Start New Search

Quick tour

How to use the database:

Choose the species, select the table your want to search, leave field as it as ('All') and type in the gene you want to search. A table listing the search results will be returned, followed by the description of the table. If no search result is returned, don't give up yet, try use the official gene symbol or gene ID (locuslink number) from NCBI Entrez Gene.

How to interpret the result:

When the promoter of a gene is occupied by CREB protein, we call it a CREB target gene. In each tissue or cell type, only a small yet distinct subset of CREB target gene are activated by cAMP. Before we look at the result, please keep in mind a few important concepts here:

1) It seems CREB occupancy is very similar across tissues. If you see CREB binding in HEK293T cells from the ChIP-on-chip result, it is likely this is also true in the tissue you are interested in.
2) However, only a small subset of CREB occupied promoters are responsive to cAMP in each tissue. If you see a CREB target gene being induced by cAMP in liver, it does not say anything about whether the gene will be incuded in brain.
3) The sensitivity of the ChIP-chip assay is only about 50%. For genes below the threshold, it could still be a perfect CREB target if there is strong evidence from other sources.

We will use some examples to give you a feeling about how to use the database and interpret the result.

a) DUSP1, a typical CREB target gene, which shows CREB occupancy on promoter, induction in HEK293T cells and presence of good CRE on the promoter.

b) NR4A2, a gene highly induced by cAMP in HEK293T cells and contain good CRE sites in the promoter. The ChIP-chip signal is not strong, but based on the other evidence, it is still a good CREB target.

c) FDPS, whose promoter is bound by CREB and contain good CRE sites. The gene in not induced by cAMP in the two tissues we have tested.

Now let's look at some real examples:

a) DUSP1, a typical CREB target gene, which shows CREB occupancy on promoter, induction in HEK293T cells and presence of good CRE on the promoter.

Let's say we want to check dual specificity phosphatase 1. To do this, just search the gene symbol DUSP1 in "human ChIP-chip" table. Two genes were returned, the first one was DUSP12, and the second hit was what we want. The p-value and binding ratio for DUSP1 in HEK293T cells are good, meaning CREB is proably on the promoter.

Now we want to see if DUSP1 is induced by cAMP in HEK293T cells. We can search DUSP1 in HEK293T_expression table, but a faster way is to click the link from the last column in the table. Now we get the expression result in HEK293T cells. There are two probe sets for DUSP1 on the Affymetrix array, we can take the first one, which has higher expression values and is therefor more reliable. The induction by cAMP agonist FSK in very big, and this effect is blocked by dominant negative ACREB.

Now we may also want to check if there is a good CREB on the promoter. To do this, search DUSP1 in CRE_Preditcion table or just click the link from the last column. We can see that there is indeed a good full site at -113 and another half site at -163, and there is a good TATA box following these sites. All look good.

b) NR4A2, a gene highly induced by cAMP in HEK293T cells and contain good CRE sites in the promoter. The ChIP-chip signal is not strong, but based on the other evidence, it is still a good CREB target.
Let's say we noticed that NR4A2 is highly induced by cAMP in HEK293T cells. Search NR4A2 in HEK293T_Expression table and you will get the following. All three probe sets show good induction by cAMP agonist FSK and this effect is blocked by dominant negative ACREB.

When we check the promoter by clicking Human_CRE_prediction at the last column, four splice variants of NR4A2 show up. They have the same start site and there are a bunch of conserved CREs on the promoter. These CREs form clusters and they are also positionally conserved.

However, when we check the ChIP-on-Chip data (click the quick link 'human_ChIP_chip' from the last column), the numbers did not meet the criteria as CREB target genes. This could be due to a variant of reasons, but considering the strong evidence from the two tables above, NR4A2 is still likely a good CREB tagets. Indeed, in our manual ChIP data, NR4A2 promoter is verified to be CREB positive.

c) FDPS, whose promoter is bound by CREB and contain good CRE sites. The gene in not induced by cAMP in the two tissues we have tested.

The promoter of FDPS shows very good ChIP-chip signal (search FDPS in human_ChIP_chip table). The values are good in all ChIP experiments we tested.

By clicking the human_CRE_prediction link at the last column, we can see that the promoter contains a conserved half CRE site.

When we check the expression of FDPS in HEK293T cells by clicking the link, however, this gene was not induced by cAMP agonist FSK. This is not surprising. Actually only a small subset of CREB target genes are induced by cAMP in each tissues. The genes that do not repsond to cAMP here may become repsonsive in another tissue, or they respond to a differnt sitmulus. Alternatively, CREB may play a role in basal transcription. Actually FDPS expression is reduced by ACREB treatment, suggesting CREB may be involvled in basel transcription of this gene.

Similarly, FDPS is not incuded by FSK in human pancreatic islet culture either.