The glutaredoxin GLRX-21 functions to prevent selenium-induced oxidative stress in Caenorhabditis elegans

Abstract

Selenium is an essential micronutrient that functions as an antioxidant. Yet, at higher concentrations, selenium is pro-oxidant and toxic. In extreme cases, exposures to excess selenium can lead to death or selenosis, a syndrome characterized by teeth, hair and nail loss, and nervous system alterations. Recent interest in selenium as an anti- tumorigenic agent has reemphasized the need to understand the mechanisms underlying the cellular consequences of increased selenium exposure. We show here, that in the nematode, Caenorhabditis elegans, selenium has a concentration range in which it functions as an antioxidant, but beyond this range it exhibits a dose- and time-dependent lethality. Oxidation-induced fluorescence emitted by the dye, carboxy-H(2)DCFDA, indicative of reactive oxygen species formation was significantly higher in animals after a brief exposure to 5mM sodium selenite. Longer-term exposures lead to a progressive selenium-induced motility impairment that could be partially prevented by coincident exposure to the cellular antioxidant-reduced glutathione. The C elegans glrx-21 gene belongs to the family of glutaredoxins (glutathione-dependent oxidoreductases) and the glrx-21(tm2921) allele is a null mutation that renders animals hypersensitive for the selenium-induced motility impairment, but not lethality. In addition, the lethality of animals with the tm2921 mutation exposed to selenium was unaffected by the addition of reduced glutathione, suggesting that GLRX-21 is required for glutathione to moderate this selenium-induced lethality. Our findings provide the first description of selenium-induced toxicity in C elegans and support its use as a model for elucidating the mechanisms of selenium toxicity.

FIG. 1.

Selenium is both protective and toxic to Caenorhabditis elegans. The effects on the survival of wild-type animals exposed to increasing Na₂SeO₃ concentrations (0–10mM) either alone (•) or in the presence of 1mM (Δ) or 2mM (▴) H₂O₂. Each dataset represents the averages of six to nine plates with 10 animals per plate exposed in liquid for 12 h and is presented as the mean percentage of dead animals ± SD. The LC₅₀ for Na₂SeO₃ exposure at 12 h was calculated to be 3.47mM. #p < 0.05, compared with 2mM H₂O₂ and 1mM Na₂SeO₃; ##p < 0.05, compared with 1mM H₂O₂ (no significant difference to 2mM Na₂SeO₃); * p < 0.01, compared with both 1mM H₂O₂ and either 0.001 or 0.01mM Na₂SeO₃; **p < 0.001, compared with 1mM H₂O₂ (no significant difference to 1mM Na₂SeO₃).

FIG. 2.

Selenium toxicity is induced by free radical formation. (A) The average luminosities of adult N2 wild-type animals exposed on agar plates for 2, 4, or 6 h to either 5mM Na₂SeO₃ (•) or 5mM Na₂SeO₃ and 10mM GSH (▴) followed by a 3-h incubation with the dye carboxy-H₂DCFDA were compared with control (□) animals mock exposed on plates for 0, 2, 4, or 6 h and then incubated for 3 h with dye. Animals that were dead or dying were eliminated before fluorescence measurements were taken as described in the “Materials and Methods” section. Each data point displayed on the graph represents the average luminosity from 10–13 animals obtained by measuring the five brightest 100 pixel dense regions of the gut located immediately posterior to the head and is represented by arbitrary units ± SD. **p < 0.001, compared with the 6-h time point for both the control and Na₂SeO₃ + GSH. (B–E) Representative photomicrographs of control (mock exposed) animals at 0 h (B) and 6 h (C) are shown for comparison to animals exposed for 6 h to either 5mM Na₂SeO₃ [(D)—box indicates region magnified in (F)] or 5mM Na₂SeO₃ and 10mM GSH (E). (F) A region of the 5mM Na₂SeO₃ exposed animal shown in (D) was magnified to 600%. The open box is representative of a 100 pixel dense regions in which the luminosity was measured to generate the graph (A).

FIG. 3.

The Caenorhabditis elegans glutaredoxin, GLRX-21 is most similar to human GLRX2. (A) Sequence alignment of all human and C elegans dithiol glutaredoxins. Compared with human GLRX2, GLRX-21 is 39% identical, whereas GLRX-22 has only 26.3% identity. The sequence used for human GLRX2 corresponds to the common domain for both the mitochondrial and nuclear isoforms (Lundberg et al., 2001). The cysteine residues at the redox active site are highlighted in red and those essential for GSH binding are in blue (Lillig et al. 2008). Other conserved residues in all the glutaredoxins are highlighted in black. The numbers on the right indicate the number of amino acid residues of each protein. (B) Phylogenetic tree of all human and C elegans dithiol glutaredoxins. The percentage bootstrap values (based on 1000 replications) are given on the nodes of the tree. Caenorhabditis elegans GLRX-21 is highlighted in red. (C) The glrx-21 gene is organized into three exons. Open boxes indicate the ORF, whereas the gray boxes designate the 5′-untranslated region (UTR) and 3′-UTR, respectively. The tm2921 deletion removes part of the proximal promoter plus the first exon and the first intron of the glrx-21 gene. Primers for RT-PCR were designed at the ATG codon and the beginning of the second exon (F1 and F2, respectively) and at the STOP codon (R1), respectively. The glrx-21 cDNA is only detected in the N2 wild-type lanes demonstrating that tm2921 is a null allele. act-1 primers were used for cDNA synthesis control.

FIG. 4.

Selenium-induced lethality of a glutaredoxin mutant is not decreased by GSH. Animals were scored for lethality after 12-h exposure to (A) 5mM Na₂SeO₃ or (B) both 5mM Na₂SeO₃ and increasing concentrations of GSH (0–10mM). (A) Comparison of the selenium-induced lethality of the wild-type strain N2 (▪), and the glutaredoxin mutant strain, glrx-21(tm2921) (x). Each data point represents the averages of four to six plates with 10 animals per plate and is presented as the mean percentage of dead animals ± SD. ** p < 0.001, for wild-type and the glrx-21 mutant strain exposed to 5mM Na₂SeO₃ and compared across all other concentrations of Na₂SeO₃ (one-way ANOVA). Note. There is no statistically significant difference in lethality between wild-type and the glrx-21 mutant strain when compared within each concentration of Na₂SeO₃ (Student’s t-test) and across all concentrations except 5mM Na₂SeO₃ (one-way ANOVA). (B) The effects of reduced GSH on the selenium-induced lethality of the wild-type strain N2 (▪), and the glutaredoxin mutant strain, glrx-21(tm2921) (x). Each data point represents the averages of 4–13 plates with 10 animals per plate and is presented as the mean percentage of dead animals ± SD. Note. There is no statistically significant difference in lethality for the glrx-21 mutant strain when compared across all concentrations of GSH (0-10mM) tested (one-way ANOVA). *p < 0.05, compared with all concentrations of GSH (0.01–10mM) exposed wild-type animals and glrx-21 mutant animals, as well as the control (0mM GSH) for the glrx-21 mutant strain; **p < 0.01, compared across all concentrations of GSH (0.01–10mM) and controls for both the wild-type and glrx-21 mutant strains.

FIG. 5.

Sodium selenite reduces motile behavior. (A) Populations of adult animals were exposed to increasing concentrations (0–20mM) of sodium selenite (Na₂SeO₃) on agar plates. Motile behavior or the ability to move in any direction was scored in individual animals at 24-h intervals over the course of four days (96 h). All datasets represent the averages of three plates with 50 worms per plate and are presented as the mean percentage motility ± SD. *p < 0.05 compared with the 24-h time point at the same concentration and with the matched time point at 0mM Na₂SeO₃. ** p < 0.001, compared with the 24-h time point at the same concentration and with the matched time point at 0mM Na₂SeO₃ (B) Adult animals exposed on agar plates to 5mM Na₂SeO₃ with 0.7, 1.3, or 2.6mM reduced GSH were scored at 24-h intervals and compared with their selenium-only controls (0mM GSH) for their motility behavior. All datasets represent the averages of three to five plates with 20 animals per plate and are presented as the mean percentage motility ± SD. *p < 0.05, compared with 0mM GSH at the same time point. **p < 0.001, compared with 0mM GSH at the same time point. (C) Populations of adult animals of the strains N2 (wild type) and VZ54 glrx-21(tm2921) were exposed to 5mM Na₂SeO₃ and tested for their motile ability after 24 and 48 h of exposure on agar plates. All datasets represent the averages of three to four plates with 20 animals per plate and are presented as the mean percentage motility ± SD. *p < 0.05, compared with wild type at 24 h. **p < 0.001, compared with the glrx-21 mutant strain at 24 h and wild type at 24 and 48 h.

FIG. 6.

Chronic exposure to selenium results in a progression of behavioral phenotypes leading to lethality. (A) Populations of adult animals placed continuously on NGM agar plates supplemented with 5mM Na₂SeO₃ were scored at 24-h intervals to determine the percentages of animals with each behavioral phenotype (normal, backing [impaired], paralyzed) and for lethality (dead). The percentages of all phenotypes for each time point (0, 24, 48, 72, and 96 h) equal 100%. Each dataset represents six plates with 20 animals per plate and is presented as the mean percentage of animals with each phenotype ± SD. * p < 0.05, compared with normal at the same time point. **p < 0.001, compared with normal at the same time point. # p < 0.05, compared with 24-h within phenotype. ## p < 0.001, compared with 24-h within phenotype. (B) Individual animals were grown singularly on 5mM Na₂SeO₃ plates for a total of 48 h. The behavioral phenotype of each individual animal was assessed every 6 h. A total of 20 individual animals were observed. N = normal (light gray), B = backing (white), P = paralyzed (dark gray), D = dead (black).

FIG. 7.

Behavioral deficits induced by acute selenium exposure are partially reversible. The behavioral phenotype of each adult animal was observed after growth for 24 h (normal, backing, paralyzed; white letters) on NGM plates supplemented with 5mM Na₂SeO₃. These animals were sorted by these initial behavioral phenotypes onto fresh NGM agar plates with no added selenium. After 24 h without selenium exposure, individual animals were again scored for their behavioral phenotype and for lethality (normal, backing, paralyzed, dead; black letters). Twenty animals per phenotype were placed on to each of three selenium-free plates for a total of 60 animals for each 24-h phenotype used. The phenotypes of animals grown for 48 h without selenium served as a control (untreated). Each graph bar represents the average population of animals with the 48-h phenotype (as indicated on the graph) and is depicted as the mean percentage ± SD.