Methods

Cell Culture and Reagents
HeLa cells and human embryonic kidney (HEK) 293 cells were maintained in DMEM (Invitrogen, Carlsbad, CA) containing 10% FBS (Japan Bioserum, Hiroshima, Japan). HUVECs (Lonza, Basel, Switzerland) were maintained in endothelial growth medium-2 (Lonza). iPSECs were established and maintained in primate ES medium (ReproCELL, Kanagawa, Japan), supplemented with 500 U/mL of penicillin/streptomycin (Invitrogen) and 4 ng/mL of recombinant human basic fibroblast growth factor (WAKO, Osaka, Japan), as previously reported.14,19,20 For use and establishment of iPS cells, the institutional ethics committees of Kyoto University reviewed and approved the present study protocols. All participants gave written informed consent; for those considered too young to consent, informed consent was given by the parent or guardian.
Recombinant human IL-1β, TGF-β, VEGF, PDGF-BB, IFN-β, and IFN-γ were obtained from PeproTech (Rocky Hill, NJ). IFN-α was obtained from BioVision (Milpitas, CA). Reagents were dissolved in 0.1% BSA (GIBCO, Carlsbad, CA) at 0 ng/mL.
For evaluation of cellular proliferation, HUVECs were seeded at 7.5×104 cells/well in 6-well plates. At 24 hours after seeding, the culture medium including the test reagent was exchanged. The numbers of viable cells were assessed and counted using trypan blue (Nacalai Tesque, Kyoto, Japan) exclusion.

Evaluation of Angiogenic Activity by Tube Formation and Migration Assays
Tube formation was assessed as described previously.14,21 HUVECs (30 000 cells/well; 0 or 1 ng/mL of IFN-β) were seeded onto 96-well plates coated with Matrigel (BD Biosciences, San Jose, CA). After incubation for 20 hours at 37°C, digital images of the tubes that formed were captured. For quantification, the area of the tube, total length of the tube, and number of tube branches were calculated using ImageJ software (National Institutes of Health, Bethesda, MD).
Migration assays were performed as described previously.22 HUVECs were grown to overconfluence in 24-well plates and then incubated overnight for serum starvation (0 or 1 ng/mL of IFN-β). After creation of a wound with a p200 pipette tip, the medium was replaced with endothelial growth medium-2 (0 or 1 ng/mL of IFN-β). The wound was allowed to narrow for healing (re-endothelialization) for 8 hours. Digital images were obtained before and after the incubation period, and the area of re-endothelialization was calculated using ImageJ software.

RNA Interference
Transfection of siRNAs was conducted using the Amaxa Nucleofector Device (Lonza), by following the manufacturer’s recommendations, as previously reported.14 We purchased and used RNF213 siRNA (sc-94184; Santa Cruz Biotechnology, Santa Cruz, CA), signal transducer and activator of transcription (STAT)1 siRNA (sc-44123; Santa Cruz Biotechnology), and control siRNA-A (sc-37007; Santa Cruz Biotechnology). Western blotting assays were conducted to monitor knockdown of gene expression.

Plasmid Construction for RNF213 WT, RNF213 R4810K, RNF213 WEQ, and RNF213 Deletion of AAA+ and Transfection
To obtain RNF213 cDNA, reverse-transcriptase polymerase chain reaction (RT-PCR) was performed using 3 sets of primers (Primer N1 to N2, M1 to M2, and C1 to C2), as shown in Table S1. We used RevTra Ace reverse transcriptase and KOD FX DNA polymerase (TOYOBO, Osaka, Japan). The 3 amplified fragments were digested with restriction enzymes (NcoI and SalI, SalI and HindIII, and HindIII and NotI, respectively). Fragments were cloned between NcoI and NotI sites of pENTR4 (Thermo Scientific, Waltham, MA) to construct full-length cDNA of RNF213. The R4810K mutation and Walker B mutation (E2488Q:WEQ) were introduced by PCR-based, site-directed mutagenesis using mutated primers using Pfu Turbo DNA polymerase (Agilent Technologies, Santa Clara, CA). These mutated primers (Primers 1 to 4) were shown in Table S1. The deleted mutation (delta AAA) was introduced by the SalI site at the 8310-nucleotide position from the start codon (primer delatSalI: AAT GTC GAC GTG ATC ACA GAA GTC CTC TGC and primer M2) by PCR using KOD FX DNA polymerase and combined with RNF213 cDNA. These entry clones were used for the LR reaction (Thermo Scientific) using a destination vector with an amino-terminal 3xFLAG tagged or enhanced green fluorescent protein (EGFP) sequence under a tetracycline-regulated cytomegalovirus promoter. The design of the tagged-RNF213 WT and mutant vectors that were used in the present study are shown in Figure1. The generated constructs were confirmed by sequencing. The plasmids were transfected into cells using an Amaxa Nucleofector Device (Lonza), following the manufacturer’s recommendations, as previously reported.14
Figure 1  Design of tag protein (FLAG or EGFP)-RNF213 WT, RNF213 R4810K, RNF213 WEQ, and RNF213 ΔAAA vector constructs. EGFP indicates enhanced green fluorescent protein; WEQ, Walker B motif; WT, wild type.

Construction of the Promoter, Transfection, and the Luciferase Assay
A DNA fragment of the human RNF213 gene promoter, including −3000 to +200 base pairs (bp) from the transcription start site (chr17: 75 846 262 to 75 849 462 from NCBI 36/hg18), was synthesized by Takara Bio Inc (Shiga, Japan). This fragment was inserted between the Kpn I and Nhe I sites in front of the luciferase reporter gene in the pGL 4.14 Luc/Hygro vector (Promega, Madison, WI) and used as the RNF213 WT promoter pGL4.14. The generated construct was confirmed by sequencing. A reporter plasmid with a mutation in the STATx-binding site (potential STAT1-binding) at −514/−505 (5′-TGCCGGGGG-3′, mutated positions are underlined), including the Not I and Sph I sites in the RNF213 promoter, was generated by GenScript Corp (Piscataway, NJ). The fragment was inserted between the Not I and Sph I sites in the RNF213 WT promoter pGL4.14 plasmid construct and used as the RNF213 STATx mutation promoter pGL4.14.
The reporter and internal control vector pGL4.74 (Promega) plasmids were transfected into cells using the Amaxa Nucleofector Device (Lonza), by following the manufacturer’s recommendations, as previously reported.14 IFN-β was added after 24 hours, and cells were harvested at 48 hours post-transfection. Luciferase activity of cell lysates was measured using the Dual-luciferase Reporter Assay System (Promega). Results were normalized by pGL4.74 luciferase activity, and the obtained values were divided by the mean value for 0 ng/mL of IFN-β.

Biochemical Assays
Real-time quantitative PCR (qPCR) and western blotting were used as biochemical assays to assess mRNA and protein expression, respectively. qPCR was performed as described previously.8 Target cDNA expression levels were normalized to corresponding expression levels of PPIA. The primer pairs that were used were as described previously.8,14
Western blotting was performed as described previously.15 Lysates were subjected to immunoblotting using an anti-RNF213 antibody that we previously generated14 or anti-STAT1 (sc-346; Santa Cruz Biotechnology), anti-FLAG (NU01102; Nacalai Tesque), anti-phospho-STAT1 (Ser727; #8826; Cell Signaling Technology, Beverly, MA), and anti-β-tubulin (sc-9104; Santa Cruz Biotechnology) antibodies. Quantification was conducted using Scion Image software (Scion Corp, Frederick, MD).

ATPase Assay
RNF213 proteins fused with EGFP were expressed in HEK293 cells, and cells were extracted with RIPA buffer. The cell extract was clarified by high-speed centrifugation and used for immunoprecipitation with anti-GFP agarose (MBL Japan, Nagoya, Japan). Immunoprecipitants were washed with RIPA buffer twice and with RIPA buffer without SDS once, and finally equilibrated with 0.5× kinase buffer (10 mmol/L of Tris-HCl [pH 7.6], 100 mmol/L of KCl, 5 mmol/L of MgCl2, and 0.025% TritonX-100). Immunoprecipitants were resuspended into 50 μL of kinase buffer, and half of this volume was subjected to SDS-PAGE and stained with GelCode staining (Thermo Scientific). To perform ATPase reactions, the indicated volume of immunoprecipitants was combined with prewarmed 1× kinase buffer with 60 μmol/L of ATP. The 50-μL ATPase reaction proceeded for 30 minutes at room temperature by adding a final concentration of Phosphate Sensor (0.5 μmol/L; Thermo Scientific). The plate was mixed and immediately read on a microplate reader at an excitation of 430 nm and emission of 450 nm.

Vascular EC- or SMC-Specific Rnf213 Tg Mouse Production
The transgene construct consisted of the following components: cysteine-adenine-guanine (CAG) promoter–LoxP–PGK promoter–Neo–3 SV40 poly(A) sequences–LoxP–mouse Rnf213 (WT or R4757K mutant) coding sequence–beta globin polyA (Figure2). The construct was generated using Gateway technology (Invitrogen). An entry vector harboring mouse Rnf213 WT was produced based on the pENTR4 vector. Five fragments of the Rnf213 coding sequence were amplified by RT-PCR from C57BL/6 mouse liver RNA using the SuperScript III One-Step RT-PCR System with Platinum Taq DNA Polymerase (Invitrogen) using primers, which are shown in Table S2. Fragments 1 and 2 were connected using the XhoI site at 3259 to 3264 bp. Fragments 2 and 3 were connected using the EcoRI site at 6080 to 6085 bp. Fragments 3 and 4 were connected using the SacII site at 8975 to 8980 bp. Fragments 4 and 5 were connected using the MluI site at 12 651 to 12 656 bp. The full-length Rnf213 coding sequence was integrated into the pENTR4 vector using NcoI and NotI sites. An allelic ortholog of human p.R4810K (p.R4757K) was introduced into Rnf213 WT using a site-directed mutagenesis kit (Invitrogen). A donation vector was generated based on the PGKneotpAlox2 vector, containing the LoxP–PGK promoter–Neo-3 SV40 poly(A)–LoxP cassette. The CAG promoter and beta-globin polyA were amplified using the pCAGGS vector as a template. The Gateway recombination cassette (attR1-ccdB-attR2) was amplified using the pDEST vector as a template. The CAG promoter was cloned into the PGKneotpAlox2 vector, upstream of the first LoxP using SacI and NotI sites. The AttR1-ccdB-attR2 cassette and beta-globin polyA were introduced into the PGKneotpAlox2 vector, downstream of the second LoxP using NheI and XhoI sites and XhoI and KpnI sites, respectively. The entry and donation vectors were converted by an LR plus clonase reaction to produce the transgene construct.
Figure 2  Schematic diagram of Rnf213 conditional expression in ECs or SMCs. CAG indicates cysteine-adenine-guanine; ECs, endothelial cells; SMCs, smooth muscle cells; WT, wild type.   The transgene constructs were digested with PvuI and KpnI, and a DNA fragment of 20 kb was purified and then microinjected into fertilized C57BL/6 mouse eggs to generate Tg mice. Genotypes of Tg offspring were determined by PCR using the primers shown in Table S3. To obtain mice harboring vascular ECs or SMCs overexpressing Rnf213, Tg founders were bred with mice expressing a Cre transgene driven by either the Tie2 kinase promoter/enhancer (Tek; strain name: B6.Cg-Tg(Tek-cre)12Flv/J) or the smooth muscle protein 22-alpha promoter (strain name: B6.Cg-Tg(Tagln-cre)1Her/J; The Jackson Laboratory, Bar Harbor, ME). Specific expression in ECs or SMCs was confirmed by western blotting. ECs were purified from lungs by magnetic cell sorting using anti-CD31/PECAM-1 antibody (Life Technology, Grand Island, NY). Aorta was used as an SMC source.

Exposure to Hypoxia and Evaluation of Cerebral Angiogenesis
Hypoxia experiments were performed in 5 groups of 3-week-old mice: (1) vascular EC-specific Rnf213 R4757K Tg mice (EC-Mut Tg); (2) vascular EC-specific WT Rnf213 Tg mice (EC-WT Tg); (3) vascular SMC-specific Rnf213 R4757K Tg mice (SMC-Mut Tg); (4) Rnf213 knockout (KO) mice17; and (5) WT mice. Each group was composed of mice with hypoxia (n=6) and with normoxia (n=6). Care of animals and all experimental procedures were in accord with the Animal Welfare Guidelines of Kyoto University (Kyoto, Japan), and animal protocols were reviewed and approved by the animal care, use and ethics committee at Kyoto University. Mice that were exposed to hypoxia were placed in an 8% oxygen chamber with nitrogen-balanced gas under normal-pressure atmosphere (Kyodo International, Kanagawa, Japan) for 2 weeks.
Hypoxia-induced cerebral angiogenesis was evaluated by staining for the blood–brain barrier using glucose transporter (GLUT)-1 immunohistochemistry.23 After hypoxic exposure, mice were anesthetized and perfused with PBS with 1 U/mL of heparin. Brains were removed, fixed in 10% formaldehyde, embedded in paraffin, and sectioned. Sections were immunostained with mouse anti-Glut1 antibody (Abcam, Cambridge, MA). We captured images at ×200 magnification in the cerebral cortex from each of the 2 sections in 6 mice per genotype in hypoxia or normoxia and counted Glut1-positive capillaries using ImageJ software.

Magnetic Resonance Imaging
Magnetic resonance angiography (MRA) was performed using a 7-T Bruker MRI (Bruker Biospin, Rheinstetten, Germany). Mice were anesthetized by inhalation of 3% isoflurane in room air, and the respiratory rate was continuously monitored. Core temperature was maintained at 30±2°C by a flow of warm air. A 3-dimensional (3D) gradient-echo sequence was used to acquire MRA images with the following parameters: field of view (FOV)=19.2×12.8×9.6 mm; matrix=192×128×96; repetition time/echo time (TR/TE)=120/4.3 ms; flip angle=60 degrees, and scan time=1 hour, 39 minutes.
To detect infarction, T2-weighted images were also acquired using a rapid acquisition with relaxation enhancement sequence with the following parameters: TE/TR=5000/15 ms; effective TE=60 ms; echo train length=8; FOV=19.2×19.2 mm2; slice thickness=0.6 mm; matrix=256×256; and the number of excitations=4.

Statistical Analysis
Results are presented as mean±SD. Number of samples are provided in the figure legends. Statistical tests on in vitro experiments were performed using the Student t test to detect the effect of treatments by comparing controls according to study designs shown below.
We first screened several angiogenic and antiangiogenic cytokines to determine whether treatment with different doses of these cytokines induced RNF213 in cultured cells. IFN-β and IFN-γ were found to induce RNF213. We then statistically compared RNF213 levels in IFN-β- or IFN-γ-treated cells with untreated cells. The mechanism of RNF213 induction with IFN-β was examined in cells transfected with RNF213 WT or mutated promoter luciferase plasmid. To test the effect of a STATx mutation, luciferase activity in cells transfected with STATx mutation promoter was compared with cells transfected with RNF213 WT promoter. Effects of treatment of IFN-β on angiogenesis were statistically evaluated by tube formation and migration assay using HUVECs and iPSECs. In experiments of target protein depletion (p-STAT1, STAT1, or RNF213) using corresponding siRNAs, we compared target protein levels in cells treated with target siRNA with cells treated with control siRNA. We also tested the effect of STAT1 and RNF213 siRNA on angiogenesis.
To evaluate the in vitro effect of various RNF213 mutations on angiogenesis, the angiogenic function of HUVECs transfected with RNF213 mutants were compared with HUVECs transfected with a control vector. In addition, ATPase activities of RNF213 mutants were compared to RNF213 WT.
For in vivo animal studies, both nonparametric methods (Kruskal–Wallis 1-way ANOVA followed by Mann–Whitney U test) and a parametric method (2-way ANOVA method) were conducted to detect effects of treatment (hypoxia), genotype, and interaction on cerebral angiogenesis. Cerebral angiogenesis was evaluated by the increase in the numbers of cerebral microvessels/mm2.
Values of P<0.05 were considered statistically significant. Statistical analyses were performed using SAS software (version 9.4; SAS Institute Inc., Cary, NC).