G-quadruplex prediction in E. coli genome reveals a conserved putative G-quadruplex-Hairpin-Duplex switch

Abstract

Many studies show that short non-coding sequences are widely conserved among regulatory elements. More and more conserved sequences are being discovered since the development of next generation sequencing technology. A common approach to identify conserved sequences with regulatory roles relies on topological changes such as hairpin formation at the DNA or RNA level. G-quadruplexes, non-canonical nucleic acid topologies with little established biological roles, are increasingly considered for conserved regulatory element discovery. Since the tertiary structure of G-quadruplexes is strongly dependent on the loop sequence which is disregarded by the generally accepted algorithm, we hypothesized that G-quadruplexes with similar topology and, indirectly, similar interaction patterns, can be determined using phylogenetic clustering based on differences in the loop sequences. Phylogenetic analysis of 52 G-quadruplex forming sequences in the Escherichia coli genome revealed two conserved G-quadruplex motifs with a potential regulatory role. Further analysis revealed that both motifs tend to form hairpins and G quadruplexes, as supported by circular dichroism studies. The phylogenetic analysis as described in this work can greatly improve the discovery of functional G-quadruplex structures and may explain unknown regulatory patterns.

Figure 1.

Phylogenetic tree of Escherichia coli HPGQs. The phylogenetic tree was constructed using ClustalW tool implementing the UPGMA method. The common predictive GQ motifs called HPG1 and HPG2 are exclusively clustered together. The first column represents the position of the HPGQs on the genome while the second column refers to genes located on the strand. The third column displays their phylogenetic distances to the nearest node as provided by ClustalW web tool and the degree of divergence from the closest HPGQ group is shown in the last column.

Figure 2.

Nucleotide alignment of HPG1 (bottom) and HPG2 (top) ‘gapped’ motifs as obtained from GLAM2. The motifs discovered using GLAM2 were aligned and GQ-forming G-rich regions are framed within the box.

Figure 3.

The conservation of the common motif among bacteria. The phylogenetic tree based on ncbi taxonomy database (left) and the negative logarithm of P-value (solid bars) and q-values (striped bars) of the best FIMO analysis matches using the common motif in Table 2 within given prokaryotic upstream sequences (right) are presented.

Figure 4.

(A) Circular dichroism (CD) and (B) melting analysis of HPG2 structures in the presence of various ions. The oligonucleotides were dissolved in a 2.5 mM Hepes buffer at pH 7.2 in the presence of either 100 mM KCl (dashed grey line), 100 mM LiCl (dashed thin black line), 10 mM MgSO₄ (thick black line), 100 mM NaCl (thick grey line) or 50 mM KCl and 50 mM LiCl (dashed thick black line). Samples were heated up to 90°C for 30 min and allowed to cool down to room temperature for 16 h. The melting curves of the HPG2 motif sequence as presented in Table 2 in the presence of K⁺ and Li⁺ according to the change at 260 nm using CD spectrophotometer indicates similar melting temperatures for both structures (73°C and 71°C, respectively).

Figure 5.

(A) Circular dichroism and (B) gel retardation analysis of HPG2 structures in the absence (solid lines) and presence (dashed lines) of complementary sequence in the presence of K⁺ (grey lines) or Mg⁺⁺ (black lines). The oligonucleotides were dissolved in a 2.5 mM Hepes buffer at pH 7.2 in the presence of either 100 mM KCl or 10 mM MgSO₄. Duplex formation was achieved by incubation equal amounts of GQ-forming strand and its complementary. Samples were heated up to 90°C for 30 min and allowed to cool down to room temperature for 16 h before measurement. The oligonucleotide concentrations were set to 1 μM for CD experiment and 5 μM for gel retardation.