G-quadruplex

Nucleic acid sequences which are rich in guanine are capable of forming four-stranded structures called G-quadruplexes (Also known as G-tetrads or G4-DNA). These consist of a square arrangement of guanines (a tetrad), stabilized by Hoogsteen hydrogen bonding. They are further stabilized by the existence of a monovalent cation (especially potassium) in the center of the tetrads. They can be formed of DNA, RNA, LNA and PNA, and may be intramolecular, bimolecular or tetramolecular. Depending on the direction of the strands or parts of a strand that form the tetrads, structures may be described as parallel or antiparallel.

Structure of a G-quadruplex. Left: a G-tetrad. Right: an intramolecular G-quadruplex

3D Structure of the intramolecular human telomeric G-quadruplex in potassium solution (PDB ID 2HY9). The backbone is represented by a tube. The center of this structure contains three layers of G-tetrads. The hydrogen bonds in these layers are represented by blue dashed lines.

Telomeric quadruplexes

Telomeric repeats in a variety of organisms have been shown to form these structures in vitro, and they have also been shown to form in vivo in some cases. The human telomeric repeat (which is the same for all vertebrates) consists of many repeats of the sequence d(GGTTAG), and the quadruplexes formed by this structure have been well studied by NMR and X-ray crystal structure determination. The formation of these quadruplexes in telomeres has been shown to decrease the activity of the enzyme telomerase, which is responsible for maintaining length of telomeres and is involved in around 85% of all cancers. This is an active target of drug discovery.

Non-telomeric quadruplexes

Recently, there has been increasing interest in quadruplexes in locations other than at the telomere. This was given a large boost by the work by Hurley on the proto-oncogene c-myc, which was shown to form a quadruplex in a nuclease hypersensitive region critical for gene activity. Since then, many other genes have been shown to have G-quadruplexes in their promoter regions, including the chicken β-globin gene, human ubiquitin-ligase RFP2 and the proto-oncogenes c-kit, bcl-2, VEGF, H-ras and N-ras. This list is ever-increasing.

Genome-wide surveys based on a quadruplex folding rule have been performed, which have identified 376,000 Putative Quadruplex Sequences (PQS) in the human genome, although not all of these probably form in vivo. A similar study has identified putative G-quadruplexes in prokaryotes.There are several possible models for how quadruplexes could control gene activity, either by upregulation or downregulation. One model is shown below, with G-quadruplex formation in or near a promoter blocking transcription of the gene, and hence de-activating it. In another model, quadruplex formed at the non-coding DNA strand helps to maintain an open conformation of the coding DNA strand and enhance an expression of the respective gene.

Model for quadruplex-mediated down-regulation of gene expression

Ligands which bind quadruplexes

One way of inducing or stabilizing G-quadruplex formation, is to introduce a molecule which can bind to the G-quadruplex structure, and a number of ligands, both small molecules and proteins, have been developed which can do so. This has become an increasingly large field of research.

A number of naturally occurring proteins have been identified which selectively bind to G-quadruplexes. These include the helicases implicated in Bloom's and Werner's syndromes and the Saccharomyces cerevisiae protein RAP1. An artificially derived three zinc finger protein called Gq1, which is specific for G-quadruplexes has also been developed, as have specific antibodies.

Cationic porphyrins have been shown to bind intercalatively with G-quadruplexes, as well as the molecule telomestatin.

Quadruplex prediction techniques

Identifying and predicting sequences which have the capacity to form quadruplexes is an important tool in further understanding of their role. A rule for predicting the formation has been proposed, where sequences are predicted to fold based on the pattern d(G₃₊N_1-7G₃₊N_1-7G₃₊N_1-7G₃₊), where N is any base (including guanine). This rule has been widely used in on-line algorithms.