|Home | About | Journals | Submit | Contact Us | Français|
Hereditary hypotrichosis simplex (HHS) is a rare autosomal dominant form of hair loss characterized by hair follicle (HF) miniaturization1, 2. Using genetic linkage analysis, we mapped a novel locus for HHS to chromosome 18p11.22, and identified a mutation (L9R) in the APCDD1 gene in three families. We show that APCDD1 is a membrane-bound glycoprotein that is abundantly expressed in human HFs, and can interact in vitro with WNT3A and LRP5, two essential components of Wnt signaling. Functional studies revealed that APCDD1 inhibits Wnt signaling in a cell-autonomous manner and functions upstream of β-catenin. Moreover, APCDD1 represses activation of Wnt reporters and target genes, and inhibits the biological effects of Wnt signaling during both the generation of neurons from progenitors in the developing chick nervous system, and axis specification in Xenopus embryos. The mutation L9R is located in the signal peptide of APCDD1, and perturbs its translational processing from ER to the plasma membrane. L9R-APCDD1 likely functions in a dominant-negative manner to inhibit the stability and membrane localization of the wild-type protein. These findings describe a novel inhibitor of the Wnt signaling pathway with an essential role in human hair growth. Since APCDD1 is expressed in a broad repertoire of cell types3, our findings suggest that APCDD1 may regulate a diversity of biological processes controlled by Wnt signaling.
Hair follicle (HF) miniaturization is a degenerative process that proportionally reduces the dimensions of the epithelial and mesenchymal compartments, and leads to the conversion of thick, terminal hair to fine, vellus hair4. HF miniaturization is most commonly observed in androgenetic alopecia, but is also characteristic of a rare, autosomal dominant form of hair loss, known as hereditary hypotrichosis simplex1 (HHS; OMIM 146520) (Supplementary Note). To gain insight into the molecular underpinnings of HF miniaturization and identify a gene underlying HHS, we performed a linkage study in two Pakistani families (HHS1 and HHS2) (Figs. 1a-d and S1a-l). After excluding the CDSN locus on chromosome 65, we used Affymetrix 10K SNP arrays for genotyping, and linkage analysis using a dominant model yielded a maximum LOD score of Z=4.6 on chromosome 18p11.22 (Fig. S1m). We narrowed the candidate interval to a 1.8 Mb region (Fig. 1e) containing 8 genes, 4 pseudogenes and 3 predicted transcripts (Fig. S1n). Direct sequencing identified a heterozygous mutation 26T>G (L9R) in the signal peptide of the Adenomatosis Polyposis Coli Down-regulated 1 (APCDD1) gene (Fig. S1o)6. The mutation L9R cosegregated with the disease phenotype in both families, and was absent in 200 unrelated, unaffected controls and in the SNP databases (Fig. S1p; data not shown). Unexpectedly, we identified the identical APCDD1 mutation in an Italian family with autosomal dominant HHS that had previously been mapped to the same region of chromosome 18p11.22 (Fig. S2)7, providing independent genetic evidence in support of this finding.
APCDD1 was abundantly expressed in both the epidermal and dermal compartments of the human HF, consistent with a role in HF miniaturization. APCDD1 mRNA and protein was present in human scalp skin by RT-PCR (Fig. S3a), and a western blot using an APCDD1 antibody (Fig. 1l). APCDD1 mRNA and protein were also highly expressed in the HF dermal papilla (DP), the matrix, and the hair shaft (Fig. 1f-j). Apcdd1 orthologs are conserved throughout vertebrate evolution (Fig. S4a,b), suggesting that a role in mouse3 and human HF growth emerged recently in mammalian species.
Several lines of evidence led us to postulate that APCDD1 may function as a negative regulator of Wnt signaling, including the observation that it is a direct target gene of Wnt/β-catenin 6; its similarity in expression pattern with another Wnt inhibitor, Wise8; the abundance of Wnt inhibitors in the HF9; and the conservation of 12 cysteine residues (Fig. S4a), a structural motif important for interaction between Wnt ligands and their receptors10,11.
To test if APCDD1 is an inhibitor of Wnt signaling, we first determined if APCDD1 interacts with ligands and receptors of the canonical Wnt pathway. No interaction was found with Fzd2, Fzd8, and Dkk4 (data not shown). In contrast, the extracellular domain of APCDD1 (APCDD1ΔTM) coprecipitated with recombinant tagged forms of Wnt3A and LRP5, two proteins important for HF induction 12 (Figs. 2a, S3b and S5), suggesting that APCDD1 can modulate the Wnt pathway via potential interactions with WNT3A and LRP5 at the cell surface. To determine the effect of APCDD1 on Wnt signaling, we performed TOP/FOP Flash Wnt reporter assays in HEK293T cells. Reporter activity induced by WNT3A alone, or in combination with LRP5/Fzd2, was downregulated ~2-fold by APCDD1 in a dose-dependent manner (Fig. 2b), indicating that APCDD1 inhibits the Wnt/β-catenin pathway.
To determine if APCDD1 can function as a Wnt inhibitor in vivo, we selected two systems in which the role of Wnt/β-catenin pathway has been well-defined: neuronal specification in the developing spinal cord13–15, and axis determination in the frog16,17. In the chick spinal cord, a Wnt/β-catenin gradient promotes proliferation of neural progenitors and generation of some neuronal classes13–15. Transfection of the Wnt reporter TOP::eGFP in the chick neural tube revealed strong activation of the pathway in the dorsal and intermediate progenitors, as previously shown15. However, overexpression of APCDD1 strongly reduced eGFP expression levels (Fig. S6a-d), decreased by ~20–30% the number of Sox3+ neural progenitors, as well as various neuronal subtypes of dorsal and ventral origin (Figs. 3a-e and S7a-d). This effect was stronger with mouse Apcdd1, a closer ortholog of the chick protein (Figs. S8a-i and S9a-e). These findings are consistent with the hypothesis that APCDD1 functions as a Wnt inhibitor.
The maternal Wnt pathway is required for the formation of dorsal and anterior structures in early Xenopus embryos 18,19. Overexpression of APCDD1 in dorsal blastomeres (n=35) reduced the anterior structures, such as the eyes and cement gland, at the tadpole stage (Fig. 4a,b), consistent with maternal Wnt inhibition. APCDD1 also inhibited transcription of the Siamois (Sia) reporter gene (Fig. 2c), activated by the maternal Wnt pathway20. A zygotic Wnt pathway is subsequently activated on the ventral side of the embryo21, and its inhibition produces secondary axes with incomplete heads16,17. Ventral overexpression of APCDD1 induced secondary axes (n=43, 28% duplicated axes, Fig. 4c,d), consistent with an inhibitory effect on zygotic Wnt signaling. The inhibition of Wnt activity by APCDD1 was also seen in transcription assays with Wnt8 RNA, but not β-catenin (Fig. 2c), indicating that it acts upstream of β-catenin.
We next investigated which domain of APCDD1 mediates its activity and in which cell APCDD1 exerts its function. First, western blot of APCDD1 expressed in HEK293T cells revealed that the protein is glycosylated and forms a dimer (Figs. 1l and S10a-c). Misexpression of mApcdd1ΔTM (lacking the transmembrane domain) in the chick neural tube mimicked the effects observed with mApcdd1 (Figs. S8j-r and S9f-j), suggesting that the Wnt inhibitory activity resides within the extracellular domain. Secondly, APCDD1 could affect either the signaling cell, by regulating Wnt secretion 22, or the receiving cell. In Xenopus transcription assays, Wnt8 RNA injected in one cell activated the Sia reporter in an adjacent cell. APCDD1 RNA inhibited transcription when coinjected with the Sia reporter, but not with Wnt8 (Fig. 4h), suggesting that APCDD1 inhibits Wnt signaling cell-autonomously in the receiving cell. Finally, since Wt-APCDD1 contains a transmembrane domain (Fig. 1k), and was localized to the plasma membrane (Fig. 2h and Fig. S11a,c,f,i), we tested whether APCDD1 undergoes cleavage to generate a diffusible inhibitor (APCDD1ΔTM), however, it was undetectable in the medium of transfected cells (Fig. S10d). Collectively, these data reveal that APCDD1 is likely a membrane-tethered Wnt inhibitor that acts as a dimer at the surface of the Wnt-receiving cell.
The L9R mutation disrupts the hydrophobic core of the signal peptide critical for co-translational processing (Fig. S4b,c)23. We analyzed protein stability and localization by western blotting and immunofluorescence in two cell lines (HEK293T or Bend3.0) transfected with either wild-type (Wt) APCDD1 or two different mutant forms (pathogenic mutation, L9R, and conservative substitution, L9V). Two fragments (68 KDa and 130 KDa) were detected in lysates of the Wt- and the L9V-APCDD1-transfected cells, whereas only a faint 68 KDa fragment was detected in the L9R mutant (Fig. 2f). In addition, Wt- or L9V-APCDD1 protein was localized to the cell membrane, while L9R-APCDD1 was retained within the ER (Figs. 2h,i and S11a-j). Furthermore, unlike the Wt isoform, N-terminally GFP-tagged L9R-APCDD1 could not be cleaved to localize at the membrane (Fig. S11l-n). Finally, when the Wt- and L9R-APCDD1 were co-transfected either in cells or injected into Xenopus embryos, some Wt protein was degraded (Fig 2e,g), and the rest sequestered in the ER along with the L9R isoform (Figs. 2j,k and S11k,o-q). Therefore, the L9R mutation likely functions in a dominant-negative manner, to destabilize the Wt protein and prevent it from reaching the plasma membrane.
We next tested if the L9R mutation affects APCDD1 protein function in vivo. In the chick neural tube, expression of L9R-APCDD1 only weakly inhibited eGFP transcription from the Wnt reporter (Fig. S6e,f), and had no effect on Sox3+ neural progenitors and neuronal subtypes (Figs. 3f-j and S7e-h), in contrast to Wt- or L9V-APCDD1 (Fig. S7m-u). Moreover, L9R-APCDD1 was able to block Wt protein function in vivo when they were co-transfected (Figs. 3k-o and S7i-l), indicative of a dominant-negative effect. The same results were observed in Xenopus, where the inhibitory effect of Wt APCDD1 on Wnt8-induced transcription was blocked by coexpression of the L9R mutant (Fig. 2d).
We then determined the consequences of Xenopus APCDD1 (Xapcdd1) protein depletion on axis formation in Xenopus embryos. Xapcdd1 mRNA is expressed maternally throughout development, with the highest levels in animal (future ectoderm) and marginal (future mesoderm) cells of stage 10 embryos (Fig. S12). Depletion of Xapcdd1 protein with a specific translation-blocking MO oligonucleotide (Fig. S13) resulted in loss of anterior and dorsal structures (Fig. 4e and Table S1). This phenotype was rescued by either injection of MO-resistant 5’ mutant Xapcdd1 RNA, or by DNWnt8 RNAs (Fig. 4f,g and Table S1), which inhibit zygotic Wnt signaling24. Therefore, the loss-of-function phenotype is consistent with ectopic activation on the dorsal side of zygotic Wnt activity, and supports the notion that endogenous APCDD1 is a Wnt inhibitor.
In conclusion, we suggest that APCDD1 may prevent formation of the Wnt receptor complex (Fig. 4i) since it interacts in vitro with LRP5 and WNT3A. The L9R mutant is unable to repress Wnt-responsive genes, by trapping the Wt protein in the ER where it may undergo degradation (Fig 4i).
Our findings underscore the requirement for exquisitely controlled regulation of the Wnt signaling pathway in HF morphogenesis and cycling 25. It is known that forced activation of Wnt signaling exclusively in the epidermis leads to increased hair follicle density and tumors26,27. We postulate that in HHS, Wnt signaling is indirectly increased through loss of the inhibitory function of APCDD1 in both the epidermal and dermal compartments of the HF, although the lack of HHS scalp samples precluded us from verifying this assumption. This notion is supported by mice with targeted ablation of another Wnt inhibitor, Klotho, which exhibit a reduction in HF density due to indirect upregulation of Wnt signaling and a depletion of HF bulge stem cells28. Since APCDD1 is expressed in both epidermal HF cells as well as the dermal papilla, we postulate that the simultaneous deregulation of Wnt signaling in both compartments may lead to the proportional reduction in organ size of the HF, resulting in miniaturization.
Our study provides the first genetic evidence that mutations in a Wnt inhibitor result in hair loss in humans. APCDD1 may be implicated in polygenic HF disorders as well, since it resides within linkage intervals on chromosome 18 in families with androgenetic alopecia29 as well as alopecia areata30. Furthermore, since APCDD1 is expressed in a broad range of cell types3, our findings raise the possibility that APCDD1 is involved in other Wnt-regulated processes, such as morphogenesis, stem cell renewal, neural development and cancer.
Genome-wide SNP-based genotyping was performed using the Affymetrix Human Mapping 10K 2.0 Array. Quality control and data analysis was performed with Genespring GT (Agilent software). Briefly, SNPs that violated a Mendelian inheritance pattern were removed from the data set prior to analysis. Haplotypes were inferred from raw genotype data. By analyzing haplotypes rather than individual SNPs, type I error introduced by linkage disequilibrium between markers is mitigated. Finally, haplotypes were analyzed for linkage under the assumption of a fully penetrant disease gene with a frequency of 0.001 transmitted by a dominant mode of inheritance.
Using the genomic DNA of the family members, all exons and exon-intron boundaries of APCDD1 gene were amplified by PCR with the gene-specific primers (Table S2). The PCR products were directly sequenced in an ABI Prism 310 Automated Sequencer, using the ABI Prism Big Dye Terminator Cycle Sequencing Ready Reaction Kit (PE Applied Biosystems). The mutation 26T>G disrupts a DdeI restriction enzyme site, which was used to screen the family members and control individuals.
Full Methods and any associated references are available in the online version of the paper at www.nature.com/nature.
Informed consent was obtained from all subjects and approval for this study was provided by the Institutional Review Board of Columbia University. The study was conducted in adherence to the Declaration of Helsinki Principles. Peripheral blood samples were collected from the family members as well as unrelated healthy control individuals of Pakistani and European origin (200 individuals each). Genomic DNA was isolated from these samples using the PUREGENE DNA isolation kit (Gentra System).
Genomic DNA from members of two Pakistani families was amplified by PCR using Platinum® PCR SuperMix (Invitrogen) and primers for microsatellite markers on chromosome 18p11. The amplified products were analyzed on 8% polyacrylamide gels.
Exon 1 and the adjacent boundary sequences of the APCDD1 gene were amplified using Platinum® Taq DNA Polymerase High Fidelity (Invitrogen). Due to the high G/C content, DMSO (final 5%) and MgSO4 (final 1.6 mM) were added to the PCR reaction. Other exons, as well as the exon-intron boundaries of the APCDD1 gene, were amplified using Platinum® PCR SuperMix (Invitrogen). Primer sequences are shown in Table S2.
In order to screen for the mutation 26T>G (L9R), a part of exon 1 and intron 1 of the APCDD1 gene was amplified by PCR using Platinum® Taq DNA Polymerase High Fidelity (Invitrogen) and the following primers: forward (5’- CCAGAGCAGGACTGGAAATG-3’), reverse (5’- CGCCAAGGGGACAGTGTAG-3’). The amplified PCR products, 191 bp in size, were digested with DdeI at 37°C overnight, and run on 2.0% agarose gels.
HEK293T (human embryonic kidney) and Bend3.0 cells were cultured in Dulbecco's modified Eagle's medium (DMEM; GIBCO) supplemented with 10% fetal bovine serum (FBS; GIBCO), 100 IU/ml penicillin, and 100 µg/ml streptomycin. For transfection experiments in HEK293T cells, dishes were coated with a coating medium containing 0.01 mg/ml of fibronectin (Sigma) and 0.03 mg/ml of type I collagen (Sigma) before seeding the cells in order to prevent detachment of the cells.
A mouse polyclonal anti-human APCDD1 antibody was purchased from Abnova Corporation. This antibody was raised against the full-length human APCDD1 protein. We performed epitope-mapping using three truncated GST-APCDD1 proteins (amino acid (aa) residues 1–171, 166–336, and 331–514), and confirmed that the epitope of the antibody exists between aa residues 166 and 336 of the human APCDD1, which corresponds to the middle portion of the extracellular domain (data not shown). This antibody recognized hair shaft and dermal papilla in human hair follicles (Fig. 1g-j), which finely overlapped with the signals detected by in situ hybridization (Fig. 1f). An affinity-purified rabbit polyclonal anti-mouse Apcdd1 antibody was produced by immunizing rabbits with the synthetic peptide, CQRPSDGSSPDRPEKRATSY (corresponding to the C-terminus of the extracellular domain of the mouse Apcdd1 protein, aa residues 441–459) conjugated to KLH (Pierce, Rockford, IL). This region is completely conserved among mouse and human APCDD1 proteins. The antibody was affinity-purified from the serum using the Sulfolink immobilization column (Pierce). This antibody strongly recognized human APCDD1 protein in western blots and immunofluorescence.
Total RNA were isolated from scalp skin and plucked scalp hairs of healthy control individuals using the RNeasy® Minikit (Quiagen). 2 µg of total RNA was reverse-transcribed with oligodT primers and SuperScript™ III (Invitrogen). The cDNAs were amplified by PCR using Platinum® PCR SuperMix and primer pairs for APCDD1, APCDD1L, keratin 15 (KRT15), LRP5, WNT3A, and β-2 microglobulin (B2M) genes (Table S2). Primers for the KRT15, LRP5, and WNT3A genes were designed as described previously31,32. PCR products were run on 1.5% agarose gels.
cDNA sequences for human APCDD1, WNT3A, CD40, and LRP5 were amplified by PCR using primers and templates shown in Table S2. The amplified products were subcloned into the mammalian expression vector pCXN2.133, a slightly modified version of pCXN234 with multiple cloning sites. The expression construct for the full-length human LRP5 was kindly provided by Dr. Patricia Ducy in Columbia University. To generate the expression construct for the mouse Frizzled 2 (mFzd2), the full-length open reading frame of the mFzd2 was purchased from Invitrogen (clone ID 6411627), which was subcloned into the NotI sites of the pCXN2.1 vector. In order to introduce a Flag-tag between amino acids 35 and 36 of the APCDD1 protein, N-terminal region of the APCDD1 was PCR-amplified using the forward primer (APCDD1-F-XhoI in Table S2) and a reverse primer (APCDD1-R-Flag-AvrII: 5’-AAAACCTAGGCTTATCGTCGTCATCCTTGTAATCATGAGACCTGCTGTCTGGAT-3’), which was followed by digestion with restriction enzymes XhoI and AvrII. The C-terminal region of the APCDD1 and the truncated APCDD1 proteins with the C-terminal HA-tag was obtained through digestion of the pCXN2.1-Wt-APCDD1-HA and the pCXN2.1-APCDD1-ΔTM-HA constructs with restriction enzymes AvrII and NheI. These two fragments were ligated with AvrII site, and subsequently subcloned into the XhoI and NheI sites of the pCXN2.1 vector. In order to generate expression constructs for N-terminal GFP-tagged APCDD1 protein, the coding region of the APCDD1 and the rabbit b-globin 3’-flanking sequences were cut out from the pCXN2.1-APCDD1 constructs with restriction enzymes XhoI and BamHI, and subcloned in frame into the pEGFP-C1 vector (Clontech). The templates were also subcloned into the XhoI and BamHI sites of pBluescript-SK (−) vector (Stratagene). pGEM Wnt8 (from R. Harland, U.C. Berkley), the Sia luciferase reporter gene (from D. Kimmelman, U. Washington), and pSP36 β-catenin (from B. M. Gumbiner, U. Virginia) have been previously described.
To generate a Xenopus expression vector for Xenopus APCDD1, we used a full length cDNA clone (BC080377, from Open Biosystems) as template and amplified the open reading frame with the primers shown in Table S2. The PCR product was inserted as ClaI/SalI fragment in CS2+2XHA (A. Vonica), resulting in CS2+Xapcdd1 HA.
The full length mouse Apcdd1 cDNA was amplified by RT-PCR from brain endothelial cells using the First Strand Synthesis Kit and High Fidelity Amplification Kit (Roche Applied Science) with primers shown in Table S2, which was subcloned into pCR®II-TOPO (Invitrogen) and pCAGGS34 vectors for in vitro transcription and chick neural tube electroporations, respectively. The Apcdd1ΔTM isoform containing the extracellular domain of mouse Apcdd1 (aa 1–486) was amplified by PCR from the full length cDNA using primers shown in Table S2 and subcloned into pCAGGS vector for chick electroporation.
The full length WT-APCDD1, L9R APCDD1, L9V APCDD1, mouse Apcdd1 or mApcdd1ΔTM isoform were subcloned into the pCAGGS vector and transfected into the chick neural tube (stage 12–13) together with nGFP vector (pCIG) using in ovo electroporation as described35. The chick embryos were grown for 3–4 additional days in the 39 °C incubator, fixed with 4% PFA/0.1M phosphate buffer, washed and cryoprotected as described35 before being processed for in situ hybridization or immunofluorescence. For the Wnt reporter assays, the TOP::eGFP reporter (M38 TOP::eGFP from Addgene) was transfected alone or in combination with Wt-APCDD1 or L9R-APCDD1. The chick embryos were grown for 12 hours in the 39 °C incubator, fixed with 4% PFA/0.1M phosphate buffer for 30 minutes, washed and cryoprotected as described35 before being processed for immunofluorescence.
Spinal cord Sox3+ progenitors, Isl1/2+ ventral motor neurons, Isl1/2+ dorsal interneurons and Chx10+ V2a interneurons were counted from 8 independent 12um thick sections of chick spinal cord from each transfected embryo. The nucleus stained with the transcription factor was considered one cell for this purpose. The cells were counted from both the electroporated side and the opposite control side. The plots were created using Sigma plot with values representing the mean for each embryo. Statistical significance was determined using the Student t-test.
HEK293T cells or Bend3.0 cells were plated in 60 mm dishes the day before transfection. Expression plasmids of APCDD1 were transfected with FuGENE® 6 (Roche Applied Science) at 60% confluency for HEK293 cells or Targefect_HUVEC for Bend3.0 cells. Total amount of transfected plasmids were adjusted with the empty pCXN2.1 vector. The cells were cultured 48 h after transfection in Opti-MEM (GIBCO). The cells were harvested and homogenized by sonication in homogenization buffer (25 mM HEPES-NaOH (pH 7.4), 10mM MgCl2, 250 mM sucrose, and 1X Complete Mini Protease Inhibitor Cocktail (Roche Applied Science)). The cell debris was removed by centrifugation at 3,000 rpm for 10 min at 4 °C, and the supernatant was collected as cell lysates. N-glycosidase (PNGase F) treatment and extraction of membrane fraction were performed as described previously33. The cultured medium with 1X Complete Mini Protease Inhibitor Cocktail was centrifuged at 1,500 rpm for 5 min at 4 °C. The supernatant was purified with 0.45 µm syringe filters (Thermo Fisher Science), and concentrated using Amicon Ultra-15 Centrifugal Filter Unit with Ultracel-10 Membrane (Millipore) according to the manufacturer’s recommendations. Total cell lysates from human scalp skin were extracted by homogenization in 50 mM Tris-HCl (pH 8.0), 150 mM NaCl, 1.0% NP40, 0.5% sodium deoxycholate, 0.1% SDS, and 1X Complete Mini Protease Inhibitor Cocktail. All samples were mixed with equal amount of Laemmli Sample Buffer (Bio-Rad Laboratories) containing 5% β-mercaptoethanol, boiled at 95 °C for 5 min, and analyzed by 10% SDS-polyacrylamide gel electrophoresis (SDS-PAGE). Western blots were performed as described previously36. The primary antibodies used were rabbit polyclonal anti-HA (diluted 1:4,000; Abcam), rabbit polyclonal anti-APCDD1 (1:20,000), mouse polyclonal anti-APCDD1 (1:1,000; Abnova), mouse monoclonal anti-Flag M2 (1:1,000; Sigma), and rabbit polyclonal anti-β-actin (1:10,000; Sigma).
HEK293T cells were seeded in 12 well dishes the day before transfection. Either 100 ng of TOPFlash (active) or FOPFlash (inactive) Wnt reporter vector (gifts from Dr. Patricia Ducy in Columbia University) was transfected into each well along with constructs for WNT3A (200 ng), Fzd2 (100 ng), LRP5 (100 ng), and/or wild type APCDD1-HA (300 ng or 800 ng) using Lipofectamine 2000 (Invitrogen). A construct for β–galactosidase reporter (100 ng) was also transfected for normalization of transfection efficiency. The cells were lysed 36 h after transfection and the signals were assayed as described previously9. The Wnt activity was measured based on the ratio of TOP/FOP luciferase activity. The results represent triplicate determination of a single experiment that is representative a total of five similar experiments.
Expression plasmids (total 4 µg) were transfected into HEK293T cells seeded on 60 mm dishes with FuGENE® 6 (Roche Applied Science) at 60% confluency. 24 h after the transfection, the cells were harvested and homogenized in lysis buffer (20 mM Tris-HCl (pH 7.5), 137 mM NaCl, 10% Glycerol, 2mM EDTA, 0.5% Triton X, and 1X Complete Mini Protease Inhibitor Cocktail). Total cell lysates were collected by centrifugation at 14,000 rpm for 15 min at 4 °C. The samples were incubated with either mouse monoclonal anti-Flag M2 agarose gel (Sigma) or mouse monoclonal anti-HA agarose gel (Sigma) for 3 h at 4 °C. The agarose beads were washed with lysis buffer for five times. The precipitated proteins were eluted with NuPAGE® LDS Sample Buffer containing Sample Reducing Agent (Invitrogen), incubated at 75 °C for 10 min, and separated on 10% NuPAGE® gels (Invitrogen). Western blots were performed using anti-HA (Abcam) and anti-Flag M2 antibodies (Sigma).
In order to express the GST fusion APCDD1 protein in bacteria, the extracellular domain of the human APCDD1 (aa residues 28–486) was PCR-amplified (Table S2), which was subcloned in-frame into the EcoRI and XhoI sites of pGEX-4T-3 vector (GE Healthcare Life Sciences). Expression of GST-fusion proteins was induced in DH5α (Invitrogen) by the addition of 0.1 mM isopropyl-β-D-thiogalactopyranoside at 37 °C for 3 h, and the fusion proteins were isolated from bacterial lysates by affinity chromatography with glutathione-Sepharose beads (GE Healthcare Life Sciences). LRP5-EC-Flag, WNT3A-HA, or CD40-EC-HA were overexpressed in HEK293T cells. GST pulldown assays were performed as described previously37. The antibodies used were: rabbit polyclonal anti-GST (1:3,000; Santa Cruz Biotechnology), anti-HA (Abcam) and anti-Flag M2 (Sigma).
A part of the human APCDD1 cDNA (GenBank Accession number, NM_153000: nt. 338–1899) was cloned into pCR®II-TOPO vector (Invitrogen). The antisense and sense DIG-labelled cRNA probes were synthesized from the linearized vectors with T7 and SP6 RNA polymerases (Roche Applied Science), respectively. In situ hybridization on dissected human hair follicles was performed following the methods described previously with minor modifications38. In situ hybridizations on chick spinal cord sections were performed as described39. The antisense mApcdd1 mRNA was generated using the In vitro transcription kit (Roche, Indianapolis, IN) with T7 RNA polymerase. The antisense chick Sim1 mRNA was generated using the T3 RNA polymerase.
IIF on cultured cells and fresh frozen sections of individually dissected hair follicles was performed as described previously36. IIF on HEK293T and Bend3.0 cells were performed 48 h after the APCDD1 expression constructs were transfected. For some stainings, cell membrane was labeled with rhodamine-phalloidin (Invitrogen). The primary antibodies used were mouse polyclonal anti-APCDD1 (diluted 1:1,000; Abnova), rabbit polyclonal anti-APCDD1 (1:4,000), rabbit polyclonal anti-pan-cadherin (1:200; Invitrogen), and goat polyclonal anti-calnexin (1:200; Santa Cruz Biotechnology). Immunofluorescence on chick spinal cord sections was performed as described40. The monoclonal antibodies against Nkx2.2, Pax6, Pax7, En-1 and Evx1 were purchased from DSHB (Iowa); rabbit anti Olig2 (Chemicon, Billerica MA), rabbit anti-Sox3 (provided by S. Wilson; University of Umea), rabbit anti Chx10, and guinea pig anti Isl1/2, sheep anti GFP (Biogenesis) and mouse anti β3-tubulin (Tuj1; Covance) were used as described40.
Based on the results of immunofluorescence with rabbit polyclonal anti-APCDD1 antibody in HEK293T cells transfected with APCDD1-expression constructs, we measured the subcellular localization of APCDD1 proteins. The cell outline was visualized using rhodamine-phalloidin (Invitrogen). Images were processed in Image J (http://rsbweb.nih.gov/ij/), splitting the channels. First, the outline of the cell was used to measure the signal within the whole cell. Second, scaling the cell frame down, the signal inside the cell was measured. For each cell, the following values were recorded: 1) the adjusted total signal in the cell (the level of fluorescence, relative to the background); 2) the adjusted signal inside the cell; 3) the adjusted signal in the membrane; and 4) the ratio between adjusted signal in the membrane and inside the cell. The adjusted signal (Sadj) was calculated by subtracting the background signal and then normalizing to the background (B) signal levels in an empty area of equal size within the same image. Sadj = (S - B)/B. For example, a reported signal Sadj = 5 indicates 5 times stronger than the background41–43. Data are represented as average ± 1 SEM (standard error of the mean). P values are reported using heteroscedastic 2-tailed t-tests, applying the Bonferroni correction to take into account the 3-way comparisons (WT versus L9R; WT versus WT + L9R; L9R versus WT + L9R). All reported values are measured in 20 cells per condition.
Xenopus laevis embryos were obtained by in vitro fertilization, cultured in 0.1X MMR and staged according to Nieuwkoop and Faber44. APCDD1 RNA was produced from the pBluescript-SK (−)-human APCDD1 constructs using the mMessage Machine in vitro T7 transcription kit (Ambion). For full length xapcdd1 RNA expression, the vector (pCMV-SPORT6) was restricted with XbaI and transcribed with the mMessage Machine in vitro SP6 transcription kit (Ambion). For Xapcdd1 HA expression, CS2+Xapcdd1 HA was restricted with NotI and transcribed with the same kit. RNA and reporter DNA injections were done at the 4 cell stage. For the effect of APCDD1 on antero-posterior patterning, APCDD1 RNA (1 ng) was injected in the marginal zone of both dorsal blastomeres at 4 cell stage. For the ventral effect of APCDD1, one ventral blastomere was injected in the marginal zone at the 4 cell stage45.
Translation blocking MO oligonucleotide (AS1 MO) was designed and synthesized by GeneTools, with the sequence: 5’-TGGTAGTTCAGCTCCAGAATGTCCT, where the nucleotide in bold is the first one in the open reading frame. The efficiency and specificity of the MO was tested on the full length mRNA (wt Xapcdd1 in Fig. S13) and the Xapcdd1 HA mRNA lacking the 5’ UTR to which AS1 MO binds (5’ mut Xapcdd1 in Fig. S13). RNA preincubation with MO and in vitro translation were performed as described previously46, using the Promega Reticulocyte Lysate System for translation in the presence of [S35]-Met. MO was injected at the 4 cell stage in both dorsal blastomeres (30 ng), alone or together with xapcdd1 RNA (300 pg) or DNWnt8 RNA (300 pg). Embryos were scored for the dorso-anterior index (DAI) at stage 41.
Injected embryos were collected at stage 9 and processed for luciferase assays (Promega) as described45. The Sia reporter gene21 was injected at 100 pg DNA. All assays were in triplicate, and each experiment was repeated three times.
Radioactive RT-PCR was performed as described previously47. mRNA was purified from whole embryos, or from fragments cut wit a hair knife, at the indicated stages with RNA-Bee (Tel-Test, Inc.), before reverse transcription with SuperScript III (Invitrogen), using Poly dT as priming oligonucleotide. The primers used for PCR were ODC: sense: 5′ CGAAGGCTAAAGTTGCAG 3′, antisense: 5′-AATGGATTTCAGAGACCA-3′; goosecoid (gsc): sense: 5’-TCTTATTCCAGAGGAACC-3’, antisense: 5’-ACAACTGGAAGCACTGGA-3’; Xapcdd1 sense: 5’-CTGGAGCTGAACTACCATGG-3’, antisense: 5’-TGACCCTCGATGTTTGGAGGC-3’.
Xenopus embryos were injected at the 4 cell stage with 1 ng RNA of wt human APCDD1 alone, or together with L9R mutant RNA (1 ng). Injections also contained 1 ng LacZ RNA as loading control. Embryos were retrieved at stage 10, homogenized in NP-40 extract buffer, mixed with LDS sample buffer and run on NuPage 4–12% gels (Invitrogen). After transfer to PVDF membrane, blots were incubated with rabbit polyclonal anti-APCDD1 (1:10,000) or anti-β-Galactosidase (1:1000; ProSci Inc.) antibodies, and stained with ECL Western Blotting Reagent (GE Healthcare).
We are grateful to the family members for their participation in this study, and to Helen Lam and Ming Zhang for technical assistance. We appreciate the collaboration and helpful discussions with Dr. Robert M. Bernstein and members of the Christiano laboratory at Columbia University. We thank Drs. Satoshi Ishii (Tokyo University, Japan) and Junichi Miyazaki (Osaka University, Japan) for the supplying pCXN2.1 vector. We thank Drs. Colin Jahoda, Andrew Tomlinson, Adrian Salic, Cassandra Extavour, Richard Vallee, Gilbert Di Paolo and Gerard Karsenty for stimulating discussions and comments on the manuscript, and Dr. Patricia Ducy for generously sharing reagents. This work was supported in part by USPHS NIH grant R01AR44924 from NIH/NIAMS (to A.M.C.). Y.S. is supported by a Research Career Development Award from the Dermatology Foundation. The work in Dr. Ben A. Barres’ laboratory (D.A. and B.A.B) was supported by grants from the Myelin Repair Foundation and the National Multiple Sclerosis Society (grant RG 3936A7/1). The work in Dr. Ali H. Brivanlou’s laboratory (A.V. and A.H.B.) was supported by NIH grants R01 HD032105 (to A.H.B.) and R03HD057334 (to A.V.).
Supplementary Information is linked to the online version of the paper at www.nature.com/nature.
Accession numbers. APCDD1 mRNA NM_153000, protein NP_694545.
Author Contributions. The study was conceived, designed and supervised by A.M.C.; laboratory work, phenotyping, and sample ascertainment were performed by Y.S., D.A., A.V., V.L., M.W. and A.B.; statistical analyses were performed by L.P.; different aspects of clinical genetics, phenotyping, and mutation screening assays were made by M.W., Y.S., A.B., S.B., A.S., and A.M.C.; and Y.S., D.A., A.V., V.L., A.H.B., B.A.B. and A.M.C. had significant input into experimental design and contributed to the preparation and editing of the manuscript.
Author Information. Reprints and permissions information is available at www.nature.com/reprints. The authors declare no competing financial interests.