Fig 2: SPARCL1 is expressed in blood vessels of metatarsals and inhibits angiogenesis ex vivo. A, Tissues covering metatarsals of WT embryos at E18.5 (n = 4) were stained for SPARCL1 (green), CD31 (red, upper), and αSMA (red, lower). Examples of co-localization of SPARCL1 and CD31 indicated by arrows and examples of co-localization of SPARCL1 and αSMA indicated by arrowheads. Draq5 [1, 5-bis{[2-(di-methylamino)ethyl]amino}-4, 8-dihydroxyanthracene-9, 10-dione] was used as counterstain. Scale bar: 50 μm. B, WT metatarsals of embryos at E18.5 were either untreated (n = 17), treated with recombinant murine SPARCL1 (mSPARCL1, 1.5 μg/mL; n = 8), or treated with vascular endothelial growth factor (100 ng/mL; n = 11) alone or in combination with mSPARCL1 (n = 10). Vessel endothelial cells are visualized by CD31 staining. One representative picture of 3 experiments is depicted. Scale bar: 500 μm. Quantification of relative numbers of branches and junctions includes pooled results of all experiments. Error bars indicate SD. *P < 0.05 and ****P < 0.0001 by 2-tailed, unpaired Student t test.

Fig 3: Hevin mutants lacking the EF-hand motif are generated in the Usp15-deficient brain. (A) The relationship between the splicing error and ER stress in the Usp15-deficient brain has not yet clarified. (B) Schematic structure of the hevin transcript in the Usp15-deficient mouse brain. The 3'-region of the hevin transcript (PSR/Junction ID: 4,776,772) is prone to be missing in the Usp15-deficient brain. (C) The putative schematic structures of Hevin mutants. (D–G) Sequence analyses of the 3'-regions of hevin transcripts. (D) DNA sequencing chromatogram of the 3'-region of hevin WT and ΔEF transcripts. (E) The 3'-region of hevin sequence. Black box indicates the lacking sequence in ΔEF mutant. Cyan indicates the EF-hand motif. (F) DNA sequencing chromatogram of the 3'-region of hevin WT and Ndufa11. (G) The 3'-region of hevin sequence. Black box indicates the lacking sequence in the Hevin-N11 mutant. Cyan indicates the EF-hand motif.

Fig 4: Wounding induces myofibroblast formation, collagen deposition, and TGF-ß target-gene expression in Ezh2 cKO but not control deciduaImplantation sites of E6.5 control and Ezh2 cKO mice were wounded by piercing them with a fine needle, through which we injected lentiviruses engineered to constitutively express GFP. The mice then received daily P4 injections until their sacrifice 48 h later.(A) Immunofluorescence detection of a-SMA+ and GFP, which marks wound sites. Representative images from n = 7 mice/group (2–8 wound sites/mouse). Scale bar, 50 µm.(B) Quantification of a-SMA expression. A blinded observer scored the quantity of a-SMA expression at wound sites from a scale of 0 (no a-SMA+ cells) to 3 (copious a-SMA+ cells) (mean ± SEM; n = 6 mice/group, 2–8 wound sites/mouse; ***p < 0.0001, Student’s t test).(C) SPARCL1 and total collagen expression at wound sites, detected by immunofluorescence (SPARCL1) and a fluorescent pan-collagen reagent (total collagen; see STAR Methods). Serial sections were stained as indicated. Representative images from n = 7 mice/group (2–5 wound sites/mouse). Scale bar, 100 µm.(D and E) Immunofluorescence detection of Ly6G+ neutrophils (D) and F4/80+ macrophages (E) at GFP+ wound sites within the Ezh2 cKO decidua. Serial sections were stained as indicated. Very few neutrophils or macrophages (arrowheads) are present. Representative images from n = 3 mice (2–3 wound sites/mouse) for the Ly6G staining and n = 5 mice (1–3 wound sites/mouse) for the F4/80 staining. Scale bars, 50 µm.(F) H&E-stained wound sites from control and Ezh2 cKO implantation sites. Serial sections were stained with GFP to identify wound sites (upper panels). Representative images from n = 4 mice/group (1–3 wound sites/mouse). Scale bars, white (immunofluorescence), 50 µm and black (H&E), 100 µm.

Fig 5: Hevin specifically induces thalamocortical synapse formation in vitro and in vivo.(A) Schematic of the cortical/thalamic neuron co-culture system. (B) Representative images of cortical neurons cultured for 14 days in vitro (DIV), either alone or in equal densities with thalamic neurons, with or without hevin treatment. Insets show individual channels for VGlut1 (green) and PSD95 (magenta) staining, as well as the merged image. Co-localized puncta (white, arrows) represent synapses. Scale bars: 10 µm (main image), 5 µm (inset). (C) Hevin did not induce VGlut1 synapse formation onto cortical neurons, with or without thalamic neurons also present (n = 30 cells per condition; p > 0.05, Student's t test between GM and hevin treatments). (D) Same as B, except VGlut2 appears as red and PSD95 in cyan. (E) Hevin strongly induces VGlut2/PSD95 synapse formation when cortical neurons are cultured together with thalamic neurons (n = 30 cells per condition; *p > 0.01, Student's t test between GM and hevin treatments). Upon hevin treatment VGlut2/PSD95 synapses are recruited heavily to neuronal soma and proximal dendrites. (F) Hevin protein was stereotactically injected directly into Layer II/III of hevin KO V1. When compared to vehicle-injected control, hevin-injected cortex had distinctly thickened VGlut2 staining throughout S/Z and upper II/III (asterisk), as well as a dense appearance of VGlut2+ axon tracks throughout II/III (arrow). Scale bar, 100 µm. (G) Hevin-injected V1 had more VGlut2/PSD95 co-localized puncta (arrowheads) than vehicle-injected controls. Scale bar, 5 µm. (H) Hevin injection significantly increased the number of VGlut2/PSD95 synapses in hevin KO cortex (n = 2 z-stacks per animal, 3 animals per treatment; p < 0.01, Student's t test).DOI: http://dx.doi.org/10.7554/eLife.04047.010