Single-Molecule Measurement of Protein Interaction Dynamics Within Biomolecular Condensates

Shawn R. Yoshida; Shasha Chong

doi:10.3791/66169

JoVE Journal > Biology

Biology

Single-Molecule Measurement of Protein Interaction Dynamics Within Biomolecular Condensates

Published: January 05, 2024

doi:

10.3791/66169

Shawn R. Yoshida^1,2, Shasha Chong¹

¹Division of Chemistry and Chemical Engineering,California Institute of Technology, ²Division of Biology and Biological Engineering,California Institute of Technology

Summary

Many intrinsically disordered proteins have been shown to participate in the formation of highly dynamic biomolecular condensates, a behavior important for numerous cellular processes. Here, we present a single-molecule imaging-based method for quantifying the dynamics by which proteins interact with each other in biomolecular condensates in live cells.

Abstract

Biomolecular condensates formed via liquid-liquid phase separation (LLPS) have been considered critical in cellular organization and an increasing number of cellular functions. Characterizing LLPS in live cells is also important because aberrant condensation has been linked to numerous diseases, including cancers and neurodegenerative disorders. LLPS is often driven by selective, transient, and multivalent interactions between intrinsically disordered proteins. Of great interest are the interaction dynamics of proteins participating in LLPS, which are well-summarized by measurements of their binding residence time (RT), that is, the amount of time they spend bound within condensates. Here, we present a method based on live-cell single-molecule imaging that allows us to measure the mean RT of a specific protein within condensates. We simultaneously visualize individual protein molecules and the condensates with which they associate, use single-particle tracking (SPT) to plot single-molecule trajectories, and then fit the trajectories to a model of protein-droplet binding to extract the mean RT of the protein. Finally, we show representative results where this single-molecule imaging method was applied to compare the mean RTs of a protein at its LLPS condensates when fused and unfused to an oligomerizing domain. This protocol is broadly applicable to measuring the interaction dynamics of any protein that participates in LLPS.

Introduction

A growing body of work suggests that biomolecular condensates play an important role in cellular organization and numerous cellular functions, e.g., transcriptional regulation¹^,²^,³^,⁴^,⁵, DNA damage repair⁶^,⁷^,⁸, chromatin organization⁹^,¹⁰^,¹¹^,¹², X-chromosome inactivation¹³^,¹⁴^,¹⁵, and intracellular signaling¹⁶^,¹⁷^,¹⁸. In addition, the dysregulation of biomolecular condensates is implicated in many diseases, including cancers¹⁹^,²⁰^,²¹ and neurodegenerative disorders²²^,²³^,²⁴^,²⁵^,²⁶. Condensate formation is often driven by transient, selective, and multivalent protein-protein, protein-nucleic acid, or nucleic acid-nucleic acid interactions²⁷. Under certain conditions, these interactions can lead to liquid-liquid phase separation (LLPS), a density transition that locally enriches specific biomolecules in membraneless droplets. Such multivalent interactions are often mediated by the intrinsically disordered regions (IDRs) of proteins¹^,²⁸^,²⁹. Biophysical characterization of these interactions at the molecular level is critical to our understanding of numerous healthy and aberrant cellular functions, given the pervasiveness of condensates across them. Although techniques based on confocal fluorescence microscopy, e.g., fluorescence recovery after photobleaching (FRAP)³⁰^,³¹^,³², have been widely used to qualitatively show that the molecular exchanges between condensates and the surrounding cellular environment are dynamic, quantifying the interaction dynamics of specific biomolecules within condensates is generally not possible using conventional confocal microscopy or single-molecule microscopy without specialized data analysis methods. The single-particle tracking (SPT) technique described in this protocol is based on live-cell single-molecule microscopy³³ and provides a uniquely powerful tool to quantify the interaction dynamics between specific proteins within condensates. The readout of SPT for such measurement is the mean residence time of a protein of interest in the condensates.

The protocol can be broken down into two parts – data acquisition and data analysis. The first step of imaging data acquisition is to express in cells a protein of interest that is fused to a HaloTag³⁴. This enables labeling of the protein of interest with two fluorophores, where a majority of the protein molecules are to be labeled with a non-photoactivatable fluorophore (e.g., JFX549 Halo ligand³⁵) and a small fraction of them are to be labeled with a spectrally distinct, photoactivatable fluorophore (e.g., PA-JF646 Halo ligand³⁶). This allows for the simultaneous acquisition of all condensate locations in the cell and the acquisition of single-molecule movies of the protein of interest binding and unbinding to the condensates. Meanwhile, the same type of cells are modified to stably express Halo-tagged H2B, a histone that is largely immobile on chromatin. The cells are then stained with the PA-JF646 Halo ligand to enable single-molecule imaging of H2B. As will be discussed in detail below, this experiment accounts for the contribution of photobleaching to enable precise quantification of the interaction dynamics of the protein of interest. Cells for imaging experiments must then be cultured on clean coverslips, stained with HaloTag ligand(s), and assembled into a live-cell imaging chamber. From there, the sample is imaged under highly inclined and laminated optical sheet (HILO) illumination on a total internal reflection fluorescence (TIRF) microscope capable of two-channel imaging and single-molecule detection. The emission is then split onto two cameras, one tracking condensate positions and one tracking single molecules. Acquisition is performed with a long integration time (on the order of hundreds of ms) to blur out freely-diffusing proteins and only capture proteins that are less mobile due to binding to stable structures in the cell³⁷.

The first step of data analysis is using an established single-particle tracking (SPT) algorithm³⁸^,³⁹ to localize individual protein molecules in each frame of the movie and assemble the localizations into a trajectory for each molecule over its detectable lifetime. The trajectories are then sorted into those representing molecules inside and those representing molecules outside the condensates by comparing the localizations of the molecules throughout their trajectories to the localizations of all the condensates at the corresponding times¹.

Next, a survival curve (1 – CDF) is generated using the lengths of all the in-condensate trajectories. The apparent mean residence time of the molecules is then extracted by fitting the survival curve to the following two-component exponential model of protein binding,

,

with A as the fraction of molecules non-specifically bound and with k_obs,ns and k_obs,s as the observed dissociation rates of the non-specifically bound and specifically bound molecules, respectively. Only k_obs,s is considered from here onward. The dynamics of both protein dissociation, k_true,s, and photobleaching of the fluorophore, k_pb, contribute to k_obs,s as

;

thus, to isolate the effects of protein dissociation, the specific dissociation rate of H2B-Halo in the cell line mentioned prior is measured.

H2B is a protein that is stably integrated into chromatin and that experiences minimal dissociation in the time scale of a single-molecule movie acquisition³⁷. Its specific dissociation rate is then equal to the photobleaching rate of the PA-JF646 Halo ligand, or

.

The mean in-condensate residence time of the protein of interest, , is then

.

Representative results from Irgen-Gioro et al.⁴⁰ are shown, where this protocol was applied to demonstrate that fusing an oligomerization domain to IDR results in longer residence times of the IDR in its condensates. This result suggests that the added oligomerization domain stabilizes the homotypic interactions of the IDR that drives LLPS. In principle, the same method with slightly modified protocols can be applied to characterize the homotypic or heterotypic interactions of any protein that participates in the formation of any types of condensates.

Protocol

1. Labeling of proteins in cells

Express the protein of interest fused to HaloTag in the desired cell line.
Stably express Halo-tagged H2B in the same type of cells as in 1.1 using transposons or viral transduction.

2. Preparation of coverslips

Before using coverslips for cell culture, clean coverslips to remove autofluorescent contaminants.
1. Mount 25 mm diameter, #1.5 coverslips on a ceramic staining rack and place into a polypropylene container.
2. Completely immerse coverslips in a 1 M solution of KOH and sonicate for 1 h.
3. Rinse coverslips three times with double-distilled water (ddH₂O).
4. Completely immerse coverslips in 100% ethanol and sonicate for 1 h.
5. Completely immerse coverslips in 20% ethanol diluted with ddH₂O and store at 4 °C until use.

3. Preparation of cells for microscopy

NOTE: Perform steps from this section in a biosafety cabinet to prevent contamination of the cells.

Plate cells on cleaned coverslips. Perform 2 days before imaging if the Halo-tagged protein is to be transiently expressed. Perform 1 day before imaging if the Halo-tagged protein is stably or endogenously expressed.
1. Use sterile forceps to place coverslips in a 6-well plate (one coverslip/well per cell sample) and allow residual ethanol to evaporate.
2. Plate each well with an appropriate number of cells to achieve 70% confluency by the time of imaging. Plate an additional well with cells stably expressing H2B-Halo with the same target confluency.
3. Follow this step only if transiently expressing the Halo-tagged protein of interest. Incubate cells for 24 h at 37 °C and 5% CO₂. Then, transfect the cells with the plasmid encoding the tagged protein of interest using appropriate transfection reagents and following the manufacturer's guidelines.
4. Incubate the cells for 24 h at 37 °C and 5% CO₂.
Stain the cells expressing the Halo-tagged protein of interest with both a non-photoactivatable Halo ligand (e.g., JFX549) and a photoactivatable Halo ligand (e.g., PA-JF646). Stain cells expressing H2B-Halo with only the photoactivatable Halo ligand.
NOTE: Concentrations of Halo ligands listed here should be used as a starting point; the optimal concentration will depend on the expression level and in-condensate local concentration of the protein of interest.
1. For each cell sample expressing the Halo-tagged protein of interest, dilute 100 nM JFX549 Halo ligand³⁵ and 20 nM PA-JF646 Halo ligand³⁶ in 500 µL of appropriate cell media. For each sample expressing H2B-Halo, dilute only 20 nM PA-JF646 Halo ligand in 500 µL of appropriate cell media.
2. For each well, aspirate existing media, add 2 mL of 1x PBS, then aspirate PBS (this is referred to as a "rinse" for the remainder of the protocol), and replace with media containing the appropriate Halo ligands.
3. Incubate for 1 h at 37 °C and 5% CO₂.
4. Rinse with PBS four times and replace with fresh media containing no Halo ligands.
5. Incubate for 15 min at 37 °C and 5% CO₂.
6. Repeat steps 3.2.4 and 3.2.5 three times (four rinse-incubation cycles total).
7. Use sterile forceps to transfer the coverslip with cells to a cell-culture coverslip chamber compatible with the microscope stage.
8. Add phenol red-free media to the chamber and proceed to imaging. Incubate the samples not immediately being imaged at 37 °C and 5% CO₂ until needed.

4. Single-molecule imaging

NOTE: Independent experiments measuring the residence times of both H2B and the protein of interest should be conducted across multiple (≥3) days to generate statistically significant results. Before imaging cell samples on the microscope, align the two cameras using 0.1 µm stained microspheres (Table of Materials) or a similar calibration standard.

Prepare the microscope.
1. Turn on the microscope and the computer that controls the microscope.
2. Turn on live-cell incubation components (heater and CO₂) on the microscope and set it to 37 °C and 5% CO₂. Allow the system to equilibrate.
3. Put a drop of the appropriate oil onto a 100x TIRF oil immersion objective on the microscope and load the cell sample onto the microscope stage.
Identify a cell to image.
1. Image the cell sample under brightfield illumination and adjust the z-position until the cells are in focus.
2. Identify a cell that exhibits morphology characteristics of healthy cells.
3. Crop the field of view (FOV) so that it only captures the target cell nucleus.
4. Use laser illumination and adjust the TIRF angle until HILO illumination⁴¹ with optimal signal-to-noise ratio of single molecules is achieved.
Image H2B-Halo in live cells.
NOTE: The laser powers used in this section are specific to this experiment and included as an example. Appropriate laser powers should be chosen according to the criteria listed in the discussion.
1. Set up an acquisition configuration as follows. Set the exposure time as 500 ms. During each 500 ms exposure, excite PA-JF646 labeled H2B-Halo molecules with a 9.1 mW, 640 nm beam. During the dead time between frames (158.01 µs on the imaging system used here), photoactivate the molecules with a 111 µW, 405 nm beam.
2. Capture 2000 frames continuously under these settings. Gradually increase the 405 nm beam's power over the course of the acquisition when the number of photoactivated molecules becomes insufficient.
Image the Halo-tagged protein of interest in live cells.
1. Use a long-pass dichroic mirror to split emission wavelengths between two cameras, one to detect PA-JF646 and the other to detect JFX549. Use appropriate emission filters in front of the cameras.
2. Use the same acquisition configuration described in 4.3.1 with the following modification. Add an additional JFX549 channel to track locations of the Halo-tagged protein condensates over time. Every 10 s, acquire one frame (500 ms, 2.1 mW, 561 nm excitation) in the JFX549 channel while keeping the acquisition of the PA-JF646 channel. This will cause bleedthrough into the PA-JF646 channel, which will be taken care of in the post-acquisition data analysis.
3. Capture 2000 frames continuously under these settings. Gradually increase the 405 nm beam's power over the course of the acquisition when the number of photoactivated molecules becomes insufficient.

5. Analysis of single-molecule imaging data

NOTE: The parameters used throughout section 5 are specific to this experiment and included as an example. Appropriate parameters should be chosen according to the criteria listed in the discussion.

Prepare raw imaging data for processing.
1. For the data of the protein of interest, convert each channel from the raw imaging data file into an independent .tif file using ImageJ or similar image processing software. For the data of H2B, convert the single channel from the raw imaging data file into a .tif file in the same way.
  NOTE: Depending on the acquisition software being used, there can be empty frames in the condensate movie, as only one condensate-tracking frame is taken every 10 s. The empty frames should be populated with the most recent condensate frame (some acquisition software automatically does this). Additionally, if there is significant bleedthrough from the condensate (JFX549) channel into the single-molecule (PA-JF646) channel during those condensate-tracking frames, the corresponding single-molecule frames should be replaced with the most recent single-molecule frame.
2. If there are any frames that need to be replaced in either movie due to the above reasons, replace them with the most recent frame from that movie by modifying (if necessary) and running pretracking_comb.txt, an ImageJ macro available in Chong et al.¹, and following the prompts.
Generate single-particle trajectories using an SPT software, e.g., SLIMfast³⁹, a GUI implementation of the MTT algorithm³⁸ available in Teves et al.⁴².
1. Load the file from 5.1.2 containing the single-molecule movie in SLIMfast by clicking on Load > Imagestack.
2. Set the parameters (localization error rate: 10^-6; deflation loops: 3) for localization by clicking OPT > Sheet: Localization.
3. Set the parameters (peak emission: 664 nm; lag time: 500 ms) for acquisition by clicking OPT > Sheet: Acquisition.
4. Visualize the localizations of all the molecules in a single frame to ensure that the parameters chosen are appropriate (TEST LOC). If inappropriate, modify parameters in steps 5.2.2 and 5.2.3 and repeat this step.
5. Generate a file containing the localizations of all the molecules in every frame (LOC ALL).
6. Load the file from 5.2.5 in SLIMfast by clicking Load > Particle Data > SLIMfast.
7. Set the parameters (maximum expected diffusion coefficient: 0.1 µm²/s; maximum number of competitors: 5) for trajectory generation by clicking OPT > Sheet: Tracking.
8. Generate a file containing the trajectories of all the molecules (GEN TRAJ).
Filter out the trajectories that are shorter than t_min, 2.5 s, using evalSPT³⁹, available in Drosopoulos et al.⁴³, to account for tracking errors that result in artificially short trajectories.
1. Load the file from 5.2.8 into evalSPT (+ > file from 5.2.8 > OK).
2. Set the parameters (min: 5 frames; max: maximum trajectory length for file from 5.2.8) to filter out the trajectories shorter than 2.5 s.
3. Generate a file with all the filtered trajectories (EXPORT DATA).
Follow this step for the trajectories of the protein of interest only. Sort the trajectories into those in the condensates and out of the condensates, then extract the mean in-condensate residence time.
NOTE: All the codes for this section are available in Chong et al.¹.
1. Threshold all the frames of the movie acquired in the JFX549 channel to generate a time-lapse movie populated with the time-evolving binary mask highlighting condensate locations by running the ImageJ macro, nucleus, and cluster mask_v2.txt, and following the prompts.
2. Reformat the single-molecule trajectories generated in 5.3.3. using the MATLAB script, ConvertASCII_SlowTracking_css3.m, and adjusting the filenames, paths, and parameters as necessary. (Parameters: Exposure, 0.5 s; Resolution, 0.16 µm/pixel).
3. Sort trajectories based on the fraction of the lifetime a given molecule spends in a condensate, F, using the MATLAB script, categorization_v4.m, and adjusting the filenames and paths as necessary.
4. Proceed using only in-condensate trajectories.
Extract k_obs,s and k_pband compute the corrected mean residence time, .
1. Extract k_obs,s from the in-condensate protein of interest trajectories using the MATLAB script, PLOT_ResidenceHist_css.m, and adjusting filenames, paths, and parameters as necessary.
  (Parameters: Exposure, 0.5 s; StartFrameForFit, 5 frames).
2. Extract k_pbfrom the H2B trajectories using the MATLAB script, PLOT_ResidenceHist_css.m, and adjusting filenames, paths, and parameters as necessary.
  (Parameters: Exposure, 0.5 s; StartFrameForFit, 5 frames).
3. Calculate the corrected mean residence time of the protein of interest specifically bound to its condensates, , as
  
  NOTE: Controls should be performed to verify that different exposure times that are long enough to blur out mobile particles, for example, 300 ms and 800 ms, still result in the same mean residence time of the protein of interest.

Representative Results

Here, we present representative results from Irgen-Gioro et al.⁴⁰, where we used this SPT protocol to compare the interaction dynamics of two proteins in their respective self-assembled LLPS condensates. TAF15 (TATA-box binding protein associated factor 15) contains an IDR that can undergo LLPS upon overexpression in human cells. We hypothesized that fusing TAF15(IDR) to FTH1 (ferritin heavy chain 1), which forms a 24-subunit oligomer, would lead to more stable homotypic protein-protein interactions that drive LLPS. To test this hypothesis, we transiently expressed in U2OS cells either Halo-TAF15(IDR) or TAF15(IDR)-Halo-FTH1 fusion and performed two-color single-molecule imaging of each protein following the above protocol. A representative frame from a TAF15(IDR)-Halo-FTH1 movie is shown in Figure 1A. Molecules detected in the PA-JF646 channel were localized, and those localizations were assembled into trajectories. The resulting trajectories were then sorted between in-condensate and out-of-condensate populations by comparison against a binary mask highlighting condensate locations (Figure 1B). While we only show a small fraction of the trajectories for good visibility, there is a clear distinction between the trajectories of molecules bound to the condensates and bound outside the condensates. Next, a survival curve of in-condensate trajectory lengths was fitted against the two-component exponential model (Figure 1C). Finally, mean residence times after correction for photobleaching were extracted and plotted for both proteins (Figure 1D). We found that the mean residence time of Halo-TAF15(IDR) in its LLPS condensates was 10.23 s ± 1.10 s while that of TAF15(IDR)-Halo-FTH1 was 64.15 s ± 11.65 s. This result suggests that the addition of the oligomerizing domain to TAF15(IDR) does indeed stabilize protein-condensate binding.

Figure 1: SPT-based method resolves differences in the residence times of proteins in their condensates. A. Representative frames from a two-color single-molecule movie of TAF15(IDR)-Halo-FTH1. Proteins were labeled with a combination of a higher concentration of non-photoactivatable dye (100 nM JFX549, yellow) to visualize the locations of condensates, and a lower concentration of photoactivatable dye (20 nM PA-JF646, magenta) to visualize individual proteins, enabling SPT. Movies under the same imaging conditions were acquired for Halo-TAF15(IDR) and single-channel, PA-JF646 movies were acquired for H2B-Halo. A white dashed line outlines the nucleus. B. Representative merged image overlaying a binary mask of condensate position (grey) and trajectories (multicolored). C. Representative survival curve of TAF15(IDR)-Halo-FTH1 fitted with the two-component exponential model. D. The mean residence time of Halo-TAF15 in was significantly shorter than that of TAF15(IDR)-Halo-FTH1 in their respective condensates. The value for each protein was averaged from 20 cells measured in independent experiments performed across three days. Error was propagated as standard error of the mean and the asterisk represents a significant difference in the residence time of the two proteins (p<0.05, Wilcoxon rank-sum test). This figure has been adapted with permission from Irgen-Gioro et al.⁴⁰. Please click here to view a larger version of this figure.

Discussion

The protocol as presented here is designed for systems like those investigated in Irgen-Gioro et al.⁴⁰. Depending on the application, some components of the protocol can be modified, e.g., the method for generating fluorescently labeled cell lines, the fluorescent labeling system, and the style of coverslip used. Halo-tagging of a protein in a cell can be done using two strategies, depending on which is more suitable for a given experiment. 1) Exogenous expression: fusing the protein of interest to a HaloTag and expressing the fusion in a target cell line either transiently using transfection or stably using transposons or viral transduction. 2) Genome editing: knock in a HaloTag at the gene locus encoding the endogenous protein of interest in a target cell line using genome editing techniques, e.g., CRISPR⁴⁴. The benefits and drawbacks of each are discussed elsewhere⁴⁵, but briefly, the exogenous expression strategy is less time-consuming but produces a population of cells with a broad range of expression levels that are often non-physiological. The genome editing strategy takes much longer, but labels endogenously expressed proteins of interest at their native expression levels.

In addition, this protocol does not specifically require the use of a HaloTag; rather, it is compatible with any tag that enables simultaneous single-molecule and ensemble labeling of the same protein in the cell. Thus, the protocol can be modified for use with other self-labeling tags like SNAP-tag⁴⁶. Finally, precleaned, commercial glass-bottom dishes for cell culture, such as MatTek dishes (MatTek Life Sciences, P35G-1.5-20-C), can be used instead of coverslip chambers provided that they are compatible with the microscope objective and immersion oil; however, coverslips can be cleaned more thoroughly immediately before use, thus are preferable.

While the above modifications to the protocol are optional, there are experimental and analysis parameters that must be optimized, namely, the concentrations of HaloTag ligands, the laser powers, the localization and tracking parameters, and t_min. The concentration of the non-photoactivatable HaloTag ligand should be chosen such that the condensate is labeled densely enough to enable mask generation. The concentration of the photoactivatable HaloTag ligand and the photoactivation laser power should be chosen such that protein molecules are labeled and photoactivated sparsely enough to generate quality single-particle trajectories, as an excess of localizations in each frame will result in inaccurate trajectory generation. For experiments shown in the representative results, there were generally fewer than five localizations per frame. The excitation laser power should be kept low enough to minimize rapid photobleaching but high enough to localize single molecules precisely. The single-molecule localization and tracking parameters should be adjusted depending on the density of excitations and the expected diffusion dynamics of the protein of interest. Finally, the values of should be computed across a range of t_min (starting from 0 s and increasing in increments of exposure time). The values of should converge above some threshold of t_min; this t_minshould be used in step 5.3.

The method outlined here quantifies the interaction dynamics of proteins within condensates with a precision that is inaccessible to conventional techniques. In addition, it can do so in live cells with minimal sample perturbation. This is critical, as the interaction behaviors of IDRs, including their condensate formation, are highly dependent on their local environment⁴⁰.

The representative results from Irgen-Gioro et al.⁴⁰ demonstrate this method's ability to extract residence times for proteins binding to their self-assembled LLPS condensates. Importantly, this method can be easily expanded to any protein binding to any type of condensates through homotypic or heterotypic interactions. Moreover, it is not limited to measuring the interaction dynamics of proteins undergoing LLPS. The fusion oncoprotein EWS::FLI1, which is known to cause Ewing Sarcoma, has been shown to form local, high-concentration hubs that play an essential role in its transcriptional activation and oncogenic transformation functions¹^,². While the formation of these hubs is driven by transient, selective, and multivalent IDR-IDR interactions of EWS::FLI1, so far, there is still no convincing evidence that they are bona fide LLPS condensates¹^,². Even so, we used the method presented here to measure the mean residence time of EWS::FLI1 at its hubs and showed that the mutation of specific residues and the deletion of its IDR significantly destabilized the binding of EWS::FLI1 to its hubs¹.

Despite the versatility of this method in characterizing the binding dynamics of proteins within condensates, caution should be exercised when choosing the model for protein binding in specific contexts. Existing biochemical data supports the idea that a two-component exponential distribution is often an appropriate model for many systems¹^,³⁷^,⁴⁰^,⁴², but the same distribution has also been shown to be an inappropriate representation of protein binding in other systems where alternative models such as three-component exponential⁴⁷^,⁴⁸^,⁴⁹, and power-law distributions⁵⁰ better match the experimental observations. To avoid drawing inaccurate conclusions, one should carefully choose and motivate a model, verify that the quality of fit supports the use of the model, and judiciously interpret the results.

Disclosures

The authors have nothing to disclose.

Acknowledgements

This work was supported by the National Science Foundation Graduate Research Fellowship under Grant No. DGE-1745301 (S.Y.), Pew-Stewart Scholar Award (S.C.), Searle Scholar Award (S.C.), the Shurl and Kay Curci Foundation Research Grant (S.C.), Merkin Innovation Seed Grant (S.C.), the Mallinckrodt Research Grant (S.C.), and the Margaret E. Early Medical Research Trust 2024 Grant (S.C.). S.C. is also supported by the NIH/NCI under Award Number P30CA016042.

Materials

0.1 µm TetraSpeck microsphere	Invitrogen	T7279	Single-molecule imaging
25 mm Diameter, #1.5 Coverslips	Marienfeld Superior	111650	Preparation of coverslips
593/40 nm bandpass filter	Semrock	FF01-593/40-25	Single-molecule imaging
676/37 nm bandpass filter	Semrock	FF01-676/37-25	Single-molecule imaging
6-Well TC Plate	Genesee	25-105MP	Preparation of cells for microscopy
Cell Line: U-2 OS	ATCC	HTB-96	Labeling of proteins in cells
ConvertASCII_SlowTracking_css3 .m			Analysis of single-molecule imaging data: Available in Chong et al., 2018
Coverglass Staining Rack	Thomas	24957	Preparation of coverslips
Deuterated Janelia Fluor 549 (JFX549)	Janelia Research Campus		Preparation of cells for microscopy
DMEM, Low Glucose	Gibco	10-567-022	Labeling of proteins in cells: Growth media used: DMEM with 5% fetal bovine serum, 1% penstrep
Eclipse Ti2-E Inverted Microscope	Nikon		Single-molecule imaging
Ethanol 200 Proof	Lab Alley	EAP200-1GAL	Preparation of coverslips
evalSPT			Analysis of single-molecule imaging data: Available in Drosopoulos et al., 2020
Fetal Bovine Serum	Cytiva	SH30396.03	Labeling of proteins in cells: Growth media used: DMEM with 5% fetal bovine serum, 1% penstrep
Fiji			Analysis of single-molecule imaging data
Ikon Ultra CCD Camera	Andor	X-13723	Single-molecule imaging
Longpass dichroic beamsplitter	Semrock	Di02-R635-25×36	Single-molecule imaging: Red/Far Red beamsplitter
LUN-F Laser Unit	Nikon		Single-molecule imaging: 405/488/561/640
MatTek glass-bottom dish	MatTek	P35G-1.5-20-C	Preparation of cells for microscopy: 35 mm, #1.5 coverslip dish for cell culture.
NIS-Elements	Nikon		Single-molecule imaging: Microscope acquisition software
nucleus and cluster mask_v2.txt			Analysis of single-molecule imaging data: Available in Chong et al., 2018
Penicillin-Streptomycin	Gibco	15-140-122	Labeling of proteins in cells: Growth media used: DMEM with 5% fetal bovine serum, 1% penstrep
Phosphate Buffered Saline	Thermo Fisher Scientific	18912014	Labeling of proteins in cells
Photoactivatable Janelia Fluor 646 (PA-JF646)	Janelia Research Campus		Preparation of cells for microscopy
PLOT_ResidenceHist_css.m			Analysis of single-molecule imaging data: Available in Chong et al., 2018
Potassium Hydroxide	Mallinckrodt Chemicals	6984-06	Preparation of coverslips
pretracking_comb.txt			Analysis of single-molecule imaging data: Available in Chong et al., 2018
SLIMfast			Analysis of single-molecule imaging data: Available in Teves et al., 2016
Stage-top incubation system	Tokai Hit		Single-molecule imaging: For live-cell imaging
TwinCam dual emission image splitter	Cairn Research		Single-molecule imaging
Ultrasonic Cleaner	Branson	5800	Preparation of coverslips

References

Chong, S., et al. Imaging dynamic and selective low-complexity domain interactions that control gene transcription. Science. 361 (6400), (2018).
Chong, S., Graham, T. G. W., Dugast-Darzacq, C., Dailey, G. M., Darzacq, X., Tjian, R. Tuning levels of low-complexity domain interactions to modulate endogenous oncogenic transcription. Molecular Cell. 82 (11), 2084-2097 (2022).
Boehning, M., et al. RNA polymerase II clustering through carboxy-terminal domain phase separation. Nature Structural & Molecular Biology. 25 (9), 833-840 (2018).
Sabari, B. R., et al. Coactivator condensation at super-enhancers links phase separation and gene control. Science. 361 (6400), (2018).
Boija, A., et al. Transcription factors activate genes through the phase-separation capacity of their activation domains. Cell. 175 (7), 1842-1855 (2018).
Levone, B. R., et al. FUS-dependent liquid-liquid phase separation is important for DNA repair initiation. Journal of Cell Biology. 220 (5), e202008030 (2021).
Kilic, S., et al. Phase separation of 53BP1 determines liquid-like behavior of DNA repair compartments. The EMBO Journal. 38 (16), e101379 (2019).
Pessina, F., et al. Functional transcription promoters at DNA double-strand breaks mediate RNA-driven phase separation of damage-response factors. Nature Cell Biology. 21 (10), 1286-1299 (2019).
Nozaki, T., et al. Condensed but liquid-like domain organization of active chromatin regions in living human cells. Science Advances. 9 (14), (2023).
Maeshima, K., et al. Nucleosomal arrays self-assemble into supramolecular globular structures lacking 30-nm fibers. The EMBO Journal. 35 (10), 1115-1132 (2016).
Strickfaden, H., Tolsma, T. O., Sharma, A., Underhill, D. A., Hansen, J. C., Hendzel, M. J. Condensed chromatin behaves like a solid on the mesoscale in vitro and in living cells. Cell. 183 (7), 1772-1784 (2020).
Gibson, B. A., et al. Organization of chromatin by intrinsic and regulated phase separation. Cell. 179 (2), 470-484 (2019).
Cerase, A., Armaos, A., Neumayer, C., Avner, P., Guttman, M., Tartaglia, G. G. Phase separation drives X-chromosome inactivation: a hypothesis. Nature Structural & Molecular Biology. 26 (5), 331-334 (2019).
Jachowicz, J. W., Strehle, M., Banerjee, A. K., Blanco, M. R., Thai, J., Guttman, M. Xist spatially amplifies SHARP/SPEN recruitment to balance chromosome-wide silencing and specificity to the X chromosome. Nature Structural & Molecular Biology. 29 (3), 239-249 (2022).
Pandya-Jones, A., et al. A protein assembly mediates Xist localization and gene silencing. Nature. 587 (7832), 145-151 (2020).
Du, M., Chen, Z. J. DNA-induced liquid phase condensation of cGAS activates innate immune signaling. Science. 361 (6403), 704-709 (2018).
Zamudio, A. V., et al. Mediator condensates localize signaling factors to key cell identity genes. Molecular Cell. 76 (5), 753-766 (2019).
Zhang, J. Z., et al. Phase separation of a PKA regulatory subunit controls cAMP compartmentation and oncogenic signaling. Cell. 182 (6), 1531-1544 (2020).
Kovar, H. Dr. Jekyll and Mr. Hyde: The two faces of the FUS/EWS/TAF15 protein family. Sarcoma. 2011, e837474 (2010).
Linardic, C. M. PAX3-FOXO1 fusion gene in rhabdomyosarcoma. Cancer Letters. 270 (1), 10-18 (2008).
Ahn, J. H., et al. Phase separation drives aberrant chromatin looping and cancer development. Nature. 595 (7868), 591-595 (2021).
Wegmann, S., et al. Tau protein liquid-liquid phase separation can initiate tau aggregation. The EMBO Journal. 37 (7), e98049 (2018).
Friedman, M. J., et al. Polyglutamine domain modulates the TBP-TFIIB interaction: implications for its normal function and neurodegeneration. Nature Neuroscience. 10 (12), 1519-1528 (2007).
Molliex, A., et al. Phase separation by low complexity domains promotes stress granule assembly and drives pathological fibrillization. Cell. 163 (1), 123-133 (2015).
Murakami, T., et al. ALS/FTD mutation-induced phase transition of FUS liquid Droplets and reversible hydrogels into irreversible hydrogels impairs RNP granule function. Neuron. 88 (4), 678-690 (2015).
Patel, A., et al. A liquid-to-solid phase transition of the ALS protein FUS accelerated by disease mutation. Cell. 162 (5), 1066-1077 (2015).
Shin, Y., Brangwynne, C. P. Liquid phase condensation in cell physiology and disease. Science. 357 (6357), (2017).
Kato, M., McKnight, S. L. A solid-state conceptualization of information transfer from gene to message to protein. Annual Review of Biochemistry. 87, 351-390 (2018).
Li, P., et al. Phase transitions in the assembly of multivalent signalling proteins. Nature. 483 (7389), 336-340 (2012).
Muzzopappa, F., et al. Detecting and quantifying liquid-liquid phase separation in living cells by model-free calibrated half-bleaching. Nature Communications. 13 (1), 7787 (2022).
Sprague, B. L., Müller, F., Pego, R. L., Bungay, P. M., Stavreva, D. A., McNally, J. G. Analysis of binding at a single spatially localized cluster of binding sites by fluorescence recovery after photobleaching. Biophysical Journal. 91 (4), 1169-1191 (2006).
Taylor, N. O., Wei, M. -. T., Stone, H. A., Brangwynne, C. P. Quantifying dynamics in phase-separated condensates using fluorescence recovery after photobleaching. Biophysical Journal. 117 (7), 1285-1300 (2019).
Liu, Z., Lavis, L. D., Betzig, E. Imaging live-cell dynamics and structure at the single-molecule level. Molecular Cell. 58 (4), 644-659 (2015).
Los, G. V., et al. HaloTag: A novel protein labeling technology for cell imaging and protein analysis. ACS Chemical Biology. 3 (6), 373-382 (2008).
Grimm, J. B., et al. A general method to improve fluorophores using deuterated auxochromes. JACS Au. 1 (5), 690-696 (2021).
Grimm, J. B., et al. photoactivatable fluorophores for single-molecule imaging. Nature Methods. 13 (12), 985-988 (2016).
Hansen, A. S., Pustova, I., Cattoglio, C., Tjian, R., Darzacq, X. CTCF and cohesin regulate chromatin loop stability with distinct dynamics. eLife. 6, 25776 (2017).
Sergé, A., Bertaux, N., Rigneault, H., Marguet, D. Dynamic multiple-target tracing to probe spatiotemporal cartography of cell membranes. Nature Methods. 5 (8), 687-694 (2008).
Normanno, D., et al. Probing the target search of DNA-binding proteins in mammalian cells using TetR as model searcher. Nature Communications. 6 (1), 7357 (2015).
Irgen-Gioro, S., Yoshida, S., Walling, V., Chong, S. Fixation can change the appearance of phase separation in living cells. eLife. 11, e79903 (2022).
Tokunaga, M., Imamoto, N., Sakata-Sogawa, K. Highly inclined thin illumination enables clear single-molecule imaging in cells. Nature Methods. 5 (2), 159-161 (2008).
Teves, S. S., An, L., Hansen, A. S., Xie, L., Darzacq, X., Tjian, R. A dynamic mode of mitotic bookmarking by transcription factors. eLife. 5, e22280 (2016).
Drosopoulos, W. C., Vierra, D. A., Kenworthy, C. A., Coleman, R. A., Schildkraut, C. L. Dynamic assembly and disassembly of the human DNA polymerase δ holoenzyme on the genome In vivo. Cell Reports. 30 (5), 1329 (2020).
Cong, L., et al. Multiplex genome engineering using CRISPR/Cas systems. Science. 339 (6121), 819-823 (2013).
Yoshida, S. R., Maity, B. K., Chong, S. Visualizing protein localizations in fixed cells: Caveats and the underlying mechanisms. The Journal of Physical Chemistry B. 127 (19), 4165-4173 (2023).
Gautier, A., et al. An engineered protein tag for multiprotein labeling in living cells. Chemistry & Biology. 15 (2), 128-136 (2008).
Hipp, L., et al. Single-molecule imaging of the transcription factor SRF reveals prolonged chromatin-binding kinetics upon cell stimulation. Proceedings of the National Academy of Sciences. 116 (3), 880-889 (2019).
Agarwal, H., Reisser, M., Wortmann, C., Gebhardt, J. C. M. Direct observation of cell-cycle-dependent interactions between CTCF and chromatin. Biophysical Journal. 112 (10), 2051-2055 (2017).
Chen, L., Zhang, Z., Han, Q., Maity, B. K., Rodrigues, L., Zboril, E., Adhikari, R., Ko, S. H., Li, X., Yoshida, S. R., Xue, P., Smith, E., Xu, K., Wang, Q., Huang, T. H., Chong, S., Liu, Z. Hormone-induced enhancer assembly requires an optimal level of hormone receptor multivalent interactions. Molecular cell. 83 (19), 3438-3456 (2023).
Garcia, D. A., et al. Power-law behavior of transcription factor dynamics at the single-molecule level implies a continuum affinity model. Nucleic Acids Research. 49 (12), 6605-6620 (2021).

Play Video

PDF

DOI

DOWNLOAD MATERIALS LIST

Cite This Article

Yoshida, S. R., Chong, S. Single-Molecule Measurement of Protein Interaction Dynamics Within Biomolecular Condensates. J. Vis. Exp. (203), e66169, doi:10.3791/66169 (2024).

Automatically Generated