Mapping and characterization of G-quadruplexes in the genome of the social amoeba Dictyostelium discoideum

Abstract

G-quadruplexes (G4) are non-canonical DNA and/or RNA secondary structures formed in guanine-rich regions. Given their over-representation in specific regions in the genome such as promoters and telomeres, they are likely to play important roles in key processes such as transcription, replication or RNA maturation. Putative G4-forming sequences (G4FS) have been reported in humans, yeast, bacteria, viruses and many organisms. Here we present the first mapping of G-quadruplex sequences in Dictyostelium discoideum , the social amoeba. ‘Dicty’ is an ameboid protozoan with a small (34 Mb) and extremely AT rich genome (78%). As a consequence, very few G4-prone motifs are expected. An in silico analysis of the Dictyostelium genome with the G4Hunter software detected 249–1055 G4-prone motifs, depending on G4Hunter chosen threshold. Interestingly, despite an even lower GC content (as compared to the whole Dicty genome), the density of G4 motifs in Dictyostelium promoters and introns is significantly higher than in the rest of the genome. Fourteen selected sequences located in important genes were characterized by a combination of biophysical and biochemical techniques. Our data show that these sequences form highly stable G4 structures under physiological conditions. Five Dictyostelium genes containing G4-prone motifs in their promoters were studied for the effect of a new G4-binding porphyrin derivative on their expression. Our results demonstrated that the new ligand significantly decreased their expression. Overall, our results constitute the first step to adopt Dictyostelium discoideum as a ‘G4-poor’ model for studies on G-quadruplexes.

Genome wide distribution of G-quadruplexes and their impact on gene expression in malaria parasites

“…G4Hunter therefore unveils 2 to 22 more sequences than previously described in P. falciparum [40]. This corresponds to densities (G4FS/kb) of 0.076 for G4H1.2 and 0.019 for G4H1.5 (Table 1), which are lower (by 1.6-fold) than in Dictyostelium discoideum, an organism whose genome harbours a similar high AT content (78%) [47,48]. G4FS were found on all chromosomes, with some chromosomic regions displaying higher G4 density, for instance chromosome ends and central regions ( Fig 1A).…”

“…falciparum [ 40 ]. This corresponds to densities (G4FS/kb) of 0.076 for G4H1.2 and 0.019 for G4H1.5 ( Table 1 ), which are lower (by 1.6-fold) than in Dictyostelium discoideum , an organism whose genome harbours a similar high AT content (78%) [ 47 , 48 ].…”

“…Analysis of gene-related features revealed that the exon regions exhibit a modest but significant enrichment of G4FS at a threshold of 1.2, while the promoter and intron regions are depleted in G4FS (Fig 2A and S3 Table). Furthermore, unlike the GC poor Dictyostelium genome [47], the G4 enrichment correlates with the relative GC content in various Plasmodium genome features. For instance, var genes exhibit the highest GC content (33%) and are associated with the highest G4 fold-enrichment, while promoters and introns have a GC content lower than genome average, which coincides with G4 depletion (Fig 2C and S3 Table).…”

“…primer extension reaction performed at 55˚C or above) and the parameters for G4 detection led this method to perform sub-optimally for the AT-rich genome of P. falciparum (~80% AT; FDR 30%). For this reason we turned to the G4Hunter algorithm, a tool that has been extensively used for the prediction of G4s in different types of genomes, including GC poor ones [41,47], to perform the most comprehensive analysis to date of G4FS in malaria parasites. To ensure that G4Hunter was indeed suitable to study the P. falciparum genome we undertook different comparison tests between our dataset and the one generated using the G4-seq method [31], and showed that former outperformed the latter in various instances.…”

RNA G-quadruplex structures exist and function in vivo

“…RG4 structures contains unique sequence features of GxLnGxLnGxLnGx (whereby G stands for Guanine; L stands for Loop, and x ≥2, n up to 15). For decades, in silico prediction based on sequence features was used to search for putative RG4s at the genome-wide scale in different organisms 6,27,37,38 . With the rise of deep sequencing methods, high throughput methods such as rG4-seq were developed to map G-rich regions with in vitro RG4 folding potential 275 throughout the transcriptome 25 .…”