(A) CKIε expression causes an electrophoretic mobility shift of SIPA1L1(FL) but not SIPA1L1(CΔ). IB, Immunoblot of whole cell lysates from HEK 293 cells expressing the indicated proteins. CKIε and α-tubulin abundance are shown in the middle and bottom panels.
(B) SIPA1L1 is phosphorylated by CKIε. Myc-SIPA1L1 plus CKIε or GFP (as the negative control for IP) were expressed in HEK 293 cells. Myc-SIPA1L1 was precipitated and treated with or without CIP.
(C) CKIε stimulates SIPA1L1 degradation. Lanes 1–5, Myc-SIPA1L1 was co-expressed with CKIε(WT) or CKIε(KD) as indicated. Lane 5 and lane 6, cells expressing Myc-SIPA1L1 were treated with IC261 (50 μM) for 5 hr. IC261 treatment increased SIPA1L1 abundance (compare lane 5 to lane3 and lane 6 to lane 2). Expression of CKIε(WT/KD) and endogenous α-tubulin abundance are shown in middle and lower panels.
(D) CKIε rescues SIPA1L1-mediated Rap1 inhibition. Myc-SIPA1L1(FL) or Myc-SIPA1L1(CΔ) was co-expressed with or without CKIε in the CHO cells, followed by GST-RalGDS pull-down assay. The abundance of GTP-Rap1 and total Rap1 was quantified by ImageQuant. The ratio of GTP-Rap1 to total Rap1 is shown in the lower panel.

(A) Expression of DN-Rap1 in Xenopus embryos causes open yolk plug at the tail region of the presumptive secondary axis. mRNA of Wnt-8 (5 pg) or CKIε (1 ng) were co-injected with DN-Rap1 (100 pg) into a single ventral cell of 4-cell stage embryos. Embryos were scored and photographed at stage 18. Whole-mount In situ hybridization for Sox-2 expression was performed on stage 14 embryos. Black arrowhead: double axis. Red arrowhead: gastrulation defect. Bottom, scoring of stage 18 embryos.
(B) Dorsal expression of DN-Rap1 alone causes a gastrulation defect in Xenopus embryos, whereas ventral expression does not affect gastrulation in the majority of embryos. DN-Rap1 RNA (500 pg) was injected into a single dorsal or ventral cell of 4-cell stage embryos. Embryos were photographed at stages 18 and 30 as indicated. Black arrowhead: normal axis. Red arrowhead: blastopore closure defect.
(C) Disruption of Rap1 activity inhibits activin-induced convergent extension of animal cap. Convergent extension assays were performed at the Xenopus embryos, dorsally injected with RNA encoding GFP, DN-Rap1 or SIPA1L1(CΔ). Animal caps were photographed when the sibling embryos reach stage 18. The Length-to-Width ratios of the activin-treated animal caps were measured with MetaMorph software and shown in the lower-right panel. Individual results for each animal cap are denoted by ◇; and the average is shown in the top and denoted by — in the graph.
(D) Xenopus embryos injected with Rap1-MO display a defect in gastrulation. Black arrowhead: normal axis. Red arrowhead: defect in blastopore closure. The CE defect could be rescued by wild-type (WT) hRap1. Control MO (40 ng) or MO against XRap1A (40 ng) and XRap1B (40 ng) were injected together into the animal cap of Xenopus at the 1-cell stage. 2 to 4 pg of HA-tagged hRap1 were injected to two of dorsal cells at 4-cell stage to rescue the Rap1-MO caused defects. Only 5% (n= 99) of the embryos exhibit gastrulation defects after rescue. Red arrowhead: blastopore closure defect.

(A) Injection of DN-Rap1 RNA, SIPA1L1(CΔ) RNA or Rap1B-MO in zebrafish embryos results in a bent and shortened AP axis. RNAs encoding GFP (500 pg; as negative control), DN-Rap1 (500 pg) or SIPA1L1(CΔ) (250 pg) were injected into zebrafish embryos at the 1–4 cell stages. Between 2–3 dpf, 42% (n=150) of DN-Rap1, 37% (n=237) of SIPA1L1(CΔ) and 58% (n=63) of Rap1B-MO injected embryos displayed a shortened and bent AP axis that varied in severity. AP defects were seen in only 6% (n=143) of GFP injected embryos. Embryos injected with either 2, 2.5 or 3.4 ng of Rap1B-MO showed similar phenotypes and were pooled for analysis.
(B) Down-regulation of Rap1 signaling in zebrafish results in cyclopia. Cyclopia (fused eyes at the midline) was seen between 2–3 dpf in a small percentage of embryos injected with DN-Rap1 (15%, n=94), SIPA1L1(CΔ) (7%, n=115) or Rap1B MO (15%, n=34). Cyclopia was seen in less than 1% of wild-type (n=226) and GFP injected (n=111) controls..
(C) The ventral midline and neural plate are broadened in Rap1 deficient embryos. In situ analysis of the ventral midline by sonic hedgehog (shh) staining and the anterior boundary of the neural plate by distalless (dlx3) staining at the 1–3 somite stages revealed a broadening of the midline and neural plate in DN-Rap1 (34%, n=29), SIPA1L1(CΔ) (22%, n=9) and Rap1B MO (42%, n=24) injected embryos. Similar defects were seen in only 14% of uninjected controls (n=77).

(A) In vivo interaction of SIPA1L1 and CKIε. Lanes 1–6 demonstrate the reciprocal co-immunoprecipitation (IP) of Myc-tagged SIPA1L1 and HA-tagged CKIε in HEK 293 cells. Lanes 7–12, immunoblot (IB) of whole cell lysates demonstrating the expression of Myc-SIPA1L1 and HA-CKIε. CKIε(WT): wild type CKIε; CKIε(KD): a kinase dead K38R mutant.
(B) SIPA1L1 binds to endogenous CKIε. Myc-SIPA1L1 was expressed in HEK293 cells and immunoprecipitated, followed by immunoblotting with a CKIε specific Ab. GFP was expressed as a negative control.
(C) The carboxyterminus of SIPA1L1 is required for interaction with CKIε. Upper and middle panels, co-IP of Myc-tagged SIPA1L1 truncated constructs and 3xflag-tagged CKIε. Bottom panel, expression of SIPA1L1 constructs. The figure is from a single exposure of the immunoblot; the lanes have been re-ordered for clarity.
(D) In vitro binding of SIPA1L1 to CKIε. Deletion constructs of SIPA1L1 were expressed in reticulocyte lysates. Recombinant 6x-His tagged CKIε was added where indicated. Luciferase was a negative control. In: Input; 20% of the amount used for binding assay.

(A) Wnt-8 stimulates CKIε-dependent phosphorylation of SIPA1L1. Protein two-dimensional electrophoresis of Myc-SIPA1L1 was performed after treatment of HEK 293 cells with (i) control medium, (ii) Wnt-8 conditioned medium, or (iii) Wnt-8 conditioned medium plus IC261. Arrows indicate Myc-SIPA1L1 with several different isoelectric points.
(B) Wnt-8 conditioned medium reduces the abundance of SIPA1L1. Upper panel, after 6-hours of conditioned medium incubation and IC261 treatment, immunoblot analysis of Myc-SIPA1L1 was performed. Lower panel, endogenous Dvl-1 abundance (as a loading control).
(C) Wnt-8 conditioned medium increases Rap1 activation while IC261 blocks basal and Wnt-8 mediated Rap1 activation. HEK293 cells were incubated with control or Wnt-8 conditioned medium and treated with or without addition of IC261 as indicated. After 6 hours, cells were lysed, followed by GST-RalGDS pull-down assay. Quantification of each lane was performed as described in Figure 2D and shown in the right panel.

(A) Dorsal expression of SIPA1L1(FL) or SIPA1L1(CΔ) alone in Xenopus embryos causes a gastrulation defect in 23% (n= 65) and 73% (n= 70) of embryos respectively. Ventral expression does not induce an obvious phenotype in the majority of injected embryos (95% and 79% unaffected, n= 39 and 38 respectively). 4-cell stage embryos were injected with 1 ng of RNAs encoding SIPA1L1(FL) or SIPA1L1(CΔ) and photographed at stage 18 and 30. Black arrowhead: normal axis. Red arrowhead: blastopore closure defect.
(B) SIPA1L1(CΔ) but not SIPA1L1(FL) inhibits gastrulation at the presumptive secondary axis. (i) Uninjected control. (ii-iv) Xenopus 4-cell stage embryos were injected dorsally with mRNA encoding (ii) CKIε (1 ng) + GFP (250 pg); (iii) CKIε + SIPA1L1(FL) (250 pg); (iv) CKIε + SIPA1L1(CΔ) (250 pg); and (v) CKIε + SIPA1L1(CΔ, mtGAP) (250 pg). In situ hybridization for Sox-2 expression was performed in stage 15 embryos. Black arrowhead: double axis. Red arrowhead: gastrulation defect. Scoring data of stage 18 embryos are shown in the lower-right panel.

(A) DN-Rap1 and SIPA1L1(CΔ) do not block dorsal gene expression but cause a expanded field of expression of β-catenin target genes. (A1 and A2) Uninjected control. (A3 and A4) Wnt-8 +GFP; (A5 and A6) Wnt-8 + DN-Rap1; (A7 and A8) CKIε + GFP; (A9 and A10) CKIε + DN-Rap1; (A11 and A12) CKIε + SIPA1L1(CΔ) mRNA were injected into a single ventral cell of 4-cell stage embryos using the same quantities as in Fig 4A and 5B. In situ hybridization assays probing Goosecoid (Gsc) and Chordin (Chd) were performed on stage 10.5 embryos.
(B) DN-Rap1 and SIPA1L1(CΔ)-mediated gastrulation defects are not rescued by β-catenin. (i) Uninjected control. (ii–iii) Xenopus embryos injected with mRNAs encoding β-catenin (100 pg ) + GFP (250 pg), (ii); β-catenin + DN-Rap1 (100 pg), (iii); β-catenin + SIPA1L1(CΔ) (250 pg), (iv). Injection, photography and scoring were performed as Figure 4A and the scoring results were shown in the right panel.
(C) Schematic model of CKIε-mediated Wnt-8 pathway. Wnt-8 signaling activates CKIε (➀; and (Swiatek et al., 2004)), leading to phosphorylation and degradation of SIPA1L1 (➁; and see the result in Figure 2A) and accumulation of β-catenin (➄; and (Gao et al., 2002)). Degradation of SIPA1L1 results in the accumulation of GTP-Rap1 (➂; and see the result of Figure 2D), promoting convergent extension during gastrulation (➃; and the results in Figures 4–6). In a separate pathway, stabilized β-catenin activates Lef1-TCF dependent gene expression (➄). Both Rap1-mediated convergent extension and β-catenin regulated gene transcription contribute to the proper development of body axis.