Genetically Encoded Tools to Monitor and Interrogate Membrane Contact Sites

Abstract

Membrane contact sites (MCSs) are dynamic subcellular compartments formed between organelles that coordinate diverse aspects of cellular communication, including signaling, metabolism, and membrane organization. Tools capable of monitoring and controlling the spatially localized and dynamic properties of MCSs are needed to dissect their regulatory mechanisms and physiological roles. Recent advances in protein engineering have begun to address this need. Proximity-based reporters, chemogenetic approaches, and optogenetic systems have been developed to enable the visualization, interrogation, and manipulation of inter-organelle contacts with improved spatial and temporal precision. This minireview highlights key developments in these molecular toolkits and their representative applications in studying MCS biology. These approaches provide new insights into organelle crosstalk and may inform future therapeutic strategies targeting MCSs.

Introduction

Membrane contact sites (MCSs) are specialized subcellular compartments formed by the close apposition of membranes from distinct organelles without membrane fusion, typically separated by a gap of approximately 10–35 nm (Scorrano et al., 2019). These inter-organellar juxtapositions serve as key hubs for cellular communication, coordinating the balanced and interdependent activities of organelles through structural tethering and regulated metabolite exchanges (Figure 1) (Kim et al., 2022). Increasing evidence has demonstrated the critical roles of MCS in regulating cellular events, including lipid metabolism, ion transport, organelle biogenesis, cell fate determination, and organelle trafficking and positioning (Voeltz et al., 2024). Interest in how altered or dysregulated formation or disassembly of membrane contact sites has expanded markedly over the past decade, with functional links emerging to neurological disorders (Kim et al., 2022), infectious diseases (Bonde and Herhaus, 2025), metabolic diseases (Yang et al., 2020), and cancers (Gil-Hernandez et al., 2020; An et al., 2024). Despite these advances, the molecular mechanisms and functional consequences of signal and metabolite exchange at some MCSs remain incompletely understood, partially owing to the high plasticity, rapid dynamics, and inherent complexity of these subcellular structures, coupled with methodological limitations in visualizing and perturbing these structures with spatiotemporal precision (Chang et al., 2017).

Figure 1.

Genetically encoded probes and actuators for studying membrane contact sites (MCSs). Schematic overview of inter-organelle membrane contact sites and representative molecular toolkits used to visualize, probe, or interrogate inter-organelle interfaces. Highlighted approaches include including proximity-based reporters, chemogenetic systems, and optogenetic actuators for controlling organelle communication and signaling at MCSs. Abbreviations: PM, plasma membrane; ER, endoplasmic reticulum; MCSs, membrane contact sites; CsFiND, complementation assay using fusion of split-GFP and TurboID; CalFlux VTN, calcium flux composed of Venus, Troponin, and NanoLuc; MAM, mitochondrial associated membrane; BiFCPL, biomolecular fluorescence complementation-based proximity labeling; PRINCESS, probe for interorganelle Ca²⁺ exchange sites based on splitFAST; FABCON, fluorogen-activated bimolecular complementation at contact sites; LaBeRling, lamin B receptor and a generic anchor protein on the partner organelle; MAPPER, membrane-attached peripheral ER; LiMETER, light-inducible membrane-tethered peripheral ER; LIT, light-inducible tethering; Opto-MLC, optogenetic mitochondria-lysosome contacts; LDLRC, light-controlled responsive crosslinker.

Microscopy-based techniques represent the classical approach for studying MCS and have long served as the cornerstone for their identification and characterization (Khaddaj and Kukulski, 2023). While electron microscopy (EM) remains the gold standard for visualizing MCS at nanometer resolution, fluorescence microscopy-based methods retain a critical role by enabling rapid imaging, deconvolution, and 3D reconstruction (Sarhadi et al., 2023). The subsequent development of super-resolution microscopy, including stimulated emission depletion (STED), structured illumination microscopy (SIM), and stochastic optical reconstruction microscopy (STORM), has advanced the accessible spatial resolution for MCS analysis (Hell and Wichmann, 1994; Rust et al., 2006; Gustafsson et al., 2008; Sahl et al., 2017; Sarhadi et al., 2023). In addition, the landscape ExM (land-ExM) technology further fills a critical gap by enabling simultaneous mapping of protein and lipid ultrastructure with a high signal-to-noise ratio (Zhuang et al., 2025). More recently, a hybrid approach combining SIM and the in situ proximity ligation assay (PLA) has been developed to characterize the subtypes of mitochondria-associated membranes (MAM) based on their spatial relationship to lipid droplets (LDs). This approach distinguishes ER-mitochondria contacts located within ∼380 nm of LDs from more conventional, or basic, MAM structures (Palard et al., 2025). Following inducible adipocyte-specific depletion of seipin in a mouse model, Palard and colleagues observed a selective reduction in ER-mitochondrial tethering proximal to LDs in adipose tissue, whereas basic MAM contacts remained largely unchanged. This finding was further validated by focused Ion Beam-Scanning Eletron Microscopy (FIB-SEM), which revealed pronounced spatial disorganization of ER-mitochondria-LD interfaces in the absence of seipin (Palard et al., 2025). This complementary strategy integrates the ultrastructural resolution of microscopy-based techniques with the protein-level specificity of PLA, thus offering a robust platform for MCS research (Palard et al., 2025). Despite these advances, most routine microscopy-based approaches remain constrained by limited spatiotemporal resolution and, in many cases, by their inability to trace dynamic processes in living cells, making it challenging to distinguish bona fide MCS assembly from mere membrane proximity. These limitations emphasize the urgent need to develop complementary techniques for studying MCS dynamics.

Although several recent reviews have covered induced proximity tools for manipulating MCS, few focus specifically on the design principles and integration of these tools with the complexities of cellular biology (Tei and Baskin, 2021; Nakatsu and Tsukiji, 2023; Gamuyao and Chang, 2024). In this more technology-oriented review, we summarize the major advances and recent improvements in MCSs visualization toolkits and showcase the latest progress through representative examples of how chemogenetic and optogenetic strategies have been employed to control MCS assembly and modulate diverse cellular processes.

MCS Visualization via Fluorescence Complementation

Early strategies based on the splitting and reconstitution of a fluorescent protein (FP), such as bimolecular fluorescence complementation (BiFC) and dimerization-dependent fluorescent protein (ddFP), represent classical spatial proximity-based techniques used in MCS studies (Tao et al., 2019). These methods rely on the principle that interactions between proteins of interest (POIs) bring together two non-fluorescent FP fragments into close proximity (Figure 2A), enabling fluorescence recovery upon reassembly (Hu et al., 2002). By fusing the non-fluorescent fragments to proteins selectively localized to different organellar membranes, BiFC- or ddFP-based toolkits hold strong potential for labeling and visualizing multiple MCSs in living cells (Ke et al., 2025a).

Figure 2.

Proximity-based tools for MCSs visualization and interrogation. Schematic illustrations of the basic principles and representative molecular tools used for MCS labeling and identification of MCS modulators. A). BiFC- and ddFP-based sensor. The two non-fluorescent halves of a fluorescent protein (FP) form a proximity-dependent heteromeric complex and restore the fluorescence upon membrane tethering. B). Contact-FP system to identify ER-mitochondria membrane contacts. The non-fluorescent halves are anchored to ER or outer mitochondrial membrane and can be reconstituted to generate fluorescence signals during ER-mitochondria MCS formation. C). Reporters building on SplitFAST. The non-function halves exhibit high and reversible apparent affinity towards each other in the presence of hydroxybenzylidene rhodamine (HBR) analogs (fluorogen), which show strong fluorescence when bound to reconstituted FAST. D). Engineering of FABCON, an inducible splitFAST variant. Addition of fluorogen brings the two non-function halves into proximity and thereby promotes the formation of ER-LD MCS. E). Design of PRINCESS, an inducible splitFAST system using CaM and M13 to probe Ca²⁺. F). Design of splitBioID and splitTurboID. The reassembly of the inactive fragments forms intact biotin ligase to induce biotinylation of nearby proteins. G). Design of the CsFiND system that specifically labels proteins within ER-mitochondria contact sites. Formation of ER-mitochondrial contacts leads to reassembly of GFP and TurboID, thereby biotinylating proteins in a proximity-dependent manner. H). APEX-based approaches to label proteins of interest. In the presence of hydrogen peroxide, APEX will convert biotin phenol to biotin phenol radicals to label nearby proteins. I). Scheme for the use of BiFCPL to label proteins located within LD-mitochondria contact sites. Reconstitution of intact GFP recruits the APEX2 fused with an anti-GFP nanobody, thereby enabling biotinylation to identify novel regulators of LD-mitochondria contact sites. Abbreviations: BiFC, bimolecular fluorescence complementation; ddFP, dimerization-dependent fluorescent proteins; MCS, membrane contact site; ER, endoplasmic reticulum; CYP2C2, cytochrome P450 2C1; CsFiND, complementation assay using fusion of split-GFP and TurboID; splitFAST, split fluorescence-activating and absorption-shifting tag; FABCON, fluorogen-activated bimolecular complementation at contact sites; PRINCESS, probe for interorganelle Ca²⁺ exchange sites based on splitFAST; CaM*: calmodulin mutation; Ca²⁺, calcium; OMM, outer mitochondrial membrane; CB5, cytochrome b5; 6xHp, 6x tandem repeats of M1 Spastin’s Hp motif; BiFCPL, biomolecular fluorescence complementation-based proximity labeling.

Building on this design, Contact-FP probes have been developed as a suite of ddFP-based sensors targeting multiple organelles to visualize the morphology and dynamics of MCSs (Miner et al., 2024). Organelle specificity was achieved using different anchoring protein pairs, such as cytochrome P450 2C1 (CYP2C1) and mitochondrial antiviral signaling protein (MAVS) for ER-mitochondria contact sites (Figure 2B) (Miner et al., 2024). Beyond visualization, implementation of Contact-FP at the LD-mitochondria interface was shown to induce MCS formation at high expression levels, while low expression allowed minimally perturbative visualization of endogenous membrane contacts (Miner et al., 2024). Similar strategies have been employed in the split-mitochondrial associated membrane (Split-MAM) (Novales et al., 2025), the split-GFP system for ER-mitochondria encounter structure (ERMES) (Kakimoto et al., 2018), vacuole-plasma membrane contact (vCOuPLE) (Shai et al., 2018), plasma membrane contact with lipid droplets (pCLIP) (Shai et al., 2018), peroxisome- plasma membrane contact site (PerPECs) (Shai et al., 2018), and peroxisome-vacuole contact (PerVale) (Shai et al., 2018). In these designs, split Venus was favored over split GFP to improve brightness and detection sensitivity (Shyu et al., 2006; Romei and Boxer, 2019).

Although BiFC- or ddFP- approaches were originally applied to label and visualize membrane tethering, the modular nature of these systems enables straightforward adaptation for probing dynamic signaling events localized to inter-membrane contact sites. For example, mitochondria-associated ER membrane calcium flux sensors (MAM-CalFluxVTN), composed of Venus, troponin, and NanoLuc, were developed by integrating BiFC with the genetically encoded ratiometric calcium (Ca²⁺) sensor CalFluxVTN (Yang et al., 2016). By incorporating Sac1 and AKAP as ER- and mitochondria-targeting sequences, respectively, MAM-CalFluxVTN enables selective monitoring of Ca²⁺ dynamics within ER-mitochondria contact sites (Cho et al., 2023). In addition to imaging, the compatibility of BiFC- and ddFP-based methods for large-scale screening has been demonstrated by combining a CRISPR-Cas9 pooled library with stable expression of a split mVenus reporter to identify novel regulators of ER-mitochondria interfaces (Wilson et al., 2024). A similar strategy was employed using split-GFP-based contact sites (SPLICS) reporters to examine the behavior of various MCSs (Cieri et al., 2018; Yang et al., 2018; Vallese et al., 2020; Giamogante et al., 2022; Giamogante et al., 2024; Olszakier et al., 2025; Yamaguchi et al., 2025). In addition to mammalian cells, split-FP reporters have been successfully applied in yeast to provide mechanistic insight into MCS dynamics. For example, the peroxisome-mitochondria contact reporter (PerMit) based on split-Venus was employed in high-content screening to identify tethers and regulators of PerMit contact sites, revealing previously unrecognized roles of Fzo1 and Pex34 in yeast (Shai et al., 2018). Similarly, split-GFP probes have been used to demonstrate how mitochondrial fusion and division contribute to the clustering of ERMES (Kakimoto-Takeda et al., 2022). Together, these studies illustrate how BiFC- and ddFP-based systems extend beyond static labeling to enable the identification of novel MCS players, as well as the interrogation of spatially confined signaling and dynamic subcellular organization.

Despite these advances, BiFC/ddFP-based techniques suffer from intrinsic limitations, such as spontaneous self-assembly (Kodama and Hu, 2010; Gookin and Assmann, 2014; Tebo and Gautier, 2019) and slow reconstitution kinetics (Robida and Kerppola, 2009; Miller et al., 2015), which sometimes may constrain their ability to faithfully capture the highly dynamic behavior of MCSs. These challenges prompted the development of the fluorogenic reporter, Fluorescence-Activating and absorption-Shifting Tag (FAST), which was adapted into a split configuration termed splitFAST (Tebo and Gautier, 2019). In this system, the non-fluorescent FAST fragments (NFAST and CFAST) reconstitute a functional fluorogen-binding pocket only in the presence of a fluorogen (Figure 2C), thereby yielding fluorescence with improved dynamics, reduced background, and enhanced reversibility (Tebo and Gautier, 2019). When anchored to distinct subcellular membranes, splitFAST provides a useful platform for tracking MCS assembly and disassembly with higher fidelity (Garcia Casas et al., 2024).

Building on a similar design, Fluorogen-Activated BiFC at CONtact sites (FABCON) was developed for quantitative visualization of LD-organelle contact sites (Li et al., 2024b). In FABCON, splitFAST fragments are anchored to lipid droplets using 6×tandem repeats of M1 spastin's Hp motif (6xHp), and to other organelles using cytochrome B5 for the ER, PMP34 for peroxisomes, and SYNJ2BP/OMP25 for mitochondria (Figure 2D) (Li et al., 2024b). FABCON has been reported to induce LD-ER membrane tethering within minutes upon ligand addition, with efficient and reversible disassembly following fluorogen washout, demonstrating robust temporal control in living cells (Li et al., 2024b). Additionally, a splitFAST-based probe, termed PRobe for INterorganelle Ca²⁺ Exchange Sites based on SplitFAST (PRINCESS), was developed to simultaneously identify MCSs and monitor localized Ca²⁺ dynamics (Garcia Casas et al., 2024). PRINCESS employs ER- or mitochondria-anchored non-fluorescent FAST fragments derived from a far-red-emitting splitFAST variant, fused to either the M13 peptide or calmodulin (CaM), endowing it with Ca²⁺-sensing capabilities at ER-mitochondria contact sites (Figure 2E) (Garcia Casas et al., 2024). PRINCESS is well suited for capturing transient Ca²⁺ spikes and short-lived signaling events at inter-organelle contact sites. While the splitFAST systems provide minimal perturbation of natural MCS formation during introducing contacts (Garcia Casas et al., 2024), several limitations should be considered, particularly in highly dynamic contexts. For example, despite the rapid complementation of splitFAST fragments, a potential mismatch between fluorogenic signal generation and actual contact dynamics may arise due to fluorogen-binding kinetics and delays between contact formation and detectable fluorescence. Such temporal offsets may complicate the interpretation of transient MCS assembly or disassembly events. An improved version, splitFAST2, introduced in 2024, exhibits higher dynamic range and reduced self-complementation (Rakotoarison et al., 2024). In addition, reconstitution efficiency can vary depending on expression levels, which may pose challenges for quantitative comparisons or large-scale screening applications in future MCS studies.

More recently, complementary reporter systems based on alternative enzyme complementation strategies have been developed to quantify inter-organelle contacts with improved signal-to-noise ratios. For example, the Mitochondria-Endoplasmic Reticulum contact reporter using NanoLuc Binary Technology (MERBiT) employs NanoBiT luciferase complementation to sensitively monitor ER-mitochondria contacts (Shiiba et al., 2025). Similarly, SpLacZ-MERCS reporters, based on complementation of the bacterial β-galactosidase encoded by the lacZ gene, provide quantitative detection of mitochondria-ER contact sites (Yang and Chan, 2024; Yang et al., 2025). These conceptually related systems improve conventional fluorescent complementation reporters by enhancing signal-to-noise ratios and offering improved reversibility, thereby expanding the toolkit for studying ER-mitochondria membrane contact sites (Yang and Chan, 2024; Shiiba et al., 2025; Yang et al., 2025).

Overall, by anchoring to specific subcellular organelles, these fluorescence complementation approaches provide selective visualization of contact sites, as well as localized monitoring of signaling events, with high spatial precision through proximity-dependent reconstitution. Their modular design and genetic encodability make them versatile tools for studying diverse MCS dynamics in living cells.

Proximity Labeling Approaches for Mapping MCS Proteomes

Another important class of proximity-based techniques relies on engineered biotin ligases, such as proximity-dependent biotin identification (BioID) derived from the R118G mutant of the Escherichia coli biotin protein ligase BirA (Sears et al., 2019). This mutant acts as a promiscuous biotin ligase that covalently labels nearby proteins with biotin, thus enabling the enrichment and specific labeling of surrounding proteins (Figure 2F). Inspired by reporter fragment complementation strategies, early work split BioID into two inactive fragments fused to POIs, allowing interaction-dependent enzyme reconstitution and proximity labeling in the presence of biotin (De Munter et al., 2017). More recently, this strategy was adapted for MCS studies by anchoring split-BioID halves to mitochondria (Mito-BioID) and LD (LD-BioID), leading to the identification of 71 proteins localized to the LD-mitochondria interface (Bezawork-Geleta et al., 2025). To address the relatively slow labeling kinetics of BioID, further engineering yielded TurboID and its truncated variant miniTurboID, which markedly accelerate labeling, reducing reaction times from hours to minutes (Branon et al., 2018). Hybrid systems such as the Complementation assay using Fusion of split GFP and TurboID (CsFiND) further extend this concept by enabling simultaneous visualization of ER-mitochondria contact sites and proteomic identification of associated proteins (Figure 2G) (Fujimoto et al., 2023). The CsFiND system has also been applied in yeast to identify new regulators of the nucleus-vacuole junction (NVJ), including the yeast INSIG homologs Nsg1 and Nsg2 and the aspartyl protease Ypf1, which participate in NVJ remodeling during glucose starvation (Fujimoto and Tamura, 2025).

In parallel, engineered ascorbate peroxidase (APEX2) has emerged as another powerful proximity-labeling platform. APEX2 catalyzes hydrogen peroxide (H₂O₂)-dependent oxidation of biotin-phenol to highly reactive radicals that rapidly label nearby proteins within about one minute. Originally developed as a genetically encoded tag for electron microscopy, APEX2 has since been widely adapted for spatially resolved proteomics (Figure 2H) (Martell et al., 2012). Building on this chemistry, the bimolecular fluorescence complementation-based proximity labeling (BiFCPL) system integrates BiFC-mediated targeting with APEX2-mediated labeling to selectively tag proteins at defined membrane contact sites (Figure 2I) (Zhou et al., 2023). Using BiFCPL, 60 high-confidence proteins were identified at mitochondria-LD membrane contacts in response to metabolic challenge, including evidence for accumulation of squalene epoxidase at these interfaces, potentially linking cholesterol synthesis to localized ATP availability (Zhou et al., 2023).

Taken together, proximity-labeling techniques have become useful tools for dissecting MCS composition by enabling selective tagging and identification of proteins within closely apposed membranes. While fluorescence complementation approaches primarily provide visualization of contact formation and dynamics in living cells, BioID-, TurboID-, and APEX2-based strategies complement these tools by resolving the local proteomes and uncovering molecular components and regulators within defined MCSs. Collectively, these approaches provide a powerful framework for identifying candidate tethering factors and signaling mediators, thereby supporting in situ mechanistic investigation of MCS biology.

Chemically Inducible Strategies for MCS Manipulation

Although proximity-based approaches have been instrumental in monitoring MCS dynamics and labeling potential regulators within various MCSs, they primarily provide correlative information and are often insufficient for dissecting the causal and mechanistic consequences of contact site perturbation. To move beyond observation toward direct functional interrogation, chemically inducible systems have been developed to actively manipulate membrane tethering and contact site dynamics. Among these, the chemically inducible dimerization (CID) system based on the FK506 binding protein (FKBP) and the FKBP-rapamycin binding (FRB) domain derived from mammalian target of rapamycin (mTOR) kinases (Choi et al., 1996) has emerged as a powerful tool for MCS studies (Jing et al., 2020). Rapamycin promotes the formation of a tight ternary complex with FKBP and FRB, with a binding affinity of approximately 12 nM (Banaszynski et al., 2005), thereby enforcing proximity between target proteins (Figure 3A). Anchoring the FKBP-FRB system to specific membranes allows inducible inter-organelle membrane tethering, allowing rapid bridging of subcellular organelles within minutes upon rapamycin addition (Komatsu et al., 2010). Using this strategy, inducible ER-mitochondria tethering was shown to be sufficient to induce rapid mitochondrial fragmentation, facilitating mechanistic studies of mitochondrial dynamics and their physiological roles (Miller and Stephens, 2014). In addition, rapamycin-induced ER relocalization toward the plasma membrane (PM) was shown to restore endomembrane ensheathing of misaligned chromosomes and partially rescue chromosome mis-segregation fidelity, demonstrating how spatial organization of endomembranes can directly influence mitotic genome stability (Ferrandiz et al., 2022).

Figure 3.

Chemically induced approaches for MCS manipulation. Overview of the principles and representative chemically inducible systems. A). Design of the rapamycin-induced heterodimerization system composed of FRB and FKBP. B). Contact-ID exemplifies the application of inducible FRB-FKBP heterodimerization. In the presence of rapamycin, ER-anchored FRB forms a complex with mitochondrial-anchored FKBP and brings the splitBioID fragments together to enable biotinylation within ER-mitochondria contact sites. C). Design of FEMP, illustrating the application of rapamycin-induced contacts formation. Upon rapamycin treatment, heterodimerization between ER-anchored FRB and OMM-targeted FKBP leads to the close proximity between the donor and acceptor fluorophores into close proximity, thereby producing a detectable emission signal from the acceptor. D). Design of the AP20187-induced homodimerization system. E). Schematic illustration of iMAPPER-633 design. Treatment with AP20187 leads to dimerization of the FKBP domain in iMAPPER-633, mimicking the store depletion induced oligomerization of STIM1 under physiological conditions. Abbreviations: FRB, FKBP12-rapamycin binding; FKBP, FK506-binding proteins; Rapa, rapamycin; FEMP, FRET-based indicator of ER-mitochondria proximity; iMAPPER, inducible membrane-attached peripheral ER; PB, polybasic domain; ORAI1, calcium release-activated calcium channel protein 1; Ca²⁺, calcium; STIM1, stromal interaction molecule 1.

Beyond inducible tethering alone, hybrid systems that integrate proximity labeling with inducible membrane contact formation have further expanded the functional scope of chemical perturbation strategies. The Contact-ID platform exemplifies this strategy by combining the FKBP-FRB inducible dimerization module with split BioID proximity labeling (Kwak et al., 2020). In this design, one BirA fragment fused to FRB is anchored to the ER membrane via Sec61β, whereas the complementary fragment fused to FKBP is targeted to the outer mitochondrial membrane (OMM) through Tom20 (Figure 3B) (Kwak et al., 2020). Using this system, 115 regulatory proteins specifically recruited during ER-mitochondrial contact formation were identified, including the proline peptide isomerase protein FKBP8, which was previously characterized as an anti-apoptotic protein rather than being linked to membrane tethering (Kwak et al., 2020). Overall, the Contact-ID system opens new avenues for exploring MCS regulators with enhanced spatial precision and reduced background labeling. A conceptually related design was employed in the FRET-based indicator of ER-mitochondria proximity (FEMP), which couples the FKBP-FRB heterodimerization with mitochondria- or ER-targeting sequences (Csordas et al., 2010). In this system, rapamycin-induced dimerization promotes ER-mitochondria juxtaposition, enabling quantification of membrane proximity through fluorescence resonance energy transfer (FRET) signals (Figure 3C). Compared with conventional colocalization analysis, FEMP provides improved spatial sensitivity for detecting ER-mitochondria contacts (Csordas et al., 2010). Application of FEMP has helped clarify the role of mitofusin 2 (Mfn2) in ER-mitochondria cross-talk, supporting its function as a bona fide ER-mitochondria tether (Naon et al., 2016). More recently, the FEMP system has been used to investigate how ER-mitochondria tethering contributes to lipid-reactive oxygen species (ROS) signaling during ferroptosis, which illuminates that PERK deficiency reverses the expansion of ER-mitochondria contacts observed during the early stages of ferroptosis (Sassano et al., 2025).

In addition to exogenous tethering modules, endogenous MCS proteins have also been engineered to create controllable tethering systems. Inspired by the ER-resident Ca²⁺ sensory protein stromal interaction molecule 1 (STIM1), a well-characterized tethering protein involved in the formation of ER-PM contact sites during store-operated calcium entry (SOCE) (Soboloff et al., 2012; Prakriya and Lewis, 2015; Nguyen et al., 2018; Wang, Zhou et al., 2010; Nonomura et al., 2025), the membrane-attached peripheral ER (MAPPER) system was developed as a convenient marker for ER-PM contact sites (Chang et al., 2013). In MAPPER, the signal peptide (SP) and single transmembrane domain of STIM1 are retained to preserve ER targeting, whereas the ER-lumenal EF-hand and sterile alpha-motif region are replaced with GFP for visualization (Chang et al., 2013). The cytosolic portion is substituted with the FRB domain and the polybasic (PB) motif from a small GTPase Rit, enabling interaction with negatively-charged phosphoinositides in the inner half-leaflet of PM (Chang et al., 2013). Using MAPPER, a recent study demonstrated increased ER-PM contact formation at the rear of migrating cells, establishing structural polarity that directs and sustains cell migration (Gong et al., 2024). A variant, inducible MAPPER-633 (iMAPPER-633), was subsequently developed by inserting tandem FKBP domains between GFP and the transmembrane domain of MAPPER (Chang et al., 2018). This design exploits AP20187-induced homodimerization of the FKBP domain (Figure 3D) (Lamb and Di Pietro, 2022), which mimics STIM1 oligomerization following ER Ca²⁺ depletion, thereby enabling its translocation toward PM to gate ORAI1 channels (Figure 3E) (Chang et al., 2018). The iMAPPER-633 system has been used to examine whether the EB1-STIM1 interaction is involved in SOCE (Wang et al., 2018), providing further insight into STIM1 PM-targeting mechanisms.

Collectively, chemically induced membrane tethering approaches provide rapid and tunable control over organelle contacts, enabling causal investigation of contact site dynamics and downstream cellular functions. However, concerns remain that forced tethering may distort native contact architecture by artificially expanding or stabilizing membrane interactions, potentially limiting their suitability for unbiased studies. In addition, the biological effects of rapamycin and other small molecules commonly used in CID systems remain potential limitations, particularly under scenarios where minimal perturbation of endogenous signaling pathways is required. Continued development of minimally disruptive tools will be important for achieving more physiologically faithful interrogation of MCS biology.

Optogenetic Control of MCS with Improved Spatial Precision and Reversibility

While chemically inducible systems provide robust temporal control of membrane tethering, their reliance on diffusible ligands can limit spatial precision and reversibility, raising concerns regarding off-target effects, systemic toxicity, and prolonged cellular perturbation (Voss et al., 2015). In contrast, light-inducible systems offer rapid, spatially confined, and reversible control of MCS dynamics and have emerged as powerful alternatives for studying inter-organelle communication (Passmore et al., 2021). The development of non-opsin photosensitive modules has significantly expanded this toolkit, enabling flexible regulation of protein-protein interactions and dissociation events with high spatiotemporal resolution (Ma et al., 2017; Lan et al., 2022; Tan et al., 2022; Wang et al., 2022; Huang et al., 2023; Li et al., 2024a; Nonomura et al., 2025; Ke et al., 2025b; Wang et al., 2025b; Guo et al., 2026). Because light stimulation operates on short timescales, these approaches are particularly well suited to the dynamic nature of many contact sites, allowing precise interrogation of MCS formation, remodeling, and function.

The genetically encoded light-oxygen-voltage-sensing domain 2 (LOV2) from oat represents a prototypical non-opsin photosensitive module, owing to its well-characterized photochemical mechanism and widespread adoption in optogenetic engineering (Ma et al., 2017; Ma et al., 2018; Hart and Gardner, 2021; He et al., 2021a; He et al., 2021b; He et al., 2021c; Tan et al., 2022; Lee et al., 2025). In general, a POI can be sterically caged by the LOV2 domain in the dark state, whereas blue light illumination induces conformational rearrangements and partial unfolding of the Jα helix, resulting in exposure of the fused functional module and restoration of its activity (Figure 4A) (Oide et al., 2018). Building on this working principle, an optogenetic tool, termed as light-inducible membrane-tethered peripheral ER (LiMETER), was engineered by replacing FRB with LOV2 in the previously described MAPPER system (Jing et al., 2015). In this design, the PB domain, initially masked by LOV2, becomes exposed upon blue light illumination, enabling its interaction with PM-embedded phosphoinositides with subsequent formation of ER-PM junctions (Figure 4B) (Jing et al., 2015). An improved variant, OptoPBer (or LiMETER-v2), was subsequently engineered by replacing Rit-PB in LiMETER with the STIM1-PB domain, resulting in reduced background tethering with enhanced signal-to-noise performance (He et al., 2017). Unlike many chemically inducible tethering systems that generate relatively stable contacts, LiMETER and OptoPBer allow reversible assembly and disassembly of targeted contact sites, thereby minimizing sustained cellular perturbation (Ke et al., 2025a). These systems have been widely applied to investigate membrane tethering between ER and PM in both mammalian (Jing et al., 2015; He et al., 2017) and plant cells (Li et al., 2025).

Figure 4.

Optogenetic strategies for MCS manipulation. Illustration of the underlying principles and representative optogenetic platforms building on light-responsive actuators. A). LOV2-based design. B). Light-responsive assembly of ER-plasma membrane contact sites induced by LiMETER. Light stimulation triggers the exposure of PB domain to restore its interaction with PM-resident PIP₂ and PIP₃, enabling photo-inducible MCS formation between ER and PM. C). Design of iLiD, the LOV2-based optical dimerization system. Under blue light illumination, uncaging of SsrA facilitates the heterodimerization between SspA and its natural binding partner sspB, thereby enabling the recruitment of proteins of interest (POIs). D). Design of the LIT system. Uncaging of SsrA from OMM-anchored LOV2 forms a complex with ER-anchored SspB, bringing the ER and mitochondrial membranes into proximity upon blue light illumination. E). CRY2-mediated light-responsive heterodimerization system. Clustering between CRY2 and CIB1/CIBN brings the target proteins together in the presence of blue light. F). Optogenetic MLC system leveraging CRY2-mediated clustering. Lysosome-anchored CRY2 forms clusters with the mitochondrial-anchored CIB under blue light stimulation, thereby inducing membrane tethering between lysosomes and mitochondria. Abbreviations: LOV2, light-oxygen-voltage-sensing domain 2; POI, protein of interest; LiMETER, light-inducible membrane-tethered peripheral ER; SP, signal peptide; PB, polybasic domain; PIP₂, phosphatidylinositol 4,5-biphosphate; PIP₃, phosphatidylinositol (3,4,5)-trisphosphate; LIT, light-inducible tethering; CB5, cytochrome b5; OMM, outer mitochondrial membrane; CRY2, cryptochrome 2; Opto-MLC, optogenetic mitochondria-lysosome contacts.

Beyond steric uncaging, light-inducible dimerization systems analogous to CID have been adapted for optical control of MCSs (Figure 4C), with the improved light-induced dimer (iLiD) system used most popularly (Guntas et al., 2015). In this system, a short peptide, SsrA, is caged by LOV2 in the dark state. Upon photostimulation, SsrA is exposed to regain its interaction with SspB, a 13kD bacterial protein which naturally engages SsrA (Guntas et al., 2015). Leveraging this concept, a genetically encoded light-inducible tethering (LIT) system was developed to control ER-mitochondria contacts with high spatiotemporal precision by anchoring iLiD and SspB to mitochondrial and ER membranes, respectively (Figure 4D) (Shi et al., 2018). A related optogenetic strategy using lysosome-anchored iLID and ER-anchored SspB has enabled interrogation of how lunapark (LNPK)-marked ER junctions and ER-lysosome contact sites spatially organize secretome mRNA translation, which reveals that lysosome-proximal ER junctions function as regulatory hubs linking ER architecture, nutrient sensing, and localized protein synthesis (Choi et al., 2026).

In addition to inducible heterodimerization systems, optogenetic approaches based on light-dependent homodimerization or oligomerization have attracted considerable attention due to their often higher efficiency in whole-cell activation experiments (Benedetti et al., 2018). One widely used module is Arabidopsis cryptochrome 2 (CRY2; Figure 4E), which undergoes blue light-dependent heterodimerization with cryptochrome-interacting basic helix-loop-helix protein (CIB1) as well as self-oligomerization (Duan et al., 2017; Dou et al., 2023; Kim et al., 2023; Ma et al., 2025; Wang et al., 2025a). This property has enabled optogenetic control of mitochondria-lysosome contacts (MLCs) by anchoring CRY2 to lysosomes via lysosomal-associated membrane protein (LAMP) and CIB to mitochondria through Tom20 (Qiu et al., 2022). Blue light-induced CRY2-CIB interaction brings lysosomes and mitochondria into proximity to form close apposition in living cells (Figure 4F) (Qiu et al., 2022). This system induces reversible mitochondrial fission and improves mitochondrial function with minimal cellular toxicity compared with chemical or genetic approaches that tend to alter the fusion-fission balance (Qiu et al., 2022). A similar CRY2-CIBN strategy was used to develop the light-inducible ER-specific mechanostimulator (LIMER), which drives ER-microtubule contact formation (Song et al., 2024). In this system, CRY2 and the N-terminal region of CIB1 (CIBN) were fused to Sec61β and kinesin family member 5A (KIF5A), positioning them to the ER and microtubule, respectively (Song et al., 2024). Compared with full-length CIB, CIBN exhibits faster diffusion and enhanced mobility, but slightly reduced coupling efficiency with CRY2 (Cui et al., 2014). Activation of LIMER triggers Ca²⁺ efflux from ER and inhibits ER-to-Golgi trafficking, thereby activating ER stress response, as evidenced by upregulation of BiP and phosphorylated eIF2α (Cui et al., 2014). These findings provide direct evidence that mechanical forces regulate ER function (Cui et al., 2014).

More recently, optogenetic strategies have expanded beyond inducible tethering to include functional disruption of contact sites, as exemplified by the light-dependent lipid droplet responsive crosslinker (LDLRC) platform (Bai et al., 2025). In this design, CRY2 was anchored to the LD membrane though perilipin-2 (Plin2), thereby promoting LD clustering upon blue light stimulation to reduce their accessibility for tethering with mitochondria (Bai et al., 2025). This disruption limits fatty acid transfer from lipid droplets to mitochondria under nutrient stress conditions, impairing cancer cell survival, and suppressing tumor growth both in vitro and in vivo (Bai et al., 2025). Of note, these findings broaden the conceptual framework of MCS tool design from mere inducible tethering toward functional modulation of organelle interactions.

Together, optogenetic approaches enable precise, reversible, and spatiotemporally controlled manipulation of diverse membrane contact sites through engineered photo-responsive modules that regulate organelle proximity, signaling, and function. Aside from facilitating studies of inter-organelle communication, these tools have advanced the field toward causal and mechanistic dissection of MCS biology, promising to illuminate how membrane tethering influences cellular signaling, metabolism, and stress responses at a systems level. Despite these advances, several challenges remain in current optogenetic approaches, including the limited tissue penetration of blue light, potential phototoxicity, heating effects, and reduced precision in deep tissues. Addressing these limitations will broaden the applicability of light-inducible approaches and enhance their translational potential by improving the safety and efficacy of optogenetic systems.

Conclusions Remarks and Future Perspective

The organelle contactome has emerged as a dynamic regulatory network that orchestrates intracellular communication and the exchange of essential molecules, including ions, lipids, and metabolites. Over the past decade, research on various contact sites and their tethering mechanisms has advanced rapidly, increasingly intersecting with synthetic biology and chemical biology approaches. Through the integration of advanced methodologies, the field has evolved from basic visualization using fluorescence complementation reporters to proteomic proximity mapping enabled by proximity-labeling approaches, which define the molecular composition and the interaction networks at MCSs. More recently, inducible tethering strategies have enabled direct manipulation of membrane contacts upon stimulation of small molecules or light. However, these approaches interrogate distinct aspects of MCS biology. Visualization-based reporters monitor the spatial occurrence of contacts, proximity-labeling methods define the molecular environment of contact sites, whereas inducible tethering systems artificially enforce membrane juxtaposition to probe functional consequences. As such, artificially induced contacts do not necessarily recapitulate native physiological parameters but instead provide complementary tools for experimentally perturbing inter-organelle communication. Collectively, these developments mark a transition from descriptive observation to mechanistic interrogation of MCS biology. We showcase herein recent progress in commonly applied engineering strategies and representative applications that allow precise and versatile control of diverse MCSs. The development of inducible molecular toolkits has shifted the field from largely descriptive characterization toward mechanistic interrogation, enabling causal analysis of MCS functions in cellular physiology.

Despite these advances, significant challenges remain. A deeper mechanistic understanding of how MCS dysregulation contributes to human disease is still needed (Cali et al., 2025), as accumulating evidence implicates altered contact site dynamics in neurodegeneration (Kim et al., 2022), metabolic disorders (Yang et al., 2020), cancer, and other pathologies (Gil-Hernandez et al., 2020; An et al., 2024). Moreover, translation of these sophisticated experimental tools into therapeutic contexts remains at an early stage, limited by challenges in delivery, specificity, and long-term safety. Looking ahead, continued integration of synthetic biology, advanced imaging, AI-guided protein design, and in vivo engineering approaches will likely accelerate both fundamental discoveries and translational opportunities. Chemically and optically inducible systems (Tan et al., 2022; Ke et al., 2025b; Wang et al., 2025b), in particular, hold strong promise for enabling spatiotemporally precise therapeutic modulation of organelle communication. By enabling controlled modulation of processes such as ER-mitochondria Ca²⁺ exchange, mitochondria-lysosome signaling, or lipid droplet-mitochondria interactions, these approaches facilitate mechanistic investigation of organelle communication and may ultimately opening new avenues for targeted intervention in complex human diseases.