3.1. Mitochondrial Morphology Changes in Response to Different Glucose Concentrations
In order to determine whether increasing glucose concentration also has an effect on mitochondrial structure and how this correlates with cellular lifespan in S. cerevisiae
, we monitored the changes in mitochondrial morphology in S. cerevisiae
cells growing in calorie-restricted (0.5%) and high glucose conditions (2% and 4%). S. cerevisiae
cells were transformed with an aconitase-GFP fusion construct, ACO1-
GFP, and expression of GFP was used to visualize mitochondrial structure. The use of this construct has been verified in [29
To ensure that mitochondrial morphology was examined at a similar growth phase, the growth of wild-type cells expressing ACO1-GFP in the different levels of glucose was monitored (). A similar growth rate was observed for all glucose conditions and cells reached stationary phase at a comparable time. The final yields of these three levels of glucose culture were also similar.
Figure 1 Growth of BY4743 overexpressing ACO1-GFP plasmid in SC medium supplemented with different concentrations of glucose. An overnight culture of cells in SC medium was diluted to an initial absorbance at 600nm of 0.1 in fresh SC medium containing (more ...)
Having determined the growth states of the cultures in the three glucose conditions, changes of mitochondrial morphology were examined. Cells were grown in synthetic medium (SC) containing 0.5%, 2%, or 4% glucose, and after 17
h growth, as cells entered the diauxic shift, mitochondrial morphology was examined using fluorescence microscopy. Remarkable differences in mitochondrial morphology were observed in response to changing glucose concentrations (). Under the standard laboratory condition with 2% glucose as a carbon source, mitochondria appeared as elongated tubular structures. However, in media containing 0.5% glucose, mitochondria displayed a highly branched, short-rod morphology similar to that observed in cells growing by respiration in ethanol medium [3
]. In the highest level of glucose tested (4%), mitochondria displayed a partial bead-thread structure with very few connections and branches. Observations using a CIT1-
DsRed construct instead of the ACO1-
GFP construct also produced the same result, indicating that the effect of glucose concentrations on mitochondrial morphology was independent of the use of the aconitase-GFP fusion (data not shown).
Figure 2 Mitochondrial morphologies of S. cerevisiae grown in different concentrations of glucose. (a) BY4743 wild-type cells transformed with an ACO1-GFP fusion construct were grown for 72 hours in three different concentrations of glucose (0.5%, 2%, and 4%), (more ...)
The difference in mitochondrial morphologies between cells grown in 2% glucose and 4% glucose was independent of osmotic stress, since addition of an equivalent molar concentration of sorbitol to 2% glucose medium did not affect the mitochondrial appearance (data not shown).
Having observed the characteristic mitochondrial morphology associated with glucose levels, we monitored the change of mitochondrial morphology in cells grown in 0.5%, 2% or 4% glucose media for 24, 48 and 72 hours (). Furthermore, to assess the structural changes of mitochondria, the percentage of cells in the population displaying total mitochondrial fragmentation, in which only punctate mitochondria with complete absence of tubular mitochondria within an individual cell, was determined ().
The percentage of wild-type and mutant cells showing completely fragmented mitochondrial morphology in cultures incubated in three different concentrations of glucose (0.5%, 2%, and 4%) at the times indicated.
Cells grown in 2% or 4% glucose displayed an increased heterogeneity in mitochondrial morphology with time, showing a progression towards punctate fragmented structures over 72 hours (). After 24 hours of growth, the culture grown in 4% glucose had the highest number of cells with totally fragmented mitochondria (10%) followed by those grown in 2% glucose (7%) and 0.5% glucose (2%) (). The percentages of cells with totally fragmented mitochondria grown in 4% and 2% glucose increased to 64% and 66%, respectively, after 72 hours. However, cells under caloric restriction showed an average of less than 3% of the population with total mitochondrial fragmentation at that time.
We monitored respiratory rate under the above conditions to determine whether this affected the morphology of mitochondria. Maximal respiratory activity was observed in cells after 12 hours of growth in 0.5% glucose medium (). This respiration peak coincided with the presence of highly branched mitochondrial morphology observed in 0.5% glucose-grown cells. However, respiratory activity in these cells decreased from 24 hours to a low level at 72 hours, yet the highly branched mitochondrial morphology was maintained throughout the 72 hours time course. Therefore, a high rate of respiration was not required to maintain the highly branched mitochondrial morphology in these cells. On the other hand, the respiratory activity of 4% glucose-grown cells was relatively low and underwent a gradual decrease throughout the 72
h incubation. Since total mitochondrial fragmentation was observed in 4% glucose-grown cells as early as 24 hours of growth, decreasing the respiratory activity could not be the cause of the onset of mitochondrial fragmentation in the presence of high glucose levels.
Glucose concentrations were measured in the supernatant collected from the different media at intervals throughout growth. The level of glucose was close to zero after 24 hours of growth in medium originally supplemented with 0.5% and 2% glucose, while cells consumed approximately half of the glucose in 4% glucose medium (). These data indicate that there was no correlation between the concentration of glucose remaining in the medium and the progression of mitochondrial fragmentation.
Figure 3 Glucose consumption of S. cerevisiae growing in 0.5%, 2%, and 4% glucose. Glucose in the culture medium was measured at intervals after incubation of the cells in SC medium containing glucose at 0.5% (closed squares), 2% (closed circles), or 4% (open (more ...)
Together, the above data showed that neither growth state nor respiratory rate, and the rate of glucose consumption correlated with mitochondrial fragmentation observed in high glucose concentrations. We then further investigate the cause of early mitochondrial fragmentation in cells grown at high glucose concentrations (2% and 4%) by analyzing the mitochondrial morphology of cells lacking genes involved in maintaining mitochondrial morphology.
3.2. Progression of Mitochondrial Fragmentation in High Glucose Is Independent of Mitochondrial Fission
Mitochondrial morphology is modulated by a delicate balance between mitochondrial fission and fusion, and in S. cerevisiae,
deletion of the DNM1
gene involved in mitochondrial fission increases replicative lifespan [4
]. We therefore determined whether the mitochondrial fragmentation observed in cells grown in a high level of glucose was regulated by factors affecting mitochondrial fission by examining mutant strains (dnm1
Δ and fis1
Δ) with known defects in the fission process. The fis1
Δ strain used in these experiments has been shown to also carry a mutation in the WHI2
gene which rescues the mitochondrial respiratory defect caused by FIS1 deficiency, which also causes a failure to suppress cell growth during amino acid deprivation [30
]. The mutant cells were transformed with the ACO1-
GFP construct and grown in 2% or 4% glucose medium under the same condition described above.
Δ mutant defective in mitochondrial fission was expected to show a reduced level of mitochondrial fragmentation [4
]; however, when grown in 4% glucose, it displayed fragmentation comparable to that of the wild type (, see also Supplementary Figure S1 available online at http://dx.doi.org/10.1155/2013/636287
). A slight reduction in the percentage of dnm1
Δ cells that harbored fragmented mitochondria was observed when cells were grown in 2% glucose. Nevertheless, mitochondrial fragmentation progressed in the dnm1
Δ strain under the high glucose conditions. Cells lacking FIS1
also showed a reduction in the percentage of cells containing mitochondrial fragmentation when grown in 2% glucose. However, similarly to dnm1
Δ, mitochondrial fragmentation was observed when fis1
Δ cells were grown in 4% glucose, resulting in 63% of cells containing completely fragmented mitochondria. These results indicated that mitochondrial fragmentation was unavoidable when cells were grown in 4% glucose, even in cells defective in mitochondrial fission. Hence, mitochondrial fragmentation observed in high glucose levels was independent of mitochondrial fission.
Cells deleted for the mitochondrial fusion gene FZO1 lack mitochondrial DNA and had severely deformed mitochondria in both glucose concentrations examined, and it was therefore difficult to determine whether there was any involvement of mitochondrial fusion in the fragmentation of mitochondria using this mutant.
3.3. Inhibition of TOR Signaling Pathway Reduces Mitochondrial Fragmentation
Since nutrient availability might play a greater role than mitochondrial fission processes in modulating mitochondrial fragmentation when cells were grown at high glucose concentration, we examined mutant strains lacking genes involved in glucose sensing (SNF3, RGT2), glucose metabolism (HXK2, GPA2, PDE1, and PDE2), and general nutrient sensing (TOR1). Mutant cells transformed with the ACO1-GFP construct were grown in 2% or 4% glucose medium as described above.
Mutants with a deletion affecting glucose sensing or glucose metabolism showed 50% to 84% of cells with totally fragmented mitochondria morphology after 72 hours of growth in either 2% or 4% glucose (). Among the mutants screened, only cells lacking the TOR1
gene showed a substantial reduction in the percentage of cells with totally fragmented mitochondria when grown in 2% or 4% glucose (; Supplementary Figure S2). Mitochondrial fragmentation of tor1
Δ mutant cells was 42% and 37% when cells were grown in 4% glucose and 2% glucose, respectively, after 72 hours. As an alternative approach to genetic manipulation of the TOR pathway, the wild-type cells were treated with 10
nM rapamycin to inhibit both TOR1
gene products. Cells treated with rapamycin showed an even greater reduction in total mitochondrial fragmentation than in the tor1
Δ strain, with only 12% to 15% of the cells showing totally fragmented mitochondria when the cells were grown for 72 hours in 2% or 4% glucose media containing 10
nM rapamycin, respectively (; Supplementary Figure S3). Hence, deletion of TOR1
only partially suppressed mitochondrial fragmentation while inhibition of the TOR pathway by rapamycin, which also inhibits TOR2,
further repressed the extent of mitochondrial fragmentation during cell growth in high glucose levels.
3.4. Autophagy Is Required to Resist Mitochondrial Fragmentation Caused by Volatile Glucose Metabolites
Since mitochondrial fragmentation occurred after cells had grown in media, we tested whether cells grown in different concentrations of glucose excreted metabolites are capable of stimulating mitochondrial fragmentation. In order to test this hypothesis, conditioned medium (in which cells had been grown in 0.5% or 4% glucose for either 24 hours or 48 hours) was collected and then used to replace the growth medium of cells grown either to exponential (6 hours) or stationary phase (48 hours).
The conditioned medium that was initially supplemented with 4% glucose (4% glucose, 48 hours) contained substances that caused mitochondrial fragmentation in exponential phase cells, regardless of the glucose concentration of the medium in which the cells were pregrown (). Mitochondrial fragmentation was observed as early as 2
h after transfer into this medium. Fragmentation also occurred for stationary phase cells pregrown in medium containing 2% and 4% glucose. In contrast, stationary phase cells pregrown in 0.5% glucose were resistant to mitochondrial fragmentation induced by the same medium. It was hypothesized that mitochondrial fragmentation was prevented in these cells because nutrients became depleted, and autophagy was activated earlier than in the other growth regimes.
Figure 4 Conditioned medium from S. cerevisiae grown in 4% glucose triggered mitochondrial fragmentation, which was delayed by autophagy. (a) The wild-type cells pregrown to exponential phase for 6 hours in 0.5%, 2%, and 4% glucose to exponential phase were transferred (more ...)
In order to investigate the involvement of autophagy in resistance to conditioned medium-induced mitochondrial fragmentation, the autophagy mutant strains Δuth1, Δatg1, and Δatg5 were grown to stationary phase in medium containing 0.5% glucose and then transferred into the conditioned medium (4% glucose, 48 hours). The mitochondria of the Δatg1 and Δatg5 mutants became fragmented, but not those of the mitophagy mutant Δuth1. These results indicated that general autophagy was important for conferring resistance to the metabolites that stimulated mitochondrial fragmentation and that starvation may be able to delay mitochondrial fragmentation. Indeed, delayed fragmentation was observed in the cells growing in 10-fold diluted SC medium containing 2% glucose compared to the cells growing in normal SC medium with 2% glucose (data not shown).
Since conditioned medium (4% glucose, 48 hours) appeared to contain a substance that stimulated fragmentation of mitochondria, it was analysed further. Treatment with diluted spent medium did not cause mitochondrial fragmentation in S. cerevisiae pregrown in any of the glucose concentrations (), indicating that the effect was probably not due to the physical disturbance of changing the medium but due to the concentration of the glucose metabolites present. These cells maintained tubular mitochondrial structure for at least 62 hours after the medium was exchanged. In addition, vacuum evaporation of the conditioned medium rendered it unable to stimulate mitochondrial fragmentation () indicating that the stimulatory substance/s were volatile. Interestingly, mitochondrial fragmentation stimulated by the conditioned medium was found to be reversible once the medium was removed ().
Figure 5 Reversibility of conditioned medium-induced mitochondrial fragmentation. (a) The mitochondria of the wild-type cells grown in 2% glucose did not fragment when they were transferred into diluted (4% 48h dilu) and evaporated (4% 48h evap) (more ...)
3.5. The Observed Mitochondrial Fragmentation Was Not due to Intracellular Acidification
Since the metabolite(s) responsible for mitochondrial fragmentation was(were) volatile, we analysed all of the 48 hours conditioned media (0.5% glucose, 2% glucose, and 4% glucose) by gas chromatography-mass spectrometry. Three volatile substances with higher concentrations in the 4% glucose-conditioned medium were detected: acetic acid, ethanol, and 2,3-butanediol (Supplementary Table S1). Of the three compounds, acetic acid was the only one that resulted in mitochondrial fragmentation when separately added to the cells (Supplementary Figure S4).
Mitochondrial fragmentation triggered by acetic acid treatment could be due to intracellular acidification caused by release of protons or to accumulation of acetate. In addition to acetic acid, benzoic acid and 2,4-dinitrophenol (2,4-DNP) also triggered mitochondrial fragmentation (Figure S1). One feature that is common to these three compounds is their ability to lead to acidification within the cells, and therefore we analyzed the intracellular concentration of acetate and intracellular pH of cells grown in different glucose concentrations. The intracellular level of acetate was higher in the wild-type cells growing in 2% and 4% glucose than those growing in 0.5% (See Supplementary Table S2). The intracellular pH of the cells grown in different concentrations of glucose was measured using the pH-sensitive GPF probe pHluorin. No significant correlation between intracellular pH and mitochondrial fragmentation was found (Supplementary Figure S5). Although mitochondrial fragmentation was already established in 2% and 4% glucose-grown cells within 24 hours of inoculation (), the intracellular pH of these cells was similar to that of 0.5% glucose-grown cells at that time. These results indicated that intracellular acidification was unlikely to be responsible for triggering mitochondrial fragmentation and that acetate or some metabolite derived from it is more likely to be responsible.
3.6. Autophagy Is Required to Reduce Mitochondrial Fragmentation
One of the many cellular processes regulated by the TOR pathway is autophagy, which recycles damaged proteins and organelles and makes amino acids and other essential metabolites to the cell [31
] available. To determine whether autophagy plays a role in the reduction of mitochondrial fragmentation under high glucose conditions, a mutant strain defective for initiation of autophagy (atg1
Δ) was transformed with ACO1-
GFP construct to examine mitochondrial fragmentation ().
Cells deleted for ATG1
displayed higher percentages (approximately 75% after 72 hours incubation) of mitochondrial fragmentation than the wild type, indicating that autophagy acts to reduce the onset of mitochondrial fragmentation in 2% and 4% glucose-grown cells. Since autophagy was important in maintaining mitochondrial morphology under these conditions, cells lacking genes affecting mitochondrial-specific autophagy, ATG32
], and UTH1
] that is also affected in cellular ageing [35
], were analyzed. Surprisingly, deletion of UTH1
did not affect mitochondrial fragmentation compared to that in the wild-type cells, indicating that mitochondrial-specific autophagy alone did not substantially suppress mitochondrial fragmentation. However, general autophagy, involving ATG1
appears to play a vitally important role for reducing mitochondrial fragmentation under higher glucose conditions since deletion of ATG1
elevated the fragmentation of mitochondria seen in cells grown on higher glucose levels.
Subsequently, we checked whether the TOR pathway regulated the function of autophagy in reducing mitochondrial fragmentation. The autophagy mutants were treated with rapamycin, and total mitochondrial fragmentation was examined. A reduction of mitochondrial fragmentation in rapamycin-treated uth1Δ was observed (), which was consistent with the finding that deletion of UTH1 did not have an impact on mitochondrial fragmentation and that the suppression effect of rapamycin observed was independent of UTH1. In the atg1Δ mutant, although treatment with rapamycin reduced mitochondrial fragmentation compared to the untreated mutant cells, the level of mitochondrial fragmentation remained much higher than in rapamycin-treated wild-type cells. Hence, rapamycin inhibition of the TOR pathway led to suppression of mitochondrial fragmentation, but this was largely dependent on the presence of a functional autophagy pathway. Therefore, it is likely that autophagy functions downstream of the TOR pathway in maintaining mitochondria in a nonfragmented state.
3.7. Role of Autophagy in Mitochondrial Fragmentation Induced by Glucose Metabolites
Having identified a cellular process that is able to prevent mitochondrial fragmentation in cells grown in high glucose concentrations, we next examined what triggered mitochondrial fragmentation in these cells. Mitochondria are the major site of reactive oxygen species (ROS) production, and an elevation of ROS could be one of the causes of mitochondrial fragmentation. We examined the levels of superoxide anion by DHE staining of cells growing in 0.5%, 2%, and 4% glucose over a 72
h time course and flow cytometry analysis to determine whether elevation in superoxide levels was correlated with the occurrence of mitochondrial fragmentation.
Cellular superoxide levels increased over time regardless of the concentration of glucose, as shown in . Cells grown in 0.5% glucose had the highest superoxide level after 24 hours growth, which is consistent with the fact that respiratory activity was the highest for these cells at that time (). After 72
h incubation, cells grown in 4% glucose had the highest level of superoxide followed by those grown in 2% glucose and then those grown in 0.5% glucose. It is therefore unlikely that an increase in ROS level triggered mitochondrial fragmentation during cell growth, since the onset of elevated levels of ROS in 0.5% glucose-grown cells did not lead to mitochondrial fragmentation.
Since fragmentation occurred 24 hours after inoculation in media, we analyzed whether the glucose metabolites accumulated in the medium during growth stimulated mitochondrial fragmentation. To test this hypothesis, conditioned medium (in which cells were grown for either 24 hours or 48 hours) originating from 4% glucose or 0.5% glucose medium was collected and then was used to replace the growth medium of cells grown to either exponential (6 hours) or stationary phase (48 hours).
The 48 hours conditioned medium that was initially supplemented with 4% glucose (4% conditioned medium) contained substances that caused mitochondrial fragmentation in wild-type cells in exponential phase, regardless of the glucose concentration of the medium in which cells were pregrown (). For instance, mitochondrial fragmentation was observed as soon as two hours after transferring cells into the 4% conditioned medium. This fragmentation was also found for stationary phase wild-type cells pregrown in medium containing 2% or 4% glucose (). In contrast, the stationary phase cells pregrown in 0.5% glucose did not display fragmented mitochondria after transfer into the 4% conditioned medium (). We hypothesized that the early nutrient depletion in 0.5% glucose-grown cells activated autophagy leading to resistance to induction of mitochondrial fragmentation.
In order to investigate the involvement of autophagy in this mitochondrial fragmentation process, the atg1Δ and uth1Δ autophagy mutants were grown for 48 hours in 0.5%, 2%, and 4% glucose and then transferred into the conditioned media (). Mitochondrial fragmentation was observed in the atg1Δ mutant cells, including those pregrown in 0.5% glucose, after transfer into the 4% glucose conditioned medium. On the other hand, uth1Δ cells pregrown in 0.5% glucose medium were partially resistant to 4% conditioned medium-induced mitochondrial fragmentation (approximately 50% of the total population displayed tubular mitochondria). These results indicated that activation of general autophagy during starvation played an important role in conferring resistance to those metabolites present in the conditioned medium that stimulated mitochondrial fragmentation.
Conversely, wild-type cells transferred into 0.5% conditioned medium displayed tubular mitochondria independent of the growth phase and the level of glucose in which the cells were pregrown (). The fragmented mitochondria in the atg1Δ, and uth1Δ mutants also returned to a tubular structure after cells were transferred into 0.5% conditioned medium, although these cells required a longer time for recovery.
The effects seen using 4% conditioned medium to stimulate fragmentation of mitochondria were not due to the physical disturbance of changing the medium. When conditioned medium was removed and fresh medium was supplemented to cells, there was no fragmentation in cells pregrown in any of the glucose concentrations used. Interestingly, mitochondrial fragmentation stimulated by the 4% conditioned medium was found to be reversible once the medium was replaced by the 0.5% conditioned medium (). The reversible nature of the process indicated that the cells were not yet committed to any deleterious effects that may result from mitochondrial fragmentation.
3.8. Mitochondrial Fragmentation and Chronological Lifespan
The above results demonstrated that S. cerevisiae
cells grown in high glucose concentrations not only possessed fragmented mitochondria but also showed higher levels of oxidative stress than those grown in calorie-restricted conditions. It is well known that S. cerevisiae
cells that are restricted in their calorie intake have longer chronological and replicative lifespans [36
], that maintenance of the morphology of mitochondria is important for cell survival since the mutants that preserve tubular mitochondrial structure (such as Δdnm1
) live longer than the wild-type cells [4
], and the mutants that progress early to mitochondrial fragmentation have shorter lifespan [38
]. This led us to investigate whether reversing fragmentation of mitochondria of cells grown in 2% and 4% glucose would extend their lifespan.
Since mitochondrial fragmentation in 4% or 2% glucose-grown cells could be reversed in 0.5% conditioned medium () and vice versa (), we determined whether chronological lifespan (CLS) could also be reversed in the same way. Cells were grown in 0.5%, 2%, or 4% glucose for 48 hours and then transferred into conditioned media as shown in , and their CLS were assessed.
As expected, cells grown in 0.5% glucose had an extended CLS compared to those grown in higher glucose concentrations (). Interestingly, their lifespan was shortened when these cells were transferred into 4% conditioned medium. On the other hand, the lifespan of cells grown in 4% glucose medium was extended following their transfer into 0.5% conditioned medium. Burtner et al. [39
] have also shown that the CLS was reversible by substituting spent growth medium in a similar way. Here, we show that the CLS of S. cerevisiae
varied depending on the type of medium into which cells were exchanged and that this correlated with the reversible changes in mitochondrial morphology.
Figure 7 The shortened chronological lifespan of S. cerevisiae cells grown in higher concentrations of glucose can be reversed without genetic manipulation. Wild-type cells were grown in media containing 0.5%, 2%, and 4% glucose for 48 hours and exchanged into (more ...)