Rho-kinase directs Bazooka/Par-3 planar polarity during Drosophila axis elongation

Abstract

Cell rearrangements shape the Drosophila embryo via spatially regulated changes in cell shape and adhesion. We show that Bazooka/Par-3 (Baz) is required for the planar polarized distribution of myosin II and adherens junction proteins and polarized intercalary behavior is disrupted in baz mutants. The myosin II activator Rho-kinase is asymmetrically enriched at the anterior and posterior borders of intercalating cells in a pattern complementary to Baz. Loss of Rho-kinase results in expansion of the Baz domain, and activated Rho-kinase is sufficient to exclude Baz from the cortex. The planar polarized distribution of Baz requires its C-terminal domain. Rho-kinase can phosphorylate this domain and inhibit its interaction with phosphoinositide membrane lipids, suggesting a mechanism by which Rho-kinase could regulate Baz association with the cell cortex. These results demonstrate that Rho-kinase plays an instructive role in planar polarity by targeting Baz/Par-3 and myosin II to complementary cortical domains.

Figure 1. baz mutants have defects in cell intercalation

(A–D) Cell behavior in time-lapse movies of wild-type and mutant embryos. (A) Germband elongation is strongly reduced in baz^GD21 and sqh¹ mutants. (B) Fewer neighbors are lost per cell in baz and sqh mutants. An average value was obtained for each embryo, error bars indicate s.e.m. across embryos (n=10 WT, 5 baz^GD21, and 2 sqh¹ embryos). WT embryos were imaged with GFP:Resille (4), GFP:Spider (3), and E-cadherin:GFP (3), baz^GD21 embryos were imaged with GFP:Spider, sqh¹ embryos were imaged with E-cadherin:GFP. t=0 is the onset of elongation in stage 7. (C) Contractile behavior in wild type is biased toward edges perpendicular to the AP axis (angles in degrees, 0 for edges parallel to the AP axis). Bars indicate 10° bins starting at 0–10°. (D) Contractile behavior in baz mutants is reduced and partially spatially deregulated: 66% of shrinking edges were oriented at ≥60° in baz mutants compared to 88% in wild type, 16% were oriented at ≤30° in baz mutants compared to <1% in wild type (n = 10 wild-type and 5 baz^GD21 embryos, 33–161 cells/embryo). (E,F) Stills from movies of wild-type (E) and baz^GD21 (F) embryos expressing Spider:GFP. Asterisks indicate defects in ventral furrow closure in baz mutants and arrows point to clusters of apically constricting cells. Scale bar = 10 μm. See Figure S1, Movies S1,2.

Figure 2. Baz/Par-3 is required for the planar polarized localization of myosin II and β-catenin

(A,D,G,J) Localization of myosin II (Zip heavy chain, red), β-catenin (green), and Baz (blue) in stage 7 wild-type (A), baz^FA50 (D), baz^GD21 (G), and sqh¹ (J) embryos. Left panels show overlays of myosin (red) and β-catenin (green). Anterior left, ventral down. Scale bar = 10 μm. (B,E,H) Myosin planar polarity was disrupted in baz^FA50 (P<0.001) (E) and baz^GD21 (P<0.001) (H) compared to wild type (B) (n=5 wild-type, 3 baz^FA50, 4 baz^GD21 embryos). (C,F,I) β-catenin planar polarity was disrupted in baz^FA50 (P<0.001) (F) and baz^GD21 (P<0.001) (I) compared to wild type (C) (n=11 wild-type, 3 baz^FA50, 4 baz^GD21 embryos). (K,L) Baz planar polarity was retained in sqh¹ mutants (L) compared to wild type (K) (n=11 wild-type, 14 sqh¹ embryos). See Figure S2.

Figure 3. Baz/Par-3 is downregulated at AP edges and upregulated at DV edges, establishing planar polarity

(A–D) Stills from a movie of intercalating cells in an embryo expressing Myo:mCherry (red) and Baz:GFP under the control of a Gal4 driver (green). t = 0 is the onset of elongation in stage 7. A −11 min, B −4 min, C 3 min, D 9 min. (B) In stage 6, Baz:GFP is displaced from and Myo:mCherry accumulates at AP edges (arrowheads). (C,D) In stage 7, Myo:mCherry and Baz:GFP occupy AP and DV edges, respectively. Scale bar = 5 μm. (E,F) Quantitation of Baz:GFP fluorescence intensity. (E) The log₂ ratio of Baz:GFP intensity at DV edges (oriented at ≤30° relative to the AP axis) relative to AP edges (oriented at ≥60° relative to the AP axis) was averaged across all cells (n = 5 embryos, 764–2142 edges/embryo). Error bars indicate s.e.m. across embryos. t = 0 is the onset of elongation in stage 7. (F) Absolute edge intensities grouped by edge orientation in a single embryo. Baz is first downregulated at AP edges (yellow, purple) and then upregulated at DV edges (dark blue, green). See Figure S3, Movie S3.

Figure 4. Rho-kinase is required for Baz/Par-3 planar polarity

(A–H) Localization of Baz (green) and β-catenin (red) in stage 7 wild-type (A–D) and DRok² mutant (E–H) embryos. Cross sections in D,H. Quantitation of Baz and Arm/β-catenin planar polarity in wild type (B,C) and DRok² (F,G). Baz and β-catenin planar polarity were strongly reduced in DRok² mutant embryos compared to wild type (P<0.001 for each result) (n=11 wild-type, 8 DRok² embryos). Two classes of defects were observed, with the more defective class shown here. See Figure S4. (I–L) Localization of Baz (green) in stage 7 embryos injected with water (I) or Y-27632 (J). Baz planar polarity was reduced in Y-27632-injected embryos (K) compared to wild type (L) (P=0.002) (n=5 control injected, 4 Y-27632-injected embryos). (M,N) Localization of Baz (green) in the dorsolateral epidermis of a stage 14 control embryo (M) or an embryo expressing activated Rho-kinase (Rok-CA, N). Scale bars = 10 μm.

Figure 5. Rho-kinase is planar polarized in intercalating cells

(A,B) Localization of Rho-kinase (HA:Rok, red) and F-actin (phalloidin, green) in stage 7 wild-type (A) or LatB-injected (B) embryos. (C–L) Localization of Rho-kinase (Venus:Rok^K116A, red) and Baz (green). Stage 7 wild-type embryo (C) or embryos injected with baz dsRNA (E), LatB (G), eve and runt dsRNA (I), or Kr dsRNA (K). Quantitation of planar polarity in wild-type (D) and injected embryos (F,H,J,L). Rho-kinase planar polarity was not affected by baz RNAi (F) (n=6 wild-type Venus:Rok^K116A expressing embryos, 7 baz dsRNA-injected embryos, P=0.29). Rho-kinase planar polarity was reduced by injection of LatB (H) (5 embryos, P<0.001), eve and runt dsRNA (J) (7 embryos, P<0.001), or Kr dsRNA (L) (4 embryos, P=0.01). Scale bar = 10 μm.

Figure 6. Rho-kinase regulates the association of the Baz/Par-3 C-terminal domain with the cortex

(A–E) Drosophila S2R+ cells expressing Baz:Venus alone (green, A) or with activated versions of Rho-kinase (Rok CA red, B), Rho1 GTPase (Rho^V14 red, C), myosin regulatory light chain (sqh^E20,E21 red, D) or Lim kinase (limK^E591 isoform C red, E). Baz:Venus was cortical in 78% of cells without Rok CA (n=216) vs. 5% with Rok CA (n=284). Expression of Rho^V14, but not sqh^E20,E21 or limK^E591, led to a loss of Baz from the cortex. (F–K) Cells expressing Baz:Venus variants alone (green, F,H,J) or with Rok CA (red, G,I,K). (F,G) The Baz C-terminus is sufficient for cortical localization (F) and is displaced from the cortex by Rok CA (G). Baz 1097-1464 was cortical in 99% of cells without Rok CA (n=190) vs. 2% with Rok CA (n=178). (H,I) Deletion of the Baz C-terminus (BazΔ1097–1464) abolished Baz cortical localization in the absence or presence of Rok CA. (J,K) An adjacent region of the Baz C-terminus (Baz 905–1221) was not affected by Rok CA. Baz 905–1221 was cortical in 91% of cells without Rok CA (n=137) vs. 97% with Rok CA (n=208). (L,M) Cells expressing a GFP:Baz variant where the C-terminus was replaced with a heterologous PH domain. BazΔ1107–1464PHP localizes to the cortex (L) but was not efficiently displaced by Rok CA (M). GFP:Baz cortical localization was reduced 5.5-fold from 94% to 17% by Rok CA. GFP:BazΔ1107–1464PHP cortical localization was reduced 1.4-fold from 74% to 54% by Rok CA (n=183–257 cells/condition). Scale bar = 5 μm. See Figures S5,6.

Figure 7. The Baz/Par-3 C-terminal domain is necessary for Baz planar polarity

(A–C) GST-Baz fragments were incubated with activated Rho-kinase (Rok-CA) and γ-³²P-ATP. (A) Domains in each fragment. (B) Total protein quantified by Coomassie staining. Molecular weight markers in kD (top to bottom): 100, 75, 50 and 37. (C) Phosphorylated proteins were detected by autoradiography. Arrowheads indicate Rok-CA, which shows weak autophosphorylation. Asterisks indicate GST-Baz fragments, wild-type chick myosin regulatory light chain (MRLC) (positive control), and unphosphorylatable chick MRLC T19A,S20A (negative control). (D–K) Localization of GFP-Baz transgenes in stage 7 embryos. Endogenous baz was knocked down by dsRNA injection and embryos were labeled with antibodies to GFP (green) and β-catenin (red). (D,E) Full-length GFP-Baz displays wild-type planar polarity (5 embryos). (F,G) GFP-BazΔ1325–1464 localizes to the cortex but fails to localize asymmetrically (4 embryos, P<0.001). (H,I) GFP-BazΔ1097–1464 lacking a larger region of the C-terminus shows reduced cortical association and fails to localize asymmetrically (3 embryos, P<0.001). (J,K) GFP-BazΔ1107–1464:PHP where the C-terminal 358 amino acids of Baz were replaced with a heterologous PH domain restores cortical localization but fails to localize asymmetrically (6 embryos, P<0.001). Scale bar = 10 μm. See Figure S6.