Visualizing molecules of functional human profilin

  1. Morgan L Pimm
  2. Xinbei Liu
  3. Farzana Tuli
  4. Jennifer Heritz
  5. Ashley Lojko
  6. Jessica L Henty-Ridilla

  • Curated by eLife

    eLife logo

    Evaluation Summary:

    This paper will be of interest to a broad audience of cell biologists and biochemists who study the cytoskeleton. It reports the development and rigorous characterization of a fully functional, fluorescently labeled version of profilin that can be used to visualize profilin's dynamic interactions in live cells. Owing to profilin's dual functions in regulating actin and microtubule assembly, this technological development will be a useful tool for a wide range of studies aimed at understanding the role of the cytoskeleton in driving fundamental cellular processes.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #1 agreed to share their name with the authors.)

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

Profilin-1 (PFN1) is a cytoskeletal protein that regulates the dynamics of actin and microtubule assembly. Thus, PFN1 is essential for the normal division, motility, and morphology of cells. Unfortunately, conventional fusion and direct labeling strategies compromise different facets of PFN1 function. As a consequence, the only methods used to determine known PFN1 functions have been indirect and often deduced in cell-free biochemical assays. We engineered and characterized two genetically encoded versions of tagged PFN1 that behave identical to each other and the tag-free protein. In biochemical assays purified proteins bind to phosphoinositide lipids, catalyze nucleotide exchange on actin monomers, stimulate formin-mediated actin filament assembly, and bound tubulin dimers (k D = 1.89 µM) to impact microtubule dynamics. In PFN1-deficient mammalian cells, Halo-PFN1 or mApple-PFN1 (mAp-PEN1) restored morphological and cytoskeletal functions. Titrations of self-labeling Halo-ligands were used to visualize molecules of PFN1. This approach combined with specific function-disrupting point-mutants (Y6D and R88E) revealed PFN1 bound to microtubules in live cells. Cells expressing the ALS-associated G118V disease variant did not associate with actin filaments or microtubules. Thus, these tagged PFN1s are reliable tools for studying the dynamic interactions of PFN1 with actin or microtubules in vitro as well as in important cell processes or disease-states.

Article activity feed

  1. Version published to 10.7554/elife.76485 on eLife
  2. Version published to 10.1101/2021.09.01.458557v3 on bioRxiv
  3. eLife

    Evaluation Summary:

    This paper will be of interest to a broad audience of cell biologists and biochemists who study the cytoskeleton. It reports the development and rigorous characterization of a fully functional, fluorescently labeled version of profilin that can be used to visualize profilin's dynamic interactions in live cells. Owing to profilin's dual functions in regulating actin and microtubule assembly, this technological development will be a useful tool for a wide range of studies aimed at understanding the role of the cytoskeleton in driving fundamental cellular processes.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #1 agreed to share their name with the authors.)

  4. eLife

    Reviewer #1 (Public Review):

    The dynamics of profilin have been insufficiently studied for multiple reasons. It is a small protein interacting with multiple partners, present in high concentration in cells, and diffusing mainly with the monomeric actin reservoir. All this makes it difficult to label.

    In this study, the authors describe a strategy to overcome all these difficulties. They express labeled profilins and first verify that these molecules retain all the major interactions of the profilin (with lipids and components of the actin and microtubule cytoskeletons). Then they show how these tools can be used in vitro and in vivo to follow the dynamics of these molecules and obtain biological information.

    Overall, these are very interesting tools because they allow to see interactions previously deduced in an indirect way. They seem particularly suitable for visualizing interactions with microtubules; for interactions with the actin cytoskeleton, their use is less clear because most interactions are transient or are with actin monomers that diffuse too rapidly. For this reason, it would be helpful if the authors would expand their discussion to explain in more detail how they envision the use of these tools in the future.

  5. eLife

    Reviewer #2 (Public Review):

    In this manuscript, Pimm and colleagues report their development of a functional probe that enables in vivo visualization of profilin, a cytoskeletal protein that interacts with both actin and microtubules. Because of its dual impacts on actin and microtubule dynamics, profilin is a critical regulatory hub for cytoskeletal assembly and cellular morphology. Through systematic characterization, the authors find that attachment of a fluorescent protein (or Halo tag) and 10 amino acid linker at the N-terminus of human profilin-1 does not impact profilin's lipid-, actin-, or poly-L-proline-binding functions, or its ability to influence actin and microtubule assembly. Further, expression of Halo-tagged profilin-1 rescues neuroblastoma cells from profilin-1 knockout-induced morphological and cytoskeletal defects. Finally, the localization of Halo-tagged profilin-1 to microtubule networks in these cells requires an intact tubulin-binding interface but is unaffected by disruption of its actin- or poly-L-proline functions.

    Strengths

    This study undertakes an extensive characterization of profilin's molecular interactions and complex effects on cytoskeletal filament assembly. Notably, the authors use a clever approach to measure profilin's affinity for actin monomers. Using both competition and direct titration assays, the authors demonstrate that binding of both untagged and fluorescently labeled profilin can be robustly detected via fluorescence anisotropy. The use of complementary biochemical and cell biological approaches demonstrates the broad utility of this tagged profilin, and the introduction of point-mutations to probe profilin's localization patterns in live cells provides insight into the dynamic nature of its interactions.

    Weaknesses

    Overall, the experiments are well designed, and the data are robust. My only recommendations seek to elicit additional information to further highlight the potential utility of this probe.

    It would be helpful for the authors to expand on the design of their tagged profilin to provide context for readers who might like to use this probe in their own studies. For example, were various linkers sampled? If so, did the authors find that the length and/or sequence impacted profilin's functions?

    It would also be helpful to have additional information about the morphology measurements carried out by the authors. For example, does the morphology index for actin and microtubules report fraction of the cell area that is occupied by polymerized actin or microtubules? Is this value normalized relative to the cell size or fluorescence intensity (i.e., density of the cytoskeletal network)? Also, do profilin knockout cells expressing Halo-G118V have altered microtubule morphologies relative to cells in which Halo-profilin-1 is expressed?

  6. Version published to 10.1101/2021.09.01.458557v2 on bioRxiv
  7. ASAPbio crowd review

    This review reflects comments and contributions by Ricardo Carvalho, Joachim Goedhart, Sónia Gomes Pereira, Pratima Gurung, Samuel Lord, Claudia Molina, Arthur Molines, Gregory Redpath, Mugdha Sathe, Sagar Varankar. Review synthesized by Ewa Sitarska.

    This preprint introduces a recombinant profilin that has a flexible linker to a genetically encoded fluorescent tag (either mApple or Halo). Fluorescent protein tagging is a popular and accepted method to study the properties of a protein of interest in solution and in cells. A careful analysis of the tagged protein relative to the untagged, native protein is crucial to understand whether the tagged protein faithfully reflects the behavior of the native protein. Therefore, studies like these are very valuable and the current manuscript is a good example of how such a study should be performed. The flexible linker presented here overcomes challenges observed in previous papers that found that linking a fluorescent protein to profilin disrupted some of its actin-related functions in cells and in vitro.

    This study is carefully conducted and nicely describes the properties of a fluorescent protein tagged profilin in a detailed manner. In particular, the authors use various in vitro assays as well as rescue experiments to demonstrate that their tagged version of profilin appears to behave similarly to wildtype profilin. The manuscript is written in a clear manner and was an enjoyable read. The comments below cover a couple of experiments, clarifications and questions for consideration to further add to the work. Regardless, this work is likely to contribute to the field, as anyone studying profilin is likely to try this construct in their future experiments.

    General comments:

    • Why was mApple chosen as a tag (as opposed to the popular and best known fluorescent protein mEGFP)?
    • The mApple is prone to photochromicity/photoswitching (https://doi.org/10.1038/nmeth.1209, https://doi.org/10.1038/nmeth.4074). This should be mentioned to warn future users of this fusion protein.
    • It would be advisable to be consistent with the naming of ‘untagged profilin’ throughout the manuscript. Currently unlabeled, untagged, wild-type or rescue are used interchangeably.
    • In Figures 4B, 5A there appears to be differences between untagged and tagged profilin in the images. Maybe a more representative image would be beneficial, where applicable.
    • Depositing the plasmids from this paper at addgene.org would be beneficial for the public (the plasmids can be deposited under condition that these will be released only after publication of this work in a peer-reviewed journal).

    Title

    ‘Functional fluorescently-tagged human profilin’– suggest clarifying in the title and throughout the text that the fluorescent tags are genetically encoded.

    Abstract

    ‘high cellular concentrations (121 µM)’– This is a very precise number for such a general statement. It seems that the number is derived from a specific cell line, so it would be beneficial to present it as a number from this cell line or change it to an approximation (~100 µM).

    Introduction

    'Some profilin outcompetes actin bound'– suggest some rewording to clarify the fragment, for example, it could be mentioned whether it refers to F-actin or G-actin.

    Results- Design of tagged profilin

    ‘Profilin is considerably smaller than the smallest fluorescent tags’ – clarifying that it is a genetically encoded fluorescent tag would be of advantage. There are no smaller fluorescent proteins (FP) yet, but genetically encoded FP of a similar size exist, for example miRFP670nano is 17kDa. https://doi.org/10.1038/s41467-018-08050-8

    ‘Traditional direct labeling approaches are cytotoxic and disrupt actin-based functions’ – is there data showing the new fluorescent profilin side-by-side with one without a flexible linker (or other version used previously) to show that the latter disrupts profilin's functions? It’s not essential but it would strengthen this point and confirm the improvement over prior work.

    ‘Estimates of cellular profilin concentration are very high depending on the cell type’ – would be nice to provide a rough estimate at this point, similarly to the introduction part.

    *‘with an mApple fluorescent probe or as Halo-tagged single molecules’ *- What is the meaning of 'single' here?

    ‘We cloned mApple- or Halo-tags fused to a ten amino acid flexible linker on the N-terminus of human profilin-1.’ – As in the introduction, it is stated that “Positioning a GFP-derived fluorescent tag on the C- or N-terminus disrupts PLP- and PIP-binding interactions, effectively rendering the fluorescent version flawed for critical measurements in cells”, it would be beneficial to state the rationale for tagging profillin at the N-terminus? Also, how was the linker composition and length determined and how is it related to other linkers used in compromised fusions of profilin?

    Figure 1

    • In panel A, it would be helpful to indicate the N-terminus, as this is the side where the fluorescent protein is attached.
    • In the legend, PFN1 is introduced for the first time and thus, it could be replaced with ‘tagged-profilin (PFN1)’.

    Results - mApple-profilin binds phosphoinositide lipids with similar affinity as untagged profilin

    ‘PIP’ - PIP is usually an abbreviation for Phosphatidyl Inositol Phosphate (which is a lipid). Phosphoinositide is the same thing, but is not abbreviated as PIP. Recommend using Phosphatidyl Inositol Phosphate = PIP. These types of lipids can be indicated as PIPs (without the addition of lipids).

    ‘Profilin also binds PI(3,5)P2 which regulates critical signal transduction events through intracellular vesicles to the early endosome’ - The prevailing consensus now is that PI(3,5)P2 is involved in late endosomal trafficking to the lysosome. This could be different for profilin specific purposes, but this statement could be updated with recent references to PI(3,5)P2.

    ‘Thus, mApple-tagged profilin retains functional interactions with two important PIP lipids.’ – Testing other phospholipids, including PIP3 (Lu et al., 1996 showed an interaction with profilin) and some negative controls would be beneficial. Covering most phosphoinositide species by using the phospholipid-binding dot blots would make this figure stronger.

    Figure 2

    • In panel B, it would be easier to compare profilin and mApple-profilin binding affinity to each of the PIP lipids. For that, profilin and mApple-profilin samples could be run side by side in the same blot. The suggestion being that panel B includes profilin and mApple-profilin incubated with PI(3,5)P2, and panel C profilin and mApple-profilin incubated with PI(4,5)P2. For clarity, can it be specified in the figure legend or in the figure that the profilin-1 lane is the negative control pellet lacking the liposomes as well as that S and P stands for supernatant.
    • Also, in panel B, what are the loading controls? The quantification and western-blots are unclear. For example, it is indicated that 1µM PFN1was used in this experiment, but the pellet band for mAp-PFN1 is not comparable to PFN1 pellet band. Despite this, quantification in 2D shows that they are similar. Please clarify and add the corresponding loading controls.
    • Panel D: Please mention what is quantified here (supernatant, pellet or overall level),
    • Panel D and E: the shades to pink may be tricky to distinguish, more contrasting colors and/or shapes might be useful.
    • It is highly appreciated that the antibodies and dilutions are mentioned in the figure legend.

    Results - Direct visualization of fluorescent-profilin with polymerizing actin filaments

    ‘We used fluorescence anisotropy to measure the binding affinity between profilin and Oregon-Green (OG)-labeled monomeric actin (Figure S2).‘ - This appears confusing with the 3D results. If a change in fluorescence anisotropy with OG-actin is not detected, then is it okay to use OG-actin for bulk polymerisation assay. Maybe the OG-tag interferes only during fluorescence anisotropy, but not during fluorescence microscopy.

    ‘Several studies demonstrate that thymosin b4 (Tb4) competes with profilin to bind actin monomers‘ – It is worth mentioning that this refers to untagged/unlabeled actin monomers.

    Figure 3

    • Please consider if it is relevant to compare the competitive and non-competitive data on the same graph.
    • In panel A description, ‘10 nM GFP-thymosin b4 (GFP-Tb4) mixed with increasing concentrations of unlabeled actin.’ – Would ‘10nM unlabelled actin monomers in presence of increasing concentration of Τβ4’ be more appropriate?
    • In panel D, the curves would be better visible when the y-axis runs from 0-30.
    • In panel E, the mApple-Profilin samples show longer filaments, maybe quantification of the filament length could be performed?
    • In panel G, are the errors bars based on the means of the technical replicates or on all aggregated data? The first option is preferred, as this plotting strategy was also used in 3F.
    • In the panel description, ‘Data were quantified from four separate reactions (FOV) each.’ – could clarify whether the data derive from 4 totally independent reactions, or from the analysis of 4 different fields of view (FOV) from the same experimental procedure?
    • In the panel description, ‘ns, not significantly different from 1 µM actin alone control; a, compared with control (p <0.05)‘ – such labeling may be confusing for the reader, it would be beneficial to state that the first ‘ns/a’ are related to the actin alone, and the second ‘ns’ are comparing labelled and unlabeled profilin or omit showing the statistical test in the plot, and show it in a table instead.

    Results - Fluorescent profilin stimulates formin-based actin filament assembly

    General: Based on the competition and interaction experiments it seems important to generate a dose-dependent inducible construct for profilin to govern the stoichiometry of interactions and study their relevance in the cells.

    ‘Similar to experiments in assessing only profilin-actin interactions (Figure 3F), we counted significantly fewer actin filaments in reactions containing actin and either profilin (Figure 4D)’ - In Figure 3F the plot shows around 45 actin filaments per FOV with 1 uM actin (20% oregon green actin). In Figure 4D the plot shows more than 100 actin filaments per FOV with 1 uM actin (10% alexa 647 actin). Surprisingly, the elongation rate are similar in Fig 3F and Fig 4D. Is the Oregon-Green actin known to be less efficient at nucleating filaments while retaining the same polymerization ability? If it is the case, it would be worth making a mention.

    ‘Thus, fluorescent profilin stimulates formin-mediated actin filament nucleation similar to the untagged version.’ - The data seem to suggest that the presence of profilin inhibits actin filament nucleation and polymerization, a clarification would be appreciated.

    ‘The ac-celerated rate of actin filament’ – Please change to ‘accelerated’.

    Results - Profilin directly binds tubulin dimers and enhances the growth rate of microtubules in vitro

    ‘microtubule stability index‘ – The values indicated on the y-axis for the plot in Fig 5E are confusing (0 / 25/ 50 / 75 / 1). Is the index expressed as a percentage and the max value supposed to be 100? Or is the index supposed to be the number of rescues per catastrophe?

    ’This suggests a mechanism where profilin accelerates microtubule polymerization by directly binding to tubulin dimers to promote microtubule assembly and then diffusing along the sides of the microtubule lattice to further stabilize microtubule growth.‘ - Microtubules are more stable the faster they grow (catastrophe frequency scales inversely with polymerization rate). In this condition where profilin increases polymerization rate by around 5x, it is unclear how much of the increased stability is due to the lattice binding. The fragment could be softened regarding the role of the transient lattice binding in microtubule stabilization.

    Figure 5

    • In panel B, intensities are quite different, would it be possible to comment on this?
    • In panels H an I, a black/magenta merge is tricky to see. Although it breaks consistency, a green/magenta or cyan/magenta merge may be more informative visually.

    Results - Profilin regulates the morphology of N2a cells through actin and microtubule crosstalk

    ‘We used quantitative western blots to determine the level of endogenous profilin as well as levels of profilin in CRISPR knockout cells following transfection with plasmids containing untagged profilin, mApple-profilin, or Halo-profilin.’ - The levels of profilin are quantified from a blot, which is a bulk measurement. The transfection of profilin will show substantial cell-to-cell variation (some cells may have much higher or lower levels than the measured average). Mentioning it and discussing its implications would be advisable.

    ‘We chose this parameter because N2a cells have unique actin filament and microtubule cytoskeletal features but do not efficiently perform other classic cell processes that require intact cytoskeletal crosstalk (i.e., migration or division).‘ – While looking at cell shape is one strategy, an additional experiment looking into a migration phenotype may strengthen this point. Another interesting experiment to strengthen this point and providing a direct measure of profilin function could be performing a pulse chase experiment using drugs to depolymerise actin/microtubules. In such an experiment, a distinct change in depolymerisation should be noted between WT, KO, and the profilin rescue cells. This could show that the mApple-profilin can substitute for WT profilin.

    ‘super resolution confocal microscopy to image fixed cells.‘ – how is super-resolution achieved here? Or is ‘super’ unnecessary here?

    ‘the ratio of endogenous cell area to other cell conditions’ - This metric is a ratio of areas and it is only valid as an assessment for shape if the cells from each condition cover similar areas. If it is not the case, then the two parameters (shape and area) are convoluted and the ratio measures both the difference in shape and in area covered. It would be good to provide the average area of the cells in each condition for clarification.

    ‘We also stained these cells for actin filaments (Figures 6H and 6I) and micro-tubules (Figure 6J and 6K) and used a similar morphology parameter to detect broad differences in cytoskeletal architecture.’ – Please clarify the reasons for using the cell area ratio metric for quantification of cell morphology. How is the quantification metric (area) used for sub-cellular network like actin and microtubule? F-actin stained with phalloidin looks different in endogenous PFN1 vs mApple-PFN1, but by using area metric there is no morphological difference. Microtubule, on the other hand, appear similar in all the cells.

    Figure 6

    • In panel A, it does not seem like mApple-PFN1 and Halo-PFN1 reach endogenous levels. Maybe a quantification would be beneficial.
    • In panel B, please report the number of independent replicates. Also, may be worth commenting on why tagged PFN1 rescue cells were not included?
    • For panel E, please provide the error bars and the legend that contains the information from where the data derived or from how many independent experiments.
    • In panel F, it may be advantageous to use red green and blue as colors for the overlay. This will generate unique colors for the different combinations of the three images.
    • For the context of panel F, unprocessed images of the tagged profilin in living cells could be presented somewhere in the main text. They could be larger than the small panels here, and not be segmented into binary images. The point of the fluorescent profilin is that it can be used for live-cell imaging without substantially disrupting the typical profilin interactions. This should be confirmed by presenting live-cell images of the profilin construct. To avoid problems with high cytoplasmic concentrations of profilin drowning out any localization signal, maybe the fluorescent version could be expressed at a very low level or the Halo version could be used with a low concentration of fluorophore.
    • In panel G, it is clear that the dots are from different cover slips, how many cells were analyzed per coverslip? Data could be shown from individual cells (not just their average). Also, please clarify if this quantification was made 24 hours after transfection as well?
    • For panel I, actin morphology was calculated from actin filament signals similar to the cell morphology index. These calculations could be explained further in the methods. Does this mean that the actin morphology index is the ratio of actin area between the two conditions? Is the actin image somehow thresholded before taking the ratio?
    • Panel L, is really appreciated and helpful to understand the "competition" between actin and microtubules for profilin. It would be also nice to represent the plasma membrane and the lipid-binding activity of profilin as well as binding to nucleation promoting factors (the proline-rich motifs of VASP and WASP), as this is mentioned throughout the paper.

    Discussion

    The fact that mApple and HaloTag both are entirely different and non-disturbing gives confidence that profilin can be fused with other tags, without losing functionality. Mentioning this in the discussion could give new insights for the readers.

    ‘Our genetic analyses in mammalian cells indicate that mApple-profilin and Halo-profilin are fully inter-changeable with the endogenous version.‘ - Authors have given the field an excelled tool which will be quite useful to study cellular functions of PFN1, its interaction with its several binding partners. Currently, cell shape is the metric used to determine if the tagged versions are fully inter-changeable with the endogenous version. Whether the tagged PFN1 can replace untagged PFN1 for other cellular functions will require further exploration. Also, high concentration of PFN1 will remain an issue even with the mApple-PFN1 developed here. Do the authors suggest mild over-expression as a strategy to go around the high concentration issue?

    ‘Based on localization experiments using the pan-formin inhibitor, SMIFH2, some interactions between profilin and the sides of micro-tubules are thought to be indirect.’ – suggest clarifying how SMIFH2 treatment leads to conclude that interactions between PFN1 and microtubules is indirect.

    Methods

    Statistical significance tests do not demonstrate that conditions are identical. That is, when two conditions are not statistically different, it is not possible to say that these are equal (https://doi.org/10.1053/j.seminhematol.2008.04.003). Suggest avoiding the use of "n.s." in graphs to indicate that the data are similar. It is clear from the data that (in many cases) the tagged and untagged profilin show similar properties. If equality needs to be demonstrated, recommend carrying out an equivalence test.

    ‘Different shades of data points show technical replicates.’ – Please rephrase to clarify, for example “Different colors represent biological replicates. Similar colored dots reflect technical replicates“. Does "technical replicate," mean repeated measurements within each independent experimental run? Or different experimental runs? If the shading is supposed to denote paired experiments (e.g. darkest shading in different conditions are from the same experimental run), that can be stated in the caption.

  8. Version published to 10.1101/2021.09.01.458557v1 on bioRxiv