Resistance to stresses
To test the sensitivity of the double mutant to different stressors, WT, Δchap1, Δskn7, and Δchap1-Δskn7 (ΔΔ) were grown on solid CMX containing either 0.75 M sorbitol – hyperosmotic stress, 0.4 M KCl – hyperosmotic and salt (ionic) stress, 20 mM H2O2 – oxidative stress, 30 μM menadione – superoxide stress or 25 mM CWS – interference with cell wall integrity (Ram & Klis, 2006) (Fig. 1). Growth of Δchap1, Δskn7, and ΔΔ in the presence of 20 mM H2O2 was completely inhibited compared with WT which showed about 40% growth relative to control (solid CMX without additives). On 0.4 M KCl, growth of the ΔΔ mutant was also inhibited compared with WT, but not completely, and it showed similar growth rate to the Δskn7 mutant. On 0.75 M sorbitol, the double mutant showed almost complete inhibition, but again similar to the Δskn7 mutant; WT and Δchap1 were also inhibited, Δchap1 more than WT, but both less than Δskn7 and the double mutant. CWS affected the growth of Δchap1 (55%) and the double mutant (47%), whereas growth of the WT and Δskn7 was less inhibited (about 65%). On menadione, the double mutant was inhibited more than the WT and Δskn7 but as much as Δchap1 (Fig. 1a).
Figure 1. Response of the mutants to different stressors. (a) WT, Δchap1, Δskn7 and ΔΔ were grown on solid CMX with different conditions for 6 days, and colony diameters were measured. Average colony diameter relative to control (solid CMX without additives) (mean and SEM for three replicates) is shown. (b) WT and mutants were grown on liquid CMX with different hydrogen peroxide concentrations; fungal mats were dried and weighed after 5 days. Dots represent growth rate calculated from the average of six biological repeats, relative to the average of nine control repeats of each strain (grown on liquid CMX without H2O2).
Download figure to PowerPoint
The double mutant (ΔΔ) is sensitive to oxidative and osmotic stresses, as well as to stressors that compromise cell wall integrity, but to the same extent as each single mutant, and there is no evidence for an additive effect on inhibition of growth. The only additive effect on growth rate was found with the cell wall stressor CWS, where the double mutant was more sensitive than either single mutant (significant at P < 0.01). WT and the mutants were also grown on liquid CMX with lower concentrations of hydrogen peroxide (0.625–10 mM) to test whether the double mutant is more sensitive than each single mutant (Fig. 1b). WT grew normally on all concentrations; apparently at these oxidant levels, WT can overcome the stress by expression of antioxidant genes, as shown previously (Lev et al., 2005). All three mutants show lower growth percentages than WT, but similar to each other.
Antioxidant gene expression changes in response to hydrogen peroxide
We tested the expression of antioxidant genes shown previously (Lev et al., 2005) to be under regulation by the transcription factor ChAP1: glutathione reductase, GLR1; thioredoxin, TRX2; thioredoxin reductase, TRR1; γ-glutamylcysteine synthetase, GSH1. In addition, we followed the expression of a superoxide dismutase gene, SOD1, and three catalase-encoding genes CAT1,2,3 (Robbertse et al., 2003). We performed quantitative real-time PCR experiments on WT, both single mutants and the double mutant after exposure to 20 mM H2O2 for 30 min in shake culture (see 'Materials and methods'). The H2O2-induced transcript levels of most of the genes tested depended strongly on ChAP1, and several required Skn7 for full induction. The gene for glutathione reductase (GLR1) was only twofold induced in Δchap1 compared with 52-fold in WT and 16-fold and Δskn7. In the double mutant, the transcript level was similar to the basal level in the untreated control, indicating that either transcription factor is sufficient only for partial expression, while both transcription factors are required for full expression (Fig. 2). TRX2 showed the same pattern as GLR1, but the additive effect was not statistically significant. The TRR1 gene is under the regulation of ChAP1 alone. While superoxide dismutase (SOD1) expression is not strongly decreased by loss of either ChAP1 or Skn7 alone, the double mutant failed to upregulate the expression of SOD1. The catalase genes CAT1 and CAT3 seem ChAP1 dependent and Skn7 independent; however, this regulation is not significant at P < 0.01 by the multiple-comparison t-test used here. CAT2 is expressed in all three mutants. The expression of γ-glutamylcysteine synthetase (GSH1) was also tested, and only minor upregulation was observed in WT and Δskn7.
Figure 2. Antioxidant gene expression in WT vs. mutants. Expression of several known antioxidant genes was measured in WT vs. Δchap1, Δskn7 and Δchap1-Δskn7 (ΔΔ) mutants in response to 30-min exposure to 20 mM hydrogen peroxide in shake culture by RT-qPCR. Transcript levels shown are the average of three biological repeats – fold induction relative to control (CMX with no H2O2) is shown (** and *** indicate significance at P < 0.01 and 0.001). CAT1 – catalase 1; CAT2 – catalase 2; CAT3 – catalase 3; GLR1 – glutathione reductase 1; SOD1 – superoxide dismutase 1; TRR1 – thioredoxin reductase 1; TRX2 – thioredoxin 2; GSH1 – γ-glutamylcysteine synthetase 1.
Download figure to PowerPoint
Plant infection assay
To test whether both ChAP1 and Skn7 contribute to virulence on the host, infection assays on maize were carried out. To inoculate undetached maize leaves, maize plants were grown in hydroponics (as described in the Materials and Methods section) for 12 days, the plants were removed from the medium and transferred into a tray where the roots were kept moist. Spores from Δchap1, Δskn7, Δchap1-Δskn7 (ΔΔ) and WT were prepared in ddW with 0.02% Tween 20; at least four plants were used for each mutant, and the second leaf was inoculated with three 7-μL droplets containing about 500 spores. Lesion areas were measured using imagej software from images taken 2 days after inoculation (Fig. 3a). Δchap1 and Δskn7 mutants were not significantly different in virulence from WT, whereas ΔΔ showed significantly smaller lesions (about 30% smaller, Fig. 3b). This demonstrates an additive contribution of the two transcription factors that are lacking in the double mutant. These contributions may promote the ability to counteract the plant's oxidative burst as well as other stresses the pathogen encounters during infection. Thus, the double mutant may be sensitive to the HR or other plant defenses, preventing spreading of the mutant and resulting in smaller lesions than those formed by the WT.
Figure 3. Plant infection assay. Second leaves of 12-day-old maize plants were inoculated with three 7-μL droplets containing about 500 spores, leaves were photographed (a) and lesion area (y-axis, plotted starting at 10 mm2) was measured after 2 days (b). Number of lesions measured: WT, 28; Δchap1, 10; Δskn7, 12; Δchap1-Δskn7, 25. WT and ΔΔ results are from two independent biological repeats combined. Bars indicate the SEM for the number of lesions (** and *** indicate significance at P < 0.01 and 0.001).
Download figure to PowerPoint
In vitro experiments showed that in response to some stressors, there is no additive contribution, whereas for others there is (Fig. 1). Loss of either of these transcription factors results in hypersensitivity to oxidants in plate assays, and the contribution of each is reflected in the expression of genes whose products allow the cell to cope with oxidative stress. ChAP1 is critical for increased expression of GLR1, TRR1, and TRX2 in response to hydrogen peroxide (Fig. 2), whereas loss of Skn7 results in a more modest decrease in expression as compared with WT. In the double mutant, expression of GLR1, for example, was about twofold lower than in Δchap1, reaching the level observed in the untreated WT control (no oxidant stress). In yeast, Yap1 and Skn7 coregulate some oxidative stress response genes (He et al., 2009), and our data provide the first genetic evidence, to our knowledge, that this mechanism acts in a filamentous fungus. Significantly, in the double mutant GLR1, TRX2 and SOD1 are not induced at all (Fig. 2). Thus, although Skn7 is not absolutely required for ChAP1 function, the combined contribution of both ChAP1 and Skn7 is needed for expression of GLR1, TRX2, and SOD1 in response to oxidant stress. The low transcript levels remaining in oxidant-stressed Δchap1 (Fig. 2) could still provide significant amounts of enzyme activity. Complete loss, in the double mutant, of oxidant-induced expression of some genes needed to cope with oxidative stress would imply that the ChAP1-dependent ROS detoxifying mechanism is severely impaired when Skn7 is absent. This prediction can be further tested at the protein abundance or enzyme activity levels. Skn7 control was most evident for the superoxide dismutase encoding gene SOD1, where ChAP1 control is minor if at all (Fig. 2). In Candida glabrata, superoxide dismutase (SOD) expression, critical for resistance to the superoxide-generating compound menadione, is independent of both CgSkn7 and CgYAP1 (Roetzer et al., 2011). The C. heterostrophus double mutant showed increased sensitivity to menadione (Fig. 1). Thus, the ‘wiring’ of the Skn7 and Yap1-dependent signaling pathways in the plant pathogen studied here is different from that in C. glabrata, but in both species SOD will be an important enzyme activity to study further.
On commercial hybrid maize cultivars Jubilee (Lev et al., 2005) and Royalty (this study), loss of ChAP1 did not compromise virulence in droplet inoculation assays. On the maize cultivar W64A, spray inoculation with Δchap1 resulted in about twofold decreased lesion size as compared with WT (Zhang et al., 2013). Necrotrophs like C. heterostrophus are thought to thrive in an oxidant-rich environment (see Heller & Tudzynski, 2011). The fungus thus must contend with ROS produced by both members of the host–pathogen pair. Skn7 senses not only oxidant stress, but also osmotic and cell wall stresses (Izumitsu et al., 2007; Oide et al., 2010; Fassler & West, 2011), and ChAP1 also appears to have redox-independent sensory functions (Shanmugam et al., 2010; Shalaby et al., 2012). In Candida glabrata, certain combinations of oxidative, nitrosative and osmotic stress were more potent than each alone (Kaloriti et al., 2012). On the plant, C. heterostrophus may encounter combined stresses; therefore, loss of virulence when both ChAP1 and Skn7 functions are missing could be related to factors other than oxidants, and the approach used in Candida could be applied to plant pathogens.
The finding (Fig. 2) that different members of the ChAP1 regulon are affected differently by loss of Skn7 suggests that a genome-wide study of these mutants will uncover classes of genes whose promoters bind different combinations of transcription factors that transduce oxidant-related signals. Furthermore, the Δskn7 mutant is highly sensitive to ROS, similar to Δchap1 (Fig. 1 and Oide et al., 2010), yet the expression of the panel of known antioxidant genes (Fig. 2) is only modestly reduced. Again, this suggests that the Skn7 regulon includes additional genes that are critical for tolerance to oxidants and other stresses. C. heterostrophus should be a good model necrotrophic pathogen in which to address these questions at the systems level, considering that the genome is being studied intensively (Ohm et al., 2012; Condon et al., 2013), as is the genetic basis for stress physiology (Lev et al., 2005; Igbaria et al., 2008; Oide et al., 2010; Wu et al., 2012; Zhang et al., 2013).