Ajuba Interacts with Isl1 (A) FLAG-HA-tagged Ajuba and Isl1 were transiently expressed in HEK293T cells. Proteins were immunoprecipitated with an anti-HA antibody and detected by immunoblot with an anti-Isl1 antibody. IP, immunoprecipitation; WB, western blot. (B) Isl1 and Myc-tagged LPP, Zyxin, LIMD1, and WTIP were transiently expressed in HEK293T cells, and immunoprecipitation with an anti-Isl1 antibody followed by immunoblot analysis with an anti-myc antibody was performed. (C) Coimmunoprecipitation of extracts from day 5 EBs differentiated from ES cells using anti-Isl1 antibody and detected with anti-Ajuba antibody. (D) Schematic representation of wild-type Isl1 and Isl1 deletion constructs (top). FLAG-HA-tagged Ajuba and Isl1 or Isl1 deletion constructs were transiently expressed in HEK293T cells, and immunoprecipitation with an anti-Isl1 antibody followed by immunoblot analysis with an anti-HA antibody was performed (bottom). (E) Schematic representation of wild-type Ajuba and Ajuba deletion constructs (top). FLAG-HA-tagged Ajuba or Ajuba deletion constructs and Isl1 were transiently expressed in HEK293T cells, and immunoprecipitation with an anti-Isl1 antibody followed by an immunoblot analysis with an anti-HA antibody was performed. NES, nuclear export sequence.

Ajuba Colocalizes with Isl1 and Represses Its Transcriptional Activity(A) Subcellular localization of Isl1 and Ajuba in HEK293T cells. FLAG-HA-tagged Ajuba and Isl1 full-length or deletion constructs were transiently expressed and detected by immunofluorescence using anti-HA or anti-Isl1 antibodies. Scale bars, 50 μm.(B and C) HEK293T (B) and NIH 3T3 (C) cells were transiently transfected with a 40 ng (B) or 50 ng (C) luciferase reporter construct containing the Mef2c cardiac-specific enhancer fragment (Dodou et al., 2004) in front of a minimal fos promoter (Mef2c-luc), alone or together with Isl1 (400 ng), Isl1 lacking the DNA-binding domain (Isl1ΔHOMEO, 400 ng), Ajuba (400 ng), or with constant amount of Isl1 (400 ng), and increasing amounts of Ajuba expression plasmid (100, 200, 300, 400 ng). The luciferase levels were normalized for the β-galactosidase activity of a cotransfected RSV-lacZ reporter and presented as fold activation relative to the luciferase levels of the Mef2c reporter construct alone. Transfections were performed at least three times in triplicates, and representative experiments with the SDs are shown.(D) ChIP of nuclear extracts from day 5 EBs, using mouse or rabbit IgG (mIgG and rIgG, respectively) as a control, anti-Isl1 and anti-Ajuba antibodies, respectively. PCRs were performed using primers flanking the two conserved Isl1-binding sites in the Mef2c AHF enhancer.(E) Confocal images of cryosections of E9 embryos stained with anti-Ajuba (green) and anti-Isl1 (red) antibodies. Scale bars, 100 μm. See also Figure S1.

Ajuba Knockdown in Zebrafish Leads to an Increased Number of Cardiomyocytes at the Arterial and Venous Poles (A) Schematic representation of the Ajuba genomic locus (left) and the exons coding the three LIM domains (drawn to scale, right). The binding sites of the AUG and the splice morpholino are shown in red. (B) Control or AUG MO-injected Tg(myl7:EGFP-HsHRAS)^s883 embryos at 48 hpf. Lateral views, anterior to the left, are in the top-left and bottom panels. Frontal views of the same embryos are in the middle and right panels. (C) Confocal images of control and Ajuba morphant Tg(myl7:EGFP-HsHRAS)^s883/Tg(5.1myl7:nDsRed2)^f2 hearts. (D) Total number of atrial cells and number of mGFP⁺DsRed⁺ and mGFP⁺DsRed cells in the atrium at 48 hpf. Error bars represent SD (n = 5). *p < 0.05; ***p < 0.001. (E) Confocal images of control and Ajuba AUG morphant ventricle at 48 hpf. (F) Total number of ventricle cells. p < 0.01. See also Figure S2 and Table S1.

Ajuba and Isl1 Restrict the Number of Isl1-Expressing Cells in the Heart (A) Confocal images of control and Ajuba morphant Tg(myl7:EGFP-HsHRAS)^s883 embryos stained with anti-GFP and anti-Isl1 antibodies at 24 somites. (B and C) Confocal images of control and Ajuba morphant Tg(myl7:EGFP-HsHRAS)^s883 embryos stained with anti-Isl1 antibodies (red) and visualized by GFP fluorescence (green) for myl7 expression at 30 hpf (B) and at 48 hpf (C). (D) Control and zAjuba mRNA-injected Tg(myl7:EGFP-HsHRAS)^s883 embryos at 48 hpf, frontal view (top); in situ hybridization for Isl1 in control or zAjuba mRNA-overexpressing embryos at 48 hpf (bottom). (E) Confocal images of hearts of control, Ajuba, Isl1, and Ajuba/Isl1 morphant Tg(myl7:EGFP-HsHRAS)^s883 embryos, stained with anti-GFP antibody. Arrows show differentiated cardiomyocytes in the pericardial wall adjacent to the heart. (F) In situ hybridization for Isl1 in control, Ajuba, Isl1, and Ajuba/Isl1-deficient embryos. (G) Immunostaining with anti-GFP and in situ hybridization for Isl1 in control (left) and Ajuba/Isl1-deficient Tg(myl7:EGFP-HsHRAS)^s883 (right) embryos from (F). See also Figure S3.

Ajuba Restricts the Number of SHF Progenitor Cells (A) Confocal images of Nkx2.5:GFP transgenic embryos stained with anti-Isl1 and anti-GFP antibodies at the 10 somite stage. The top-middle panel shows a higher magnification of the image in the top-left panel; the lines indicate the position of the transverse (xz) optical sections presented in top-right panel. NT, neural tube; Tg, trigeminal placode. (B) In situ hybridization of control and Ajuba morphant embryos at 3, 5, 10, and 15 somite stage with an Isl1 probe. (C) In situ hybridization of control and Ajuba morphant embryos with probes for Mef2ca, Mef2cb, Tbx20, Tbx1, and Bmp4. (D) Relative expression level of cardiac progenitor and cardiomyocyte marker mRNA normalized to GAPDH mRNA in day 5 EBs overexpressing either GFP only (control) or both GFP and Ajuba. Error bars represent SDs derived from two independent pools of control and Ajuba-overexpressing ES cells subjected to differentiation in EBs. See also Figures S4, S5, and S6.

The Isl1-Ajuba Transcriptional Complex Negatively Regulates Isl1 Expression in an RA-Dependent Mechanism (A) Relative expression levels of cardiac progenitor marker mRNA normalized to GAPDH mRNA in day 5 EBs either nontreated or treated with 0.1 or 0.3 µM RA at day 3. Data are mean ± SD (n = 2). *p < 0.05; **p < 0.01; ***p < 0.001. (B) Western blot analysis of nuclear extracts from day 5 EBs either nontreated or treated with 0.1 or 0.3 μM RA at day 3.(C) ChIP of nuclear extracts from day 5 EBs either nontreated or treated with 0.1 μM RA at day 3, using anti-Isl1 and anti-Ajuba antibodies or IgG as a control. PCRs were performed using primers flanking a conserved Isl1 binding site in the Isl1 enhancer. (D and E) P19 cells were transiently transfected with 25 ng Isl1-luc reporter construct (Isl1-F; Kang et al., 2009), alone or together with Isl1 (400 ng), Ajuba (900 ng), or with constant amount of Isl1 (400 ng) and increasing amounts of Ajuba expression plasmid (500 and 900 ng). P19 cell were either nontreated (D) or treated with RA (E). Data are mean ± SD (n = 4). p < 0.01; ***p < 0.001. (F) In situ hybridization for the atrial marker atrial myosin heavy chain (Amhc; top panels) and Isl1 (bottom panels) at 48 hpf in control and zebrafish embryos, treated with RA for 1 hr. WT, wild-type. (G) In situ hybridization for Isl1 in control and Ajuba morphant and RA-treated control or Ajuba morphant zebrafish embryos at 48 hpf. See also Figure S7 and Tables S2 and S3.

Ajuba Regulates Nkx2.5 Levels(A) In situ hybridization for Nkx2.5 in control and Ajuba morphant embryos at 10 and 15 somites.(B) Relative expression level of Nkx2.5 normalized to Rpl13α in pools of control and Ajuba morphant embryos at 10 and 15 somites. Error bars represent SDs derived from three independent pools of control (n = 10) and Ajuba morphant embryos (n = 10). **p < 0.01.(C) FLAG-HA-tagged Ajuba and Nkx2.5-myc expression plasmids were transiently expressed in HEK293T cells, and immunoprecipitation with an anti-HA antibody followed by immunoblot analysis with an anti-myc antibody was performed.(D) Western blot analysis of lysates from day 5 EBs overexpressing either GFP only or both GFP and Ajuba using Nkx2.5 antibody. The two lanes were spliced from the same exposure of the same blot.(E) HEK293T cells were transfected with constant amounts of Nkx2.5-myc (5 μg) and FLAG-H3.3 (1 μg, used as a control for transfection efficiency) expression plasmids and increasing amounts of an Ajuba plasmid (0.25, 0.5, 0.75 μg). Immunoblot analysis of equal amounts of total protein extracts was performed using either anti-myc, anti-HA or anti-FLAG antibodies.(F) Western blot analysis of whole-cell lysates from HEK293T cells treated with either DMSO or MG-132 using Nkx2.5 antibody. Ponceau staining served as loading control.(G) In situ hybridization of control and Nkx2.5/Nkx2.7 morphant (Nkx-deficient) embryos at 3, 5, 10, and 15 somite stages and at 48 hpf with an Isl1 probe.(H) Model for the control of the SHF progenitor pool by Ajuba.

(related to Figure 2). Isl1 co-localizes with Ajuba and WTIP

(A) Isl1 and FLAG-HA-tagged Ajuba lacking the LIM3 domain, which is not responsible for its interaction with Isl1, were transiently expressed in HEK 293T cells and immunofluorescent staining with anti-HA or anti-Isl1 antibodies was performed. The co-expression of Isl1 and the AjubaΔLIM3 protein affected the localization of both proteins: Isl1 was retained in the cytoplasm, and concomitantly, a significant amount of the Ajuba protein localized in the nucleus. (B) Subcellular localization of Isl1 and the members of the Ajuba protein family, WTIP and LIMD1, in HEK 293T cells. Myc-tagged WTIP, LIMD1 and Isl1 were transiently expressed, and immunofluorescence with anti-myc and anti-Isl1 antibodies was performed. Strong co-localization of WTIP with Isl1 was observed, but no co-localization was observed for LIMD1. (C) Subcellular localization of Isl1 and Zyxin in HEK 293T cells. Myc-tagged Zyxin and Isl1 were transiently expressed, and immunofluorescence with anti-myc or anti-Isl1 antibodies was performed. No co-localization was observed. (D) Subcellular localization of Isl1 and Ajuba from Danio rerio in HEK 293T cells. FLAG-HA-tagged Ajuba (D. rerio) and Isl1 (D. rerio) were transiently expressed, and immunofluorescence with anti-HA and anti-Isl1 antibodies was performed. Co-transfection of Ajuba and Isl-1 led to substantial co-localization.

(related to Figure 3). Confirmation of the efficiency and the specificity of morpholino targeting

(A) Schematic representation of the Ajuba genomic locus and the binding sites of the AUG and the splice morpholino (shown in red). Arrows indicate the positions of the primer pair used for the RT-PCR analysis in (C), (B) To confirm the efficiency of the AUG morpholino we constructed a pCS2-Ajuba-GFP plasmid, in which the binding sequence of the AUG morpholino was fused in frame with the GFP cDNA. Embryos injected with 200 ng Ajuba-GFP mRNA, synthesized from the pCS2-Ajuba-GFP plasmid, showed GFP expression. In contrast in embryos injected with 200 ng Ajuba-GFP mRNA and 5 ng Ajuba AUG MO, GFP fluorescence was not observed. (C) The efficiency of Ajuba sp-MO was analyzed by PCR using primers spanning the first intron of the Ajuba gene. This analysis indicated loss of spliced Ajuba mRNA in embryos injected with the Ajuba sp-MO. (D) Control or Ajuba sp-MO injected Tg(myl7:EGFP-HsHRAS)^s883 embryos at 48 hpf. Frontal views, showing an enlarged atrium and reversed left–right asymmetry of the heart.

(related to Figure 4). Ajuba proteins regulate the number of Isl1-expressing cells in the heart

(A) Number of Isl1⁺ cells at the venous pole in and at the periphery of the heart at different developmental stages. (B) Number of Isl1⁺ cells at the venous pole in and at the periphery of the heart in control and Ajuba morphant embryos at 30 and 48 hpf. n = 3; *p<0.05; **p<0.01. (C) Number of Isl1⁺ cardiomyocytes at the venous pole in control and Ajuba morphant embryos at 48 hpf. n = 3; **p<0.01. For these experiments (A, B, C), confocal images of control and Ajuba-morphant Tg(myl7:EGFP-HsHRAS)^s883 embryos stained with anti-Isl1 antibodies (red) and visualized by GFP fluorescence (green), were projected using ImageJ and Isl1⁺ cells or Isl1⁺ cardiomyocytes were counted. (D) In situ hybridization for Isl1 of control, LIMD1 and WTIP morphant embryos at 48 hpf. (E) Total number of cardiomyocytes in control, Ajuba, Isl1 and Isl1/Ajuba-deficient zebrafish embryos at 48 hpf. n = 3; ***p<0.001. (F) In situ hybridization for Isl1 of control, mIsl1 and mIsl1ΔLIM2 mRNA injected embryos at 48 hpf.

(related to Figure 5). Expression patterns of Ajuba, Isl1 and Nkx2.5

(A, A′, A′′) In situ hybridization for Ajuba in wild-type zebrafish embryos at 5, 10 and 15 somites. Dorsal views, anterior to the top. (B, B′, B′′) In situ hybridization for Ajuba in wild-type zebrafish embryos at 5, 10 and 15 somites. Dorsoventral views, anterior to the left. (C, C′, C′′) In situ hybridization for Isl1 in wild-type zebrafish embryos at 5, 10 and 15 somites. Dorsal views, anterior to the top. (C′′′) In situ hybridization for Isl1 in wild-type zebrafish embryos at 5 somites. Dorsoventral views, anterior to the left. (D, D′) In situ hybridization for Nkx2.5 in wild-type zebrafish embryos at 10 and 15 somites. Dorsal views, anterior to the top.

(related to Figure 5). Isl1 expression in cardiac progenitor cells.

(A, A′, A′′) Expression of the Nkx2.5:GFP transgene in anterior lateral plate mesoderm, adjacent to pharyngeal endoderm, marked by her5:mCherry (Tallafuss and Bally-Cuif, 2003) at different developmental stages. (B) In situ hybridization for Nkx2.5 and immunostaining for GFP (Nkx2.5), showing a similar domain of expression of the Nkx2.5:GFP transgene and the Nkx2.5 mRNA (C, C′, C′′, C′′′). Confocal images of Nkx2.5:GFP transgenic embryos stained with anti-Isl1 and anti-GFP antibodies at the 15 somite stage. C′′ shows the position of the transverse (xz) optical sections presented in C′′′1, 2, 3. (D, D′, D′′, D′′′) Confocal images of Nkx2.5:GFP transgenic embryos stained with anti-Isl1 and anti-GFP antibodies at the 18 somite stage. D′′ shows the position of the transverse optical sections presented in D′′′1, 2, 3. (E, E′, E′′) Confocal images of Nkx2.5:GFP transgenic embryos stained with anti-Isl1 and anti-GFP antibodies at the 21 somite stage. (F, F′, F′′) Confocal images of Tg(myl7:EGFP-HsHRAS)^s883 embryos stained with anti-Isl1 and anti-GFP antibodies at the 24 somite stage. NT, neural tube, Tg, trigeminal placode.

Ajuba deficiency does not affect Isl1⁺ cardiac progenitor cell proliferation.

Confocal images of control (lower magnification: A, A′, A′′, higher magnification C, C′, C′′) and Ajuba morphant embryos (lower magnification: B, B′, B′′ higher magnification D, D′, D′′) stained with anti-Isl1 and anti-phospho-H3 antibodies at the 10 somite stage.

(related to Figure 6). Retinoic acid restricts the number of Isl1⁺ cardiac progenitors

(A) Relative expression level of Isl1 and Nkx2.5 normalized to Rpl13α in pools of control embryos (n=10) and embryos, treated with 0.1 μM RA (n=10) or 0.3 μM RA (n=10) at 10 somites. Error bars represent standard deviations derived from two independent pools of control and RA treated embryos. (B) In situ hybridization for Isl1 in control and RA treated zebrafish embryos at 10 somites. (C) In situ hybridization for Isl1 in control and DEAB treated zebrafish embryos at 10 somites.

Unillustrated author statements