Quantifying molecular- to cellular-level forces in living cells

Abstract

Mechanical cues have been suggested to play an important role in cell functions and cell fate determination, however, such physical quantities are challenging to directly measure in living cells with single molecule sensitivity and resolution. In this review, we focus on two main technologies that are promising in probing forces at the single molecule level. We review their theoretical fundamentals, recent technical advancements, and future directions, tailored specifically for interrogating mechanosensitive molecules in live cells.

Keywords: FRET; atomic force microscopy; conformational imaging; mechanobiology; molecular forces.

Figure 1.

General mechanism of AFM and force-distance (FD) curves. (a) AFM measures the interactions between the cantilever tip and the cell surface through laser deflection onto a photodiode. Shown here is the AFM scanning the tip across the cell to obtain a topography image. Single-molecule force-spectroscopy (SMFS) involves recording the force felt by the functionalized tip as it approaches and retracts from the surface of the cell. This interaction results in FD curves. Most common investigations include ligand-receptor interactions which separate with a single bond rupture. Reprinted by permission from Springer Nature Customer Service Centre GmbH: [Springer Nature] [Nature Chemical Biology] [5] (2009). (b), and membrane proteins that display a characteristic sawtooth pattern while unfolding. (c) Typical FD curves obtained from SMFS are shown with annotated images. Reprinted from [6], Copyright (2014), with permission from Elsevier.

Figure 2.

Exemplary studies using live cell AFM to observe functional changes of membrane proteins. ((a) – (d); blue) An example of MR-AFM on living cells to observe functional changes of membrane proteins. (a) Als5p proteins on living fungal cell surface were tagged with V5 epitopes. AFM tip was terminated with anti-V5 antibodies. The anti-V5 tip is capable of dual detection. Blue curves show single weak adhesion of the antibody to the V5 epitope. Red curves show unfolding of the Als5p domains via ligand binding. (b) Topographic image of the cell with two distinct marked locations. First, anti-V5 adhesion is mapped on an untouched cell (c). After some time, another adhesion map is performed over the same area of the cell (d). The nanoscale clusters of Als5p are outlined with dashed lines. ((e)–(h); magenta) peak-force tapping (PFT)-AFM performed on a Saccharomyces cerevisiae cell expressing His-tagged Mid2 sensors recorded with Ni²⁺-nitrilotriacetate (NTA) group functionalized cantilever tips (e). (f) Low resolution topographic image of living yeast cell. Figures (g) and (h) show adhesion force maps obtained simultaneously to topography. Green marks the detection of a single sensor. Figure (g) was generated with conventional force-volume imaging (20 min for 32 × 32 pixels). Figure (h) was of the same dimensions obtained using PFT-AFM (8 min for 512 × 512 pixels). Here we see the difference in speed and resolution of the two imaging modes. (e)–(h) Reprinted with permission from [21]. Copyright (2012) American Chemical Society. ((i)–(l); green) Demonstration of a bifunctionalized cantilever interacting with PAR1 binding sites. Two ligands were tethered to the tip: tris-NTA and a thrombin receptor-activating peptide. (i) Schematic demonstrates the approach (blue) and retraction (red) cycle to collect FD curves (j) AFM topography and simultaneous adhesion map (k) of PAR1 proteoliposome. The adhesion map displays three different force regions: the red signals the peptide-PAR1 bond and green is the tris-Ni(ii)-NTA-His10-tag bond and blue is the force filter. (l) Specific interactions detected in three consecutive recordings. Interaction sites are colored with the same key as before and numbers show the number of times detected. While this image was from a lipid bilayer on mica, this technique can also be applied on live cell membranes. (i)–(l) Reproduced from [46]. CC BY 4.0. ((m)–(p); yellow) Mapping virus binding on live Madin–Darby canine kidney (MDCK) cells expressing tumour virus receptor A (TVA) receptors with confocal microscopy and FD-based AFM. Figure (m) shows a schematic of the sinusoidal oscillation of the AFM cantilever during scanning. A sample FD curve of the recorded tip-sample interactions is shown. (n) TVA-mCherry confocal image showing MDCK cells expressing TVA receptors (red). Figures (o) and (p) are height and adhesion images respectively, acquired through AFM over the region outlined in a white box from (n). Fluorescence intensity of mCherry and adhesion detection were directly correlated. (m)–(p) Reprinted by permission from Springer Nature Customer Service Centre GmbH: [Springer Nature] [Nature Nanotechnology] [47] (2017).

Figure 3.

Different AFM modalities and their experiment outputs. (a) SCFS involves attaching a live cell to a tip-less cantilever. A diagram shows a typical SCFS measurement. The cantilever approaches the sample (i) and is put in contact (ii) for a period of time before retracting (iii) and (iv). In contrast with SMFS, cell probes experience the unbinding of multiple bonds and nanotubes before fully detaching from a substrate. Reprinted by permission from Springer Nature Customer Service Centre GmbH: [Springer Nature] [Nature Chemical Biology] [5] (2009). (b) Multiparametric imaging with SCFS probe. Corneocyte cells were imaged with Staphylococcus aureus bacteria on a cantilever. Left and right are height and adhesion images recorded with a S. aureus Newman cell probe respectfully. Reproduced from [73] with permission of The Royal Society of Chemistry. (c) Correlative AFM-Super resolution microscopy setup. The inverted optical microscope is mounted on an x/y-translation stage. The optical path is aligned with the AFM cantilever by adjusting the stage. (d) AFM/photo activated light microscopy (PALM) images of a living CHO-K1 cell. Left is a time-resolved AFM topography image showing filopodia protrusion with lamellopodia extension. Right is a time-resolved PALM image showing the reorganization of the paxillin-mEos2 clusters in a zoomed in area of the AFM image. Reprinted with permission from [74]. Copyright (2015) American Chemical Society. (e) Schematic of high-speed (HS)-AFM configured for tapping mode imaging. Feedback control is optimized to minimize delay time and the xyz-scanner is modified for fast scanning, among other improvements. Reprinted from [78], Copyright (2018), with permission from Elsevier. (f) Time series of topographical images of living hippocampal neurons. White arrows indicate growing filopodia on the dentrite. Reproduced from [75]. CC BY 4.0. (g) Schematic of experiment setup for scanning near field ultrasonic holography (SNFUH). The probe and sample are oscillated at ultrasonic frequencies f _c and f _s respectively. Deflection signal from the cantilever is collected using lock-in amplification referenced at |f _c – f _s|. Reprinted from [77], Copyright (2010), with permission from Elsevier. (h) High-resolution images of nanoparticles in fixed red blood cells. AFM topography (left) and SNFUH phase images (right) show particles of size 80–100 nm resolved through SNFUH but not tapping mode AFM. Reprinted by permission from Springer Nature Customer Service Centre GmbH: [Springer Nature] [Nature Nanotechnology] [76] (2008).

Figure 4.

The fundamental physics of FRET and various genetically encoded tension sensors. (a) Fluorescence resonance energy transfer (FRET) is a non-radiative transfer of energy from a donor molecule to an acceptor molecule. During donor excitation, electrons are excited to higher energy levels as the energy due to vibrational movements of the molecule quickly relaxes. If the conditions for FRET are met, the energy from the excited donor will non-radiatively (no emission of a photon) transfer to the acceptor. The acceptor excitation results in electrons moving to higher energy levels before relaxing and emitting a photon, this emission from the acceptor is called ‘sensitized emission’. Measurement can made by observing the decline in fluorescence lifetime of the donor. This way of measuring FRET is among many techniques. Reproduced from [132]. CC BY-SA 3.0. (b) The orientation factor, κ², is a value that represents the relative orientation between the emission dipole of the donor and the absorbance dipole the acceptor. This value is often difficult to measure as the dipoles can be continuously rotating and therefore is typically assigned a value of 2/3 (although this assignation is not always appropriate.). θ_D and θ_A are the angles between the dipoles with an axis ‘r’ joining the two. φ is the angle between the two plane on which the dipoles lie. Reproduced from [133]. CC BY 4.0. (c) The Forster distance (R₀), the distance at which 50% of the energy is transferred, has an inverse sixth dependence- illustrating that FRET is position-dependent. FRET is most effective in a 10–100 angstrom range. Distances less than 10 A are subject to bleed-through from other energy transfer mechanisms and distances over 100 A have very little efficiency, essentially useless for many experiments. The Forster distance is dependent on the relative orientation of the transition dipoles and the spectral overlap between the emission of the donor and the absorbance of the acceptor. It is important to choose donors and acceptors with the correct R₀ to respect whatever distance is desired in the experiment. For example, spectral overlap can be adjusted by choosing the right donors and acceptors. The emission spectrum of the donor and the absorbance spectrum of the acceptor must overlap to cause FRET. The larger the overlap, calculated by the integral J, the larger the energy transfer, and thus a larger R₀. Reproduced with permission from [130]. (d) The tension sensor module (TSM) consists of two fluorophores (donor and acceptor) attached by a flexible linker peptide that extends under tension when encoded into proteins. (e) The TSM is placed in a protein of interest (POI), particularly in a place where the function and structure of the POI will be conserved. (f) To ensure changes in FRET are only dependent on the tension of the protein, a series of zero-force controls can be put in place. (g) The wild-type (wt) control of only the donor is necessary to measure fluorescence intensity of the donor in the absence of the acceptor (another requirement for the FRET efficiency calculation). (h) The intermolecular forces affecting the TSM can be measured by pairing the donor wt- control with the analogous acceptor-only wt-control. When both wt-controls are utilized and placed in the cell, effects of intermolecular force can be exposed and accounted for. Reproduced from [131]. CC BY 4.0.

Figure 5.

FRET Images in cell–cell junctions, cell-matrix adhesions, and cell cortex. (a) Adjacent MCDK cells with a custom genetic tension sensor produce the FRET emission on the right. Background and bleed-through are corrected on the right [142]. (b) Bovine aortic endothelial cell (BAEC) junctions with custom tension sensor, depicted with reduced background and an index image from gray scale. Reproduced from [143]. CC BY 4.0. (c) Triangular pattern of talin in cell-matrix imaged, forces were larger for triangular shaped arrangements as compared to round. Reprinted by permission from Springer Nature Customer Service Centre GmbH: [Springer Nature] [Nature Methods] [145] (2017). (d) In BAECs in cell-matrix, vinculin show low FRET ratios in focal adhesions near protruding edges correlating to higher tensions. The retracted regions have higher FRET rations indicating lower tensions. Reprinted by permission from Springer Nature Customer Service Centre GmbH: [Springer Nature] [Nature] [146] (2010). (e) FRET index of Beta-spectrin (a molecule that can cause defects in neuron morphology) captured by FRET of a donor-excited teal fluorescent protein and Venus acceptor. The boxed regions represent areas of interest regarding mechanical tension sensation being dependent on spectrin in the cell cortex. To the right is a fluorescence lifetime imaging microscopy of the UNC-70 tension sensor used to image the Beta-spectrin. Reproduced from [144]. CC BY 4.0. (f) BAEC’s FRET images by a circularly permutated single stranded FRET fluorescence. By using circularly permutated Cerulean and Venus monomers and taking advantage of orientation factor in FRET, images of BAEC in the cell cortex. This can be compared to the middle image where spectrin is attached to the C-terminal where there is no tension exerted. Similarly, the circularly permutated single stranded FRET fluorescence images can be compared to spectrin where the N and C terminal are next to each other. Reproduced from [147]. CC BY 4.0.