G-quadruplexes as novel cis-elements controlling transcription during embryonic development

Abstract

G-quadruplexes are dynamic structures folded in G-rich single-stranded DNA regions. These structures have been recognized as a potential nucleic acid based mechanism for regulating multiple cellular processes such as replication, transcription and genomic maintenance. So far, their transcriptional role in vivo during vertebrate embryonic development has not yet been addressed. Here, we performed an in silico search to find conserved putative G-quadruplex sequences (PQSs) within proximal promoter regions of human, mouse and zebrafish developmental genes. Among the PQSs able to fold in vitro as G-quadruplex, those present in nog3, col2a1 and fzd5 promoters were selected for further studies. In cellulo studies revealed that the selected G-quadruplexes affected the transcription of luciferase controlled by the SV40 nonrelated promoter. G-quadruplex disruption in vivo by microinjection in zebrafish embryos of either small ligands or DNA oligonucleotides complementary to the selected PQSs resulted in lower transcription of the targeted genes. Moreover, zebrafish embryos and larvae phenotypes caused by the presence of complementary oligonucleotides fully resembled those ones reported for nog3, col2a1 and fzd5 morphants. To our knowledge, this is the first work revealing in vivo the role of conserved G-quadruplexes in the embryonic development, one of the most regulated processes of the vertebrates biology.

Figure 1.

Selection strategy of conserved PQSs in silico. (A) Schematic representation of conserved PQS selection strategy. (B) Chart representing the number of genes retrieved in each step of the bioinformatic search for the analyzed species. GC contents (%) were obtained from Kai et al. (31). (C) The three lists of genes containing PQSs in their PPRs obtained for the three species were intersected using BioVenn software.

Figure 2.

In vitro analysis of G-quadruplex formation by ThT assays and circular dicroism. (A) Bar graph of fluorescence enhancement (F/F₀) of ThT in the presence of the different oligonucleotides representing human and zebrafish selected PQSs. Each bar represents the mean of three technical repeats and error bars correspond to standard deviation (SD). (B) CD spectra of oligonucleotides representing human PQSs of the seven selected genes positive for ThT assay (apba1, col2a1, fzd5, map3k1, nog3, nr2e1 and shox2) and two control genes negative for ThT assay (jun and gnaq). (C) CD spectrum of oligonucleotides representing zebrafish PQSs of the seven selected genes positive for ThT assay (apba1, col2a1, fzd5, map3k1, nog3, nr2e1 and shox2) and two control genes negative for ThT assay (jun and gnaq).

Figure 3.

Col2a1, fzd5 and nog3 PQSs and their role in transcriptional expression control in cellulo. (A) Chart presenting the sequences, strand locations and relative positions for the PQSs present in the PPRs of the three selected human and zebrafish genes. G-tracts are bold and underlined. Arrowheads point the G to A replacements in mutated PQSs. *Relative positions are considering TSS as +1. (B) Luciferase assay performed in Neuro-2a cells transfected with pGL3-promoter vector plasmid (no PQS) or pGL3-promoter vector plasmid containing the wild type (wild-type PQS) or mutated (mutated PQS) sequence of human or zebrafish PQSs of the three selected genes upstream the basal promoter SV40. Each bar represents the luciferase activity normalized to β-galactosidase activity and relativized to that for the unmodified pGL3-promoter vector plasmid. Bars represent the mean of three independent experiments and error bars correspond to standard deviation (SD). **P < 0.01, ***P < 0.001, t-Student test.

Figure 4.

Disruption of G-quadruplexes by G-quadruplex ligands and ASO injection affects transcriptional expression in zebrafish embryos. (A) Relative abundance of transcripts of col2a1, fzd5, nog3 and the control actb2 genes measured by RT-qPCR in 48-hpf embryos injected with the drugs TMPyP2 and TMPyP4. (B) Relative abundance of transcripts of col2a1, fzd5, nog3 and the control actb2 genes measured by RT-qPCR in 90% epiboly, 30-hpf and 48-hpf staged embryos, respectively, injected with the ASO complementary to the corresponding PQS. In the case of actb2 gene, two ASOs were used, one complementary to the template (actb2(+)-ASO) and the other to the coding strand (actb2(-)-ASO). In all cases, three biological and technical repeats were performed for each condition, resulting in similar trends. Bars represent the mean of the three technical repeats for one representative biological repeat. Error bars correspond to standard deviation (SD) of the three technical repeats. *P < 0,05, **P < 0.01, ***P < 0.001, t-Student test.

Figure 5.

Effect of col2a1 G-quadruplex disruption by col2a1-ASO injection in zebrafish embryos. (A) Numbers and percentages of embryos showing dead, deformed, normal or ventral curvature phenotypes are shown in a table and represented in a stacked bar graph for 24-hpf staged embryos injected with CTRL or col2a1-ASO. ***P < 0.001, chi-square test. (B) Representative picture of 4-dpf staged larvae injected with CTRL or col2a1-ASO and stained with Alcian Blue to determine the body length. Lateral views, anterior to the left. (C) Box-plot of the relative body length of 4-dpf staged larvae injected with CTRL or col2a1-ASO. *P < 0.05, t-Student test. (D) WISH assessing the expression of col2a1 mRNA in 90% epiboly (i and ii) and 10-somite (iii and iv) staged embryos injected with CTRL (i and iii) or col2a1-ASO (ii and iv). Arrowheads point regions of lower expression. In lateral views anterior is to the left, and in dorsal and fronto-dorsal views anterior is up. Numbers and percentages of embryos/larvae with the shown phenotype are indicated in each panel. n: notochord. Scale bars (200 μm) are represented in B, Di and Diii.

Figure 6.

Effect of fzd5 G-quadruplex disruption by fzd5-ASO injection in zebrafish embryos. (A) Numbers and percentages of embryos showing dead, deformed, normal or reduced-eye phenotypes are shown in a table and represented in a stacked bar graph for 30-hpf staged embryos injected with CTRL or fzd5-ASO. ***P < 0.001, chi-square test. (B) Representative picture of 30-hpf staged embryos injected with CTRL or fzd5-ASO used to determine eye diameter. Lateral views, anterior to the left. (C) Box-plot of the relative eye diameter of 30-hpf staged embryos injected with CTRL or fzd5-ASO. ***P < 0.001, t-Student test. (D) WISH assessing the expression of fzd5 mRNA in 2-somite (i and ii), 10-somite (iii and iv), 15-somite (v and vi) and 30-hpf (vii and viii) staged embryos injected with CTRL (i, iii, v and vii) or fzd5-ASO (ii, iv, vi and viii). Arrowheads point regions of lower expression. In lateral views anterior is to the left, and in fronto-dorsal views anterior is up. Numbers and percentages of embryos/larvae with the shown phenotype are indicated in each panel. e: eye; vdc: ventral diencephalon. Scale bars (200 μm) are represented in B, Di, Diii, Dv and Dvii.

Figure 7.

Effect of nog3 G-quadruplex disruption by nog3-ASO injection in zebrafish embryos. (A) Numbers and percentages of embryos showing dead, deformed, normal or small head phenotypes are shown in a table and represented in a stacked bar graph for 48-hpf staged embryos injected with CTRL or nog3-ASO. ***P < 0.001, chi-square test. (B) Representative picture of Alcian Blue staining of craniofacial cartilages in 4-dpf staged larvae injected with CTRL or nog3-ASO. Ventral views, anterior to the left. (C) Box-plots of the relative ceratohyal cartilages angle and ceratobranchial cartilages number of 4-dpf staged larvae injected with CTRL or nog3-ASO. *P < 0.05, t-Student test. (D) WISH assessing the expression of nog3 mRNA in 56-hpf staged larvae injected with CTRL (i) or nog3-ASO (ii). Arrowheads point regions of lower expression. In lateral and dorsal views anterior is to the left. Numbers and percentages of embryos/larvae with the shown phenotype are indicated in each panel. ca: ceratohyal cartilages angle; cb (3-7): ceratobranchial cartilages 3 to 7; ch: ceratohyal cartilage; pa: pharyngeal arches; pf: pectoral fin; tc: trabeculae cranii. Scale bars (200 μm) are represented in B and Di.