Download Supplementary Data

Please select the files you want to download:

1. Supplementary Tables:

Table S1 (Table 2 and 3 in PNAS paper) Promoter prediction of CRE sites for RefSeq genes	Read Legends	Download File a. human; b. mouse
Table S2 (Table 4 in PNAS paper) GO analysis of putative CREB Target genes	Read Legends	Download File
Table S3 (Table 5 in PNAS paper) Details about 40 full CRE containing genes in HEK293T cells	Read Legends	Download File
Table S4 (Table 6 in PNAS paper) ChIP-Chip data of CREB binding in human tissues	Read Legends	Download File
Table S5 (Table 7 in PNAS paper) cAMP responsive genes in human and mouse tissues	Read Legends	Download File
Table S6 (Table 8 in PNAS paper) Other TF sites that co-seggregate with CREB binding sites	Read Legends	Download File

Download Legends for all Tables.

2. Taining sequence and profile Hidden Markov Models (pHMMs) for Full and Half CREs
Download ZIP File
Known CREB target genes (from literature or from our unpublished ChIP results) were chosen and the CRE sites and flanking sequences were used as training sets. A total of 36 full CRE sites and 50 half CRE sites were used. For each training set, CRE sites and any conserved flanking sequences were discovered by MEME . In MEME, the default parameters were used, except for half CRE sites, for which the minimum width was set to 5. The output alignment from MEME was used to generate the pHMM using the HMMER package . Default parameters were used and the models are calibrated. The training set sequences, MEME output, alignments used as input for HMMER, and the final HMM models were available in supplementary file NewHMMModels.zip.

3. Primers used for Manual ChIPs
Download File
For each gene, the two primers are named Symbol_F and Symbol_R.
For CREs occur at intergenic regions, the chromosome location is used as symbol (chr_position), and "c" at the end of the symbol means it is conserved in rodents.

4. Supplementary Methods
Download File

Legends for Supplementary Tables

Tabel S1. Computational prediction of functional CREs on promoters of human (a) and mouse (b) RefSeq genes.
Download File a. human; b. mouse

The names for most columns are self-explanatory; Details about columns for computational results are listed below:

CRE Flag: Flags, or markers for the types of CREs on the promoter (-3Kb to 300bp from transcription start site). "F/f" means full site, "H/h" means half site, with upper case representing conserved CRE. "FH" is a full site CRE in the species studied but only appear as half site in other species. "T/t" at the end means the presence of a TATA box less than 300bp downstream of the CRE.
Full Site/Half Site: All occurrences of full site (TGACGTCA) or half site(TGACG/CGTCA) CREs in -5Kb-1Kb region of the transcription start site (TSS). Here the position of each CRE relative to the TSS is shown following its flag.
Conserved CRE: All the conserved full and half site are shown in this column since they are the most likely functional CREs. These CREs are shown similar to the format used by Full site/Half site, with the addition of the distance to closest downstream TATA box, shown as the last number.
CRECluster: Here all the CRE clusters are shown; In a CRE cluster, neighbors must be closer than 50bp to each other; If more than two clusters appear in a promoter, they are separated by "|".
pHMM+Pos: If there is a positional conserved CRE identified by pHMM search in human and mouse orthologs, it will be shown here. The values shown are 1) models used(F: full site model; H: half site model) in human and mouse, respectively; 2) Human match score and position; 3) Mouse match and position. Here the position is relative to the start codon.
Chrom/Strand/Pos: Chromosomal position of the TSS and the strand of the transcript based on hg16 human genome assembly or mm3 mouse genome assembly from the UCSC Genome Bioinformatics Site.

Table S2. a) GO categories that are enriched in CRE containing genes. Download File
Gene categories which are over-represented in the CRE_all list comparing to all RefSeq genes (P<0.005) are shown with the CRE containing genes in the categories listed. The "RefSeq Total" consists of all RefSeq genes which are annotated within a given system of classifying genes (e.g. the "Molecular Function" branch of the Gene Ontology). Therefore the population can change from one system to the next, depending on the coverage of annotations. "Hits from RefSeq" refers to genes falling within the gene category in question. "CRE_All Total" refers to the subset of CRE containing genes from " RefSeq Total", and "Hits from CRE_All" refers to the number of CRE containing genes that fall within a specific category.
b). Representative list of transcription factors that are potential CREB targets.

Table S3. The methylation state, CREB occupancy and cAMP responsiveness of 40 full CRE containing genes in HEK293T cells. Download File

Table S4. Genome-wide location analysis (ChIP-chip) of CREB binding in HEK293T cells and in hepatocytes. Download File
Both CREB and P-CREB ChIPs performed in each tissue types except for pancreatic islets, where only P-CREB yielded useful data. In HEK293T cells, three time points, 0hr, 1hr and 4hr after forskolin stimulation are shown, as well as the average of the three time points for CREB ChIPs. Cutoff used to choose ChIP positive genes are: p-value<=0.001 (in red) and binding ratio>=2 (in blue). For RefSeq genes whose promoters have been analyzed (Table S1), the prediction of CRE sites is shown. The last column shows locations of probes that have been mapped to the May 2004 assembly of the human genome. The list is sorted by the p-values from the average of CREB ChIP in HEK293T cells.

Table S5. The lists of cAMP responsive genes from HEK293T, Min6, hepatocytes and pancreatic islets based on gene profiling using Affymetrix arrays.Download File
The values are lower bounds of the 90% confidence intervals of fold changes; In other words, if the lower bound value is 1, one can say with 90% probability that the gene expression is increased (fold change >=1). Gene expression induced by cAMP treatment (forskolin) is shown in "FSK induction" columns; while inhibition by dominant negative A-CREB is shown A-CREB Inhibition columns. For hepatocytes, in addition to forskolin treatment in tissue culture, liver samples from fasting and feeding animals are compared. Each tissue is in a separate worksheet.

Table S6. Transcription binding sites that tend to occur together with CREB binding sites.Download File
a. Conserved transcription factor binding sites that occur much more often in the CRE_All list than in all RefSeq genes are listed in the order of their P-values. The factor names, consensus sequences in IUPAC code, and matrix IDs are from the TransFac database. b. Conserved transcription factor binding sites that occur much more often in ChIP-chip positive genes than in all genes on the promoter array are listed in the order of their P-values. CREB ChIP positive genes from HEK293T were used.