Quantitation of CYP24A1 Enzymatic Activity With a Simple Two-Hybrid System

Abstract

Context:

Mutations of the CYP24A1 gene encoding the 24-hydroxylase (24OHase) that inactivates metabolites of vitamin D can cause hypercalcemia in infants and adults; in vitro assays of 24OHase activity have been difficult.

Objective:

We sought an alternative assay to characterize a CYP24A1 mutation in a young adult with bilateral nephrolithiasis and hypercalcemia associated with ingestion of excess vitamin D supplements and robust dairy intake for 5 years.

Methods:

CYP24A1 exons were sequenced from leukocyte DNA. Wild-type and mutant CYP24A1 cDNAs were expressed in JEG-3 cells, and 24OHase activity was assayed by a two-hybrid system.

Results:

The CYP24A1 missense mutation L409S was found on only one allele; no other mutation was found in exons or in at least 30 bp of each intron/exon junction. Based on assays of endogenous 24OHase activity and of activity from a transiently transfected CYP24A1 cDNA expression vector, JEG-3 cells were chosen over HepG2, Y1, MA10, and NCI-H295A cells for two-hybrid assays of 24OHase activity. The apparent Michaelis constant, Km(app), was 9.0 ± 2.0 nm for CYP24A1 and 8.6 ± 2.2 nm for its mutant; the apparent maximum velocity, Vmax(app), was 0.71 ± 0.055 d⁻¹ for the wild type and 0.22 ± 0.026 d⁻¹ for the mutant. As assessed by Vmax/Km, the L409S mutant has 32% of wild-type activity (P = .0012).

Conclusions:

The two-hybrid system in JEG-3 cells provides a simple, sensitive, quantitative assay of 24OHase activity. Heterozygous mutation of CYP24A1 may cause hypercalcemia in the setting of excessive vitamin D intake, but it is also possible that the patient had another, unidentified CYP24A1 mutation on the other allele.

Solar UVB radiation (280–300 nm) converts 7-dehydrocholesterol in human skin to vitamin D₃, which is then activated by sequential 25-hydroxylation in the liver to yield 25-hydroxyvitamin D (25OHD) and 1α-hydroxylation in the kidney to yield 1,25-dihydroxyvitamin D [1,25(OH)₂D; calcitriol] (reviewed in 1). Hepatic production of 25OHD can be catalyzed by several enzymes, principally microsomal CYP2R1, but also by mitochondrial CYP27A1 and possibly other hepatic cytochrome P450 (CYP) enzymes (1). The 25-hydroxylating systems are very efficient, converting most vitamin D to 25OHD on a single pass through the liver. Vitamin D binding protein binds 25OHD much more avidly than 1,25(OH)₂D, dramatically increasing its serum half-life, so that serum concentrations of 25OHD exceed those of 1,25(OH)₂D by about 1000-fold. 25OHD is activated to 1,25(OH)₂D by CYP27B1, principally in the kidney, but also in keratinocytes, brain, testis, macrophages, osteoblasts, and other tissues. Many patients with hypocalcemia and mutations in CYP27B1 have been described, usually presenting at 1–2 years (1, 2) but occasionally presenting later with partial deficiencies of CYP27B1 activity. Two families have been described with 25OHase deficiency due to mutations in CYP2R1 (3).

The principal enzyme inactivating 1,25(OH)₂D is mitochondrial CYP24A1, which 24-hydroxylates both 25OHD to 24,25-dihydroxyvitamin D [24,25(OH)₂D] and 1,25(OH)₂D to 1,24,25-trihydroxyvitamin D [1,24,25 (OH)₃D] (4, 5), primarily in the kidney and intestine; CYP24A1 also catalyzes the 23-hydroxylase pathway that degrades 25OHD and 1,25(OH)₂D (6). 24,25(OH)₂D is inactive and is the most abundant vitamin D metabolite in circulation. To mediate catalysis, mitochondrial P450 enzymes, including CYP27B1 and CYP24A1, must receive electrons from reduced nicotinamide adenine dinucleotide phosphate via ferredoxin reductase and ferredoxin (reviewed in 7).

Whereas defects in the activating 1α-hydroxylase, CYP27B1, cause hypocalcemia, defects in the inactivating 24-hydroxylase (24OHase), CYP24A1, cause hypercalcemia. Initial reports of 24OHase deficiency described infants less than 1 year of age with weight loss, failure to thrive, hypercalcemia, hypercalciuria and/or nephrocalcinosis, normal 25OHD levels, normal to moderately elevated 1,25(OH)₂D levels, low 24,25(OH)₂D levels, and low PTH (8, 9). Most were receiving vitamin D supplementation before diagnosis. Serum 24,25(OH)₂D was low when measured (9). About 21 patients with hypercalcemia and CYP24A1 mutations have been reported (8–15) (Table 1), but only a few CYP24A1 mutations have been characterized functionally, in part because the described assay requires complex HPLC equipment and expensive radiolabeled 1,25(OH)₂D substrate. We found the CYP24A1 mutation L409S in a 21-year-old man with hypercalcemia and nephrocalcinosis in the setting of excessive vitamin D intake and assayed the activity of the mutant in a novel, simple system.

Table 1.

Reported Patients With CYP24A1 Mutations

Patient	Mutations	Activity	Age and Sex	Highest Serum Ca (mmol/L)	25OHD (ng/mL)1,25(OH)2D (pg/mL)	iPTH pg/mL	Stones or ↑UCa	Vit D Suppl	Ref
1 (1.1)	A475fsX490	Nil	6 mo F	4.0	50	<1	Yes	500 U/d	8
	A474fsX490	Nil			65
2 (2.1)	E143del	Nil	6 mo M	4.2	27	5	Yes	500 U/d	8
	E151X	Nil			57
3 (2.2)	E143del	Nil	Asymptomatic	3.7	27	4	Yes	500 U/d	8
	E151X	Nil	M		43
4 (3.1)	L409S	5.3 ± 0.3%	8 mo M	4.3	64	<1	Yes	500 U/d	8
	R396W	Nil			79
5 (3.2)	L409S	5.3 ± 0.3%	Asymptomatic	2.4	33	ND	Yes	None	8
	R396W	Nil	M		ND
6 (4.1)	E149del	Nil	11 mo F	4.3	68	2	Yes	500 U/d	8
	R159Q	Nil			ND
7 (5.1)	E322K	Nil	7 mo F	3.5	176	ND	Yes	600 000 × 3	8
	R396W	Nil			129
8 (6.1)	E322K	Nil	3.5 mo M	3.5	ND	ND	Yes	600 000 × 2	8
	R396W	Nil			ND
9 (7.1)	R396W	Nil	7 wk F	4.1	178	ND	Yes	600 000 × 1	8
	R396W	Nil			ND
10 (8.1)	Deletion	ND	5 wk F	5.0	31	ND	Yes	600 000 × 1	8
	Deletion	ND			ND
11	E147del	ND	47 yo M	11.4 (mg/dL)	77–124	4–8	Yes	Not mentioned	9
	E147del	ND			83–118
12	E143 del	ND	10 mo M	3.8	43	<3	Yes	No	10
	E143 del	ND			33
13	IVS5 + 1G>A	ND	44 yo M	10.4 (mg/dL)	50	8.1 pg/mL	Yes	Not mentioned	11
	IVS6–2A>G	ND			123	(nl 15–65)
14 (II.1)	Del ex9-11	ND	6 mo M	3.68	140 nmol/L	1 ng/mL	Yes	1900 U/d	12
	Del ex9-11	ND			96 pmol/L	(nl 10–55)
15 (II.2)	Del ex9-11	ND	Asymptomatic	2.62	90 nmol/L	4 ng/mL	No	400 U/d	12
	Del ex9-11	ND	M		116 pmol/L	(nl 10–55)
16 (1.1)	E143del	ND	48 yo M	12.7 (mg/dL)	63.5	<10	Yes	Amount not stated	13
	E143del	ND			71.8
17 (2.1)	L409S	ND	31 yo M	11/2 (mg/dL)	58	<1	Yes	Amount not stated	13
	W268X	ND			71.4
18 (2.2)	L409S	ND	13 yo M	11 (mg/dL)	50	<3	Yes	Amount not stated	13
	W268X	ND			114.2
19	W210R	ND	3 mo M	4.2	123	ND	Yes	800 U/d	14
	W210R	ND			ND
20	E143del	ND	73 yo M	15.8	20	ND	Yes	None	15
	E143del	ND			48
21	L409S		21 yo M	15.4 (mg/dL)	43.9	8	Yes	3200 U/d	This report
			None		46

Abbreviations: ND, not done.

Subject and Methods

Subject

A 21-year-old male presented elsewhere with bilateral nephrolithiasis, hypercalcemia to 15.4 mg/dL, and renal insufficiency (creatinine, 2 mg/dL). He had been hospitalized twice for hypercalcemia and received extracorporeal shock wave lithotripsy once. Renal biopsy showed nephrocalcinosis. For 5 years before presentation, he took multiple nutritional supplements containing >3200 IU vitamin D daily, and he consumed 2–3 gallons of milk weekly (∼1600 IU vitamin D/gallon). He had episodes of nausea, vomiting, and polyuria during this time. There was no history of consanguinity, infantile hypercalcemia, nephrolithiasis, or hypercalcemia in the patient's six siblings. The patient's father had a history of nephrolithiasis in his 20s. The patient stopped all supplements and milk intake 1 year before presentation to our institution; calcium was 10.6 mg/dL (normal range, 8.6–10.5); bicarbonate, 26 mEq/L (24–32); creatinine, 1.4 mg/dL (0.44–1.27); phosphorus, 3.3 mg/dL (2.4–5.0); PTH, 8 pg/mL (12–88); PTHrP, < 2.4 pmol/L (<4.0); 25OHD, 43.9 ng/mL (30–100); and 1,25(OH)₂D, 46 pg/mL (17–74). His 24,25(OH)₂D was undetectable (<0.2 ng/mL), both by the competitive binding assay and the liquid chromatography-tandem mass spectrometry assay at Heartland Assays. His urinary calcium was 203 mg/24 h (50–400), with a creatinine level of 1950 mg/24 h (800–2000).

DNA preparation, PCR amplification, and DNA sequencing

With Institutional Review Board approval, DNA was prepared from whole blood, and the 12 CYP24A1 exons were amplified in seven segments by PCR. Each exon was sequenced on both strands (Supplemental Methods and Supplemental Tables 1 and 2). Putative mutations were checked twice by PCR amplification and sequencing of both strands of the affected exons.

Preparation of CYP24A1 expression vectors

Human liver HepG2 cells were cultured as described (16) and treated with 0.1 μm 1,25(OH)₂D₃ for 24 hours (5), and RNA was isolated. CYP24A1 cDNA was generated by RT-PCR and cloned in pcDNA3.1, and reverse transcription errors were corrected by site-directed mutagenesis (see Supplemental Methods and Supplemental Tables 3 and 4). The L409S mutation was created in this vector by oligonucleotide-mediated mutagenesis (Supplemental Methods).

Establishment of the two-hybrid assay and measurement of enzyme kinetics

Various cell lines were treated with 10 or 500 nm 1,25(OH)₂D for 24 hours, and quantitative RT-PCR measurement of CYP24A1 mRNA was done using 40 ng cDNA per reaction (Supplemental Methods).

To measure CYP24A1 activity, we used a mammalian two-hybrid assay consisting of the retinoid-X receptor (RXR) and vitamin D receptor (VDR) (17). Human RXR cloned in pCMV-BD, human VDR cloned in pCMV-AD, and pFR-Luc were generously supplied by Dr Peter Jurutka, University of Arizona. This system was used in JEG-3 cells to measure the kinetics of wild-type and L409S CYP24A1 (Supplemental Methods).

Results

DNA sequencing

CYP24A1 exons were sequenced in their entirety, including at least 30 bp of intronic DNA at each exon/intron junction. The patient was heterozygous for the previously described mutation L409S. No other mutation was found in the patient, and no mutation was found in DNA from the patient's mother; DNA from his father and siblings was not available for analysis. The patient's undetectable 24,25(OH)₂D suggests that he carried another, undetected CYP24A1 mutation, but asymptomatic heterozygous relatives of patients homozygous for CYP24A1 mutations have been reported with low or undetectable 24,25(OH)₂D (11); thus, the undetectable serum 24,25 (OH)₂D only suggests, but does not prove, that there was another, undetected mutation.

Cell line selection

The two-hybrid system uses RXRα cloned into pCMV-BD (the “bait”), which contains a heterologous DNA binding domain. When the RXR hybridizes with activated VDR (the “prey”), it can bind to and activate expression of luciferase in the pFR-Luc reporter gene; luciferase activity is easily quantitated in a luminometer. Expression of the VDR construct is activated by binding of 1,25(OH)₂D; when CYP24A1 is present, 1,25(OH)₂D is degraded to 1,24,25(OH)₃D, thus decreasing transcriptional activation. Thus, although CYP24A1 is vigorously induced by 10 nm 1,25(OH)₂D in a variety of cell types (16), the host cell should have low endogenous levels of CYP24A1 expression that are unresponsive to the exogenously added 1,25(OH)₂D substrate, thus minimizing background activity and maximizing the signal-to-noise ratio. The host cell should also express relatively high levels of the electron-transfer proteins ferredoxin reductase and ferredoxin, which are required for catalysis by CYP24A1 (and all other mitochondrial P450 enzymes). We previously showed that the use of steroidogenic cell lines that express ferredoxin reductase and ferredoxin can increase the activities of transfected mitochondrial P450 enzymes by 10- to 20-fold (18). Thus, we used RT-PCR to assay the endogenous expression of CYP24A1 in human placental JEG-3, mouse adrenal Y1, mouse testis MA10, and human adrenal NCI-H295A cells (Figure 1, A and B). Both JEG-3 and MA-10 cells had similar, low levels of CYP24A1 mRNA that were unresponsive to exogenously added 1,25 (OH)₂D; hence, we used JEG-3 cells to utilize a human system and for ease of growth and transfection.

Figure 1.

Expression and kinetics of CYP24A1. A and B, CYP24A1 expression in various cultured cell lines. In wholly separate experiments, cells were treated without and with 10 nm 1,25(OH)₂D (A) and with 500 nm 1,25(OH)₂D (B) for 24 hours, and CYP24A1 mRNA was quantitated by RT-PCR; the results without 1,25(OH)₂D are arbitrarily set at 1.0. Each incubation was done in triplicate; data shown are means ± SD. C, kinetics of CYP24A1 activity as assessed by the two-hybrid assay in JEG-3 cells. Data are presented as Lineweaver-Burke plots; data are means ± SEM from three independent experiments, each performed in triplicate. Data for wild-type CYP24A1 are shown with open symbols, and data for the L409S mutant are shown with closed symbols.

Assays of enzymatic activity

The RXR-VDR two-hybrid system provides a direct and accurate measurement of 24OHase activity (17). Because the measurements are done in whole cells and not with pure proteins, the resultant data for the Michaelis constant (Km) and maximum velocity (Vmax) are only apparent values, designated Km(app) and Vmax(app). Using this system in JEG-3 cells and treating the cells with a broad range of concentrations of 1,25(OH)₂D (1.0–100 nm), we obtained good kinetic data, indicating that wild-type CYP24A1 has a Km(app) of 9.0 ± 2.0 nm, and the L409S mutant has a Km(app) of 8.6 ± 2.2 nm (Figure 1C). Because incubations were done for 1 day, the Vmax(app) data are per day; the wild-type Vmax(app) was 0.71 ± 0.055 d⁻¹, and the mutant Vmax(app) was 0.22 ± 0.026 d⁻¹. The catalytic efficiency [Vmax(app)/Km(app)] of the wild type was 0.2, and of the mutant was 0.07. Thus, the L409S mutant has about 32% of wild-type 24-hydroxylase activity (P = .0012).

Discussion

The RXR/VDR two-hybrid system provides an assay for 24OHase activity that does not require expensive radioactive reagents or complex equipment; this system can also be used to assess the 1α-hydroxylase activity of CYP27B1 (17). In this assay, wild-type CYP24A1 had a Km(app) of 9.0 ± 2.0 nm, consistent with the nanomolar circulating concentrations of 1,25(OH)₂D; Km and Vmax values of CYP24A1 and its mutants have not been reported in prior studies (8–15). Although the structure for human CYP24A1 has not been reported, it has been reported for rat CYP24A1 (19), with which it shares 82% identity and 90% similarity. Residue 409 is leucine in both proteins; in rat, CYP24A1 L409 lies in β-sheet 2–2, which does not contribute to the substrate-binding pocket, the substrate access channel, or the redox-partner binding site. Furthermore, leucine and serine occupy similar volumes; thus, it appears that the L409S mutant probably alters protein conformation by substituting a polar residue for an aliphatic one.

Infantile hypercalcemia is a rare disorder associated with anorexia, emesis, lethargy, constipation, polyuria, irritability, seizures, and mental retardation. Infantile hypercalcemia may be associated with supravalvular aortic stenosis and “elfin” facies (Williams-Beuren syndrome), calcium-sensing receptor mutations, neonatal familial hyperparathyroidism, thyroid disease, elevated PTHrP, Jansen's metaphyseal dysplasia, subcutaneous fat necrosis, phosphate depletion, and vitamin A and D toxicity. When these disorders are excluded, the diagnosis is usually “idiopathic infantile hypercalcemia.” The incidence of idiopathic infantile hypercalcemia in the United Kingdom was estimated at approximately 18 cases per year (1 in 47 000 births) from about 1960–1980. However, from January 1953 to June 1955, 216 cases (7.2 per month) were reported in the United Kingdom, ostensibly secondary to excessive fortification of milk and cereal with vitamin D, resulting in intakes of about 4000 U/d (20). Causality was not established, but after reduction in dietary vitamin D supplementation in 1957, the incidence of hypercalcemia dropped to 3.0 per month in 1960 (20). The reason that only a few children became profoundly hypercalcemic when given excess vitamin D, but most did not, remains unknown. Infants who became hypercalcemic may have had CYP24A1 lesions because such mutations can cause both infantile and adult hypercalcemia (8–15). Most reports indicate autosomal recessive inheritance, which would be consistent with the rarity of this disorder, although autosomal dominant inheritance with partial penetrance has been suggested (11). Our patient ingested large amounts of dietary supplements, including 3200 U/d of vitamin D. This greatly increased load of vitamin D apparently exceeded the capacity of his 24OHase activity, apparently leading to manifesting heterozygosity with hypercalcemia and nephrocalcinosis. When not taking vitamin D supplements, the patient had mild hypercalcemia (10.6 mg/dL), normal serum 25(OH)D and 1,25(OH)₂D levels, and low 24,25(OH)₂D. It is not yet known whether CYP24A1 mutations were responsible for the idiopathic infantile hypercalcemia associated with excessive vitamin D supplementation in the United Kingdom in the 1950s.