Elsevier

Methods

Volume 31, Issue 1, September 2003, Pages 90-95
Methods

In vivo protein–protein and protein–DNA crosslinking for genomewide binding microarray

https://doi.org/10.1016/S1046-2023(03)00092-6Get rights and content

Abstract

Chromatin immunoprecipitation (ChrIP or ChIP) has commonly been used to map protein–DNA interaction sites at specific genomic loci through use of formaldehyde-induced crosslinking. However, formaldehyde alone has proved inadequate for crosslinking of certain proteins such as the yeast histone deacetylase Rpd3. We report here a modified crosslinking procedure that includes a protein–protein crosslinking agent in addition to formaldehyde. Using this double crosslinking method, we have successfully mapped Rpd3 binding sites in vivo. We also describe the use of ChrIP in combination with DNA microarrays (ChrIP-array) to determine the pattern of Rpd3 binding genomewide. This approach couples the versatility of ChrIP with that of microarrays to identify binding patterns that would otherwise be hidden in a gene-by-gene survey.

Introduction

In eukaryotes, chromatin is composed of DNA wrapped around an octameric core of histones. Far from being a simple repetitive scaffold, chromatin contains gene-specific and chromosomal domain specific histone modification patterns. These play a dynamic role in chromosomal functions that include heterochromatin formation, transcription, and replication. Therefore, the proteins and enzymes that modify chromatin are important factors that establish and maintain a specific chromatin state.

One class of chromatin-modifying enzymes is represented by the histone deacetylase Rpd3, which confers gene-specific repression of transcription when recruited by a DNA-binding protein to a particular upstream regulatory sequence. For instance, Ume6 binds to the INO1 promoter and recruits the Rpd3 complex to repress INO1 activity [1], [2]. Despite this targeted recruitment, Rpd3 also deacetylates large regions of chromatin including promoters and coding regions that do not contain UME6 binding sites in a process termed “global deacetylation [3].” Like local deacetylation of single genes, the global effect of Rpd3 represses transcription but may function over larger chromosomal domains [3]. The role of Rpd3 in genomewide regulation of transcription is further compounded by the fact that the Rpd3 complex is very large (∼1 MDa) and appears heterogeneous [4], suggesting that there may be site-specific utilization of certain components of the complex. Together, these observations indicate that genomewide approaches are needed to illustrate the variety of ways by which Rpd3 recognizes chromatin, modifies histones, and regulates gene expression. Genomewide expression microarrays have provided a partial picture of histone deacetylase function by revealing genes that are de-repressed when RPD3 is deleted [5], [6]. However, the expression of genes can be indirectly affected by the absence of the ubiquitous deacetylase. Examining whether genes in microarrays are hyperacetylated on RPD3 disruption can be integrated with expression arrays to greatly decrease apparent indirect effects evident in the microarrays [7]. However, we have found that there are many cases in which multiple deacetylases have redundant effects at specific genes, thereby hiding the role of Rpd3. Moreover, deacetylases may interact with regulators of active genes only to repress transcription later in the life cycle of yeast. Therefore, enzyme binding arrays are also needed to directly determine the genomic regions that are regulated by Rpd3.

In a common technique used to probe DNA–protein interactions, chromatin-bound proteins are first crosslinked to DNA by formaldehyde and then immunoprecipitated by specific antibodies to a protein of interest (chromatin immunoprecipitation, ChrIP). The immunoprecipitated DNA is purified and semiquantitatively amplified by PCR [8], [9]. The enrichment of specific sequences relative to an internal control reflects the sites of protein–DNA interaction(s) in vivo. However, formaldehyde has proved inadequate for crosslinking of Rpd3 to chromatin. Therefore, we describe here a method for in vivo determination of Rpd3 binding sites, which combines a novel protein–protein (DMA) crosslinking method with formaldehyde and DNA microarrays to provide a genomewide binding map [10], [11], [12]. Such a panoramic map has uncovered distinct chromosomal domains affected by Rpd3, the regulation of groups of genes based on common function that would otherwise be concealed in a gene-by-gene survey, and the presence of novel Rpd3 recruitment mechanisms. In addition, binding maps of other members of the large Rpd3 complex elucidate their limited or widespread use genomewide.

Section snippets

Method

This section contains mainly the protocol for double crosslinking using whole cells and a brief subsection on application of ChrIP to DNA microarrays (for more detail, see Robyr and Grunstein in this issue [22]). The basic steps include: (1) protein–protein DMA crosslinking followed by formaldehyde-induced protein–DNA crosslinking; (2) ChrIP: shearing of chromatin, immunoprecipitation with specific antibodies, reversal of crosslink, and purification of immunoprecipitated DNA; (3)

Conclusion

We have employed a modified crosslinking method for ChrIP that includes a protein–protein crosslinking agent in addition to formaldehyde to map Rpd3 binding in yeast for the first time. We have also shown that protein–protein crosslinking agents other than DMA may be used for Rpd3 ChrIP. Considering the large variety of protein–protein crosslinking agents, the double crosslinking method should, therefore, increase the utility of ChrIP, especially for proteins that fail to crosslink to DNA by

Acknowledgements

S.K.K. is a Howard Hughes Medical Institute Physician Postdoctoral Fellow. This work was supported by Public Health Service grants (GM23674 and GM42421) of the National Institutes of Health to M.G.

References (22)

  • D Kadosh et al.

    Cell

    (1997)
  • T.R Hughes et al.

    Cell

    (2000)
  • D Robyr et al.

    Cell

    (2002)
  • A Hecht et al.

    Methods Enzymol.

    (1999)
  • N Suka et al.

    Mol. Cell

    (2001)
  • F.C Holstege et al.

    Cell

    (1998)
  • J.D Hughes et al.

    J. Mol. Biol.

    (2000)
  • S.E Rundlett et al.

    Nature

    (1998)
  • M Vogelauer et al.

    Nature

    (2000)
  • M.M Kasten et al.

    Mol. Cell. Biol.

    (1997)
  • B.E Bernstein et al.

    Proc. Natl. Acad. Sci. USA

    (2000)
  • Cited by (103)

    • State-of-the-art techniques to study epigenetics

      2023, Progress in Molecular Biology and Translational Science
    • An Optimized Chromatin Immunoprecipitation Protocol for Quantification of Protein-DNA Interactions

      2020, STAR Protocols
      Citation Excerpt :

      This means that proteins that do not directly contact DNA can only be cross-linked to DNA through other proteins, which is inefficient. Use of a protein-protein cross-linker such as for example dimethyl adipimidate or disuccinimidyl glutarate together with formaldehyde leads to more efficient ChIP for these proteins (Kurdistani and Grunstein, 2003; Tian et al., 2012). For a complete list of protein cross-linkers and their properties see (Kurdistani and Grunstein, 2003; Tian et al., 2012).

    • Quantitative telomeric chromatin isolation protocol for human cells

      2017, Methods
      Citation Excerpt :

      However, to crosslink transient interactors, larger multiprotein complexes and proteins that are more distant from the core chromatin, longer cross-linking times should be employed. The co-immunoprecipitation efficiency of these proteins may be improved when in addition to formaldehyde, a protein-protein crosslinking agent is included [54,55]. For QTIP we obtained good results when combining formaldehyde with EGS, an amine-reactive cross-linker with a 12-atom spacer arm of 16 Å, which can be cleaved by treatment with hydroxylamine at pH 8.5.

    View all citing articles on Scopus
    View full text