The CYP2R1 Enzyme: Structure, Function, Enzymatic Properties and Genetic Polymorphism

Since the discovery of its role in vitamin D metabolism, significant progress has been made in the understanding of gene organisation, protein structure, catalytic function, and genetic polymorphism of cytochrome P450 2R1 (CYP2R1). Located on chromosome 11p15.2, CYP2R1 possesses five exons, unlike most other CYP isoforms that carry nine exons. CYP2R1 crystal structure displays a fold pattern typical of a CYP protein, with 12 -helices as its structural core, and -sheets mostly arranged on one side, and the heme buried in the interior part of the protein. Overall, CYP2R1 structure adopts a closed conformation with the B′ helix serving as a gate covering the substrate access channel, with the substrate vitamin D3 occupying a position with the side chain pointing toward the heme group. In liver, CYP2R1 25-hydroxylates vitamin D and serves as an important determinant of 25(OH)D level in the tissue and in circulation. While substrate profile has been well studied, inhibitor specificity for CYP2R1 requires further investigation. Both exonic and non-exonic single nucleotide polymorphisms (SNPs) have been reported in CYP2R1, including the CYP2R1*2 carrying Leu99Pro exchange, and a number of non-exonic SNPs with variable functional consequences in gene regulation. A non-exonic SNP, rs10741657, has its causal relationship with diseases established, including that of rickets, ovarian cancer, and multiple sclerosis. The role of other CYP2R1 SNPs in vitamin D deficiency and their causal link to other traits however remain uncertain currently and more studies are warranted to help identify possible physiological mechanisms underlying those complex traits.


INTRODUCTION
Cytochrome P450 (CYP) enzymes are a superfamily of haem-containing isoenzymes. These enzymes carry out the biotransformation of a diverse number of exogenous and endogenous compounds, including steroids, fatty acids, vitamins, leukotrienes, and prostaglandins, and contribute to drug metabolism and detoxification [1]. In the human genome, there are 57 genes which encode active CYP enzymes, and these are further classified into 18 gene families [2]. Of these 57 isoenzymes, 51 are found in microsomes and 6 are located in mitochondria. CYP isoforms from families 1 to 3 have major roles in the biotransformation and bioinactivation of xenobiotics, including drugs and toxins. On the other hand, another set of CYP enzymes, many of which are expressed extrahepatically, display important endogenous functions, for instance, in regulating the levels of bile acids, lipids, steroid hormones and other signalling molecules.
The Human Genome Project has allowed the identification of some previously uncharacterised extrahepatic CYPs, the so-called 'orphan' CYPs, including CYP2R1, CYP2S1, CYP2U1 and CYP2W1. Like their counterparts in the liver, these CYPs has been shown to biotransform endogenous and/or foreign compounds in a tissue-specific manner. CYP2R1 is a major vitamin D 25hydroxylase [3] whereas CYP2S1 has been identified as the main metabolizing enzyme for all-trans retinoic acid and naphthalene [4,5]. Likewise, CYP2U1 is known to metabolise arachidonic acid, docosahexaenoic acid (DHA) and other long chain fatty acids [6] while CYP2W1 has exhibited distinctive expression pattern in fetal and tumour tissues with pivotal role in localized retinoid metabolism [7].
With full sequence of the human genome completed and execution of various initiatives related to this project, significant progress has been made in the characterization of the structure, function, catalytic properties, and genetic polymorphism of CYP2R1. This review aims to summarize the recent understanding and comprehension of the gene organisation, protein structure and catalytic function of CYP2R1, and the influence of its gene polymorphism on occurrence of inherited rickets and other diseases.

CYP2R1 gene and protein structure Gene organisation
The CYP2R1 gene is localized on chromosome 11p15.2 and has five exons, spanning across a region of approximately 15.5 kb (Figure 1). The exon number is different from many other CYP2 members where nine exons is reported. The only exception to this is CYP2U1 which also possesses 5 exons. Both CYP2U1 and CYP2R1 have introns 1, 3, 4, 5, and 7, as found in all other CYP2 genes, removed, leaving only introns 2, 6 and 8, indicating that the two genes must have shared a common ancestral gene that has undergone partial processing and reinsertion in their genome structures. Since then, both CYP2U1 and CYP2R1 have independently acquired an additional intron, hence the unique 5-exon structure presently noted in both genes. An overview of the relationships of the 57 CYPs found in human is shown in the phylogenetic tree in Figure 2. CYP2R1 and CYP2U1 show clear orthologous relationships, and are closely related to CYP2J2, CYP2S1, CYP2W1 and CYP2D6. CYP2R1 is highly conserved among living kingdom with human CYP2R1 sharing 89% and 66% sequence homology to mouse and fugu fish CYP2R1, respectively, at primary sequence level. Highly conserved CYP enzymes are usually linked closely in metabolism of endogenous substrates, such as sterols, vitamins, and fatty acids, and CYP2R1 fits this pattern. CYP2R1 also displays high sequence homology with other CYP2 isoforms, with CYP2J2, CYP2U1 and CYP2D6 showing the highest homology among the 57 known human isoforms (41.2%, 38.6% and 38.5%, respectively). The sequence alignment of the abovementioned CYPs with CYP2R1 ( Figure 3) showed that CYP2R1 has 501 residues and has similar overall sequence features with other isoforms, including a lipophilic N-terminus, regions that interacts with the redox partner, NADPH cytochrome P450 oxidoreductase, and the cysteine ligand of the prosthetic group heme [10]. There is a region upstream of the transmembrane helix (TMH) in CYP2R1 comprising approximately 20 amino acids in its N-terminal sequence, which is not observed in other isoforms with the exception of CYP2U1 and CYP2J2 that show additional 40 and 20 pre-TMH residues, respectively ( Figure 3). It is possible that these additional sequences affect intracellular targeting, membrane incorporation, or structure of CYP2R1, CYP2U1 and CYP2J2. The tissue distribution of CYP2R1 mRNA in humans and mice has been investigated. In humans, the descending order of CYP2R1 mRNA levels was pancreas, liver, and kidney. In mice, liver and testis demonstrated higher mRNA levels, with lower levels reported in other tissues such as brain, kidney, muscle, and skin [3]. A search in the web-based atlas of human gene expression, Amazonia! [11], for the distribution of CYPs in the human body, showed ubiquitous distribution of CYP2R1 in many tissues with testes recorded the highest mRNA levels, and quite high levels detected in many other organs. As for cDNA cloning, the mouse CYP2R1 cDNA was cloned from a cDNA library constructed from CYP27A1-deficient mice. The cDNA-expressed CYP2R1 showed capability to perform 25hydroxylation in both vitamin D2 and vitamin D3 [3,12].

Protein structure
A number of crystal structures of human CYP2R1 have recently been determined and published, including 3C6G (complexation with vitamin D3), 3CZH (complexation with vitamin D2), 2OJD (complexation with vitamin D3), and 3DL9 (complexation with 1--hydroxy-vitamin D2). The first structure (PDB code 3C6G) was solved by Strushkevich and coworkers [12]. The CYP2R1 existed as homodimer in solution and crystallized in a dimeric complex with the detergent molecule 2hydroxypropyl-b-cyclodextrin bound. It has a fold pattern typical of a CYP protein, comprising 12 helices (A to L), a cluster of -sheets mostly located on one side, and the heme group that was buried in the protein interior ( Figure 4A). The core of the protein was formed by helices E, I, J, K and L. Helices F and G formed part of the homodimer interface and bordered the active site partly. The long I helix ran across the entire protein ( Figure 4A), no distortion of this helix was noted in CYP2R1, even though it is observed in many mammalians and bacterial CYPs where a kink of the I-helix, a result of local distortion of intra-helical hydrogen bonding, was found to be important for CYP catalytic cycle in the delivery of proton to the O2 molecule [13]. Nevertheless, CYP2R1 carried the highly conserved Thr314 which may be involved in proton transfer cascade to the heme-bound oxygen. The heme was tightly bound within the active site. Residues involved in this included Trp133, Arg137, Arg446 which were hydrogen bonded to the D-ring propionate, and those interacted with the A-ring propionate were Arg109, His381 and Ser442.
Overall, CYP2R1 structure adopts a closed conformation with the B′ helix serving as a gate covering the substrate access channel. The active site is bordered by B′ helix, F helix, I helix, the loop between C-terminal β-sheet 4, and the loop between helix K and β-sheet 1. The size of CYP2R1 active site, as determined from the crystal structure, is compared to a number of CYP crystal structures that have been published ( Table 1). The active site volume of CYP2R1 (979 Å 3 ) is intermediate between those of CYP2A6, CYP2A13, CYP2B4 and CYP2E1 (<500 Å 3 ), that are known to only bind small substrates, and the isoforms receiving large substrates (volume in the range of 1000 -2000 Å 3 ), such as CYP2C8 [19]. Its volume is comparable to those of CYP2J2 or CY2U1, therefore CYP2R1 is classified under those having a moderately sized active site cavity. Vitamin D3 was demonstrated to enter the active site and was bound, in an elongated conformation, between the B' helix/B-C loop and helices G and I with the side chain directed toward the heme group ( Figure 4B). The lining of the active site was formed mainly by hydrophobic residues. The 3-OH group of the A-ring of vitamin D3 faced the protein surface and adopted a chair conformation with an axial orientation of the 3-OH group. The 3-OH was hydrogen bonded to the amide function of Ala250 of the helix G in addition to a watermediated hydrogen bond with Asn217 in helix F. The A-ring was surrounded by residues with aromatic side chains. Phe214 of helix F formed van der Waals interaction with the conjugated triene of vitamin D3 and adopted a planar geometry perpendicular to the heme. Another hydrophobic residue, Phe115 of helix B' made hydrophobic contact with the C-18 methyl group of vitamin D3. The 17-aliphatic side-chain of vitamin D3 was oriented within the active site cavity that placed C-25 toward the heme iron, for hydroxylation but at a distance of 6.5 Å, in contrast to a distance of 5 Å or less observed in other CYPsubstrate structures such as CYP2C9 [25]. Taken together, substrate selectivity in CYP2R1 largely reflects interactions with the protein near the entrance of the substrate binding site where the secosterol portion of the substrate molecule is sequestered. Similar to other microsomal CYPs, CYP2R1 utilizes both electrostatic interactions in complex formation with the redox partner NADPHcytochrome P450 reductase. The interacting residues have been mapped and include residues in C helix (Arg131, Arg137, Arg138 and Arg145), meander region (Lys434, Lys435 and Arg445), as well as L helix (Arg455). secondary structure elements are coloured (helices in blue, β-sheets in red). Predicted substrate recognition sites (SRS) of CYP2 family enzymes [9] are underlined and designated. Strongly conserved residues are indicated below the alignment with asterisks, colons indicate the residues variation occurring within the strongly conserved groups and dots indicating the residues variation occurring within weaker conserved residue groups. The sequences were taken from UniProtKB database with the following accession numbers (Q6VVX0 -CYP2R1, Q96AQ9-CYP2S1, Q7Z449-CYP2U1, Q8TAV3-CYP2W1, P10635-CYP2D6 and P51589-CYP2J2).

CYP2R1 activity and function Vitamin D metabolic pathway
Vitamin D plays a central role in calcium homeostasis and bone metabolism. Its production in the skin involves non-enzymatic reactions ( Figure 5). Vitamin D3 (cholecalciferol, the major secosteroid precursor in human body) is a breakdown product of 7-dehydrocholesterol (7-DHC) via a two-step process involving breaking down of a 7-DHC internal ring by the sun light, in particular radiation from UV light (spectrum 280-320 UVB), to form pre-vitamin D3. The pre-cursor subsequently forms vitamin D3 via isomerization in a thermosensitive but noncatalytic process. The rate of vitamin D3 formation therefore depends on the intensity of UVB and skin pigmentation level [26]. An alternative source of vitamin D is dietary factor. Fatty fish is rich in vitamin D but level in most other foods is low unless fortified. The form in the fish is vitamin D3, whereas vitamin D2 (ergocalciferol) is usually used in fortified foods. Vitamin D2 is produced by UVB irradiation of the ergosterol in plants and fungi such as mushrooms [27]. Both of these vitamins D lack intrinsic biological activity and are considered as hormone precursors.
Vitamin D is bioactivated through two sequential hydroxylation steps mediated by 25hydroxylase and 1-hydroxylase to generate 25hydroxyvitamin D (25(OH)D3 or 25(OH)D2) in the liver and later, the hormonally active 1,25dihydroxyvitamin D (1,25(OH)2D3 or 1,25(OH)2D2) mainly in the kidney, respectively ( Figure 5). 1,25(OH)2D has a shorter half-life and a lower concentration than 25(OH)D (1000-fold lower), it is however the most active form of vitamin D and contributes to the bulk of the biological function of the vitamin. There is another hydroxylation step involving inactivation of 25(OH)D, by 24hydroxylase, to form biologically inactive calcitroic acid. Distinct CYP isoforms are involved in production of the active hormone from inactive precursors and the subsequent biotransformation and inactivation. These enzymes include both microsomal (e.g., CYP2R1) or mitochondrial (e.g., CYP24A1, CYP27A1, and CYP27B1) isoforms. 25(OH)D, the major circulating form, is the most commonly used marker for vitamin D status, reflecting vitamin D input, from all sources including the diet and environment [28]. Also, in epidemiological studies, both 25(OH)D and 1,25(OH)2D are the most common reported biomarkers.
The liver is a well-established source for 25(OH)D production from vitamin D. Initial studies of the hepatic 25-hydroxlase found activity in both endoplasmic reticulum and mitochondrial fractions, and studies carried out subsequently have identified several CYPs to possess 25-hydroxylase activity. Thus far, the only mitochondrial 25-hydroxylase identified is CYP27A1, and it is widely distributed in the body inclusive of the liver [29]. Subsequent to identification of CYP27A1, CYP2R1 was identified as an active hepatic microsomal isoform in 25hydroxylation in mice [3]. This isoform catalysed hydroxylation of both vitamin D2 and D3 with similar kinetics, unlike CYP27A1 that only hydroxylated vitamin D3. Its expression was mainly in the liver and testes. The 25-hydroxylase activity has been found in CYP isoforms other than CYP27A1 and CYP2R1 ( Figure 5). CYP3A4, the major drug-metabolizing isoform preferentially expressed in liver and intestine, possessed 25-hydroxylase activity [30]. CYP3A4 preferred 1(OH)D to both vitamin D2 and D3 as the substrate. CYP2J3 was also identified as the rat liver isoform with 25-hydroxylase activity, but CYP2J2, the human homolog to CYP2J3 and a primary isoform in the heart, was found to have lower 25-hydroxylase activity, and functioned mainly as an arachidonic acid epoxygenase [31]. CYP2C11, expressed in the liver of male rats, was demonstrated to have 25-hydroxylase activity for vitamin D3 and D2 as well as the 1(OH)D analogues but it was better established for its role in testosterone hydroxylations [32]. It is generally believed that CYP2R1 is the major 25-hydroxylase, but other enzymes having the same activity may play a role as determinants of the 25(OH)D level within a distinct tissue and possibly its circulating levels.
The second activation step in vitamin D metabolism involves 1α-hydroxylation by the mitochondrial CYP27B1 ( Figure 5) [33,34]. CYP27B1 is widely distributed in many different tissues but generally at low levels. CYP27B1 catalyzed formation of 1,25(OH)2D at the proximal tubule of the kidney, it primary site of action, from which its product entered the circulation and served its endocrine function [35]. Biologically active vitamin D forms have in general short half-lives in target cells. These active vitamin D forms attenuate their activity by auto-induction of metabolism via CYP24A1. CYP24A1 was shown to repeatedly attack the C20-27 side chain and was the only identified 24-hydroxylase in vitamin D metabolic pathway [36,37]. The 24-hydroxylase pathway was responsible for a cascade of sequential metabolic events resulting in a variety of products with increasing polarity, including the biologically inactive calcitroic acid, and eventual loss of hormonal activity [38].

CYP2R1 catalytic properties
As discussed earlier, CYP2R1 was first identified as microsomal vitamin D 25-hydroxylase in mouse and human via screening of a liver cDNA library from gene knockout mice for cyp27a1 using a vitamin D receptor ligand activation assay [3]. The 25hydroxylase activity of CYP2R1 was confirmed in the radioimmunoassay for detection of 25(OH)D3 as well as on thin layer chromatography plate based on co-migration with authentic standard. It was also shown that CYP2R1 worked together with CYP27B1 in generating 1,25(OH)2D3 and their activity was counteracted by co-expression of CYP24A1, the isoform implicated in the degradation of the active ligand.
Notably, CYP2R1 catalyzed 25 hydroxylation on both vitamin D3 and vitamin D2 at comparable rate, a feature that was not observed in CYP27A1 [39]. CYP2R1 has subsequently been expressed in yeast cells, where the microsomal fraction was isolated and examined for activity. This yeast-expressed CYP2R1 can hydroxylate both vitamin D3 and vitamin D2 at C25, with hydroxylase activity for vitamin D3 being about double of the activity for vitamin D2 [40]. Strushkevich and coworkers who first described CYP2R1 crystal structure (see section above) have demonstrated that the E. coli-expressed and purified CYP2R1 was able to metabolize vitamin D3, 1(OH)D3 and 1(OH)D2 at similar rates to those reported for the yeast system [12]. In a more recent work, CYP2R1 was expressed in E. coli and purified by nickel affinitychromatography and assayed for activity in a reconstituted membrane system comprising phospholipid vesicles supplemented with human NADPH cytochrome P450 reductase. CYP2R1 25hydroxylated vitamin D3 showing the typical Michaelis-Menten kinetics. The kinetic parameters for both CYP2R1 and CYP27A1 were investigated in the reconstituted mixtures under identical conditions. Lower kcat and Km than CYP27A1were observed for CYP2R1, leading to an overall 17-fold higher catalytic efficiency (kcat/Km), confirming the major role of CYP2R1 as the vitamin D 25hydroxylase. Moreover, CYP2R1 was able to metabolize 20(OH)D3, the main product of an alternative pathway mediated by CYP11A1, to 20,25(OH)2D3, with catalytic efficiency similar to that for the 25-hydroxylation of vitamin D3 [41]. It is apparent that the substrate specificity of CYP2R1 has been well characterized thus far, the inhibitor specificity however has not been investigated and reported, therefore this represents an area worthy of further exploration. In

Vitamin D signalling system
The hormonally active form 1α,25-dihydroxyvitamin D, 1,25(OH)2D3 or 1,25(OH)2D2, are similar functionally, and regulate the expression of a number of genes after binding to the vitamin D receptor (VDR). VDR is a widely expressed steroid hormone nuclear receptor family member serving as a transcription factor. Using real-time quantitative RT-PCR in mice, VDR was found to express mainly in the kidney and gastrointestinal tract, moderate levels in the thyroid, skin and bone, lower levels in most other organs, and very low levels in liver [42]. Generally, vitamin D exhibits genomic action similar to other steroid hormones. 1,25(OH)2D binds, with high affinity, in the hydrophobic pocket of the VDR, thereby inducing a conformational change leading to formation of the transcriptional complex [43]. The activated complex then binds to the retinoid X receptor (RXR) and the resulting heterodimer translocates to vitamin D response elements, specific hexameric DNA sequence response elements, in the regulatory region of the target genes. This is followed by recruitment of coregulatory complexes eventually resulting in its genomic activity [44]. VDR activation leads to modulation of the gene expression in a cellspecific way. A large number of genes may be involved with up to 5% of the genome affected, and more than 900 genes are known to activated directly by vitamin D [45,46]. Among these are genes important for calcium homeostasis and skeletal integrity, including osteopontin, osteocalcin, calbindin, calcium channels TRPV5 and TRPV6, and RANKL [35]. In addition, vitamin D responsive genes are involved in modulating many physiological and pathophysiological systems including, among others, inflammation, cell proliferation and differentiation, metabolism, immune responses, response to oxidative stress, production of antibiotic peptides, as well as mental and cognitive function [46][47][48][49].

CYP2R1 polymorphism and health Genetic Polymorphism of CYP2R1
An evidence for pivotal role of CYP2R1 in vitamin D activation and signalling as the 25-hydroxylase was the finding of a CYP2R1 polymorphism carrying Leu99Pro mutation in a Nigerian family that was presented with vitamin D-dependent rickets (VDDR) type 1B [50]. This disease was postulated four decades ago [51]. The genetic basis of the Leu99Pro mutation of CYP2R1 was however only elucidated by Cheng and colleagues [52] a decade after the initial report of the Nigerian rachitic patient [50]. This has made follow-up with the patient and family difficult. Cheng and colleagues also, in the same study, performed genetic analysis of CYP2R1 in 50 Nigerian individuals [52] and revealed one heterozygote with the Leu99Pro mutation, therefore pointing to a founder gene effect in the Nigerian population in which there was a high prevalence of vitamin D deficiency [53]. Leu99Pro exchange is the result of a single nucleotide polymorphism (SNP) involving 6359T>C mutation in exon 2 of CYP2R1 gene, and the allele has been designated as CYP2R1*2 (accessed in June 2020) on the Human Cytochrome P450 Allele Nomenclature Committee home page (https://www.pharmvar.org/genes). The molecular basis for the 25-hydroxylase deficiency involving the mutation has been elucidated in the CYP2R1 crystal structure. In the structure [12], Leu99, a B helix residue, was closely oriented to the heme binding loop but no direct substrate binding was observed. A rigid network of hydrogen bonds and structural water molecules was demonstrated to be important for structural integrity of this region in CYP proteins. The Leu99 carboxyl group was found to bond with Arg445 amide nitrogen via a water molecule. The Arg residue was located in critical part of the active site and was three residues from the heme coordinating Cys448 ( Figure 6). The Leu99Pro mutation is believed to be able to disrupt the H-bond network and sterics of the helix in the region since proline residue is a helix breaker. Therefore, the mutation has likely caused protein misfolding that may adversely affect the enzymatic activity. Consistent with this notion, experiments using heterologous expression showed that the mutated CYP2R1 enzyme showed had negligible activity toward vitamin D3 [52] or failed to express [12]. Studies using gene knockout technique revealed that serum 25(OH)D level was reduced by half compared with wild-type or heterozygous littermates in CYP2R1 knockout mice [29], implying that although CYP2R1 is a major physiologically relevant vitamin D 25-hydroxylase, redundancy exists in the vitamin D 25-hydroxylase "family" of enzymes, and there was compensation mechanism for the deletion of CYP2R1. Subsequent to the discovery of CYP2R1*2, over 20 SNPs of the CYP2R1 gene associated with alteration in vitamin D levels have been identified in a number of genome-wide association studies (GWAS) carried out in European populations [54][55][56][57]. Numerous replication studies have also been conducted in populations of other ethnicity [58,59]. Among the identified SNPs, association between a SNP located in the non-coding region 5′-UTR, rs10741657, and 25(OH)D levels has been established [60][61][62][63] where individuals with the GG genotype demonstrated a decreasing trend of 25(OH)D levels in comparison to the no-risk genotype AA. It is believed that SNP in this 5′-UTR region is likely to regulate gene expression, and therefore the levels of activity and expression of 25hydroxylase may be affected [64]. Two other SNPs, the rs10766197 and rs12794714, have been identified in the coding region of introns, and these polymorphisms may be associated with changes in selective splicing regulation and gene regulation [64]. Subsequent GWAS however did not show any link between the SNPs and 25(OH)D levels in the total or subgroup populations. As rs12794714 and rs10766197 are in the intron region, these synonymous mutations do not change the protein sequence, hence minimal effect on the 25hydroxylase activity and function [65]. To date, fairly large number of polymorphisms of the CYP2R1 gene, including the above-mentioned SNPs, have been reported in various SNP databases including Sanger COSMIC, Ensembl, dbSNP, and 1000 Genomes. Figure 6. Positioning of Leu99 in the CYP2R1 structure. The heme is in light magenta, vitamin D3 in yellow, and the water molecule is shown as red sphere.

Implications in health
As a steroid pro-hormone and fat-soluble vitamin, vitamin D plays a pivotal role in bone mineralization. Calcium and phosphorus levels in both blood and extracellular tissues are regulated by this fat-soluble vitamin, and adequacy of the mineral levels is vital for the deposition of calcium hydroxyapatite to the bone matrix. Association of vitamin D deficiency with osteomalacia in adults and rickets in children has long been established [66]. Other than musculoskeletal disorders, common diseases, many of which are of important public health concerns, have been connected with vitamin D level. These include cardiovascular disease, cancer, infectious disease, autoimmune inflammatory disorder, and diabetes [67][68][69][70]. Meta-analyses from a number of epidemiological studies showed a close relationship between levels of the precursor 25(OH)D and the incidence and severity of the abovementioned diseases that might be improved by adequate supplementation with vitamin D [71,72]. Treatment using vitamin D analogues works successfully in indications like type I rickets, secondary hyperparathyroidism, osteoporosis, and hyperproliferative skin conditions such as psoriasis [73,74]. Furthermore, experimental and clinical evidence has emerged that vitamin D analogues should work in some of the autoimmune diseases and malignancies [75,76].
The causal role of vitamin D in bone-related disorders is well established, and its importance beyond bone health has begun to emerge in recent years. Findings from GWAS, whole genome sequencing (WGS), as well as Mendelian randomization (MR) studies have shed light on genetic determinants of circulating vitamin D, including SNPs associated with CYP2R1, and their putative causal relationship with various disorders and traits. GWAS of the genetic determinants of serum 25(OH)D concentrations [54,60] concluded that variants at the chromosomal site for CYP2R1 (11p15.2) (SNP rs10741657) was the second strongest association of only four sites, with vitamin D-binding protein (DBP, formerly known as GC), 7dehydrocholesterol reductase (DHCR7), and CYP24A1, being the others (rs2282679, rs17216707 and rs12785878, respectively). These four susceptibility loci have exhibited a critical role whereby their common genetic variants have been associated to altered regulation of circulating 25(OH)D concentrations. This is not unexpected because the four loci are all in or near genes encoding crucial steps in vitamin D synthesis, transportation and bioinactivation. Interestingly, variants of the other 25-hydroxylases, such as CYP27A1, did not demonstrate any significant association with serum 25(OH)D levels, again arguing that it plays negligible role in 25-hydroxylation of vitamin D at physiologically relevant concentrations.
One recent GWS study has investigated the impact of low-frequency and rare variants (minor allele frequency, MAF < 5% and < 1%, respectively) in 25(OH)D, using WGS data from the UK10K program involving 2619 individuals, and deepimputation data from 39,655 individuals with genome-wide genotypes [65]. Meta-analysis has identified a synonymous CYP2R1 variant with lowfrequency occurrence (MAF=2.5%, rs117913124) among the 19 participating cohorts. This SNP was able to modulate 25(OH)D levels substantially, showing four times stronger impact and the change was independent of the influence of another reported SNP rs10741657 in CYP2R1. This SNP has also been found to double the risk of vitamin D insufficiency, a magnitude that is likely to be of clinically relevance. Another recent GWS study investigated the vitamin D modulating genes in early childhood for a cohort of 761 healthy Finnish children who participated in a vitamin D intervention study [77]. The study has involved a genome-wide analysis of almost 700,000 genetic variants and demonstrated that genetic variation in DBP and CYP2R1 were important determinants for serum 25(OH)D concentration. Genetic variants within the DBP gene also demonstrated impact on how the children responded to vitamin D supplementation. In addition, data from the study suggested that in 2-year-old children, vitamin D level, even when within the normal range, influenced bone strength as children with genetic constellations linked to lower vitamin D concentration and poorer response to vitamin D supplementation also showed lower bone strength, as judged from skeletal peripheral quantitative computed tomography parameters. These results, combined with data from other GWASs, are consistent with an oligogenetic architecture for vitamin D. This indicates that vitamin D is affected by a few genetic variants, with some having relatively large effects, therefore indicating an oligogenic feature [78].
MR is an approach that assesses the causal effect of a risk factor on an outcome from observational data where genetic variants are used as instrumental variables [79]. MR analyses have provided some evidence for a role of the genetically predicted level of 25(OH)D in certain traits or diseases. One MR study, looking into the GWAS summary statistics from ovarian cancer cases, has provided evidence for a causal effect of vitamin D in the cancer. The odds ratio was estimated to be 1.27 (1.06-1.51) per 20 nmol/L decline in 25(OH)D concentration for epithelial ovarian cancer risk, and the risk was higher (1.54 [1.19-2.01]) for high-grade serous subtype [80]. These results are supported by a Danish MR study which showed a causal effect of low circulating 25(OH)D with cancer-specific mortality at a borderline significance [1.03 (1.00-1.06)] [81]. The investigated gene loci include CYP2R1 SNPs, rs10741657 and rs12794714, alongside with those of DHCR7 and DBP. Despite these positive relationships observed, no significant effect of 25(OH)D on the canser risk was reported in the MR studies conducted in other cancer types, including oral, oesophageal, lung, pancreatic, breast, prostate, colorectal, neuroblastoma, glioma, melanoma, and non-melanoma skin cancer [82][83][84][85][86][87][88][89][90][91]. These data imply that although some degree of causal effect of 25(OH)D with maglinancies cannot be ruled out, there may be a limit on the potential benefits of 25(OH)D on lowering cancer risk.
The causal link between vitamin D and metabolic diseases or traits remains unclear at this stage. So far, only one MR study has reported a change of 0.029 mmHg in diastolic blood pressure, and an 8.1% decreased odds ratio in hypertension [0.92 (0.87-0.97)] for each 10% increase in genetically instrumented 25(OH)D level [92]. Another MR study has demonstrated a link between 50% reduction in genetically predicted 25(OH)D level with 6.0% lower level (p=0.001) of highdensity lipoprotein [93]. No obvious effect of genetically predicted vitamin D level has been found for type 2 diabetes [62], body mass index [94], cardiovascular diseases [95,96], metabolic syndrome [97] or other metabolites [98,99,100].
A causal role has been well established in the autoimmune neurodegenerative disorder, multiple sclerosis (MS), for vitamin D. Four independent MR studies have clearly demonstrated a positive link between genetically predicated 25(OH)D and MS.  [106]. A more recent MR study investigating the causal association between DBP levels and Alzheimer's disease showed that a standard deviation increase of the genetically DBP levels was significantly associated with reduced AD risk [odds ratio = 0.63 (0.45-0.89); p = 0.009 [107]. In this study, the DBP rs2282679 variant was selected as the instrumental variable, and the dataset was extracted from the multiple GWAS summary statistics data from International Genomics of Alzheimer's Project (IGAP). The positive associations observed above for MS and Azheimer's disease were however not evidenced across autoimmune and neurodegenerative diseases. No significant finding has been revealed for asthma [108], rheumatoid arthritis [109], Parkinson's disease [110], cognition status [111,112] or amyotrophic lateral sclerosis [113].
As a summary, examination of the genetic determinants of vitamin D could expand our knowledge to the disease itself, assist in the screening of vitamin D deficiency, and also promote personalized recommendations related to the use of supplementation. Outcomes assessed by GWAS, WGS and MR studies of vitamin D have been mixed, with both positive and negative outcomes noted. There is a need for findings from these studies to be interpreted in the context of other evidence relevant to the particular issues under investigation. The results derived from the studies may help with planning of randomised clinical trials on target populations or diseases, and also help to prioritize vitamin D treatment to the most promising target diseases. Moving forward, well-powered populationbased whole genome sequencing and exome sequencing studies are required to reveal the role of rare and low-frequency variants in vitamin D-related disorders. Identification of new promising susceptibility loci would provide an improved instrument strength and accuracy of estimation, and will further enhance our knowledge of genetic regulation of vitamin D homeostasis.

CONCLUSION
Since the identification of CYP2R1 as the major microsomal 25-hydroxylase for vitamin D [3], significant progress has been made in the characterization of gene organisation, protein structure and catalytic function of CYP2R1, and the relationship of its genetic polymorphism with inherited rickets and other vitamin D-deficiency diseases. The CYP2R1 gene is localized on chromosome 11p15.2 and has five exons, unlike most other CYP isoforms that carry nine exons. From the X-ray crystallographic studies of the protein, CYP2R1 displays a fold pattern typical of a CYP protein, consisting of 12 -helices, with -sheets mostly on one side, and the heme buried deeply in the protein interior [12]. Overall, CYP2R1 structure adopts a closed conformation with the B′ helix serving as a gate covering the substrate access channel, which defines the substrate entrance point. Co-crystallized vitamin D3 in the CYP2R1 occupied a position with the side chain pointing toward the heme group. In liver, CYP2R1 25-hydroxylates both vitamin D2 and D3 with comparable kinetics and plays pivotal role in determining 25(OH)D level in both the tissue and circulation. While substrate profile has been well studied, inhibitor specificity for CYP2R1 however has not been well investigated to a significant extent to date. Like most other CYP isoforms, several polymorphisms have been identified in CYP2R1 gene. Both exonic and nonexonic SNPs have been reported, including the CYP2R1*2 which carries a SNP resulting in Leu99Pro exchange in protein sequence [52], and a number of non-exonic SNPs with variable functional consequences in gene regulation. A non-exonic SNP, rs10741657, located in the 5'-UTR, is established as an important genetic determinant of circulating vitamin D level, and has its causal relationship with diseases established, including that of rickets, ovarian cancer and MS [54,60,80,101,102,103]. The role of other CYP2R1 SNPs in vitamin D deficiency and their causal link to other traits however remain uncertain currently and more studies are needed to unveil the gene-disease association for CYP2R1.

ACKNOWLEDGMENTS
The planning and preparation of this review were supported by the FRGS Grant (no. FRGS/1/2014/SKK03/MUSM/02/1) of the Ministry of Higher Education, Malaysia. Phylogenetic analysis of CYP sequences was accomplished using Phylogeny.fr, a web-based phylogenetic inference software developed and managed by Dr Sebastien Santini (CNRS/AMU IGS UMR 7256) of the French National Centre for Scientific Research, France. All authors were responsible for the concept, design and writing of the article, as well as reviewing, and final proofing of this article and graphics performance.