secondary digestion that should neither show a high proportion of mononucleosome fragments nor no apparent change on digestion. Select appropriate samples and use 20 ml of each DNA for secondary digestion and indirect end labeling. A schematic representation of the method is shown in Fig. 1 and a typical result in Fig. 2. Monitoring Factor Binding in Vivo by Dimethyl Sulfate Footprinting

FIG. 1. Procedure for DNase I chromatin analysis of isolated yeast nuclei.

In addition to the direct determination of the chromatin structure of a particular locus it is valuable to be able to correlate a defined nucleosomal organization with the presence or absence of a factor bound to the DNA in that region. In this respect the use of DMS in yeast has been particularly successful (e.g., 24, 25). Unlike DNase I, DMS is able to freely diffuse through the yeast cell wall and membrane and can be used directly on intact cells, thus obviating the need for the preparation of nuclei. Unfortunately only methylated guanines and, to a much lesser extent, adenines can be assayed for, which presents a technical problem should the protein of interest bind to a site with few or no guanine residues. For example,

298

¨ RZ GREGORY, BARBARIC, AND HO

binding of Pho4 to UASp1 of the PHO5 promoter results in the protection of a single guanine, and binding of Pho2 at this promoter cannot be detected as its binding site contains almost no guanine residues (4, 13). However, the speed of the DMS reaction, its application to whole cells, coupled with the fact that it appears to be effectively

blind to the presence or absence of nucleosomes (26), makes it a powerful and convenient method for the investigation of factor binding in vivo. Such binding may result in both protection and enhancement of the level of methylation (see Fig. 3B). Protection is presumably due to the physical presence of the protein, whereas enhancement of

FIG. 2. DNase I analysis of the two chromatin states of the yeast PHO8 promoter. (A) Nuclei from repressed or activated cells were digested with 0.5 (lanes 1 and 7), 1.5 (lanes 2 and 6), or 3.5 (lanes 3 and 5) U/ml DNase I for 20 min at 37°C. DNA was isolated, digested with BglII, analyzed on a 1.5% agarose gel, blotted, and hybridized with a PvuII/XhoI fragment as probe. A mixture of double digests of purified genomic DNA with BglII and EcoRV (E), HpaI (H), NheI (N), RsaI (R), HindIII (D), or XhoI (X) shown in lane 4 serve as markers. (B) Schematic representation of the chromatin organization of the repressed (1Pi) or active (2Pi) PHO8 promoter. The positions of the Pho4 binding sites UASp1 and UASp2 are shown. Stable nucleosomes (filled circles), slightly unstable nucleosomes (hatched circles), and highly labilized nucleosomes (open circles) are indicated. Also shown are the position of the probe, the restriction sites used to generate the marker lane, and the location of the BglII site used for indirect end labeling.

CHROMATIN ANALYSIS AND FACTOR BINDING IN YEAST

methylation may be due to altered DNA topology on factor binding or by the formation of a hydrophobic pocket concentrating the DMS next to the DNA (27). We have employed the technique given in detail below, which is similar to the previously described method of Giniger et al. (28), to the study of the Pho4 protein at both the PHO5 (4, 12) and PHO8 promoters (unpublished data and Fig. 3B) in their chromosomal locus. However, while the study of the gene in its native locus is preferable we have also successfully obtained clear DMS footprints when the gene is carried by an ARS-

299

CEN plasmid (unpublished observations). A schematic of the protocol is given in Fig. 3A, and a typical result from the use of this procedure to monitor binding of the Pho4 protein at the PHO8 promoter is shown in Fig. 3B. Protocol DMS TREATMENT

AND

DNA PURIFICATION

Grow the desired yeast strain in a 500-ml culture to approximately 2– 4 3 10 cells/ml, as for the nuclei

FIG. 3. In vivo DMS footprinting. (A) Schematic representation of the procedure. The guanine residues preferentially methylated by DMS treatment are shown, and the extent of methylation is indicated by the size of the methyl group or thickness of the resulting fragment. (B) In vivo footprint of Pho4 at UASp2 of the PHO8 promoter. Strains lacking (lane 1), overexpressing (lane 2), or expressing the Pho4 protein at wt levels were grown under activating conditions, treated with DMS, and analyzed with a PHO8 UASp2-specific primer. The positions of the guanine residues within the UASp2 element are marked with dots on the gel and on the sequence to the right in which the in vitro DNase I footprint of Pho4 is boxed (10). The thick arrow represents a guanine residue that is hypersensitive on Pho4 binding, observed as an enhancement of the signal from this residue. The short arrows indicate protected guanines, and the medium arrows indicate residues whose reactivity with DMS is unaffected relative to the pattern obtained in the absence of the Pho4 protein.

300

¨ RZ GREGORY, BARBARIC, AND HO

preparation (see earlier protocol). Collect the cells by centrifugation, and resuspend by adding 1.5 ml of warm growth medium per 1 3 10 cells/ml. Following resuspension, divide into aliquots of 1.5 ml in 50-ml tubes. To each aliquot add 2 ml DMS and incubate the cells at room temperature for 5–20 min. The optimal time required can depend on the growth conditions and on whether or not the protein of interest is bound. In the study of the Pho4 protein, for example, we have found that the optimal length of DMS treatment for cells under derepressing conditions brought about by growth in synthetic phosphate-free medium is ;20 min, whereas repressed cells grown in high-phosphate medium require ;10 min. Pho4 is, however, apparently rather stably bound to its sites throughout the time course of the DMS treatment. Should the DNA-binding protein of interest show a faster off-rate this may inhibit detection of binding by DMS. In such cases, it may be helpful to use higher concentrations of DMS with shorter time points. A time course of DMS treatment is therefore recommended to determine the optimal treatment for the specific conditions employed. Stop the DMS reaction by adding 40 ml of ice-cold TEN buffer [10 mM Tris–HCl (pH 7.5), 1 mM EDTA, 40 mM NaCl], and collect the cells by centrifugation. Resuspend the cell pellet in a total volume of 1 ml of SCED buffer [1 M sorbitol, 0.1 M sodium citrate (pH 5.8), 10 mM EDTA, 2 mM dithiothreitol (DTT)], transfer to 1.5-ml microfuge tubes, and digest with 1 mg/ml Zymolyase 100T (ICN) for 30 – 45 min at 37°C. The efficiency of the Zymolyase treatment should be monitored as described for the preparation of nuclei. Recover yeast spheroplasts by centrifugation, and resuspend each aliquot in 20 mM EDTA/50 mM Tris–HCl (pH 8.0) to a total volume of 1 ml. To lyse the cells add 50 ml 20% SDS and incubate at 65°C with shaking for 30 – 45 min. To precipitate cell debris add 0.4 ml 5 M potassium acetate (pH 8.0) to each tube, place the tubes on ice for 1 h, and centrifuge for 10 min at 4°C and maximum speed in a microfuge. Remove the supernatant carefully, divide into two microfuge tubes, and mix with 0.6 vol of isopropanol in a new tube to precipitate the nucleic acids. Centrifuge for 10 min at room temperature, wash the DNA pellet with cold 70% ethanol, air-dry for 5–10 min, and combine the previously divided samples by resuspension in a total volume of 0.6 ml TE. Remove RNA by digesting with 4 mg/ml RNase A (DNase-free) at 37°C for 1 h. Precipitate the nucleic acids by adding 2 ml 5 M NaCl and 0.4 ml isopropanol at room temperature, and

wash the pellets with cold 70% ethanol. Resuspend the pellet in 0.4 ml restriction enzyme buffer, and digest with an appropriate restriction enzyme to decrease viscosity. We generally use EcoRV; however, as long as there are no sites between the priming site and the site of interest other enzymes can be used. If the distance between the primer and a specific restriction site is 300 –500 bp, the resulting DNA fragment in untreated DNA samples can then be used as a control to evaluate how well the Taq polymerase extension works. Ethanol-precipitate the DNA in each tube, wash in cold 70% ethanol, and dissolve the DNA in 0.1 ml TE. Earlier methods employed a piperidine cleavage reaction prior to primer extension (22). However, as reported by others (29) this step is effectively redundant when the primer extension is carried out with Taq polymerase. PRIMER EXTENSION USING Taq POLYMERASE The primer should be about 16 –20 nucleotides long and should lie approximately 40-100 bp from the site of interest. Preferably gel-purified primers should be used to ensure that no shorter or truncated by-products of primer synthesis are present in the reaction. Label 5 pmol of the primer using T4 polynucleotide kinase and 10 pmol [g-32P]ATP (6000 Ci/mmol), and then purify from unincorporated label, for example, through a Stratagene Nuctrap probe purification column. An important factor in primer labeling is the resulting specific activity. Should a higher specific activity be required increase the concentration of the radiolabel to 30 pmol and reduce the oligonucleotide concentration to 3 pmol. Under these conditions almost every molecule of oligonucleotide becomes radiolabeled. Mix approximately 8 –10 ml of the isolated DNA (;5 mg DNA), or 5 ng plasmid DNA that has been treated with DMS in vitro for a free DNA control, with ;2 3 105 cpm of the radiolabeled primer in 50 ml commercial polymerase chain reaction (PCR) buffer (Boehringer: 103 buffer 1 1.5 mM MgCl2) containing 0.25 mM of each dNTP. Add 2.5–5 units Taq polymerase, and incubate the samples for 30 cycles of 95°C for 1 min, 45°C for 1 min, and 70°C for 2 min. The annealing temperature given here has proved successful with a number of different primers (13). However, should a nonspecific signal appear this may be abated by raising the annealing temperature accordingly. After the primer extension reaction add 6.6 ml of 1% SDS, 100 mM EDTA, 1 mg/ml proteinase K, and incubate the samples for 30 min at 45°C. Pre-

CHROMATIN ANALYSIS AND FACTOR BINDING IN YEAST

cipitate the DNA by adding 2 ml 5 M NaCl, mixing well, and adding 0.15 ml ethanol and incubating at 220°C overnight. Centrifuge at maximum speed for 10 min at 4°C and wash each pellet with cold 70% ethanol. Dissolve the precipitated DNA in a suitable volume of 13 TBE and add commercial denaturing gel loading buffer (stop solution) to allow approximately half the sample to be loaded onto the gel. Prior to loading, boil the samples for 3 min, place on ice for 2 min, and run the samples on a 6 –10% polyacrylamide sequencing gel. The gel can be either dried or simply covered with plastic wrap and then exposed to X-ray film with an intensifying screen at 270°C. Depending on the specific activity of the primer, the methylation pattern can often be seen after an exposure time of 12–24 h. The DMS-treated DNA that is used in the extension reaction is relatively impure, and occasionally, some inhibitors of the Taq polymerase are present in the DNA samples. This may lead to a lane being totally blank, but may be cured by use of half the amount of DNA in the extension reaction. Alternatively, the DNA can be further purified using commercially available DNA purification kits. To assist in the proper interpretation of the pattern obtained by in vivo DMS footprinting a number of controls are suggested. The use of an in vitro-treated DNA sample provides the ‘‘naked’’ DNA methylation intensities, from which it may be possible to observe protection or hypersensitivity. For inducible genes it may be possible to observe a change in the binding pattern on induction and, as with PHO5 and PHO8, correlate binding with the perturbation of chromatin structure across the promoter. Assuming previous work has identified a candidate gene encoding the putative DNA binding factor, the recombinant protein may used in vitro to determine the qualitative footprint for the factor, or a yeast strain lacking the protein assayed in vivo to correlate loss of gene with loss of footprint. In the latter case, however, it is still formally possible that the absence of the protein indirectly affects the ability of a second protein to generate a footprint.

have been extensively employed in the study of transcriptional regulation. However, the use of DNase I to identify positioned nucleosomes and/or hypersensitive sites within chromatin provides structural data that may be essential to the understanding of other cellular processes such as recombination and repair (30). One limitation of this technique is perhaps the inherent difficulty in quantifying the extent to which the chromatin structure has changed. In this regard the use of restriction enzymes with isolated nuclei can provide meaningful quantitative data on DNA accessibility, and as such is a valuable complementary technique to those described here (31). In vivo chromatin analysis techniques, in conjunction with genetic approaches made possible by the sequence of the entire yeast genome, make this model system ideal for the study of chromatinregulated processes in the eukaryotic cell.

ACKNOWLEDGMENTS We thank M. Mu¨nsterko¨tter for careful reading of the manuscript. This work was supported by grants from the European Commission Human Capital and Mobility Network (ERBCHRXCT940447), the Deutsche Forschungsgemeinschaft (SFB190), and Fonds der Chemischen Industrie.

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8.

CONCLUDING REMARKS The techniques described provide a basis for studying the organization of chromatin and the factors that associate with this structure in vivo and

301

9. 10. 11.

Wolffe, A. P. (1992) Chromatin: Structure and Function, Academic Press, San Diego. Edmondson, D. G., and Roth, S. Y. (1996) FASEB J. 10, 1173–1182. Gregory, P. D., and Ho¨rz, W. (1998) Curr. Opin. Cell Biol. 10, 339 –345. Venter, U., Svaren, J., Schmitz, J., Schmid, A., and Ho¨rz, W. (1994) EMBO J. 13, 4848 – 4855. Burns, L. G., and Peterson, C. L. (1997) Mol. Cell. Biol. 17, 4811– 4819. Brownell, J. E., and Allis, C. D. (1996) Curr. Opin. Genet. Dev. 6, 176 –184. Wolffe, A. P. (1997) Nature 387, 16 –17. Almer, A., Rudolph, H., Hinnen, A., and Ho¨rz, W. (1986) EMBO J. 5, 2689 –2696. Gregory, P. D., Schmid, A., Zavari, M., Lin, L., Berger, S. L., and Ho¨rz, W. (1998) Mol. Cell 1, 495–505. Barbaric, S., Fascher, K. D., and Ho¨rz, W. (1992) Nucleic Acids Res. 20, 1031–1038. Pavlovic, B., and Ho¨rz, W. (1988) Mol. Cell. Biol. 8, 5513– 5520.

302

¨ RZ GREGORY, BARBARIC, AND HO

12. Svaren, J., Schmitz, J., and Ho¨rz, W. (1994) EMBO J. 13, 4856 – 4862. 13. Barbaric, S., Mu¨nsterko¨tter, M., Goding, C. R., and Ho¨rz, W. (1998) Mol. Cell. Biol. 18, 2629 –2639. 14. Wintersberger, U., Smith, P., and Letnansky, K. (1973) Eur. J. Biochem. 33, 123–130. 15. Aris, J. P., and Blobel, G. (1991) Methods Enzymol. 194, 735–749. 16. Lohr, D. (1988) in Yeast: A Practical Approach (Campbell, I., and Duffus, J. H., Eds.), pp. 125–145, IRL Press, Oxford. 17. Fedor, M. J., Lue, N. F., and Kornberg, R. D. (1988) J. Mol. Biol. 204, 109 –127. 18. Hull, M. W., Thomas, G., Huibregtse, J. M., and Engelke, D. R. (1991) Methods Cell Biol. 35, 383– 415. 19. Huibregtse, J. M., and Engelke, D. R. (1991) Methods Enzymol. 194, 550 –562. 20. Kent, N. A., Bird, L. E., and Mellor, J. (1993) Nucleic Acids Res. 21, 4653– 4654. 21. Schmid, A., Fascher, K. D., and Ho¨rz, W. (1992) Cell 71, 853– 864.

22. Svaren, J., Venter, U., and Ho¨rz, W. (1995) Microbial Gene Techniques 6, 153–167. 23. Ho¨rz, W., and Altenburger, W. (1981) Nucleic Acids Res. 9, 2643–2658. 24. Gross, D. S., English, K. E., Collins, K. W., and Lee, S. W. (1990) J. Mol. Biol. 216, 611– 631. 25. Ganter, B., Tan, S., and Richmond, T. J. (1993) J. Mol. Biol. 234, 975–987. 26. Axelrod, J. D., Reagan, M. S., and Majors, J. (1993) Genes Dev. 7, 857– 869. 27. Ogata, R. T., and Gilbert, W. (1979) J. Mol. Biol. 132, 709 – 728. 28. Giniger, E., Varnum, S. M., and Ptashne, M. (1985) Cell 40, 767–774. 29. Brewer, A. C., Marsh, P. J., and Patient, R. K. (1990) Nucleic Acids Res. 18, 5574. 30. Suter, B., Livingstone, Z. M., and Thoma, F. (1997) EMBO J. 16, 2150 –2160. 31. Gregory, P. D., Barbaric, S., and Ho¨rz, W. (1998) Methods Mol. Biol., in press.