Yeast mitochondrial adaptation to freeze thaw stress depends on pre freeze fermentation state

Abstract

During frozen dough production, transportation, and processing, temperature fluctuations expose yeast to freeze-thaw (FT) stress, causing yeast cellular damage. This study investigated the effects of FT cycles on mitochondrial structure and function in yeast under different fermentation processes. As FT cycles increased, cell mortality, mitochondrial membrane potential, and superoxide anion (O₂^.-) generation rate significantly increased, with the highest values observed in the FT6 group. Cytochrome c content and cytochrome c oxidase activity varied among groups, while isocitrate dehydrogenase activity was highest in FT-0. Acetyl-CoA synthetase activity increased with FT cycles, and α-ketoglutarate dehydrogenase activity was higher in FT6 than in FT3. Gene expression of ACS1 and KGD increased with cycles, whereas YNO1, AGT32, and MDM34 were lower in FT3 than in FT0 and FT6. Compared with the control, expression of IDH2, CYC7, NDI1, AIF1, DNM1, AGT11, AGT8, and MDM12 was upregulated under FT stress. Collectively, FT stress induces yeast cell death and impairs mitochondrial function by modulating key enzyme activities, promoting O₂^.- production, altering membrane potential, and upregulating genes involved in mitochondrial division and autophagy.

Introduction

Frozen dough could supply a high-efficiency method to preserve and offer fresh-tasting foods, retaining their nutritional benefits. Nowadays, there are three techniques for making frozen dough on the market: raw frozen dough, pre-fermented frozen dough, and fermented frozen dough¹. While the dough’s quality and stability face challenges when the temperature fluctuates during processing, freezing, storage, and transporting². Yeast in the dough will be affected by freeze-thaw stress, which leads to gas production decreases, moisture migration, and protein structure disruption in the dough, mainly due to the decrease in yeast cell viability and changes in metabolic production³. Mitochondria, as cellular energy factories, are crucial for the metabolic homeostasis and viability of eukaryotic organisms⁴. Despite significant advances in freeze-thaw protection technology in recent years including manipulating intracellular metabolites⁵, altering membrane composition⁶, managing antioxidant defenses⁷, regulating aquaporin expression, and employing adaptive evolutionary and breeding techniques⁸, the effects of freeze-thaw cycles under different production process conditions on mitochondrial function, especially the activities of citrate synthase, isocitrate dehydrogenase, and α-ketoglutarate dehydrogenase and the dynamic balance of the mitochondrial metabolic network by adjusting the related genes expression have not been fully explored.

Important protein complexes such as ATP synthase and the electron transport chain in the inner mitochondrial membrane might be damaged or inactivated by environmental stress, affecting ATP synthesis and the efficiency of electron transport^9,10. The tricarboxylic acid cycle is the main pathway for yeast cells to synthesize ATP. The tricarboxylic acid cycle depends on the supply of the precursor substance acetyl coenzyme A (acetyl-CoA). Acetyl coenzyme A synthetase is involved in acetyl coenzyme A synthesis in the cytoplasm, and although it does not act directly on the mitochondrial energy cycle, the acetyl coenzyme A produced is the onset point in the mitochondrial TCA cycle and is regulated by the ACS1 gene¹¹. Acetyl coenzyme A synthetase converts acetic acid to acetyl coenzyme A, which is encoded by the ACS1 gene¹². α-Ketoglutarate dehydrogenase and isocitrate dehydrogenase are directly involved in the tricarboxylic acid cycle (TCA cycle) within the mitochondria, catalyzing the conversion of α-ketoglutarate to succinyl coenzyme A and isocitrate to α-ketoglutarate, respectively, with the production of NADH by both processes. These reactions are the core steps of cellular energy production (ATP), with the energy conversion being achieved through the process of oxidative phosphorylation. They are regulated by the IDH1/2 and KGD1 genes, respectively^13–15. The mitochondrial electron transport chain and tricarboxylic acid cycle, related to ROS production, are closely related¹⁶. In addition, NADPH oxidase encoded by the YNO1 gene catalyzes the movement of electrons from NADPH to FADH2, which results in superoxide production. Yno1 protein Overproduction leads to a burst of reactive oxygen species (ROS), leading to cell damage and even death¹⁷.

Mitochondria not only perform the core metabolic and energy production tasks in yeast cells, but also play a complex and fine regulatory role in responding to injury, maintaining their homeostasis, and participating in cellular life and death decisions. Mitochondria in yeast cells maintain highly dynamic properties, adjusting their morphology, number, and functional status through successive fusion and division processes to adapt to changing cellular demands and environmental stresses. This dynamic equilibrium is precisely regulated by a specific set of proteins: Fis1, Dnm1, and other proteins are responsible for driving mitochondrial division and regulating the number and morphology of mitochondria¹⁸. Yeast cells have a specialized mechanism, mitochondrial autophagy, to identify and remove damaged or degraded mitochondria. This process relies on proteins encoded by the Autophagy-related (ATG) gene family, such as ATG8, ATG11, and ATG32^19–21. Through mitochondrial autophagy, yeast can effectively remove abnormal mitochondria that may produce ROS, disrupt energy metabolism, or cause cytotoxicity, thus maintaining the overall quality of mitochondria, preventing oxidative stress caused by ROS accumulation, and safeguarding the health and survival of the cell. Meanwhile, MMM1, MDM34, MDM11, and MDM10 genes take part in maintaining the mitochondrial intermembrane stability (i.e., the contact site between the inner and outer membranes) and the dynamic equilibrium of the mitochondrial network^22–24. Studying the functions of these genes under the freeze-thaw cycle could help to reveal the response mechanisms of mitochondrial biogenesis, functional maintenance, and dynamic changes in yeast cells.

This study aimed to compare the activities of key mitochondrial enzymes and related genes (ACS1, IDH1/2, KGD1, YNO1) encoding the enzymes and mitochondrial damage and repair-related genes (CYC1, AIF1, CYC7, NDI1, FIS1, DNM1, ATG8, ATG11, ATG32, MMM1, MDM34, MDM11, MDM10) of yeast cells during freeze-thaw cycling, it will help to deepen the understanding of mitochondrial stress response and metabolic regulation mechanism and provide a new perspective to reveal the survival strategy of yeast cells adapting to freeze-thaw stress. This study will reveal the effects of freeze-thaw cycles under different fermentation processes on the activities of key mitochondrial enzymes and gene expression of yeast, and provide valuable information and insights for improving yeast activity and the quality of frozen fermented dough.

Results

Yeast cell viability, mitochondrial integrity, and redox status

As shown in Fig.1A, B, the mortality rate of yeast cells and cytochrome c content exhibited distinct responses to freeze-thaw (FT) cycles, dependent on the fermentation protocol prior to freezing. The mortality rate increased with the number of FT cycles across all groups, with the FT6-50 group displaying the highest level of cell death. With the increase of freeze-thaw cycles, the cellular structure of yeast was more severely damaged, and the changes in mortality rate were basically consistent (Fig. S1). Concurrently, a significant increase in cytochrome c content was observed in most FT-treated groups compared to the control, indicating mitochondrial membrane disruption and release of cytochrome c into the cytoplasm²⁵. Notably, the cytochrome c content in the FT0-50 group was not significantly different from that of the control. In contrast, the FT3-0 group exhibited a pronounced elevation in cytochrome c content at the 3rd FT cycle. Interestingly, by the 6th FT cycle, the FT6-50 group exhibited a lower cytochrome c content compared to the FT3-50 group. This pattern suggests that under prolonged freeze-thaw stress, surviving cells may undergo metabolic reprogramming, including the synthesis of protective substances like glycerol and trehalose, downregulation of mitochondrial respiration, and a consequent reduction in cytochrome c synthesis. Furthermore, the accumulation of ROS over extended stress periods could deplete cytochrome c pools, and the inhibition of mitochondrial biosynthesis by freeze-thaw stress may further contribute to the reduced total cytochrome c content²⁶.

Fig. 1. Effects of key substances and key enzyme activities in mitochondria in yeast cells under freeze–thaw cycles of different fermentation processes.

A Yeast cell mortality (%), B cytochrome c content (mg/mgprot), C mitochondrial membrane potential (%), D cytochrome c oxidase activity (U/mgprot), E O₂^- generation rate (nmol/mg min), F acetyl coenzyme A synthetase activity (U/mgprot), G isocitrate dehydrogenase (U/mgprot), H α-ketoglutarate dehydrogenase (U/mgprot). F0/3/6-50: Fermentation for 50 min, freezing for 24 h, freeze-thawing for 0,3,6 times; F0/3/6-20: First ferment for 20 min, freeze for 24 h, thawed, then fermented for 30 min, freeze–thaw for 0,3,6 times; F0/3/6-0: First frozen for 24 h, thawed, then fermented for 50 min, freeze–thaw 0,3,6 times. Different lowercase letters indicate significant differences (p < 0.05) among groups within the same freeze–thaw cycle.

The average mitochondrial membrane potential (ΔΨm), assessed by rhodamine 123 (RH123) fluorescence, displayed a non-monotonic trend with increasing FT cycles (Fig. 1C). An increase in ΔΨm was observed up to the 3rd cycle, with the FT0-20 group showing a significantly higher signal than both the FT0-50 and FT0-0 groups. It is critical to note that RH123 staining provides a population-weighted average and cannot distinguish between viable cells with functional mitochondria and dead or permeabilized cells that nonspecifically retain the dye²⁷. Given the substantial cell mortality in later cycles (e.g., FT6-50; Fig. 1A), the measured ΔΨm likely represents a composite signal dominated by the surviving subpopulation, particularly in resilient groups like FT0-20 and FT3-20. This interpretation is reinforced by the observation that the FT6-0 group exhibited a significantly lower ΔΨm than the FT6-50 and FT6-20 groups, despite their comparable mortality rates. This indicates that the functional outcome is determined by the metabolic state of the surviving cells, not merely by the survival rate itself. This cellular heterogeneity is consistent with the established understanding that freeze-thaw stress induces divergent fates: a subset of cells activates protective pathways, while others suffer irreversible damage²⁸.

This interpretation is consistent with the observed changes in cytochrome c oxidase (COX) activity (Fig. 1D). COX activity was significantly elevated in the FT0-50 and FT0-20 groups compared to the control. Similarly, at the 3rd and 6th FT cycles, the FT3-20 and FT6-20 groups maintained higher COX activity than their respective FT3-50/0 and FT6-50/0 counterparts. The coordinated trends in membrane potential and COX activity suggest a stress-responsive modulation of mitochondrial function that helps maintain energy homeostasis and mitigate oxidative damage during freeze-thaw cycles^29,30. The elevated COX activity in adapted cells (e.g., FT0-20, FT3-20) likely plays a role in maintaining redox homeostasis and preventing excessive hydrogen peroxide accumulation, as defects in COX assembly are known to lead to increased hydrogen peroxide production and impaired cell proliferation³¹. Mitochondrially produced NADPH may further support this defense against oxidative stress and potentially stabilize COX activity through enhanced TCA cycle enzyme activity⁹.

As presented in Fig. 1E, the O₂^.⁻ production rate was significantly elevated in all FT-treated groups compared to the control, and this rate generally increased with the number of FT cycles. The FT0-50 group exhibited a much higher O₂^.⁻ production rate than the FT0-0 and FT0-20 groups. At the 3rd and 6th FT cycles, the FT3/6-0 groups showed higher O₂^.⁻ production than the FT3/6-50 groups. Despite the overall increase in O₂^.⁻ at the 6th cycle, the FT6-20 group displayed a significantly lower production rate compared to the FT6-50 and FT6-0 groups. Freeze-thaw cycling disrupts intracellular redox balance, leading to increased electron leakage from the electron transport chain and consequently a higher rate of O₂^.⁻ production³². The repeated formation of ice crystals causes physical and mechanical damage to yeast cells, which can trigger intracellular oxidative stress and further elevate O₂^.⁻ levels. This physical damage may compromise cell membrane integrity, destabilizing the intra- and extracellular environment. Since O₂^.⁻ is primarily generated in the mitochondria during electron transfer in the respiratory chain, mitochondrial dysfunction or damage under freeze-thaw stress conditions can result in increased electron leakage and a higher O₂^.⁻ production rate³³. The lower O₂^.⁻ levels in the FT6-20 group, coupled with its higher COX activity, suggest that an intermediate fermentation state (20 min) before freezing better preserves mitochondrial integrity and function, thereby minimizing ROS generation during subsequent stress.

Key enzymes activity in the mitochondria of yeast cells

As shown in Fig. 1F, Acetyl-CoA synthetase (ACS) activity in the FT0-20 group was not significantly different from the control. In contrast, both FT0-50 and FT0-0 exhibited significantly higher ACS activity than the control. At the 3rd FT cycle, ACS activity increased across all groups compared to control, with the FT3-20 group showing the highest activity among the FT3 series. By the 6th FT cycle, ACS activity declined markedly compared to the 3rd cycle across all fermentation conditions. Notably, at both the 3rd and 6th cycles, FT-20 and FT-0 groups consistently displayed higher ACS activity than their FT-50 counterparts. This pattern suggests that the metabolic state at freezing influences the regulation of acetyl-CoA supply, with pre-fermentation freezing associated with elevated ACS activity under repeated stress.

Isocitrate dehydrogenase (ICDH) activity (Fig. 1G) decreased in most FT-treated groups relative to the control. The FT0-0 group showed significantly higher ICDH activity than FT0-20 and FT0-50. At the 3rd FT cycle, no significant differences in ICDH activity were observed among the FT3-50, FT3-20, and FT3-0 groups. However, at the 6th cycle, the FT6-0 group exhibited significantly higher ICDH activity than both FT6-20 and FT6-50. Overall, the data indicate that ICDH activity is suppressed by FT stress, and this suppression is more pronounced in cells frozen after partial or full fermentation (20 min or 50 min) compared to those frozen before fermentation (0 min). Given that ICDH catalyzes the oxidative decarboxylation of isocitrate to α-ketoglutarate—the first NADH-generating step of the TCA cycle—its reduced activity may reflect disruption of early-cycle flux under cryostress³⁴.

α-Ketoglutarate dehydrogenase (α-KGDH) activity (Fig. 1H) was significantly elevated in the FT0-50 group but significantly reduced in FT0-20 and FT0-0 compared to the control. At the 3rd FT cycle, α-KGDH activity in FT3-50 was lower than in FT0-50. Strikingly, at the 6th FT cycle, α-KGDH activity increased compared to the 3rd cycle across all groups, with FT6-20 and FT6-0 showing significantly higher activity than FT6-50. This rebound in α-KGDH activity coincides with the observed decline in ICDH activity (Fig. 1G), suggesting a potential compensatory adjustment in TCA cycle flux. The inverse relationship between ICDH and α-KGDH activities across treatment groups is consistent with metabolic re-routing, where reduced conversion of isocitrate to α-ketoglutarate may be offset by enhanced downstream oxidation of α-ketoglutarate to succinyl-CoA³⁵. This pattern aligns with the earlier observations of preserved mitochondrial membrane potential and COX activity in FT-20 groups (Fig. 1C, D), implying that cells frozen during active glycolysis retain greater capacity for regulated TCA cycle operation under repeated freeze–thaw stress.

Collectively, the enzyme activity profiles indicate that the timing of freezing relative to fermentation critically shapes mitochondrial metabolic responses. Cells frozen at the 20-min fermentation point (FT-20) exhibit a distinct signature: moderate ACS activity, suppressed ICDH, but relatively high α-KGDH and COX activity—suggesting a reconfigured, yet functional, respiratory metabolism that correlates with improved survival and reduced ROS generation^34–36.

Expression of relevant genes in yeast cell mitochondria

The pre-freeze metabolic condition profoundly shapes this transcriptional journey, with the FT-20 groups consistently exhibiting a more resilient profile. FT-0 (frozen without fermentation): these starved cells mount a strong but ultimately stressed response. They show the highest initial ACS1 and YNO1 expression (Fig. 2A, E), reflecting an urgent need for energy and a severe oxidative challenge, after yeast cells were subjected to freezing and freeze-thawing cycles, it possibly enhanced the demand for fatty acids, which activated the ACS1 gene expression³⁷. By FT6, they maintain high ACS1 and KGD1 (Fig. 2A, D) but suffer from a sharp drop in IDH2 (Fig. 2C), aligning with their low mitochondrial membrane potential (ΔΨm, Fig. 1C) and high O₂^.⁻ production (Fig. 1E). FT-50 (frozen after 50-min fermentation): Cells frozen in a high-metabolic state display a blunted adaptive capacity. They have the lowest ACS1 and KGD1 expression at FT6 and show a significant decline in IDH1 (Fig. 2B) from FT3 to FT6. This transcriptional inflexibility correlates with their poor physiological outcomes, including the highest mortality (Fig. 1A). FT-20 (frozen after 20-min fermentation): This group demonstrates the most favorable and balanced adaptation. At FT3, it exhibits the peak IDH2 expression, indicating a robust early TCA cycle engagement. At FT6, it maintains the highest KGD1 levels alongside strong ACS1 expression, a profile that directly supports its superior mitochondrial function, as evidenced by high COX activity (Fig. 1D) and the lowest O₂⁻· production among FT6 groups (Fig. 1E).

Fig. 2. Relative expression of mitochondrial metabolic processes related genes in yeast cells under freeze–thaw cycles of different fermentation processes.

A Relative expression of ACS1, B relative expression of IDH1, C relative expression of IDH2, D relative expression of KGD1, E relative expression of YNO1. F0/3/6-50: Fermentation for 50 min, freezing for 24 h, freeze-thawing for 0,3,6 times; F0/3/6-20: First ferment for 20 min, freeze for 24 h, thawed, then fermented for 30 min, freeze–thaw for 0,3,6 times; F0/3/6-0: First frozen for 24 h, thawed, then fermented for 50 min, freeze–thaw 0,3,6 times. Different lowercase letters indicate significant differences (p < 0.05) among groups within the same freeze–thaw cycle.

The expression of YNO1, a gene involved in cellular redox modulation, provides a direct link between metabolic status and oxidative stress^17,38. Its expression is overwhelmingly dominated by the FT-0 groups, peaking in FT6-0 (Fig. 2E). This strong induction is consistent with the severe oxidative phenotype of these cells—high ROS and low ΔΨm—suggesting that Yno1 is part of a stress-response network activated under conditions of extreme redox imbalance. In conclusion, yeast cells do not respond to repeated freeze-thaw stress with a static defense but rather execute a dynamic, multi-stage metabolic strategy. This strategy evolves from an initial attempt to boost core energy metabolism to a later, more refined program focused on maintaining essential metabolic fluxes and managing oxidative damage. The success of this program is decisively influenced by the cell’s metabolic state at the moment of the first freeze, with a 20-min pre-freeze fermentation providing an optimal balance that enables sustained mitochondrial health and stress resilience.

Respiratory chain-related genes expression in yeast cell mitochondria

The transcriptional regulation of key mitochondrial respiratory chain genes in response to freeze-thaw (FT) stress is highly dependent on the cell’s physiological state at the time of freezing, defined by pre-FT fermentation duration (0, 20, or 50 min). This state dictates a cell’s capacity to maintain energy metabolism and manage stress-induced damage. Genes encoding isoforms of cytochrome c (CYC1 and CYC7), which are integral to Complex III function, exhibited a strong initial upregulation, particularly in cells not frozen in a late-fermentation state. At FT0, CYC1 expression was significantly elevated in both FT0-0 and FT0-20 groups (Fig. 3A), suggesting an immediate compensatory mechanism to bolster electron transfer efficiency in response to cryo-damage³⁹. A similar, though less pronounced, trend was observed for CYC7 across all FT0 groups (Fig. 3B). By the 3rd cycle (FT3), the FT3-20 group displayed the highest CYC1 expression, significantly outperforming FT3-50 and FT3-0. This highlights the advantage of being frozen during early active growth, where cells retain the metabolic flexibility to mount a robust protective response. However, this adaptive capacity waned with chronic stress. At FT6, CYC1 expression was universally downregulated compared to FT3, indicating severe mitochondrial impairment after repeated cycles. Notably, the FT6-20 group maintained the highest relative CYC1 levels, reinforcing its status as the most resilient condition. The stabilization of CYC7 expression from FT3 to FT6 (Fig. 3B) suggests that its role may be more constitutive, providing a baseline level of Complex III function under sustained stress⁴⁰.

Fig. 3. Relative expression of respiratory chain related genes in yeast cells under freeze–thaw cycles of different fermentation processes.

A Relative expression of CYC1, B relative expression of CYC7, C relative expression of NDI1, D relative expression of AIF1. F0/3/6-50: Fermentation for 50 min, freezing for 24 h, freeze-thawing for 0,3,6 times; F0/3/6-20: First ferment for 20 min, freeze for 24 h, thawed, then fermented for 30 min, freeze–thaw for 0,3,6 times; F0/3/6-0: First frozen for 24 h, thawed, then fermented for 50 min, freeze–thaw 0,3,6 times. Different lowercase letters indicate significant differences (p < 0.05) among groups within the same freeze–thaw cycle.

The gene NDI1, encoding the primary NADH dehydrogenase of the yeast alternative respiratory pathway, showed a clear pattern of progressive suppression, reflecting a decline in core mitochondrial bioenergetic capacity. While NDI1 was initially higher than the control in all FT0 groups (Fig. 3C), its expression was consistently lowest in the FT-50 groups, whose cells were frozen in a high-metabolic, vulnerable state. A dramatic and significant downregulation of NDI1 occurred in all groups by the 6th cycle (FT6). This collapse in NDI1 transcription is a hallmark of severe mitochondrial dysfunction, leading to restricted energy metabolism, a disrupted cellular redox balance, and a shift towards a pro-apoptotic state⁴¹. The apoptosis-inducing factor gene AIF1 displayed a distinct regulatory pattern that directly correlated with the risk of cell death. In the initial phase (FT0 and FT3), AIF1 expression was highest in the dormant FT-0 groups (Fig. 3D). This may represent an early attempt to eliminate severely damaged cells via regulated cell death to preserve the population’s overall fitness^41,42. In stark contrast, during the chronic phase (FT6), AIF1 expression surged in the FT6-50 group—the group with the highest mortality—while it was downregulated in the resilient FT6-20 group. This late-stage upregulation in FT6-50 likely signifies a final, failed effort to manage catastrophic damage or is a direct consequence of the irreversible commitment to apoptosis.

The transcriptional dynamics of the respiratory chain provide a molecular narrative of yeast’s struggle against freeze-thaw stress. An early, adaptive upregulation of Complex III genes (CYC1/CYC7) is a key survival strategy, most effectively deployed by cells in the FT-20 state. This is followed by a progressive failure of core bioenergetics, marked by NDI1 suppression, and culminates in divergent apoptotic signaling: an early, controlled response in dormant cells (FTx-0) versus a terminal, damage-driven response in metabolically exhausted cells (FT-50). The consistent association of the FT-20 condition with a gene expression profile favoring mitochondrial integrity and reduced apoptosis underscores its critical role in conferring freeze-thaw tolerance.

Mitochondrial division-related genes expression in yeast cells

Fis1 and Dnm1 are core regulatory factors of mitochondrial division in yeast, and their expression patterns undergo dynamic changes under various stress conditions, closely related to oxidative stress levels, cell death regulation, and peroxisome division^43,44. The expression patterns of mitochondrial fission genes FIS1 and DNM1 were profoundly shaped by the physiological state at freezing, defined by pre-freeze fermentation duration (0, 20, or 50 min), rather than solely by the number of freeze-thaw (FT) cycles. In the FT0-0 group, DNM1 expression was significantly elevated, whereas FIS1 remained low and comparable to the control (Fig. 4A). This suggests that starved, dormant cells initiate rapid mitochondrial fission upon initial freeze stress, likely to generate functional mitochondrial fragments for energy maintenance under metabolic constraint^45,46. However, by the 6th cycle (FT6-0), both DNM1 and FIS1 were markedly downregulated (Fig. 4A, B), indicating a shift toward mitochondrial quiescence or fragmentation arrest, consistent with the observed low respiratory activity (Fig. 1D, F, H) and high O₂^.⁻ production (Fig. 1E) in this group, suggesting inefficient, leaky fission rather than productive biogenesis. At FT3, DNM1 was significantly upregulated in both FT3-20 and FT3-50 (Fig. 4B), while FIS1 was suppressed in FT3-50 and FT3-20 but elevated in FT3-0 (Fig. 4A). This divergence implies that cells frozen in active growth phases (20/50 min) delay fission initiation, possibly prioritizing repair over division during early stress. By FT6, DNM1 expression stabilized near FT3 levels (no further increase), whereas FIS1 was significantly upregulated in FT6-20 and FT6-50 (Fig. 4A). This late-phase FIS1 induction—coinciding with peak COX activity (Fig. 1D), elevated acetyl-CoA synthetase (Fig. 1F), and lowest O₂^.⁻ among FT6 groups (Fig. 1E), which suggests a protective fission program: FIS1-mediated division may facilitate segregation of damaged mitochondrial components, supporting mitophagy and preserving a functional mitochondrial network^47,48. Notably, FIS1 is also involved in peroxisomal fission⁴⁵; its upregulation may reflect a broader organelle quality-control response to chronic oxidative stress.

Fig. 4. Relative expression of mitochondrial division related genes in yeast cells under freeze–thaw cycles of different fermentation processes.

A Relative expression of FIS1, B relative expression of DNM1. F0/3/6-50: Fermentation for 50 min, freezing for 24 h, freeze-thawing for 0,3,6 times; F0/3/6-20: First ferment for 20 min, freeze for 24 h, thawed, then fermented for 30 min, freeze–thaw for 0,3,6 times; F0/3/6-0: First frozen for 24 h, thawed, then fermented for 50 min, freeze–thaw 0,3,6 times. Different lowercase letters indicate significant differences (p < 0.05) among groups within the same freeze–thaw cycle.

FT6-0 shows high mortality resistance but suppressed fission genes and low respiratory capacity, indicating survival via dormancy, not dynamic remodeling; FT6-50 exhibits high mortality yet elevated DNM1/FIS1—likely reflecting stress-induced, dysregulated fission in a damaged subpopulation; Only FT6-20 displays the optimal combination: moderate mortality + induced FIS1 + sustained respiratory function, suggesting that timely, FIS1-dependent fission is a hallmark of resilient adaptation. Mitochondrial fission in freeze-thaw-stressed yeast is state-dependent and biphasic: an early DNM1-driven burst in dormant cells (FT-0), followed by a delayed FIS1-dominated, quality-control-oriented fission in metabolically primed cells (FT-20/50). The coordinated upregulation of FIS1 at FT6 in FT6-20/50—paired with enhanced respiratory enzyme activities—supports its role not merely in division, but in maintaining mitochondrial fitness under recurrent cryostress^46,48. Future studies should couple these transcriptional data with mitochondrial morphology (e.g., TEM or confocal imaging of mtDNA/nucleoids) to distinguish productive fission from pathological fragmentation.

Mitochondrial autophagy-related genes expression in yeast cells

The transcriptional response of yeast to freeze-thaw (FT) stress involves a coordinated regulation of mitochondrial quality control pathways, including mitophagy and ER-mitochondria communication, which is critically dependent on the cell’s physiological state at the time of freezing (defined by pre-FT fermentation duration: 0, 20, or 50 min). As FT cycles progressed, the expression of core mitophagy genes ATG32, ATG11, and ATG8 exhibited complex, non-uniform patterns (Fig. 5A–C), indicating that mitophagy is not simply induced as a blanket response to damage but is finely tuned based on initial cellular conditions. In the initial (FT0) phase, ATG11 and ATG8 were broadly upregulated across all groups, suggesting a general activation of autophagic machinery upon first exposure to cryostress. Notably, ATG32 was significantly downregulated in the FT0-20 group, implying that cells frozen during early active growth may initially rely less on receptor-mediated mitophagy. By the 3rd cycle (FT3), a key divergence emerged: ATG32 expression was highest in the metabolically dormant FT3-0 group, while it remained low in the active FT3-50 group. Concurrently, ATG8 expression was lowest in FT3-20. This suggests that the strategy for managing mitochondrial damage differs; starved cells (FT-0) may prioritize selective mitophagy, whereas cells with recent metabolic activity (FT-20/50) may employ alternative repair or tolerance mechanisms. At the 6th cycle (FT6), the pattern shifted again. The FT6-50 group, which suffered the highest mortality (Fig. 1A), showed the highest ATG32 but the lowest ATG11 expression. This uncoupling of the Atg32-Atg11 axis—a critical step for targeted mitophagy—suggests a potential breakdown in the selective autophagy pathway under severe, chronic stress. In stark contrast, the FT6-0 group exhibited the highest ATG8 levels, possibly reflecting a sustained, albeit non-selective, autophagic effort to maintain homeostasis in the absence of fermentative energy^49,50.

Fig. 5. Relative expression of mitochondrial autophagy related genes in yeast cells under freeze–thaw cycles of different fermentation processes.

A Relative expression of ATG32, B relative expression of ATG11, C relative expression of ATG8, D relative expression of MMM1, E relative expression of MDM10, F relative expression of MDM12, G relative expression of MDM34. F0/3/6-50: Fermentation for 50 min, freezing for 24 h, freeze-thawing for 0,3,6 times; F0/3/6-20: First ferment for 20 min, freeze for 24 h, thawed, then fermented for 30 min, freeze–thaw for 0,3,6 times; F0/3/6-0: First frozen for 24 h, thawed, then fermented for 50 min, freeze–thaw 0,3,6 times. Different lowercase letters indicate significant differences (p < 0.05) among groups within the same freeze–thaw cycle.

The endoplasmic reticulum-mitochondria encounter structure (ERMES) complex, encoded by MMM1, MDM10, MDM12, and MDM34, plays a pivotal role in maintaining mitochondrial integrity and facilitating inter-organelle communication during stress^51,52. Its gene expression profile mirrored the mitophagy dynamics, further underscoring the link between organelle contact sites and quality control. At FT0, MMM1, MDM10, and MDM12 (Fig. 5D–F) were significantly upregulated in the FT0-0 and FT0-20 groups, but MMM1 was downregulated in FT0-50. This indicates that cells not in a late-fermentation state can rapidly reinforce ER-mitochondria contacts as a protective measure⁵³. A striking observation was the consistent and significant upregulation of MDM34 in the FT-50 groups across all FT cycles (Fig. 5G). Given that Mdm34 is crucial for stabilizing ERMES and regulating lipid transfer, this sustained induction likely represents a specific adaptive mechanism in cells that were metabolically active at freezing⁵⁴. By strengthening ER-mitochondrial tethering, these cells may enhance their capacity for phospholipid exchange, calcium signaling, and coordinated stress responses, thereby mitigating ROS-induced damage^55,56. The expression of MMM1 and MDM10 showed more variable trends, being suppressed in some FT3 groups but induced in FT6-20. This complexity suggests that the stoichiometry and assembly of the ERMES complex are precisely regulated, with different subunits playing distinct roles at various stages of the stress timeline. The freeze-thaw stress response in yeast involves a sophisticated interplay between mitochondrial autophagy and ER-mitochondria communication. Cells that were in an active metabolic state prior to freezing (FT-50) appear to mount a robust, long-term adaptive strategy characterized by the sustained upregulation of MDM34 and a controlled mitophagy response. Conversely, cells frozen in a dormant state (FT-0) rely on a strong initial mitophagy burst (ATG32) and high basal autophagic flux (ATG8). The data collectively demonstrate that the efficacy of mitochondrial quality control pathways is a major determinant of yeast survival and fitness during recurrent cryopreservation cycles.

In conclusion, this work provides a comprehensive mechanistic framework for understanding how the physiological state of yeast at the onset of freeze-thaw stress dictates its long-term survival and mitochondrial health. An intermediate metabolic state—achieved by a 20-min pre-freeze fermentation—optimally primes the cell for resilience. In stark contrast, cells frozen in a starved (FT-0) or highly active (FT-50) state fail to coordinate these responses effectively, leading to either a stressed dormancy or a catastrophic collapse of mitochondrial function, respectively. Therefore, successful adaptation to recurrent cryostress is not merely a function of generic stress defenses but is a direct consequence of the cell’s initial metabolic “readiness,” which sets the trajectory for a successful integrated stress response. This insight has profound implications for optimizing cryopreservation protocols in both industrial and research settings.

Methods

Culture medium and cultivation of yeast

Saccharomyces cerevisiae SC-8532 (screened from fermented dough and preserved in our laboratory) was activated and cultivated using the method of Xie et al.⁵⁷. Yeast cells (2.0 g, about 70% water content) were inoculated into the simulated dough medium of 100 mL. The simulated dough medium (glucose 40 g L⁻¹, ammonium sulfate 2.5 g L⁻¹, urea 5 g L⁻¹, potassium dihydrogen phosphate 16 g L⁻¹, disodium hydrogen phosphate 5 g L⁻¹, magnesium sulate 0.6 g L⁻¹, nicotinic acid 22.5 mg L⁻¹, pantothenic acid 5.0 mg L⁻¹, vitamin B1 2.5 mg L⁻¹, vitamin B6 1.25 mg L⁻¹, and vitamin B2 1.0 mg L⁻¹, folic acid 0.5 mg L⁻¹) was modified from that described by Xie et al.⁵⁸.

The yeast-inoculated simulated dough medium was placed in an incubator and fermented (30 °C, humidity 85%, 50 min), determined as the control group; the yeast-inoculated simulated dough medium was fermented at 30 °C, humidity 85% for 50 min, frozen for 24 h, and then thawed in the incubator at 25 °C, humidity 85%, as FT0-50; the yeast-inoculated simulated dough medium was fermented for 20 min at 30 °C, humidity 85% frozen for 24 h, thawed at 25 °C, humidity 85%, and the thawed fermentation broth was ferment for 30 min again at 30 °C, humidity 85%, as FT0-20; the yeast-inoculated simulated dough medium was frozen (24 h), thawed in an incubator (25 °C, humidity 85%), and the thawed fermentation liquid was fermented for 50 min at 30 °C and humidity 85%, as FT0-0. After the above process was completed, the medium was put into the refrigerator at −18 °C and frozen for 24 h. The medium was placed in a thermostatic incubator to thaw until free of ice, recorded as the 1st freeze-thaw cycle (FT1). The yeast-inoculated simulated dough medium was taken sequentially for the third and sixth freeze-thaws, recorded as FT3/6-50, FT3/6-20, and FT3/6-0, respectively (Fig. S2).

Determination of the mortality of yeast cells

The reaction between the fermentation solution and an equal volume of Meilan stain was left to stand for 5 min, and the stained and unstained cells were observed under the microscope.

Mortality rate = stained cells number /total number of cells×100% (1)

Determination of mitochondrial membrane potential

An appropriate amount of fermentation broth was sampled, and centrifuged (2310 × g, 4 °C, 10 min), supernatant fluid was dropped, RH-123 was added and loaded (30 °C, 30 min), and a fluorescence microscope was used to observe and calculate the percentage of green fluorescent yeast was observed under⁵⁹.

Sample preparation of yeast intracellular material

Yeast fermentation broth (100 mL) was centrifuged at 2310 × g for 10 min. The supernatant fluid was dropped, and the collected cells were washed with physiological saline twice and transferred into the mortar with physiological saline. The collected cells were broken by an Ultrasonic cell pulverizer (JY99-IIDN, Ningbo Scientz Biotechnology Co., Ltd, Ningbo, China). After vortexing and centrifuging (13,300 × g, 4 °C) for 10 min, the supernatant was collected as the yeast intracellular sample. The protein content of intracellular material was determined using Coomassie Brilliant Blue.

Preparation of mitochondrial content sample

Isolation of mitochondria was according to the method of Liao et al.⁶⁰. The mitochondria content sample was lysed by adding lysis solution (1% X-tristate, 50 mmol/L Tris–HCl pH 8.5) at 4 °C for 10 min and then centrifuged (11,000 × g, 4 °C) for 10 min to save supernatant fluid as the sample.

Determination of cytochrome c content

Cytochrome c content was determined according to the method of Xie et al.⁶¹. The intracellular sample was mixed with an appropriate amount of pH 7.3 PBS buffer and trichloroacetic acid was put in and shaken (15 min), 1 M NaOH was added and centrifuged (9300 × g, 10 min), the supernatant was taken and mixed with appropriate amount of 0.4 M Na₂S₂O₃ solution, and absorbance value (OD₅₂₀) was measured, and cytochrome c was calculated according to the standard curve (Fig. S3).

Measurement of cytochrome c oxidase activity

The sample of mitochondrial content was taken and reduced with Na₂S₂O₃ solution with 0.02% cytochrome c and mixed according to 1:1. The amount of change in absorbance value within 1 min (OD₅₅₀) was determined and the results were calibrated against the concentration of protein in the sample of the test solution⁶².

Measurement of acetyl-Coenzyme A synthase (ACS) activity

The intracellular sample and the mixture (pH=8.0, 100 mM Tris-HCl 10 mM NAD, 0.2 mM Co-A, 10 mM L-malic acid, MDH (300 μ/ml), citrate synthase (5.5 μ/ml), 20 mM CH₃COOK, 20 mM 5-adenosine triphosphate disodium salt trihydrate) were mixed in a 1:1 mixture. Absorbance value change within 1 min (OD₃₄₀) was determined, and the results were calibrated against the concentration of protein in the sample of the solution to be tested⁶³.

Determination of pyruvate dehydrogenase activity

The mitochondrial content sample and the reaction mixture (50 mM PBS, pH 7.4, 2 mM MgCl₂, 0.1 mM CoA, 6 mM L-Cys, 0.4 mM thiamine pyrophosphate, 4 mM NAD⁺, and 4 mM sodium pyruvate) were taken and mixed according to a volume ratio 1:1. The amount of change in absorbance value within 1 min (OD₃₄₀) was determined and the results were calibrated against the concentration of protein in the sample of the solution to be measured⁶⁴.

Measurement of α-ketoglutarate dehydrogenase activity

The mitochondrial content sample and the mixture (2.5 mM NAD, 0.2 mM thiamine pyrophosphate, 0.1 mM acetyl coenzyme A sodium salt, 0.3 mM DTT, 1 mM MgCl₂) were taken and mixed according to a volume ratio 1:1. The amount of change in absorbance value within 1 min (OD340) was determined, and the results were calibrated against the concentration of protein in the sample of the solution to be measured⁶⁵.

Calculation of isocitrate dehydrogenase activity

The mitochondrial content sample and the mixing mixture (50 mM Tris-HCl, 3.34 mM NAD, 1.66 mM AMP, 6.6 mM MgCl₂, and 6.6 mM trisodium isocitrate salt) were taken and mixed according to a volume ratio of 1:1. The amount of change in absorbance value within 1 min (OD340) was determined, and the results were calibrated against the concentration of protein in the sample of the solution to be measured⁶⁶.

Measurement of relative gene expression

RNA was extracted and purified using the Kit (B518657, Sangon Biotech (Shanghai) Co., Ltd, Shanghai). The extracted total RNA was operated using the kit (R333, Nanjing Novozymes Bioscience & Technology Co., Ltd). A real-time fluorescence quantitative PCR instrument (T100, Bio-Rad Laboratories, Inc., USA) was used to determine the reaction system. The experimental steps were performed according to the kit (Q712, Nanjing Novozymes Bio-technology Co., Ltd) instructions. ACT1 was selected as a reference gene; gene expression was analyzed using total RNA extracted from the entire cell population, including viable, injured, and dead cells. This approach captures the integrated transcriptional response of the heterogeneous population subjected to freeze–thaw stress. All primers were provided by Sangon Biotech (Shanghai) Co., Ltd, and the primer sequences are shown in Table 1. The results were calculated using 2^-∆∆CT for gene expression.

Table 1.

Gene name and primer sequence

Gene name		Primer sequence
ACT1	FORWARD	TGGATTCTGAGGTTGCTGCTTTGG
	REVERSE	TGGTGTCTTGGTCTACCGACGATAG
KGD1	FORWARD	GAAATTGTTTGGTGCCAGGAAGAGC
	REVERSE	TGTGTAAGCGTGGTTCTGTGTATGC
IDH1	FORWARD	TGAATCCGTCCCTGGTGTAGTGG
	REVERSE	CTTGGCGAAGTCAAAGGCAAATCTG
IDH2	FORWARD	GAGATGCCTCTGAGCGAGTCATTAG
	REVERSE	ACAATAACTCTTGGTCTGCCGATGG
ACS1	FORWARD	GCCTCATTCCCCTTCTTCGGTATTG
	REVERSE	TCTGCGTGGCTGGTGTTAAGTTC
AIF1	FORWARD	GGAACTGGGTGGGACATACGC
	REVERSE	ACGGCAGGATCGACTTGGTATAATC
CYC1	FORWARD	GAATTCAAGGCCGGTTCTGCTAAG
	REVERSE	TGCAAGTTTGGACCAACCTTATGTG
CYC7	FORWARD	ACGGGATTCAAACCAGGCTCTG
	REVERSE	CAACTTTGTTAGGACCACCCTCTTC
NDI1	FORWARD	AGTCAGATTCGCTTCCACCAGATC
	REVERSE	TCGCTGTGCTGAGGGTCAATG
FIS1	FORWARD	CCCACCGCTACCATACAGTCAAG
	REVERSE	TTTCACACCAAGCCTTTCGTCATTC
MDM34	FORWARD	TTCCGCTATACCTGGTTGCTTGG
	REVERSE	TGGTTGTGGAAATGATTGCGAAGAG
MDM12	FORWARD	GTGCCGACCTTGTGCTGAATTATC
	REVERSE	CAACAATGCAAAGGGAATGGAGACC
MDM10	FORWARD	CTTCGTTGTCGCTTGGTGCTG
	REVERSE	GGTCGTCCTGTGTTTGTAGAATGTG
MMM1	FORWARD	CGCTAATTGGCTCACGGTCTAAAC
	REVERSE	TTGGAATCTTGGTTCAACGCATCG
ATG32	FORWARD	ATTCCGCAACATTCCCGCAATATAC
	REVERSE	TGGCAGATCGGTATTACAGGCTTG
ATG11	FORWARD	GCCTATGCCCTCGCCTTTGG
	REVERSE	ACCAAGGTTCGTCGTCAGTAAGC
ATG8	FORWARD	AGGAAGGCGGAGTCGGAGAG
	REVERSE	CTACGGTAAGGTCAGCAGGAACTAG
DNM1	FORWARD	GCGAGGATAGCCGGTAAGGATAAG
	REVERSE	TCCCAGGCAAATCTACTAGCGTTAG

Data analysis

All data in this experiment were biologically replicated at least three times, and intracellular substance content and enzyme activity data were analyzed for significance using Duncan’s one-way analysis of variance (ANOVA) in SPSS 20. There is a statistically significant difference with P < 0.05. Data were plotted using Origin 2018 (OriginLab, Northampton, MA, USA). Gene data were determined as 2^-ΔΔCT, and one-way ANOVA in GraphPad 9 (GraphPad Software, Boston, MA) was used to calculate and compare the gene expression.

Author contributions

Dongdong Xie: Conceptualization, Methodology, Resources, Project administration, Funding acquisition, Writing-review & editing, Supervision, Formal analysis. Xing Li: Investigation, Data curation, Writing-Original draft, Visualization. Jiaxin Zheng: Software, Visualization. Zhao Liang: Software.Yuxin Xue: Software.

Data availability

The datasets generated and/or analyzed during the current study are not publicly available due to an unsubmitted graduation thesis but are available from the corresponding author on reasonable request.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41538-026-00921-5.