Identifying transcriptional targets

Abstract

Identifying the targets of transcription factors is important for understanding cellular processes. We review how targets have previously been isolated and outline new technologies that are being developed to identify novel direct targets, including chromatin immunoprecipitation combined with microarray screening and bioinformatic approaches.

Figure 1

Four established techniques that are used to identify transcription-factor targets. These methods all compare mRNAs extracted from two populations of cells, one of which has the transcription factor in question overexpressed or knocked out. (a) Differences in the levels of specific candidate target genes in the two populations can be analyzed by reverse-transcriptase-coupled (RT-)PCR (for example, see [1,40]). (b) Any mRNAs that are equally expressed in both populations are subtracted, or removed, by cDNA-RNA hybridization. The remaining cDNAs are derived from mRNAs that are differentially expressed in one of the populations, and these can then be cloned and sequenced [3]. (c) With differential display, partial cDNA sequences are amplified from mRNA pools by RT-PCR. One primer - (T)_nNN - binds to the polyadenylated tail of a subset of mRNAs that is defined by the two bases immediately 53 to the tail. The other binds to short sequences (6 or 7 base-pairs) that will occur with moderate frequency within the transcriptome. The products are radiolabeled and analyzed by polyacrylamide gel electrophoresis. Short cDNAs present in only one population can be isolated and sequenced [41,42]. (d) In serial analysis of gene expression (SAGE), cDNA is synthesized from mRNA and cleaved by a restriction enzyme that recognizes a 4 nucleotide sequence. The 33 end of the cleaved cDNA is isolated using beads that bind to oligo-dT, and 53 linkers are ligated to the restriction sites. These linkers contain type-IIS restriction sites, which are recognized by endonucleases that cleave a defined distance away (up to 20 base-pairs). This produces short DNA tags whose sequence and position are sufficient to identify the original transcript, provided cDNA sequences or expressed sequence tags (ESTs) are already known. The tags can be concatenated and sequenced, providing quantitative analysis of many transcripts simultaneously [43].

Figure 2

Experimental procedures for identifying transcription-factor targets in vivo by chromatin immunoprecipitation (ChIP) and Dam methylase identification (DamID), using microarrays. (a) In ChIP, formaldehyde is used to fix proteins bound to DNA in vivo. The DNA is then isolated and sheared by sonication into fragments of 200-700 base-pairs. An antibody against the transcription factor of interest is used to immunoprecipitate the factor and associated chromatin; or, if an epitope-tagged version of the protein is expressed in cells, an antibody can be used that is specific to the epitope. Protein is then removed from the DNA by reversal of the crosslinks and digestion with proteinase K. At this point, the isolated DNA can be used to verify targets by PCR or dot blot, or the DNA may be sub-cloned and sequenced to identify new targets [44]. For ChIP array analysis, the purified DNA is amplified by PCR and then labeled with a fluorophore, such as Cy3. As a reference for background binding, input DNA that is not enriched by immunoprecipitation is also amplified and labeled with another fluorophore, such as Cy5 [16,18]. Alternatively, non-enriched reference DNA is isolated after immunoprecipitation from cells that do not contain the transcription factor of interest [15]. The two populations of DNA are then hybridized to a microarray containing genomic sequences, and target sequences bound by the factor are identified according to the relative fluorescent intensity of each spot. (b) With DamID, the transcription factor of interest is fused to the Escherichia coli enzyme DNA adenine methylase (Dam). The fusion protein is expressed in vivo and Dam methylates DNA in the immediate vicinity of the binding site of the transcription factor, specifically acting on adenines in the sequence GATC. Dam alone is also expressed in cells as a reference, to identify background binding and methylation. Given that endogenous methylation of adenine does not occur in the DNA of most eukaryotes, methylated DNA can then be digested with Dpn1 (which cuts at the sequence GA^meTC) and isolated from uncut genomic DNA by size fractionation. The resulting DNA can then be analyzed by Southern blot to verify putative targets [14]. For genome-wide analysis, DNA from the experimental and reference samples is labeled with two different fluorophores (such as Cy3 and Cy5) and hybridized to a microarray [17,28,29].