Mitochondrial-derived compartments buffer outer membrane protein load during acute mitochondrial adaptation

Reviewer #1 (Public review):

Summary:

This study investigated the formation of mitochondrial-derived compartments (MDCs) under metabolic adaptations. They hypothesized that MDCs may play a role in regulating the mitochondrial proteome under these conditions by removing excess and superfluous membrane proteins that may challenge mitochondrial proteostasis. They found that glucose restriction, carbon-source switching, and osmotic stress can stimulate MDC formation. Underlying these stressors is a common signaling pathway that involves Snf1-dependent derepression of mitochondrial biogenesis and rapid synthesis and trafficking of nuclear-encoded proteins into mitochondria. They then showed that MDC formation is attenuated in tom70/tom71 mutants, suggesting that the delivery of these proteins to mitochondria is critical. Data also suggested that HAP4-stimulated mitochondrial biogenesis promotes MDC formation, which is further enhanced by glucose restriction and is suppressed after prolonged adaptation.

Strengths:

The genetically amenable yeast system allowed the authors to generate convincing data showing the rapid formation of MDCs under physiologically relevant conditions where the mitochondrial proteome needs to be expanded to accommodate increasing metabolic function. MDCs therefore function to buffer spillovers of outer membrane proteins upon an abrupt protein influx. Overall, the data presented are of high quality. The conclusion is strongly supported by the data.

I think this is a significant study as (1) it supported MDCs as a physiologically relevant mechanism of mitochondrial proteostasis; and (2) it offers a common mechanistic framework explaining the MDC phenomenon under many other conditions such as TOR inhibition and hydrophobic protein overloading previously published by this group. Although the precise mechanism of MDC formation and how MDC formation contributes to the overall proteostasis of mitochondria remain unknown, as the authors stated in the manuscript, the current work is a clearly identifiable milestone in this specific area of investigation.

Weaknesses:

Although the data are overall strong, weaknesses are mainly related to potential misinterpretation of the data.

(1) I have reservations regarding the interpretation of some results. First, the authors concluded that MDC biogenesis is activated when glycolytic metabolism is altered. I disagree with this. The authors should distinguish between "loss of glycolysis" and "loss of glucose repression". The yeast S288C strains are GAL2 and can ferment galactose. Likewise, glycolysis is also supported by raffinose and sucrose. In a broad sense, these carbon sources do support glycolysis as long as sugar influx is maintained at a high level. However, these alternative carbons do not repress mitochondrial respiration like glucose. It is likely the derepression of mitochondrial respiration (which is stated in some sections of the manuscript) instead of loss of glycolytic metabolism that stimulates MDC formation. This needs to be made clear throughout the manuscript. As such, the statement that "Carbon-source switching" stimulates MDCs is not accurate and needs to be re-interpreted.

(2) The explanation for the requirement of low glucose levels could be misleading. A complete lack of carbon sources and high concentrations of 2-DG may shut down global protein synthesis, cell cycle progression, and many other processes, including mitochondrial biogenesis. Glucose at 0.02% is not sufficient to cause glucose repression, as only the high-affinity but low-influx transporters are functioning. Under the low glucose conditions, mitochondrial respiration is also derepressed. In this scenario, low glucose simply plays a role in supporting cell growth without causing the repression of mitochondrial biogenesis.

(3) The idea that MDCs are formed when protein load exceeds the capacity the organelle can accommodate is attractive. Early studies have shown that the mitochondrial compartment is expanded by several folds in volume when yeast cells are switched from fermentative to oxidative metabolism. Perhaps, space expansion takes longer than protein influx increase. It would be interesting to see whether there is a correlation between MDC frequencies and the delay in volume expansion. Long-term adaptation would solve this challenge, as it allows the cell to complete volume expansion.

(4) HAP4 may primarily activate OXPHOS genes but not some MDC cargo proteins. The requirement for "metabolic remodeling" for full induction of MDC formation may be an overstatement. The authors should either reexamine the proteomic data to see whether known MDC cargos are not subject to HAP4 activation or have this discussed in the manuscript.

Reviewer #2 (Public review):

Summary:

Price et al. present new work providing insight into the function and mechanisms of mitochondrial-derived compartments (MDCs) in yeast. The Hughes lab previously established that these large ~micron-sized structures are formed under a variety of conditions including amino acid stress (rapamycin, conA, cycloheximide), alterations in mitochondrial metabolites and lipids, or acute expression of specific outer membrane proteins. These stressors lead to the sequestration of outer membrane proteins (that can include mistargeted inner membrane proteins) that extend or tubulate into large multilamellar structures that ultimately target the vacuole in an ATG5/Dnm1 dependent autophagy related pathway for degradation. Initially reported in aging yeast a decade ago, it has now been accepted as a mechanism to remove excess mitochondrial proteins as a pathway distinct from mitophagy or the extraction of stalled precursors from the import translocon.

In this study, the authors examined additional metabolic transitions they suspected would drive increased mitochondrial protein expression and promote MDC formation. Indeed, they show that glucose-restricted conditions (or a switch to galactose or incubation with 2DG) induced MDCs within 2 hours. This correlated with increased transcription/translation of mitochondrial precursors porin and OM45. Similar results were seen with osmotic shock, a process previously shown to induce mitochondrial gene expression. The metabolic or osmotic shift was shown to activate a yeast AMPK-type kinase called Snf1, which phosphorylates a key substrate Mig1 - an established repressor of mitochondrial gene expression. Loss of these pathways abolished the generation of MDCs under these conditions. As the key novel finding in the study, the authors explored the relationship/requirement for Snf1 and Mig1 using multiple approaches in different backgrounds and employing auxin-inducible degron tools for acute depletion. These data further support the hypothesis that excess mitochondrial outer membrane proteins result in MDC formation to facilitate their removal, at least transiently until the import machinery can adapt to the increased import demand. To test this more directly, they generated an inducible yeast strain to express a canonical transcription factor Hap4 that induces mitochondrial gene expression. In this system, induction of Hap4 expression also resulted in MDC formation. While not all previously reported MDC inducers act through Snf1/Mig1, the common feature is the transcriptional induction of mitochondrial protein expression.

Strengths:

The important aspect of this work is that the authors dissected the transcriptional signaling pathway that induces MDCs in a much more physiological metabolic transition, which complements the more common use of chemical compounds. They had previously shown that overexpression of individual outer membrane proteins could lead to MDCs, but here the Hap4 expression offers a new condition to show that the canonical induction of mitochondrial biogenesis leads to MDC shedding. Overall, the data are of high quality, the findings are clear, and the work provides important new insights into the regulation of MDC formation.

Weaknesses:

There are a few points that should be addressed.

(1) MDCs are almost exclusively monitored through GFP-tagged TOM70, and the authors do not show the inclusion of any endogenous cargo. The evidence for their fate in the vacuole is through the appearance of cleaved, free GFP after 6 hours that is dependent on ATG5, Dnm1, Pep4, etc. Can the authors demonstrate the appearance of MDCs without expressing any GFP tags and instead monitor known outer membrane cargoes? In the case of Hap4 expression, the proteomics identifies some very highly induced mitochondrial proteins, and there surely must be some with antibodies that can detect the protein by IF and Western blot.

(2) There is a very unexpected ~10X increased in a sporulation factor SPO21 upon induction of Hap4. I see no evidence of sporulation, and it's not long enough for stationary phase. Is the increased mitochondrial biogenesis driving a specific metabolic state of these cells that is signaling to other biology?

(3) It is important to understand the kinetics and stoichiometry of outer membrane loading that drives MDCs, and their transit to the vacuole. This is why it would be highly informative to monitor some endogenous cargoes (previous point). In the review the authors cite (NRMBC, Pfanner lab 2019), it was stated that the import machinery is not generally increased upon metabolic induction of mitochondrial gene expression. Therefore, (pre-MDCs) the field concluded that the import machinery has a very high capacity for the rapid biogenesis of newly synthesized proteins, along with regulation through the phosphorylation of import receptors (ie; the work of Meisenger). Consistent with this, the Hap1 proteomics did not show any increases in the core import machinery, while ETC subunits and a large swath of mitochondrial proteins were elevated over 2-fold (I looked carefully through the Excel sheet). Since MDCs are induced transiently about 2 hours after glucose deprivation, and fully dependent on de-repression of Mig1, the authors are right to imply that this is coupled to the import of newly synthesized proteins.

However, it seems to me that MDCs are being formed at very early stages of mitochondrial protein expression, not after they have necessarily "overloaded" the outer membrane. The Hap4 proteomics after 3.5hr of induction would suggest that the bulk of the mitochondrial proteins have been successfully inserted (no import failure) and are likely already functional (metabolizing). I'm trying to understand the percentage of the proteins that would be incorporated within MDCs, as the mitochondria appear to handle the bulk of their newly inserted proteins without issue. How can the authors adapt their "free GFP" assay to understand the stoichiometry of the transport of endogenous, newly imported outer membrane proteins to the vacuole?

(4) As a last theoretical point for discussion: Can the authors exclude that MDCs are not functional or play a signaling role? Given the emerging work on SPOTs (Lena Pernas), and from the new evidence from Craig Thompson's lab that there can be very specific functional mitochondria (oxidizing vs reducing), it is possible that MDCs are not simply there to be degraded. They last at least 3 hours, which is a long time for yeast (budding cycle 90 min, 3 hours in glucose deprivation). Taking the data presented here very objectively, there is no direct evidence that the cargoes within MDVs reflect any failure to import, or that they are damaged in any way. The deletions of Tom70/71 have way too many pleotropic effects and essentially demonstrate only that the MDC cargoes came from the mitochondria. It could be helpful if the discussion also positioned these MDC mechanisms within the context of other aspects of selective mitochondrial-related compartments that have been emerging in the literature.

Reviewer #3 (Public review):

Summary:

In this manuscript, Price et al. report the physiological conditions and proteins involved in the formation of mitochondria-derived compartments (MDCs), specialized domains exclusively containing outer mitochondrial membrane (OMM) proteins, in budding yeast. Hughes and his colleagues have previously established MDCs as unique multilamellar membrane structures derived from the OMM that arise with both mitochondrial metabolic perturbation and hydrophobic OMM protein load. Whether cells undergo MDC formation in response to physiological changes in mitochondrial biogenesis remains to be explored. In this study, the authors sought to test if glucose restriction, carbon-source switching, and salt stress can induce MDC formation, and found that these situations, which naturally promote acute mitochondrial biogenesis concomitantly with metabolic transitions, trigger MDC induction. Under these conditions, loss of Snf1, an AMP-activated protein kinase that facilitates mitochondrial biogenesis under metabolic stress, almost completely abolished MDC formation. Snf1 induces MDC induction under metabolic stress via phosphorylating (suppressing) Mig1, a transcriptional repressor of mitochondrial biogenesis. Consistent with this idea, loss of Mig1 mostly rescues MDC formation under glucose restriction or salt stress in cells lacking Snf1. The authors further found that acute induction of Hap4, a core activator of mitochondrial biogenesis, is sufficient to trigger MDC formation even without metabolic stress. Finally, cells lacking Tom70 and Tom71, protein receptors of the TOM (translocase of the outer membrane) complex that mediate targeting of hydrophobic mitochondrial proteins, almost failed to form MDCs under glucose restriction. Correctively, the authors propose that MDCs act in the reduction of OMM protein load upon metabolic stress-induced acute mitochondrial biogenesis.

Strengths:

The experiments for this study are well-designed, and the resulting data are mostly convincing, with proper controls and significant statistics to support their conclusions. The paper potentially provides new insights into the physiology of MDC formation.

Weaknesses:

There are only a few new mechanistic advancements in this paper.