Genome-wide detection and analysis of hippocampus core promoters using DeepCAGE

Abstract

Finding and characterizing mRNAs, their transcription start sites (TSS), and their associated promoters is a major focus in post-genome biology. Mammalian cells have at least 5-10 magnitudes more TSS than previously believed, and deeper sequencing is necessary to detect all active promoters in a given tissue. Here, we present a new method for high-throughput sequencing of 5' cDNA tags-DeepCAGE: merging the Cap Analysis of Gene Expression method with ultra-high-throughput sequence technology. We apply DeepCAGE to characterize 1.4 million sequenced TSS from mouse hippocampus and reveal a wealth of novel core promoters that are preferentially used in hippocampus: This is the most comprehensive promoter data set for any tissue to date. Using these data, we present evidence indicating a key role for the Arnt2 transcription factor in hippocampus gene regulation. DeepCAGE can also detect promoters used only in a small subset of cells within the complex tissue.

Figure 1.

Preparation of DeepCAGE cDNA libraries. cDNA is produced with reverse transcriptase using random priming to maximize chances to reach the cap sites and to include non-polyadenylated RNAs. The cap site is biotinylated, followed by cleavage of single-strand RNA (incomplete cDNA). After recovery of the cDNA, a linker is ligated on the 5′-end, which carries the MmeI restriction site, which cleaves 20/21 nt inside the 5′-end of the cDNA. After a second linker ligation and PCR amplification, restricted digested tags are purified and finally concatenated together with the “454” linkers A and B. (SMB) Streptavidin-coated magnetic beads; (B) biotin.

Figure 2.

Exploration and validation of identified core promoters. (A) Exploration of tissue usage in all core promoters having more than 30 tags per million using hierarchical clustering, with CAGE tag expression data from the actual core promoters. Preferential usage for a certain tissue (fraction of tags belonging to the tissue in question) is color-coded as shown in the legend. Rows represent individual promoters (the row dendrogram is omitted because of the large number of rows), while columns are the different tissues. (B) Number of core promoters used preferentially in just one tissue (PEPs, as defined in Methods) and the locations of these core promoters relative to known genes. (C,D) Examples of discovery and validation of novel intergenic promoters expressed preferentially in hippocampus. All hippocampus tags are shown as red bar plots (one for each strand). Clusters of tags that are preferentially expressed in hippocampus (hippocampus PEPs) are shown as blue fragments where the color intensity is proportional to the fraction of tags in the cluster from hippocampus vs. other tissues. RACE products are in black. (C) The validation of a proximal alternative promoter to the Bai3 gene. (D) The validation of a promoter upstream of the Arpc5 gene but on the opposite strand; RACE as well as a human orthologous transcript show that this is a distal upstream promoter for the Rgl1 gene.

Figure 3.

Alternative promoters preferentially expressed in different brain tissues. (A) The Venn diagram shows the number of genes having at least one preferentially expressed promoter (PEP) from any of the four sampled brain tissues, or any combination PEPs of the four tissues. For example, there are 15 genes that have at least one hippocampus PEP and one cerebellum PEP. (B) The Dlgap1 gene has four PEPs, one from each brain tissue. All of these are overlapping known 5′-ends inferred by cDNAs. There are five 14-amino-acid-long repeats in the N-terminal end of the corresponding protein that affects the ability to interact with other proteins. Repeats are indicated as triangles. Note that the cerebellum PEP is downstream from these and will give a protein that lacks these repeats.

Figure 4.

Examples of changes of domain content for genes by use of hippocampus PEPs. Hippocampus preferentially expressed promoter (PEP) locations are shown as red triangles. Locations of predicted protein domains are shown as colored blocks (note that domains spanning more than exons are extended over the intron(s). In all of these cases, at least one domain is upstream of the PEP, which means that this domain is not included in the isoform expressed in hippocampus. Known cDNA locations are shown below: transcription is right-to-left. BAI1 is a membrane protein whose N-terminal domain is extracellular, with a transmembrane region just downstream from the GPS domain (data not shown). The extracellular part can be cleaved off at the GPS domain, releasing a tumor-suppressing peptide; however, the hippocampus PEP is just downstream from the GPS domain, presumably giving a BAI1 variant that is attached to the membrane but without the extracellular domains, which lack the tumor-suppression capability (Kaur et al. 2005). Similarly, the PEPs in Myo10 confirm a previous study showing the neuronal expression of an isoform lacking the Myosin head domain (Sousa et al. 2006). In Pclo, the zf-piccolo domain cannot be included when using the hippocampus PEP.

Figure 5.

Transcription factor genes with preferential expression in hippocampus. (A) The relation between expression strength (number of hippocampus CAGE tags/million) vs. the “tissue specificity” (fraction of hippocampus tags vs. all brain tags, normalized for library size) of all known transcription factor genes. Note that only a few transcription factor genes are both highly expressed in and highly specific for hippocampus. For example, both in situ hybridization and CAGE data show that the Nr3c2 gene (the mineralocorticoid receptor) is not highly expressed in brain, but almost exclusively expressed in hippocampus, while the Aes gene is expressed in the whole brain with a preference for hippocampus. (C-H and Aes, right) In situ images for some of these factors. (B) In situ hybridization images (Lein et al. 2007) showing the hippocampus expression of the Arnt2 gene. The three images show the Allen reference atlas, the ISH in situ image, and corresponding expression heatmaps.

Figure 6.

CAGE identifies promoter activity from small subpopulations of hippocampal cells. Examples of correspondence between CAGE tags and signal detected by in situ hybridization, ordered from relatively high expression (from the top left quadrant), expressed as the number of CAGE tags from hippocampus mapping to the gene, to low expression (the lowest right quadrant). On the left the original in situ signal is shown; on the right the in situ hybridization signal is quantified with pseudo-colors where red corresponds to high expression. In situ hybridization images were obtained from the Allen Brain Institute (Lein et al. 2007). Notice that the signal or tags corresponding to less than 5/1.4 × 10⁶ mapped tags correspond to RNAs that are expressed only in a specific subset of cells.