HIF-1 Has a Central Role in Caenorhabditis elegans Organismal Response to Selenium

Abstract

Selenium is a trace element for most organisms; its deficiency and excess are detrimental. Selenium beneficial effects are mainly due to the role of the 21^st genetically encoded amino acid selenocysteine (Sec). Selenium also exerts Sec-independent beneficial effects. Its harmful effects are thought to be mainly due to non-specific incorporation in protein synthesis. Yet the selenium response in animals is poorly understood. In Caenorhabditis elegans, Sec is genetically incorporated into a single selenoprotein. Similar to mammals, a 20-fold excess of the optimal selenium requirement is harmful. Sodium selenite (Na₂SeO₃) excess causes development retardation, impaired growth, and neurodegeneration of motor neurons. To study the organismal response to selenium we performed a genetic screen for C. elegans mutants that are resistant to selenite. We isolated non-sense and missense egl-9/EGLN mutants that confer robust resistance to selenium. In contrast, hif-1/HIF null mutant was highly sensitive to selenium, establishing a role for this transcription factor in the selenium response. We showed that EGL-9 regulates HIF-1 activity through VHL-1, and identified CYSL-1 as a key sensor that transduces the selenium signal. Finally, we showed that the key enzymes involved in sulfide and sulfite stress (sulfide quinone oxidoreductase and sulfite oxidase) are not required for selenium resistance. In contrast, knockout strains in the persulfide dioxygenase ETHE-1 and the sulfurtransferase MPST-7 affect the organismal response to selenium. In sum, our results identified a transcriptional pathway as well as enzymes possibly involved in the organismal selenium response.

Keywords: CYSL-1; Caenorhabditis elegans; EGL-9; HIF-1; selenite; selenium; stress; sulfide.

Figure 1

EGL-9 protein domains and C. elegans egl-9 transcripts representation. (A) Scheme of the primary structure of EGL-9 protein (isoform a), highlighting the regions that constitute the hydroxylase and MYND domains. aa: amino acids. (B) Different egl-9 transcripts reported (F22E12.4a-e). The coding region is represented in orange, UTR sequences in gray and introns as lines. The coding regions for the hydroxylase domain and the MYND domain are indicated in green and yellow, respectively. The position and identity of egl-9 mutant alleles isolated in this study (allele zf150 and zf151), as well as the location of the previously reported egl-9(sa307) allele (243 bp deletion) are indicated.

Figure 2

egl-9 mutant strains are resistant to toxic selenite concentration. Locomotor activity refers to the motility of a population of worms, relative to the basal activity measured before the addition of the compound of interest, as detailed in methods. (A) Locomotor activity of egl-9(zf150) mutant worms and wild-type (WT) in 0, 10, and 20 mM of sodium selenite (Na₂SeO₃) for 16 h. Points indicate the average of locomotor activities measured every 15 min. AU: arbitrary units. (B–D) Relative locomotor activity (Se/vehicle) of egl-9(zf150), egl-9(zf151), egl-9(sa307), and WT worms at the endpoint of incubation (16 h). Columns indicate the average locomotor activity of Na₂SeO₃-treated worms relative to the activity of the control without Na₂SeO₃ (0 mM) for each strain. Error bars (only + shown) indicate standard deviation. Variance analysis test was performed [one-way ANOVA, p = 1.77E-9 (B) and p = 5.86E-7 (D), and Welch F test, p = 5.92E-5 (C)] followed by Tukey test. Different lowercase letters denote significant differences obtained by Tukey test (the statistical analysis and p values obtained are shown in Supplementary Data ). Each graph corresponds to a representative experiment with four wells per condition per strain (80 worms per well). Three biological replicates were performed.

Figure 3

CYSL-1 functions as a selenium sensor upstream of EGL-9 leading to increased HIF-1 activity. Locomotor activity refers to the motility of a population of worms, relative to the basal activity measured before the addition of the compound of interest, as detailed in methods. Error bars (only + shown) indicate standard deviation. Different lowercase letters denote significant differences obtained by Tukey test (the statistical analysis and p values obtained are shown in Supplementary Data ). (A) HIF-1 activation mechanism involving CYSL-1. CYSL-1 negatively regulates EGL-9 by protein-protein interaction and promotes the HIF-1 activity (Ma et al., 2012). (B) Locomotor activity of hif-1(ia04) and WT animals in 0, 5, and 10 mM of Na₂SeO₃ for 16 h. Points indicate the average of locomotor activities measured every 15 min. AU: arbitrary units. (C) Locomotor activity of hif-1(ia04) Na₂SeO₃-treated worms relative to the activity of the control without Na₂SeO₃ (0 mM) at the endpoint of incubation (16 h). Variance analysis test was performed (Welch F test p = 1.66E-5) and subsequent Tukey test. The graph corresponds to a representative experiment with four wells per condition per strain (80 worms per well). Three biological replicates were performed. (D) Survival of WT, hif-1(ia04), and hif-1(ia04); Exhif-1::gfp strains in 0 and 5 mM of Na₂SeO₃. Columns indicate the percentage of live adult worms after 20 h of incubation. The graph corresponds to three independent experiments with one plate per strain (30–40 worms per plate). (E) Locomotor activity of vhl-1(ok161), swan-1(ok267), and WT strains in 0, 5, and 10 mM Na₂SeO₃ relative to the activity in 0 mM after 16 h of incubation. Variance analysis test was performed (One-way ANOVA, p = 2.39E-14) and subsequent Tukey test. (F) Survival of the WT, vhl-1(ok161), and vhl-1(ok161); Exvhl-1::gfp strains in 0, 10, and 20 mM of Na₂SeO₃ after 48 h of incubation. Columns indicate the percentage of live adult worms. A Kruskal-Wallis test was performed (p = 2.2E-7) followed by Mann-Whitney pairwise comparisons. The graph corresponds to three independent experiments with two plates per strain (20 worms per plate). (G, I) Locomotor activity of cysl-1(ok764) (G) and cysl-1(ok764); egl-9(sa307) (I) mutant strains in 0, 5, and 10 mM Na₂SeO₃ relative to the activity in 0 mM after 16 h. Variance analysis test was performed [Welch F test, p = 8.53E-9 (G) and one-way ANOVA, p = 3.75E-5 (I)], followed by Tukey test. Each graph corresponds to a representative experiment with four wells per condition per strain (80 worms per well). Three biological replicates were performed. (H) Survival of WT and cysl-1(ok762) in 0 and 5 mM of Na₂SeO₃. Columns indicate the average of live adult worms after 20 h of incubation. The graph corresponds to three independent experiments with one plate per strain (30–40 worms per plate).

Figure 4

Persulfide dioxygenase (ETHE-1) and sulfurtransferase (MPST-7) are involved in selenite metabolism. Locomotor activity refers to the motility of a population of worms, relative to the basal activity measured before the addition of the compound of interest, as detailed in methods. Error bars (only + shown) indicate standard deviation, unless otherwise specified. Different lowercase letters denote significant differences obtained by post hoc test (the statistical analysis and p values obtained are shown in Supplementary Data ). (A) Locomotor activity of sqrd-1(tm3378) and WT in 0, 5, and 10 mM Na₂SeO₃ relative to the activity in the control (0 mM) after 16 h of incubation. Variance analysis test was performed (Welch F test, p = 7.33E-14) and subsequent Tukey test. The graph corresponds to the mean of four experiments and error bars (only + shown) indicate standard error of the mean. Each experiment includes four wells per condition per strain (80 worms per well). (B) Locomotor activity of sqrd-1(tm3378) and WT in 0, 2, 10 mM Na₂S relative to the activity in the control (0 mM) after 16 h of incubation. Variance analysis test was performed (Kruskal-Wallis, p = 1.08E-6) and subsequent Mann-Whitney test. The graph corresponds to one representative experiment. Each experiment includes four wells per condition per strain (80 worms per well). (C) H₂S oxidation mechanism. SQRD-1 catalyzes the H₂S oxidation. The sulfur, as sulfone, is transferred to an acceptor molecule. Glutathione (GSH) and sulfite (SO₃ ^2-) have been proposed as alternative acceptor molecules. The persulfide dioxygenase (ETHE-1) catalyzes the synthesis of sulfite using glutathione persulfide (GSSH) as precursor and the preferential reaction catalyzed by the sulfur transferase (MPST-7) is the formation of thiosulfate (SSO₃²⁻) using sulfite as a precursor (wider line). The sulfite oxidase (SUOX-1) catalyzes the formation of sulfate (SO₄ ²⁻) using sulfite (Filipovic et al., 2018). (D) Survival of worms of the WT strain with the suox-1 gene expression interfered (suox-1) and the negative control (empty). Columns indicate the percentage of live adult worms after 20 h in 2 mM of Na₂SeO₃ and 4 mM of sodium sulfite (SO₄²⁻). The graph corresponds to two experiments with one plate each (30–40 worms per plate). (E, F) Locomotor activity of mpst-7(gk14674) (E) and ethe-1(tm4101) (F) in 0, 5, and 10 mM Na₂SeO₃ relative to the activity in the control 0 mM after 16 h. Analysis of variance test was performed [Welch F test, p = 1.06E-5 (E) and p = 4.06E-6 (F)] and subsequent Tukey test. Three biological replicates were performed with similar results.

Figure 5

Selenium-triggered transcriptional response mechanism model and possible compounds involved in MPST-7 catalyzed reaction. (A) Transcriptional response to selenium involving DAF-16/FOXO, SKN-1/NRF2, and HIF-1/HIF has been identified. The HIF-1 pathway involves CYSL-1, which detects selenium and inhibits EGL-9 (this study). HIF-1 activation would result in a gene expression change responsible for the selenium organismal response. The same pathway was previously proposed for sulfur in reference (Budde and Roth, 2011). (B) MPST-7, catalyze the conversion of sulfite (SO₃ ^2-) and glutathione persulfide (GSSH) to thiosulfate (SSO₃ ^2-) and glutathione (GSH) (black) (Filipovic et al., 2018). Possible MPST-7 selenium substrates and products are shown in gray.