Here we are presenting a chromatin immunoprecipitation (ChIP) procedure for genome-wide location analysis of protein isoforms that differ in a histone-binding domain. We are applying it to ChIP-Seq analysis to identify the targets of the KDM5A/JARID1A/RBP2 histone demethylase.
Recruitment of transcriptional and epigenetic factors to their targets is a key step in their regulation. Prominently featured in recruitment are the protein domains that bind to specific histone modifications. One such domain is the plant homeodomain (PHD), found in several chromatin-binding proteins. The epigenetic factor RBP2 has multiple PHD domains, however, they have different functions (Figure 4). In particular, the C-terminal PHD domain, found in a RBP2 oncogenic fusion in human leukemia, binds to trimethylated lysine 4 in histone H3 (H3K4me3)1. The transcript corresponding to the RBP2 isoform containing the C-terminal PHD accumulates during differentiation of promonocytic, lymphoma-derived, U937 cells into monocytes2. Consistent with both sets of data, genome-wide analysis showed that in differentiated U937 cells, the RBP2 protein gets localized to genomic regions highly enriched for H3K4me33. Localization of RBP2 to its targets correlates with a decrease in H3K4me3 due to RBP2 histone demethylase activity and a decrease in transcriptional activity. In contrast, two other PHDs of RBP2 are unable to bind H3K4me3. Notably, the C-terminal domain PHD of RBP2 is absent in the smaller RBP2 isoform4. It is conceivable that the small isoform of RBP2, which lacks interaction with H3K4me3, differs from the larger isoform in genomic location. The difference in genomic location of RBP2 isoforms may account for the observed diversity in RBP2 function. Specifically, RBP2 is a critical player in cellular differentiation mediated by the retinoblastoma protein (pRB). Consistent with these data, previous genome-wide analysis, without distinction between isoforms, identified two distinct groups of RBP2 target genes: 1) genes bound by RBP2 in a manner that is independent of differentiation; 2) genes bound by RBP2 in a differentiation-dependent manner.
To identify differences in localization between the isoforms we performed genome-wide location analysis by ChIP-Seq. Using antibodies that detect both RBP2 isoforms we have located all RBP2 targets. Additionally we have antibodies that only bind large, and not small RBP2 isoform (Figure 4). After identifying the large isoform targets, one can then subtract them from all RBP2 targets to reveal the targets of small isoform. These data show the contribution of chromatin-interacting domain in protein recruitment to its binding sites in the genome.
The protocol was originally adapted from B. Ren, 2001. It represents a slight modification of the protocol by Odom et al.5 that can be found at http://jura.wi.mit.edu/cgi-bin/young_public/navframe.cgi?s=22&f=appendices_downloads
1. Pre-block and binding of antibody to magnetic beads (Should be performed the night before next step)
2. Cell cross-linking
3. Cell Sonication
4. Chromatin Immunoprecipitation
5. Amplification of genomic DNA
6. Gel purification of amplified products
7. Representative Results
Figure 1. The overall goal of the following experiment is to identify genomic targets of KDM5A/JARID1A/RBP2 histone demethylase.
(P1) This is achieved by preparation of cross-linked chromatin that is sonicated to produce DNA fragments of appropriate size. (P2) As a second step, RBP2 bound DNA fragments are immunoprecipitated with RBP2 antibodies. (P3) Next, the recovered DNA fragments are repaired at the ends and adapters are ligated onto the ends of genomic DNA to prepare libraries of genomic DNA for analysis on the flow cells in the Illumina Cluster Station. The DNA libraries are amplified by PCR using low number of cycles. (P4) Finally, illumina sequenced short reads are uniquely aligned to the human genome, significant peaks are identified and annotated to the closest genes in order to identify RBP2 enriched regions. (P5) Results are obtained that show molecular functions associated with RBP2 target genes based on genome-wide identification of RBP2 enriched regions.
Figure 2. Checking the results of chromatin sonication To check the sizes of the DNA fragments produced by the sonication procedure, 5 μg (1/10) of the purified input sample was loaded on 1.8% agarose gel. As expected, our sonication procedure produced fragments in the range of 150-350 bp. However, if the fragments do not fall in this range, the sonication procedure should be adjusted accordingly, by varying the number of bursts and the amplitude.
Figure 3. Gel purification of amplified products. The PCR products are run on a gel to remove the adapters and select a size-range of templates for the cluster generation platform. This gel shows that fragments in the range between 150 and 350 base pairs were produced. They represent genomic DNA fragments between 50 and 250 base pairs in length and an adapter. We excise the selected region of gel with a scalpel. Care should be taken to avoid the adapter-adapter band, which runs at about 120 base pairs.
Figure 4. Isoform-specific antibody allows distinguishing between large and small isoforms of RBP2. RBP2 protein structure is presented in domain view. RBP2 contains several domains: the catalytic histone demethylation JmjC domain and associated JmjN domain, an ARID domain capable of sequence-specific DNA binding, a C5HC2 zinc finger that may potentially interact with DNA or other proteins, and multiple PHD domains. The two anti-RBP2 antibodies that allow distinguishing between RBP2 isoforms were derived against the RBP2 fragments indicated by thick lines.
Figure 5a. Overview of RBP2 binding regions along with chromosome and Ensembl genes. Genomic coordinates of identified enriched RBP2 (all isoforms and large isoform) binding regions are presented chromosome wise. Occupances of Ensembl genes (version 54; hg18) in the chromosomes (chromosome 10 in this picture) are presented as black bars (top panel). Each RBP2 binding region is represented as a vertical line, where the middle panel shows all RBP2 isoforms binding regions on chromosome 10, and the bottom panel shows RBP2 large isoform binding regions. The middle and bottom panels give overview of overlapping and specific occupancy of RBP2 isoforms on chromosome 10.
Figure 5b. Functional enrichment analysis of RBP2 target genes. Heatmap showing FDR corrected significantly (p-value ≤0.05) enriched GO Molecular Function categories among the genes bound (closest genes to the RBP2 binding region) by RBP2 (all and large isoforms). Colors toward red indicate high statistic significance, yellow indicates low statistic significance, and gray indicates no statistic significance. Enrichment analysis shows the overlapping and isoform specific molecular functions of RBP2 target genes.
Functional differences between RBP2 isoforms have not been determined. We have used a comprehensive approach to identify genomic regions bound by RBP2 isoforms and define the functional categories that these regions represent. This was accomplished by ChIP-Seq analysis followed by bioinformatics analysis of obtained sequence reads.
RBP2 modifies methylated lysine residues on histone tails. We found that RBP2 large isoform containing the recognition module for methylated histone lysine and RBP2 small isoform lacking this module bind to different regions in the human genome (Figure 5A). Importantly, isoform-specific regions and overlapping regions belong to genes with different molecular functions (Figure 5B). For example, “chromatin binding” and “transcription factor binding” functions can be ascribed to gene targets of RBP2 small isoform but not of RBP2 large isoform (Figure 5B). By comparing actual gene sets generated for all isoforms and RBP2 large isoform (data not shown), we can also define if the large isoform is specifically recruited to certain genes.
The authors have nothing to disclose.
This work was supported by 115347-RSG-08-271-01-GMC from ACS, and by CA138631 grant from NIH.
Material Name | タイプ | Company | Catalogue Number | Comment |
---|---|---|---|---|
37% formaldehyde solution | Fisher | F79500 | ||
BSA | USB | 10852 | ||
chloroform/isoamyl alcohol | Sigma | C0549 | ||
Dynabeads Protein G | Invitrogen | 100.04D | ||
EDTA | Fisher | BP120-500 | ||
EGTA | Sigma | E3889 | ||
Genomic DNA Sample Prep Kit | Illumina | FC-102-1001 | ||
glycerol | Sigma | G5516 | ||
glycogen | Boehringer | 901393 | ||
Hepes | Invitrogen | 15630080 | ||
KOH | Fisher | P250-500 | ||
LiCl | Sigma | L9650 | ||
Low molecular weight DNA ladder | NEB | N3233S | ||
sodium deoxycholate | Sigma | D6750 | ||
N-lauroyl sarcosine | MP Biomedicals | 194008 | ||
NP-40 | Sigma | I3021 | ||
PBS | Fisher | mt21040cv | ||
phenol | Sigma | P-4557 | ||
Phusion DNA Polymerase | NEB | F-540S | ||
proteinase K | Gibco | 25530-049 | ||
QIAquick PCR purification Kit | QIAGEN | 28104 | ||
MinElute PCR purification Kit | QIAGEN | 28004 | ||
QIAquick Gel Extraction Kit | QIAGEN | 28704 | ||
RNAse A | Sigma | R4642 | ||
SDS | Invitrogen | 24730020 | ||
TE (endofree for maxi-prep kit) | QIAGEN | 19048 | ||
Tris | Sigma | 154563 | ||
Triton X-100 | Fisher | BP151100 | ||
Ultra Agarose, Certified low-range | BIO-RAD | 161-3106 |
Oligonucleotide sequences 2006 Illumina, Inc. All rights reserved.