The P. anserina wild-type (WT) strain “s” (Rizet, 1953), the previously described reporter strain PaSod1::Gfp (Zintel et al., 2010), the PaClpP deletion strain (ΔPaClpP; Luce and Osiewacz, 2009), and the following recently generated strains were used: ΔPaSnf1, ΔPaSnf1/Flag::PaSnf1, ΔPaSnf1/PaSod1::Gfp, ΔPaSnf1/ΔPaClpP, and ΔPaSnf1/ΔPaClpP/PaSod1::Gfp. Double mutants were obtained by crossing single-mutant strains. The triple-mutant ΔPaSnf1/ΔPaClpP/PaSod1::Gfp was obtained by crossing of ΔPaSnf1/ΔPaClpP with PaSod1::Gfp. All mutant strains were constructed in the genetic background of the WT “s.” Unless otherwise stated, strains were grown on complete medium (BMM) containing cornmeal extract (Esser, 1974). To induce the formation of fruiting bodies, cultures with opposite mating types were placed on solid semi-synthetic M2 agar plates (M2) (Osiewacz et al., 2013) and incubated for 10–14 days at 27°C and constant light. For spore germination, the spores were incubated for 2 days at 27°C in the dark on BMM containing 60 mM ammonium acetate. Strains were grown on M2 (0.025% KH₂PO₄, 0.03% K₂HPO₄, 0.025% MgSO₄, 0.05% urea, 1% dextrin, 2% agar), M2 + glycerol (0.025% KH₂PO₄, 0.03% K₂HPO₄, 0.025% MgSO₄, 0.05% urea, 1% glycerol, 2% agar), or M2-N (0.025% KH₂PO₄, 0.03% K₂HPO₄, 0.025% MgSO₄, 1.5% dextrin, 2% agar) or in liquid complete medium [CM: 0.1% KH₂PO₄, 0.05% KCl, 0.05% MgSO₄, 1% glucose, 0.37% NH₄Cl, 0.2% tryptone, 0.2% yeast extract, 0.1% stock solution A (0.1% ZnSO₄, 0.1% FeCl₂, 0.1% MnCl₂, pH 6.5 (KOH)), pH 6.5 (KOH)] at 27°C under constant light (Osiewacz et al., 2013).

To analyze putative changes in the abundance of metabolites in the PaClpP deletion mutant compared to the wild type, a metabolomic profiling was performed with 5- and 20-day-old wild-type and ΔPaClpP strains (n = 4 each). Mycelial extracts were obtained from strains grown on M2 agar plates at 27°C and constant light for a specific time. Subsequently, the mycelium of the aged strains (17-day-old strains) and young strains (2-day-old strains, directly used from the BMM plates with ammonium acetate) were inoculated on M2 agar plates, covered with cellophane, and allowed to overgrow the plates for three additional days at 27°C and constant light. The agar pieces were removed, the mycelium was harvested and portioned in 50-mg aliquots into a 1.5-ml tube. The samples were frozen in liquid nitrogen and stored at −80°C until the final analysis. The analysis was performed by Metabolomic Discoveries (Metabolomic Discoveries GmbH, Potsdam, Germany). The used optimized method allows the unambiguous characterization and relative quantification of the requested metabolites. Metabolite analysis by LC–tandem mass spectrometry was performed using a LCMS-8050 by Shimadzu triple-quadrupole mass spectrometer equipped with an electrospray ionization (ESI) source and operated in the multiple reaction mode (MRM).

All strains were constructed in the genetic background of the wild-type strain “s” (Rizet, 1953). To generate the ΔPaSnf1 strain, we proceeded as described by Hamann et al. (2005). First, the flanking regions of PaSnf1 were amplified with the following oligonucleotides: For the 5′-flanking region, Snf1KO1 (CAAGGGCACGGTGCTGTC, Biomers, Ulm, Germany) and Snf1KO2 (TTAAGCTTCATGCTGGCGACGGCTATC, HindIII site is underlined, Biomers, Ulm, Germany) were used. For the 3′-flank, Snf1KO3 (CGCGACTAGTGCGGATGCGGA TTGATTATTG, BcuI site is underlined, Biomers, Ulm, Germany) and Snf1KO4 (AAGCGGCCGCTGTCGATGATA CAGATAG, NotI site is underlined, Biomers, Ulm, Germany) were used. These amplified flanking regions were cloned in the plasmid pKO4 (Hamann et al., 2005; Luce and Osiewacz, 2009), using the restriction enzymes KpnI, BcuI, NotI, and HindIII (Thermo Scientific, Waltham, MA, United States, ER0521, ER1251, ER0591, ER0501; KpnI site is part of the 5′-flank and thus has not to be added through the oligonucleotide). The wild-type PaSnf1 gene in a cosmid selected from a representative cosmid library (Osiewacz, 1994) was replaced in Escherichia coli by a combined blasticidin/phleomycin resistance cassette as described by Hamann et al. (2005). The resulting replacement cosmid was used to transform P. anserina wild-type spheroplasts (Osiewacz et al., 1991). Transformants were selected on a phleomycin- (Genaxxon, Ulm, Germany, M3429) containing medium. For the complementation of the PaSnf1 deletion mutant, and additionally to mark the PaSNF1 protein with a His::FLAG epitope, the sequence of PaSnf1 was fused with the sequence encoding this epitope. To achieve this, the sequence encoding the His::FLAG epitope was introduced in the plasmid pKO7 (Kunstmann and Osiewacz, 2009), which contains a hygromycin B resistance cassette. To include the PaSnf1 gene in this pHisFLAG vector, the gene was amplified in two parts. The first part consists of the promoter region, which was amplified with the oligonucleotides Snf1-4 (AAGCGGCCGCCATGGAAATAAGGACCCAG, NotI site is underlined) and Snf1-11 (TTGGATCCCATGCTGGCGACG GCTAT, BamHI site is underlined). The second part contains the PaSnf1 ORF and 500 bp of the terminator region, which were amplified with the oligonucleotides Snf1-12 (ATGGCCCAGGCCTACGACGACGA, this oligonucleotide contains a phosphate residue at the 5′-ending) and Snf1-9 (ACCTGCAGTGCCAGAATGGCCAGCCTTATG, PstI site is underlined). The PCR product of the PaSnf1 ORF and the pHisFLAG vector was digested with the restriction enzymes SmaI and PstI (Thermo Scientific, Waltham, MA, United States, ER0661, ER0611). The constructed vector pHis-Flag::PaSnf1-1 contains the PaSnf1 ORF fused with a sequence encoding a HIS::FLAG epitope at the 5′-ending. In a second step, the promoter region of PaSnf1 was integrated into the newly constructed plasmid pHis-Flag::PaSnf1-1 using the restriction enzyme sites of NotI and BamHI (Thermo Scientific, Waltham, MA, United States, ER0591, ER0051). The final plasmid pFlag-Snf1 contains the sequence of a HIS::FLAG epitope between the sequence of the promoter region of PaSnf1 and the PaSnf1 ORF with the terminator region and was transformed into P. anserina wild-type spheroplasts as previously described (Osiewacz et al., 1991), and hygromycin B [Merck Millipore (Calbiochem), Burlington, MA, United States, 400051]-resistant transformants were selected. Finally, the newly generated strain Flag::PaSnf1 was crossed with the PaSnf1 deletion mutant to remove the endogenous PaSnf1 sequence. The new ΔPaSnf1/Flag::PaSnf1 strain was selected by the phleomycin and hygromycin B resistance and verified by Southern blot analysis.

Total DNA was isolated as described before (Lecellier and Silar, 1994). Restriction of the DNA, gel electrophoresis, and Southern blotting were performed in accordance with standard protocols (Sambrook et al., 1989). Hybridization and detection were accomplished as described in the manufacturer’s protocol with digoxigenin-labeled hybridization probes (DIG DNA Labeling Mix, Sigma-Aldrich, St. Louis, MO, United States, 11175033910). The PaSnf1-specific probe was generated by PCR with the oligonucleotides Snf1-1 (GCATCGGCGCCTACAACATTGTC, Biomers, Ulm, Germany) and Snf1-2 (GAGACCGAAATCGGCGATCTTG, Biomers, Ulm, Germany). The phleomycin (Ble)-specific probe was generated from the plasmid pKO4 (Luce and Osiewacz, 2009), which contains the phleomycin resistance cassette. The respective fragment was recovered after digestion with BamHI. The resulting 1.3-kb fragment was extracted and purified from the agarose gel. The PaClpP-specific probe was generated according to Fischer et al. (2013).

For isolation of total RNA, strains were allowed to overgrow a cellophane foil-covered M2 petri dish for 3 days at 27°C and constant light. The mycelium was subsequently transferred into 2-ml tubes with 0.5-mm glass beads (Bertin Technologies, Montigny-le-Bretonneux, France, KT03961-1-004.2) and disrupted in a Precellys 24 homogenizer (Bertin Technologies; peqlab; Precellys 24 lysis and homogenization, Montigny-le-Bretonneux, France) for 2 × 25 s at 5800 rpm and a 10-s break in between. The isolation was performed using the NucleoSpin RNA Plant kit (Macherey Nagel #740949.250, Düren, Germany) according to the manufacturer’s instruction. Samples of wild type, ΔPaSnf1, ΔPaClpP, and ΔPaSnf1/ΔPaClpP strains were analyzed. 1 μg RNA was used for the reverse transcription with a Revert Aid Reverse transcriptase and Ribolock RNase inhibitor (Thermo Scientific, Waltham, MA, United States, EP0441, EC0381) and oligo(dT)₁₈ (Thermo Scientific, SO132). 2 μl of the resulting cDNA was used for the quantitative real-time PCR (qRT-PCR) reaction (IQ Sybr Green SuperMix, Bio-Rad, Hercules, CA, United States). To determine the expression of PaMdm38, the primers Pa_1_1789-1 (GCCCAGAAGGAAGAGTTTAC) and Pa_1-1780-2 (GTTGGAGACACCATCACTAC) were used. For each gene, the PCR efficiency was determined according to Pfaffl (2001). The relative expression level was normalized to the expression of PaPorin which was determined with the oligonucleotides Porin-RT-for (TCTCCTCCGGCAGCCTTG, Biomers, Ulm, Germany) and Porin-RT-rev (CGGAGGCGGACTTGTGAC, Biomers, Ulm, Germany). The relative expression level was calculated with the following formula:

The lifespan and growth rate of monokaryotic isolates at 27 and 35°C were determined using race tubes containing M2 medium as described previously (Osiewacz et al., 2013). Lifespan analyses on M2 with glycerol instead of dextrin, or on the autophagy-inducing M2 medium without the nitrogen source urea but with the 1.5-fold concentration of dextrin (M2-N), were performed in petri dishes. The petri dishes and the race tubes were incubated at 27 or 35°C and constant light. The lifespan is defined as the time period in days (d) of linear growth. The growth rate is defined as the growth in centimeters (cm) per day (d). For measurement, the growth front of the strains was marked every 2–3 days until death.

To analyze the content of the fruiting bodies and the spore morphology, confrontational crosses of different strains were performed on M2 medium. The strains were placed 3 cm apart of each other and were incubated for 12 (wild type × wild type; wild type × ΔPaSnf1; ΔPaSnf1/Flag::PaSnf1 × ΔPaSnf1; ΔPaSnf1 × ΔPaSnf1) and 14 days (ΔPaSnf1 × ΔPaSnf1) at 27°C and constant light. The content of a fruiting body was placed in a 20-μl H₂O droplet on a slide and covered with a cover slip. The samples were observed with the Leica Biomed Microscope (Leica Biosystems Nussloch GmbH, Nußloch, Germany) and documented with the camera DFC7000 T. To analyze the fruiting body content, the 10×/0.25 objective lens (506088) was used. To visualize spore morphology, the 40×/0.65 objective lens (506099) was used. For quantification of fruiting body content, spores of 10 fruiting bodies of each cross were analyzed (wild type × wild type n = 1691 spores, ΔPaSnf1 × ΔPaSnf1: n = 760 spores; wild type × ΔPaSnf1: n = 1595 spores; ΔPaSnf1/Flag::PaSnf1 × ΔPaSnf1: n = 1423 spores).

Determination of male (♂) or female (♀) fertility was performed by spermatization experiments. Cultures derived from mononucleate ascospores of the wild type (n = 3 of each mating type) and ΔPaSnf1 (n = 3 of mating type “minus” and n = 2 of mating type “plus”) were grown for 7 days on M2 agar plates at 27°C and constant light. To harvest spermatia, the mycelium was overlaid with 5 ml sterile water and incubated for 5 min. Afterward the water containing spermatia was collected from the plate. For spermatization, 300 μl of this suspension was dropped on mycelium of the opposite mating type (n = 5 technical replicates each). After 5 min of incubation, the droplets were removed and the plates were incubated for 5 days at 27°C and constant light. The number of perithecia obtained from wild-type crosses was set to 100%. For each case [wild type (♂) × wild type (♀); wild type (♂) × ΔPaSnf1 (♀), ΔPaSnf1 (♂) × wild type (♀); ΔPaSnf1 (♂) × ΔPaSnf1 (♀)], six different spermatizations were used for quantification.

For this experiment, P. anserina strains were cultivated on glass slides with a central depression containing 160 μl M2 medium for 1 day at 27°C and constant light in a moisture-prone chamber. For microscopic analysis, the Zeiss Cell Observer SD-Spinning Disk Confocal Microscope (Carl Zeiss Microscopy, Jena, Germany) with a 100×/1.46 oil objective lens (Carl Zeiss Microscopy, Jena, Germany, 420792-9800) was used. For visualization of mitochondria, the hyphae were stained with 100 μl of 2 μM MitoTracker Red (Invitrogen, Carlsbad, CA, United States) for 15 min in the dark. Subsequently, the hyphae were washed with 100 μl water. The mitochondria were visualized with the appropriate excitation and emission filters. For image processing, ZEN 2.5 (blue, Carl Zeiss Microscopy, Jena, Germany) was used.

The measurement of oxygen consumption rate was performed at 27°C by high-resolution respirometry (Oxygraph-2k series C and G, OROBOROS Instruments, Innsbruck, Austria). Cultures derived from mononucleate ascospores of wild type, ΔPaSnf1, ΔPaClpP, and ΔPaSnf1/ΔPaClpP were grown on M2 plates covered with a cellophane foil for 2 days, shifted into CM liquid medium, and further incubated for 3 days at 27°C under shaking and constant light. A small piece of mycelium was placed into the Oroboros respirometer chamber into 2 ml fresh CM media. The oxygen consumption rate was determined as previously described in Fischer et al. (2015a). After the measurement, the mycelium was placed into a 2-ml tube and boiled for 10 min at 95°C. Subsequently, the mycelium was air-dried for 2 days, and dry weight was determined. Finally, basal oxygen consumption per mg mycelium was calculated. For the analysis of the data, the DatLab6 software from Oroboros (Innsbruck, Austria) was used.

For isolation of total protein extracts, the strains were allowed to age on M2 plates at 27°C under constant light, after spore germination on BMM plates with 60 mM ammonium acetate for 2 days at 27°C in the dark. Mycelia from the different P. anserina strains and of different age were allowed to overgrow the cellophane foil-covered solid M2 medium for 3 days at 27°C and finally shifted to CM liquid medium and further incubated for 3 days at 27°C under constant light and shaking. To harvest the mycelium, the cultures were filtered over two layers of gauze. Approximately 250 mg mycelia was transferred in 2-ml tubes with 0.5-mm glass beads (Bertin Technologies, Montigny-le-Bretonneux, France, KT03961-1-004.2), and two volumes of protein extraction buffer (50 mM HEPES, 100 mM NaCl, 5 mM DTT, pH = 7.4) were added. The samples were kept on ice. For homogenization of the mycelia, the Precellys homogenizer (Bertin Technologies; peqlab; Precellys 24 lysis and homogenization, Montigny-le-Bretonneux, France) was used (2 × 25 s at 5800 rpm with a 10-s break in between). Afterward, the samples were centrifuged for 5 min at 9300 × g and 4°C. Subsequently, the supernatant was transferred to a new 1.5-ml reaction tube. The isolation of mitochondrial protein extracts was performed as previously described by Kunstmann and Osiewacz (2008). For western blot analysis, 50–150 μg protein extract was mixed with SDS-loading buffer (6% TRIS/HCl pH 6.8, 48% glycerol, 9% β-mercapto-ethanol, 0.03% bromophenol blue, 6% SDS) and heated up for 10 min at 95°C. The separation of proteins was accomplished by SDS-PAGE by standard protocols (Brust et al., 2010), and the proteins were subsequently transferred to a PVDF membrane (Immobilon-FL PVDF, Millipore, Burlington, MA, United States, IPFL00010). Blocking and antibody treatment of the membrane were performed according to protocols of the Odyssey^® Imaging System Western blot analysis (LI-COR Biosciences, Bad Homburg, Germany). The following primary antibodies were used for this study: anti-Phos-αAMPK (Phospho-AMPKα (Thr172) (40H9), rabbit mAb (dilution 1:300, Cell Signaling Technology, Danvers, MA, United States, #2535L), anti-FLAG (mouse, dilution 1:1,000, Sigma, F3165), anti-GFP (mouse, dilution 1:10,000, Sigma-Aldrich, St. Louis, MO, United States, G6795), and anti-PaDNM1 (rabbit, dilution 1:5,000, Sigma (NEP), polyclonal antibody raised against synthetic peptide “Ac-DKGTPEKESIAIRKC-amide,” affinity purification). IRDye^® 680RD anti-mouse (goat, dilution 1:15,000, LI-COR Biosciences, Bad Homburg, Germany, 926-68070) and IRDye^® 680LT anti-rabbit (goat, dilution 1:20,000, LI-COR Biosciences, Bad Homburg, Germany, 926-68021) were used as secondary antibodies.

The statistical analysis of lifespans was performed with the IBM SPSS statistics 19 software package, and a Kaplan–Meier survival estimate was generated. Significance was evaluated using three independent statistical tests (Breslow, generalized Wilcoxon; log rank, Mantel–Cox; Tarone-Ware). In the figure legends, only the lowest p-values are indicated. A table with the exact values can be found in the Supplementary Tables 2, 3. To determine the significance of the other experiments, two-tailed Student’s t-test was used (^∗∗∗: p < 0.001; ^∗∗: p < 0.01; ^∗: p < 0.05).

The previous demonstration of potential substrates of the mitochondrial PaCLPXP complex, the majority of which are associated with metabolic pathways in mitochondria (Fischer et al., 2015b), and the observation that the ablation of the protease and chaperone component of the complex leads to a pronounced lifespan extension (Fischer et al., 2013) directed our interest to more carefully investigate the impact of metabolic processes on aging of P. anserina.

In a first series of experiments, we investigated the metabolome of the ΔPaClpP mutant of 5- and 20-day-old cultures and compared it with that of the wild type of the same chronological age. In addition, the metabolomes of the wild type of 5- and 20-days of age were analyzed. At note, since the two strains are characterized by a different lifespan, their physiological status (i.e., their biological age) at a given chronological age is not identical. For instance, while 20-day-old wild-type strains are almost senescent and functionally impaired, ΔPaClpP mutant strains are rather unaffected.

The metabolome analysis was performed with cell extracts isolated from mycelia of four biological replicates of each strain and age and the relative amount of 50 metabolites was determined by triple-quadrupole mass spectrometry. The complete data set is presented in Supplementary Table 1 and visualized in a heat map (Figure 1). Most striking alterations occur in the abundance of components of the TCA cycle, of nucleotides, glycolysis metabolites, and amino acids between young (5-day-old) mutant and wild-type cultures (Figure 1; left). Interestingly, some of these alterations are also observed in wild-type cultures of different age (Figure 1; WT 20 day vs. 5 day). In particular, the abundance of glycolysis metabolites and of nucleotides changes during aging in the same direction compared to the situation in young mutant vs. young wild-type strains. It appears that these alterations are rather not responsible for the observed lifespan extension of the mutant. In contrast, changes in amino acid and TCA cycle metabolite levels differ in the aging wild type compared to young mutant vs. young wild type. Especially a significant increase in citrate/isocitrate and aconitate levels is observed in the 5-day-old mutant compared to wild type of the same chronological age. No change in abundance of these metabolites is observed in wild-type cultures of 5 and 20 days of age. Moreover, the abundance of several amino acids (arginine, proline, glycine, histidine, and threonine) is significantly lower in 5-day-old mutants compared to wild types of the same chronological age but not in 20-day-old wild types compared to 5-day-old wild-type cultures. In addition, during wild-type aging several metabolite levels (especially glycolysis metabolites) significantly decreased, while during ΔPaClpP aging (20 days vs. 5 days, Figure 1; right) rather an overall increase in metabolite levels is observed. This effect may result from the upregulation of non-selective autophagy in the ΔPaClpP mutant that was observed in a previous study (Knuppertz and Osiewacz, 2017) as a compensatory response to mitochondrial impairments in the mutant. This idea is supported by the observed changes in the abundance of several nucleotides. Most interestingly, AMP is increased while ATP levels decrease in the young mutant compared to young wild type. This increase in the AMP/ATP ratio directed us to a potential induction of AMPK signaling in the ΔPaClpP mutant, which may explain its lifespan-extended phenotype. To address this possibility experimentally, we set out to delete the gene encoding the catalytic α-subunit of the AMPK complex, termed SNF1 in fungi, which becomes activated by an increased AMP/ATP ratio. During the course of our investigation, Li et al. (2020) reported the generation and characterization of such a deletion mutant of P. anserina (in wild-type background “S”) with special emphasis on lignocellulose degradation. In their study, the impact of PaSNF1 on aging and lifespan control was not addressed.

Deletion of PaSnf1 was performed following the procedure described in Hamann et al. (2005) in wild-type strain “s” Rizet (1953). The genomic sequence Pa_2_770 annotated to encode PaSNF1 (UniProt B2B4C1) was replaced by a phleomycin resistance cassette. A double mutant, lacking PaSnf1 and PaClpP, was selected from the progeny of crosses of the newly isolated PaSnf1 mutant with the previously isolated ΔPaClpP strain (Fischer et al., 2013). The correct genotype of strains was verified by Southern blot analysis (see Supplementary Figure 1A). In addition, western blot analysis using a commercially available anti-phospho-α-AMPK antibody failed to detect phosphorylated and thus activated AMPK in the PaSnf1 deletion strain (Supplementary Figure 1B). As control, a plasmid encoding FLAG::PaSNF1 was introduced in the ΔPaSnf1 strain. The resulting ΔPaSnf1/Flag::PaSnf1 strain was verified by Southern blot analysis (Supplementary Figure 1C), and production of the FLAG-tagged PaSNF1 protein was demonstrated by western blot analysis (Supplementary Figure 1D).

Except for a significantly reduced growth rate, the PaSnf1 deletion strain does not show any obvious macroscopic differences (i.e., pigmentation, aerial hyphae formation) compared to the wild type grown on M2 standard medium (Figures 2A,B). As described earlier, the ΔPaClpP strain has a wild-type-like growth rate (Fischer et al., 2013). In the double mutant ΔPaSnf1/ΔPaClpP, growth rate is reduced like in ΔPaSnf1 (Figure 2B), suggesting that growth impairment of ΔPaClpP is epistatic. Introduction of the Flag-tagged PaSnf1 copy into the PaSnf1 deletion strain restored the wild-type growth rate (Figures 2A,B). In addition to the reduction in growth rate, the analysis of sexual reproduction revealed severe impairments in the ΔPaSnf1 strain. In our analyses, we investigated the fertility of PaSnf1 deletion strains originating from monokaryotic ascospores. These cultures were either of mating type “plus (+)” or “minus (−)” and form fruiting bodies only after contact with a partner of the opposite mating type. At note, mating of P. anserina is possible by placing mycelial pieces of the two strains of opposite mating type on opposite positions on one agar plate. After the partners contact each other, the male gametes (spermatia) of each of the two strains fertilize the female gametangia (protoperithecia) and fruiting bodies (perithecia) are subsequently formed. This kind of “confrontational cross” allows to investigate the efficiency and morphological differences in the formation of reproductive structures (perithecia, ascospores) but not the discrimination of the functionality of male gametes and female gametangia, respectively. Such discrimination is possible by spermatization of a growing culture of one mating type (the “acceptor” strain) with spermatia of the opposite mating type isolated from the “donor” strain grown on a separate agar plate. In this case, fruiting bodies are only formed when spermatia are poured on the “acceptor” strain and female and male fertility can be clearly attributed and measured.

In our study, we first investigated the content of the fruiting bodies derived from confrontational crosses of different strains by light microscopy (Figure 2C). In crosses of two wild-type strains of the opposite mating type, normal mature spores occur after 12 days of incubation at 27°C (constant light). In contrast, after this time, only few mature spores are visible in fruiting bodies of ΔPaSnf1 × ΔPaSnf1 crosses. However, after two additional days of incubation, in the mutant’s fruiting bodies more mature spores occur (Figure 2C, right) demonstrating that ablation of PaSNF1 delays spore development. Even more, the morphology of the ascospores in these crosses is clearly distinct from those in which one parent is wild type. The ascospores derived from a ΔPaSnf1 × ΔPaSnf1 cross are rather round instead of elliptical, suggesting that PaSNF1 is involved in spore morphology regulation. The presence of one PaSnf1 gene within the fruiting bodies is sufficient (e.g., in a wild type × ΔPaSnf1 cross) to induce normal spore morphology (Figures 2C,D). Complementation with an ectopically integrated PaSnf1 gene fused to a Flag tag restores spore morphology (Figures 2C,D).

Next, we performed spermatization experiments and found that female fertility is strongly affected in PaSnf1 deletion strains. While spermatia of the ΔPaSnf1 strain are readily able to induce fruiting body formation on wild-type acceptor strains, the spermatization of ΔPaSnf1 by spermatia of the wild type (reciprocal cross) leads to a reduction in fruiting body formation by 57% (Figure 2E). In crosses of male and female ΔPaSnf1, the number of fruiting bodies is not further reduced, indicating that it is female fertility that is impaired in the mutant.

Finally, we investigated the impact of PaSnf1 deletion on lifespan. We analyzed individual cultures derived from monokaryotic ascospores grown at 27°C on the M2 medium containing glucose as a carbon source (Figure 3A and Supplementary Table 2). As described earlier (Fischer et al., 2013), ΔPaClpP is long-lived. Compared to the wild type, the PaSnf1 deletion strain shows a 22% mean lifespan extension (Figure 3B and Supplementary Table 2, mean lifespan 28 days vs. wild-type mean lifespan of 23 days). Complementation of ΔPaSnf1 with an ectopically integrated Flag-tagged PaSnf1 gene restores wild-type lifespan (Supplementary Figure 1E and Supplementary Table 3). Most surprisingly, in contrast to our initial expectation, the ΔPaClpP/ΔPaSnf1 double mutant did not display a wild-type like lifespan. In fact, the simultaneous deletion of the two genes was found to be synergistic leading to a spectacular extension of the mean lifespan to more than 186 days (Figures 3A,B and Supplementary Table 2). This is an increase of at least 174 and 564%, respectively, in comparison to the PaClpP and PaSnf1 single deletion strains. At note, the exact values could not be calculated yet, since some of the double mutant strains are still alive. It is striking that in the double mutant the mortality rate is strongly reduced in particular after about 50% of the investigated individual cultures died (Figure 3A).

This biphasic progression of the survival curve is rather unusual, and therefore, we examined the lifespan data in more detail. In P. anserina, it is known that the mating type has an impact on lifespan. In most cases (e.g., wild type), this impact is rather moderate. However, in combination with some genes it can be stronger (Belcour et al., 1991; Contamine et al., 1996, 2004; Sellem et al., 2005; Adam et al., 2012). In fact, this mating type-linked effect is due to two alleles of the rmp1 gene that are located in very close proximity to the mating-type alleles and therefore are not separated by intrachromosomal recombination during sexual reproduction and co-inherited with the mating type of the parental strain (Contamine et al., 2004). We therefore decided to split the combined lifespan curves of Figure 3A in which mating type “plus” and mating type “minus” isolates are both presented into two mating type-specific curves (Figures 3C,D). Significantly, we found a strong effect of the mating type on lifespan of ΔPaClpP/ΔPaSnf1. While the lifespan of mating-type “plus” double mutants are only moderately increased, in mating-type “minus” background lifespan is dramatically increased. Thus, the original biphasic progression of the survival curve is due to the fact that the “plus” strains have a shorter lifespan than the “minus” strains. In contrast, the effect of the mating type on lifespan of the wild type and the PaSnf1 and PaClpP single deletion strains was rather moderate (Figures 3C,D and Supplementary Table 3). Because of this observation, we provide lifespan values for both combined mating types (Supplementary Table 2) and lifespans separated for the two mating types (Supplementary Table 3).

From a previous study, we know that the observed lifespan extension may only be valid under the specific laboratory conditions applied in the corresponding study (Weil et al., 2011). In particular, incubation temperature, which – under natural conditions – strongly varies during the day, may have a clear impact. In order to investigate such an impact, we analyzed cultures derived from monokaryotic spores of all strains incubated at 35°C (Figures 3E,F and Supplementary Tables 2,3). All strains showed a strong reduction in lifespan when grown at this temperature. The mean lifespan of wild type, ΔPaClpP, and ΔPaSnf1 is 10, 9, and 10 days, respectively (Figure 3E and Supplementary Table 2) compared to 23, 68, and 28 days at 27°C (Figure 3B and Supplementary Table 2). In wild type and ΔPaSnf1, the lifespan at 35°C is not affected by the mating type. However, interestingly, strains ablated for PaCLPP with rmp1-2 (mating type “plus”) have a strongly reduced lifespan (Figure 3F). In contrast, the ΔPaClpP strain with rmp1-1 allele (mating type “minus”) has a longer lifespan than wild type “minus” (Figure 3F).

AMP-activated protein kinase is known to activate autophagy through different mechanisms, e.g., directly by phosphorylation of ULK1 and BECLIN1 (the mammalian homologs of ATG1 and ATG6, Kim et al., 2011, 2013) and indirectly by inhibiting TOR (Russell et al., 2014). Therefore, we expected that ablation of PaSNF1 results in impaired autophagy. To test this possibility, we measured autophagy in the wild type, ΔPaSnf1, and ΔPaSnf1/ΔPaClpP using the previously established reporter degradation assay (Knuppertz et al., 2014). In this assay, after autophagic degradation of the cytosolic reporter protein PaSOD1::GFP, the GFP moiety remains stable in the fungal vacuole and can be detected by western blot analyses. The results obtained with this reporter are independent of the mating type, and therefore, we presented the data of “plus” and “minus” isolates together. As expected, in 7-day-old ΔPaSnf1 strains autophagy is significantly impaired and reduced amounts of free GFP are found (Figures 4A,B). At 20 days of age, this difference is reduced and no longer significant, suggesting that aging induces PaSNF1-independent autophagy. Surprisingly, in 7-day-old ΔPaSnf1/ΔPaClpP autophagy is indistinguishable from that of 7-day-old wild type. In contrast, in 20-day-old cultures autophagic flux is significantly lower in the double mutant compared to wild type. Thus, a beneficial induction of autophagy, as it is known for the single PaClpP deletion mutant (Knuppertz and Osiewacz, 2017), is not responsible for the lifespan extension of the double mutant.

Next, we tested whether or not the observed impairment in autophagy of the PaSnf1 deletion strain affects the ability to respond to nitrogen limitation, a well-known condition inducing autophagy in different organisms including P. anserina (Todde et al., 2009; Mukaiyama et al., 2010; Knuppertz et al., 2014; Li et al., 2015). We determined the lifespan of all investigated strains on M2 medium without urea as nitrogen source and supplemented with additional glucose (Figures 4C,D). Nitrogen limitation leads to a pronounced lifespan extension in the wild type (40 days, Figure 4D, vs. 23 days on M2, Figure 3B, respectively, Supplementary Table 2), which is independent of the mating type (Figure 4D and Supplementary Table 3). The mean lifespan of ΔPaClpP is slightly reduced compared to the lifespan on M2 media with nitrogen (62 days, Figure 4D, vs. 68 days on M2, Figure 3B, Supplementary Table 2, respectively) but still longer than that of the wild type. In contrast, under nitrogen-replete conditions both ΔPaSnf1 and ΔPaSnf1/ΔPaClpP are characterized by a strongly reduced lifespan, which is 7 days for both strains (Figure 4D and Supplementary Table 2) compared to 28 days and at least 186 days on M2 (Figure 3B and Supplementary Table 2). These results identified a pivotal role of PaSNF1 in lifespan control under nitrogen-limiting conditions. Moreover, under these conditions, the deletion of PaSnf1 is epistatic over the deletion of PaClpP again indicating a link of PaClpP and PaSnf1 controlled pathways.

Since AMPK is known to be involved in the regulation mitochondrial homeostasis via mitochondrial biogenesis and dynamics (reviewed in Herzig and Shaw, 2018), we next investigated the impact of the deletion of PaSnf1 on mitochondria. To this end, we analyzed the mitochondrial morphology of wild type, the two single deletion mutants, and the ΔPaSnf1/ΔPaClpP double mutant by fluorescence microscopy (Figure 5A). As described earlier, during aging of P. anserina mitochondria become fragmented (Scheckhuber et al., 2007). In 20-day-old wild-type hyphae all mitochondria are fragmented, while in ΔPaClpP some mitochondria remain filamentous (Figure 5A). Interestingly, we found that both ΔPaSnf1 and ΔPaSnf1/ΔPaClpP still contain a substantial fraction of filamentous mitochondria even at 20 days of age – independent of the mating type. Thus, as described in other systems (Toyama et al., 2016; Willis et al., 2018; Chen et al., 2019; Wu et al., 2020), PaSNF1 appears to be crucial for fission of P. anserina mitochondria.

Recently, in HeLa cells the translocation of AMPK from the cytosol to mitochondria was demonstrated upon energy stress, thereby regulating mitochondrial division and autophagy (Hu et al., 2020). We wondered whether ablation of PaSNF1 might impair the recruitment of the mitochondrial fission factor PaDNM1 to mitochondria leading to impaired mitochondrial fission. Scheckhuber et al. (2007) showed that the PaDnm1 gene is only poorly expressed in young cultures. Therefore, we first analyzed the amount of PaDNM1 in total protein extracts of 20-day-old cultures (Figures 5B,C). Interestingly, in all mutants, the amount of PaDNM1 in total protein extracts is higher than in the wild type. However, only the increase in the double mutant is significant (Figure 5C). Subsequently, we investigated whether recruitment of PaDNM1 to mitochondria is altered in the mutants in comparison to the wild type. Surprisingly, on first glance, in purified mitochondria of 20-day-old ΔPaSnf1 the amount of PaDNM1 appears to be elevated (Figures 5D,E). However, the different biological replicates show a high variability which is not dependent on the mating type. The differences in the abundance of PaDNM1 are not significant, and the ablation of PaSNF1 thus does not impair PaDNM1 recruitment to the mitochondrial fraction. Nonetheless, PaDNM1 function appears to be affected in the mutant strains. While in all young cultures and in 20-day-old wild types only one PaDNM1 protein of about 120 kDa is visible, in 20-day-old mutant strains an additional band of about 125 kDa can be detected (Figure 5D). This band is very prominent in ΔPaSnf1, suggesting that PaSNF1 ablation may affect post-translational modification of PaDNM1. It thus is possible that this modification may lead to inactivation of PaDNM1 and, as a consequence, to reduced mitochondrial fission and to filamentous mitochondria even in 20-day-old cultures. In the double mutant as well as in ΔPaClpP, both PaDNM1 bands are present in equal amounts. Thus, the observed differences in mitochondrial morphology between these two strains cannot be explained by distinct post-translational modification of PaDNM1.

In search of a possible explanation for the observed differences in mitochondrial morphology, we carefully inspected the transcriptome data of the PaSnf1 deletion mutant which was recently published by Li et al. (2020). Interestingly, one gene, PODANS_1_1780, was found to be strongly upregulated in ΔPaSnf1 compared to the wild type. The corresponding protein, PODANS_1_1780 (hereafter termed PaMDM38), is a homolog of yeast MDM38 and human LETM1 protein (Nowikovsky et al., 2007; Nakamura et al., 2020), proteins which are implicated in the regulation of mitochondrial morphology. Therefore, we investigated PaMdm38 transcription by qRT-PCR and found that the gene is indeed upregulated in ΔPaSnf1 (Supplementary Figure 2), but not in other mutants, suggesting that post-translational modification of PaDNM1 and PaMdm38 expression is linked.

Our results indicate that ablation of PaCLPP is able to counteract changes in PaDNM1 recruitment and post-translational modification as well as the induction of PaMdm38 expression, suggesting that PaSnf1 affects mitochondrial pathways. We therefore tested mitochondrial function in the different strains by measuring oxygen consumption of fungal mycelium at three different age stages (7, 12, and 20 days) in fresh liquid medium containing 1% glucose (Figure 6A). We found that aging does not significantly affect the oxygen consumption rate (OCR) in the investigated strains. Markedly, the OCR of ΔPaClpP is lower than the OCR of the wild type as described before (Knuppertz and Osiewacz, 2017). In contrast, at these conditions, OCR is not significantly affected in ΔPaSnf1 (Figure 6A). A small but significant increase in OCR is observed in 12-day-old ΔPaSnf1/ΔPaClpP (Figure 6A). This moderate impact on OCR seen upon PaSNF1 ablation is in agreement with the observation that yeast SNF1 only becomes activated upon glucose depletion (Vallejo et al., 2020). If this response is also true in P. anserina, ablation of PaSNF1 should not affect OCR at high glucose concentrations. However, it can be expected that on medium containing non-fermentable glycerol as the sole carbon source, PaSNF1 is important. To test this possibility, we performed a lifespan analysis on M2 medium containing glycerol. On this substrate, all strains show a reduction in mean lifespan. It is 23 days in wild type, 9 days in ΔPaClpP, 10 days in ΔPaSnf1, and 7 days in ΔPaSnf1/ΔPaClpP, compared to 23, 68, 28, and at least 186 days on M2 with 1% glucose (Figures 6B,C vs. Figure 3A,B and Supplementary Table 2). We also tested whether or not rmp1 affects survival on glycerol medium (Supplementary Table 3). Indeed, similar to the situation at 27 and 35°C with 1% glucose, also growth on glycerol is affected by the rmp1 allele (Figure 6C and Supplementary Table 3). Most intriguingly, the lifespan of wild-type strains is clearly higher in the presence of rmp1-1 (mating type “minus”) than rmp1-2 (mating type “plus”). On this medium, rmp1-1 or rmp1-2 has no effect in the mutant strains ΔPaClpP, ΔPaSnf1, and ΔPaSnf1/ΔPaClpP. Thus, rmp1-1 ensures the ability to use the non-fermentable carbon source glycerol in the wild type. This ability is reduced in all deletion mutants, clearly demonstrating that they are characterized by impaired mitochondrial function although mitochondria morphology is filamentous.

From these observations, several conclusions can be drawn: First, the lifespan of P. anserina carrying the rmp1-2 allele is reduced on the medium, which requires functional mitochondria indicating a critical lifespan-limiting role of the rmp1 gene. Second, survival of ΔPaClpP is significantly more strongly compromised on glycerol than survival of the wild type. This demonstrates that mitochondrial function is impaired in this mutant, which is in accordance with the observed reduction in OCR. Thus, ablation of PaCLPP affects mitochondrial function independent of the available carbon source. Third, although ablation of PaSNF1 alone does not reduce OCR at 1% glucose, it strongly impairs survival on glycerol containing M2 (Figures 6B,C). Thus, PaSNF1 seems to control mitochondrial function only in the glycerol medium, which suggests that PaSNF1 is required for the metabolic switch from glycolysis to respiration. Obviously, PaCLPP and PaSNF1 affect mitochondrial function independently. Overall, our results uncover a hitherto unknown mechanism by which PaSNF1, RMP1, and PaCLPP synergistically interact with each other and affect the lifespan of P. anserina.