Abstract
The actin cytoskeleton plays multiple critical roles in cells, from cell migration to organelle dynamics. The small and transient actin structures regulating organelle dynamics are difficult to detect with fluorescence microscopy. We generated fluorescent protein-tagged actin nanobodies targeted to organelle membranes to enable live cell imaging of sub-organellar actin dynamics with high spatiotemporal resolution. These probes reveal ER-associated actin drives fission of multiple organelles including mitochondria, endosomes, lysosomes, peroxisomes, and the Golgi.
Introduction
The critical role of the actin cytoskeleton in organelle dynamics is largely accepted, but poorly understood. The precise spatiotemporal dynamics of actin at organelle membranes remain particularly unclear due to the combined limitations of currently available actin probes and imaging approaches. For fluorescence microscopy approaches, imaging smaller actin structures in the cell suffers from an enormous background signal issue - the high signal from the dense meshwork of actin filaments throughout the cell overwhelms the signal from the relatively small, transient actin structures associated with organelle dynamics. Furthermore, the limitations in resolution make it difficult to determine whether any actin filaments are physically associated with the organelle. Meanwhile, EM approaches can be used to visualize actin filaments with extremely high resolution1,2. However, these techniques only capture single timepoints, making it difficult to determine the precise state or dynamics of the actin filaments or their associated organelles at the time of fixation. All these limitations ultimately preclude a solid understanding of the mechanisms by which actin regulates organelle dynamics in health or disease. Here we employ fluorescent protein-tagged actin nanobodies, aka “actin chromobodies” (AC)3,4, fused to organelle membrane targeting sequences to facilitate live cell imaging of sub-organellar actin dynamics with high spatiotemporal resolution. Using these probes we discovered that ER-associated actin accumulates at all ER-organelle contacts during organelle fission.
Results
We hypothesized that AC probes with organelle membrane targeting sequences could be used to monitor actin dynamics exclusively within a ~10nm distance from the target organelle membrane. To test this hypothesis, we used the yeast Fis1 mitochondrial outer membrane and Cytb5ER endoplasmic reticulum (ER) minimal C-terminus tail membrane targeting sequences fused to the cytoplasm-facing actin nanobody and tagGFP (“AC-mito” and “AC-ER”)5. Live cell Airyscan imaging of Hap1 or U2OS cells expressing AC-mito counterstained with MitoTracker dye revealed strikingly specific regions of AC-mito enrichment on the surfaces of mitochondria (Figure 1 and Supplementary Movies 1-2). Similarly, AC-ER expressing cells revealed significant accumulation of AC-ER at specific ER-mitochondria contact sites (Figure 1 and Supplementary Movies 3-4). Interestingly, the AC-ER probe often revealed significant actin accumulation at ER-mitochondria contacts (Figure 1a).
To rule out the possibility that the membrane targeting sequences we used were causing the probe to accumulate in specific regions independent of actin-binding activity, we generated and co-transfected AC-mito or AC-ER with mCherry-tagged versions of the mitochondrial (Fis1) and ER (Cytb5ER) membrane targeting sequences (mCherry-mito and mCherry-ER). As expected, the mCherry-mito and mCherry-ER signals were evenly distributed along their respective organelle membranes, with no obvious enrichment in any specific regions. In contrast, the co-expressed AC-mito and AC-ER constructs were uniquely enriched in specific regions on their respective organelles (Figure 1b). mCherry-ER showed some accumulation at ER-mitochondria intersections, indicative of stable contacts at these sites6. However, the AC-ER probe showed particularly enhanced enrichment at ER-mitochondria contacts, revealing significant accumulation of actin specifically at these contact sites (Figure 1b). These probes thus facilitate unprecedented resolution (10nm from the membrane) and signal-to-background imaging of mitochondria- and ER-associated actin filaments in live cells, validating previous reports and models that actin accumulates at ER-mitochondria intersections, presumably to drive mitochondrial fission1,7–11. It should be noted that most (but not all – see Moore et al.12) previous live imaging of actin and mitochondria was performed during overexpression or stress conditions in order to increase mitochondrial fission and/or the associated actin signal at mitochondria7-11,13, whereas these data were obtained under normal physiological conditions.
We considered it important to determine whether the Fis1 membrane targeting sequence in our AC-mito probe might induce dominant negative effects on mitochondrial dynamics by outcompeting endogenous Fis1 for mitochondrial outer membrane localization. Since the AC-mito construct uses only the C-terminal membrane targeting domain of Fis1, an anti-Fis1 antibody against the cytoplasmic N-terminal domain should only detect endogenous protein. We found no change in endogenous Fis1 localization to mitochondria in AC-mito expressing cells compared to neighboring non-transfected cells (Supplementary Figure 1). Finally, to determine whether our localization results were specific to the Fis1 membrane targeting sequence or to actin nanobodies in particular, we generated variants of our AC-mito probe using a different membrane targeting sequence (Cytb5mito5) and/or a different F-actin probe (LifeAct14). All four of our probes yielded very similar results, revealing specific sub-mitochondrial regions of actin accumulation (Supplementary Figure 2). However, the LifeAct probes displayed a more diffuse localization on the mitochondrial outer membrane. In contrast, our AC probe seemed to better highlight specific subdomains of actin enrichment, consistent with previous comparative studies suggesting that actin nanobodies have superior F-actin labeling capabilities4. Thus, we decided to mainly use our AC probes for the majority of our experiments. To further test whether AC-mito accumulation on mitochondria is dependent on F-actin, we treated AC-mito expressing cells with the F-actin depolymerizing drug Latrunculin B (LatB), which significantly reduced the AC-mito signal. Most of the remaining signal appeared as punctate structures, matching the punctate F-actin structures observed with the pan-actin probe AC-tagRFP (“pan-AC”) (Supplementary Figure 3). Similar results were observed after treatment with cytochalasin D (Supplementary Movie 5). Taken together, these results strongly support the conclusion that our AC probes are labeling F-actin on their respective membrane surfaces.
In our hands, pan-actin probes such as phalloidin fail to clearly reveal sites of actin accumulation on mitochondria or the ER under normal conditions (Supplementary Figure 4). Furthermore, given the abundance of actin signal throughout the cell and the limitations of resolution in fluorescence microscopy, it is extremely difficult to assess whether colocalization of a pan-actin signal and an organelle membrane is a true actin-membrane association or not. Since our AC probes are small proteins tethered to their respective organelle membranes, we can conclude that all observed actin accumulation is within a maximum distance of 10nm from the organelle membrane (although the distal ends of associated actin filaments may extend far beyond 10nm). Since phalloidin is widely considered the gold standard for F-actin labeling4, we attempted to colocalize AC-mito and AC-ER with phalloidin staining in fixed cells (Supplementary Figure 5). Interestingly, we found that fixation with 4% PFA causes a dramatic loss in AC-mito enrichment as well as an overall decrease in AC-mito or AC-ER signal, perhaps reflective of the relatively small and transient nature of these membrane-associated actin structures. However, we were still able to detect some co-accumulation of our probes with phalloidin in fixed cells by adjusting the image contrast. Notably, the contrast in the phalloidin signal needed to be extremely high in order to visualize the signal at AC enriched sites (Supplementary Figure 5), causing neighboring structures to be completely saturated or clipped. These results highlight both the value of live imaging labeling approaches as well as why similar actin enriched sites were not previously detected with phalloidin staining. Indeed, only with the AC signal serving as a guide could confidently identify sub-organellar regions of actin enrichment in our phalloidin signal.
Previous studies have demonstrated a role for actin at ER-mitochondria contact sites in driving mitochondrial fission1,7-11,13,15. However, due to the abundance of actin in the cytoplasm, it is difficult to detect actin accumulating near ER-mitochondria contact sites, or whether any such actin directly associates with either organelle using live imaging. To determine whether mitochondria-or ER-associated actin accumulates during ER-mediated fission, we performed live imaging of cells co-expressing AC-mito and mCherry-tagged Drp1, an actin-binding dynamin-related GTPase protein essential for mitochondrial fission8,16,17 (Figure 2, Supplementary Figure 6, and Supplementary Movies 6-9). To avoid potential overexpression artifacts, we used the low expression ubiquitin promoter C to drive expression of our AC probes18. We found that AC-mito accumulates at or immediately adjacent to mitochondrial fission sites prior to Drp1-mediated fission (Figure 2a, Supplementary Movie 6). Interestingly, we frequently observed elongated regions of AC-mito enrichment crossing over mitochondrial constriction sites, consistent with recent reports of similar actin structures observed with platinum replica electron microscopy of membrane extracted cells1. We similarly observed accumulation of AC-ER at mitochondrial fission sites prior to Drp1-mediated fission (Figure 2b, Supplementary Movie 7). Co-expression of AC-mito or AC-ER with mCherry-ER showed that both mitochondria- and ER-associated actin accumulates at ER-tubules driving mitochondrial fission (Figure 2c-d, Supplementary Movies 8-9). These results identify a subpopulation of actin filaments specifically associated with the mitochondrial outer membrane and the ER prior to and during ER- and Drp1-mediated mitochondrial fission.
Recent studies have suggested an analogous role for the ER and actin in mediating endosomal fission19,20. However, direct evidence is missing. Using our AC-ER probe, we observed actin enrichment at ER-associated endosomal fission sites (Figure 3b, Supplementary Movies 10-11). Given the ER regularly contacts most organelles in the cell more than 90% of the time21, we hypothesized that actin accumulates at all ER-organelle contact sites to drive their fission. Live imaging of cells co-expressing AC-ER and markers for mitochondria, endosomes, peroxisomes, lysosomes, and the Golgi all revealed accumulation of ER-associated actin at fission sites for each of these organelles (Figure 3a-e, Supplementary Movies 10-14). Scanning electron microscopy (SEM) imaging of cells treated with saponin and cytoskeleton stabilizing buffer shows actin filaments associated with mitochondria in similar patterns observed with AC-mito (Figure 3f). Finally, AC-ER reveals a striking network of ER-associated actin bundles on the nucleus (Figure 3g, Supplementary Movie 15). SEM imaging of non-transfected cells treated with saponin and cytoskeleton stabilizing buffer, we observed a similar network of actin fibers accumulating on the nucleus, validating the localization pattern observed in our AC-ER fluorescence data (Figure 3g). A perinuclear actin cap structure has been previously reported and is known to have important roles in multiple cellular processes22-27. However, as with other actin structures, the actin filaments specifically associated with the nucleus have been difficult to visualize with pan-actin probes due to the overwhelming signal from surrounding actin stress fibers and cortical actin. Our data show that AC-ER labeling may provide a more specific labelling approach for visualisation of perinuclear actin cap structures.
Conclusion
Taken together, our data reveal a novel role for the ER in recruiting actin filaments to all ER-organelle membrane contact sites, most likely in order to drive membrane remodeling and scaffolding for dynamin-mediated organelle fission. Given the important role of both organelle-organelle contact sites and the actin cytoskeleton in health and disease, the apparent ability of the ER to drive actin accumulation at organelle contacts has broad implications for both cell biology and biomedical research. Finally, the ability to use membrane-anchored AC probes as “proximity sensors” for sub-cellular actin dynamics provides a novel tool for studying the role of actin in a wide range of cell biological processes. The variety of subcellular compartments, protein targets, and corresponding subcellular or sub-organellar processes that can be studied with such probes will increase as additional nanobodies and targeting sequences are developed.
Methods
Cell culture
U2OS and Hap1 cells were purchased from ATCC. Cells were grown in DMEM supplemented with 10% fetal bovine serum at 37°C with 5% CO2. Cells were transfected with Lipofectamine 2000 (ThermoFisher). Cells were plated onto either 8-well #1.5 imaging chambers or #1.5 35mm dishes (Cellvis) that were coated with 10μg/mL fibronectin in PBS at 37°C for 30 minutes prior to plating. 50nM MitoTracker Deep Red (ThermoFisher) was added for 30 minutes then washed for at least 30 minutes to allow for recovery time before imaging in FluoroBrite (ThermoFisher) media.
Airyscan confocal imaging
Cells were imaged with a 63x 1.4NA oil objective on a ZEISS 880 LSM Airyscan confocal system with an inverted stage and heated incubation system with 5% CO2 control. The GFP channels were imaged with a 488nm laser line at ~500nW laser power. The mCherry or tagRFP channels were imaged with 561nm laser at ~1μW laser power. The MitoTracker Deep Red channel was imaged with ~250nW laser power. For timelapse imaging, the zoom factor was set between 3x-6x to increase the frame rate. In all cases, the maximum pixel dwell time (~0.684μs/pixel) and 2x Nyquist optimal pixel size (~40nm/pixel) was used.
Antibodies
We used the rabbit anti-Fis1 antibody against the N-terminus cytoplasmic facing side of the human Fis1 protein, made by Prestige Antibodies Powered by Atlas Antibodies (Sigma-Aldrich, catalog #: HPA017430). The amino acid sequence of the antigen is MEAVLNELVSVEDLLKFEKKFQSEKAAGSVSKSTQFEYAWCLVRSKYNDDIRKGIVLLEELLPKGS KEEQRDYVFYLAVGNYRLKEYEKALKYVRGLLQTEPQNNQAKELERLIDKAMKKD.
Immunofluorescence
Cells were washed in PBS then fixed with 4% PFA for 30 minutes before permeabilization with 0.1% Triton X-100 for 30 minutes. Cells were then blocked overnight with 4% BSA at 4°C. Cells were then incubated with primary antibody for 2 hours, rinsed 3x with PBS for 10 minutes each, then incubated with secondary antibodies (Jackson Immunoresearch Laboratories) for 1 hour, rinsed 3x with PBS for 10 minutes each, then counterstained with Alexa405-phalloidin (ThermoFisher) for 30 minutes, then rinsed with PBS 3x for 10 minutes each, then mounted with ProLong Glass antifade reagent (ThermoFisher).
Scanning Electron Microscopy
Cells were briefly washed with 0.1 M phosphate-buffered saline (PBS) (37oC) to remove culture media and immediately treated with a membrane extraction solution containing 1% Triton X-100, 100 mM PIPES (pH 7.2), 1 mM EGTA, 1 mM MgCl2, 4.2 % sucrose, 10 μM taxol (Thermo-Scientific) and 10 μM phalloidin (Sigma) for 10 min with gentle rocking at room temperature. Then the samples were washed twice for 5 min in PBS and fixed with 2% glutaraldehyde, 2% paraformaldehyde (Electron Microscopy Sciences - EMS), in 0.1 M sodium cacodylate buffer for 30 min at RT. Samples were washed in the same buffer, post-fixed with 1% OsO4 (EMS) and 1% tannic acid (EMS) for 1h each, dehydrated with a graded ethanol series until absolute and critical point-dried (Leica CPD 030). The samples were coated with a thin platinum layer (4 nm) (Leica EM SCD500) and imaged on a Zeiss Sigma-VP scanning electron microscope at 5 kV.
Image processing and analysis
After acquisition, images were Airyscan processed using the auto-filter 2D-SR settings in Zen Blue (ZEISS). All images were post-processed and analyzed using Imaris (BITPLANE) and Fiji software28.
Plasmids
Drp1-mCherry was a kind gift from Gia Voeltz (Addgene plasmid #49152). mCherry-Cytob5RR was a gift from Nica Borgese29. EGFP-LifeAct was a gift from Jennifer Rohn 30. Lamp1-mCherry, SiT-mApple, Sec61-mCherry, and dsRed-Skl were all generous gifts from the Lippincott-Schwartz lab. Rab5a-mCherry was a generous gift from the Merrifield lab (Addgene plasmid #27679). All custom actin nanobody probes were generated starting from the commercial vector of actin chromobody-tagGFP or actin chromobody-tagRFP (ChromoTek) and cloned via the BglII and NotI restriction sites. The following amino acid sequences were attached to the C-terminal of the actin chromobody probes to target the protein either to mitochondria or the ER:
Fis1 (AC-mito and LifeAct-GFP-Fis1)
IQKETLKGVVVAGGVLAGAVAVASFFLRNKRR5
Cytb5mito (aka “Cytob5RR”) (AC-GFP-Cytb5mito and LifeAct-GFP-Cytb5mito)
FEPSETLITTVESNSSWWTNWVIPAISALVVALMYRR31
Cytb5ER (AC-ER)
IDSSSSWWTNWVIPAISAVAVALMYRLYMAED5
LifeAct-GFP-Fis1, LifeAct-GFP-Cytb5mito, and AC-GFP-Cytb5mito were generated using PFU Ultra II for megaprimer PCR insertion32. The PCR primers, intended modifications, insert templates, and destination plasmids are listed in Supplemental Table 1 below. All constructs were sequenced completely across their coding region.
Supplementary Movie 1|Live cell imaging of mitochondria-associated actin. A Hap1 cell expressing AC-mito counterstained with MitoTracker shows dynamic subdomains of actin enrichment on mitochondria.
Supplementary Movie 2|Live cell imaging of mitochondria-associated actin. A Hap1 cell expressing AC-mito counterstained with MitoTracker shows dynamic subdomains of actin enrichment on mitochondria (example 2).
Supplementary Movie 3|Live cell imaging of ER-associated actin. A Hap1 cell expressing AC-ER counterstained with MitoTracker shows dynamic subdomains of actin enrichment on the ER, in particular at ER-mitochondria contact sites.
Supplementary Movie 4|Live cell imaging of ER-associated actin. A Hap1 cell expressing AC-ER counterstained with MitoTracker shows dynamic subdomains of actin enrichment on the ER, in particular at ER-mitochondria contact sites (example 2).
Supplementary Movie 5|Live imaging of AC-mito expressing cells before and after Cytochalasin D treatment. Live imaging of Cytochalasin D untreated (left) vs. treated (right) AC-mito expressing U2OS cells.
Supplementary Movie 6|Live imaging of AC-mito and Drp1 during mitochondrial fission. Live imaging of AC-mito and Drp1-mCherry in U2OS cells counterstained with MitoTracker reveals accumulation of mitochondria-associated actin prior to Drp1-mediated fission.
Supplementary Movie 7|Live imaging of AC-ER and Drp1 during mitochondrial fission. Live imaging of AC-ER and Drp1-mCherry in U2OS cells counterstained with MitoTracker reveals accumulation of ER-associated actin prior to Drp1-mediated fission.
Supplementary Movie 8|Live imaging of AC-mito and the ER during mitochondrial fission. Live imaging of AC-mito and mCherry-ER in U2OS cells counterstained with MitoTracker reveals accumulation of mitochondria-associated actin prior to ER-mediated fission.
Supplementary Movie 9|Live imaging of AC-ER and the ER during mitochondrial fission. Live imaging of AC-ER and mCherry-ER in U2OS cells counterstained with MitoTracker reveals accumulation of ER-associated actin prior to ER-mediated fission.
Supplementary Movie 10|ER-associated actin accumulates at mitochondrial fission sites. AC-ER accumulation at a fission site in MitoTracker-labeled mitochondria. Scale bar: 1 µm
Supplementary Movie 11|ER-associated actin accumulates at endosomal fission sites. AC-ER accumulation at two fission sites in Rab5a-mCherry-labeled endosomes. Scale bar: 1 µm
Supplementary Movie 12|ER-associated actin accumulates at peroxisomal fission sites. AC-ER accumulation at a fission site in dsRed-Skl-labeled peroxisomes. Scale bar: 1 µm
Supplementary Movie 13|ER-associated actin accumulates at lysosomal fission sites. AC-ER accumulation at a fission site in LAMP1-mCherry-labeled lysosomes. Scale bar: 1 µm
Supplementary Movie 14|ER-associated actin accumulates at Golgi fission sites. AC-ER accumulation at a fission site in SiT-mApple-labeled Golgi. Scale bar: 1 µm
Supplementary Movie 15|ER-associated actin forms a network on the nucleus. 3D rendering of a U2OS cell expressing AC-ER reveals accumulation around the surface of the nucleus. Cyan: AC-ER. Orange: MitoTracker.
Acknowledgements
The Waitt Advanced Biophotonics Center is funded by the Waitt Foundation and Core Grant applications NCI CCSG (CA014195) and NINDS Neuroscience Center (NS072031). This work was supported by the Transgenic Core Facility of the Salk Institute with funding from NIH-NCI CCSG: P30 014195. R.G. laboratory is funded by grants from HFSP RGP0021/2016 and the Cluster of Excellence CIBSS-Centre for Integrative Biological Signalling Studies. O.A.Q. lab is supported by NIGMS Grant R15 GM119077 and by funding from the University of Richmond School of Arts & Sciences.