We investigated the toxicity of DNP and glyphosate on the growth of the BY4741 wild type strain of S. cerevisiae (Fig. 1, statistical analysis in Fig. S1). We used 3 concentrations of glyphosate (150; 300 and 600 mg/L) in the range of the sprayed commercial herbicides and 5 concentrations of DNP (0.5; 1; 2; 2.5 and 5 mM) corresponding to the range of effective decoupling activity in yeast.
Growth curves in the presence of DNP revealed a clear dose-dependent toxicity of this pollutant from 2 to 5 mM which led to total inhibition of growth culture (Fig. 1A1). We confirmed these results studying the viability of yeast cells after 3 and 6 h of culture. No effect was observed below 1 mM DNP whereas a dramatic decrease in living cells was observed for the higher DNP concentrations (Fig. 1A3 and Fig. S1). The absence of effects at 0.5 and 1 mM was surprising given the strong decoupling effect of DNP observed on yeast mitochondrion at these concentrations. Contrarily to readily accessible membrane of purified mitochondria, yeast cells could display a low permeability to this weak acidic molecule (pKa 4.08). DNP would cross the plasma membrane much more easily in its liposoluble acid form as described for other acid phenols as 2-methyl-4-chlorophenoxyacetic acid (MCPA) and 2,4-dichlorophenoxyacetic acid (2,4-D). We chose to use the non-toxic concentrations 0.5 and 1 mM DNP to further analyze the synergistic effect of DNP and glyphosate.
None of the glyphosate concentrations tested in this study exhibited a toxic effect on yeast growth, whether following the population growth or counting the living cells (Fig. 1A1 and A3). This is in agreement with the previous observations of Braconi et al. showing an inhibitory effect of the commercial formulation Silglif™ on yeast growth and metabolism but not of its active compound glyphosate. These results led us to test the 3 concentrations of glyphosate (ie. 150, 300 and 600 mg/L) in interaction with the two non-toxic concentrations of DNP.
In the presence of 0.5 mM DNP, 600 mg/L glyphosate drastically slowed the growth of the yeast population (Fig. 2B1). The analysis of cell viability confirmed this effect. A significant decrease in living cells was observed after 3 h (Fig. 2B3) and 6 h (Fig. 2B4) treatment compared to the control treated only with DNP. After 6 h in presence of 0.5 mM DNP, the yeast growth with 600 mg/L glyphosate is also significantly lower compared to 150 and 300 mg/L (Fig. 2B3 and B4). This synergy of DNP and glyphosate observed with the highest concentration of the latter is even much more pronounced when the effect of glyphosate is tested in the presence of 1 mM DNP (Fig. 2B2–B4). In these conditions, all 3 concentrations of glyphosate used showed a toxic effect on yeast growth (Fig. 2B2). Cell viability decreased after 3 h of treatment with increasing concentration of glyphosate (except at 3 h when both 300 mg/L and 600 mg/L glyphosate gave a comparable result) (Fig. 2B3 and B4). We quantified this cocktail effect calculating the generation time from the growth curves by comparing the detrimental effect of added glyphosate compared to conditions with DNP alone (Fig. 3). We observed a generation rate lowering by 1.5 times with added 150 mg/L glyphosate compared to the condition with 1 mM DNP alone. The generation rate lowered 6 times when the glyphosate concentration increases from 150 mg/L to 300 mg/L (Fig. 3). 100% growth inhibition was observed when applying 600 mg/L glyphosate in the presence of 1 mM DNP. Such a huge cocktail effect was previously observed on mammalian CHO-K1 cells treated with a combination of glyphosate, the herbicide atrazine, and their main breakdown products. To our knowledge, the toxic synergy of glyphosate with another single pollutant has not been described before.
We tested if the toxic synergetic effect was a consequence of increased oxidative stress. Glyphosate was found to induce oxidative stress in Arabidopsis thaliana plants and in yeast. A mild mitochondrial decoupling activity observed with low concentrations of DNP (1 nM) is supposed to decrease the mitochondrion-generated oxidative stress and to increase yeast life span but high concentrations could have the opposite effect as suggested with millimolar range concentrations on Hordeum vulgare plants. Last but not least, the cocktail effect of glyphosate, atrazine and their breakdown products on CHO-K1 cells could result from ROS production.
As hydrogen peroxide is described as a major redox molecule playing a central role in the interconversions and detoxification of the different ROS species during oxidative stress, we measured the H2O2 concentrations in the culture medium of yeast treated with glyphosate in the presence or in absence of 1 mM DNP (Fig. 4). No trend in variations can be evidenced with increasing glyphosate concentration, suggesting that H2O2 oxidative stress is not associated with the synergistic effect of glyphosate and DNP. However, we cannot totally exclude that other forms of oxidative stress specifically mediated by superoxide ion or hydroxyl-radical can be involved in this toxic synergy.
To confirm this observation, we also tested the synergistic effect of glyphosate and DNP on the double mutant cta1∆/ctt1∆ defective for the H2O2 detoxification catalase activity. Both catalases Cta1p and Ctt1p are important for resistance to H2O2 stress and pharmacological inhibition of catalase activity enhances the mitochondrial oxidative stress induced by Ca2+. Yeast cta1∆/ctt1∆ double mutant display increased sensitivity to H2O2 and oxidative stress-generating conditions (, our observations with the COREPS students).
In order to compare the WT strain to the double mutant cta1∆-ctt1∆, we calculated the generation time (G) for the cta1∆-ctt1∆ mutant. In normal growth conditions, we showed a weak increase of G (1.4 fold, see legend of Fig. 3) in the double mutant compared to the WT. Similarly, in the presence of 1 mM DNP, G is also 1.4 times higher, (see Fig. 3) in the cta1∆/ctt1∆ genotype compared to the wild type, whatever the glyphosate concentration. These results indicate that the cta1∆/ctt1∆ mutant is not more sensitive than the wild type to the glyphosate-DNP cocktail effect. Oxidative stress might not be involved in this observed glyphosate-DNP cocktail effect and that the mechanistic process of this toxic synergy remains to be elucidated.