Mitochondrial protein and organelle quality control – lessons from budding yeast

Summary

Mitochondria are essential for normal cellular function and have emerged as key aging determinants. Indeed, defects in mitochondrial function have been linked to cardiovascular, skeletal muscle and neurodegenerative diseases, premature aging and age-linked diseases. Here, we describe mechanisms for mitochondrial protein and organelle quality control. These surveillance mechanisms mediate repair or degradation of damaged or mistargeted mitochondrial proteins, segregate mitochondria based on their functional state during asymmetric cell division, and modulate cellular fitness, the response to stress and lifespan control in yeast and other eukaryotes.

Introduction

Mitochondria are essential for cellular functions including aerobic energy generation, central metabolism, production of indispensable biomolecules, calcium storage and signal transduction (1–3). Indeed, damage to mitochondria is associated with premature aging (4–8) and age-related diseases including muscle and neurodegenerative diseases (6, 9–14). Here, we describe recent findings obtained using the budding yeast Saccharomyces cerevisiae on mitochondrial protein and organelle quality control mechanisms that are conserved, essential for normal cell function and lifespan control, and compromised in human disease.

Mitochondrial protein quality control

Mitochondrial proteins are challenged by folding and assembly stressors not found in other organelles. They are encoded by two genomes (nuclear and mitochondrial DNA), targeted to four submitochondrial compartments, assembled into some of the largest membrane protein complexes found in eukaryotic cells, and subject to damage by mitochondrial reactive oxygen species (ROS). Indeed, defects in folding or assembly of the human respiratory chain megacomplex I₂III₂IV₂, which contains 140 integral and peripheral membrane proteins (15), play a major role in the pathology of proteinopathies including Alzheimer’s disease, Parkinson’s disease, and amyotrophic lateral sclerosis (16).

Mitochondrial proteostatic stress, i.e. accumulation of misfolded mitochondrial proteins within the organelle or the cytosol, can activate the mitochondrial unfolded protein response pathway (UPRmt). The UPRmt is a signaling pathway that induces expression of mitochondrial protein quality control proteins (e.g. chaperones and proteases), down-regulates transcription and translation of oxidative phosphorylation system (OXPHOS) genes (to reduce the protein folding burden on the organelle), and modulates expression of genes that promote mitochondrial function and support cell survival and adaptation (17).

In yeast, the UPRmt is functionally related to the unfolded protein response activated by mistargeting of proteins (UPRam). This pathway is induced by impaired import of proteins into mitochondria and accumulation of misfolded or unimported mitochondrial proteins in the cytosol. UPRam activates gene expression that decreases translation and elevates proteasomal activity (18). Indeed, the transcriptional remodeling in response to inhibition of mitochondrial protein import is mediated by ‘wideband’ transcription factors Hsf1 and Rpn4 and includes targeted downregulation of the oxidative phosphorylation components by inactivation of the HAP complex (19). Both the UPRmt and UPRam pathways are predicted to be responses to mitochondrial precursor over-accumulation stress (mPOS), a condition that activates expression of genes that inhibit general translation, stimulates stress-resistant protein translation and ultimately leads to cell death if the proteostatic stress associated with defects in mitochondrial protein import is not alleviated (20, 21).

Although the upstream regulators of the UPRmt pathways are not fully conserved between yeast and mammals, the targets of those pathways are conserved (17). The UPRmt and related stress response systems activate expression of mitochondrial chaperones (e.g. mitochondrial Hsp70, Hsp60, Hsp10), mitochondrial proteases (AAA, CLP and Lon proteases) and the proteasome to promote both re-folding and degradation of misfolded mitochondrial proteins. Peptides produced by proteolysis of proteins within mitochondria are exported into the cytosol by the ATP-binding cassette transporter HAF-1 (22, 23). Up-regulation of the proteasome in response to mitochondrial damage or dysfunction (18) is critical for degradation of misfolded or mistargeted mitochondrial proteins in the cytosol by the mitochondria-associated degradation pathway (MAD) and related pathways (see below) (24, 25).

Recent studies also revealed a possible role for mitochondria-derived vesicles (MDVs) in mitochondrial quality control in mammalian cells. Mitochondrial proteins and lipids that are damaged by exposure to stressors, including mild oxidative stress, are packaged into 70-150 nm vesicles (MDVs) that can be targeted to lysosomes or peroxisomes for degradation (26). Mitochondrial-derived compartments (MDCs) have been identified in yeast (27). However, these structures are physically and functionally distinct from MDVs. MDCs are larger than MDVs and may function in reducing the toxicity associated with amino acids and not in protein quality control (28).

Finally, mitochondria that are damaged beyond repair can be degraded by mitophagy. Mitophagy is a conserved and selective form of macroautophagy that is mediated by the mitophagy receptor gene ATG32 and other ATG genes. During this process, a double-membrane structure, the pre-autophagosomal structure or phagophore assembly site (PAS), envelops targeted mitochondria to form the autophagosome. Autophagosomes are then targeted to and degraded by the lysosome (vacuole in yeast) (29). Both defective and excessive mitophagy have been linked to age-associated neurodegenerative diseases (e.g. Parkinson’s and Alzheimer’s diseases), metabolic diseases, myocardial injury, muscle dystrophy, and liver disease (30).

Interestingly, several lines of evidence indicate that mitophagy regulates mitochondrial abundance and function in response to cellular or environmental cues and may not play a major role in removal of damaged mitochondria in yeast. Indeed, mitophagy is not induced by mitochondrial stressors that activate the UPRmt in yeast, including loss of mitochondrial membrane potential or oxidative stress (31, 32). Rather, yeast mitophagy is induced by limitations in nutrient availability and regulated by the Target of Rapamycin (TOR) pathway (33). Specifically, mitophagy is triggered under conditions in which mitochondrial oxidative phosphorylation and abundance are down-regulated, including changes in carbon source that trigger a shift from mitochondrial respiration to glycolysis, nitrogen starvation or entry into stationary phase (34). Each of these nutrient limitation conditions results in TOR-mediated release of repression of ATG32 by the transcription factor Ume6-Sin3-Rpd3, causing up-regulation of mitophagy (33).

Interestingly, emerging evidence support a role for mitophagy in damage-independent regulation of mitochondria in mammalian cells. First, in mouse tissues, mitophagy occurs under basal, non-stressed conditions and in a manner that is not dependent on PINK1, the PTEN-induced putative protein kinase that activates mitophagy in response to defects in mitochondrial function (35). Second, a compound that induces the autophagic adaptor molecule P62/SQSTM1 activates mitophagy without damaging mitochondrial membrane potential or accumulation of Parkin, a Pink1-regulated a ubiquitin ligase that activates mitophagy in response to mitochondrial damage, to the organelle (36, 37). These studies raise the possibilities that 1) Pink1 and Parkin do not regulate mitophagy under basal conditions in vivo, and 2) mitophagy can be induced by mechanisms other than mitochondrial damage in mammalian cells.

MAD in turnover of mitochondrial outer membrane proteins

MAD is analogous to other organelle quality control pathways, including ER-associated degradation (ERAD) (38, 39), endosome and golgi-associated degradation (EGAD) (40), and chloroplast-associated protein degradation (CHLORAD) (41, 42). In all of these pathways, unfolded proteins are identified, tagged by ubiquitination, extracted from organelles by a segregase complex and degraded by the UPS.

MAD and its related pathways have emerged as key players in mitochondrial quality control in yeast (Fig. 1–2). Specifically, our previous studies indicate that deletion of yeast genes encoding non-essential mitochondrial proteases and chaperones, or components that mediate mitophagy or macroautophagy, does not increase the sensitivity of yeast to exposure to mild oxidative stress in mitochondria (32). In contrast, inhibition of MAD or UPS increases stress-induced defects in growth and mitochondrial redox state, and causes accumulation of ubiquitinated proteins in mitochondria. Moreover, inhibition of MAD or the UPS decreases lifespan (32, 43). Thus, MAD, and not mitophagy or non-essential mitochondrial proteases and chaperones, is critical for the cellular response to oxidative stress.

Figure 1. MAD and MAD-related proteasomal degradation of proteins in mitochondrial OM and inner compartments.

Please refer to the text for a description of these pathways. OM: outer membrane; IMS: intermembrane space; IM: inner membrane; Star: damaged mitochondrial protein; U: ubiquitin.

Figure 2. MAD-related pathways for mitochondrial protein quality control.

Please refer to the text for a description of mitoTAD (mitochondrial protein translocation-associated degradation), Pth2-Dsk2, mitoCPR (mitochondrial compromised protein import response pathway) and mitoRQC (mitochondrial ribosome-associated quality control) pathways. OM: outer membrane; IMS: intermembrane space; IM: inter membrane; U: ubiquitin.

The segregase complex functions in MAD, ERAD, EGAD and CHLORAD. It is targeted to specific organelles, functions in extracting damaged proteins from those organelles, and targeting of damaged proteins to the proteasome for degradation. Cdc48 (VCP/p97 in mammals and TER94 in Drosophila), a core component of the segregase complex, uses the energy of ATP hydrolysis to extract proteins from membranes and protein complexes (40, 43–46). It is an abundant, essential AAA-ATPase that contains two ATPase domains and forms a hexameric structure with a central pore (47). ATP hydrolysis by Cdc48 generates a pulling force that drives transport of unfolded proteins into its central pore, which results in the extraction of the bound protein (48). The extracted proteins are then delivered to the proteasome, where they are deubiquitinated and degraded (49).

In yeast, the Cdc48-containing segregase complex is associated with cofactors that facilitate substrate processing including ubiquitin-binding proteins Npl4 and Ufd1. Not surprisingly, since Cdc48 and the segregase function in proteostasis control for multiple organelles, Cdc48 also binds to adapters that target the segregase to different subcellular sites (40, 43–46). Targeting of Cdc48 and the segregase to mitochondria in MAD is mediated by Doa1 (Ufd3, PLAA or PLAP in mammals), Ubx2, and Vms1. These adapter proteins bind to Cdc48 and to ubiquitin adducts on mitochondrial proteins, recruiting the segregase to mitochondria and mitochondrial substrates in response to stress (43, 44) (Fig. 1).

Doa1 is physically associated with both Cdc48 and damaged mitochondrial proteins (Fig. 1A). Moreover, deletion of DOA1 results in accumulation of damaged, ubiquitinated proteins in mitochondria, increased sensitivity to mitochondria-targeted oxidative stress and reduced cellular lifespan in yeast (32, 43). Ubx2, a protein initially identified as an adaptor protein for Cdc48 in ERAD, has also been detected in mitochondria. Ubx2 contains the UBX domain, which is a docking site for Cdc48 and recognizes defective Sam35 and Sen2, peripheral mitochondrial outer membrane (OM) proteins, in conjunction with Doa1 in MAD (49, 50) (Fig. 1B). Finally, Vms1 translocates to mitochondria in yeast under oxidative stress induced by the accumulation of an oxidized sterol, ergosterol peroxide, in mitochondria (44, 51) (Fig. 1C). However, the role of Vms1 in MAD is under debate, because Vms1 has also been implicated in the MAD-related mitoRQC pathway described below (43, 52).

Tail-anchored (TA) ER proteins that are mistargeted to the mitochondrial OM are also degraded by MAD-related proteasomal degradation. These proteins are recognized and extracted from the organelle by Msp1, a mitochondrial AAA-ATPase and hexameric complex with a central core (53–57). ATP hydrolysis by Msp1 drives extraction of target proteins and insertion of those proteins into the central pore of Msp1 (56, 57). Thereafter, mistargeted proteins are transferred to and inserted into the ER by the guided entry of TA protein (GET) pathway, ubiquitinated by E2 conjugating enzymes Ubc6 and Ubc7 and the E3 ligase Doa10, and targeted for proteasomal degradation by a segregase complex containing Cdc48, Npl4 and Ufd1. Thus, the degradation of TA ER proteins that are mistargeted to the mitochondrial OM occurs through the coordinated actions of a MAD-related pathway and ERAD (58–60) (Fig. 1D).

Interestingly, under basal, non-stressed conditions, Fzo1, a GTPase of the dynamin family that mediates mitochondrial fusion, is ubiquitinated, recognized by Doa1, removed from the OM by the segregase and sent to the proteasome for degradation (43). Ubx2 has been implicated as an adaptor for ubiquitinated Fzo1 in this MAD-mediated event (61). Other OM proteins including Tom70 (a component of the translocase of outer membrane (TOM) complex that mediates import of proteins in mitochondria) and Mdm34 (a mitochondrial component of the ER and mitochondria encounter structure (ERMES) complex) are degraded by MAD under basal, non-stressed conditions (43). Since these MAD-associated protein degradation events occur in the absence of excessive stress, it is possible that the targeted proteins are not necessarily damaged or unfolded. Rather, MAD-induced degradation of Fzo1, Tom70 and Mdm34 may regulate mitochondrial fusion, protein import, and interaction of mitochondria with the ER, respectively, in response to cellular or environmental cues and not in response to damage or misfolding of those proteins.

MAD-related mitochondrial quality control pathways that respond to defects in targeting and import of proteins into mitochondria

Over 90% of the proteins within mitochondria are encoded by nuclear genes, contain mitochondrial targeting sequences, and are imported into the organelle. Mechanisms for import of proteins into mitochondria have been characterized extensively in model organisms including S. cerevisiae. In yeast and other eukaryotes, nuclear-encoded mitochondrial proteins are imported into the organelle through channels in the mitochondrial OM and inner membrane, the TOM and translocase of the inner membrane (TIM) complexes, respectively. Moreover, import of proteins into the organelle often occurs co-translationally: nuclear-encoded mitochondrial proteins are translated on mitochondria-associated ribosomes and translocated across one or both mitochondrial membranes during translation (62).

Three MAD-related pathways, mitoTAD (mitochondrial protein translocation-associated degradation), the Pth2-Dsk2 pathway, and mitoCPR (mitochondrial compromised protein import response pathway), respond to defects in import of proteins into mitochondria and the associated accumulation of mitochondrial proteins at the TOM complex or translation machinery in the cytosol (Fig. 2). In these pathways, impaired mitochondrial proteins are marked by a post-translational modification. In many cases, proteins are marked by ubiquitination and degraded by the proteasome. However, in some cases, compromised proteins are imported into mitochondria, where they are degraded by proteases within the organelle.

mitoTAD and the Pth2-Dsk2 pathway function under non-stressed conditions that target proteins that accumulate at the TOM complex for degradation by the proteosome. In mitoTAD, the Cdc48-containing segregase is recruited to mitochondria by binding of the segregase adapter protein Ubx2 to the TOM complex (49). This leads to segregase-mediated targeting of proteins that accumulate at the TOM complex to the proteasome (Fig. 2A). The Pth2-Dsk2 pathway promotes proteasomal degradation of proteins that accumulate at the TOM complex; however, it does so in a manner that is independent of both Cdc48 and Ubx2 (63) (Fig. 2B). In this pathway, an E3 ubiquitin ligase, Rsp5, catalyzes ubiquitination of mislocalized precursor proteins. Pth2, a mitochondrially-localized peptidyl-tRNA hydrolase that contains a ubiquitin-like (UBL) domain, then binds to marked proteins (64). Dsk2, which contains UBL and UBA (ubiquitin-associated domain) domains, binds to Pth2 and shuttles ubiquitinated proteins to the proteasome for degradation (65, 66).

mitoCPR also responds to accumulation of proteins at the TOM complex. However, in contrast to mitoTAD and the Pth2-Dsk2 pathway, mitoCPR is a stress-activated pathway that can induced by overexpression of nuclear-encoded mitochondrial proteins (67). During mitoCPR, Cis1 is up-regulated by the transcription factor Pdr3, binds to the Tom70 component of the TOM complex and recruits Msp1 to mitochondria. Msp1 then extracts non-imported proteins from the TOM complex, leading to their degradation by the proteasome (Fig. 2C).

Finally, during co-translational import of proteins into mitochondria, ribosomes engaged in translation of nuclear-encoded mitochondrial proteins can stall as a result of errors including STOP codon errors, secondary structures in the mRNA or deficiencies in amino acids or tRNAs (68). This results in accumulation of nascent peptides of nuclear-encoded mitochondrial proteins on stalled ribosomes, and leads in turn to proteostatic stress. The mitochondrial ribosome-associated quality control pathway (mitoRQC) responds to this import defect by mediating degradation of nascent peptides that accumulate on stalled ribosomes, using either the proteasome or mitochondrial proteases (Fig. 2D).

In mitoRQC, mistargeted nascent peptides are released from stalled ribosomes but remain associated with the 60S ribosomal subunit. The ribosome quality control complex component Rqc2, which binds to ribosomes, mediates addition of alanine and threonine residues to the C-terminus of the nascent chain (C-terminal alanine-threonine tails, CAT-tails). Addition of CAT-tails may expose lysine residues on nascent peptides that are hidden in the exit tunnel of the 60s ribosomal subunit. Exposed lysine residues are ubiquitinated by the E3 ubiquitin ligase Ltn1, which targets the peptide to the proteasome (69–71). If a lysine is not accessible, mistargeted nascent peptides are imported into mitochondria where they are degraded by proteases within the organelle. Vms1, a protein implicated as a segregase adapter for MAD, inhibits addition of CAT-tails and functions as a peptidyl-tRNA hydrolase to release the nascent chain from the 60S ribosomal subunit (69, 72, 73). Released nascent peptides are then imported into mitochondria and degraded by mitochondrial proteases (72).

MAD function in turnover of damaged proteins within mitochondria

Although MAD has mechanistic and functional similarities with ERAD, most studies focus on MAD function in the turnover of mitochondrial OM proteins. However, ERAD surveillance extends to proteins within the ER lumen: unfolded ER luminal proteins are retrotranslocated across the ER membrane to the cytosol where they are degraded by the proteasome (74–77). Similarly, in CHLORAD, damaged within the chloroplast in plant cells can be degraded in the cytosol (45).

Our studies revealed that MAD functions in protein quality control, not just on the mitochondrial OM, but within the mitochondrial matrix and inner membrane (IM) under basal and oxidative stress conditions (32). Specifically, we find that two proteins that localize to the mitochondrial matrix and are highly sensitive to oxidative damage, Kgd1 (a subunit of the tricarboxylic acid (TCA) cycle component α-ketoglutarate dehydrogenase) and Pim1 (the Lon protease of yeast), are degraded by MAD. Indeed, deletion of DOA1 results in an increase in the steady-state levels and ubiquitination of Kgd1 and Pim1, defects in assembly of Kgd1 into TCA cycle supercomplexes, increased sensitivity to oxidative stress and reduced lifespan (32). These studies also revealed that mitochondrial IM proteins are ubiquitinated in response to oxidative stress (32). Therefore, MAD functions beyond mitochondrial OM proteostasis and has broader functions in mitochondrial quality control than was previously appreciated (32).

Supporting this study, MAD has been implicated in the degradation of the mitochondrial IM protein NADH dehydrogenase (Nde1) (78). Notably, Nde1 has two topomers. One form serves as an NADH dehydrogenase in the electron transport chain (ETC) and localizes to the intermembrane space (IMS). Another form of Nde1 functions as a signal for apoptosis in respiration-deficient cells and is exposed to the cytosol. In healthy cells, the cytosol-exposed form of Nde1 is degraded by the proteasome in a MAD-dependent manner and Nde1 within mitochondria is degraded by i-AAA proteases. In contrast, in respiration-deficient cells, Nde1 is stabilized, which leads to induction of apoptosis (78).

In ERAD, Derlin and the VCP-interacting protein VIMP associate with the segregase and mediate retrotranslocation of unfolded ER proteins across the ER membrane to the cytosol for degradation by the proteasome (79). How MAD substrates are retrotranslocated from internal mitochondrial compartments to the cytosol is still unknown. However, previous studies revealed retrotranslocation of the misfolded mitochondrial IMS proteins that are imported into the organelle by the mitochondrial intermembrane space assembly (MIA) pathway (80). During normal, MIA-dependent import, precursor IMS proteins are maintained in a reduced form in the cytosol and undergo oxidative folding to their final active conformation after import into the IMS. Proteins that are misfolded as a result of defects in this process, including reduction of disulfide bonds in proteins undergoing oxidative folding, are retrotranslocated from the IMS through TOM channels in the mitochondrial OM. Although the fate of retrotranslocated IMS proteins in the cytosol and the role MAD function in degradation of misfolded IMS proteins are not well understood, these findings raise the possibility that import channels in the OM can also serve as bidirectional channels that retrotranslocate misfolded proteins from mitochondria to the cytosol.

Organelle quality control mechanisms for surveillance and segregation of mitochondria during cell division in yeast

An intuitive concept in human experience is that babies are born young, largely independent of the age of their parents. This process, mother-daughter age asymmetry, has been documented from bacteria to humans, and is driven by segregation of aging determinants between mother and daughter cells during cell division (Fig. 3). Age asymmetry is best understood in budding yeast, a species that undergoes asymmetric cell division during vegetative growth. As a result, buds or daughter cells that are produced from an asymmetrically dividing mother cell are not identical to their mother cell. Indeed, factors that promote cellular aging and preserve a memory of age, including damaged proteins, extrachromosomal rDNA circles (ERCs), and lower-functioning organelles, are retained within mother cells and excluded from buds (6, 81–83). On the other hand, rejuvenating factors, including higher-functioning mitochondria and vacuoles (the yeast lysosome), are preferentially inherited by daughter cells (84). Moreover, catalase, an antioxidant, is selectively activated in daughter cells immediately after they undergo cytokinesis and separation from yeast mother cells (82). As described below, asymmetric inheritance of aging and rejuvenating factors is necessary for normal cellular fitness and lifespan and for mother-daughter age asymmetry (85–88).

Figure 3. Asymmetric inheritance of aging determinants during cell division in budding yeast.

Please refer to the text for a description of segregation of aging and rejuvenating determinants in budding yeast.

Importantly, asymmetric inheritance of aging determinants also occurs in mammalian stem cells, which also undergo asymmetric cell division. In asymmetric stem cell division, one daughter cell retains stem cell qualities, while the other becomes a tissue progenitor cell. For example, young and old mitochondria are segregated during asymmetric cell division of human mammary stem-like cells. Daughter cells that inherit young mitochondria become stem cells, while those that inherit old mitochondria become tissue progenitor cells (89). Similarly, segregation of mitochondria has been documented during asymmetric division of hematopoietic stem cells (HSCs) and the associated production of mature blood and immune cells. During this process, lower-functioning mitochondria are retained in HSCs, which serves as a memory of HSC history, ultimately limits their stem-like function and promotes HSC attrition (90). Thus, asymmetric inheritance of mitochondria occurs in mammalian stem cells, where it affects stem cell fitness and fate. Since stem cells are critical for cell and tissue renewal not just during normal homeostasis but also as tissues and organs age, asymmetric inheritance of mitochondria in stem cells affects lifespan control at the organismal level.

Segregation of mitochondria based on their functional state during cell division is a form of mitochondrial quality control that is critical for cellular fitness and lifespan control. However, in contrast to mitochondrial proteostasis control, which occurs in all mitochondria within a cell and does not respond to cell polarity cues, segregation of the organelle during asymmetric cell division is a quality control event that ensures that higher-functioning mitochondria localize to specific sites within the asymmetrically dividing cell. The mechanisms underlying asymmetric inheritance of mitochondria in mammalian cells are not well understood. However, as described below, studies in yeast have led to the identification of patterns of asymmetric mitochondrial inheritance, mechanisms underlying this process, and its impact on cell fitness and lifespan.

Asymmetric inheritance of mitochondria during yeast cell division

In budding yeast, mitochondria can be a rejuvenating or aging factor depending on their quality and functional state. Using biosensors to study mitochondrial function in living cells, we found that mitochondria in developing daughter cells contain less ROS including superoxides and hydrogen peroxide, are more reduced, and have a higher membrane potential (ΔΨ) compared to mitochondria in mother cells (6, 91, 92) (Fig. 4). Moreover, fluorescence loss in photobleaching (FLIP) and network analysis experiments revealed that mitochondria in large buds are a continuous reticulum that is physically and functionally distinct from mitochondria in mother cells. These studies also indicate that deletion of mitochondrial fusion mediators (Fzo1 or Mgm1) leads to decreased accumulation of mitochondria at the bud tip, defects in inheritance of fitter mitochondria by buds and reduced lifespan (5). Thus, our studies revealed that daughter cells selectively inherit fitter mitochondria and identified mechanisms to ensure that mitochondria in the bud are physically and functionally distinct from mitochondria in mother cells. Interestingly, the volume ratio of mitochondria inherited by the bud remains high and relatively unchanged during the aging process in yeast; however, the volume ratio of mitochondria that are retained in the mother cell declines with age (93).

Figure 4. Mechanisms for mitochondrial quality control during cell division in yeast.

RACF (retrograde actin cable flow) and anchorage of mitochondria at the bud tip and the distal tip of the mother cell contribute to the asymmetric segregation of higher- from lower-functioning mitochondria during asymmetric cell division. Please refer to the text for further details.

Equally important, inheritance of fitter mitochondria by yeast daughter cells is critical for normal cellular function and lifespan control. In yeast, there are two models for lifespan control. Chronological lifespan, the survival time of stationary phase, non-dividing yeast cells, is a model for stress resistance in postmitotic cells. On the other hand, replicative lifespan is a model for aging of cell division-competent cells and is measured as the number of times that a cell can divide prior to senescence. We find that mutations or conditions that inhibit inheritance of fitter mitochondria by yeast daughter cells result in defects in overall mitochondrial function, reduced replicative lifespan and loss of mother-daughter age asymmetry (91, 94, 95). Thus, inheritance of fitter mitochondria by yeast daughter cells rejuvenates and resets the aging clock of buds, and enables them to have a full replicative lifespan largely independent of the age of their mother cells. Conversely, lower-functioning mitochondria that are retained in the mother cell are aging determinants that preserve the memory of age of the mother cell and ensure that mother cells continue to age as they give rise to daughter cells with a full replicative lifespan.

Finally, there is evidence for functional segregation of mitochondria in response to cellular positional cues, not just between mother cells and buds, but within mother cells in budding yeast (Fig. 4). As a result of inheritance of fitter mitochondria by buds, mother cells are left with a mixed population of high- and low-functioning mitochondria, leading to a lower overall level of mitochondrial function compared to mitochondria in the bud. Interestingly, a small population of higher-functioning mitochondria accumulates at the distal tip of the mother cell. This population of mitochondria is less oxidized compared to other mitochondria in the mother cell and the bud. Moreover, region-specific retention of some higher-functioning mitochondria in the mother cell is necessary for mother cell fitness and replicative lifespan in yeast (91, 95).

Role of cytoskeleton-driven organelle motility in mitochondrial quality control during asymmetric cell division in yeast

Asymmetric inheritance of mitochondria and aging determinants in yeast depends on actin cables, which are the cytoskeletal tracks for polarized movement of virtually all organelles and cargoes in yeast (Fig. 4). Actin cables are bundles of actin filaments (F-actin) that assemble in the bud, extend along the mother-bud axis and disassemble in the mother cell tip (96–98). The type V myosin motor protein (Myo2) associates with mitochondria and uses the energy of ATP binding and hydrolysis to generate forces for anterograde, bud-directed movement of the organelle using actin cables as tracks (99–107). Our studies revealed that the Arp2/3 complex is also associated with mitochondria in yeast and that Arp2/3 complex-driven actin polymerization at the interface between mitochondria and actin cables also drives mitochondrial movement and inheritance in yeast (108–112) (Fig. 4). Thus, mitochondria undergo linear, polarized movement and inheritance that is mediated by actin cables and two conserved force generators during asymmetric cell division in yeast.

In contrast to most cytoskeletal tracks, which are stationary, our previous studies revealed that actin cables are dynamic structures that undergo retrograde actin cable flow (RACF), treadmill-like movement away from the bud and towards the mother cell (113, 114). As a result, mitochondria that are using actin cables as tracks for movement from mother cells to buds are effectively “swimming upstream” against the opposing force of RACF from the bud to mother cells. Indeed, RACF serves as a filter to prevent the transport and inheritance of lower-functioning mitochondria from mother to daughter cells. Below, we describe the mechanisms underlying RACF and evidence for a role of RACF in mitochondrial quality control and lifespan in budding yeast.

The mechanism for RACF in yeast is conserved and underlies the retrograde flow of actin networks at the leading edge of motile mammalian cells (113, 114). This process is driven, in part, by the pushing force of elongation of actin cables at their assembly site in the bud (Fig. 5). During RACF, the two formin proteins of yeast, Bni1 and Bnr1, localize to the bud tip and bud neck respectively, and nucleate new actin filaments at those sites (115). Newly polymerized actin filaments are then stabilized and assembled into parallel bundles by tropomyosins, Tpm1 and Tpm2, and actin-bundling proteins, fimbrin (Sac6) and Abp140. Finally, the newly assembled bundles are inserted into the end of the actin cable in the bud such that the fast-growing plus ends of F-actin in the bundles face the plasma membrane in the bud tip. This insertion pushes actin cables, along the mother-bud axis, from their site of assembly in the bud toward the distal tip of the mother cell. A second force, which is a pulling force, also drives RACF (Fig. 5). This pulling force is generated by the type II myosin motor Myo1, which localizes to the bud neck, binds to actin cables that extend across the bud neck and pulls those cables across the bud neck and toward the distal tip of the mother cell (116).

Figure 5. Mechanisms underlying actin cable assembly and force generation on actin cables during retrograde actin cable flow (RACF).

Please refer to the text for a detailed description of these mechanisms.

Importantly, decreasing the rate of RACF, by deletion of the type II myosin that provides RACF pulling forces, results in inheritance of lower-functioning mitochondria by the daughter cell. This defect in asymmetric inheritance of mitochondria reduces daughter cell fitness and results in premature aging. Conversely, deletion of the tropomyosin Tpm2 increases the rate of RACF by promoting Myo1 binding to actin cables. As a result of the increase in RACF rates in tpm2∆ mutants, there is an increase in the quality of mitochondria that are inherited by daughter cells, an increase in daughter cell fitness, and extension of replicative lifespan (94). Thus, the membrane-cytoskeleton interactions that drive mitochondrial movement and inheritance during asymmetric cell division in yeast promote inheritance of fitter mitochondria by daughter cells and retention of lower-functioning mitochondria in mother cells, which in turn affects yeast cell fitness, replicative lifespan and mother-daughter age asymmetry.

Consistent with this, our recent studies revealed that actin cables become functionally impaired with age and that stabilization of actin cables increases yeast cell fitness and extends lifespan (117). These studies led to the identification of a novel modulator of the actin cytoskeleton that influences mitochondrial quality control and the aging process. Deletion of this gene, AAN1 (Actin, Aging, and Nutrient modulator protein 1), increases actin cable stability and abundance and results in an increase in mitochondrial fitness and extended replicative lifespan (117). Intriguingly, aan1Δ cells show altered branched-chain amino acid (BCAA) metabolism. Moreover, genetically modulating BCAA metabolism or decreasing leucine levels in the growth medium results in increased actin cable stability and mitochondrial quality (117). Thus, Aan1 is a previously undescribed aging determinant that impacts yeast replicative lifespan and mitochondrial quality control in response to nutrient cues through effects on actin cable stability. These studies provide additional evidence for a role of the actin cytoskeleton in lifespan control, and are supported by findings that stabilization of the actin cytoskeleton affects cellular fitness and lifespan in C. elegans (118).

Mitochondrial anchorage in mitochondrial quality control

Organelles are not always motile. Mitochondria are immobilized or anchored at specific subcellular locations, which in turn contributes to allocation of mitochondria during cell division and coordination of mitochondrial quality control pathways (86, 119, 120). Indeed, we find that higher-functioning mitochondria accumulate at opposite poles of the budding yeast cell, the bud tip and mother cell tip, and are anchored at those sites by tethers that link mitochondria to other organelles. Moreover, defects in this process affect mitochondrial distribution, quality control and asymmetric inheritance, and ultimately affect overall cell fitness and lifespan.

Three major region-specific mitochondrial retention mediators have been identified in budding yeast: Mmr1-mediated anchorage of mitochondria in the bud tip, tethering of mitochondria to the mother cell cortex by Num1, and anchorage of mitochondria at the mother cell tip by Mfb1. Below, we describe these tethering mechanisms and how they contribute to mitochondrial quality control during inheritance and replicative lifespan in yeast.

Mmr1, a tether for anchorage of mitochondria in the bud tip:

Mmr1 was originally identified as an adapter protein that localizes to mitochondria and links mitochondria to myosin motors for control of mitochondrial movement during cell division-associated inheritance (102, 103, 121, 122). However, there is also evidence that Mmr1 functions to retain mitochondria in the bud by tethering them at the bud tip. First, Mmr1 is a member of the DSL (Delta/Serrate/lag-2) family of tethering proteins (123). In addition, MMR1 mRNA and protein are preferentially localized to the bud tip (123, 124), indicating that the protein encoded by the MMR1 gene functions in the bud tip where mitochondria are anchored and not in the mother cell, where the mitochondria are highly motile. Furthermore, deleting MMR1 leads to defects in accumulation of mitochondria at the bud tip and defects in lifespan control (123), while overexpression of the protein leads to increased accumulation of mitochondria in the bud (5). Importantly, experiments carried out using a temperature-sensitive mmr1-5 mutant revealed a direct role for Mmr1 as a tether for mitochondria in the bud tip: temperature-induced loss of Mmr1 function results in a release of mitochondria from the bud tip (5). Thus, the higher-functioning mitochondria that are selected by RACF for transport to the bud are retained in the bud by Mmr1-mediated anchorage at the bud tip.

Num1, a tether for mitochondria in the mother cell cortex:

Num1 is a phospholipid-binding protein that anchors mitochondria to the cortex of the mother cell as a component of the mitochondria-ER-cortex-anchor (MECA) (125, 126). Num1 localizes primarily to the mother cell during the early stages of the cell cycle (127, 128). Deletion of NUM1 results in increased mitochondrial motility, defects in mitochondrial distribution, and increased inheritance of mitochondria by yeast daughter cells (125, 129, 130). Furthermore, rapid loss of Num1 using an artificially inducible degradation system leads to release of mitochondria from the cortex of the mother cell (131). However, Num1 is multifunctional: it has also been implicated as a tether for the molecular motor dynein at the yeast cell cortex, which contributes to nuclear migration during cell division (127, 128, 132, 133). As a result, the specific role of Num1-mediated tethering of mitochondria in cell division-linked mitochondrial quality control is not well understood. Nonetheless, available evidence indicates that Num1 is a mitochondrial tether that is critical for retention of mitochondria in the mother cell during yeast cell division.

Mfb1, a tether for anchorage of mitochondria in the mother cell tip:

As described above, a small population of higher-functioning mitochondria accumulates in the mother cell tip of dividing yeast cells (91, 119). Mfb1, a mitochondrial F-box protein, was originally identified as a protein that localizes to mitochondria and is required for normal mitochondrial morphology (134, 135). Further characterization revealed that Mfb1 localizes to a subset of mitochondria in yeast: it preferentially localizes to mitochondria that are anchored in the mother cell tip. Consistent with this, analysis of MFB1 deletion mutants revealed that Mfb1 is required for accumulation of mitochondria in the mother cell tip, mitochondrial quality control during yeast cell division and lifespan control (91). Indeed, overall mitochondrial quality in the mother cell and replicative lifespan are significantly compromised in mfb1Δ cells. Thus, Mfb1 contributes to lifespan control in yeast through its function as a tether that anchors higher-functioning mitochondria in the tip of the mother cell in dividing yeast cells.

Intriguingly, although Mfb1 functions as a mitochondrial quality control mediator, it does not sense or detect mitochondrial function. Indeed, disrupting mitochondrial quality by treatment with agents that dissipate mitochondrial membrane potential or induce oxidative stress does not alter localization of Mfb1 to mitochondria or its function in anchoring mitochondria to the distal tip of the mother cell (91). These findings are consistent with the model that higher-functioning mitochondria localize to the mother cell tip in an Mfb1-independent manner but are retained at that site by Mfb1-mediated tethering activity.

Role for the cell polarity machinery in asymmetric inheritance of mitochondria and lifespan control

Recent work revealed a role for the cell polarity machinery for enrichment and anchorage of higher-functioning mitochondria at both poles of the dividing yeast cell. During asymmetric cell division in yeast, the site for formation and development of a bud is selected on the surface of a mother cell. The cell polarity machinery is then activated, which induces reorganization of the cytoskeleton and directed transport of cellular constituents to the selected bud site. This process is mediated by conserved polarity mediators. Bud1/Rsr1, a Ras-type GTPase, is activated at the selected bud site, which recruits Cdc24, a guanine nucleotide exchange factor (GEF) for Cdc42, to that site (136). This, in turn, results in region-specific activation of Cdc42, and Cdc42-dependent assembly of actin cables at that site for transport of cellular constituents, including higher-functioning mitochondria, toward the selected bud site leading to bud formation and growth (97, 137). Polarized actin cables also mediate the transport and further enrichment of Cdc42p at the bud tip (138).

Our studies revealed a role for the cell polarity machinery in mitochondrial positional and quality control during cell division, and showed that the bud site selection machinery controls Mfb1 localization and function in anchorage of higher-functioning mitochondria in the mother cell tip. Interestingly, although Mfb1 localizes to the mitochondria at the mother cell tip throughout the cell cycle, some Mfb1 is detected at the bud tip prior to the cytokinesis (Fig. 6). Mfb1 that localizes to the bud tip mediates anchorage of mitochondria, in conjunction with Mmr1, at that site (91).

Fig. 6. Cell polarity- and cell cycle-linked changes in Mfb1 localization and function during the bud-to-mother transition in yeast.

Please refer to the text for further details.

These studies also revealed how activation of the cell polarity machinery in the bud tip can lead to localization and function of Mfb1 at the distal tip of the mother cell, the opposite pole of the dividing yeast cell. During asymmetric cell division in haploid yeast, newborn cells bud at sites adjacent to the site of the previous cytokinesis (139, 140). As a result, the bud tip in the developing daughter cell becomes the mother tip in the mature cell (Fig. 6). Based on these findings, we propose that enrichment of Mfb1 at the bud tip late in the cell cycle serves two important purposes. First, it contributes to anchorage of higher-functioning mitochondria in the bud tip late in the cell division cycle. Second, it enables Mfb1 and its associated higher-functioning mitochondria to localize to the mother cell tip after the bud separates from its mother cell and becomes a mother cell in the next round of cell division (Fig. 6).

Consistent with this, our recent findings support a role for the cell polarity machinery in lifespan control and in Mfb1 localization and mitochondrial quality control during asymmetric cell division (95). Specifically, we find that the function of the cell polarity machinery, polarized localization of Mfb1 and mitochondrial quality in the mother cell decline with age in yeast. Moreover, the deletion of polarity-establishing genes (BUD1, BUD2, and BUD5) disrupts Mfb1 localization and its function in mitochondrial distribution and quality control, and results in reduced replicative lifespan (95). Taken together, these findings support a mechanism whereby localized Mfb1 enrichment, orchestrated by the cell polarity and cell cycle machinery, plays a direct role in anchoring higher-functioning mitochondria at the distal tip of the mother cell to maintain the quality of mitochondria, cellular fitness and lifespan.

Conclusions

The budding yeast Saccharomyces cerevisiae has been a fruitful model system for mitochondrial protein and organelle quality control, leading to the discovery of mechanisms that are critical for cellular function and lifespan. Moreover, these pathways are mediated by proteins that are conserved and targets for mutation in human disease. This underscores the importance of mitochondrial function in yeast and other eukaryotes. Still, many questions remain, including the precise mechanisms underlying these diverse pathways for mitochondrial quality control; how they are coordinated and regulated in response to cellular and environmental cues; how they change during the aging process and their precise impact on lifespan control and mother-daughter age asymmetry; and whether interventions that improve mitochondrial quality control can promote human health or extend lifespan.

List of Abbreviations:

AAN1

actin, aging, and nutrient modulator protein 1

BCAA

branched-chain amino acid

CAT-tail

C-terminal alanine-threonine tail

CHLORAD

chloroplast-associated protein degradation

ΔΨ

mitochondrial membrane potential

DSL

Delta/Serrate/lag-2

EGAD

endosome and golgi-associated degradation

ERAD

ER-associated degradation

ERC

extrachromosomal rDNA circle

ERMES

ER and mitochondria encounter structure

ETC

electron transport chain

F-actin

filamentous actin

FLIP

fluorescence loss in photobleaching

GEF

guanine nucleotide exchange factor

GET

guided entry of TA protein

HSC

hematopoietic stem cell

IMS

intermembrane space

MAD

mitochondria-associated degradation pathway

MDC

mitochondrial-derived compartment

MDV

mitochondria-derived vesicle

MECA

mitochondria-ER-cortex-anchor

MIA

mitochondrial intermembrane space assembly

mitoCPR

mitochondrial compromised protein import response pathway

mitoRQC

mitochondrial ribosome-associated quality control pathway

mitoTAD

mitochondrial protein translocation-associated degradation

mPOS

mitochondrial precursor over-accumulation stress

OM

outer membrane

OXPHOS

oxidative phosphorylation system

PAS

phagophore assembly site

RACF

retrograde actin cable flow

ROS

reactive oxygen species

TA

tail-anchored

TCA

tricarboxylic acid

TIM

translocase of inner membrane

TOM

translocase of the outer membrane

TOR

target of rapamycin

UBA

ubiquitin-associated

UBL

ubiquitin-like

UPRam

unfolded protein response activated by mistargeting of proteins

UPRmt

mitochondrial unfolded protein response

UPS

ubiquitin-proteasome system