Case studies on potential G-quadruplex-forming sequences from the bacterial orders Deinococcales and Thermales derived from a survey of published genomes

Abstract

Genomes provide a platform for storage of chemical information that must be stable under the context in which an organism thrives. The 2'-deoxyguanosine (G) nucleotide has the potential to provide additional chemical information beyond its Watson-Crick base-pairing capacity. Sequences with four or more runs of three G nucleotides each are potential G-quadruplex forming sequences (PQSs) that can adopt G-quadruplex folds. Herein, we analyzed sequenced genomes from the NCBI database to determine the PQS densities of the genome sequences. First, we found organisms with large genomes, including humans, alligators, and maize, have similar densities of PQSs (~300 PQSs/Mbp), and the genomes are significantly enriched in PQSs with more than four G tracks. Analysis of microorganism genomes found a greater diversity of PQS densities. In general, PQS densities positively tracked with the GC% of the genome. Exceptions to this observation were the genomes from thermophiles that had many more PQSs than expected by random chance. Analysis of the location of these PQSs in annotated genomes from the order Thermales showed these G-rich sequences to be randomly distributed; in contrast, in the order Deinococcales the PQSs were enriched and biased around transcription start sites of genes. Four representative PQSs, two each from the Thermales and Deinococcales, were studied by biophysical methods to establish the ability of them to fold to G-quadruplexes. The experiments found the two PQSs in the Thermales did not adopt G-quadruplex folds, while the two most common in the Deinococcales adopted stable parallel-stranded G-quadruplexes. The findings lead to a hypothesis that thermophilic organisms are enriched with PQSs as an unavoidable consequence to stabilize thermally their genomes to live at high temperature; in contrast, the genomes from stress-resistant bacteria found in the Deinococcales may utilize PQSs for gene regulatory purposes.

Figure 1

Characteristics of a G-quadruplex. (A) General sequence formula for a PQS. (B) Structure of a G-tetrad. (C) Cartoon representations for a parallel and antiparallel G4.

Figure 2

The density of PQSs in the genomes of select organisms as a function of the genome length, GC%, and percentage of PQSs with >4 G tracks.

Figure 3

Plots of PQS densities and GC% for the phyla of bacteria with representative genomes found in the NCBI database. There exist six phyla of bacteria (Caldiserica, Calditrichaeota, Chrysiogenetes, Dictyoglomi, Elusimicrobia, and Thermodesulfobacteria) that have fewer than four representative genomes sequenced; these data can be found in the supporting information (Fig. S2). Three panels represent superphyla of bacteria that include the FCB group (phyla = Fibrobacteres, Chlorobi, and Bacteroidetes), PVC group (phyla = Planctomycetes, Verrucomicrobia, Chlamydiae, and Lentisphaerae), and the Terrabacteria group (phyla = Cyanobacteria, Chlorflexi, and Deinococcus-Thermus).

Figure 4

Profiles of the PQSs found in the phylum Deinococcus-Thermus. (A) Plot of densities of PQSs vs. GC%. The red box plots represent the theoretical PQS densities determined from analysis of randomized genomes with defined GC%. (B) Distribution of PQSs around TSSs binned at 30 nts. (C) Distributions of PQSs on the coding or template strand in the orders Deinoccocales and Thermales. (D) Distributions of >4 G-track PQSs on the coding or template strand in the orders Deinoccocales and Thermales.

Figure 5

Hierarchical classification in the phylum Deinococcus-Thermus on the basis of the distribution of PQSs around TSSs. Unsupervised hierarchical classification was performed on the vectors of the PQS distributions around the TSSs. The bacterial names were colored coded after classification.

Figure 6

Representative PQSs selected for biophysical characterization to establish G4 folding potential. (A) The sequences selected for study by (B) ¹H-NMR, (C) CD spectroscopy, and (D) thermal melting analysis (i.e., T_m).