A number of FRET-based force biosensors have recently been developed, enabling the protein-specific resolution of intracellular force. In this protocol, we demonstrate how one of these sensors, designed for the linker of the nucleoskeleton-cytoskeleton (LINC) complex protein Nesprin-2G can be used to measure actomyosin forces on the nuclear LINC complex.
The LINC complex has been hypothesized to be the critical structure that mediates the transfer of mechanical forces from the cytoskeleton to the nucleus. Nesprin-2G is a key component of the LINC complex that connects the actin cytoskeleton to membrane proteins (SUN domain proteins) in the perinuclear space. These membrane proteins connect to lamins inside the nucleus. Recently, a Förster Resonance Energy Transfer (FRET)-force probe was cloned into mini-Nesprin-2G (Nesprin-TS (tension sensor)) and used to measure tension across Nesprin-2G in live NIH3T3 fibroblasts. This paper describes the process of using Nesprin-TS to measure LINC complex forces in NIH3T3 fibroblasts. To extract FRET information from Nesprin-TS, an outline of how to spectrally unmix raw spectral images into acceptor and donor fluorescent channels is also presented. Using open-source software (ImageJ), images are pre-processed and transformed into ratiometric images. Finally, FRET data of Nesprin-TS is presented, along with strategies for how to compare data across different experimental groups.
Force-sensitive, genetically encoded FRET sensors have recently emerged as an important tool for measuring tensile-based forces in live cells, providing insight into how mechanical forces are applied across proteins1,2,3,4. With these tools, researchers can non-invasively image intracellular forces in living cells using conventional fluorescent microscopes. These sensors consist of a FRET-pair (donor and acceptor fluorescent proteins, most frequently a blue donor and yellow acceptor) separated by an elastic peptide3. In contrast to C- or N-terminal tagging, this sensor is inserted into an internal site of a protein to measure the mechanical force transmitted across the protein, behaving as a molecular strain gauge. Increased mechanical tension across the sensor results in an increased distance between the FRET-pair, resulting in decreased FRET3. As a result, the FRET is inversely related to tensile force.
These fluorescent-based sensors have been developed for focal adhesion proteins (vinculin3 and talin4), cytoskeletal proteins (α-actinin5), and cell-cell junction proteins (E-Cadherin6,7, VE-Cadherin8, and PECAM8). The most frequently used and well-characterized elastic linker in these biosensors is known as TSmod and consists of a repetitive sequence of 40 amino acids, (GPGGA)8, which was derived from the spider silk protein flagelliform. TSmod has been shown to behave as a linear elastic nano-spring, with FRET responsiveness to 1 to 5 pN of tensile force3. Different lengths of flagelliform can be used to alter the dynamic range of TSmod FRET-force sensitivity9. In addition to flagelliform, spectrin repeats5 and villin headpiece peptide (known as HP35)4 have been used as the elastic peptides between FRET-pairs in similar force biosensors4. Finally, a recent report showed that TSmod can also be used to detect compressive forces10.
We recently developed a force sensor for the linker of the nucleo-cytoskeleton (LINC) complex protein Nesprin-2G by using TSmod inserted into a previously developed truncated Nesprin-2G protein known as mini-Nesprin2G (Figure 2C), which behaves similarly to endogenous Nesprin-2G11. The LINC complex contains multiple proteins that lead from the outside to the inside of the nucleus, linking the cytoplasmic cytoskeleton to the nuclear lamina. Nesprin-2G is a structural protein binding to both the actin cytoskeleton in the cytoplasm and to SUN proteins in the perinuclear space. Using our biosensor, we were able to show that Nesprin-2G is subject to actomyosin-dependent tension in NIH3T3 fibroblasts2. This was the first time that force was directly measured across a protein in the nuclear LINC complex, and it is likely to become an important tool to understand the role of force on the nucleus in mechanobiology.
The protocol below provides a detailed methodology of how to use the Nesprin-2G force sensor, including the expression of the Nesprin tension sensor (Nesprin-TS) in mammalian cells, as well as the acquisition and analysis of FRET images of cells expressing Nesprin-TS. Using an inverted confocal microscope equipped with a spectral detector, a description of how to measure sensitized emission FRET using spectral unmixing and ratiometric FRET imaging is provided. The output ratiometric images can be used to make relative quantitative force comparisons. While this protocol is focused on the expression of Nesprin-TS in fibroblasts, it is easily adaptable to other mammalian cells, including both cell lines and primary cells. Furthermore, this protocol as it relates to image acquisition and FRET analysis can readily be adapted to other FRET-based force biosensors that have been developed for other proteins.
1. Obtain Nesprin-2G Sensor DNA and Other Plasmid DNA
2. Transfect Cells with Nesprin-2G and Other Plasmid DNA
3. Verify Transfection Efficiency
4. Capture Spectral Fingerprints of mTFP1 and Venus Fluorophores for Spectral Unmixing
5. Capture Unmixed Images
6. Image Processing and Ratio Image Analysis
Following the protocol above, plasmid DNA was acquired from the DNA repository and transformed into E. coli cells. E. coli expressing the sensor DNA were selected from LB/Ampicillin plates and amplified in a liquid LB broth. Following the amplification of the vectors, DNA plasmids were purified into TRIS-EDTA buffer using a standard, commercially available DNA isolation kit. Using a spectrophotometer, purified DNA was quantified into a standard concentration of µg/mL (Figure 1A, Days 1-2).
Using NIH3T3 fibroblasts, cells were grown to 70-90% confluence in a 6-well plastic cell culture dish. To insert plasmid DNA into the NIH3T3 cells, a lipid-mediated plasmid transfection was performed. Commercially available transfection reagent (see the Table of Materials) was used as the lipid vessel for DNA. Two replicates of the Nesprin-TS and Nesprin-HL sensors were transfected into 4 cell wells (Figure 1B). In preparation for FRET imaging, single fluorophore constructs of mTFP1 and venus were transfected into NIH3T3 cells (Figure 1B). Following transfection, cells were transferred to glass-bottom viewing dishes compatible with high magnification, inverted confocal microscopes.
Before proceeding to confocal imaging, a simple inverted wide-field fluorescent microscope was used to verify the transfection efficiency. Using a 10X objective with a numerical aperture of 0.25, many cells in a typical field-of-view expressed the Nesprin-TS (Figure 2). Generally, the sensor localized around the nuclear envelope, consistent with the endogenous localization of Nesprin-2G2. The no-force control sensor, Nesprin-HL, followed a similar expression pattern2.
After the confirmation of the transfection efficiency of Nesprin sensors TS and HL, live cells expressing single fluorophore constructs (mTFP1 or venus) were imaged with an inverted confocal microscope. Using the spectral detection mode, mTFP1 and venus spectral fingerprints were recorded in the spectral unmixing database (Figure 3, Arrow-2 and Figure 4A). Following fingerprinting, the capture parameter was switched to unmixing mode, with the mTFP1 and venus spectra loaded as the components (Figure 4D). Raw images were acquired as spectrally resolved image stacks (Figure 4).
Finally, the spectrally resolved image stacks were imported into ImageJ open-source software for processing. The images were pre-processed, and ratio images were computed using established procedures15.
Since the FRET ratio is a relative measurement, every experiment must compare two or more conditions. Additionally, because laser and camera detector settings can influence FRET, all conditions should be acquired in the same imaging session. Frequently, we use a force-insensitive headless or tailless control, which represents a zero-force control, as a base-line for our experiments. Because FRET ratio images can be noisy, it is often necessary to take multiple images (5-20) for each condition. For conditions where there are large differences in FRET, this will be visually discernable in the FRET ratio images. The most dramatic difference in FRET for the Nesprin tension sensor was between unpatterned and elongated cells (Figure 5). Although there is variation between cells, the change in FRET between the two conditions is so dramatic that further analysis may not be required. However, other conditions often have minimal changes in FRET; they may not be visually discernable between individual images, but rather require the data to be analyzed in aggregate. To determine if these changes are significant, the median FRET ratio for each image is computed and visually expressed as a set of histograms. By comparing histograms between conditions, it becomes easier to see subtler changes in FRET between experimental groups. In Figure 6A and 6B, the FRET histograms of each individual cell are represented by a color in the stacked histogram. Calyculin A-treated cells (Figure 6B) exhibit a leftward-shifted FRET histogram relative to cells treated with myosin inhibitor ML7 (Figure 6A).
Figure 1: Experimental Time-based Flow Chart. (A) Starting with acquired sensor plasmid DNA, E. coli vectors are transformed with DNA, selected, and amplified (Days 1-2). From concentrated E. coli LB broth, plasmid DNA is isolated and quantified (Day 3). Using purified DNA plasmids, NIH3T3 fibroblasts are transfected in a six-well format (B) and transferred onto glass-bottom dishes that are compatible with an inverted confocal microscope (Day 3). To verify the success of the transfection, cells are viewed under a widefield fluorescent microscope equipped with the appropriate filter sets (Day 4). Contingent upon a successful transfection, cells are imaged on a confocal microscope equipped with a spectral detector. Using spectral unmixing, mTFP1 and venus channels are separated during the acquisition and saved as spectrally resolved image stacks (Days 4-5). Lastly, images are preprocessed in ImageJ, and the ratio-images are computed. Please click here to view a larger version of this figure.
Figure 2: A Representative Image of Transfected NIH3T3 Fibroblasts. (A) 10X magnification of transfected NIH3T3 fibroblasts expressing Nesprin-TS using a GFP filterset. (B) An exploded view of the bounding box from part A. Sensor expression follows the nuclear envelope and often extends into the cytoplasm. (C) Nesprin-TS and -HL Control and how they bind to the actin and SUN binding domains. Please click here to view a larger version of this figure.
Figure 3: Spectral Fingerprinting Tutorial Using Confocal Software. Normalized emission spectra obtained from 20-pixel radii in the highlighted regions-of-interest delineated by colored crosshairs. (Arrow-1) Spectral detection mode to capture the image. (Arrow-2) "Spectral Unmixing" setting to add normalized spectra to the spectral database. (Arrow-3) The excitation frequency. Please click here to view a larger version of this figure.
Figure 4: Live Spectral Unmixing and Output of Confocal Software. (A) Critical imaging parameters to reproduce live, unmixed images. (Arrow-1) The excitation frequency. (Arrow-2) Live spectral unmixing mode. (B) Nuclei image in unmixed venus channel. (C) Nuclei image in unmixed mTFP1 channel. (D) Combined image of venus and mTFP1. (Arrow-3) The nuclear membrane of the cell where Nesprin-TS is being expressed. Please click here to view a larger version of this figure.
Figure 5: Nesprin-TS FRET ratio images in patterned and unpatterned Madin-Darby Canine Kidney (MDCK) cells. MDCK cells that stably express Nesprin-TS were imaged on 10 µm-wide fibronectin lines (A) or unpatterned polydimethylsiloxane (PDMS) membranes (B). Nuclear membranes were visually masked and processed into FRET ratios. The colored scale bar represents the FRET ratio amplitude for unpatterned and patterned nuclei. Please click here to view a larger version of this figure.
Figure 6: Processed Ratiometric FRET Images from Nesprin-TS-Expressing Cells and Histogram Analysis of Aggregated Data. (A) A representative, normalized stacked histogram of FRET ratio images of Nesprin-TS treated with ML7 myosin inhibitor, color-coded by individual cell nuclei. (B) A representative, normalized stacked histogram of FRET ratio images of Nesprin-TS treated with Calyculin A, an activator of cell contractility. The legend correlates each analyzed nucleus to its location on the histogram. Please click here to view a larger version of this figure.
A method and demonstration of live cell imaging of mechanical tension across Nesprin-2G, a protein in the nuclear LINC complex, was outlined above. Prior to this work, various techniques, such as micropipette aspiration, magnetic-bead cytometry, and microscopic laser-ablation, have been used to apply strain on the cell nucleus and to measure its bulk material properties16,17,18. However, until our recent work, no studies had directly shown that tensile forces are transferred directly onto an LINC complex protein in live cells2.
Previously, our laboratory demonstrated that a modified version of Nesprin-2G, known as mini-Nesprin2G, could be engineered to express a FRET-force probe known as TSmod (Figure 2C)2. TSmod was previously shown to dynamically measure tensile forces in live cells and in various load-bearing proteins3,4,6,8. It was further shown that mini-Nesprin-2G was indistinguishable from endogenous Nesprin-2G with respect to its localization, known effects on the cytoskeleton, and ability to rescue nuclear positioning11,19. Using the mini Nesprin-2G protein with the TSmod (Nesprin-TS), it was shown that NIH3T3 fibroblasts expressing the sensor have a baseline level of tension relative to Nesprin-HL, which cannot bind to actin2. Furthermore, the modulation of cellular contractility using inhibitors or activators of myosin can be detected by Nesprin-TS2.
There are a number of critical steps in this protocol. Confusing Nesprin-TS with Nesprin-HL (DNA or expressing cells) is not readily detected, as both similarly localize to the nuclear membrane, and will lead to confusing results. Since Nesprin-HL is a no-force control, the measured FRET should be higher than Nesprin-TS on average. To be cautious, sequencing isolated DNA constructs and carefully labeling cells is advised. Second, it is important to carefully optimize the detector gain and laser power on the confocal microscope to improve the signal-to-noise ratio and to minimize bleaching. To properly measure FRET, these parameters must remain constant during the image acquisition. Furthermore, optimizing the detector gain or power for dim or very bright expressing cells is suboptimal, because median expressing cells will tend to fall out of the dynamic range of the detector. Certain cells that express a high concentration of sensor tend to fluoresce in the nucleoplasm and endoplasmic reticulum, making it difficult to resolve the nuclear membrane, which is presumably the more interesting region with regards to force. Selecting cells with slightly lower expression levels tends to resolve this issue of sensor localization.
While ratiometric spectral imaging is a relatively simple and quick way to measure FRET, it does not output units of FRET efficiency due to acceptor signal bleed-through20. Without units of FRET efficiency, it is not possible to convert the FRET ratio index into a calibrated force measurement. Due to this limitation, only relative quantitative force comparisons can be made. FRET efficiency can be determined by more complicated methods, such as fluorescence-lifetime imaging microscopy (FLIM) or acceptor photobleaching. Absolute force (at the pN level) can be estimated from FRET efficiency, as previously described3,8.
In contrast to previous methods to measure the mechanical properties of the nucleus or the displacement of the nucleus based on externally applied forces, measuring the FRET from Nesprin-TS has the advantage of non-invasively probing tension on a component of the nuclear membrane16,17,18. The sensor outputs a fluorescent signal that is correlated with the strain on individual Nesprin-2G molecules. In contrast, micropipette aspiration requires the use of delicate tools when working with live cells. Confocal laser ablation causes local cellular damage and requires a pulsed laser source. Magnetic twisting cytometry transfers forces onto the nuclear membrane via the cytoskeleton and cannot be used to measure forces on specific proteins, unless it were to be coupled with the Nesprin-2G tension sensor.
While it has been demonstrated that actin-based tensile forces are transferred to Nesprin-2G, a component of the LINC complex, it remains unknown how much tensile or compressive forces are exerted onto other structural LINC proteins, such SUN-1 or SUN-2. Future work will be directed at cloning FRET-force probes into these putative load-bearing LINC proteins to further elucidate nuclear mechanics with protein-level resolution. Another issue that presents itself is that changes in FRET may be unrelated to changes in tension but rather reflect either changes in Nesprin-2G oligomerization or conformational changes of Nesprin-2G. Changes in cellular pH can influence the fluorescence of the sensor and change the measured FRET. The Nesprin-2G HL sensor represents a good control for some of these changes (as FRET changes observed with this sensor are most likely unrelated to force), and intermolecular FRET control constructs can be developed to assess oligomerization (see below). Additionally, the expression of nesprin-2G TS is driven by a CMV promoter, which results in overexpression, and this high level of expression may potentially induce biological artifacts. Recent advances with clustered regularly interspaced short palindromic repeats (CRISPR) have allowed for homology-directed repair (HDR) approaches to insert larger DNA sequences into the genome21. This approach could be used to modify endogenous Nesprin-2G (or other proteins) to include TSmod. This would provide expression of the force sensor using the appropriate endogenous promotor, reducing the potential for overexpression artifacts.
TSmod and other elastic FRET probes can also be used to develop new sensors for other proteins, which could then be used in a similar protocol as that described in this article. A major concern with the development of any new force sensor is determining the internal insertion site for the FRET sensor. First, the site of insertion must be in a region of the protein subjected to mechanical load. These regions can be identified by examining where cytoskeletal-connected proteins interact with the protein. Second, the insertion of the large (~50 kDa) TSmod must not disrupt the biological function of the protein being studied. A previously developed "mini" form of Nesprin-2G showed that actin-binding CH domains bridged to the SUN-binding KASH domain were sufficient for Nesprin-2G nuclear movement in wounded monolayers11,19, suggesting that the insertion of TSmod would likely not disrupt protein function. Confirmation of the biological function of the sensor requires a functional assay for protein function (in which the force sensor can be shown to rescue a specific function in cells depleted of endogenous protein). Third, changes in protein oligomerization can alter the FRET between proteins (termed intermolecular FRET22), which would result in FRET changes that are not related to force. A careful analysis of this often requires specialized intermolecular FRET constructs that specifically report oligomerization.
FRET-based force sensors have enabled the spatial and temporal analysis of mechanical forces with protein-specific resolution. Although this protocol details the use of the Nesprin-2G tension sensor, it is applicable to a number of existing sensors available for proteins in focal adhesions and cell-cell junctions. The use of these sensors for Nesprin-2G and other load-bearing proteins will provide significant insight into cell mechanics and mechanobiology.
The authors have nothing to disclose.
This work was supported by the Thomas F. and Kate Miller Jeffress Memoria Trust (to DEC) and NIH grant R35GM119617 (to DEC). The confocal microscope imaging was performed at the VCU Nanomaterials Characterization Core (NCC) Facility.
Nesprin-TS DNA | Addgene | 68127 | Retrieve from https://www.addgene.org/68127/ |
Nesprin-HL DNA | Addgene | 68128 | Retrieve from https://www.addgene.org/68128/ |
mTFP1 DNA | Addgene | 54613 | Retrieve from https://www.addgene.org/54613/ |
mVenus DNA | Addgene | 27793 | Retrieve from https://www.addgene.org/27793/ |
TSmod DNA | Addgene | 26021 | Retrieve from https://www.addgene.org/26021/ |
Competent Cells | Bioline | BIO-85026 | |
Liquid LB Media | ThermoFisher | 10855001 | https://www.thermofisher.com/order/catalog/product/10855001 |
Solid LB Bacterial Culture Plates | Sigma-Aldrich | L5667 | http://www.sigmaaldrich.com/catalog/product/sigma/l5667?lang=en®ion=US |
Ampicillin | Sigma | A9518 | |
Spectrophotometer | Biorad | 273 BR 07335 | SmartSpec Plus |
quartz cuvette | Biorad | 1702504 | Cuvette for SmartSpec Plus |
DNA isolation kit | Macherey-Nagel | 740412.5 | NucleoBond Xtra Midi Plus |
6-well cell culture dish | Falcon-Corning | 353046 | Multiwell 6-well Polystrene Culture Dish |
Dulbecco's Modified Eagle Medium, (DMEM) cell media | Gibco | 11995-065 | DMEM(1x) |
Bovine Serum | Life Technologies | 16170-078 | |
reduced serum cell media | Gibco | 31985-070 | Reduced Serum Medium, "optimem" |
Lipid Carrier Solution | invitrogen | 11668-019 | Lipid Reagent, "Lipofectamine 2000" |
1.5mL sterile plastic tube | Denville | c2170 | |
Trypsin | Gibco | 25200-056 | 0.25% Trypsin-EDTA (1x) |
glass-bottom microscope viewing dish | In Vitro Scientific | D35-20-1.5-N | 35mm Dish with 20mm Bottom Well #1.5 glass |
Fibronectin | ThermoFisher | 33016015 | fibronectin human protein,plasma |
Phosphate Buffered Saline (PBS) | Gibco | 14190-144 | Dulbecco's Phosphate Buffered Saline |
15mL sterile centrifuge tube | Greiner bio-one | 188261 | |
swinging rotor centrifuge | Thermo electron | Centra CL2 | Swinging rotor thermo electron 236 |
cell culture biosafety hood | Forma Scientific | 1284 | |
climate controlled cell culture incubator | ThermoFisher | 3596 | |
inverted LED widefield fluorescent microscope | Life technologies | EVOS FL | |
Clear HEPES buffered imaging media | Molecular Probes | A14291DJ | |
Fetal bovine Serum | Life technologies | 10437-028 | |
Temperature Controlled-Inverted confocal w/458 and 515nm laser sources | Zeiss | LSM 710-w/spectral META detector | |
Outgrowth Media | Newengland Biolabs | B9020s | |
NIH 3T3 Fibroblasts | ATCC | CRL-1658 |