Fig 1.
Ift88 expedites cell migration in ciliated MDCK cells.
(A) Ift88-i1 cells were grown to confluency for 7–8 days (ciliated stage) and subjected to wounding. Cells depleted of Ift88 by inducible shRNA (+Tet) migrate slower compared with non-induced control cells (-Tet). The leading edge is shown after 2h and 6h. Scale bars: 100μm. (B) Quantification of migration speed in different cell lines expressing different tetracycline inducible shRNAs. No significant (n.s.) reduction in migration speed is observed in cells expressing unspecific shRNA (control; -Tet: 111.5 ±5.8 μm2/minute vs. 105.3 ±10.2 μm2/minute, p = 0.46, n = 4). Two independent cell lines expressing different shRNAs against Ift88 show significantly reduced migration speed: Ift88-i1: -Tet: 111.0 ±4.7 μm2/minute vs. +Tet: 64.6 ±10.3 μm2/minute, p<0.01, n = 4; Ift88-i2: -Tet: 124.7 ±1.2 μm2/minute vs. +Tet: 56.7 ±6.8 μm2/minute, p<0.01, n = 3. Migration speed is partially restored in Ift88-i1 cells expressing non-degradable Ift88 mRNA (Ift88-i1.rescue): -Tet: 116.5 ±5.5 μm2/minute vs. +Tet: 85.8 ±4.3 μm2/minute, p<0.01, n = 4.
Fig 2.
MDCK cells lose cilia during sheet migration.
(A) Confluent layers of ciliated MDCK cells (7 days after seeding) were scratched, fixed immediately (0h) or 6h after wounding, then stained with antibodies against acetylated Tubulin for cilia (magenta), Cep164 for the basal body (green) and Hoechst for nuclei (blue). Cilia occur at different lengths, basal bodies devoid of magenta are unciliated (squared). White lines indicate the leading edge. Scale bars: 10μm. (B) Quantification of ciliated cells at various time points in areas away from the leading edge and at the leading edge. 0h: 56.7%±3.3% vs. 31.0%±4.5%, 2h: 58.1%±1.7% vs. 20.2%±4.7%, 4h: 53.3%±2.1% vs. 15.0%±3.0%, 6h: 54.6%±3.1% vs. 7.5%±2.7% (asterisk: p < 0.01). n = 10 fields of view in two independent experiments. Note that the number of cilia at time point 0h at the leading edge is decreased compared to distant cells due to mechanical injury from wounding. (C) Analysis of proliferation in cells after wounding. MDCK cells stably expressing Fucci cell-cycle indicators were monitored after 7 days post seeding after scratch wounding. Nuclei of non-proliferating cells in G1 express RFP (red), nuclei of proliferating cells (S/G2/M) express GFP (green). Scale bars: 100μm. (D) Quantification of proliferating cells at the leading edge and cells away from the wound over 21 hours after scratch. n = 10 fields of view from two independent experiments.
Fig 3.
The number of ciliated MDCK cells after seeding increases over time.
(A) Cells were fixed at respective time points and stained with antibodies against acetylatedTubulin for cilia (magenta), Cep164 for the mother centriole (green) and Hoechst for nuclei (blue). Arrows point to mother centrioles which do not have cilia. (B) The number of ciliated cells was quantified over time: 6h: 0.0% ±0.0%, 3420 cells; 1d: 0.0% ±0.0%, 9347 cells; 2d: 0.1% ±0.1%, 18600 cells; 3d: 3.5% ±1.7%, 18309 cells; 7d: 43.2% ±4.7%, 17418 cells and 10d: 51.9% ±3.6%, 17475 cells. n = 3 (32 fields of view per N).
Fig 4.
Ift88 expedites cell migration in previously unciliated MDCK cells.
(A) Ift88-i1 cells were grown to confluency for 2 days (unciliated stage, compare Fig 3) and subjected to wounding. Cells depleted of Ift88 by inducible shRNA (+Tet) migrate more slowly compared with non-induced control cells (-Tet). The leading edge is shown after 2h and 6h. Scale bars: 100μm. (B) Quantification reveals no significant reduction of migration speed for a control cell line: -Tet: 157.1 ±12.8 μm2/minute vs. +Tet: 144.8 ±10.6 μm2/minute (n.s.; p = 0.5). Migration speed is reduced in two independent cell lines after induced depletion of Ift88. Ift88-i1: -Tet: 133.9 ±7.0 μm2/minute vs. +Tet: 50.4 ±2.5 μm2/minute, p<0.01, Ift88-i2: -Tet: 193.0 ±7.4 μm2/minute vs. + Tet: 86.0 ±2.9 μm2/minute, p<0.01. Migration speed is restored in Ift88-i1 cells expressing non-degradable Ift88 mRNA (Ift88-i1.rescue): -Tet: 162.0 ±9.8 μm2/minute vs. +Tet: 154.6 ±10.0 μm2/minute, p = 0.63. All n = 3. (C) Sparsely seeded Ift88-i1 cells were stimulated with HGF and the morphology was analyzed. Ift88 depleted cells (+Tet) frequently exhibit an unpolarized pancake shape, while non-induced controls (-Tet) more often show a leading and a trailing edge (A representative cell stained for α-Tubulin and a Golgi marker is shown in S1E Fig). Phase contrast images. Scale bars: 10μm. (D) Quantification of the proportion of pancake-shaped cells. No significant increase is seen after exposure to tetracycline in a control cell line: -Tet: 18.3 ±1.9% vs. +Tet: 17.8 ±0.7%, p = 0.87, n = 4, a total of 269 and 392 cells counted. Depletion of Ift88 increases the number of unpolarized cells: -Tet: 19.9 ±2.4% vs. +Tet: 50.2 ±2.9%, p<0.01, n = 4, a total of 454 and 362 cells counted. (E) Representative trajectories of single cell migration over 5 hours after stimulation of sparsely seeded cells with HGF (10ng/ml). Scale Bar: 300μm. (F) Quantification of track lengths over 5h. In control cells no significant reduction of track lengths is observed after incubation with tetracycline: -Tet: 360.4 ±27.0 μm vs. +Tet: 348.9 ±24.3 μm, p = 0.76, n = 40 tracks from four independent experiments. A reduction is observed in Ift88 depleted cells: -Tet: 426.7 ±28.5 μm vs. +Tet: 213.8 ±21.3 μm, p<0.05, n = 40 tracks from four independent experiments.
Fig 5.
Ift88 facilitates polarization in migrating cells.
(A) Ift88 depleted (+Tet) or control cells (-Tet) were subjected to wound healing as in Fig 4A. After 6 hours cells were fixed, stained with antibodies against Scrib (green) and Hoechst for nuclei (blue), and the leading edge was imaged. Scale bars: 10μm. White arrows point to the leading edge in absence of Scrib. (B) Score based quantification of Scrib localization (0 (not polarized), 1 (partially polarized) or 2 (fully polarized)). After tetracycline induced depletion of Ift88 fewer cells have a fully polarized score (23.5%) compared to the absence of tetracycline (49.2%), n = 3, -Tet 478 cells/ +Tet 468 cells. (C) Ift88-i1 cells were subjected to wound healing assay as in Fig 4A. After 6 hours cells were fixed, stained against Golgin-97 for the Golgi apparatus (magenta) and Hoechst for nuclei (blue), and the leading edge was imaged. Red arrows show the direction of polarization of the Golgi apparatus relative to the nucleus and the leading edge within ±60°; white arrows illustrate unpolarized Golgi (not within ±60°). Scale bars: 10μm. White line illustrates the leading edge. (D) Quantification of Golgi apparatus polarization (for details see methods) showed a significant decrease of polarized Golgi in Ift88-knockdown conditions (+Tet: 37.5 ±4.8%) compared to non-induced controls (-Tet: 53.6 ±3.3%), p<0.05 (Chi2-Test), n = 3, -Tet: 258 cells/ +Tet: 269 cells. (E) After 6 hours of migration Ift88-i1 cells were fixed, stained against α-Tubulin (white) and Hoechst for nuclei (blue) showing reduced numbers of MTs in the leading edge in Ift88-knockdown conditions (+Tet). Scale bars: 10μm. (F) Quantification of the MTs in the leading edge (conducted in Ift88-i1/ α-Tubulin-YFP cells, see also Fig 6A) revealed significantly fewer MT in Ift88 deficient cells (-Tet: 23.7±3.4 vs. +Tet: 12.4±2.5, p<0.05, n = 10/11 cells in two independent experiments).
Fig 6.
Ift88 does not influence MTOC activity.
(A) Migrating Ift88-i1 cells with stably transduced α-Tubulin-YFP (Ift88-i1.Tubulin-YFP) were subjected to TIRF microscopy and MT were assessed. The number of MT probing into the leading edge is reduced upon Ift88-knockdown compared to non-induced controls, but (B) no changes in MT dynamic instability were observed. (C) The amount of γ-Tubulin at the centrioles appears similar in Ift88-knockdown cells (+Tet) compared to non-induced controls (-Tet). Lower panel shows magnifications of the white squares. (D) Quantification of γ-Tubulin fluorescence intensity reveals no significant difference between -Tet (188.3 ±27.3 a.u.) and +Tet (201.5 ±25.1 a.u.) conditions. p = 0.72, n = 6, 19 fields of view each, 556/461 cells. (E, F) Migrating Ift88-i1 cells as in Fig 4A were subjected to MT depolymerization with nocodazole and subsequently MT regrowth was analyzed. Dotted line illustrates the leading edge. No significant difference was observed in aster intensity (-Tet: 84.0 ±9.9 a.u. vs. +Tet: 65.2 ±11.0 a.u., p = 0.27, n = 3, total of 33/34 centrosomes). Ift88-knockdown was confirmed by staining for Ift88 comparing fluorescence intensity at the mother centriole in ±Tet conditions (-Tet: 84.1 ±11.8 a.u vs. +Tet: 34.8 ±11.0 a.u., p<0.05, n = 3). Scale bars: 10μm.