The following P. anserina strains were used: Wild-type strain “s” (Rizet, 1953), the PaCypD-mutants ΔPaCypD, PaCypD_OEx (Brust et al., 2010a), the metacaspase deletion mutants ΔPaMca1, ΔPaMca2, ΔPaMca1/ ΔPaMca2 (Hamann et al., 2007), the autophagy-deficiency mutant ΔPaAtg1 (Knuppertz et al., 2014), the PaSod3-overexpressing mutant PaSod3_OEx and the previously described mutants PaSod1::Gfp, PaSod3::Gfp (Zintel et al., 2010). These strains were used for the generation of the new mutants ΔPaCypD/PaSod3::Gfp, PaCypD_OEx/PaSod3::Gfp, ΔPaCypD/ PaSod1::Gfp, PaCypD_OEx/PaSod1::Gfp in addition to PaCypD_OEx/ ΔPaAtg1 and PaCypD_OEx/ΔPaMca1/ ΔPaMca2, which were verified by Southern blot and western blot analysis (Supplementary Figures 1–8). All strains were constructed in the genetic background of the wild-type strain “s.” Strains were grown on complete medium (BMM) containing cornmeal extract (Esser, 1974) at 27°C under constant light. For spore germination, the BMM medium was supplemented with 60 mM ammonium acetate and spores were incubated at 27°C in the dark for 2 d. All used strains originated from monokaryotic ascospores isolated from irregular asci. For subsequent experiments strains were either grown on solid semi-synthetic M2 medium (M2) or in liquid complete medium (CM) at 27° under constant light (Osiewacz et al., 2013).

Total DNA was isolated according to standard protocols (Lecellier and Silar, 1994). DNA restriction, gel electrophoresis, and Southern blotting were performed according to standard protocols (Sambrook et al., 1989). As a hybridization probe specific for the phleomycin resistance gene (Ble), the 1293 bp BamHI-fragment (ER0051, Thermo Scientific) of the plasmid pKO4 (Hamann et al., 2005; Luce and Osiewacz, 2009) was used. The 727 bp NcoI/ClaI-fragment (ER0571, ER0141, Thermo Scientific) of the plasmid pKO7 (Kunstmann and Osiewacz, 2009) was used as a hybridization probe specific for the hygromycin resistance gene (Hph). For Southern blot hybridization and detection of hybridizing bands, the instructions of the supplier (Roche, Germany) were followed.

The lifespan and growth rate of monokaryotic isolates was determined using race tubes containing M2 medium as described (Kunstmann and Osiewacz, 2008). The lifespan of P. anserina is defined as the time period in days (d) of linear hyphal growth whereas the growth rate is defined as the measured growth in centimeters (cm) per time period in days (d). For the analysis of the growth rate and lifespan the growth front was marked every 1–3 days until death of the individuals. From these data the mean lifespan was calculated as average of all individual isolates from each strain as previously described (Osiewacz et al., 2013). To determine the lifespan under oxidative stress or in presence of cyclosporine A, petri dishes containing 30 ml M2 medium with 80 or 160 μM of paraquat (Sigma-Aldrich, 856177), CuSO₄, a combination of both or 0.05 μg/ml CsA were inoculated with mycelium from germinated spores. Plates were incubated at 27°C under constant light or in the dark in case of plates containing light sensitive cyclosporine A.

For the generation of extracts of 11 days old cultures mycelium from 2 days old germinated spores of four different isolates of each strain (= biological replicates) were grown on M2 medium for 5 days at 27°C under constant light. M2 medium petri dishes covered with cellophane were inoculated with pieces of mycelium that derived from the growth front and incubated for additional 2 days. Afterwards the grown mycelia was transferred into CM-liquid medium and incubated for 2 further days at 27°C under constant light and shaking at 180 rpm. The harvested mycelia were ground in liquid nitrogen and used for the isolation of total protein extracts and measurement of the arginine-specific peptidase activity using the fluorochrome-coupled peptide Z-GGR-Amc as a substrate as described (Hamann et al., 2007). Peptidase activity in total protein extracts from each strain was measured twice.

Cultivation of at least three different isolates of each P. anserina strain was carried out as described above for the determination of the metacaspase activity. For the cultivation of 6 days old strains M2 medium petri dishes covered with cellophane were directly inoculated with mycelium from germinated spores. In case of growth under oxidative stress 60 μM paraquat (Sigma-Aldrich, 856177) was added to CM liquid medium. Harvested mycelia were ground in liquid nitrogen and used for the isolation of total protein extracts with 1:100 protease inhibitor cocktail set IV (Calbiochem) as described (Osiewacz et al., 2013). One hundred micrograms total protein extracts were fractionated by 2-phase SDS-PAGE (12% separating gels) according to the standard protocol (Brust et al., 2010a). After electrophoresis, proteins were transferred to PVDF membranes (Millipore, IPFL00010). Blocking, antibody incubation, and washing steps were performed according to the Odyssey western blot analysis handbook (LI-COR Biosciences, Bad Homburg, Germany). The following primary antibodies were used: Anti-GFP (mouse, 1:10,000 dilution, Sigma-Aldrich, G6795), anti-PaCYPD (rabbit, 1:5000 dilution) previously described in Brust et al. (2010a). In all analyses, secondary antibodies conjugated with IRDye680RD (1:15,000 dilution, goat anti-mouse: LI-COR Biosciences, 926–68,070) or IRDye800CW (1:15,000 dilution, goat anti-rabbit: LI-COR Biosciences, 926–32,211) were used. The Odyssey infrared scanner (LI-COR Biosciences) was used for detection and quantification using the manufacturer's software.

For statistical analyses of growth rate, metacaspase activity, and quantification of the GFP protein level in western blot analyses the 2-tailed student's t-test was used. For statistical analysis of lifespan the 2-tailed Wilcoxon rank-sum test was used. For all analyses statistical significance was calculated and defined as not significant (P > 0.05); as significant (^*P < 0.05); as highly significant (^**P < 0.01); very highly significant (^***P < 0.001).

Previously it was shown that PaCYPD is involved in the induction of PCD. The analysis of a PaCypD-overexpressing strain with an ectopic integration of the plasmid pCypDEx1 containing PaCypD under the control of the constitutive promoter of the metallothionein gene PaMt1, identified an accelerated cytochrome c release, nuclear condensation and a decrease in membrane potential as markers for an apoptotic type of PCD (Brust et al., 2010a). However, the study did not address the dependence of the underlying processes on the two P. anserina metacaspases PaMCA1 and PaMCA2 and thus not explicitly demonstrated a metacaspase-dependent ‘type I’ PCD. In order to investigate this role, we constructed a PaCypD_OEx/ΔPaMca1/ΔPaMca2 triple mutant in which both PaMca genes are deleted in the PaCypD overexpression background. Subsequently, we determined the arginine-specific peptidase activity of 11 days old cultures of the wild type, the PaCypD overexpressor, and the PaMca1/PaMca2 double deletion in the wild-type and the PaCypD overexpression background (Figure 1A). The arginine-specific peptidase activity of the ΔPaMca1/ΔPaMca2 deletion strain is very low [214 ± 27 RFU/(min^*mg)] compared to the activity of the wild type [2514 ± 376 RFU/(min^*mg)] indicating that the activity mainly results from the activity of PaMCA1 and PaMCA2. Strikingly, the PaCypD_OEx mutant is characterized by a considerably lower arginine-specific activity [981 ± 359 RFU/(min^*mg)] than the wild type, while the activity of the PaCypD_OEx/ΔPaMca1/ΔPaMca2 triple mutant [258 ± 82 RFU/(min^*mg)] is comparable to that of the ΔPaMca1/ΔPaMca2 deletion strain. From these data the metacaspase activity (Figure 1B) was calculated as the difference in the arginine-specific peptidase activity between wild type and ΔPaMca1/ΔPaMca2 strain or PaCypD_OEx and PaCypD_OEx/ΔPaMca1/ΔPaMca2 triple mutant. Surprisingly, compared to the wild type of the same age, the metacaspase activity in 11 days old PaCypD overexpressors is strongly reduced.

Next we determined the growth rate (Figure 1C) and lifespan (Figure 1D) of the different strains on semi-synthetic M2 medium. On this medium the growth rate and mean lifespan (Supplementary Figure 9) of the wild type and the ΔPaMca1/ΔPaMca2 strain do not significantly differ from each other. In contrast, overexpression of PaCypD in the wild-type and the ΔPaMca1/ΔPaMca2 genetic background leads to a similar reduction of the growth rate and the mean lifespan. Overall these data demonstrate that the accelerated aging process of PaCypD overexpressors does not rely on metacaspase-activity and “type I” PCD.

Next we addressed a link of PaCYPD and autophagy. For the demonstration of the vacuolar degradation of mitochondrial and cytosolic proteins via mitophagy and general autophagy, we utilized a recently established modified assay (Knuppertz et al., 2014) that originally was developed for the yeast S. cerevisiae (Meiling-Wesse et al., 2002; Welter et al., 2010). In this assay the degradation of reporter proteins is analyzed. In our analysis, mitochondrial PaSOD3::GFP served as a marker for mitophagy and cytosolic PaSOD1::GFP for general autophagy. After autophagy-mediated degradation of parts of the fusion proteins in the vacuole, “free GFP” remains stable and can be quantified by western blot analysis. We analyzed “free GFP” in total protein extracts of two different age stages of the wild type, the PaCypD overexpressor and the PaCypD deletion strain, which was obtained by homologous replacement of PaCypD by a selection marker containing a blasticidin resistance (Bsd) and phleomycin resistance (Ble) gene for selection in Escherichia coli and P. anserina (Brust et al., 2010a; Figures 2A,C). Compared to the wild type, the amount of “free GFP” derived from the PaSOD3::GFP mitophagy reporter protein was significantly lower in total protein extracts of ΔPaCypD of both 6 and 13 days old cultures. In striking contrast, compared to 6 days old cultures of the wild type, a 10-fold increase of “free GFP” was detected in extracts of PaCypD overexpressors of the same chronological age. Also, at the age of 13 days PaCypD overexpressing strains exhibit a higher amount of “free GFP” (Figure 2B). In addition, the analysis of “free GFP” resulting from degradation of the general autophagy reporter PaSOD1::GFP revealed that 6 days old wild-type and ΔPaCypD cultures display equal amounts of “free GFP,” indicating that the basal level of general autophagy is independent of PaCYPD. In striking contrast, we verified the age-related increase in general autophagy first reported for the wild type of P. anserina (Knuppertz et al., 2014) and found that it is PaCYPD-dependent. In ΔPaCypD strains “free GFP” of 11 days old cultures is significantly lower than in wild-type strains of the same age and no significant difference of “free GFP” is observed in 6 and 11 days old ΔPaCypD cultures. In addition, compared to 6 days old wild-type cultures, PaCypD overexpressors display a 3.5-fold increase of “free GFP” in 6 days cultures and a 4.6-fold increase at an age of 11 days (Figure 2D). The mitophagy reporter PaSOD3::GFP may also be degraded (together with cytosolic compounds) unspecifically to some extent in the PaCypD overexpressor. However, since the increase in degradation of PaSOD3::GFP is much higher than the increase in PaSOD1::GFP degradation (10-fold vs. 3.5-fold) a preferential degradation of PaSOD3::GFP via mitophagy can be concluded.

Next we investigated the impact of autophagy on growth and lifespan of PaCypD overexpressors. Ablation of PaATG1, an essential component of the autophagic machinery controlling early steps in autophagosome formation, impairs the pro-survival function of autophagy and results in a reduction of the growth rate and a short-lived phenotype (Knuppertz et al., 2014). Most strikingly, we found that growth rate and lifespan of the PaCypD_OEx/ΔPaAtg1 double mutant is not further reduced in comparison to the two single mutants (Figures 2E,F, Supplementary Figure 10), indicating an epistatic relation of PaAtg1 and PaCypD. Thus, a pro-survival role of autophagy, as it is found in the wild type, can be excluded in the PaCypD overexpressor. We conclude that lifespan reduction in this mutant is caused by the CYPD-dependent strong induction of general autophagy and mitophagy. PaCYPD thus appears to be involved in the switch from a pro-survival to a pro-death role of autophagy. Moreover, as indicated by the lifespan reduction in the PaCypD_OEx/ΔPaAtg1 double mutant, a complete suppression of autophagy in the PaCypD overexpression background is also not beneficial clearly demonstrating that balancing of autophagy flux is important for cellular homeostasis.

After the identification of a role of PaCYPD in the induction of autophagy we set out to investigate the underlying mechanisms in more detail. Since a role of ROS in aging and in the regulation of CYPD-induced mPT is well documented (Linard et al., 2009; Davalli et al., 2016), we next analyzed the impact of oxidative stress on the growth rate of the wild type and different PaCypD mutants (Figure 3A). We induced global oxidative stress in all cellular compartments via the addition of CuSO₄ to the growth medium. In addition, the herbicide paraquat was used to induce mitochondrial stress. Under standard conditions without the application of the two stressors, the growth rate of ΔPaCypD and PaCypD_OEx strains is slightly reduced compared to the growth rate of the wild type. On medium containing 80 and 160 μM CuSO₄, respectively, the growth rate of ΔPaCypD and the wild type do not differ. Growth of PaCypD_OEx is only slightly decreased compared to the wild type, indicating that the tolerance to Cu²⁺-induced global cellular oxidative stress is independent from PaCYPD. In contrast, compared to the wild type, 80 and 160 μM paraquat in the growth medium resulted in a significant increase of the growth rate of ΔPaCypD and a decreased growth of PaCypD_OEx on medium containing 80 μM paraquat. It appears, that ΔPaCypD strains are more tolerant to paraquat-mediated mitochondrial ROS stress. While all three investigated strains were able to grow on 160 μM CuSO₄ or 160 μM paraquat, respectively, only the wild type and the PaCypD deletion strain were tolerant to the addition of 80 μM of both ROS stressors. Under these conditions, growth of PaCypD_OEx was almost completely inhibited. These data are consistent with observations in higher eukaryotes in which CYPD facilitates ROS-induced mPTP-opening leading to the induction of cell death.

To validate this latter role in P. anserina, we tested whether the reduced lifespan of a mutant overexpressing PaSod3, encoding the mitochondrial superoxide dismutase, can be restored by the CYPD-inhibitor cyclosporine A. The mutant is characterized by increased ROS stress (Grimm and Osiewacz, 2015). Upon application of 0.05 μg/ml cyclosporine A to the growth medium, the growth rates of the wild type and PaSod3_OEx are significantly decreased, but do not differ from each other (Figure 3B). The cyclosporine A-mediated impairment of the growth rate is consistent with earlier findings reported for Neurospora crassa (Tropschug et al., 1989; Bardiya and Shiu, 2007) and was assumed to result from a PaCYPD/CsA interaction product, since the growth rate of ΔPaCypD is unaffected by CsA (Brust et al., 2010a). Moreover, chronic treatment with cyclosporine A does not affect the mean lifespan of the wild type. In striking contrast, the lifespan of PaSod3_OEx is significantly increased from 16.3 ± 2 to 20.4 ± 3 days (Figure 3C, Supplementary Figure 11). Together these data suggest that PaCYPD-inhibition or deficiency mediates an increased tolerance to mitochondrial but not to global oxidative stress. In addition, premature death caused by high mitochondrial oxidative stress can be counteracted by the CYPD-inhibitor cyclosporine A.

Since CYPD-mediated opening of the mPTP can be induced by ROS and since we recently found that autophagy is also induced by ROS in P. anserina (Knuppertz et al., submitted), we next analyzed whether the induction of autophagy depends on PaCYPD. In this study we compared “free GFP” levels in 6 days old cultures of wild-type and ΔPaCypD strains challenged by 60 μM paraquat (Figure 4A). As a control we used the strains of the same age grown in medium without paraquat from the age-related study described above (Figure 2D). In the wild type, paraquat led to a two-fold increase of “free GFP” (Figure 4B). Although paraquat also led to increased “free GFP” in ΔPaCypD, the level is significantly lower than in the wild type indicating that the paraquat-mediated induction of autophagy is indeed, at least partially, dependent on PaCypD.

Next we investigated whether the beneficial autophagy-dependent effect of mild paraquat stress identified in a parallel study (Knuppertz et al., submitted) also depends on PaCYPD. We compared the lifespans of the wild type and the ΔPaCypD mutant grown on standard growth medium containing increasing amounts of paraquat (Figure 4C). Strikingly, we found that the lifespan-extending effect of paraquat observed for the wild type is strongly reduced in the ΔPaCypD strain. On medium containing 20 μM paraquat the difference in mean lifespan is significant (Figure 4D). In contrast, at 80 μM the difference in mean lifespan is not significant and under harsh paraquat stress (i.e., 150 μM) the lifespan is strongly reduced in both strains. These results are consistent with previous investigations in other organisms like S. cerevisiae or Caenorhabditis elegans (Cypser and Johnson, 2002; Mesquita et al., 2010; Yang and Hekimi, 2010) and provide mechanistic clues explaining the beneficial effect of mild stress via the induction of adaptive responses leading to increased stress tolerance. They support a beneficial role of PaCYPD in the mediation of mitohormesis.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.