Advancing Epigenetics Towards Systems Biology

In vivo dual cross-linking chromatin immunoprecipitation: detecting chromatin proteins not directly bound to DNA (Prot 29)

Antonio Porro and Giovanni Perini

Introduction

Cross-linking Chromatin Immunoprecipitation (ChIP) has become a popular method to detect the in vivo binding of proteins to DNA. In general the protein/DNA complex is fixed with formaldehyde, a cross-linking agent that, because of its short spacer arm (about 2 A), generates reversible covalent links, mainly between proteins and DNA. Subsequently, chromatin is fragmented by sonication and DNA protein complexes are selectively immunoprecipitated with antibodies raised against proteins of interest. Covalent links are reversed and proteins are removed from DNA through phenol/chloroform extraction. Following DNA purification, specific DNA regions supposed to be engaged with proteins of interest are analyzed by semi-quantitative PCR or real time PCR. Although very powerful, this method may present some limits when applied to nuclear proteins that, although being tightly associated with chromatin DNA, are not directly bound to it. For example, recent studies have shown that c-Myc can repress transcription of the p21CIP/WAF gene without binding DNA directly but through interaction with different transcription factors such as Sp1 and Miz-1 that recognize specific DNA sequences in the p21CIP/WAF promoter (Gartel et al., 2001 and Wu e al., 2003). N-Myc, another member of the Myc family, is expressed essentially during development of the nervous system and in some rare pediatric tumors such as neuroblastoma, medulloblastoma and rhabdomyosarcoma. We have recently defined some parts of the network of genes that N-Myc can regulate in human neuroblastoma (Perini et al., 2005). Many target genes are positively regulated by N-Myc; nonetheless, we have found that several genes are also negatively regulated by N-Myc, including p21CIP/WAF (unpublished data). The majority of the promoters of these genes do not carry canonical E-Box sites, thus suggesting that N-Myc, like c-Myc, may repress their transcription through indirect binding to promoters. To specifically address this problem, we thought of combining the use of formaldehyde with other cross-linking agents to improve the formation of covalent links between proteins and stabilize the association of proteins to promoters, including those that are not directly bound to DNA. A few papers had been recently published describing the efficacy of several cross-linking agents in ChIP to detect binding of the transcription factors to target genes (Kurdistani et al., 2002; Noma et al., 2006; Nowak et al., 2005; Zeng et al., 2006) Based on these studies we set up a dual cross-linking ChIP protocol that we have successfully employed to improve immunoprecipitation of complexes in which tested factors are not in direct contact with DNA. Specifically, dual cross-linking ChIP was used to check whether N-Myc may associate with the p21CIP/WAF promoter to repress transcription. In Figure 1 we compare results obtained using standard ChIP or the dual cross-linking ChIP. As it can be observed in the Figure, dual cross-linking ChIP significantly improves recovery of p21CIP/WAF promoter and confirms the association of N-Myc to it. Importantly, the sequential use of two cross-linking agents does not affect the recovery of other DNA regions bound by N-Myc directly (see ChIP results for the Apex-1 gene promoter).

In conclusion the protocol described below, can be used to improve in vivo detection of proteins that do not contact DNA directly, though being part of chromatin. Moreover, results suggest that the use of novel cross-linking agents may help increase ChIP sensitivity.

PDF version

Antonio Porro and Giovanni Perini

University of Bologna - Department of Biology - University of Bologna - Via Francesco Selmi 3 - 40126 Bologna, Italy

Antonio Porro