Fixation can change the appearance of phase separation in living cells
Curation statements for this article:-
Curated by eLife
eLife assessment
Chemical fixation of cells is ubiquitous in microscopy. However, fixation artifacts can lead the incorrect interpretations of biological processes. In here, Irgen-Gioro et al. show that in the context of liquid condensates formed by liquid-liquid phase separation (LLPS), paraformaldehyde (PFA) fixation can lead to artifacts such as changes in the number, appearance, or disappearance of liquid condensates, when comparing fixed to live cells. This will be of great interest not only for those in the LLPS field but for cell biologists, in general, using fixed samples for microscopy.
This article has been Reviewed by the following groups
Listed in
- Evaluated articles (eLife)
Abstract
Fixing cells with paraformaldehyde (PFA) is an essential step in numerous biological techniques as it is thought to preserve a snapshot of biomolecular transactions in living cells. Fixed-cell imaging techniques such as immunofluorescence have been widely used to detect liquid–liquid phase separation (LLPS) in vivo. Here, we compared images, before and after fixation, of cells expressing intrinsically disordered proteins that are able to undergo LLPS. Surprisingly, we found that PFA fixation can both enhance and diminish putative LLPS behaviors. For specific proteins, fixation can even cause their droplet-like puncta to artificially appear in cells that do not have any detectable puncta in the live condition. Fixing cells in the presence of glycine, a molecule that modulates fixation rates, can reverse the fixation effect from enhancing to diminishing LLPS appearance. We further established a kinetic model of fixation in the context of dynamic protein–protein interactions. Simulations based on the model suggest that protein localization in fixed cells depends on an intricate balance of protein–protein interaction dynamics, the overall rate of fixation, and notably, the difference between fixation rates of different proteins. Consistent with simulations, live-cell single-molecule imaging experiments showed that a fast overall rate of fixation relative to protein–protein interaction dynamics can minimize fixation artifacts. Our work reveals that PFA fixation changes the appearance of LLPS from living cells, presents a caveat in studying LLPS using fixation-based methods, and suggests a mechanism underlying the fixation artifact.
Article activity feed
-
-
-
Author Response
Reviewer #2 (Public Review):
“To describe LLPS or to distinguish between polymer-polymer phase separation and LLPS, recent studies have used single particle tracking, a technique allowing to follow the dynamics of individual proteins in living cells (https://doi.org/10.7554/eLife.60577; https://doi.org/10.7554/eLife.69181; https://doi.org/10.7554/eLife.47098). The authors should mention that such an approach can be a good alternative to avoid the artefact of fixation. Using techniques such as single particle tracking or FCS, it is possible to estimate the effective diffusion coefficient of protein-living cells. When a liquid phase separation is formed, it is also possible to estimate the diffusion coefficient of the protein of interest (POI) inside versus outside of the LLPS.”
We thank the reviewer for their insight …
Author Response
Reviewer #2 (Public Review):
“To describe LLPS or to distinguish between polymer-polymer phase separation and LLPS, recent studies have used single particle tracking, a technique allowing to follow the dynamics of individual proteins in living cells (https://doi.org/10.7554/eLife.60577; https://doi.org/10.7554/eLife.69181; https://doi.org/10.7554/eLife.47098). The authors should mention that such an approach can be a good alternative to avoid the artefact of fixation. Using techniques such as single particle tracking or FCS, it is possible to estimate the effective diffusion coefficient of protein-living cells. When a liquid phase separation is formed, it is also possible to estimate the diffusion coefficient of the protein of interest (POI) inside versus outside of the LLPS.”
We thank the reviewer for their insight and fully agree that live-cell techniques like SPT and FCS are valuable for investigating LLPS while avoiding fixation artifacts. We have added discussion emphasizing this fact and incorporated the citations recommended by the reviewer in Paragraph 1 on Page 15: “Live imaging techniques that allow estimation of protein diffusion coefficients within specific cellular compartments, e.g., SPT (Hansen et al., 2018 and Heckert et al., 2022) and fluorescence correlation spectroscopy (Lanzanò et al., 2017), can be useful alternative approaches for diagnosing LLPS in vivo without the potential artifact of fixation, as diffusion dynamics are recently shown to be affected by LLPS (Heltberg et al., 2021; McSwiggen et al., 2019a; Miné-Hattab et al., 2021; Chong et al., 2022; and Ladouceur et al., 2020).”
“The authors say that less dynamic interactions are better captured by PFA fixation. In the simulation part, would it be possible to predict from the diffusion coefficients of the POI inside a condensate the effect of the PFA fixation? […] In the simulation part, they could try to incorporate the diffusion coefficient of the protein of interest and see if it is possible to predict the effect of fixation as a function of the diffusion coefficient.”
We thank the reviewer for pointing out the absence of this critical piece that connects our experimental observations to our kinetic model. Our model considers association/dissociation rates rather than diffusion coefficients to describe interaction dynamics, but the reviewers’ point is still very insightful and important. As described in Response 2, we compared two proteins: Halo-TAF15(IDR), which is poorly preserved by fixation, and TAF15(IDR)-Halo-FTH1, which is well preserved by fixation. We used SPT to measure the dissociation rates of Halo-TAF15(IDR) and TAF15(IDR)-Halo-FTH1 and showed that the dissociation rate of Halo-TAF15(IDR) from its puncta is much faster than that of TAF15(IDR)-Halo-FTH1, demonstrating more stable homotypic interactions of the latter than the former. The observation that TAF15(IDR)-Halo-FTH1 has less dynamic interactions and is better preserved by fixation compared to Halo-TAF15(IDR) agrees with our model’s prediction that less dynamic interactions are better captured by fixation. Please see Response 2 for more details. Our new data and discussion have been added to the revised manuscript in Paragraph 3 on Page 13 and in Figure 3B, Figure 3E, Figure 6, and Video 2.
“Finally, the authors propose that in the future, it will be important to design novel fixatives with significantly faster cross-linking rates than biomolecular interactions to eliminate fixation artifacts in the cell. It would be even more interesting if the authors could propose some ideas of potential novel fixatives. Did they test several concentrations of PFA, for example? Did they test different times of PFA incubation? Did they test cryofixation and do they know what would be their effect on LLPS? Do they have novel fixatives in mind? […] To strengthen the manuscript, the authors should try more protocols of fixation.”
We thank the reviewer for these good questions. As described in Response 1, we have done additional quantification of the change of LLPS appearance in cells upon treatment of 0% PFA (only PBS buffer), 1% PFA, 2% PFA, and 8% PFA as well as 4% PFA supplemented by 0.2% GA. We saw statistically significant changes in the LLPS-describing parameters upon all the PFA and PFA/GA treatments except the 0% PFA control. To examine how fixation artifacts depend on the time of PFA incubation, we acquired a time-lapse movie of a cell overexpressing EGFP-FUS(IDR) immediately after 4% PFA treatment and quantified the number of puncta over time (Video 1). We showed that fixation is complete (the number of puncta becomes constant) by roughly 100 seconds (Figure 1 – figure supplement 2). Our new data also justified our choice of a 10-minute PFA incubation time for analyzing fixation-induced change of LLPS appearance in the rest of the paper. Please see Response 1 for more details. Our new data and discussion have been added to the revised manuscript in Paragraph 3 on Page 3 and in Figure 1 - figure supplement 2 (time dependence of fixation artifacts), Figure 1 - figure supplement 3 (fixation artifact at various PFA concentrations), and Figure 1 - figure supplement 4 (fixation artifact upon treatment of 4% PFA supplemented with 0.2% GA).
We agree that testing more cell fixation protocols such as cryofixation on LLPS appearance would be interesting. However, given the complexity of novel fixation protocols like cryofixation and highly specialized equipment and reagents they require, testing widely how different fixation methods might change LLPS appearance would be a tremendous amount of work that is enough to fill a separate paper. These experiments would be much more appropriate for a separate study in the future.
Reviewer #3 (Public Review):
“Understanding whether/how fixation methods affect the detection of biomolecular condensates is of broad interest given the importance of LLPS in regulating different aspects of cell biology. However, in this manuscript, the authors use only paraformaldehyde as a fixation method and study only fluorescently-labelled IDR proteins. The work would benefit from a comparison between living cells and cells fixed with other fixation methods.”
We appreciate the reviewer for this suggestion and agree that more fixation protocols should be investigated. As described in Response 1 and Response 18, besides examining PFA fixation, we have quantified how fixation using 4% PFA supplemented by 0.2% GA changes LLPS appearance in cells. We saw statistically significant changes in all the LLPS-describing parameters upon PFA/GA treatments. Please see Response 1 and Response 18 for details. Our new data and discussion have been added to the revised manuscript in Paragraph 3 on Page 3 and in Figure 1 - figure supplement 4.
“In addition, it would be useful to test the impact of these fixation methods on the detection of endogenous proteins or IDR proteins without fluorescent tag.”
We appreciate the reviewer for this suggestion and have now investigated an endogenous IDR-containing protein in the revised manuscript. Specifically, we quantified the effect of 4% PFA fixation on endogenously expressed EWS::FLI1 in an Ewing sarcoma cell line A673, which is an oncogenic fusion transcription factor that causes Ewing sarcoma (Grünewald et al., 2018) and known to form local, high-concentration hubs at target genes associated with GGAA microsatellites (Chong et al., 2018). We previously Halo-tagged endogenous EWS::FLI1 in A673 cells using CRISPR/Cas9-mediated genome editing (Chong et al., 2018). Here, we quantified the effect of PFA fixation on endogenous EWS::FLI1 puncta in this knock-in cell line and found no significant difference in the distribution of EWS::FLI1 upon fixation. This result suggests that PFA fixation does not change the intracellular distribution of all proteins. Our new data and discussion have been added to the revised manuscript in Paragraph 1 on Page 8 and in Figure 3C.
Unfortunately, testing fixation artifacts of IDR-containing proteins without a fluorescent tag has been infeasible as we rely on fluorescence from a tag on the protein of interest to quantitatively compare LLPS appearance in live and fixed cells. Although we have considered using non-fluorescent methods, e.g., phase contrast microscopy, to visualize putative LLPS in cells, its lack of specificity in imaging proteins or cellular structures makes the type of quantification we do for fixation artifact characterization inaccessible.
-
eLife assessment
Chemical fixation of cells is ubiquitous in microscopy. However, fixation artifacts can lead the incorrect interpretations of biological processes. In here, Irgen-Gioro et al. show that in the context of liquid condensates formed by liquid-liquid phase separation (LLPS), paraformaldehyde (PFA) fixation can lead to artifacts such as changes in the number, appearance, or disappearance of liquid condensates, when comparing fixed to live cells. This will be of great interest not only for those in the LLPS field but for cell biologists, in general, using fixed samples for microscopy.
-
Reviewer #1 (Public Review):
The authors investigate whether and how PFA fixation affects the structures formed by some proteins that undergo LLPS. They do that by over-expressing a number of constructs in cells and imaging them by live cell fluorescence microscopy, after which they fix the cells and image the same cell after fixation. Their results clearly show that, for different proteins and with different tags, there is a non-systematic alteration of the LLPS structures.
In parallel to this experimental work, the authors also present and analyze a dynamic computational model in which they investigate how different fixation rates for those proteins inside and outside the condensates can lead to alterations in the overall condensate organization after fixation. Their model shows that if the fixation rate inside the condensate is …
Reviewer #1 (Public Review):
The authors investigate whether and how PFA fixation affects the structures formed by some proteins that undergo LLPS. They do that by over-expressing a number of constructs in cells and imaging them by live cell fluorescence microscopy, after which they fix the cells and image the same cell after fixation. Their results clearly show that, for different proteins and with different tags, there is a non-systematic alteration of the LLPS structures.
In parallel to this experimental work, the authors also present and analyze a dynamic computational model in which they investigate how different fixation rates for those proteins inside and outside the condensates can lead to alterations in the overall condensate organization after fixation. Their model shows that if the fixation rate inside the condensate is different than outside the condensate, and if the dynamics of protein exchange in/out of the condensates are fast enough, fixation artifacts are to be expected.
It remains to be seen whether the alterations in condensate structures after fixation (as seen experimentally) are caused by different fixation rates (as shown computationally).
Overall, this manuscript puts forward an important observation, on how chemical fixation can alter cellular structures, such as those of membraneless organelles.
-
Reviewer #2 (Public Review):
The manuscript entitled "Fixation Can Change the Appearance of Phase Separation in Living Cells" discussed the different fixation artefacts that can change the appearance of LLPS. The manuscript points out a fundamental question in the field of phase separation which is rarely discussed. The authors found that PFA fixation can both enhance and diminish putative LLPS behaviors; in some cases, it can also create condensates that did not exist in living cells. Using a simple but elegant model, they found that protein localization in fixed cells depends on an intricate balance of protein-protein interaction dynamics, the overall rate of fixation, and notably, the difference between fixation rates of different proteins. They conclude that less dynamic interactions are better captured by PFA fixation. The text is …
Reviewer #2 (Public Review):
The manuscript entitled "Fixation Can Change the Appearance of Phase Separation in Living Cells" discussed the different fixation artefacts that can change the appearance of LLPS. The manuscript points out a fundamental question in the field of phase separation which is rarely discussed. The authors found that PFA fixation can both enhance and diminish putative LLPS behaviors; in some cases, it can also create condensates that did not exist in living cells. Using a simple but elegant model, they found that protein localization in fixed cells depends on an intricate balance of protein-protein interaction dynamics, the overall rate of fixation, and notably, the difference between fixation rates of different proteins. They conclude that less dynamic interactions are better captured by PFA fixation. The text is clearly written, the experiments are well designed and the simulations give an interesting explanation of the different artefacts observed after fixation.
To describe LLPS or to distinguish between polymer-polymer phase separation and LLPS, recent studies have used single particle tracking, a technique allowing to follow the dynamics of individual proteins in living cells (https://doi.org/10.7554/eLife.60577; https://doi.org/10.7554/eLife.69181; https://doi.org/10.7554/eLife.47098). The authors should mention that such an approach can be a good alternative to avoid the artefact of fixation.
Using techniques such as single particle tracking or FCS, it is possible to estimate the effective diffusion coefficient of protein-living cells. When a liquid phase separation is formed, it is also possible to estimate the diffusion coefficient of the protein of interest (POI) inside versus outside of the LLPS. The authors say that less dynamic interactions are better captured by PFA fixation. In the simulation part, would it be possible to predict from the diffusion coefficients of the POI inside a condensate the effect of the PAF fixation?Finally, the authors propose that in the future, it will be important to design novel fixatives with significantly faster cross-linking rates than biomolecular interactions to eliminate fixation artifacts in the cell. It would be even more interesting if the authors could propose some ideas of potential novel fixatives. Did they test several concentrations of PFA, for example? Did they test different times of PFA incubation? Did they test cryofixation and do they know what would be their effect on LLPS? Do they have novel fixatives in mind?
Adding some precisions about these points in the simulation and in the fixation protocol would increase the impact of the manuscript. Otherwise, the study is interesting and thought-provoking.
-
Reviewer #3 (Public Review):
The authors compare the detection of biomolecular condensates in living cells overexpressing fluorescently tagged IDR proteins and upon fixation with paraformaldehyde (PFA). Given that they observe differences in the number and size of the condensates in the fixed versus living cells the authors conclude that the fixation method can introduce an artifact in the visualization of these condensates. Next, through kinetic modeling simulations, the authors propose a model in which the extent of the artifact introduced by PFA fixation correlates with the strength of the protein-protein interaction: artifacts are lower when the protein‐protein interactions are stable and less dynamic compared with the overall fixation rate. Based on their comparative analysis of PFA fixation and the kinetic modeling the authors …
Reviewer #3 (Public Review):
The authors compare the detection of biomolecular condensates in living cells overexpressing fluorescently tagged IDR proteins and upon fixation with paraformaldehyde (PFA). Given that they observe differences in the number and size of the condensates in the fixed versus living cells the authors conclude that the fixation method can introduce an artifact in the visualization of these condensates. Next, through kinetic modeling simulations, the authors propose a model in which the extent of the artifact introduced by PFA fixation correlates with the strength of the protein-protein interaction: artifacts are lower when the protein‐protein interactions are stable and less dynamic compared with the overall fixation rate. Based on their comparative analysis of PFA fixation and the kinetic modeling the authors strongly recommend caution in the interpretation of data obtained in PFA-fixed cells and suggest that parallel studies with living cells should be performed.
Understanding whether/how fixation methods affect the detection of biomolecular condensates is of broad interest given the importance of LLPS in regulating different aspects of cell biology. However, in this manuscript, the authors use only paraformaldehyde as a fixation method and study only fluorescently-labelled IDR proteins. The work would benefit from a comparison between living cells and cells fixed with other fixation methods; in addition, it would be useful to test the impact of these fixation methods on the detection of endogenous proteins or IDR proteins without fluorescent tag.
-