Cell-cycle-dependent TGFβ-BMP antagonism regulates neural tube closure by modulating tight junctions

Abstract

Many organs form by invaginating and rolling flat epithelial cell sheets into tubes. Invagination of the ventral midline of the neural plate forms the median hinge point (MHP), an event that elevates the neural folds and is essential for neural tube closure (NTC). MHP formation involves dynamic spatiotemporal modulations of cell shape, but how these are achieved is not understood. Here, we show that cell-cycle-dependent BMP and TGFβ antagonism elicits MHP formation by dynamically regulating interactions between apical (PAR complex) and basolateral (LGL) polarity proteins. TGFβ and BMP-activated receptor (r)-SMADs [phosphorylated SMAD2 or SMAD3 (pSMAD2,3), or phosphorylated SMAD1, SMAD5 or SMAD8 (pSMAD1,5,8)] undergo cell-cycle-dependent modulations and nucleo-cytosolic shuttling along the apicobasal axis of the neural plate. Non-canonical TGFβ and BMP activity in the cytosol determines whether pSMAD2,3 or pSMAD1,5,8 associates with the tight junction (PAR complex) or with LGL, and whether cell shape changes can occur at the MHP. Thus, the interactions of BMP and TGFβ with polarity proteins dynamically modulate MHP formation by regulating r-SMAD competition for tight junctions and r-SMAD sequestration by LGL.

Keywords: Apicobasal polarity; Hinge point; LGL; Midbrain; Neural tube closure defects; Organogenesis; PAR3.

Fig. 1.

Cell-cycle-dependent modulation and nucleo-cytoplasmic shuttling of pSMAD2,3 and pSMAD1,5,8. (A,C) Low-power hemi-sections of HH7 neural plates showing pSMAD2,3 (A) and pSMAD1,5,8 modulation (C) along the apicobasal (a↔b) axis. Blue arrowhead, ventral midline. The location of pHH3+ mitotic (m) cells and interphase (i) nuclei is indicated. Boxed areas in A,C include the MHP and are magnified in B–B″ and D–D″, respectively. The scale bar for A and C is shown in A. The lateral neural plate region demarcated by the blue brackets is magnified in Fig. S1. (B–B″) Colabeling of pSMAD2,3 with pHH3 (B′) and DAPI (B″) shows that pSMAD2,3 expression is high and nuclear during interphase (yellow arrowheads), and low and cytosolic during mitosis (white arrowheads). (D–D″) pSMAD1,5,8 expression is low and nuclear during interphase (yellow arrowheads), and high and ubiquitous (chromatin-associated and cytosolic) during mitosis (white arrowheads). The scale bar for B–B″ and D–D″ is shown in B″. (E,G) Low-power hemi-sections of HH6 neural plates showing colabeling of pSMAD2,3 (E) and pSMAD1,5,8 (G) with acetylated tubulin. Boxed areas in E,G (MHP) are magnified in F–F″ and H–H″, respectively. Blue arrowhead, ventral midline. The scale bar for E and G is shown in E. (F–F″) DAPI, pSMAD2,3 and acetylated tubulin colabeling demonstrate that pSMAD2,3 expression is predominantly nuclear during interphase (yellow arrowhead, inset i in F″) and both chromatin-associated and cytosolic during mitosis (white arrowhead, inset m in F″). (H–H″) DAPI, acetylated tubulin and pSMAD1,5,8 colabeling demonstrate that pSMAD1,5,8 expression is nuclear during interphase (yellow arrowhead, inset i in H″), and both chromatin-associated and cytosolic during mitosis (white arrowhead, inset m in H″). The scale bar for F–F″ and H–H″ is shown in F. (I,J) Quantification of pSMAD fluorescence at the HH6–HH7 MHP; data are presented as mean±s.e.m.; n=35 cells from six brains. pSMAD2,3 (I) and pSMAD1,5,8 (J) fluorescence in the nucleus and cytoplasm of interphase and mitotic cells was compared using the Mann–Whitney test as follows: (1) pSMAD2,3 at mitosis, cytosolic versus chromatin-associated: 4.83×10⁴±9.25×10² versus 3.9×10⁴±5.56×10²; *P≤0.05); (2) pSMAD2,3 at interphase, cytosolic versus nuclear: 3.24×10⁴±1.33×10³ versus 1.44×10⁵±5.11×10²; ****P≤0.0001); (3) pSMAD1,5,8 at mitosis, cytosolic versus chromatin-associated: 1.81×10⁵±5.04×10³ versus 2.37×10⁵±1.59×10⁴; **P≤0.01); (4) pSMAD1,5,8 at interphase, cytosolic versus nuclear: 3.60x10⁴±1.33x10³ versus 1.01×10⁵±2.63×10², ****P≤0.0001). Other relevant comparisons are indicated in the text.

Fig. 2.

TGFβ signaling induces MHP formation by suppressing BMP activity. (A–D) TGFβ misexpression is necessary and sufficient for MHP formation. (A–A″) EGFP electroporations (ep) at HH5 do not alter the apical contours of the lateral midbrain at HH7. The embryonic ages noted on all micrographs reflect the time points at which brains were examined. The boxed areas in the insets in A–D are magnified in adjacent panels. White and yellow dotted lines mark the apical and basal surfaces. Scale bars for A–B″ and C–D″ are shown in A″ and C″, respectively. (B–B″) caSmad2 misexpression (green) at HH5 induces ectopic hinge points in lateral midbrain (arrowhead). (C–C″) EGFP electroporations at HH5 showing the normal contours of the ventral midline (arrowhead). (D–D″) Lefty2 misexpression (green) at the ventral midline (white arrowhead) flattens the endogenous MHP. NC, notochord. (E–G) Cartoons depicting the opposite effects of TGFβ and BMP signaling on MHP induction. Black arrowheads, ventral midline; red arrowhead, ectopic hinge point. BMP data summarized from Eom et al., 2011. (H–K) Cross-inhibition of r-SMAD phosphorylation underlies TGFβ–BMP antagonism. Top panel, western blotting of electroporated whole-cell midbrain lysates. Electroporated DNA concentration (µg/µl) is indicated above each lane. Bottom panel, loading controls (α-tubulin). (H) Compared to EGFP controls, caSmad2 electroporations reduce, and Lefty2 and BMP4 electroporations increase pSMAD1,5,8 levels. (I) caSmad2 and Noggin electroporations increase, and Lefty2 and BMP4 reduce pSMAD2,3 levels. (J) The total levels of SMAD1 and SMAD2 proteins remain unchanged following BMP and TGFβ manipulations. (K) Cartoon illustrating that TGFβ and BMP phosphorylate their own r-SMADs, while inhibiting the phosphorylation of the r-SMADs from the other pathway.

Fig. 3.

TGFβ signaling regulates cell shapes and fates at the MHP. (A–B‴) Compared to controls (A–A‴), caSmad2+mRFP electroporated mitotic cells (B–B‴) display basally located nuclei (arrowhead) and apical constriction, demonstrated by a reduced apical width (aw):widest width (ww) ratio (see Materials and Methods). The boxed area in A″ is magnified in A‴ and shows mitotic cells in cytokinesis (left) and metaphase (right) stages. Scale bar for A and B is shown in A, and for A‴ and B‴ in A‴. (C) Quantification of basal nuclear localization in HH7 brains. Data are presented as mean±s.e.m. Nuclear distances from the apical surface for control (9.43 µm, n=85 cells from 17 brains) and caSmad2 electroporated cells (17.8 µm, n=100 cells from 17 brains) are significantly different. ****P≤0.0001 (Mann–Whitney test). (D) Quantification of apical constriction in HH7 brains. Data are presented as mean±s.e.m. The apical width:widest width (aw/ww) ratio of control cells=0.43 (n=70 cells/17 brains) and caSmad2-electroplated cells=0.3 (n=89 cells/17 brains) are significantly different. ****P≤0.0001 (Mann–Whitney test). (E–F‴) Compared to controls (E–E‴), midbrains electroporated with caSmad2 at HH4–HH5 display ectopic hinge points (arrowhead, F–F‴) with robust FOXA2 expression. E′–E‴ and F′–F‴ are magnified views of the boxes in E and F, respectively. Scale bar for E and F is shown in F, and for E′–E‴ and F′–F‴ is shown in E‴. See also Fig. S2E–H‴.

Fig. 4.

The TGFß signaling cascade interacts with the PAR polarity complex and regulates the subcellular localization of polarity proteins. (A–B′) Control electroporations (ep) with low, non-phenotypic (1 µg/µl) levels of Lgl1–GFP show a smooth PAR3+ apical contour and complete segregation between the apical (PAR3+) and the basolateral (LGL–GFP+) compartments. The inset in A″ is magnified in B,B′. Note the absence of PAR3 in EEA1+ endosomes (arrowheads) in B–B′. (C–D′) Co-electroporations of caSmad2 with low amounts of Lgl1–GFP result in ectopic hinge points (white arrowheads, C–C″) accompanied by the loss of apical PAR3, ectopic apical LGL–GFP and/or apical overlap between PAR3 and LGL–GFP. The inset in C″ is magnified in D,D′ and unlike controls (B,B′), shows colocalization of PAR3 with EEA1+ endosomes (yellow arrowheads, C″–D′). The scale bar for A–A″ and C–C″ is shown in C, and for B,B′ and D,D′ in B. (E–G′) Biochemical interactions between pSMAD2,3 and the PAR complex in EGFP-electroporated whole-cell lysates immunoprecipitated with anti-PAR3 (E), anti-PAR6 (F) and anti-aPKC (G) antibodies and immunoblotted with the anti-pSMAD2,3 antibody. The co-immunoprecipitations were reversed in E′–G′. (H,H′) Biochemical interactions between pSMAD2,3 and PAR3 in wild-type whole-cell lysates are weaker than those between pSMAD1,5,8 and PAR3. (I–I‴) Non-phenotypic PAR3–GFP electroporations followed by pSMAD2,3 immunohistochemistry show partial overlap between pSMAD2,3 and PAR3–GFP (arrowhead, I‴). The inset in I‴ shows a magnified view of the PAR3-GFP+ apical compartment in the cell marked by arrowhead.

Fig. 5.

TGFβ and BMP signals modulate pSMAD–PAR3 interactions in opposite directions. Each panel (A–F′) represents a single co-immunoprecipitation (IP) and western blotting (WB) experiment on cytosolic extracts, with lanes cut and aligned vertically for clarity. Experimental lanes in each panel should be compared to their own EGFP controls (top row) and not to EGFP controls in adjacent panels. Co-immunoprecipitation and western blotting in A–F were reversed in A′–F′, respectively. ep, electroporation. (A,A′) Compared to EGFP, caSmad2 increases and Lefty2 reduces PAR3–pSMAD2,3 interactions. (B,B′) caSmad2 reduces and Lefty2 increases PAR3–pSMAD1,5,8 interactions. (C–D′) BMP4 reduces PAR3–pSMAD2,3 interactions (C,C′) and increases pSMAD1,5,8–PAR3 interactions (D,D′). (E–F′) Noggin electroporations increase PAR3–pSMAD2,3 interactions (E,E′) and reduce PAR3–pSMAD1,5,8 interactions. (G) Top panel, western blotting of whole-cell lysates electroporated with EGFP, caSmad2, Lefty2, BMP4 or Noggin display similar levels of PAR3 protein. Bottom panel, loading controls. (H) Western blotting with anti-H1b (nuclear) and anti-α-tubulin (cytosolic) antibodies demonstrating the purity of cytosolic extracts prepared from EGFP and caSmad2-electroporated midbrains. (I) Cartoon summarizing the results shown in A–F.

Fig. 6.

LGL1–pSMAD interactions occur in the cytosol and are ligand dependent. (A) Absence of colocalization with DAPI demonstrates the cytosolic restriction of electroporated (ep) Lgl1–GFP. (B,C) Cytosolic extracts from Lgl1–GFP (1 µg/µl) electroporated cells display biochemical interactions between LGL1–GFP and pSMAD2,3 (B), and LGL1–GFP and pSMAD1,5,8 (C). Lgl1–GFP electroporations combined with pSMAD immunohistochemistry display cytosolic overlap between LGL1–GFP and pSMAD proteins (arrowheads, B,C). (D–K) Panels on the left represent single co-immunoprecipitation (IP) and western blotting (WB) experiments on cytosolic extracts, with lanes cut and aligned vertically for clarity. Experimental lanes in each panel should be compared to their own Lgl1–GFP controls (top row) and not to Lgl1–GFP controls in adjacent panels. Panels on the right provide immunohistochemical evidence of overlap (arrowheads) between LGL1–GFP and pSMAD2,3 or pSMAD1,5,8. (D,E) Compared to Lgl1–GFP controls, caSmad2 reduces and Lefty2 increases LGL1–GFP–pSMAD2,3 interactions. (F,G) Compared to controls (top panel), caSmad2 increases and Lefty reduces pSMAD1,5,8–LGL1–GFP interactions. (H,I) BMP4 increases LGL1–GFP–pSMAD2,3 interactions (H) and reduces pSMAD1,5,8–LGL1–GFP (I) interactions. (J,K) Noggin reduces pSMAD2,3–LGL1–GFP interactions (J) and increases pSMAD1,5,8–LGL1–GFP interactions (K). The scale bar in B applies to all immunohistochemical panels (B–K). (L) Cartoon summarizing data shown in B–K.

Fig. 7.

Cartoon summarizing cell-cycle-dependent TGF–BMP polarity interactions in regulating MHP formation. BMP signaling induces SMAD1,5,8 phosphorylation and reduces SMAD2,3 phosphorylation. This increases pSMAD1,5,8–PAR-complex interactions and reduces pSMAD2,3–PAR-complex interactions at apical junctions. BMPs also increase pSMAD2,3–LGL interactions and reduce pSMAD1,5,8–LGL interactions. Together these interactions create a stable epithelium. TGFβ signaling has the opposite effect and can disrupt epithelial organization and tight junction integrity either by increasing pSMAD2,3–PAR-complex interactions or decreasing pSMAD2,3–LGL interactions. Each cell cycles between the two states in a cell-cycle-dependent manner, giving rise to a dynamic epithelium capable of cell-shape changes, while retaining overall epithelial integrity. In addition, BMP and TGFβ signaling can modulate SLUG and NCAD expression in opposite directions, potentially through canonical mechanisms.