In vivo analysis of formin dynamics in the moss P. patens reveals functional class diversification

Abstract

Formins are actin regulators critical for diverse processes across eukaryotes. With many formins in plants and animals, it has been challenging to determine formin function in vivo We found that the phylogenetically distinct class I integral membrane formins (denoted For1) from the moss P.patens enrich at sites of membrane turnover, with For1D more tightly associated with the plasma membrane than For1A. To probe formin function, we generated formin-null lines with greatly reduced formin complexity. We found that For1A and For1D help to anchor actin near the cell apex, with For1A contributing to formation of cytosolic actin, while For1D contributes to plasma membrane-associated actin. At the cortex, For1A and For1D localized to motile puncta and differentially impacted actin dynamics. We found that class I cortical formin mobility depended on microtubules and only moderately on actin, whereas class II formin (denoted For2) mobility solely depended on actin. Moreover, cortical For2A tightly correlated with the puncta labeled by the endocytic membrane dye FM4-64, and null mutants in class I formins did not affect uptake of a similar dye, FM1-43, suggesting that class I and II formins are involved in distinct membrane trafficking pathways.

Keywords: Actin; Endocytosis; Exocytosis; Formin; Microtubule; Plant.

Fig. 1.

Subcellular localization of endogenously tagged For1A and For1D. (A) Immunoblot performed with an antibody to GFP on whole-cell extracts from the For1A or For1D lines where the endogenous locus was tagged in frame with sequences encoding three tandem GFP molecules (GFP). The molecular mass in kDa is indicated. Expected sizes for unmodified proteins are: For1A–GFP, 232 kDa; For1D–GFP, 204 kDa. (B,C) Spinning-disc confocal images of For1A–GFP and For1D–GFP in the apical cell (B) and in a dividing cell (C, For1D–GFP) of a protonemal moss filament. (C, For1A–GFP). Laser scanning confocal images (de-noised with NIS elements software) during cell division. Maximum projections from confocal z-stacks are shown in B and C, and the medial plane is also shown in B. Arrows indicate cytoplasmic For1A–GFP accumulations and puncta. Large globular structures are chloroplasts, which display autofluorescence under these imaging conditions. Scale bars: 5 µm. (D) VAEM images of the cell cortex in the For1A–GFP and For1D–GFP lines. Scale bars: 2 µm. See also Movie 1.

Fig. 2.

Subcellular localization of endogenously tagged For1A and For1D during cell division. Laser-scanning confocal microscope images of For1A–GFP (A) and For1D–GFP (B) with Lifeact–mCherry in dividing cells. Images are a single focal plane acquired every 20 s and deconvolved with NIS elements software (type Richardson–Lucy). White arrows indicate cytoplasmic For1A–GFP accumulations and puncta. Large globular structures are chloroplasts (yellow arrows), which display autofluorescence under these imaging conditions. Scale bar: 5 µm. See also Movie 2.

Fig. 3.

Analysis of formin localization with respect to endocytic activity labeled with FM4-64. (A) Laser-scanning confocal images of the medial plane of an apical cell growing in a microfluidic imaging chamber. Images were de-noised with NIS elements software. Large structures in the For1A–GFP image are chloroplasts, which display autofluorescence under these imaging conditions. Scale bars: 5 µm. See also Movie 3. (B) Simultaneous VAEM imaging of endogenously tagged formins as indicated at the cell cortex. See also Movie 4. Scale bar: 2 µm. For all images, the formin is green and FM4-64 is magenta in the merged images. (C) Pearson's correlation coefficient comparing the formin and FM4-64 channels acquired with VAEM. Letters above the bars indicate statistical groups with α<0.05 from one-way ANOVA with a Tukey HSD post hoc test (For1A, n=10 cells; For1D, n=11 cells; For1F, n=12 cells; For2A, n=12 cells). (D) Time course of FM1-43 uptake in control (Lifeact–mRuby) and formin-null lines as indicated. Single focal plane laser scanning confocal images of cells are shown immediately after putting cells in FM1-43 (0 min) and at 15 min intervals. Scale bar: 5 µm. (E) Box plot depicting the quantification of the ratio of the intensity of the plasma membrane to the intensity within the cell within 5–7 µm from the cell tip. The box encloses 50% of the data with the median value drawn as a line. Lines extending from the box mark minimum and maximum values for the data set except for data sets with outliers. Outliers falling outside of the upper quartile are value that are 1.5×interquartile distance, and are depicted as open circles. No significant differences were found at each time point using an one-way ANOVA with a Tukey HSD post hoc test (n=10 cells for all lines).

Fig. 4.

Simultaneous imaging of For1A–GFP or For1D–GFP with Lifeact–mCherry. (A) Laser-scanning confocal images from the medial plane of an apical cell growing in a microfluidic imaging chamber. Images were de-noised with NIS elements software. Arrows indicate regions of the highest formin signal intensity. Scale bars: 5 µm. See also Movie 5. The time projection is a maximum intensity projection of 24 frames from 8 min of the time-lapse acquisition. (B) Laser-scanning confocal images of Lifeact–mRuby in control and formin-null lines as indicated. The single frame is a timepoint from a time-lapse acquisition, and is a maximum intensity projection of three confocal z-stacks taken near the medial section of the cell. The time projection is a maximum intensity projection of 121 frames from 20 min of the time-lapse acquisition. Scale bar: 5 µm. See also Movie 6. (C) Dot plot of the normalized area of the Lifeact–mRuby signal from the time projections. To normalize, the area was divided by the distance grown over the time-lapse acquisition. An one-way ANOVA with a Tukey HSD post hoc test was performed. *P<0.05; ***P<0.001.

Fig. 5.

Characterization of cortical actin in endogenously tagged lines and formin-null lines. (A) Simultaneous VAEM imaging of For1A–GFP or For1D–GFP with Lifeact–mCherry at the cell cortex. Scale bars: 2 µm. See also Movie 7. White arrows indicate examples of formin cortical foci overlapping with actin filaments; yellow arrows indicate examples of formin foci that are not associated with actin filaments. For all images, formin is green and Lifeact–mCherry is magenta in the merged images. (B) VAEM images of cortical actin labeled with Lifeact–mRuby in control and the indicated formin-null lines. Scale bar: 5 µm. See also Movie 8. (C) Quantification of cortical actin dynamics under the indicated conditions. The correlation coefficient between two images was calculated at all possible temporal spacings (time interval). Error bars represent s.e.m. (Lifeact–mRuby, n=10; Δfor1ABCE2B-84, n=13; Δfor1BCDE2B-111, n=14; Δfor1ABCDE2B, n=10).

Fig. 6.

For1A and For1D differentially contribute to cytosolic actin structures. Apical cells expressing (A) For1A–GFP (green) or (C) For1D–GFP (green) with Lifeact–mCherry (magenta) growing in microfluidic imaging chambers were treated with 12.5 µM oryzalin and imaged on a laser-scanning confocal microscope. Images are single focal planes acquired every 5 s and deconvolved with NIS elements software (type Richardson–Lucy). The time-lapse images show accumulations of For1A–GFP that closely correlate temporally and spatially with generation of cytosolic actin foci. See also Movie 9. Whereas For1A–GFP strongly accumulates in the cytoplasm in the absence of microtubules, the For1D–GFP signal (remains low in the cytoplasm and mostly stays at the plasma membrane. See also Movie 10. (B,D) Quantification of time-lapse images in A and C. Foci enriched with Lifeact–mCherry were tracked using TrackMate and the mean intensity of For1A–GFP and Lifeact-mCherry in A, and For1D–GFP and Lifeact-mCherry in C were plotted over time. Numbers above the peaks (1*–4*) in the plots correspond to the bursts of actin polymerization indicated in A and C. Time stamps represent min:s. Scale bars: 10 µm.

Fig. 7.

Formins are differentially affected by cytoskeletal inhibitors. (A) VAEM imaging of endogenously tagged formins at the cell cortex. See also Movie 11. Three frames taken every 200 ms from a time-lapse acquisition were false-colored red, green and blue and then merged into a single image. Movement of cortical dots from one frame to the next appear colored in the merge. If the particle does not move, then the red–green–blue merge results in a white particle. Scale bar: 2 µm. (B) Quantification of cortical formin dynamics under the indicated conditions. The correlation coefficient between two images was calculated at all possible temporal spacings (time interval). Error bars represent s.e.m. (n=10 cells except for For1A control, n=9; For1D LatB, n=8; For2A LatB, n=11).