Figure 1.
Characterization of AQP9-induced filopodia.
(A) Representative confocal images of HEK-293 cells transfected with tagRFP-AQP9 or empty vector together with GFP-Mem to label the membrane. Intensities have been adjusted linearly to visualize the relative expression and localization of both fluorophores. Scalebar 10 µm. (B) Quantification of peripheral filopodia in HEK-293 cells transfected with GFP-AQP9 or GFP-Mem. The data is presented as mean number of filopodia/µm perimeter(±SEM; n = 34–43 cells/group). (C) Representative confocal images of HEK-293 cells transfected with tagRFP-AQP9 and GFP-Mem. Images are pseudo-colored in fire scale to visualize the differences between the two vectors in the filopodia. The intensities have been adjusted linearly to visualize the relative distribution of both fluorophores. The lower panel represents enlargement of the green box. Scalebar 10 µm. (D) Ratiometric measurements of mean fluorescence intensity (MFI) in the filopodial membrane divided by MFI in the cell body membrane in HEK-293 cells transfected with both tagRFP-AQP9 and GFP-Mem. Measurement areas are illustrated in the schematic image. The data is presented as mean (± SEM, n = 51 filopodia/group). (E) Montage of a representative confocal time-lapse of a HEK-293 cell overexpressing GFP-AQP9 pseudo-colored in fire scale to visualize AQP9 localization in growing filopodia. The linear intensity has been adjusted to visualize differences in fluorescent intensity. Scalebar 2 µm. (F, left panel) An enlarged image from (E) showing the points of measurements for the profile plots presented in the right panel. (F, right panel) Intensity profile plots of filopodia during growth to visualize AQP9 accumulation in filopodial tips.
Figure 2.
Localization of tubulin, myosin X and BAIAP2 in GFP-AQP9-transfected HEK-293 cells.
(A) Quantification of peripheral filopodia in GFP-transfected HEK-293 cells before and after fixation. Data are presented as mean (± SEM, n = 12–43 cells/group). (B) Images captured at the basal part of a HEK-293 cells co-expressing AQP9 and other filopodia-associated proteins fused to GFP or tagRFP. Linear intensities have been adjusted to visualize the relative distribution of both fluorophores. The zoom panel illustrating AQP9 and tubulin is split to emphasize the lack of tubulin in filopodia. Scalebar 10 µm.
Figure 3.
Addition of H2O to the medium causes filopodial bleb-like protrusions.
(A) Confocal time-lapse montage of a HEK-293 cell stably overexpressing GFP-AQP9. During acquisition, 20 µl of H2O was added to the medium (2 ml) with a pipette directed towards the cell, yielding a rapid but transient reduction in local osmolarity. The images are pseudo-colored in fire scale to visualize variations in fluorescence intensity. White arrows are pointing towards a representative bleb-like protrusions formed during image acquisition. Scale bar 10 µm. (B, upper panel) An enlarged image of a single filopodium during acquisition, before and after the addition of H2O. The white arrow shows the direction and length of measurement presented in the lower panel. The images are linearly adjusted and pseudo-colored in fire scale to visualize variations in fluorescent intensity. (B, lower panel) Intensity profile plots measured along the filopodia as shown by the white arrow. The red arrows are pointing towards peaks in fluorescent intensity before and after the addition of H2O. (C) Quantification of the percentage of filopodia that developed filopodial bleb-like protrusions subsequent to the addition of 20 µl of H2O after pre-treatment with AQP9-inhibitors. HEK-293 cells overexpressing GFP-AQP9 were pretreated with 1, 5 and 10 µM Hg2+ or with 25 µM of HTS13286. Control cells represents untreated HEK-293 cell overexpressing GFP-AQP9. Data is presented as mean (±SEM, n = 4–7 experiements/group). (D, left panel) Phase contrast images of primary human macrophages. The cell in the lower panel is treated with 25 µM of the novel AQP9 inhibitor HTS13286. (D, right panel) Cropped and inverted time lapse montage of the cells displayed in the left panel. During image acquisition 20 µl of H2O was added to the medium (2 ml). Magenta arrows are pointing towards filopodial bleb-like protrusions. Scale bar 10 µm.
Figure 4.
Interplay of filopodial actin and AQP9 reveals temporal changes in their relative distribution.
(A) Representative confocal images of a HEK-293 cell stably overexpressing GFP-AQP9 and transfected with the actin-filament labeling probe tagRFP-LifeAct. Intensities have been adjusted linearly to visualize the relative distribution of both fluorophores. Scalebar 10 µm. (B, left panel) Enlarged area of the yellow box in A. Arrow shows the direction and filopodia subjected for measurement in the right panel. (B, right panel) Relative intensity profile plot of the filopodia presented in the left panel. The data is presented as a moving average of 3 adjacent values. (C) A cropped confocal time-lapse montage of tagRFP-LifeAct (upper panel, red in lower panel) and GFP-AQP9 (middle panel and green in lower panel) distribution during filopodial growth. To highlight differences in spatial distribution, the tip of the actin filaments were labeled with a red arrow and the tip of the AQP9 fluorescent intensity is labeled with a green arrow. Linear intensities have been adjusted to visualize the relative distribution of both fluorophores. (D) A representative image showing the analysis of the relative distribution of GFP-AQP9 and tagRFP-LifeAct along the length of a filopodium during growth. The data is presented as a moving average of 3 adjacent values. (E) Illustration of the space between the filopodial tip in the GFP channel versus the RFP channel in four filopodia imaged in HEK-293 cell overexpressing both GFP-AQP9 and tagRFP-LifeAct. The initial point represents the initiation of growth, and the terminal point represents the end of growth. Each colored line represents the distance between AQP9 and actin in individual filopodia. The data is presented as a moving average of 3 adjacent values.(F) Similar analysis as in E, using mRFP-UtrCH showing the gap between AQP9 fluorescence and actin filaments was not due to interference of polymerization or loss of binding by the actin labeling probe in E.
Figure 5.
Disruption of actin dynamics inhibits the formation of new filopodia.
(A) Representative confocal images of HEK-293 cells stably overexpressing GFP-AQP9 before and 15 min after treatment with 1 µM Cyt D or 500 nM Jasplakinolide. The red arrows point towards distended filopodia. Scalebar 10 µm. (B) Quantification of peripheral filopodia before, and 10–15 min after treatment with 1 µM Cyt D. The data is presented as mean (±SEM, n = 5–43 cells/group). (C, left panel) Quantification of the relative filopodial tip area of GFP-AQP9 expressing cells before (Ctrl), and 15 min after the addition of 1 µM Cytochalsin D. The filopodial tips are defined by the fluorescent area occupied in a 2×2 µm ROI of the filopodial tips. The data is presented as mean (±SEM) of fold change compared to untreated cells (Ctrl; n = 13–22 filopodia/group).(C, right panel) Representative examples of a filopodia before and after treatment with Cyt D. The red box illustrates the area of measurement for the data presented in the left panel. (D) A confocal time-lapse montage of GFP-AQP9 fluorescence, pseudo-colored in fire scale, in HEK-293 cells 10 min after treatment with 10 µM Cytochalsin D. The images illustrate a bleb-like protrusion that recoils back towards the cell body after treatment with actin dynamics inhibitors. The linear intensity is adjusted to visualize differences in fluorescence intensity. Scalebar 1 µm.
Figure 6.
Size and location of H2O-induced bleb like protrusions.
(A) Time lapse montage of the formation of filopodial bleb-like protrusions. H2O was delivered with a micropipette in close proximity to the cell. The pressure (4000hPa) was applied to the micropipette for 1, 2, 4 or 8 s to the same cell. The white arrow is pointing toward a filopodial bleb-like protrusion. Magenta arrow is pointing toward a bleb-like protrusion originating from the cell body. (A, lower panel) Following an 8 s localized water release in close vicinity of a cell that was pre-treated with 25 µM HTS13286 no bleb-like formations were observed. Scale bar 5 µm. (B) Quantification of the filopodial diameter at the site of the bleb-like protrusion. Time 0 equals the image before localized H2O release. The lower graph shows the mean diameter for 8 s water release to cells untreated (green line), or treated with 25 µM of HTS13286 (black line).The data is displayed as mean±SEM, n = 8–18 filopodia/water release period. (C) Quantification of mean percentage of bleb-like protrusions originating from filopodia or the cell body after 1, 2, 4 or 8 s injection of micropipette-delivered H2O (n = 3 experiments).
Figure 7.
Water fluxes across AQP9 induce blebs.
(A) A confocal time-lapse montage of HEK-293 cells stably overexpressing GFP-AQP9. The images are linearly adjusted and pseudo-colored in fire scale to visualize fluctuations in fluorescence intensity. The arrow is pointing towards GFP-AQP9 accumulation in the membrane Scalebar 10 µm. The red box represents the area of measurement for MFI in (B). A smoothing filter was applied to this image to reduce background. (B) MFI-measurement of the blebbing membrane throughout the time-lapse. The area of measurement is presented in (A). (C) Zoom in panel of the blebbing membrane presented in (A). (D) Representative confocal time-lapse montage of a blebbing HEK-293 cell expressing both GFP-AQP9 and tagRFP-LifeAct. Scalebar 5 µm. (E) The same montage as presented in (D) showing a longer acquisition time to illustrate the complete lifetime of the same bleb. (F) Ratio measurements obtained from the sequence presented in (E). The ratio is measured as AQP9 MFI in the bleb membrane divided by submembraneous actin MFI showing a rapid initial increase in ratio due to the absence of actin during bleb formation. The ratio subsequently decreases due to increasing actin fluorescence in the bleb. The insert is a representative image showing the area of measurement for AQP9 and actin.
Figure 8.
Model for AQP9-induced membrane protrusion.
(A) A migrating cell with lamellipodia, filopodia, and blebs where an increased influx of water corresponds to a darker blue tone. (B1) Local accumulation of AQP9 by vesicle transport and/or lateral membrane diffusion enables a localized increased influx of water across the cell membrane. The influx is driven by an osmotic gradient, likely created by the transmembrane ion distribution (not shown). (B2) The rapid influx of water creates a localized hydrostatic pressure between the membrane and the cytoskeleton pushing the membrane outwards, thus initiating a membrane protrusion. (B3) The influx of water increases the hydrostatic pressure locally. In parallel, actin polymerization is promoted by the exposure of previously membrane-anchored barbed ends and the rapid diffusion of actin monomers in the now diluted, less viscous cytoplasm leading to an elongating filopodium. (B4) Then the rapid water-induced elongation reaches a critical distance from the actin, resulting in termination of the filopodial elongation likely due to equilibration of the water along the filopodium and loss of counter-pressure obtained from the actin cytoskeleton. (B5) The rate of the actin polymerization catches up with the water-induced protrusion and thereby stabilizes the structure. Based on the rate of water flux and equilibration, the filopodium can either protrude once more, or remain at its present length.