Peroxisomal catalase deficiency modulates yeast lifespan depending on growth conditions.

We studied the role of peroxisomal catalase in chronological aging of the yeastHansenula polymorpha in relation to various growth substrates. Catalase-deficient (cat) cells showed a similar chronological life span (CLS) relative to the wild-type control upon growth on carbon and nitrogen sources that are not oxidized by peroxisomal enzymes. However, when media contained methylamine, which is oxidized by peroxisomal amine oxidase, the CLS of cat cells was significantly reduced. Conversely, the CLS of cat cells was enhanced relative to the wild-type control, when cells were grown on methanol, which is oxidized by peroxisomal alcohol oxidase. At these conditions strongly enhanced ROS levels were observed during the exponential growth phase of cat cells. This was paralleled by activation of the transcription factor Yap1, as well as an increase in the levels of the antioxidant enzymes cytochrome c peroxidase and superoxide dismutase. Upon deletion of the genes encoding Yap1 or cytochrome c peroxidase, the CLS extension of cat cells on methanol was abolished. These findings reveal for the first time an important role of enhanced cytochrome c peroxidase levels in yeast CLS extension.

INTRODUCTION mitochondrial redox balance [18], which furthermore implicates a role of peroxisomal catalase in aging.
Saccharomyces cerevisiae is widely used as a model organism to study the molecular mechanisms of aging [19]. In contrast to mammals, this yeast species contains two catalase genes encoding peroxisomal catalase A (Cta1) and cytosolic catalase T (Ctt1), respectively [20,21]. Deletion of CTA1 was shown to cause a decrease in the chronological lifespan (CLS) [22]. However, in another study deletion of CTA1 caused an increase in CLS. This was explained by the elevation of H 2 O 2 levels, which triggered expression of superoxide dismutase (SOD2) thereby decreasing the level of superoxide anions [23]. Moreover, overexpression of CTA1 was shown to decrease the CLS of S. cerevisiae. Both observations would be in line with the redox stress hypothesis.
To better understand the role of peroxisomal catalase in aging, we used the yeast Hansenula polymorpha as a model organism. Akin to mammalian cells, this organism has only one catalase, which is peroxisomal [24,25]. In contrast to S. cerevisiae which contains only one peroxisomal oxidase, acyl CoA oxidase, H. polymorpha peroxisomes contain in addition multiple other oxidases, like in mammals.
In this study we analyzed the CLS of wild-type (WT) and catalase-deficient (cat) H. polymorpha cells upon cultivation on media containing different carbon and nitrogen sources that do or do not involve peroxisome function. During growth on glucose or glycerol as carbon source in the presence of ammonium sulfate as nitrogen source peroxisomal enzymes are not required for the metabolism of the primary carbon and nitrogen sources. However, when media contain methylamine as sole nitrogen source or methanol as carbon source, the peroxisomal oxidases amine oxidase (AMO) and alcohol oxidase (AO) are required for growth. Our data indicate that the effects of the absence of peroxisomal catalase on the CLS is highly variable (ranging from a negative to a positive effect) and depends on the growth substrates.

Peroxisomal metabolism triggers lifespan changes in catalase deficient H. polymorpha
To analyze the role of peroxisomal catalase in the CLS of H. polymorpha, CLS measurements were performed using WT and cat strain grown on different carbon and nitrogen sources. Survival measurements started (T = 0 h) when the cultures had entered the stationary phase (Fig. S1ABCD). Very similar CLS curves and mean and maximal lifespans were observed for WT and cat culture grown on glucose/ammonium sulfate (Glc/AS) or glycerol/ammonium sulfate (Gly/AS) (Fig. 1AB, Table 1).

www.impactaging.com
When we used glucose/methylamine media (Glc/MA), cat cultures showed a significant reduction in both median and maximum lifespan compared to the WT control (Fig. 1C, Table 1). This reduction was not related to growth differences, as the doubling times and growth yields of both cultures were identical (Fig. S1B). Upon growth on glycerol/methanol/ammonium sulfate (Gly/MeOH/AS) the median lifespans were similar, but the maximum lifespan of cat cultures was significantly enhanced relative to the WT control (Fig. 1D, Table 1). This effect was not related to differences in methanol consumption in both cultures (Fig. S1E). The final optical density of the cat culture on Gly/MeOH/AS was however lower relative to that of WT culture, as reported previously [26] (Fig. S1D).

Catalase deletion affects intracellular ROS levels
To check the impact of CAT deletion on ROS levels, we determined the levels of these compounds using the fluorescent dye dihydrorodhamine 123 (DHR) and FACS (Fig. 2). As shown in Figure 2A cells in the exponential growth phase on Glc/AS, Gly/AS and Glc/MA showed very similar, low ROS levels. Upon growth of WT cells on Gly/MeOH/AS ROS levels were slightly increased. However, in cat cells grown on Gly/MeOH/AS ROS levels were strongly enhanced relatively to that in cat cells grown on the other substrates as well as compared to the WT control grown on Gly/MeOH/AS (>13 x).
During chronological aging ROS levels increased in all cultures with time. No differences in ROS levels were observed during chronological aging of Glc/AS or Gly/AS grown WT and cat cultures (Fig. 2BC). However, in cat cultures grown on Glc/MA and Gly/MeOH/AS ROS levels increased to higher levels compared to the WT controls ( Fig 2DE).
In the Gly/MeOH/AS culture the ROS levels were lower at day 1 (40 h after inoculation; Fig. 2E) relative to the levels observed in the exponential cultures (16 h after inoculation; Fig. 2A) most likely related to the fact that in the stationary phase methanol oxidation had ceased (Fig. S1E).

Activation of stress adaptation pathways
The high ROS levels observed during the exponential growth phase of Gly/MeOH/AS cat cultures may have induced stress response genes, which are often linked to lifespan extension [27][28][29]. To test this we analyzed the resistance of WT and cat cells to externally added H 2 O 2 or acrolein. Acrolein is a toxic byproduct of lipid peroxidation, which occurs in living cells under conditions of oxidative stress.  www.impactaging.com Independent of the cultivation conditions used, cat cells displayed decreased resistance to externally added H 2 O 2 in comparison to the WT (Fig. 3A), which most likely can be fully explained by the absence of catalase activity in these cells. However, cat cells grown on Gly/MeOH/AS were more resistant to H 2 O 2 than Glc/AS, Glc/MA or Gly/AS grown cat cells, suggesting that a catalase independent H 2 O 2 defense pathway is induced in these cells. Both WT and cat cultures showed lowest acrolein resistance upon growth on Glc/AS, whereas the resistance slightly increased for both strains upon growth on Glc/MA or Gly/AS, and further increased upon growth on Gly/MeOH/AS. Interestingly, cat cells grown on Gly/MeOH/AS showed higher acrolein resistance than WT cells (Fig. 3B).  [32], S. pombe [33] and S. cerevisiae [34]. The putative H. polymorpha Yap1 homologue showed 35% sequence identity to P. pastoris Yap1, 25% to S. pombe Yap1 and 24% to S. cerevisiae Yap1. YAP1 (accession number EFW96135) encodes a protein of 420 amino acids containing a predicted nuclear localization signal (NLS) [35] at the N-terminus (amino acids 30-50) and a leucine rich nuclear export signal (NES) [36] at the C terminus (amino acids 386-392).
Moreover, it contains 6 cysteines, two of which are present within the NES. Sequence alignments indicated that those features are conserved (Fig. S2).

Yap1 is involved in H 2 O 2 and acrolein resistance of H. polymorpha cat cells
We first analyzed the role of Yap1 in stress resistance using yap1 and cat yap1 deletion strains. These strains showed similar growth profiles as cat and WT cultures (compare Fig. S1) in media containing Glc/AS, Glc/MA or Gly/AS (Fig. S3ABC). In Gly/MeOH/AS medium, cat yap1 cells showed a slightly enhanced doubling time and a reduced yield relative to cat cultures ( Fig. S3D; compare Fig. S1D). Analysis of the resistance to H 2 O 2 and acrolein, revealed that this was reduced in cat yap1 cells relative to the cat controls (Fig. 4AB).  Next, we tested whether Yap1 plays a role in CLS determination. No strong differences in CLS between yap1 and cat yap1 cultures were observed upon growth on Glc/AS or Gly/AS (Fig. 4CD). Moreover, the curves were similar to those observed for WT and cat (Fig. 1). On Glc/MA, the yap1 curve resembled that of WT, whereas the cat1 yap1 survival curve was similar to that of cat cells (Fig. 4E, compare Fig. 1C), indicating that Yap1 is not an important player in determining the CLS at these conditions. The CLS of yap1 cultures on Gly/MeOH/AS was comparable to that of the WT control (Fig. 4F, compare Fig. 1D). In contrast a very rapid decrease in survival was observed for Gly/MeOH/AS grown cat yap1 cells resulting in less than 10% survival within 12 h after shifting the cells to this medium (Fig. 4F, inset). These data indicate that Yap1 is of major importance for survival of Gly/MeOH/AS grown cat cultures.
To seek further evidence for the function of Yap1, we performed localization experiments using a gene encoding an N-terminal fusion of mGFP with Yap1 under control of the YAP1 promoter. Growth and CLS experiments revealed that the fusion protein fully complemented the yap1 deletion strain (data not shown). Upon growth on Glc/AS, Glc/MA or Gly/AS, mGFP-Yap1 was predominantly localized to the cytosol in WT and cat cells (data not shown). However, after the shift of Glc/AS-grown cells to Gly/MeOH/AS, mGFP-Yap1 rapidly migrated (within two hours) to the nucleus of cat cells, but not of WT cells (Fig. 4GH). These data reveal rapid Yap1 activation upon exposure of cat cells to methanol.
Cytochrome c peroxidase and superoxide dismutase, but not glutathione reductase, are induced in Gly/MeOH/AS grown cat cells Induction of antioxidant enzymes has been implicated in yeast lifespan extension. We determined the activities of cytochrome c peroxidase (CCP), superoxide dismutase (SOD) and glutathione reductase (GLR) in crude extracts of WT and cat cells grown for 16 h on Glc/MA or Gly/MeOH/AS. As shown in Figure 5A, CCP is induced 3-fold in Gly/MeOH/AS-grown cat cells relative to the WT control. During growth on Glc/MA CCP activities were similar in both strains (Fig. 5A).
To check whether induction of CCP depends on Yap1, we also measured CCP activities in Gly/MeOH/AS grown yap1 and cat yap1 cells, using WT and cat cells as controls. Deletion of YAP1 alone did not affect activity of this enzyme relative to WT. CCP activity was enhanced in cat cells but not in cat yap1 cells relative to the WT control (Fig. 5B).
SOD isozyme profiling of stationary WT cells grown on Glc/AS medium revealed the presence of 3 bands (Fig.  5C). The bands of lowest (designated SOD1) and highest (designated SOD3) relative mobility were inactivated by H 2 O 2 and KCN and therefore represent Cu/Zn containing SOD enzymes. The middle band (designated SOD2) is most likely a Mn containing SOD, as it appeared to be resistant to both compounds.
Comparison of isozyme profiles of WT and cat cells grown on Glc/MA or Gly/MeOH/AS, revealed that cat cells grown on Gly/MeOH/AS showed the highest SOD1 activity (Fig. 5D), which was inhibited by H 2 O 2 and KCN (Fig. 5E).
Because cat yap1 cells very rapidly die on Gly/MeOH/AS media (Fig. 4F), SOD activities were determined in these cultures obtained 4 hours after the shift from Glc/AS to Gly/MeOH/AS. At this time point the activity of SOD1 had not yet increased in cat cells.
In cat yap1 cells however SOD levels were increased (Fig. 5F), suggesting that SOD1 expression is not controlled by Yap1.
Enzyme activities of GLR were similar in WT and cat cells grown on Glc/MA or Gly/MeOH/AS (Fig 5E).
Together our data indicate that in cat cells grown on Gly/MeOH/AS, but not on Glc/MA, genes are expressed that enhance stress resistance and extend the CLS. One of these genes encodes CCP and is induced in a Yap1 dependent way, another one is SOD1, which is not under control of Yap1.

CCP plays a role in lifespan extension of cat cells grown on Gly/MeOH/AS
The increase in lifespan of cat cells on Gly/MeOH/AS depends on Yap1 and is paralleled by an increase in CCP activity, which also depends on Yap1. To test whether indeed enhanced CCP levels are essential for lifespan extension, CLS experiments were performed with a ccp deletion strain and a cat ccp double deletion strain. As shown in Figure 5H, CCP is essential for survival of cat cells in Gly/MeOH/AS medium. Deletion of CCP alone did not have a significant effect on CLS (Fig. 5H). These data suggest that in cat cells in Gly/MeOH/AS medium Yap1 mediated induction of CCP, is essential for lifespan extension. www.impactaging.com

AMO activity is partially reduced in Glc/MA grown cat cells
We recently showed that the presence of MA instead of AS as sole nitrogen source in H. polymorpha cultures extends the CLS, because MA serves as extra energy source in stationary phase cultures [37]. Since the lifespan of cat cells in Glc/MA is reduced relative to the WT control, but similar to that of Glc/AS cultures (Fig.  1AC) we investigated whether this is related to the absence of amine oxidase (AMO) activities, which would prevent the generation of energy from MA www.impactaging.com during chronological aging. As shown in Figure 6, AMO activity was still present albeit reduced by 53% in cat cells in comparison to WT in stationary cultures (Fig. 6A). However, the residual AMO activity most likely still allows to metabolize MA during the stationary phase. Western blot analysis revealed that AMO protein levels were only slightly reduced, indicating that part of the AMO enzyme is inactivated in cat cells (Fig. 6BC).

The activation of stress responsive genes does not alleviate the negative effect of MA on the CLS of cat cells
Our data suggested that upon growth on Gly/MeOH/AS, but not on Glc/MA, resistance genes are induced in cat cells. To test whether the induction of these genes can overcome the negative effect of MA on cell survival, we compared the CLS of WT and cat cells grown on Gly/MeOH supplemented with MA as sole nitrogen source. In agreement with our previous data addition of MA extended the lifespan of WT cells, however cat cultures showed a strongly reduced mean and maximum lifespan relative to the WT control (Fig. 6D)

DISCUSSION
In yeast two types of aging have been defined, namely chronological and replicative aging. The chronological lifespan is the time cells can survive in a non-dividing state, whereas the replicative lifespan represent the number of buds a mother cell produces before it dies. These two types of aging are regulated by partially overlapping regulatory mechanisms. Several recent studies focused on the role of acidification, which accelerates both modes of aging in yeast as well as in mammalian cells [38,39]. Here we studied the role of another important factor in aging, namely reactive oxygen species (ROS).
We have analyzed the role of catalase, a major peroxisomal antioxidant enzyme, on the chronological lifespan of H. polymorpha. Previously, several reports appeared on the function of this enzyme in chronological aging of S. cerevisiae. In one of them, deletion of CTA1, the gene encoding peroxisomal catalase, was shown to lead to shorter lifespan of S. cerevisiae upon growth on 2% glucose. [22], whereas in another study this increased the CLS [23]. A complicating factor in the analysis of the role of peroxisomal catalase in chronological aging of S. cerevisiae is that baker's yeast also contains a cytosolic isoenzyme, Ctt1. By contrast, the H. polymorpha has only one catalase gene, which encodes a peroxisomal protein, like in man.
Our data indicate that in H. polymorpha the effect of catalase deficiency on CLS varies with the cultivation conditions and ranges from a negative effect (Glc/MA) via no effect (Glc/AS and Gly/AS) to a positive effect (Gly/MeOH/AS). An important difference between MA and MeOH utilization is the amount of H 2 O 2 that is generated during the growth phase. This is higher for MeOH, which is used as carbon source, relative to MA, which serves as nitrogen source (C/N ratio of H. polymorpha cells = 7). www.impactaging.com During growth on Glc/AS and Gly/AS no peroxisomal enzymes are involved in the metabolism of the carbon and nitrogen source. Hence, no peroxisomal hydrogen peroxide is produced. This may explain why deletion of CAT has no effect on the CLS upon growth on these substrates (Fig. 1AB). Also in these media the ROS levels were similar in WT and cat cells, both in the exponential growth phase ( Fig. 2A) and during chronological aging (Fig. 2BC).
We recently showed that growth of WT H. polymorpha cells on Glc/MA extends the CLS relative to Glc/AS because of the generation of NADH from formaldehyde, the MA oxidation product, during the stationary phase [37]. However, we now observed that the CLS of the cat cells grown on Glc/MA is reduced relative to the WT control. MA metabolism during the stationary phase results in H 2 O 2 production. In the absence of catalase H 2 O 2 will be decomposed by processes that require reducing equivalents, e.g. via CCP. Hence, it is likely that the net NADH gained by MA metabolism is lost due to NADH requiring H 2 O 2 degradation in cat cells resulting in median and maximal lifespans similar to that observed for Glc/AS grown cat cells.
The extended CLS of the Gly/MeOH/AS-grown cat strain was related to the presence of MeOH in the media during the growth phase, because the CLS of cat and WT cells grown on Gly/AS was identical. A positive effect of NADH generation from formaldehyde generated from methanol oxidation during chronological aging can be ruled out as methanol was consumed at the initiation of the CLS experiment (Fig. S1E).
Our data suggest that enhanced ROS levels during the exponential growth phase on Gly/MeOH/AS ( Fig. 2A) is beneficial and trigger the induction of stress response genes [23]. In line with this, Gly/MeOH/AS grown cat cells showed enhanced resistance towards H 2 O 2 and acrolein relative to Gly/AS grown cat cells. Indeed enzyme activity measurements revealed the induction of CCP and SOD. The enhanced resistance against acrolein may be explained by inducting of old yellow enzyme (OYE) [40].
Yap1 has been described as important transcription factor in regulating stress responsive genes in different yeast species and Yap1 binding sites have been identified in promoter regions of genes involved in ROS scavenging. Using the Yeastract tool [41] we analyzed promoter regions of candidate H. polymorpha genes for putative Yap1 binding sites [34,42,43] (Table 2). These motifs were indeed identified in H. polymorpha genes encoding CCP, SOD and OYE (HYE1, HYE3), but not in Cu/Zn SOD (Table 2). In line with this is the observation that CCP is not up-regulated in a cat yap1 double deletion strain (Fig. 5B), confirming the role of Yap1 in CCP induction. Also, acrolein resistance was impaired in the double deletion strain (Fig. 4B), pointing to a role for Yap1 in OYE induction. As expected SOD induction was not affected by YAP1 deletion (Fig. 5F).
The important role of CCP in CLS extension of cat cells was demonstrated by the observation that cat ccp cells, but not ccp cells, have a very short CLS on Gly/MeOH/AS (Fig. 5G).
Summarizing, our findings indicate that upon growth of cat cells on Gly/MeOH/AS, but not on Glu/MA, several stress responsive genes are induced, among others CCP, SOD and OYE, of which CCP and OYE induction is Yap1 dependent. Yap1 activation is most likely caused by the high ROS levels during the exponential growth on Gly/MeOH/AS, which is not observed in Glu/MA cells. Consistent with this is the observed migration of Yap1 to the nucleus upon transfer of glucose-grown cat cells to these media.
Finally, we show for the first time the important role of CCP in CLS extension as no CLS extension was observed in cat ccp1 cells grown on Gly/MeOH/AS. This makes CCP another important player in yeast CLS next to SOD.
Taken together, we have shown that peroxisomal catalase is important in regulating lifespan. However, the actual effect strongly depends on the carbon and nitrogen sources used for growth.

METHODS
Strains and growth conditions. The H. polymorpha strains used in this study are listed in Table 3. Cells were grown on mineral medium (MM) [44] supplemented with different carbon sources: 0.25% glucose (Glc), 0.05% glycerol (Gly) or a mixture of 0.05% glycerol and 0.5% methanol (Gly/MeOH), and nitrogen sources: 0.25% ammonium sulfate (AS) or 0.25% methylamine(MA). 6 mM K 2 SO 4 was added when methylamine was used as nitrogen source. When required, leucine was added to a final concentration of 60 μg/ml. Biochemical methods. Preparation of cell extracts [46], determination of methanol concentrations [47], enzyme assays for AMO [48], superoxide dismutase [49,50] and cytochrome c peroxidase [51] were performed as described earlier. Glutathione reductase activity was determined using the glutathione reductase assay kit (Sigma-Aldrich). Protein samples for SDS-PAGE gels were prepared as described before [52] and separated on 10% polyacrylamide gels. Proteins were transferred to nitrocellulose membranes using the semi-dry blotting method and probed with specific polyclonal rabbit anti-AMO or anti-pyruvate carboxylase (Pyc1; loading control) antibodies. Blots were quantified using ImageJ.
ROS measurements. ROS accumulation was measured using dihydrorhodamine 123 (DHR, Invitrogen). 10 7 cells were harvested and stained in 50 mM potassium phosphate buffer pH=7.0 containing 20 µg/ml DHR for 30 minutes at room temperature in the dark. Cells were washed once and resuspended in the same buffer. Cells treated in same way without dye were used as controls for fluorescence background. Fluorescence signal of individual cells was captured in a FACS Aria II Cell sorter (BD Biosciences) for 10.000 events at the speed of 500-1000 events per second using a 488nm laser, 505nm long pass mirror and 525/50nm band-pass filter. FACSDiva software version 6.1.2 was used for data acquisition and analysis. The presented data represent differences in mean fluorescence between stained cells and the background.
Fluorescence microscopy. Fluorescence microscopy images were captured using a Zeiss Axioskop 50 with a 100x 1.30 NA Plan Neofluar objective using MetaVue software and a digital camera (model 1300Y; Princeton Instruments). GFP signal was visualized with a 470⁄40 nm bandpass excitation filter, a 495 nm dichromatic mirror, and a 525⁄50-nm bandpass emission filter. ImageJ and Adobe Photoshop CS2 were used for image analysis and figure preparation. In overlay figures bright field images were false colored in blue to mark cell edges.