G-Quadruplex

A G-quadruplex (G4) is a non-canonical four-stranded secondary structure formed by guanine-rich nucleotides in single-stranded DNA or messenger RNA sequences matching the canonical consensus motif G₃₊N_1–7G₃₊N_1–7G₃₊N_1–7G₃₊. Unlike canonical right-handed Watson–Crick B-DNA duplexes, a G-quadruplex is built from two or more stacked planar G-quartets (also termed G-tetrads), in which four guanine bases associate in a cyclic square-planar array via eight Hoogsteen hydrogen bonds. These stacked aromatic planes are electrostatically stabilized by dehydrating monovalent cations—following the thermodynamic selectivity series K⁺ > Na⁺ >> Li⁺—that coordinate the eight inward-pointing O6 carbonyl oxygens located within the central axial channel.

Across the human genome, high-throughput G4-seq mapping and predictive scoring algorithms such as G4Hunter identify more than 700,000 putative G-quadruplex-forming sequences. Rather than distributing randomly across intergenic space, G-quadruplex motifs exhibit strong evolutionary enrichment within telomeric (TTAGGG)_n repeats, human subtelomere regions (Chr1p36.33, Chr16p13.3), proto-oncogene promoter elements (MYC, KRAS, KIT, BCL2), 5′ untranslated regions, and GC-rich expansion segments of eukaryotic 28S ribosomal RNA (RN28S1). Within living cells, folded G-quadruplexes act as reversible biophysical switches that regulate transcription, translation, and telomere maintenance, while unscheduled G4 persistence stalls DNA replication forks unless unwound by specialized ATP-dependent helicases (PIF1, WRN, BLM, and FANCJ).

Hoogsteen G-Quartets and Central Monovalent Ion Channels

The fundamental building block of every G-quadruplex is the G-quartet, a coplanar macrocycle of four guanine bases in which each guanine donates two hydrogen bonds from its Watson–Crick face (the exocyclic C2-amino group and the endocyclic N1 imino proton) to the Hoogsteen face (the N7 nitrogen and the O6 carbonyl oxygen, respectively) of its neighbor. This cyclic arrangement generates eight hydrogen bonds per quartet and positions four electronegative O6 keto oxygens toward the center of the tetrad, creating a deep negative electrostatic potential well along the central helical axis.

Without counterion shielding, electrostatic repulsion among the eight converged O6 carbonyl oxygens of two adjacent G-quartets prevents stable folding at physiological temperature. Monovalent cations enter the central channel to coordinate these carbonyl oxygens in either a square-antiprismatic geometry (intercalated between two adjacent G-quartet planes) or a coplanar geometry. Potassium ions (K⁺, ionic radius ~1.38 Å) possess the optimal ionic radius to sit equidistantly between two stacked G-quartets with an eight-coordinate O6–K⁺ bond distance of ~2.7–2.9 Å and a comparatively low hydration free energy penalty, yielding thermal melting temperatures (T_m) often exceeding 65–85 °C in 100 mM KCl. Sodium ions (Na⁺, ionic radius ~1.02 Å) are small enough to reside within or slightly above a single G-quartet plane, but incur a larger dehydration enthalpy penalty, reducing T_m by 15–30 °C relative to K⁺. Lithium ions (Li⁺, ionic radius ~0.76 Å) carry a high charge density and tight hydration shell that resists dehydration, failing to stabilize the central channel.

Strand Stoichiometry, Glycosidic Torsion Angles, and Loop Topologies

G-quadruplexes exhibit remarkable structural polymorphism governed by strand stoichiometry (intramolecular unimolecular folds versus intermolecular bimolecular or tetramolecular assemblies), relative strand polarity, glycosidic torsion angles (syn versus anti), and the length and nucleotide composition of intervening loops (N_1–7). In a parallel G-quadruplex, all four G-tracts run in the same 5′ → 3′ orientation, forcing every deoxyguanosine or guanosine residue to adopt the anti glycosidic conformation and connecting the top of one G-tract to the bottom of the next via double-chain-reversal (propeller) loops. RNA G-quadruplexes (rG4s) fold almost universally into parallel topologies because the 2′-hydroxyl group of ribose sterically favors the C3′-endo sugar pucker, which restricts purines to the anti glycosidic angle.

In contrast, DNA G-quadruplexes readily adopt antiparallel or hybrid (3+1) topologies in which at least one G-tract runs opposite to the others. Antiparallel strands require alternating syn and anti glycosidic bonds around each G-quartet to preserve cyclic Hoogsteen hydrogen bonding, and connect adjacent G-tracts via edgewise (lateral) loops or diagonal loops spanning across the terminal quartet. Human telomeric repeats (TTAGGG)₄ illustrate this sensitivity to chemical environment: the same 22-nt sequence folds into an antiparallel basket structure in Na⁺ solution, a parallel propeller structure in crowded crystalline environments, and mixed hybrid-1/hybrid-2 (3+1) conformations in physiological K⁺ solution.

The G4Hunter Algorithm and Compositional Skew Scoring

Early bioinformatic surveys searched genomes using strict regular expressions (G{3,5}N{1,7}G{3,5}N{1,7}G{3,5}N{1,7}G{3,5}), identifying ~376,000 canonical motifs in the human reference assembly. However, biophysical assays demonstrated that many stable G-quadruplexes contain interrupted G-runs (bulges) or loops longer than 7 nt, while C-rich tracts embedded inside a window compete via Watson–Crick base pairing to trap the sequence in a standard B-DNA duplex. To overcome these limitations, Laurent Lacroix, Jean-Louis Mergny, and colleagues developed G4Hunter, a sliding-window quantitative algorithm that scores both G-richness and G/C skewness simultaneously.

In G4Hunter, each nucleotide position within a sliding window (typically 25 nt) is assigned a run-length-weighted score: isolated guanines receive +1, a dinucleotide GG receives +2 per base, GGG receives +3 per base, and G_n (with n ≥ 4) is capped at +4 per base. Cytosines receive symmetric negative penalties (−1, −2, −3, or −4) to model competing duplex stem-loop formation, while adenines and thymines score 0. The window score is the arithmetic mean of its constituent base scores. Windows with |G4Hunter score| ≥ 1.2 moderately favor quadruplex formation, whereas windows exceeding ≥ 1.5 (and especially ≥ 1.75 to 2.0) fold into thermally stable G-quadruplexes in vitro with >90% positive predictive precision.

Subtelomeric Clusters, Oncogene Promoters, and 28S rDNA Expansion Segments

Applying G4Hunter and high-resolution potassium/PDS-stalled sequencing (G4-seq) across the telomere-to-telomere human genome reveals sharp positional clustering. Within human subtelomere regions—such as the terminal 500 kb of Chr1p36.33, Chr7p22.1, and Chr16p13.3—G-quadruplex density rises 4-fold to 8-fold above the genome-wide median, often co-localizing with alternating purine–pyrimidine Z-DNA tracts inside GC-rich segmental duplications and variable number tandem repeats (VNTRs). At chromosome ends, telomeric repeat-containing RNA (TERRA) transcribed from subtelomeric promoters folds into parallel RNA G-quadruplexes that bind the telomere shelterin protein TRF2 and chromatin modifiers to maintain heterochromatic compaction.

Within protein-coding loci, more than 40% of human gene promoters harbor at least one G4 motif within 1 kb upstream of the transcription start site. In the MYC proto-oncogene, the nuclease hypersensitive element III₁ (NHE III1) contains a 27-nt purine-rich tract that forms a parallel G-quadruplex with a T_m > 85 °C, functioning as a transcriptional silencer when stabilized. Similarly, eukaryotic 45S ribosomal DNA arrays (rDNA) and mature 28S rRNA molecules carry dense clusters of G4 motifs in their GC-rich expansion segments (ES7L, ES15L, ES27L), where they create solvent-exposed structural scaffolds on the periphery of the 60S cytosolic ribosome.

Replication-Fork Stalling and Specialized G4 Helicases

During S-phase DNA replication or active RNA Polymerase II elongation, transient unwinding of the B-DNA duplex exposes single-stranded lagging-strand templates or non-template coding strands, promoting spontaneous G-quadruplex folding behind the advancing replisome. Because standard replicative DNA polymerases (Pol δ and Pol ε) cannot traverse a potassium-stabilized G4 knot, unresolved quadruplexes arrest fork progression, induce single-stranded gaps, and trigger DNA double-strand breaks and epigenetics memory loss due to uncoupled histone recycling.

Cells deploy a specialized repertoire of ATP-dependent G4 helicases to unwind these structures prior to fork collapse. The 5′ → 3′ superfamily 1B helicase PIF1 recognizes single-stranded 5′ tails adjacent to G-quadruplexes and repetitively unfolds the tetrad core using energy derived from ATP hydrolysis. Two RecQ-family 3′ → 5′ helicases—WRN (mutated in Werner premature aging syndrome) and BLM (mutated in Bloom syndrome)—contain conserved RQC (RecQ C-terminal) and HRDC domains that bind G4 planes and unwind telomeric and subtelomeric quadruplexes during lagging-strand synthesis. Meanwhile, the iron–sulfur (Fe-S) cluster 5′ → 3′ helicase FANCJ (BRIP1), associated with the Fanconi anemia DNA repair pathway, collaborates with the translesion polymerase REV1 to replicate through G4 obstacles and preserve local DNA methylation and histone modification patterns across cell divisions.

Small-Molecule G4 Stabilizers in Oncology

Because cancer cells frequently exhibit heightened oncogene transcription (MYC, KRAS) or deficiencies in homologous recombination (BRCA1/BRCA2 loss) and G4 helicase pathways, synthetic small-molecule G4 ligands have emerged as targeted therapeutic agents in drug discovery. Pyridostatin (PDS) and Phen-DC3 are planar heteroaromatic compounds engineered with cationic side chains that π-stack onto the external 5′ or 3′ G-quartet faces with nanomolar dissociation constants (K_d ≈ 10–100 nM), clamping the quadruplex against helicase unwinding and inducing replication-dependent DNA damage at telomeres and oncogene promoters.

Clinically advanced G4 ligands exploit synthetic lethality in DNA-repair-deficient tumors. CX-5461 (pidnarulex) and CX-3543 (quarfloxin) stabilize G-quadruplexes at rDNA promoters and replication forks; in BRCA1/BRCA2-deficient breast and ovarian carcinomas or PALB2-mutant tumors, replication forks stalled at CX-5461-trapped G-quadruplexes cannot be restarted via homologous recombination, leading to selective tumor cell apoptosis.

Why do potassium ions stabilize G-quadruplexes much more effectively than sodium or lithium ions?

Potassium ions (K⁺, ionic radius ~1.38 Å) fit cleanly into the bipyramidal antiprismatic cavity between two stacked G-quartets, coordinating all eight inward-pointing O6 carbonyl oxygens simultaneously. More importantly, K⁺ has a lower free energy of hydration (ΔG_hydr ≈ −295 kJ mol⁻¹) than Na⁺ (−365 kJ mol⁻¹) or Li⁺ (−475 kJ mol⁻¹). Because cations must shed their bound water molecules before entering the narrow central channel, the favorable balance between electrostatic coordination and easier dehydration makes K⁺ the strongest thermodynamic stabilizer.

How do RNA G-quadruplexes differ structurally from DNA G-quadruplexes?

RNA G-quadruplexes (rG4s) contain a 2′-hydroxyl (2′-OH) group on every ribose ring. This 2′-OH group sterically locks the ribose sugar into the C3′-endo (North) conformation, which energetically disfavors the syn glycosidic bond angle required for antiparallel strand folding. Consequently, unmodified RNA G-quadruplexes adopt parallel topologies with all-anti guanosines and exhibit higher thermal stability (ΔT_m of +10 to +25 °C) than homologous DNA sequences due to additional 2′-OH intramolecular hydrogen-bonding networks.

What is the G4Hunter threshold for biologically stable G-quadruplexes?

The G4Hunter algorithm assigns positive scores (+1 to +4) to consecutive guanines and negative scores (−1 to −4) to consecutive cytosines across a 25-nt sliding window. While a score of |G4H| ≥ 1.2 identifies candidate sequences capable of transient quadruplex formation, biophysical validation by circular dichroism and thermal difference spectroscopy demonstrates that windows scoring |G4H| ≥ 1.5 to 2.0 reliably fold into stable G-quadruplexes under physiological potassium concentrations (100–150 mM K⁺).

1. Burge, S., Parkinson, G. N., Hazel, P., Todd, A. K., & Neidle, S. (2006). "Quadruplex DNA: sequence, topology and structure." Nucleic Acids Research, 34(19), 5402–5415. https://doi.org/10.1093/nar/gkl655
2. Bedrat, A., Lacroix, L., & Mergny, J. L. (2016). "Re-evaluation of G-quadruplex propensity with G4Hunter." Nucleic Acids Research, 44(4), 1746–1759. https://doi.org/10.1093/nar/gkw006
3. Chambers, V. S., Marsico, G., Boutell, J. M., Di Antonio, M., Smith, G. P., & Balasubramanian, S. (2015). "High-throughput sequencing of DNA G-quadruplex structures in the human genome." Nature Biotechnology, 33(8), 877–881. https://doi.org/10.1038/nbt.3295
4. Varshney, D., Spiegel, J., Zyner, K., Tannahill, D., & Balasubramanian, S. (2020). "The regulation and functions of DNA and RNA G-quadruplexes." Nature Reviews Molecular Cell Biology, 21(8), 459–474. https://doi.org/10.1038/s41580-020-0236-x