Greglist: a database listing potential G-quadruplex regulated genes.

The double helix is a conformation that genomic DNA usually assumes; under certain conditions, however, guanine-rich DNA sequences can form a four-stranded structure, G-quadruplex, which is found to play a role in regulating gene expression. Indeed, it has been demonstrated that the G-quadruplex formed in the c-MYC promoter suppresses its transcriptional activity. Recent studies suggest that G-quadruplex motifs (GQMs) are enriched in human gene promoters. To facilitate the research of G-quadruplex, we have constructed Greglist, a database listing potentially G-quadruplex regulated genes. Greglist harbors genes that contain promoter GQMs from genomes of various species, including humans, mice, rats and chickens. Many important genes are found to contain previously unreported promoter GQMs, such as ATM, BAD, AKT1, LEPR, UCP1, APOE, DKK1, WT1, WEE1, WNT1 and CLOCK. Furthermore, we find that not only protein coding genes, 126 human microRNAs also contain promoter GQMs. Greglist therefore provides candidates for further studying G-quadruplex functions and is freely available at http://tubic.tju.edu.cn/greglist.

INTRODUCTION

The double helix structure is a conformation that genomic DNA usually assumes; however, DNA can form other non-classical structures as well (1). For instance, under certain conditions, guanine-rich DNA sequences can form a special structure called G-quadruplex. The discovery of G-quadruplex can be traced back to G-quartets, planar arrays of four guanines held together by hydrogen bonds, which were found by Davies and coworkers (2) about 5 decades ago. Later Sen and Gilbert (3) discovered G-quadruplex, a four-stranded structure that is stabilized by G-quartets. As an example, readers may visit www.rcsb.org to view the 3-dimensional (3D) structure of a G-quadruplex (PDB code: 1XAV), which is formed in the promoter regions of the c-MYC gene (4). Sequences with high potential to form G-quadruplex have been found in many different genomic regions, suggesting diverse roles of G-quadruplexes (5–11). For instance, telomeric repeats in virtually all eukaryotes have the ability to form G-quadruplexes (10,11), offering a protection for the telomere 3′ overhang (12,13), which is essential for cell survival.

Recent interests on G-quadruplexes have been focused on its role in transcriptional regulation. By using electron microscopy, Maizels and coworkers (14) observed that the G-quadruplex structure is formed cotranscriptionally in vivo. Indeed, Hurley and coworkers have demonstrated that the region upstream of the c-MYC promoter forms a G-quadruplex, removal of which results in an increase, whereas its stabilization results in a decrease in basal transcriptional activity of this promoter, suggesting promoter G-quadruplexes as transcriptional repressor elements (15).

Sequences containing G-quadruplex motifs (GQMs) in promoter regions have only been reported for about 10 genes, including c-MYC (20,15–17), VEGF (18), BCL-2 (19), c-KIT (21) and some others (22,23). Recent bioinformatics studies, however, showed that GQMs are prevalent in the human genome (24,25). Furthermore, GQMs were found to be highly enriched in human gene promoters with more than 40% promoters containing at least 1 GQM (26).

To facilitate the study of the role of promoter G-quadruplexes, we constructed Greglist, a database listing potential G-quadruplex REgulated Genes, i.e. genes that contain promoter GQMs. The database provides detailed information about the number, the position and the sequence of promoter GQMs from genes of various species. Many important genes are found to contain previously unreported promoter GQMs, such as ATM, BAD, AKT1, LEPR, UCP1, APOE, DKK1, WT1, WEE1, WNT1 and CLOCK. Furthermore, we found that not only protein coding genes, 126 human microRNAs also contain promoter GQMs. Greglist contains candidates for further studying G-quadruplex functions and is another device added to the existing online G-quadruplex toolbox.

DATABASE CONSTRUCTION AND DESCRIPTION

Greglist of the current version contains genes that have promoter GQMs in the genomes of human, mouse, rat and chicken. Table 1 provides a descriptive statistics of the content of the database. We generally defined sequences 1 kb upstream of transcription start site (TSS) as promoter regions. These sequences were downloaded from Ensembl using the software BioMart. The dataset used was Ensembl 45 and human, mouse, rat and chicken genome sequences were based on the versions of NCBI36, NCBIM36, RGSC3.4 and WASHUC2, respectively. The software Quadparser (26) was used to find the promoter GQM, which is G3+N1−7G3+N1−7G3+N1−7G3+, where N denotes any nucleotide. In addition, the G-quadruplex structure can be formed on either of the two DNA strands; therefore the motif of C3+N1−7C3+N1−7C3+N1−7C3+ was also used, which suggests the capability of the G-quadruplex formation on the complementary strand.

Table 1.

Descriptive statistics of genes in Greglist

Species (Latin name)	Species	Genome version	Number of genes having promoter GQMs	Total gene number	Percentage of genes having promoter GQMs	Average GQMs a gene has	GQM density in promoter regions (GQMs/Kb)	Average GQM length (mean ± SD)
Homo sapiens	Human	NCBI36	10 277	31 524	32.60%	1.93	0.63	29.19 ± 13.57
Mus musculus	Mouse	NCBIM36	8962	28 390	31.57%	1.61	0.51	28.18 ± 12.57
Rattus norvegicus	Rat	RGSC3.4	7013	27 302	25.69%	1.43	0.37	26.39 ± 8.82
Gallus gallus	Chicken	WASHUC2	5949	17 438	34.12%	1.75	0.60	28.70 ± 14.44

So far, only about 10 genes have been reported to contain promoter GQMs. In Greglist, however, a lot more genes that contain promoter GQMs are listed. For instance, these genes include ATM, BAD, AKT1, LEPR, UCP1, APOE, DKK1, WT1, WEE1, WNT1, CLOCK, ATF1 and BMP2, which have critical functions in various cellular processes, such as apoptosis and transcriptional regulation. Table 2 lists a sample of 30 genes that contain promoter GQMs with the position of GQMs and gene functions.

Table 2.

A list of 30 human genes that have not been previously reported to contain promoter G-quadruplex motifs

No.	Abbreviation	Gene name	Ensembl ID	Function or associated disease	Reference	Number of GQM	Distance to TSS
1	WNT1	Wingless-type MMTV integration site family, member 1	ENSG00000125084	The Wnt signaling pathway, CNS development	(30)	1	193
2	WNT5A	Wingless-type MMTV integration site family, member 5A	ENSG00000114251	The Wnt signaling pathway, vertebrate development	(31)	2	567, 936
3	LEPR	LEPTIN receptor	ENSG00000116678	Energy metabolism	(32)	3	310, 372, 495
4	UCP1	Uncoupling protein 1	ENSG00000109424	Energy metabolism	(33)	2	89, 224
5	APOE	Apolipoprotein E	ENSG00000130203	Alzheimer's disease	(34)	4	46, 65, 407, 739
6	ATM	Ataxia telangiectasia mutated	ENSG00000149311	Ataxia telangiectasia	(35)	1	59
7	PAX8	Paired box gene 8	ENSG00000125618	Permanent congenital hypothyroidism	(36)	1	133
8	SOX1	SRY (sex determining region Y)-box 1	ENSG00000203883	Lens development	(37)	3	80, 726, 826
9	SOX10	SRY (sex determining region Y)-box 10	ENSG00000100146	Waardenburg–Hirschsprung disease	(38)	2	130, 313
10	HDAC1	Histone deacetylase 1	ENSG00000116478	Histone modification	(39)	1	34
11	TGFβ1	Transforming growth factor, beta 1	ENSG00000105329	TGFβ signaling	(40)	1	151
12	SMAD2	MAD homolog 2	ENSG00000175387	TGFβ signaling	(41)	2	235, 450
13	DKK1	Dickkopf homolog 1	ENSG00000107984	TGFβ signaling	(42)	1	136
14	CLOCK	Clock homolog	ENSG00000134852	Circadian rhythms	(43)	3	147, 341, 692
15	WEE1	WEE1 homolog	ENSG00000166483	Cell cycle control	(44)	1	542
16	BAD	BCL2-antagonist of cell death	ENSG00000002330	Apoptosis	(45)	3	116, 628, 756
17	AKT1	V-akt murine thymoma viral oncogene homolog 1	ENSG00000142208	Apoptosis	(46)	1	61
18	GATA4	GATA-binding protein 4	ENSG00000136574	Heart development	(47)	1	314
19	MYOD1	Myogenic differentiation 1	ENSG00000129152	Muscle development	(48)	2	128, 216
20	WT1	Wilms tumor 1	ENSG00000184937	Kidney development	(49)	2	168, 900
21	GDF1	Growth differentiation factor 1	ENSG00000135414	Left–right patterning	(50)	4	78, 166, 327, 766
22	BMP2	Bone morphogenetic protein 2	ENSG00000125845	Bone development	(51)	1	163
23	MEF2D	MADS box transcription enhancer factor 2D	ENSG00000116604	Heart development	(52)	4	18, 85, 169, 232
24	STAT6	Signal transducer and activator of transcription 6	ENSG00000166888	Immunity	(53)	1	505
25	SOCS1	Suppressor of cytokine signaling 1	ENSG00000185338	Immunity	(54)	5	112, 211, 534, 578, 758
26	MMP2	Matrix metallopeptidase 2	ENSG00000167346	Function of extracellular matrix	(55)	1	576
27	MAPK2	Mitogen-activated protein kinase 2	ENSG00000162889	MAP kinase pathway	(56)	2	100, 137
28	ATF1	Activating transcription factor 1	ENSG00000123268	Transcriptional regulation	(57)	1	36
29	TAF2	TAF2 RNA polymerase II	ENSG00000064313	Transcriptional regulation	(58)	1	296
30	RING1	Ring finger protein 1	ENSG00000204227	Transcriptional regulation	(59)	4	501, 559, 677, 938

In addition, we found that not only protein coding genes, many microRNAs, such as hsa-mir-639 and hsa-mir-381, also contain promoter GQMs. Totally 126 human microRNAs were found to have promoter GQMs. To get a full list of these microRNAs, refer to the Supplementary Table 1. MicroRNAs have emerged as important regulators of gene expression. The finding that promoter regions of microRNA genes contain GQMs necessitates further studies to address the role of G-quadruplexes in microRNA regulation.

Of note, the presence of a GQM only suggests the potential of a sequence to form G-quadruplex. In addition, the G-quadruplex structure is a dynamic structure that is formed upon denaturation of the DNA duplex. Therefore caution must be taken to interpret the data in Greglist. In other words, gene records in Greglist provide a starting point for further analysis of the potential G-quadruplex structure in these genes. Furthermore, Huppert et al. (26) reported that more than 40% of human genes contain promoter GQMs, however, in Greglist, ∼32% human genes do. This is likely because in Ref. (26), only less than 20 000 known genes were used, whereas in the current study, more than 30 000 human genes, including those classified as novel and those encode RNAs were included. Therefore, Greglist is made to be inclusive, not exclusive.

Gene names, Ensembl IDs, RefSeq IDs, numbers of GQMs, distance of the GQM to TSS, functional description of gene ontology, sequences containing the GQM and coding sequences of the gene, were extracted from Ensembl database and Quadparser output files. All the data were then organized by using an open-source management system, MySQL, which allows rapid data retrieval. All gene records have been linked directly to corresponding entries in Ensembl. Users can browse each entry or download all records. Because of the large volume of data, a good searching function is important for this database. In Greglist, users can perform searches by inputting gene accession numbers or names at the homepage, and then click ‘Go’. To perform more detailed searches, users can click ‘Search’, and then in the new page, more detailed searching options are provided. For instance, users can search by gene ontology terms to get a list of genes that have desired functions. To further facilitate searching the gene of interest, we installed Blast program locally. So users can input the coding sequence of their gene of interest and perform Blast searches to find homologous ones.

Many online resources for G-quadruplexes are available. These include G4P calculator (14), QGRS Mapper (27), Quadfinder (28), which are online programs or web servers for predicting G-quadruplexes. GRSDB (29) is a database of quadruplex forming G-rich sequences in alternatively processed mammalian pre-mRNA sequences. Greglist is another device added to the existing online G-quadruplex toolbox.

We plan to include more species in future versions of Greglist. In addition, with the availability of more experimental data, we plan to integrate experimental evidence in corresponding entries. Furthermore, although the GQM used in Quadparser is quite commonly used, there are other motifs that have potential to form G-quadruplexes, and we also plan to include these motifs in future versions of the database. We welcome users’ comments, corrections and new information, which will be used for updating.

Greglist is freely available at the website: http://tubic.tju.edu.cn/greglist, and should be cited with the present publication as reference.

SUPPLEMENTARY DATA

Supplementary Data are available at NAR Online.