Department of Medicine, Stanford University School of Medicine, Stanford, California, USA
Address reprint requests to: David Feldman, MD, Stanford University School of Medicine, Division of Endocrinology, Gerontology and Metabolism, Stanford University Medical Center, Room S025, 300 Pasteur Drive, Stanford, CA 94305-5103, USA
A young Chilean boy with severe rickets was found to have hereditary vitamin D-resistant rickets without alopecia. He had a unique insertion/substitution mutation in the ligand-binding domain of the vitamin D receptor. The in-frame mutation disrupted ligand binding and co-activator binding and resulted in 1,25(OH)2D3 resistance.
Introduction: Hereditary vitamin D-resistant rickets (HVDRR) is a genetic disorder caused by mutations in the vitamin D receptor (VDR). In this study, we examined the VDR in a young boy who exhibited the typical clinical features of HVDRR but without alopecia.
Materials and Methods: The patient's VDR was studied using cultured dermal fibroblasts, and the recreated mutant VDR was analyzed in transfected cells.
Results: The patient's fibroblasts were resistant to 1,25-dihydroxyvitamin D [1,25(OH)2D3], exhibiting only a slight induction of 24-hydroxylase gene expression when treated with 1 μM 1,25(OH)2D3. [3H]1,25(OH)2D3 binding was absent in cell extracts from the patient's fibroblasts. Sequence analysis of the VDR gene uncovered a unique 5-bp deletion/8-bp insertion in exon 4. The mutation in helix H1 of the ligand-binding domain deletes two amino acids (H141 and T142) and inserts three amino acids (L141, W142, and A143). In transactivation assays, the recreated mutant VDR was 1000-fold less active than the wildtype (WT) VDR. In glutathione S-transferase (GST) pull-down assays, the mutant VDR bound GST-retinoid X receptor (RXR) weakly in the absence of 1,25(OH)2D3; however, the binding did not increase with increasing concentrations of ligand. The mutant VDR did not bind to GST-vitamin D receptor interacting protein (DRIP) 205 at concentrations up to 1 μM 1,25(OH)2D3. We also examined effects of the three individual mutations on VDR transactivation. Only the insertion of A143 into the WT VDR disrupted VDR transactivation to the same extent observed with the natural mutation.
Conclusion: We describe a novel insertion/substitution mutation in helix H1 of the VDR ligand-binding domain (LBD) that abolishes ligand binding and result in the syndrome of HVDRR. This is the first time an insertion/substitution has been found as the defect-causing HVDRR.
THE PLEIOTROPIC ACTIONS of 1,25-dihydroxyvitamin D [1,25(OH)2D3] are mediated by the vitamin D receptor (VDR), a member of the steroid/nuclear receptor superfamily of ligand-activated transcription factors.(1–3) The VDR is a member of the subfamily of receptors that requires heterodimerization with the retinoid X receptor (RXR) to activate gene transcription. The VDR is composed of an N-terminal DNA binding domain (DBD) and C-terminal ligand-binding domain (LBD). The DBD enables the VDR to interact with vitamin D response elements (VDRES) in promoters of target genes. The LBD binds 1,25(OH)2D3 and consists of 13 α-helices (H1, H2n, H2na, H3-H12) and 3 β-sheets.(4) Ligand binding elicits a series of molecular events leading to the activation of vitamin D responsive genes. Ligand binding also induces the formation of a hydrophobic groove on the receptor surface that involves helices H3 and H4 on one side and helix H12 on the other side. Studies have shown that the hydrophobic groove on the nuclear receptor surface serves as a binding site for a conserved LXXLL motif present in the nuclear receptor binding domain of co-activator proteins. The co-activator proteins include the p160 proteins (SRC1/NcoA1, TIF2/GRIP1/NcoA2/SRC2, pCIP/RAC3/ACTR/AIB1/SRC3) and the vitamin D receptor interacting protein (DRIP) complex. The co-activator proteins associate with the VDR in a ligand-dependent manner and enhance transactivation by modifying chromatin. They also link the VDR to the pre-initiation complex and RNA polymerase II.
Mutations in the VDR gene cause the rare genetic disorder hereditary vitamin D-resistant rickets (HVDRR), also known as vitamin D-dependent rickets type II (VDDR II).(2) Patients with HVDRR display a number of clinical features including early onset rickets, hypocalcemia, and secondary hyperparathyroidism. Many patients also have total body alopecia. The types of mutations identified in the VDR gene include missense mutations, nonsense mutations, splicing mutations, and a partial gene deletion.(2) Mutations in the DBD affect VDR-DNA interactions but not ligand binding, and they result in total loss of VDR transactivation.(5–9) Mutations in the VDR LBD have been shown to cause defects in ligand binding, RXR heterodimerization, and co-activator interaction, and they result in partial or total hormone unresponsiveness.(10–14) There is one report in which a patient with the classical signs of HVDRR including alopecia did not have a mutation in the VDR.(15) The cause of 1,25(OH)2D3 resistance in this case has been attributed to the abnormal expression of a hormone response element binding protein that inhibits VDR transactivation.(16) In this report, we describe the first insertion/substitution mutation in the VDR that results in the syndrome of HVDRR.
MATERIALS AND METHODS
Fibroblast cultures were established from skin biopsy specimens and grown in DMEM containing 10% iron-supplemented calf serum (Hyclone, Logan, UT, USA) and antibiotics as previously described.(17) Cells were incubated at 37°C under a 5% CO2 atmosphere.
1,25(OH)2D3 induction of 24-hydroxylase mRNA
Cultured fibroblasts were grown to confluence and treated with 1,25(OH)2D3 for 6 h in medium containing 1% FBS. Total RNA was prepared using TRIZOL reagent (Invitrogen, Carlsbad, CA, USA), and 5 μg of RNA was electrophoresed on 1% agarose gels, transferred to nylon filters, and immobilized by UV crosslinking. The filters were hybridized with cDNA probes for 24-hydroxylase and L7 ribosomal protein. The probes were labeled with Redivue [α32P]dCTP using the Rediprime DNA labeling system (Amersham, Piscataway, NJ, USA). L7 has been shown in multiple experiments to be unaffected by 1,25(OH)2D3 treatment, and therefore, serves as a control for loading and transfer efficiency.
Fibroblasts were cultured from a skin biopsy. The cells were resuspended in M-PER extraction buffer (Pierce Biotechnology, Rockford, IL, USA) containing 300 mM KCl, 5 mM dithiothreitol, and a complete protease inhibitor tablet (Roche Molecular Biochemicals, Indianapolis, IN, USA), and incubated at ambient temperature for 10 minutes on a rotating mixer. Cell debris was removed by centrifugation at 13,000 rpm for 15 minutes at 4°C. The crude cell extracts were incubated with [3H]1,25(OH)2D3 (Amersham) with or without 250-fold excess of radioinert 1,25(OH)2D3 as previously described.(17) Hydroxylapatite was used to separate bound and free hormone. Protein concentrations were determined by the Bradford method.(18)
Gene amplification and DNA sequencing
Exons 2-9 of the VDR gene were amplified by PCR and sequenced at the Stanford Protein and Nucleic Acid Facility.
Restriction fragment length polymorphism analysis
Exon 4 of the VDR gene was amplified from DNA using the oligonucleotide primers 4U and 5L. After an initial 5-minute denaturation at 95°C, the samples were cycled at 94°C for 30 s and 68°C for 30 s for a total of 35 cycles. PCR products were digested with the restriction endonuclease BslI at 55°C according to the manufacturer (New England Biolabs, Beverly, MA, USA) and analyzed on 2% agarose 1000 gels in Tris-acetate-EDTA (TAE) buffer. Gels were stained with ethidium bromide, visualized by UV light, and photographed.
Site-directed mutagenesis and plasmid construction
Site-directed mutagenesis of the wildtype (WT) VDR cDNA in pSG5 (Stratagene, La Jolla, CA, USA) was performed using the Gene Editor system (Promega, Madison, WI, USA) as previously described.(12) The mutant oligonucleotide used was 5′-ACTGCTGGACGCCCACCACTTATGGGCCTACG-ACCCCACCTACT-3′. Clones were sequenced to confirm the presence of the mutation.
COS-7 cells were grown to 60-80% confluence in 12-well tissue culture plates. Cells were transfected with either 125 ng WT or mutant VDR expression plasmid and 250 ng rat 24-hydroxylase promoter VDRE-luciferase plasmid using Polyfect (Qiagen, Valencia, CA, USA). A renilla-luciferase plasmid (pRLnull; 5 ng) served as an internal control for transfection efficiency. After a 16-h transfection, the cells were incubated in DMEM containing 1% FBS with or without 1,25(OH)2D3 for 24 h. The cells were washed and lysed in 250 μl of Passive Lysis Buffer (Promega). Aliquots were assayed for luciferase activities using the Dual Luciferase Assay (Promega) and a Turner Design luminometer (Turner Design, Sunnyvale, CA, USA).
Glutathione S-transferase-pull down
Glutathione S-transferase (GST) fusion proteins (RXRα and DRIP205) were expressed in E. coli BL21(DE3) after induction with 0.1 mM isopropyl-β-D-thiogalactopyranoside (IPTG) for 3 h at 37°C. Proteins were extracted by incubating the cells in B-PER extraction reagent (Pierce Biotechnology) containing 100 mM KCl, 10 mM MgCl2, 1 mM dithiothreitol, 10% glycerol, and a complete protease inhibitor tablet (1 tablet = 50 ml; Roche Molecular Biochemicals) for 10 minutes at ambient temperature with gentle shaking. Cell debris was removed by centrifugation at 12,500g for 20 minutes at 4°C. WT and mutant VDRs were labeled with [35S]-methionine (Amersham) by in vitro transcription/translation using the TNT coupled system (Promega). For binding assays, E. coli extracts containing GST-fusion proteins were mixed with glutathione agarose at 4°C for 16 h and washed. [35S]-labeled VDRs and 1,25(OH)2D3 were added to the beads in GST-binding buffer (50 mM Tris buffer [pH 7.5] containing 100 mM KCl, 10 mM MgCl2, 0.3 mM dithiothreitol, 0.1% NP-40, and 10% glycerol) and incubated at 4°C for 16 h. The agarose beads were washed three times with binding buffer. Bound proteins were eluted in 2× lithium dodecyl sulfate (LDS) sample buffer, heated at 70°C for 5 minutes, and electrophoresed on 10% SDS-NuPAGE gels (Invitrogen). Gels were fixed in 50% methanol and 10% acetic acid for 10 minutes and incubated in Amplify (Amersham) for 15 minutes. Gels were dried and exposed to Hyperfilm (Amersham) at −80°C. Nonspecific binding was determined using extracts containing GST alone.
Table Table 1.. Clinical Data
The patient is a Chilean boy first evaluated in the Endocrine Unit at Pontificia Universidad Católica de Chile, Santiago, Chile, at 32 months of age. He had severe rickets and growth retardation, several episodes of pneumonia, and seizures. There is no known consanguinity between the parents. The patient has an unaffected half-brother. On examination, the patient had bowed legs, wide metaphysis, rachitic rosary, tooth enamel hypoplasia, and normal scalp and body hair. Laboratory testing showed the following levels: calcium, 4.48 mg/dl; phosphate, 3.2 mg/dl; alkaline phosphatase, 3876 IU/liter; urinary cyclic adenosine monophosphate (cAMP), 62 nM/mg creatinine; 1,25(OH)2D, >2400 pmol (Table 1). On X-ray, the patient showed metaphyseal irregularities, osteopenia, long bone and joint deformities of rickets, and old multiple nonhealing fractures. His rickets did not improve despite 3 g of calcium carbonate orally and 6 μg calcitriol per day. Through a central intravenous catheter, he was started on a calcium infusion of 1 g/10 h (night) and oral phosphate at 860 mg/day, initially, as a daily dose, and after 1 year, every other day. After 1 month of treatment, the patient showed remarkable improvement, with fracture consolidation, healing of rickets, and biochemical improvement. After 1 year, calcium infusion was discontinued because of several catheter infections. Afterward, he was maintained on 600,000 IU of vitamin D twice a week and 2 g of calcium orally per day. During this time, he showed progressive radiological deterioration of his rickets with increased alkaline phosphatase levels; however, he was fracture-free and remained normocalcemic for 2 years. The patient was then lost to follow-up but maintained the same oral therapy. At 15 years of age, the patient returned to the clinic with his mother. At this time, the patient had severe growth retardation, seizures, bone pain, severe hypocalcemia (normal albumin level), radiological rickets, and osteopenia. He was again treated with intravenous calcium infusion. After 18 months of therapy, he showed biochemical improvement, healing of rickets, remarkable increase of BMD (assessed by DXA; Table 1), and catch-up growth of 11 cm.
Studies of patient's fibroblasts
We cultured dermal fibroblasts from a skin biopsy from the patient and used these cells to examine VDR function. We first tested the cells for 1,25(OH)2D3 responsiveness by analyzing the induction of the 24-hydroxylase gene. As shown in Fig. 1, the cells were unresponsive to doses up to 100 nM 1,25(OH)2D3 and showed only a slight response at 1 μM 1,25(OH)2D3 (not readily visible on photo reproduction). These findings clearly showed that the patient's cells are resistant to high doses of 1,25(OH)2D3. We next examined the ligand-binding properties of the patient's VDR using [3H]1,25(OH)2D3 binding assays. As shown in Fig. 2, no specific [3H]1,25(OH)2D3 binding was evident in the patient's fibroblasts at the concentrations tested.
We amplified the VDR gene and directly sequenced individual exons 2-9. A novel 5-bp deletion/8-bp substitution was identified in exon 4 (Fig. 3). A 5-bp sequence TAAGA starting at nucleotide 420 was deleted and was replaced with the 8-bp sequence CTTATGGG. The sequence CTTATGG seems to have been copied from the bottom strand of DNA because of a crossing-over event (Fig. 3). The addition of an additional guanine nucleotide at the end of the sequence followed by a second crossover event put the sequence back into the correct reading frame. The mutation changed the normal amino acid sequence from 138 AHHKTYN 144 to 138 AHHLWAY 144 (Fig. 3). We refer to this mutation as the LWA mutation.
The LWA mutation created a unique BslI restriction site in exon 4. We amplified exons 4-5 and used BslI to genotype the patient, the patient's mother, and the patient's half-brother. As shown in Fig. 4, BslI digestion of exons 4-5 from a normal control shows four bands, whereas the patient shows five bands, including smaller size fragments of 280 and 39 bp, consistent with the presence of a new BslI site. The data show that the patient is homozygous for the mutation. The patient's mother and half-brother exhibit both normal and mutant bands, showing that they are heterozygous for the LWA mutation. These results also show that the LWA mutation is recessive, because the mother and half-brother show no signs of the disease. DNA from the father was not available.
Analysis of defects in the mutated VDR
The insertion/substitution mutation occurs in helix H1 in the LBD and has a dramatic effect on ligand binding (Fig. 2). Because ligand binding is critical for VDR transactivation, RXR heterodimerization, and repositioning of helix H12 for co-activator recruitment, we next examined the effects of the LWA mutation on the ability of the VDR to carry out these functions. We recreated the LWA mutation in the pSG5-VDR cDNA expression vector and transfected COS-7 cells for analysis. As shown in Fig. 5, the LWA mutant was unresponsive up to 10 nM 1,25(OH)2D3 in transactivation assays using a 24-hydroxylase promoter-reporter containing both distal and proximal VDREs. The LWA mutant began to respond at 100 nM 1,25(OH)2D3 and achieved maximal response at 1 μM 1,25(OH)2D3. We also determined the effects of the three individual mutations (K141L, T142W, and the insertion of A143) on transactivation to determine whether any one of these three mutations alone could cause 1,25(OH)2D3 resistance. As shown in Fig. 5, the K141L mutation had no effect on transactivation, whereas the T142W mutant was about 50-fold less active. The most dramatic affect was observed with the insertion of A143 that was about 1000-fold less active than the WT VDR; it exhibited similar activity to the LWA mutant VDR. This finding suggests that the A143 insertion may shift the position of all downstream amino acids, thus preventing normal contacts with the ligand and resulting in 1,25(OH)2D3 resistance.
Because the LWA mutation caused defective ligand binding and transactivation, we wanted to determine whether the mutation affected VDR-RXRα heterodimerization. As shown in Fig. 6A, when GST-pull-down assays were used in the absence of 1,25(OH)2D3, the WT VDR exhibits a minimal amount of RXR binding. However, in the presence of 1,25(OH)2D3, RXR binding increases concomitant with increasing concentrations of 1,25(OH)2D3. Like the WT VDR, the LWA mutant showed minimal binding to RXR in the absence of 1,25(OH)2D3. However, when 1,25(OH)2D3 was added, RXR binding did not increase, showing that ligand binding is critical for maintaining a strong interaction with RXR.
We then determined if co-activator interactions were disrupted by the mutation. In GST-pull-down assays, the WT VDR binds DRIP205 in a 1,25(OH)2D3 dose-dependent manner (Fig. 6B). In contrast, the LWA mutant VDR exhibited only trace amounts of binding to DRIP205 at 1 μM 1,25(OH)2D3, showing the requirement for ligand binding in recruiting co-activators. Although we were able to show transactivation activity by the LWA mutant at 100 nM and 1 μM 1,25(OH)2D3 (Fig. 5), we did not detect a substantial increase in RXR binding or DRIP205 binding at these concentrations. This may be a result of the increased sensitivity of the luciferase assay used for transactivation compared with the GST-pull-down assay.
The child described here exhibits the classical clinical pattern of HVDRR. He initially presented with early onset and severe rickets, secondary hyperparathyroidism, hypocalcemia, and elevated serum 1,25(OH)2D3 levels. Unlike many patients with HVDRR, the child described here did not have alopecia. We identified a novel 5-bp deletion/8-bp insertion in the VDR to be the molecular cause of HVDRR in this patient. The mutation disrupts 1,25(OH)2D3 binding and causes 1,25(OH)2D3 resistance. This is the first example of an insertion/substitution mutation causing HVDRR.
Crystallographic studies on the VDR LBD show that the LBD is composed of 13 α-helices and 3 β-sheets.(4) The mutation described here is located in helix H1 in the LBD. Mutations in helix H1 may compromise the correct folding of the ligand-binding pocket, because the pocket is formed by helixes H1, H3, H4, H5, H7, H8, H9, H10, and H11.(4) Helix H1 makes contacts with other helices particularly through residues A138, H139, and T142. The insertion/substitution replaces a positively charged lysine residue (K141) with a small hydrophobic leucine residue (K141L) and a hydrophilic threonine residue (T142) with a bulky hydrophobic tryptophan residue (T142W). Our analysis showed that the K141L mutation alone has no effect on transactivation. Although the T142W mutation alone might be expected to disrupt the formation of the ligand-binding pocket by preventing H1 contact with H3, transactivation by the T142W mutant VDR was only modestly compromised. On the other hand, the insertion of the alanine residue (A143) alone had the most dramatic effect on transactivation. A143 displaces a tyrosine residue (Y143) that is important in forming a hydrogen bond with the 3-OH group of 1,25(OH)2D3.(4) The insertion of an extra amino acid before Y143 apparently alters ligand binding by shifting the position of the contact residue to Y144. Thus, the combined effect of the T142W mutation and the insertion of the extra amino acid caused by the LWA mutation ultimately causes 1,25(OH)2D3 resistance.
In addition to the dramatic effect of the LWA mutation on ligand binding, the mutation also affected RXR heterodimerization. It is interesting to note that RXR binding by the LWA mutant VDR was similar to the WT VDR in the absence of 1,25(OH)2D3. However, the LWA mutation eliminated the ligand's ability to further induce RXR binding. The ability of the LWA mutant VDR to heterodimerize with RXR in the absence of ligand may be a key reason why the patient did not have alopecia, because disruption of the heterodimer by other mutations causing HVDRR is associated with alopecia.(2)
Our data also showed that the LWA mutant VDR exhibited only trace amounts of binding to the co-activator DRIP205 at high concentrations of 1,25(OH)2D3. Because the LWA mutation disrupted ligand binding, it prevented the ligand-dependent interaction with DRIP205. The LWA mutation is located in helix H1, not in helices H3 and H12, which are thought to be involved in the formation of the co-activator binding site. The alteration in 1,25(OH)2D3 binding caused by the LWA mutation further shows the requirement for ligand binding to form the co-activator binding site. We did not analyze steroid receptor co-activator 1 (SRC-1) binding to the LWA mutant VDR. However, we suspect that SRC-1 would not bind to the VDR, because SRC-1 binding is ligand dependent, and it shares the same binding site on the VDR as DRIP205. In any event, the loss of DRIP205 binding may be sufficient to result in the loss of transactivation.
Alopecia has been found in many cases of HVDRR that have been reported to date. Recently, a number of cases have been described that did not have alopecia. In each case, the cause of HVDRR was a mutation in the LBD. Among these cases, one patient had a mutation in the contact point for the 1-hydroxyl group (R274L),(10) and one patient had a mutation in the contact point for the 25-hydroxyl group (H305Q).(12) Two other patients that did not have alopecia were described with mutations (I314S and W286R) that affected ligand binding.(11,19) Also, a patient with an E420K mutation in helix H12, which had no effect on ligand binding or RXR heterodimerization but prevented co-activator binding to the VDR, did not have alopecia.(14) The patient described here with the LWA mutation did not have alopecia. The LWA mutation disrupted ligand binding and co-activator binding, further showing that ligand binding and co-activator interactions are not critical functions of the VDR in its role in the hair cycle.
In conclusion, we have identified a novel insertion/substitution mutation in the VDR LBD that causes HVDRR. This is the first description of this type of mutation in the VDR.
We thank Dr Len Freedman for the GST-DRIP205 plasmid, Dr Hector DeLuca for the 24-hydroxylase luciferase plasmid, Dr J Wesly Pike for the VDR cDNA, and Dr Judith Campisi for the L7 plasmid. This study was supported by National Institutes of Health Grant DK42482 (DF).