Figure 1.
Generation of NOSIP knockout mice.
(A) Schematic presentation of the gene trap strategy. The NOSIP gene is located on mouse chromosome 7 and consists of 9 exons. LTR-mediated insertion of a trapping cassette flanked by a splice acceptor (SA) and splice donor (SD) site into the second intron leads to functional disruption of the NOSIP gene. The trapping cassette comprises the neomycin resistance gene (NeoR), polyadenylation signal (pA), phosphoglycerolkinase promotor (PGK) and an unrelated sequence from Burton’s tyrosine kinase (select), used for identification of the site of insertion. (B) Genotyping is carried out by PCR with two gene-specific primers (500 bp WT fragment) and one cassette-specific primer (200 bp KO fragment) as indicated in A. (C) Absence of NOSIP protein from heads of NOSIP KO embryos is confirmed by immunoblotting with a NOSIP-specific antiserum. Immunoblot using a vinculin-specific antibody is shown as loading control. (D) Immunohistochemistry staining with a NOSIP-specific antiserum (middle and right) or without primary antibody for control purposes (left) on coronal head sections of E12.5 embryos of the indicated genotype showing expression of NOSIP in the telencephalon (lv = lateral ventricle).(E) Significant weight reduction in NOSIP knockout embryos at E13.5 (left) and E18.5 (right). Statistical significance was analysed by One-way ANOVA with Bonferroni post-test, 95% confidence interval. (B–E) Mice were backcrossed to C57Bl/6J for ten generations.
Figure 2.
NOSIP KO mice show craniofacial anomalies.
(A–E) Comparison of embryos at E15; D and E are frontal and lateral images of the same embryo. Bars in A and B indicating the distance between the eyes. (F–H) Comparison of embryos at E13.5; G and H are lateral and frontal view of the same embryo. Arrow in H points to the single eye. (I/J) Comparison of HE-stained coronal sections through the eye at E15.5. Site of missing lens indicated by asterisk in J. It should be noted, that the liver, as the site of Epo-dependent erythropoiesis, appeared smaller and paler in NOSIP KO embryos in comparison to the WT littermates (G), indicating a defect in embryonic erythropoiesis. This is in accordance with the previously proposed role of NOSIP in Epo signal transduction [14]. el (eye lid), ff (fetal fissure), inl (inner neuroblastic layer), l (lense), nfl (nerve fibre layer), onl (outer neuroblastic layer), rpe (retinal pigmented epithelium). (A–J) Mice were backcrossed to C57Bl/6J for ten generations.
Figure 3.
NOSIP KO mice display holoprosencephaly and clefting of face and palate.
(A–D) Comparison of embryos at E18.5; A and C are lateral and frontal view of the same embryo. (E–H) HE-stained coronal head sections of embryos at E15.5 of the same phenotype category as shown in A–D; asterisk in F indicates cleft palate, arrow in H points to the ectopic eye, consisting of pigment and neural layer, but no lens. (I–L) HE-stained longitudinal sections of embryos at E14.5 with equal grade of phenotype as shown in A–D; cleft indicated by asterisk in J. (M/N) Comparison of SEM images of the palate of embryos shown in A/B at E18.5. (O/P) Comparison of isolated brains at E13.5, dorsal view, frontal side up. (Q/R) Comparison of brain and olfactory bulbs at E18.5. (S/T) HE-stained coronal sections of embryos at E12.5; arrow pointing to the site of the missing eye in T. All embryos shown for comparison are littermates from NOSIP HET x HET intercrosses. 3v (third ventricle), bb (basisphenoid bone), c (midfacial cartilage), d (diencephalon), dm (dorsal midline), e (eye), ftv (fused telencephalic ventricle), ll (lower lip), lge (lateral ganglionic eminence), lv (lateral ventricle), mge (medial ganglionic eminence), np (nasopharynx), ob (olfactory bulb), oe (olfactory epithelium), p (palate/palatal shelf), t (tongue), tel (telencephalon), ul (upper lip). Mice were backcrossed to C57Bl/6J for six (E–H) or ten generations (all other panels).
Figure 4.
PP2A is a novel interacting protein of NOSIP.
(A) PP2A subunits found to interact with NOSIP in an affinity chromatography approach followed by mass spectrometry analysis. (B) GST-NOSIP was used for GST-pulldown experiments using HeLa cell lysates endogenously expressing PP2A. 0.5% of the volume of the cell lysate used for pulldown is shown for comparison of protein levels (0.5% input). Proteins are detected by immunoblotting with PP2A subunit-specific antibodies. (C) Co-immunoprecipitation of endogenous PP2A subunits with endogenous NOSIP from HeLa cell lysates using a polyclonal NOSIP-specific antiserum for immunoprecipitation (IP). Precipitation with sepharose beads only served as specificity control. 0.2–0.8% of the volume of the cell lysate used for IP is shown for comparison of protein levels (% input). Proteins are detected by immunoblotting with PP2A subunit-specific antibodies and a NOSIP-specific antiserum.
Figure 5.
NOSIP monoubiquitinates the catalytic subunit of PP2A.
(A) In vitro ubiquitination assay. An assay without E1 enzyme served as negative control. Ubiquitinated PP2A is detected with α-PP2Ac and α-Flag (Flag-tag of ubiquitin), IB with α-NOSIP served as control. Arrowheads indicate monoubiquitinated PP2Ac; * heavy chain of the antibody; ** originates from the commercial Flag-ubiquitin; the upper bands detected in both lanes are considered unspecific reactivity of the Flag-antibody. (B) In vivo ubiquitination assay with rescue of PP2Ac monoubiquitination by the expression of NOSIP. MEFs of the indicated genotype were infected with Flag-ubiquitin (Flag-ubi) plus NOSIP or empty vector (control). Immunoprecipitation (IP) was performed with α-PP2Ac and proteins were detected by IB with α-Flag (Flag-tag of ubiquitin). IB using α-PP2Ac, α-NOSIP and α-vinculin served as controls. Arrowhead indicates monoubiquitinated PP2Ac; * heavy chain of the antibody (C) In vivo ubiquitination assay with proteasomal inhibition. WT MEFs were infected with Flag-ubiquitin (Flag-ubi) and either left untreated (−) or pre-treated with 20 µM MG-132 for 5 h (MG) and monoubiquitinated PP2Ac was detected as described above. IB using α-ubiquitin served as control for proteasome inhibition. Arrowhead indicates monoubiquitinated PP2Ac; * heavy chain of the antibody. (D) Protein levels of different PP2A subunits in MEFs from WT, HET and NOSIP KO mice were analysed by IB with PP2A subunit-specific antibodies and α-NOSIP. IB using α-GAPDH is shown as loading control. IB using α-ubiquitin served as control for proteasome inhibition. MEFs were left untreated (−), treated with 20 µM MG-132 for 5 h (MG) or treated with 100 nM okadaic acid for 3 h (OA).
Figure 6.
PP2A activity is increased in the absence of NOSIP.
(A) Protein levels in MEFs from WT and NOSIP KO mice were analysed by IB with antibodies against PP2Ac, the methylated form of PP2Ac (α-PP2Ac-m), LCMT, PME, PTPA, alpha4 and NOSIP. IB using α-vinculin is shown as loading control. (B) PP2A activity was measured in MEFs isolated from WT, HET and NOSIP KO mice. Values are means ± SEM, n = 6; statistical significance was analysed by One-way ANOVA with Bonferroni post-test, 95% confidence interval (top). Following the activity measurement all samples were subjected to IB with α-PP2Ac to determine the amount of immunoprecipitated PP2A and the activity was normalised to relative intensity of α-PP2Ac IB. One representative experiment is shown (bottom). (C) PP2A activity was measured in lung, palate and midfacial tissue (excluding palate, maxilla and mandible) isolated from WT and NOSIP KO embryos at E18.5 backcrossed to C57Bl/6J for ten generations. Values are means ± SEM, n = 7; statistical significance was analysed by Wilcoxon matched pairs test, two tailed, 95% confidence interval (top). Following the activity measurement all samples were subjected to IB with α-PP2Ac to determine the amount of immunoprecipitated PP2A in the activity assay and one representative experiment is shown (bottom). Activity data are normalised to relative intensity of α-PP2Ac IB.