Topological rules and anomalies in intramolecular G-quadruplex folding: a comprehensive study

G-quadruplexes (G4s) make knots in DNA and RNA. One-block G4s can adopt eight different topologies, yet their folding rules are not precisely known. A systematic study of 353 intramolecular G4s enabled us to establish certain characteristics that rule their topology. In the absence of restraints imposed by a ligand or modified guanosines, when two loops are one nucleotide-long (nt), the topology is parallel, regardless of the third loop length. When loops 1 and 3 are 2-nt long, loop 2 length (L2) determines the G4 topology: parallel for L2 = 1, parallel or antiparallel-chair for L2 = 2, antiparallel-chair for L2 = 3, and antiparallel-basket for L2 = 4-5 nts. Diagonal loops require at least 4 nts; the only three structures with a diagonal loop of 3 nts present multiple anomalies. Only three topologies can accommodate a diagonal loop; the other five cannot. This shows the importance of distinguishing between the eight different topologies. We explain why parallel and hybrid topologies form preferentially three-tetrad G4s, while antiparallel topologies form preferentially two-tetrad G4s. We also explain why bulges between the G4 stem and the hairpin loop in quadruplex-duplex hybrids are sometimes needed. These findings clarify G4 folding principles and topological restraints.