Mitochondria–membranous organelle contacts at a glance Open Access

In the 1950s, the association between mitochondria and the endoplasmic reticulum (ER) was first described (Bernhard et al., 1952; Bernhard and Rouiller, 1956), marking the beginning of the exploration of mitochondria–organelle membrane contact sites (MCSs), or mitochondria-associated membranes (MAMs), which are specialized areas where the outer mitochondrial membrane (OMM) is juxtaposed to the membrane of a different organelle, without fusion of the two heterotypic membranes. Two decades after their discovery, these initial observations of mitochondria–ER contacts were supported by biochemical evidence (Lewis and Tata, 1973; Shore and Tata, 1977), and in the 1990s their role in lipid and Ca²⁺ exchange was functionally annotated (Rizzuto et al., 1998; Vance, 1990). New techniques have helped to identify the protein–protein and protein–lipid interactions that facilitate the tethering of the two membranes, leading to the discovery and characterization of new mitochondrial contacts. Currently, mitochondria are considered ‘social’ organelles, communicating with almost every other organelle through physical contacts, which have been implicated in several cellular and pathophysiological processes (Perrone et al., 2020; Zung and Schuldiner, 2020).

The various interorganellar mitochondrial contacts can differ in number, length, protein composition and the distance of the gap between membranes, both among different cell types and among contacts in the same cell. Usually, the distance between two tethered membranes ranges from 10 nm to 80 nm, reaching over 300 nm in certain cases (Giacomello and Pellegrini, 2016). More than 5000 different proteins have been reported to reside in these compartments (Pan et al., 2024). These proteins fall into one or more of four major categories: tethers, which are proteins that physically bridge the two membranes; functional proteins, which facilitate specific functions at MAMs; regulator proteins, which influence the extent and distance of the contact site and/or the function of the other proteins residing there; and sorter or recruiter proteins, which affect the lipidome and proteome of the MAM (Scorrano et al., 2019). Table 1 lists some of the known functional proteins, regulators and sorting proteins found in mitochondrial MCSs. Although operational criteria have been identified to determine whether proteins residing at interorganellar interfaces function as tethers (Scorrano et al., 2019), defining molecularly whether such proteins are true tethers remains challenging. Measuring tethering forces is not straightforward, and many tethers and tethering complexes appear to be redundant, further complicating these types of analyses at the whole-cell level. Moreover, mitochondrial MCSs adapt to metabolic transitions and environmental cues (Casas-Martinez et al., 2024; Gordaliza-Alaguero et al., 2019), supporting the notion that they are dynamic structures that respond to changes in cellular function to regulate key processes.

In this Cell Science at a Glance article and the accompanying poster, we summarize known tethers found between mammalian mitochondria and other membranous organelles, their regulators and their main roles in cell physiology.

Mitochondria–ER contacts (MERCs), perhaps the best characterized mitochondrial MCSs, are crucial hubs for lipid trafficking and Ca²⁺ flux in physiology and disease (de Ridder et al., 2023; Loncke et al., 2021; Sassano et al., 2022; Vance, 2020), and they also play roles in aspects of protein homeostasis and quality control, stress responses, control of cell death, and organelle dynamics (see poster).

The sorting protein PACS-2 was the first molecular regulator of MERCs to be discovered (Simmen et al., 2005). PACS-2 participates in Ca²⁺ transfer, which might depend on PACS-2 protein-sorting function: the Ca²⁺ regulatory protein calnexin (CNX) as well as polycystin-2, which interacts with the ER Ca²⁺-release channel inositol 1,4,5-trisphosphate receptor (IP₃R; herein referring collectively to ITPR1, ITPR2 and ITPR3) and mitofusin 2 (MFN2), are both cargos of PACS-2, which controls their distribution to MAMs (Kuo et al., 2019; Myhill et al., 2008; Sammels et al., 2010). In addition, PACS-2 regulates the levels of some lipid-synthesizing enzymes, such as PSS1 (also known as PTDSS1) and FACL4 (also known as ACSL4), at MAMs through a still unknown mechanism (Simmen et al., 2005). PACS-2 silencing results in cleavage of the integral ER membrane protein BAP31 (also known as BCAP31) to p20BAP31, followed by mitochondrial fragmentation and uncoupling from the ER (Simmen et al., 2005). BAP31 has more recently been shown to form a tethering complex with the mitochondrial fission protein FIS1 that acts as a platform for procaspase 8 activation at MAMs and propagates the apoptotic signal from mitochondria to ER (Iwasawa et al., 2011). BAP31 also plays a role in targeting of mitochondrial precursor proteins through an ER stress-sensitive interaction with the mitochondrial import receptor subunit TOMM40 at MAMs. This complex favors translocation of NADH dehydrogenase enzymes NDUFS4 and NDUFB11 to mitochondria, thereby regulating complex I activity (Namba, 2019). Thus, the role of PACS-2 in controlling the extent of MERCs appears indirect (that is, not mediated by an interaction in trans with another protein on the OMM).

In 2006, a tripartite complex formed by the OMM voltage-dependent channel VDAC1 with IP₃R, stabilized by the OMM-associated fraction of the chaperone GRP75 (also known as HSPA9), was described (Szabadkai et al., 2006) (see poster). All IP₃R subtypes localize at MERCs (Bartok et al., 2019) and are tightly controlled by many MERC-residing proteins (Table 1). IP₃Rs are central in Ca²⁺ transfer between compartments but also play a Ca²⁺-independent structural role in forming MERCs (Katona et al., 2022). Opening of IP₃R on the ER leads to efflux of Ca²⁺, which is then taken up into the mitochondrial matrix via VDAC1 in the OMM and the mitochondrial Ca²⁺ uniporter (MCU) in the inner mitochondrial membrane (IMM) (de Stefani et al., 2011; Rizzuto et al., 1998). These proteins, strategically placed at MERCs, provide the spatial unit that makes mitochondrial Ca²⁺ uptake possible. Indeed, the Ca²⁺ affinity of the MCU is more than an order of magnitude higher than bulk cytosolic Ca²⁺ concentrations. MERCs thus create a separate microdomain of high Ca²⁺ concentration at levels permissive for uptake by mitochondria (Csordas et al., 2010; Giacomello et al., 2010). Despite the importance of this complex, as with PACS-2, it also has no direct effect on the proximity of the two tethered organelles.

MFN2 was the first bona fide structural mitochondria–ER tether identified in mammalian cells. A fraction of MFN2, which is a dynamin-related protein that promotes mitochondrial fusion, localizes in MAMs and engages in homotypic or heterotypic interactions with other MFN2 molecules or mitofusin 1 (MFN1) on the OMM (de Brito and Scorrano, 2008). In addition, deletion of MFN2 increases the distance between ER and mitochondria and diminishes agonist-stimulated ER-to-mitochondria Ca²⁺ transfer. Whereas some studies have challenged the role of MFN2 as a tether (Cosson et al., 2012; Filadi et al., 2015; Leal et al., 2016), many others support this function (Chen et al., 2012; Göbel et al., 2020; Li et al., 2015; Liao et al., 2024; Naon et al., 2016; Schneeberger et al., 2013; Sebastian et al., 2012), highlighting the complexity of these interorganellar interfaces. MFN2-mediated tethering is regulated by a variety of post-translational modifications and protein interactions, further supporting its essential role in maintaining MERCs (Table 1). However, how such an important cellular function could be accomplished by an ectopically localized fraction of primarily mitochondrially localized MFN2 remained unclear until the recent discovery of purely extramitochondrial splice variants of the MFN2 gene, dubbed ERMIN2 and ERMIT2 (Naon et al., 2023) (see poster). ERMIN2 retains the part of the MFN2 GTPase domain essential for promoting mitochondrial fusion and can restore normal ER morphology in MFN2-deficient cells, whereas ERMIT2, which lacks the GTPase domain but retains the CC2 domain essential for interactions with the CC1 or CC2 domains of mitochondrial mitofusins, tethers the ER to mitochondria (Naon et al., 2023).

Several other physical tethers at MERCs have been identified and characterized. The interaction between VAPB on the ER and PTPIP51 (also known as RMDN3) on the OMM establishes a tethering complex that regulates Ca²⁺ homeostasis (de Vos et al., 2012). VAPB and PTPIP51 are often found at interorganellar interfaces at synapses, and their depletion reduces synaptic activity, a pathological feature in several neurodegenerative diseases (Gomez-Suaga et al., 2019). PTPIP51 phosphorylation and subsequent binding to VAPB is increased by mitochondrial reactive oxygen species (ROS), and PTPIP51 performs an antioxidant function by removing lipid radicals from mitochondria via MERCs through the lipid radical-transfer activity of its tetratricopeptide repeat (TPR) domain (Shiiba et al., 2025). Recently, the lipid transfer protein (LTP) VPS13D has been shown to bind to VAPA and VAPB and tether the ER to mitochondria in a manner dependent on mitochondrial Rho GTPase (Miro; herein referring to Miro1 and Miro2, also known as RHOT1 and RHOT2, respectively) (Guillen-Samander et al., 2021). PTPIP51 also interacts with the LTPs ORP5 (OSBPL5) and ORP8 (OSBPL8) at MAMs, and loss of these proteins leads to mitochondrial defects (Galmes et al., 2016). ORP5 and ORP8 were initially described at ER–plasma membrane (PM) contacts, where they transfer phosphatidylserine (PS) from the cortical ER to the PM in counter-exchange for phosphatidylinositol-4-phosphate (PI4P) from the PM (Chung et al., 2015). In MAMs, ORP5 and ORP8 have been shown to mediate the transport of PS from the ER to mitochondria by cooperating with two protein complexes that bridge the IMM and OMM, the mitochondrial intermembrane space bridging (MIB) complex and the mitochondrial contact sites and cristae junction organizing system (MICOS) complex (Monteiro-Cardoso et al., 2022), and to control lipid droplet (LD) biogenesis by recruiting seipin at MAMs (Guyard et al., 2022). Because the TPR domain of PTPIP51 can transfer phosphatidic acid (PA) (Yeo et al., 2021), its interaction with ORP5 and ORP8 might also mediate PS trafficking.

PERK (also known as EIF2AK3), an effector of the unfolded protein response (UPR), recruits and scaffolds the ER protein ESYT1 at MERCs (Sassano et al., 2023) to regulate phospholipid (PL) transfer, and disruption of this interaction or the lipid transfer function of ESYT1 causes mitochondrial respiration and lipid transfer defects. Depletion of PERK but not ESYT1 has been shown to reduce the number of MERCs per cell. ESYT1 forms a tethering complex with the OMM protein SYNJ2BP (Janer et al., 2024). Depletion of either protein has been shown to reduce the number and length of MERCs and impair mitochondrial Ca²⁺ uptake and lipid homeostasis, whereas overexpression of SYNJ2BP promotes MERC formation. SYNJ2BP also interacts with the ribosome-binding protein RRBP1, forming a mitochondria–rough ER contact of ∼45 nm (Hung et al., 2017); however, whether this represents a bona fide tethering complex remains to be ascertained.

Recently, PERK has also been found to interact with ATAD3A (Brar et al., 2024), an AAA+ ATPase that spans the intermembrane space (IMS), with its C terminus in the matrix and N terminus anchored in the OMM (Gilquin et al., 2010). This interaction tethers the ER to mitochondria during ER stress and protects key mitochondrial proteins from PERK-mediated translational repression. In breast cancer cells, a complex containing ATAD3A, WASF3 and BIP (HSPA5) stabilizes WASF3 in the OMM and creates a bridge to the ER (Teng et al., 2016). ATAD3A supports several MERC-related functions, such as steroid and lipid synthesis, mitochondrial DNA maintenance, and regulation of mitochondrial morphology (Gerhold et al., 2015; Gilquin et al., 2010; He et al., 2007; Issop et al., 2015), as well as regulation of mitochondrial ribosome assembly and stability, and cristae morphology (Rigoni et al., 2025). It is therefore possible that all these functions are coordinated by ATAD3A-mediated tethering at MERCs.

Another UPR and ER stress regulator localized at MAMs is PIGBOS (also known as PIGBOS1), an OMM microprotein (a small peptide or protein translated from a small open reading frame). Interaction of PIGBOS with the ER protein CLCC1 at MAMs is necessary for the function of PIGBOS in the regulation of the UPR and ER stress-induced apoptosis (Chu et al., 2019). However, depletion or overexpression of PIGBOS has no effect on MERC formation.

The ER transmembrane protein PDZD8 has also been identified as a MERC tether (Hirabayashi et al., 2017). PDZD8 contains a synaptotagmin-like mitochondrial lipid-binding protein (SMP) domain that is functionally orthologous to yeast Mmm1, a subunit of the ER–mitochondria encounter structure (ERMES) that tethers mitochondria to ER in yeast (Kornmann et al., 2009). Its long-sought tethering partner on the OMM has recently been found to be FKBP8 (Nakamura et al., 2025). PDZD8 influences mitochondrial morphology in a FKBP8-dependent manner, suggesting that this tethering complex regulates mammalian mitochondrial shape. In neurons, PDZD8 is required for Ca²⁺ uptake by mitochondria after synaptic-induced Ca²⁺ release from the ER (Hirabayashi et al., 2017). Another potential MERC tether is REEP1, which contains a receptor expression-enhancing protein (REEP) homology domain, shown to insert into membranes as a hairpin and modulate membrane curvature. REEP1 also contains subdomains for mitochondrial and ER localization and is found in MAMs, where it facilitates MERC formation (Lim et al., 2015). Lastly, many MERC tethering proteins also participate in formation of tripartite contacts between mitochondria, ER and other organelles (Box 1; see poster).

Mitochondria and peroxisomes are functionally interconnected, co-participating in α- and β-oxidation, bile acid synthesis, steroid biosynthesis, ROS metabolism, glyoxylate detoxification, and anti-viral signaling and response (Wanders et al., 2023). Like mitochondria, peroxisomes are morphologically dynamic organelles, and both organelles share core fission machinery: FIS1, mitochondrial fission factor (MFF), ganglioside-induced differentiation-associated protein 1 (GDAP1) and dynamin-related protein 1 (DRP1, also known as DNM1L) (Subramani et al., 2023). The first mitochondria–peroxisome tethering complex to be identified was found in Leydig cells and includes the acyl-CoA-binding protein ACBD2/ECI2 isoform A (encoded by ECI2), which localizes in both peroxisomes and mitochondria via competitive binding between PEX5 on peroxisomes and TOMM20 on mitochondria, bringing the two organelles close (see poster) (Fan et al., 2016). Ectopic expression of ACBD2/ECI2 isoform A in MA-10 cells, derived from a mouse Leydig cell tumor, increases steroid biosynthesis, suggesting that mitochondria source the cholesterol required for this process via peroxisome contacts.

PEX11β (also known as PEX11B), a key regulator of peroxisomal membrane dynamics and division (Koch et al., 2010; Schrader et al., 2022), has also been identified as a potential mitochondria–peroxisome tether (Kustatscher et al., 2019). PEX11β is co-regulated with subunits of the mitochondrial ATP synthase and other components of the electron transport chain (Kustatscher et al., 2019). Its expression induces the formation of peroxisomal membrane protrusions, facilitating interactions with mitochondria. However, it remains unknown whether PEX11β tethers the membranes of the two organelles or what the identity of its partners on mitochondria might be.

Mitofusins also contribute to co-clustering of mitochondria and peroxisomes as a potential tether between these organelles (Huo et al., 2022). MFN2 is enriched at mitochondria–peroxisome contacts, and upregulation of MFN2 via leflunomide treatment stimulates contact formation, whereas the expression of a dominant-negative MFN2 mutant inhibits it. In yeast, modulation of levels of the mitofusin ortholog Fzo1 by the ubiquitin–proteasome system and the desaturation status of fatty acids (FAs) regulates the formation of homotypic Fzo1-mediated mitochondria–peroxisome contacts to facilitate the transfer of peroxisomal citrate to mitochondria (Alsayyah et al., 2024).

Mitochondria and endolysosomes also exhibit crosstalk to regulate key cellular processes, including autophagy, nutrient and energy homeostasis, stress and immune responses, proliferation, apoptosis, cell death signaling, and iron and heme metabolism (see poster), as evidenced by the impact of mitochondrial defects on lysosomal biogenesis and function, and vice versa (Deus et al., 2020). The physical contacts between mitochondria and lysosomes (mitochondria–lysosome contacts, MLCs) in mammals display a distance of ∼10 nm (Wong et al., 2018) and vary in duration (Han et al., 2017). Deregulation of MLCs is involved in the pathogenesis of several diseases, including Parkinson's disease, Charcot–Marie–Tooth (CMT) disease and lysosomal storage disorders such as mucolipidosis type IV (Cantarero et al., 2021; Kim et al., 2021; Peng et al., 2020; Rizzollo and Agostinis, 2025; Wong et al., 2019; Xie et al., 2024).

The small GTPase Rab7 (herein referring to both Rab7A and Rab7B) promotes MLC formation in its lysosomal GTP-bound form, and tethering to mitochondria is probably mediated through Rab7 effector proteins (Wong et al., 2018). Untethering is mediated by Rab7 GTP hydrolysis stimulated by TBC1D15, a Rab7 GTP-hydrolysis-activating protein (GAP) that is recruited to mitochondria by FIS1 (Onoue et al., 2013). This mitochondria–lysosome tethering thus regulates mitochondrial fission (Wong et al., 2018). Although TBC1D15 has no known role in contact formation, reducing its activity has proven beneficial in disease models (Kim et al., 2021; Sun et al., 2022; Yu et al., 2020). Active Rab7 interacts with the NPC1 cholesterol transporter and stimulates lysosomal cholesterol export, a process regulated by the trimeric Mon1-Ccz1-C18orf8 (MCC) complex, which acts as a Rab7 guanine-nucleotide-exchange factor (GEF) (van den Boomen et al., 2020). Downstream of mTOR complex 1 (mTORC1) signaling, Rab7A controls cholesterol trafficking from lysosomes to mitochondria by regulating the interaction between translocator protein (TSPO) on the OMM and NPC1 on lysosomes (Lin et al., 2023). Interestingly, in the absence of NPC1, lysosomes form extensive contacts with mitochondria that are dependent on the LTP STARD3 (Hoglinger et al., 2019). In hepatocellular carcinoma, Rab7 interacts with a phosphorylated form of DRP1 (p-DRP1^S616) at MLCs, triggers PINK1–parkin (PRKN)-dependent mitophagy (the selective degradation of mitochondria via autophagy) and promotes cell survival. The PP2A phosphatase subunit B56γ (also known as PPP2R5C) negatively regulates MLC formation via dephosphorylation of p-DRP1^S616, leading to apoptosis and sensitization of cells to chemotherapy (Che et al., 2022). Interestingly, the mouse DRP1_ABCD isoform, which is enriched in brain and is involved in mitochondrial and peroxisomal division, localizes to the interface between mitochondria and lysosomes or late endosomes without participating in contact formation (Itoh et al., 2018). Rab7A is also found on late endosomes that interact with RNA granules, which often dock at mitochondria in axons, where they control the synthesis of pro-survival proteins (Cioni et al., 2019). These RNA-bearing late endosomes also associate with ribosomes, and disruption of their function affects the local translation efficiency of mRNAs essential for mitochondrial function, but the exact role of these mitochondria–endosome contacts is not known.

MLCs also modulate intracellular Ca²⁺ fluxes via the direct transfer of lysosomal Ca²⁺ to mitochondria, mediated by the interaction of the nonselective cation channel TRPML1 (also known as MCOLN1) on lysosomes with VDAC1 on the OMM (Peng et al., 2020). However, whether TRPML1 and VDAC1 act as a tether is unclear because expression of a dominant-negative nonconducting TRPML1 pore mutant increases the incidence of stable MLCs (Peng et al., 2020).

GDAP1, an atypical glutathione S-transferase (GST) located on the OMM and at MAMs, can interact with the lysosome-associated membrane protein LAMP1, forming a redox-sensitive tether between mitochondria and lysosomes in neurons (Cantarero et al., 2021). Depletion of GDAP1 reduces MLCs and glutathione levels, resulting in lysosomal and mitochondrial network abnormalities. Pathogenic variants of GDAP1 that cause CMT disease display different effects on MLCs as well as on cellular and mitochondrial Ca²⁺ levels, suggesting a potential role for MLCs in the observed differences in severity between dominant and recessive forms (Cantarero et al., 2023).

Lastly, in erythroid cells, MLCs are assembled by MFN2 to facilitate transferrin receptor 2 (TfR2)-dependent transferrin (Tf) delivery to lysosomes. MFN2 knockdown reduces MLC numbers, total heme content and erythroid differentiation (Khalil et al., 2017).

Melanosomes are organelles responsible for the synthesis, storage and transport of melanin. MFN2 tethers mitochondria to melanosomes through fibrillar bridges (Daniele et al., 2014) (see poster), and its knockdown reduces both formation of mitochondria–melanosome contacts and activation of melanogenesis. Close mitochondria–melanosome contact might facilitate ATP supply for melanosome biogenesis. However, a recent study has demonstrated an opposing role for MFN2 as a negative regulator of melanogenesis (Tanwar et al., 2022).

Mitochondrial–lipid droplet contacts (MLDCs) form to channel FAs between the two organelles or to channel ATP from mitochondria to LDs to support lipid synthesis. On LDs, perilipins play a crucial role in establishing MLDCs (see poster). The first tether described to link LDs to mitochondria and regulate oxidative LD hydrolysis and FA flux was perilipin 5 (Wang et al., 2011a), the C terminus of which is necessary and sufficient to mediate MLDCs. In myoblasts, perilipin 5 interacts with the acyl-CoA synthetase FATP4 (also known as SLC27A4) on the OMM, and this interaction is specifically required for FA channeling (Miner et al., 2023). Upon starvation in skeletal muscle, perilipin 5 interacts with GTP-bound, active Rab8A and recruits adipose triglyceride lipase (ATGL, also known as PNPLA2) to promote long-chain FA (LCFA) transfer into mitochondria for β-oxidation (Ouyang et al., 2023). Perilipin 5 and another member of the perilipin family, perilipin 1, interact with MFN2 in human bone osteosarcoma epithelial cells (U2OS cells; Miner et al., 2023) and adipose tissue, forming contacts that assemble in response to adrenergic stimulation and couple triglyceride hydrolysis to FA oxidation (Boutant et al., 2017). In hepatocytes exposed to aflatoxin B1, perilipin 2 forms a tethering complex with mitochondria-translocated p53 (TP53) that inhibits lysosome-associated lipophagy, leading to lipid accumulation and lipotoxicity (Che et al., 2023). Removal of perilipins via chaperone-mediated autophagy is indeed necessary for initiation of lipolysis by lipases or lipophagy (Kaushik and Cuervo, 2015).

The synaptosomal-associated protein receptor (SNARE) protein SNAP23, found on LDs, modulates both MLDC formation and β-oxidation in NIH3T3 fibroblasts. However, these two effects appear not to be linked, and the molecular mechanism by which SNAP23 functions is still unknown (Jagerstrom et al., 2009). During glucose deprivation, when FA oxidization is stimulated, SNAP23 and VAMP4 associate with ACSL1 on mitochondria, tethering LDs to the OMM (Young et al., 2018). ACSL1 has been shown to be recruited to mitochondria by TBK1, which operates as a molecular rheostat to control hepatic FA oxidation (Huh et al., 2020). However, SNAP23 tethering is not essential for the flow of FAs from LDs to mitochondria. FA transfer from LDs also occurs via membrane remodeling facilitated through association between VPS13D and the endosomal sorting complex required for transport (ESCRT) protein TSG101 at MLDCs: the N-terminal domain of VPS13D targets mitochondria, whereas the C terminus targets LDs and the Vps13 adaptor-binding (VAB) domain interacts with TSG101 (Wang et al., 2021b). Suppression of VPS13D, but not TSG101, significantly reduces the incidence of MLDCs (Wang et al., 2021b).

Physical tethering between mitochondria and the nucleus participates in regulation of cholesterol trafficking, protein distribution and transcriptional activity (see poster). During mitochondrial retrograde response (MRR) signaling, mitochondria–nucleus tethering occurs through the interaction of TSPO, which recruits ACBD3 to mitochondria, and protein kinase A (PKA), which also interacts with the A-kinase anchoring protein AKAP95 (also known as AKAP8) on the nuclear membrane (Desai et al., 2020). In breast cancer cells, formation of these nucleus-associated membranes facilitates trafficking of cholesterol to the nucleus, sustaining a pro-survival response by blocking NF-κB deacetylation. In response to proliferative stimuli, MFN2 also tethers mitochondria to the nucleus in a noncanonical pathway for importing the IMM pyruvate dehydrogenase complex (PDC) into the nucleus. Nuclear PDC retains its enzymatic activity, interacts with the nuclear lamina matrix protein lamin A and possibly participates in protein acetylation at nucleoplasmic hubs (Zervopoulos et al., 2022).

In the past two decades, our understanding of organellar communication and functional coordination has significantly evolved. Organelles, once considered isolated units, are now recognized as components of a physically and functionally interconnected dynamic network. However, despite considerable advances, several challenges remain. Foremost is the development of efficient, user-friendly techniques and tools for visualizing contact sites and identifying the molecular tethers that stabilize them. Following the consensus on the ontology of tethers reached in 2019 (Scorrano et al., 2019), the field must also reach agreement on methodologies and data interpretation.

As additional mitochondrial contacts with various organelles are discovered, the gaps in our knowledge regarding the (co-)regulation of tethering machineries become more apparent. Recent studies have revised our notion of organellar tethers from stable, static bridges to dynamic, orchestrated connections applied by many functional interactions in response to specific stimuli. Stimulus-triggered rearrangements in mitochondrial contacts might involve single or multiple tethers, last for differing durations, and lead to distinct outcomes depending on the duration or intensity of stimuli. This structural and temporal flexibility ensures that MCSs can adapt to changing needs without compromising cell function. For example, a transient increase in MERCs facilitates Ca²⁺ transfer from the ER to mitochondria, stimulating metabolic enzymes and ATP production, whereas prolonged tethering might lead to Ca²⁺ overload and apoptosis. Similarly, transient ER stress increases MERCs and leads to enhanced ATP production (Bravo et al., 2011), whereas chronic UPR activation has the opposite effect (Wang et al., 2011b).

The plasticity of MCSs in response to nutrient availability is particularly well documented (Benador et al., 2018; Boutant et al., 2017; Honscher et al., 2014; Sood et al., 2014; Theurey et al., 2016; Young et al., 2018) and is typically periodic – MCSs adjust their number, extent and structure to meet metabolic demands before returning to a steady state. These shifts occur on the order of hours (Sood et al., 2014), in contrast to Ca²⁺-related remodeling, which can occur in seconds (Yi et al., 2004). High-speed molecular tracking of VAPB has recently demonstrated the dynamic nature of mitochondrial contacts by revealing rapid diffusion of VAPB molecules in and out of a long-lasting contact site that quickly remodels in response to metabolic cues (Obara et al., 2024). In addition, LTPs such as ESYT1, ORP5 and ORP8 form bridges by binding to specific lipids in organelle membranes (Galmes et al., 2016; Saheki and De Camilli, 2017) or by tethering different organelles to allow them to utilize specific fuels, as in the case of VPS13D (Wang et al., 2021b), suggesting that precise tethering mechanisms function to meet metabolic demands. Mitochondrial contacts are also remodeled in response to physiological processes such as mitosis (Yu et al., 2024; Zhao et al., 2024), and these rearrangements are transient, lasting only until the completion of the process. However, key questions regarding MCS dynamics remain, including how the proteins localized in these specialized subcompartments coordinate to maintain or dissolve contacts and how universal tethers, such as MFN2, are regulated.

Finally, perturbations in mitochondrial contacts and their tethering protein machinery have been linked to multiple diseases (Cisneros et al., 2022; Fan and Tan, 2024; Makio and Simmen, 2024; Morcillo et al., 2024; Rizzollo and Agostinis, 2025; Wilson and Metzakopian, 2021). Clarifying the molecular mechanisms and identifying the specific players involved in these pathological conditions is crucial for developing effective treatments. The challenge remains to specifically target these perturbations without disrupting the processes that occur at contact sites.

We thank the members of the Scorrano lab for helpful discussions.