Hydroxylation of CYP11A1-derived products of vitamin D3 metabolism by human and Mouse CYP27B1

Article abstract of DOI:10.1124/dmd.113.050955

CYP11A1 can hydroxylate vitamin D3 at carbons 17, 20, 22, and 23, producing a range of secosteroids which are biologically active with respect to their ability to inhibit proliferation and stimulate differentiation of various cell types, including cancer cells. As 1a-hydroxylation of the primary metabolite of CYP11A1 action, 20S-hydroxyvitamin D3 [20(OH)D3], greatly influences its properties, we examined the ability of both human and mouse CYP27B1 to 1a-hydroxylate six secosteroids generated by CYP11A1. Based on their kcat/Km values, all CYP11A1-derived metabolites are poor substrates for CYP27B1 from both species compared with 25-hydroxyvitamin D3. No hydroxylation of metabolites with a 17a-hydroxyl group was observed. 17a,20-Dihydroxyvitamin D3 acted as an inhibitor on human CYP27B1 but not the mouse enzyme. We also tested CYP27B1 activity on 20,24-, 20,25-, and 20,26-dihydroxyvitamin D3, which are products of CYP24A1 or CYP27A1 activity on 20(OH)D3. All three compounds were metabolized with higher catalytic efficiency (kcat/Km) by both mouse and human CYP27B1 than 25-hydroxyvitamin D3. CYP27B1 action on these new dihydroxy derivatives was confirmed to be 1ahydroxylation by mass spectrometry and nuclear magnetic resonance analyses. Both 1,20,25- and 1,20,26- trihydroxyvitamin D3 were tested for their ability to inhibit melanoma (SKMEL-188) colony formation, and were significantly more active than 20(OH)D3. This study shows that CYP11A1-derived secosteroids are 1ahydroxylated by both human and mouse CYP27B1 with low catalytic efficiency, and that the presence of a 17a-hydroxyl group completely blocks 1a-hydroxylation. In contrast, the secondary metabolites produced by subsequent hydroxylation of 20(OH)D3 at C24, C25, or C26 are very good substrates for CYP27B1.

Full text of DOI:10.1124/dmd.113.050955

1521-009X/41/5/1112–1124$25.00
http://dx.doi.org/10.1124/dmd.113.050955
DRUG METABOLISM AND DISPOSITION
Drug Metab Dispos 41:1112–1124, May 2013
Copyright ª 2013 by The American Society for Pharmacology and Experimental Therapeutics
Hydroxylation of CYP11A1-Derived Products of Vitamin D3
Metabolism by Human and Mouse CYP27B1
Edith K. Y. Tang, Jianjun Chen, Zorica Janjetovic, Elaine W. Tieu, Andrzej T. Slominski, Wei Li,
and Robert C. Tuckey
School of Chemistry and Biochemistry, The University of Western Australia, Crawley, Western Australia, Australia (E.K.Y.T., E.W.T.,
R.C.T.); and Department of Pharmaceutical Sciences, College of Pharmacy (J.C., W.L.), and Department of Pathology and
Laboratory Medicine (Z.J., A.T.S.) and Department of Medicine (A.T.S.), University of Tennessee Health Science Center, Memphis,
Tennessee
Received January 12, 2013; accepted February 28, 2013
ABSTRACT
CYP11A1 can hydroxylate vitamin D3 at carbons 17, 20, 22, and
23, producing a range of secosteroids which are biologically
active with respect to their ability to inhibit proliferation and
stimulate differentiation of various cell types, including cancer
cells. As 1a-hydroxylation of the primary metabolite of CYP11A1
action, 20S-hydroxyvitamin D3 [20(OH)D3], greatly influences its
properties, we examined the ability of both human and mouse
CYP27A1 activity on 20(OH)D3. All three compounds were metab-
olized with higher catalytic efficiency (kcat/Km) by both mouse and
human CYP27B1 than 25-hydroxyvitamin D3. CYP27B1 action
on these new dihydroxy derivatives was confirmed to be 1a-
hydroxylation by mass spectrometry and nuclear magnetic res-
onance analyses. Both 1,20,25- and 1,20,26- trihydroxyvitamin
D3 were tested for their ability to inhibit melanoma (SKMEL-188)
CYP27B1 to 1a-hydroxylate six secosteroids generated by CYP11A1. colony formation, and were significantly more active than 20(OH)D3.
Based on their kcat/Km values, all CYP11A1-derived metabolites This study shows that CYP11A1-derived secosteroids are 1a-
are poor substrates for CYP27B1 from both species compared hydroxylated by both human and mouse CYP27B1 with low
with 25-hydroxyvitamin D3. No hydroxylation of metabolites with
catalytic efficiency, and that the presence of a 17a-hydroxyl
a 17a-hydroxyl group was observed. 17a,20-Dihydroxyvitamin D3 group completely blocks 1a-hydroxylation. In contrast, the sec-
acted as an inhibitor on human CYP27B1 but not the mouse
enzyme. We also tested CYP27B1 activity on 20,24-, 20,25-, and
20,26-dihydroxyvitamin D3, which are products of CYP24A1 or
ondary metabolites produced by subsequent hydroxylation of
20(OH)D3 at C24, C25, or C26 are very good substrates for
CYP27B1.
Introduction
This has led to the search for new vitamin D analogs that retain the
antiproliferative and prodifferentiative properties of 1,25(OH)2D3 but
are not calcemic (Masuda and Jones, 2006; Takahashi and Morikawa,
2006; Lee et al., 2008; Slominski et al., 2010). Vitamin D analogs that
have such properties are produced by the action of cytochrome
P450scc (CYP11A1) on vitamin D (Slominski et al., 2010, 2011;
Wang et al., 2012), and several of them show antimelanoma activity in
cell culture models (Janjetovic et al., 2011; Slominski et al., 2012a,
2013b). CYP11A1 can act on vitamin D3, producing at least eight
different mono-, di-, and trihydroxyvitamin D derivatives, with the
major products being 20S-hydroxyvitamin D3 [20(OH)D3] and 20,23-
dihydroxyvitamin D3 [20,23(OH)2D3] (Fig. 1) (Guryev et al., 2003;
Slominski et al., 2005, 2012b; Tuckey et al., 2008a, 2011). Several
of these secosteroids appear to be produced in vivo, with 20(OH)D3,
22(OH)D3, 20,23(OH)2D3, and 17,20,23(OH)3D3 being detected
in cultured keratinocytes without the addition of exogenous vitamin
D3 as a substrate (Slominski et al., 2012b). To date, 20(OH)D3 and
20,23(OH)2D3, as well as 20-hydroxyvitamin D2 [20(OH)D2], the
CYP27B1 is the mitochondrial cytochrome P450 that catalyzes
the final step of vitamin D activation, the 1a-hydroxylation of 25-
hydroxyvitamin D3 [25(OH)D3], to give 1,25-dihydroxyvitamin
D3 [1,25(OH)2D3]. 1,25(OH)2D3 maintains calcium and phosphorus
homeostasis, plays an important role in regulating the immune system
and displays antiproliferative and prodifferentiative properties on many
different cell types (Prosser and Jones, 2004; Bouillon et al., 2006;
Masuda and Jones, 2006; Takahashi and Morikawa, 2006; Holick,
2007; Bikle, 2009). Its potential for the treatment of cancer and
immune disorders is limited, however, due largely to the toxic
hypercalcemia that occurs at the supraphysiological doses required.
This work was supported by the National Institutes of Health [Grant
R01AR052190] (to A.T.S.) and [Grant 1S10RR026377-01]; the University of
Western Australia; and the College of Pharmacy at the University of Tennessee
Health Science Center.
dx.doi.org/10.1124/dmd.113.050955.
ABBREVIATIONS: E. coli, Escherichia coli; HPLC, high-performance liquid chromatography; HSQC, heteronuclear single-quantum correlation;
NMR, nuclear magnetic resonance; 17,20(OH)2D2, 17a,20-dihydroxyvitamin D2; 17,20(OH)2D3, 17a,20-dihydroxyvitamin D3; 17,20,23(OH)3D3,
17a,20,23-trihydroxyvitamin D3; 20(OH)D2, 20-hydroxyvitamin D2; 20(OH)D3, 20S-hydroxyvitamin D3; 20,23(OH)2D3, 20,23-dihydroxyvitamin D3;
20,24(OH)2D3, 20,24-dihydroxyvitamin D3; 20,25(OH)2D3, 20,25-dihydroxyvitamin D3; 20,26(OH)2D3, 20,26-dihydroxyvitamin D3; 25(OH)D3,
25-hydroxyvitamin D3; 1,20(OH)2D3, 1a,20-dihydroxyvitamin D3; 1,25(OH)2D3, 1a,25-dihydroxyvitamin D3; 1,20,23(OH)3D3, 1a,20,23-trihydroxyvi-
tamin D3; 1,20,24(OH)3D3, 1a,20,24-trihydroxyvitamin D3; 1,20,25(OH)3D3, 1a,20,25-trihydroxyvitamin D3; 1,20,26(OH)3D3, 1a,20,26-trihydroxyvita-
min D3; TOCSY, total correlation spectroscopy.
1112

Activity of CYP27B1 on CYP11A1-Derived Vitamin D Metabolites
1113
Fig. 1. Pathways for the enzymatic synthesis of vitamin D analogs tested as substrates for CYP27B1. The substrate, vitamin D3, and the major product of CYP11A1 action
on it, 20(OH)D3, are enclosed in boxes. All reactions shown by arrows without a label are catalyzed by CYP11A1. CYP27A1 and CYP24A1 also hydroxylate 20(OH)D3, as
shown in the lower part of the figure.
major product of CYP11A1 activity on vitamin D2 (Slominski et al., (25-hydroxylase; CYP27A1) and inactivation (24-hydroxylase;
2006; Nguyen et al., 2009), are the most extensively studied for CYP24A1) of vitamin D3 (Tieu et al., 2012a,b). Specifically, CYP27A1
their biologic effects. Via binding to the vitamin D receptor, these hydroxylates 20(OH)D3 at either C25 or C26, producing 20,25-
compounds can inhibit keratinocyte, melanoma, leukemia, breast, and dihydroxyvitamin D3 [20,25(OH)2D3] and 20,26-dihydroxyvitamin
liver cancer cell proliferation; promote differentiation; and display anti- D3 [20,26(OH)2D3] (Tieu et al., 2011). CYP24A1 also hydroxylates
inflammatory activity by decreasing nuclear factor-ĸB activity (Zbytek 20(OH)D3 at C25, producing 20,25(OH)2D3. Additionally, it hydroxy-
et al., 2008; Janjetovic et al., 2009, 2010, 2011; Li et al., 2010; lates at the expected C24 position, producing 20,24-dihydroxyvitamin
Slominski et al., 2010, 2011, 2012a; Tang et al., 2010a; Tuckey et al., D3 [20,24(OH)2D3] (Fig. 1) (Tieu et al., 2012b). All three of these
2011; Kim et al., 2012; Lu et al., 2012; Wang et al., 2012). Importantly, products are more potent than the parent, 20(OH)D3, at inhibiting
all three of these derivatives lack calcemic activity in rodents, even at colony formation by melanoma cells in soft agar (Tieu et al., 2012b).
relatively high doses (Slominski et al., 2010, 2013a; Wang et al., 2012).
Previously, we reported that purified mouse CYP27B1 can 1a-
However, the addition of a hydroxyl group to 20(OH)D3 at the 1a- hydroxylate the two major products of CYP11A1 action on vitamin
position (producing 1a,20-dihydroxyvitamin D3 [1,20(OH)2D3]) D3, 20(OH)D3 and 20,23(OH)2D3, and one of the major products of
causes the appearance of some calcemic activity, although less than CYP11A1 action on vitamin D2, 20(OH)D2, with the products
that for 1,25(OH)2D3 (Slominski et al., 2010). The therapeutic displaying altered biological activity compared with the parent
potential of the noncalcemic, CYP11A1-derived secosteroids is secosteroids (Tang et al., 2010a,b, 2012; Slominski et al., 2011).
illustrated by the marked reduction in collagen synthesis and clinical Recently, we successfully expressed human CYP27B1 in Escherichia
signs of bleomycin-induced scleroderma in mice following 20(OH)D3 coli (E. coli), extracted and partially purified the enzyme, and carried
treatment (Slominski et al., 2013a).
out a detailed characterization of its catalytic activity on its classic
Recently, we have shown that 20(OH)D3 can be hydroxylated substrates (Tang et al., 2012). To further characterize CYP27B1,
by cytochrome P450s that are involved in the preliminary activation especially the human isoform, and to further characterize the ability of

1114
Tang et al.
(3-[(3-Cholamidopropyl)dimethylammonio]-1-propanesulfonate) and purified by
nickel affinity and octyl Sepharose chromatography (Tang et al., 2010b, 2012).
20(OH)D3, 22-hydroxyvitamin D3 [22(OH)D3], 20,22-dihydroxyvitamin
D3 [20,22(OH)2D3], 20,23(OH)2D3, 17a,20-dihydroxyvitamin D3 [17,20(OH)2D3],
and 17,20,23(OH)3D3 were generated by the action of CYP11A1 on vitamin
D3 solubilized in 2-hydroxypropyl-b-cyclodextrin and purified by thin-layer
chromatography and high-performance liquid chromatography (HPLC)
(Guryev et al., 2003; Tuckey et al., 2008a, 2011). 20,24(OH)2D3 and 20,25(OH)2D3
the CYP11A1-derived secosteroids to be metabolized by cytochromes
P450, we examined the ability of both human and mouse CYP27B1 to
metabolize these secosteroids, including those further hydroxylated by
CYP27A1 and CYP24A1 (Fig. 1).
Materials and Methods
Preparation of Enzymes and Vitamin D Analogs. Mouse and human were produced by the action of rat CYP24A1 on 20(OH)D3 (Tieu et al., 2012b),
adrenodoxin and human adrenodoxin reductase were expressed in E. coli and
purified as before (Woods et al., 1998; Tuckey and Sadleir, 1999; Tang et al.,
2010b, 2012). Mouse and human CYP27B1 were coexpressed in E. coli with
whereas 20,26(OH)2D3 was generated by the action of human CYP27A1 on
20(OH)D3 (Tieu et al., 2012b).
Measurement of Human CYP27B1 Activity in Phospholipid Vesicles.
the chaperonins, GroEL/ES, to facilitate correct protein folding, as described Phospholipid vesicles comprising dioleoyl phosphatidylcholine (Sigma-
before (Tang et al., 2010b, 2012). The cDNA constructs for mouse and human
CYP27B1 encoded a 4- and 6-histidine tag, respectively, at the C terminus,
Aldrich, St. Louis, MO), bovine heart cardiolipin (Sigma-Aldrich), and the
vitamin D3 analogs under study were prepared by sonication in a bath-type
and the N-terminal mitochondrial-targeting sequences were removed. The sonicator as described in detail before (Lambeth et al., 1982; Tuckey and
expressed mouse and human CYP27B1 were extracted using CHAPS detergent Kamin, 1982; Tuckey et al., 2008b). The incubation mixture comprised vesicles
Fig. 2. Chromatograms showing metabolism of 25(OH)D3 (A), 20,22(OH)2D3 (B), 20,25(OH)2D3 (C), and 17,20,23(OH)3D3 (D) incorporated into phospholipid vesicles
by human CYP27B1. Small unilamellar phospholipid vesicles were prepared with 85% dioleoyl phosphatidylcholine (DOPC) and 15% cardiolipin with the substrates at
a ratio of 0.02 mol/mol phospholipid. CYP27B1 (0.1 mM) was added to the vesicles, and reactions were carried out in the presence of human adrenodoxin reductase (0.4
mM) and adrenodoxin (15 mM) for 1 hour at 37°C. The secosteroids were extracted with dicholoromethane and analyzed by reverse-phase HPLC on a C18 column. In
control incubations, adrenodoxin was omitted (top panels).

Activity of CYP27B1 on CYP11A1-Derived Vitamin D Metabolites
1115
(510 mM phospholipid), CYP27B1 (5–800 nM), adrenodoxin (15 mM), soft agar, as described in detail before (Slominski et al., 2012a; Tieu et al.,
adrenodoxin reductase (0.4 mM), glucose 6-phosphate (2 mM), glucose-6- 2012b). Secosteroids were added in 100 ml of media to final concentrations of
phosphate dehydrogenase (2 U/ml), and NADPH (50 mM) in the same buffer 0.1 or 10 nM. Each condition was tested in quadruplicates. Cells were grown at
used for vesicle preparation (20 mM HEPES, pH 7.4, 100 mM NaCl, 0.1 mM
dithiothreitol, 0.1 mM EDTA). Samples (typically 0.25–2.0 ml) were pre-
incubated for 4 minutes at 37°C, and then the reaction was initiated by the
addition of adrenodoxin. Samples were incubated at 37°C with shaking for
2 minutes (unless otherwise stated), and then reactions were terminated by the
addition of 2 volumes of ice-cold dichloromethane and vortexing (Guryev
et al., 2003). After phase separation by centrifugation, the lower organic phase
was retained, and the upper aqueous phase was extracted twice more with two
volumes of dichloromethane. The dichloromethane was subsequently removed
under nitrogen, and the residual sample was dissolved in 64 or 75% methanol
in water or 100% ethanol for analysis. Product formation was measured
by reverse-phase HPLC using either a PerkinElmer Biocompatible Binary
Pump 250/TurboChrom (PerkinElmer, Waltham, MA) or HP Agilent 1050/
ChemStation system with a GraceSmart C18 column (4.6 Â 150 mm, particle
size 5 mm; Agilent Technologies, Palo Alto, CA) as described before (Tang
et al., 2010b).
37°C with 5% CO2 over 2 weeks, with secosteroids in fresh media (100 ml)
being added after every 72 hours. Colonies were then scored and stained with
methylthiazol-tetrazolium reagent (Promega, Madison, WI), and then counted
under the microscope. Data were analyzed with Student’s t test (for two groups)
using Prism 4.0 (GraphPad Software, San Diego, CA).
Analysis of Kinetics Data. The experimental data were fitted to the
Michaelis-Menten equation. If substrate inhibition was observed, data were also
fitted to the inhibition model equation that we have previously described (Tang
et al., 2010b), using Kaleidagraph 4.1 (Abelbeck Synergy Software, Reading,
PA). This equation is derived from a model where two molecules of substrate
bind to the enzyme substrate complex simultaneously with equal Ki values to
form an ES3 (enzyme bound to 3 molecules of substrate) complex that is
catalytically inactive. As well as the Km and kcat values, this equation also
provides the inhibitor constant for substrate (Ki).
Large-Scale Preparation of 1a-Hydroxylated Metabolites. To produce
metabolites on a scale that permitted structure determination by nuclear mag-
netic resonance (NMR) (more than 30 mg), mouse CYP27B1 (0.3 mM) was
incubated with 20,24(OH)2D3, 20,25(OH)2D3, or 20,26(OH)2D3 incorporated
into phospholipid vesicles (0.025 or 0.03 mol/mol phospholipid) in 10–25-ml
incubations comprising vesicle preparation buffer (as described earlier), 15 mM
mouse adrenodoxin, 0.4 mM human adrenodoxin reductase, 2 mM glucose-6-
phosphate, 2 U/ml glucose-6-phosphate dehydrogenase, and 50 mM NADPH.
The reactions were performed for 1 hour at 25°C with shaking. Reactions were
stopped with 2 volumes of ice-cold dichloromethane, and products were ex-
tracted in a scaled-up version of that described earlier. The products were
purified by reverse-phase HPLC using a GraceSmart column (PerkinElmer
Biocompatible Binary Pump/TurboChrom system) with a gradient of 64%
methanol in water at a flow rate of 0.5 ml/min for 10 minutes (Guryev et al.,
2003; Tuckey et al., 2008a). The yield of products, subsequently shown
to be the expected 1,20,24-trihydroxyvitamin D3 [1,20,24(OH)3D3], 1,20,25-
trihydroxyvitamin D3 [1,20,25(OH)3D3], and 1,20,26-trihydroxyvitamin D3
[1,20,26(OH)3D3] (see Results), was determined spectrophotometrically at 263
nm using an extinction coefficient of 18,000 M-1cm21× (Hiwatashi et al., 1982).
Mass Spectroscopy. Mass spectra were acquired in a Bruker Esquire-Liquid
Chromatography/Mass Spectrometry system (Bruker Daltonics, Billerica, MA),
using the ionization source of electrospray ionization with nitrogen as the
nebulizing gas. Data were collected and processed by an ACD mass processor
(Advanced Chemistry Development, Toronto, ON, Canada).
NMR Spectroscopy. NMR measurements were performed using an inverse
triple-resonance 3-mm probe on a Varian Unity Inova 500 MHz spectrometer
(Agilent Technologies). The sample was dissolved in methanol-d4 and transferred
to a 3-mm Shigemi NMR tube (Shigemi Inc., Allison Park, PA). Temperature was
regulated at 22°C, and was controlled with an accuracy of 60.1°C. Chemical
shifts were referenced to residual solvent peaks for CD3OD (3.31 ppm for proton
and 49.15 ppm for carbon). Standard two-dimensional NMR experiments [1H-1H
correlation spectroscopy, 1H-1H total correlation spectroscopy (TOCSY;
Fig. 3. The effect of substrate concentration on the rate of hydroxylation of different
substrates by human CYP27B1. Reaction rates were determined for 20,26(OH)2D3
(top), 20,24(OH)2D3 (middle), and 20(OH)D3 (bottom) using substrates incor-
porated into phospholipid (PL) vesicles prepared from dioleoyl phosphatidylcholine
(DOPC) and cardiolipin. CYP27B1 was incubated with the vesicle-associated
substrates for 2 minutes in the presence of adrenodoxin and adrenodoxin reductase.
Substrates and products were extracted and analyzed by reverse-phase HPLC. Data
for high substrate concentration where inhibition was observed were excluded from
the Michaelis-Menten curve fits (solid lines). Data for 20,26(OH)2D3 were also
fitted to a model describing substrate inhibition as described in Materials and
Methods (dashed line).
mixing time
= C heteronuclear single-quantum
80 milliseconds), 1H-13
correlation spectroscopy (HSQC), and 1H-13C heteronuclear multiple-bond
correlation spectroscopy] were conducted to fully elucidate the structures of the
metabolites. All data were processed using ACD software, with zero filling in
the direct dimension and linear prediction in the indirect dimension.
Measurement of Cell Proliferation in Soft Agar. The effect of the
secosteroids hydroxylated by CYP27B1 on the tumorigenicity of SKMEL-188
melanoma cells was determined by assaying their ability to form colonies in

1116
Tang et al.
TABLE 1
Kinetic parameters for metabolism of various substrates incorporated into phospholipid vesicles by mouse and human CYP27B1
Activity of human and mouse CYP27B1 was determined toward various substrates incorporated into small unilamellar phospholipid (PL) vesicles comprising 15 mol% cardiolipin and 85 mol%
dioleoyl phosphatidylcholine (DOPC). Data are the mean 6 S.E. of the curve fit for representative experiments. Kinetic parameters were determined from fitting the Michaelis-Menten equation to
the data.
Mouse
kcat
Human
kcat
Substrate
Km
kcat/Km
Km
kcat/Km
mmol/mol PL
min21
mmol/mol PL
min21
25(OH)D3a,b
5.9 6 1.2
49 6 21
41 6 3
8.5 6 1.7
6.9 6 1.9
0.17 6 0.11
5.4 6 0.8
19.5 6 3.0
55.4 6 4.4
4.7 6 0.2
10.3 6 2.3
0.24 6 0.05
20(OH)D3b
22(OH)D3
Low activity detected
No activity detected
1.8 6 0.2
8.1 6 0.3
No activity detected
39.5 6 1.3
Low activity detected
No activity detected
Low activity detected
4.1 6 0.2
No activity detected
41.3 6 0.7
17,20(OH)2D3
20,22(OH)2D3c
20,23(OH)2D3
17,20,23(OH)3D3
20,24(OH)2D3c
20,25(OH)2D3c
20,26(OH)2D3c
53.2 6 8.4
12.1 6 1.6
0.03 6 0.01
0.67 6 0.12
9.7 6 1.8
0.43 6 0.10
1.4 6 0.2
3.4 6 0.3
2.9 6 0.6
28.7 6 5.1
18.6 6 2.1
20.9 6 5.4
1.4 6 0.1
2.2 6 0.4
3.9 6 0.3
29.1 6 2.8
29.6 6 6.5
18.3 6 1.9
63.9 6 1.6
60.7 6 3.2
63.7 6 3.0
72.1 6 1.8
a
Compounds where substrate inhibition was observed for mouse and human CYP27B1.
Kinetic parameters from previously published data (Tang et al., 2010b, 2012), except for human CYP27B1 with 20(OH)D3.
Compounds where substrate inhibition was observed for mouse CYP27B1 only.
b
c
Other Procedures. The concentration of CYP27B1 was determined from
the CO-reduced minus reduced difference spectrum using an extinction coefficient
of 91,000 M-1cm21 for the absorbance difference between 450 and 490 nm
(Omura and Sato, 1964). The concentrations of all hydroxyvitamin D3 stock
solutions were determined using an extinction coefficient of 18,000 M-1cm21 at
263 nm (Hiwatashi et al., 1982).
or CYP27A1 on 20(OH)D3 [20,24(OH)2D3, 20,25(OH)2D3, and
20,26(OH)2D3]. Products were identified from their retention times
compared with authentic standards in the case of 1,20(OH)2D3 and
1a,20,23-trihydroxyvitamin D3 [1,20,23(OH)3D3] (Tuckey et al.,
2008a; Tang et al., 2010a), and by mass spectrometry and NMR in
the case of the products from 20,24(OH)2D3, 20,25(OH)2D3, and
20,26(OH)2D3 (described below). For 22(OH)D3 and 20,22(OH)2D3,
where limited substrate was available and the hydroxylation rate was
Results
Kinetics of the Metabolism of CYP11A1-Derived Vitamin D3 low, insufficient product was generated by CYP27B1 to identify
Analogs by CYP27B1. CYP27B1 activity was measured with sub- their structures by NMR, but it is presumed that they are the 1a-
strates incorporated into phospholipid vesicles, a system that mimics hydroxy derivatives. For all substrates where the metabolite was
the native environment of the cytochrome in the inner mitochondrial identified, there was a single 1a-hydroxylated product with both the
membrane, which we have used previously with CYP27B1 (Tang mouse and human enzymes. This is illustrated in Fig. 2 for 25(OH)D3,
et al., 2010b, 2012). Substrates tested included the primary pro- 20,22(OH)2D3, and 20,25(OH)2D3. No products were detected for
ducts of CYP11A1 action of vitamin D3 [20(OH)D3, 22(OH)D3, 17,20(OH)2D3 or 17,20,23(OH)3D3 (Fig. 2D), indicating that the pre-
20,22(OH)2D3, 20,23(OH)2D3, 17,20(OH)2D3, and 17,20,23(OH)3D3], sence of the 17a-hydroxyl group prevented hydroxylation by CYP27B1.
as well as secondary products generated by the action of CYP24A1 This was also the case for the metabolism of 17a,20-dihydroxyvitamin
Fig. 4. Chromatograms illustrating the difference in CYP27B1 activity toward 20(OH)D3 (A) and 22(OH)D3 (B). Substrates were incorporated into phospholipid vesicles at
a ratio of 0.025 mol/mol phospholipid and incubated with mouse CYP27B1 (0.8 mM) in the presence of adrenodoxin reductase and mouse adrenodoxin for 1 hour at 37°C.
The secosteroids were extracted with dicholoromethane and analyzed by reverse-phase HPLC on a C18 column. Control incubations where adrenodoxin was omitted are also
shown (top panels).

Activity of CYP27B1 on CYP11A1-Derived Vitamin D Metabolites
1117
observed for 20,26(OH)2D3 (as shown in Fig. 3 for the human
enzyme), giving a Ki value of 133 mmol/mol phospholipid, which is
30 times higher than its Km. High concentrations of 20,24(OH)2D3
inhibited the mouse enzyme, giving a Ki value of 241 mmol/mol
phospholipid, which is 166-fold higher than its Km. 20,22(OH)2D3
also showed substrate inhibition with mouse CYP27B1 (Ki = 37.4
mmol/mol phospholipid), but due to low activity, kinetic parameters
were not measured for the human enzyme. In each case where
inhibition was observed, the data fit well to a one catalytic site/two
inhibitory site model (see Materials and Methods). Since the Ki values
were very high relative to the Km values, and the Km and kcat values
calculated from a Michaelis-Menten curve fit (using data at lower
substrate concentrations) were indistinguishable from those deter-
mined with the inhibition model, only the Michaelis-Menten data are
shown (Table 1). 20(OH)D3 was metabolized relatively poorly by
human and mouse CYP27B1, displaying higher Km and lower kcat
values than for 25(OH)D3. 22(OH)D3 was an even poorer substrate
for both mouse and human CYP27B1, as shown in a 1-hour incubation
of mouse CYP27B1 with this secosteroid compared with 20(OH)D3
(Fig. 4). The low activity with 22(OH)D3 and the limited amount of
this secosteroid available prevented a full kinetic analysis from being
carried out.
When the vitamin D3 side chain contained both 20- and 22-
hydroxyl groups [as in 20,22(OH)2D3], activity was also low, with the
kcat being 5-fold lower than that for 20(OH)D3 for the mouse enzyme,
and too low to quantitate for the human enzyme. Movement of
the second hydroxyl group on the side chain to the 23 position
[20,23(OH)2D3] markedly increased the kcat/Km for both the mouse
and human enzymes, with values being higher than those for 20(OH)D3.
Shifting the second hydroxyl group to C24, C25, or C26 dra-
matically improved the ability of CYP27B1 to 1a-hydroxylate the
analog, with Km values being similar to or lower than those for
25(OH)D3 and the kcat/Km values being higher. The highest kcat/Km
value for any of the substrates analyzed with the mouse enzyme
was for 20,24(OH)2D3, and for the human enzyme was equal for
20,24(OH)2D3 and 20,25(OH)2D3. Generally, values for both kcat and
Fig. 5. Analysis of the ability of 17,20(OH)2D3 to inhibit CYP27B1 activity. (A)
Mouse CYP27B1 or (B) human CYP27B1was incubated with 25(OH)D3 at a concen-
tration of 0.025 mol/mol phospholipid in vesicles for 2 minutes with 17,20(OH)2D3
present at the ratios indicated. Data represent the means 6 S.E.M. (n = 3), and were
analyzed using the Student’s t test on GraphPad Prism. *P , 0.05; **P , 0.01;
***P , 0.001.
D2 [17,20(OH)2D2] by CYP27B1, where in contrast to 20(OH)D2 Km were reasonably similar between the mouse and human enzymes,
(Tang et al., 2010b), no hydroxylation was observed (unpublished with the largest differences being for 20(OH)D3, where the human
data).
enzyme showed both kcat and Km values approximately half those of
We have previously reported that inhibition of both mouse and the mouse enzyme.
human CYP27B1 occurs with high concentrations of 25(OH)D3 (Tang
17,20(OH)2D3 Is a Weak Inhibitor of Human CYP27B1 Activity.
et al., 2010b, 2012). In the current study, no inhibition of mouse or Our results show that both human and mouse CYP27B1 are inactive
human CYP27B1 activity was observed at high concentrations of toward 17,20(OH)2D3 and other 17a-hydroxylated secosteroids
20(OH)D3 or 20,23(OH)2D3, but weak substrate inhibition was (Table 1). This could be due to very poor binding of the analogs to
Fig. 6. Mass spectra with electrospray ionization of product from CYP27B1 action on 20,24(OH)2D3 (A), 20,25(OH)2D3 (B), and 20,26(OH)2D3 (C).

1118
Tang et al.
Fig. 7. NMR spectra of 1,20,24(OH)3D3. (A) One-dimensional proton, (B) 1H-13C HSQC, and (C) 1H-1H TOCSY.
the active site of CYP27B1, or binding in a position that prevents
Determination of the Structure of Products Resulting from
hydroxylation. To distinguish between these possibilities, we tested the Action of CYP27B1 on 20,24(OH)2D3, 20,25(OH)2D3, and
the ability of 17,20(OH)2D3 to inhibit the 1a-hydroxylation of 20,26(OH)2D3. Mass spectrometry confirmed that the metabolites
25(OH)D3. Significant inhibition of human CYP27B1 activity was generated by CYP27B1 action on 20,24(OH)2D3, 20,25(OH)2D3,
seen at 17,20(OH)2D3:25(OH)D3 ratios of 0.5:1, 1:1, and 2:1, but no and 20,26(OH)2D3 were all trihydroxy derivatives. The observed
inhibition was observed for the mouse enzyme at these ratios (Fig. 5). molecular ion for each compound had a mass of 455 [M + Na]+,
This suggests that some binding of 17,20(OH)2D3 to the human enzyme giving a molecular weight of 432 (Fig. 6). The site of hydroxylation
occurs, permitting it to act as an inhibitor, presumably competitive. on the three compounds by CYP27B1 was unambiguously assigned to
Insufficient 17,20(OH)2D3 was available to permit a full analysis of the be at the 1a-position based on the NMR spectra for these metabolites.
type of inhibition or the Ki.
First, for the product of 20,24(OH)2D3 and 20,25(OH)2D3, none of
Synthesis of CYP27B1 Metabolites of 20,24(OH)2D3, the four methyl groups (18, 21, 26, and 27) are hydroxylated based on
1
20,25(OH)2D3, and 20,26(OH)2D3 for Structure Determination. 1H NMR (Figs. 7A and 8A). H-13C HSQC revealed the presence of
Mouse CYP27B1 was used to scale up reactions to produce sufficient a new methine group at 4.35 or 4.36 ppm (13C at 71.3 ppm; Figs. 7B
products from hydroxylation of 20,24(OH)2D3, 20,25(OH)2D3, and and 8B). 1H-1H TOCSY (Figs. 7C and 8C) demonstrated that the new
20,26(OH)2D3 to enable structure determination by mass spectrometry methine at 4.35 or 4.36 ppm is in the same spin system as that of
and NMR. Complete conversion of 20,24(OH)2D3, 20,25(OH)2D3, 2-CH2 (1H at 1.89 ppm), 3-CH (1H at 4.13 ppm), and 4-CH2 (2.53 and
and 20,26(OH)2D3 was achieved when 0.3 mM mouse CYP27B1 was 2.27 ppm), indicating the hydroxylations occurred in the 1 position.
incubated with the substrates in phospholipid vesicles at a concentra-
With 20,26(OH)2D3, similar to the product of the other two
tion of 0.025 or 0.03 mol/mol phospholipid in a 1-hour incubation. dihydroxy derivatives, C18, C21, and C27 were not hydroxylated
1
1
The major limitation in producing the metabolites was the availability based on H NMR (Fig. 9B). H-13C HSQC revealed the presence of
of substrates which required enzymatic synthesis of 20(OH)D3 by a new methine group at 4.36 ppm (13C at 71.3 ppm; Fig. 9B). H-1H
1
CYP11A1 and its subsequent hydroxylation by purified CYP27A1 or TOCSY (Fig. 9C) showed that the methine at 4.36 ppm has
CYP24A1 (see Materials and Methods). Overall, 30–60 mg of purified correlations to 2-CH2 (1H at 1.89 ppm), 3-CH (1H at 4.13 ppm),
products were prepared for each compound, which was adequate for and 4-CH2 (2.52 and 2.26 ppm), indicating the hydroxylation occurred
structure determination.
in the 1 position. Overall, the previous analysis shows that the

Activity of CYP27B1 on CYP11A1-Derived Vitamin D Metabolites
1119
Fig. 8. NMR spectra of 1,20,25(OH)3D3. (A) 1D proton, (B) 1H-13C HSQC, and (C) 1H-1H TOCSY.
hydroxylation sites can be assigned to the 1 position unambiguously. et al., 2011; Szyszka et al., 2012), a deadly disease for which
The full assignments for the CYP27B1-hydroxylated metabolites and therapy is still unsatisfactory (Hauschild et al., 2012). Therefore,
the dihydroxy derivatives are summarized in Tables 2–4.
we tested whether the 1a-hydroxylation of 20,25(OH)2D3 and
The new hydroxyl group at C1 is in the 1a configuration (i.e., the 20,26(OH)2D3, both good substrates for CYP27B1, altered their
remaining proton attached to C1 has 1b configuration) for all three ability to inhibit melanoma colony formation in soft agar. This assay
CYP27B1-derived metabolites based on the analysis of the chemical shift was previously used to characterize the parent compounds and is
and 1H-1H coupling constants between protons attached to C1 and C2, as a good measure of tumorigenicity. Although 20,24(OH)2D3 is also an
described in the following text using 1,20,26(OH)3D3 as an example. excellent substrate for CYP27B1, this had not been synthesized at the
First, the proton chemical shift (4.35 ppm) at C1 (Fig. 10) is very similar time of testing.
to that of the 1b-H (4.29 ppm) in 1,25(OH)2D3 measured in the same
The parent compounds and their 1a-hydroxy derivatives at 10 nM
solvent (CD3OD) (Eguchi and Ikekawa, 1990). Second, the vicinal significantly inhibited colony formation compared with the vehicle
1H-1H coupling constant (3JH-1, H-2a = 3JH-1, H-2b = 5.9 Hz) between the control (P , 0.0002). Excluding 1,20,26(OH)3D3, they also in-
proton at C1 and the two protons at C2 of this metabolite is essentially hibited colony formation significantly more than 20(OH)D3 (Tieu
3
the same as that in 1,25(OH)2D3 (3JH-1b, H-2a = JH-1b, H-2b = 6.0 Hz) et al., 2012b). These significant differences were also seen when the
(Eguchi and Ikekawa, 1990). This pseudo-triplet splitting pattern is secosteroid concentration was decreased to 0.1 nM (Fig. 11). At 0.1 nM,
only possible when the proton directly attached to C1 is in the 1b 1,20,25(OH)3D3 caused significantly greater inhibition of colony for-
configuration due to similar vicinal coupling constants between H-1b mation than 20,25(OH)2D3, whereas at 10 nM, 1,20,26(OH)3D3 caused
and H-2a (3Jee) or H-2b (3Jea). If the proton directly attached to C1 was significantly less inhibition than 20,26(OH)2D3.
in the 1a configuration, the values of the two vicinal coupling constants
would be very different (3Jaaꢀ11Hz, Jaeꢀ 6Hz), resulting in a clear
3
Discussion
doublet of doublet instead of a pseudo-triplet as observed. Taken
together, the previous analysis shows that the hydroxyl group at C1 must
be in the 1a configuration for all three CYP27B1-derived products.
The ability of CYP11A1 to hydroxylate the vitamin D3 side chain
at C20 and C22 (Tuckey et al., 2011) has enabled us to look at the
The Effect of 1a-hydroxylation on Antimelanoma Activity. effect of the position of the side chain hydroxyl group on CYP27B1
Targeting of the vitamin D receptor represents a promising strategy for activity. The current study shows that when the hydroxyl group on
_
treatment of melanoma (Pinczewski and Slominski, 2010; Brozyna the vitamin D side chain is moved from C25 to C20, the catalytic

1120
Tang et al.
Fig. 9. NMR spectra of 1,20,26(OH)3D3. (A) 1D proton, (B) 1H-13C HSQC, and (C) 1H-1H TOCSY.
efficiency (kcat/Km) of the 1a-hydroxylase decreases by 40- to 45-fold, addition of a second hydroxyl group on the side chain close to the “D”
due to both an increase in Km and a decrease in kcat. When the side ring, for example, at C22, gave a very low kcat/Km value for the mouse
chain hydroxyl group is at C22, rather than at C20, the activity is even enzyme, lower than that for 20(OH)D3, and too low to measure for the
lower, not permitting measurement of its kinetic parameters.
human enzyme. When the second hydroxyl group is in the 17a
Recently, we reported that 20(OH)D3 can be hydroxylated by two position of the D ring [17,20(OH)2D3], no activity was observed even
vitamin D–metabolizing enzymes, human CYP27A1 (Tieu et al., at high CYP27B1 concentrations. The ability of the 17a-hydroxyl
2012a) and rat CYP24A1 (Tieu et al., 2012b). The main products group to completely inhibit CYP27B1 activity was also seen with the
of CYP24A1 action on 20(OH)D3 are 20,24(OH)2D3 and 20,25(OH)2D3, trihydroxy derivative, 17,20,23(OH)3D3, and the vitamin D2 de-
and for CYP27A1 are 20,25(OH)2D3 and 20,26(OH)2D3. Along rivative, 17,20(OH)2D2. We observed that 17,20(OH)2D3 could
with the primary products of CYP11A1 action on vitamin D3, inhibit 25(OH)D3 metabolism by human CYP27B1, but not the
17,20(OH)2D3, 20,22(OH)2D3, and 20,23(OH)2D3, we have a unique mouse enzyme, suggesting that it can bind to the active site of the
set of derivatives all containing a 20-hydroxyl group plus one ad- human enzyme, but in an orientation that precludes hydroxylation.
ditional hydroxyl group at every position on the side chain, except
Overall, the kinetic data presented in this study indicate that the
C21 (note that C26 and C27 are equivalent). All of these 20-hydroxy position of one or more hydroxyl groups on the vitamin D side chain
derivatives have been shown to be biologically active (Zbytek et al., dramatically influences the ability of both mouse and human
2008; Janjetovic et al., 2009, 2010, 2011; Li et al., 2010; Slominski CYP27B1 to hydroxylate the derivative. Yamamoto et al. (2004,
et al., 2010, 2011, 2012a; Tang et al., 2010a; Tuckey et al., 2011; Kim 2005) prepared a modeled structure of CYP27B1 and used muta-
et al., 2012; Lu et al., 2012; Wang et al., 2012). The low catalytic genesis to conclude that Thr409 of the human CYP27B1 hydrogen
efficiency of CYP27B1 toward 20(OH)D3 is greatly enhanced when bonds to the 25-hydroxyl group of 25(OH)D3, playing a critical role
the second hydroxyl group is added toward the end of the side chain, in substrate binding. Since we observed that the Km values for
as seen for 24,20(OH)2D3, 20,25(OH)2D3, and 20,26(OH)2D3. All 20,24(OH)2D3, 20,25(OH)2D3, and 20,26(OH)2D3 were comparable
of these compounds displayed higher kcat/Km values than that for to or lower than that for 25(OH)D3, it would seem likely that a hy-
25(OH)D3, primarily due to their lower Km values. Conversely, the droxyl group at C24 or C26, such as at C25, can also hydrogen bond to

Activity of CYP27B1 on CYP11A1-Derived Vitamin D Metabolites
1121
TABLE 2
NMR chemical shift assignments for 1,20,24(OH)3D3 analysis of its two-dimensional NMR (solvent: methanol-d4)
The assignments are compared with that for its dihydroxy derivative prior to hydroxylation by CYP27B1.
20,24(OH)2D3 1,20,24(OH)3D3
C Atom
Structure
1H
13C
1H
13C
1
2
3
4
5
6
7
8
2.11a, 2.41b
1.96a, 1.53b
3.76
2.52a, 2.18b
NA
6.22
6.03
NA
1.68a, 2.85b
NA
1.55a, 1.69b
1.36a, 2.10b
NA
2.01
1.77
1.68a, 1.77b
1.69
0.7
4.75, 5.04
NA
1.24
33.5
36.5
70.4
46.8
136.4
122.5
119.3
141.1
28.4
146.2
24.2
42.3
45.4
57.6
23.1
23
60.4
14.1
112.4
74.7
25.9
39.6
41
4.36
1.89
4.13
2.52a, 2.27b
NA
71.7
44
67.7
46.4
136.4
125.2
119.6
NI
6.32
6.08
NA
9
1.69a, 2.86b
NA
1.56a, 1.67b
1.38a, 2.11b
NA
2.01
1.49a, 1.77b
1.68
1.68
0.72
4.90, 5.29
NA
1.24
30.1
NI
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
24.8
42.6
47.3
58.2
23.3
24.6
60.8
14.4
112.3
76.3
26.2
41.4
30.5
78.6
35.3
19.6
18.3
1.36
1.74, 1.38
3.22
1.62
0.93
1.71, 1.38
1.33
3.21
1.63
0.91
78.2
33.5
19.3
18.5
0.92
0.92
NA, not applicable (ternary carbons); NI, not identifiable (due to low resolution).
Thr409. The higher Km values observed with 20(OH)D3 and Importantly in terms of their potential for therapeutic applications,
20,22(OH)2D3 suggest that their side chain hydroxyl groups are too such as in the treatment of scleroderma (Slominski et al., 2013a), it has
far from Thr409 to form this hydrogen bond. Differences in the been shown that, unlike 1,25(OH)2D3, 20(OH)D3 lacks calcemic
observed kcat values between different substrates are likely due to the activity in rodents at doses up to 30 mg/kg, whereas severe hyper-
distance of C1 from the activated oxygen bound to the heme group, or calcemia was detected with 1,25(OH)2D3 at a dose of only 2 mg/kg
steric effects of the substrate on the optimal conformation of the (Wang et al., 2012). However, when a 1a-hydroxyl group was added
enzyme for catalysis.
to 20(OH)D3 to produce 1,20(OH)2D3, moderate calcemic activity
The present work shows that the position and number of hydroxyl was observed (Slominski et al., 2010). This suggests that the presence
groups on the side chain of the vitamin D analog do not alter the site of of a 1a-hydroxyl group is necessary for the vitamin D analog to
hydroxylation by CYP27B1, which is always the C1a position. This regulate serum calcium levels. Thus, the relatively poor ability of
suggests that the A ring of vitamin D analogs must bind to the enzyme CYP27B1 to hydroxylate CYP11A1-derived hydroxyvitamin D3
active site in a specific orientation so that only the 1a-hydroxy analogs may be important in vivo for preventing them from causing
derivative is produced. Thus, CYP27B1 is highly specific for 1a- hypercalcemia. It has to be noted that both 20(OH)D3 and 20(OH)D2
hydroxylation, consistent with the limited literature available where do not require hydroxylation to express phenotypic activity (Slominski
there are no reports of this enzyme hydroxylating at the C1b position et al., 2011). Nevertheless, there is evidence for at least some 1a-
or at adjacent carbon atoms. However, low 25-hydroxylase activity of hydroxylation of 20(OH)D3 and 20,23(OH)2D3 in vivo, as their 1a-
CYP27B1 has been observed when 1a-hydroxyvitamin D3 is used as hydroxylated products have been detected in placental explants
substrate, requiring binding of the secosteroid to the active site of the incubated with vitamin D3 and in cultured keratinocytes (Slominski
enzyme in the opposite orientation to that for 1a-hydroxylation et al., 2012b). The CYP11A1-derived secosteroids, 17,20(OH)2D2,
(Uchida et al., 2004; Tang et al., 2010b).
17,20(OH)2D3, and 17,20,23(OH)3D3, may provide some therapeutic
It has been documented that the CYP11A1-derived vitamin D3 advantage over the other CYP11A1-derived secosteroids because
analogs, 20(OH)D3 and 20,23(OH)2D3, display similar properties to of the inability of CYP27B1 to hydroxylate them and, therefore, alter
1,25(OH)2D3, such as inhibiting cell proliferation, promoting dif- their properties.
ferentiation of keratinocytes and leukemia cells, and suppressing
In this study, we performed an initial investigation of the anti-
tumorigenicity of melanoma (Zbytek et al., 2008; Janjetovic et al., proliferative activity of two of the new trihydroxy derivatives
2009, 2011; Li et al., 2010; Slominski et al., 2010, 2012a; Tuckey generated with high catalytic efficiency by CYP27B1, 1,20,25(OH)3D3
et al., 2011; Kim et al., 2012; Lu et al., 2012; Wang et al., 2012). and 1,20,26(OH)3D3, and compared this to the activity of their parent

1122
Tang et al.
TABLE 3
NMR chemical shift assignments for 1,20,25(OH)3D3 analysis of its two-dimensional NMR (solvent: methanol-d4)
The assignments are compared with that for its dihydroxy derivative prior to hydroxylation by CYP27B1.
20,25(OH)2D3 1,20,25(OH)3D3
C Atom
Structure
1H
13C
1H
13C
1
2
3
4
5
6
7
8
2.12a, 2.41b
1.97a, 1.54b
3.76
2.54a, 2.19b
NA
6.22
6.03
NA
1.69a, 2.86b
NA
1.54a, 1.68b
1.39a, 2.10b
NA
2.02
1.78
1.68a, 1.56b
1.7
0.7
4.74, 5.04
NA
1.26
33.5
36.6
70.5
47
4.35
1.89
4.13
2.53a, 2.27b
NA
71.3
43.6
67.2
46
135.4
124.8
119.2
NI
29.8
149.4
24.2
42
46.9
57.7
22.8
22.7
59.8
14
111.8
75.9
25.9
45.1
19.8
45.2
71.1
29.1
29.1
136
122.6
119.4
141.1
29.7
145.9
22.7
42.2
45.4
57.8
22.9
24.4
59.9
14
112.64
74.4
26.1
45.4
19.9
45.5
70
6.33
6.08
NA
9
1.69a, 2.87b
NA
1.58a, 1.69b
1.38a, 2.11b
NA
2
1.77
1.70, 1.48
1.7
0.72
4.90, 5.30
NA
1.26
1.47, 1.37
1.41
1.41
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
1.48, 1.35
1.4
1.41
NA
NA
1.19
1.19
29.2
29.2
1.18
1.18
NA, not applicable (ternary carbons); NI, not identifiable (due to low resolution).
secosteroids, 20,25(OH)2D3 and 20,26(OH)2D3. As reported before 2013a,b). This will be complemented by testing the toxicity of these
(Tieu et al., 2012b), the parent secosteroids displayed a significantly modified compounds.
higher inhibitory effect on colony formation by melanoma cells grown
In conclusion, this study clearly defines the kinetic parameters for
in soft agar than either 20(OH)D3 or 1,25(OH)2D3 at concentrations the action of CYP27B1 on a range of secosteroids produced by
of both 0.1 and 10 nM. Interestingly, the antiproliferative activity of CYP11A1 in a vesicle-reconstituted system that resembles the inner
1,20,25(OH)3D3 and 1,20,26(OH)3D3 on melanoma cells (SKMEL- mitochondrial membrane. All of the secosteroids produced directly
188) was found to be significantly different from that of their parent by CYP11A1 action on vitamin D3 are relatively poor substrates for
compounds, with a greater effect seen for 1,20,25(OH)3D3 and CYP27B1. In contrast, the secondary metabolites produced by the
a lesser effect for 1,20,26(OH)3D3 [although still significantly more action of CYP27A1 or CYP24A1 on 20(OH)D3, which have a hydroxyl
than for 20(OH)D3]. The increased potency resulting from the 1a- group close to the end of the side chain, are very good substrates for
hydroxylation of 20,25(OH)2D3, with the product completely inhibit- CYP27B1.
ing colony formation at 10 nM and inhibiting by 66% at 0.1 nM,
suggests that it is a good candidate for further testing of its
Acknowledgments
The authors thank Alex Hua for help in preparing the figures.
antimelanoma activity. The previous data are also consistent with
recent clinicopathologic analyses that clearly demonstrate a decrease
in CYP27B1 expression during progression of melanoma, as well as
Authorship Contributions
Participated in research design: Tang, Tuckey, Slominski, Li.
with the positive correlation between shorter survival of patients and
Conducted experiments: Tang, Janjetovic, Chen.
low CYP27B1 expression in melanoma specimens (Broz_yna et al.,
Contributed new reagents or analytic tools: Tieu.
2013). Thus, the current and previous studies suggest that CYP27A1
Performed data analysis: Tang, Janjetovic, Chen, Tuckey, Slominski, Li.
Wrote or contributed to the writing of the manuscript: Tang, Tuckey,
Slominski, Chen, Li.
and CYP24A1 can increase the antimelanoma potency of 20(OH)D3,
with further modification by CYP27B1. Based on our previous studies
(Kim et al., 2012; Slominski et al., 2012a), we expect that these effects
are mediated through the activation of the vitamin D receptor. How-
ever, the final proof would require additional testing using models
described previously (Kim et al., 2012), and an extended panel of
melanoma lines available in our laboratories, and comparison with
References
Bikle D (2009) Nonclassic actions of vitamin D. J Clin Endocrinol Metab 94:26–34.
Bouillon R, Eelen G, Verlinden L, Mathieu C, Carmeliet G, and Verstuyf A (2006) Vitamin D
and cancer. J Steroid Biochem Mol Biol 102:156–162.
Broz_yna AA, Jozwicki W, Janjetovic Z, and Slominski AT (2011) Expression of vitamin D
_
normal melanocytes (Brozyna et al., 2012; Slominski et al., 2012a,
receptor decreases during progression of pigmented skin lesions. Hum Pathol 42:618–631.

Activity of CYP27B1 on CYP11A1-Derived Vitamin D Metabolites
1123
TABLE 4
NMR chemical shift assignments for 1,20,26(OH)3D3 analysis of its two-dimensional NMR (solvent: methanol-d4)
The assignments are compared with that for its dihydroxy derivative prior to hydroxylation by CYP27B1.
20,26(OH)2D3 1,20,26(OH)3D3
C Atom
Structure
1H
13C
1H
13C
1
2
3
4
5
6
7
8
2.12a, 2.42b
1.97a, 1.53b
3.77
2.54a, 2.20b
NA
6.22
6.03
NA
1.69a, 2.86b
NA
1.56a, 1.67b
1.38a, 2.10b
NA
2.01
1.5
1.72a, 1.79b
1.68
0.7
4.75, 5.05
NA
1.25
33.4
36.3
70.5
46.8
4.36
1.89
4.13
2.52a, 2.26b
NA
71.3
43.6
67.4
46.2
135.9
124.8
119.2
NI
29.8
149.6
23.2
42.1
46.8
57.7
23.2
23.2
59.9
13.8
111.8
76
136
122.6
119.3
140.4
29.6
145.7
24.3
42.1
45.5
57.7
22.7
22.7
59.9
13.9
112.5
74.5
26
6.32
6.08
NA
9
1.69a, 2.87b
NA
1.55a, 1.67b
1.37a, 2.07b
NA
2.01
1.47
1.68a, 1.78b
1.68
0.71
4.90, 5.29
NA
1.24
1.50, 1.33
NA*
1.40,1.08
1.58
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
25.9
45.1
NI
34.8
36.6
68.2
16.8
1.34, 1.51
1.43, 1.29
1.41, 1.07
1.59
3.36
0.91
44.9
22.3
34.9
36.7
68.4
16.8
3.36
0.91
NA, not applicable (ternary carbons); NI, not identifiable (due to low resolution).
Broz_yna AA, Jóꢀzwicki W, Janjetovic Z, and Slominski AT (2013) Expression of the vitamin
D-activating enzyme 1a-hydroxylase (CYP27B1) decreases during melanoma progression.
Hum Pathol 44:374–387.
Eguchi T and Ikekawa N (1990) Conformational analysis of 1a,25-dihydroxyvitamin D3 by
nuclear magnetic resonance. Bioorg Chem 18:19–29 DOI: 10.1016/0045-2068(90)90012-T.
Guryev O, Carvalho RA, Usanov S, Gilep A, and Estabrook RW (2003) A pathway for the
metabolism of vitamin D3: unique hydroxylated metabolites formed during catalysis with
cytochrome P450scc (CYP11A1). Proc Natl Acad Sci USA 100:14754–14759.
Hauschild A, Grob J-J, Demidov LV, Jouary T, Gutzmer R, Millward M, Rutkowski P, Blank
CU, Miller WH, Jr, and Kaempgen E, et al. (2012) Dabrafenib in BRAF-mutated metastatic
melanoma: a multicentre, open-label, phase 3 randomised controlled trial. Lancet 380:
358–365.
Hiwatashi A, Nishii Y, and Ichikawa Y (1982) Purification of cytochrome P-450D1 (25-
a
hydroxyvitamin D3-1 a-hydroxylase) of bovine kidney mitochondria. Biochem Biophys Res
Commun 105:320–327.
Holick MF (2007) Vitamin D deficiency. N Engl J Med 357:266–281.
Janjetovic Z, Brozyna AA, Tuckey RC, Kim TK, Nguyen MN, Jozwicki W, Pfeffer SR, Pfeffer
LM, and Slominski AT (2011) High basal NF-kB activity in nonpigmented melanoma cells
is associated with an enhanced sensitivity to vitamin D3 derivatives. Br J Cancer 105:1874–
1884.
Janjetovic Z, Tuckey RC, Nguyen MN, Thorpe EM, Jr, and Slominski AT (2010) 20,23-
dihydroxyvitamin D3, novel P450scc product, stimulates differentiation and inhibits pro-
liferation and NF-kappaB activity in human keratinocytes. J Cell Physiol 223:36–48.
Janjetovic Z, Zmijewski MA, Tuckey RC, DeLeon DA, Nguyen MN, Pfeffer LM, and Slominski
AT (2009) 20-Hydroxycholecalciferol, product of vitamin D3 hydroxylation by P450scc,
decreases NF-kappaB activity by increasing IkappaB a levels in human keratinocytes. PLoS
ONE 4:e5988.
Kim T-K, Wang J, Janjetovic Z, Chen J, Tuckey RC, Nguyen MN, Tang EKY, Miller D, Li W,
and Slominski AT (2012) Correlation between secosteroid-induced vitamin D receptor activity
in melanoma cells and computer-modeled receptor binding strength. Mol Cell Endocrinol 361:
143–152.
Lambeth JD, Kitchen SE, Farooqui AA, Tuckey R, and Kamin H (1982) Cytochrome P-450scc
-
substrate interactions. Studies of binding and catalytic activity using hydroxycholesterols.
J Biol Chem 257:1876–1884.
Lee HJ, Paul S, Atalla N, Thomas PE, Lin X, Yang I, Buckley B, Lu G, Zheng X, and Lou Y-R,
et al. (2008) Gemini vitamin D analogues inhibit estrogen receptor-positive and estrogen
receptor-negative mammary tumorigenesis without hypercalcemic toxicity. Cancer Prev Res
(Phila) 1:476–484.
Li W, Chen J, Janjetovic Z, Kim T-K, Sweatman T, Lu Y, Zjawiony J, Tuckey RC, Miller D,
and Slominski A (2010) Chemical synthesis of 20S-hydroxyvitamin D3, which shows anti-
proliferative activity. Steroids 75:926–935.
Fig. 10. Segment of 1H NMR spectrum for 1,20,26(OH)3D3.

1124
Tang et al.
Slominski A, Semak I, Zjawiony J, Wortsman J, Li W, Szczesniewski A, and Tuckey RC (2005)
The cytochrome P450scc system opens an alternate pathway of vitamin D3 metabolism. FEBS
J 272:4080–4090.
Slominski AT, Janjetovic Z, Fuller BE, Zmijewski MA, Tuckey RC, Nguyen MN, Sweatman T,
Li W, Zjawiony J, and Miller D, et al. (2010) Products of vitamin D3 or 7-dehydrocholesterol
metabolism by cytochrome P450scc show anti-leukemia effects, having low or absent calcemic
activity. PLoS ONE 5:e9907.
Slominski AT, Janjetovic Z, Kim T-K, Wright AC, Grese LN, Riney SJ, Nguyen MN,
and Tuckey RC (2012a) Novel vitamin D hydroxyderivatives inhibit melanoma growth and
show differential effects on normal melanocytes. Anticancer Res 32:3733–3742.
Slominski AT, Kim T-K, Janjetovic Z, Tuckey RC, Bieniek R, Yue J, Li W, Chen J, Nguyen MN,
and Tang EKY, et al. (2011) 20-Hydroxyvitamin D2 is a noncalcemic analog of vitamin D with
potent antiproliferative and prodifferentiation activities in normal and malignant cells. Am J
Physiol Cell Physiol 300:C526–C541.
Slominski AT, Kim T-K, Shehabi HZ, Semak I, Tang EKY, Nguyen MN, Benson HAE, Korik E,
Janjetovic Z, and Chen J, et al. (2012b) In vivo evidence for a novel pathway of vitamin D₃
metabolism initiated by P450scc and modified by CYP27B1. FASEB J 26:3901–3915.
Slominski AT, Kim TK, Zmijewski MA, Janjetovic Z, Li W, Chen J, Kusniatsova EI, Semak I,
Postlethwaite AE, and Miller D, et al. (2013b) Novel vitamin D photoproducts and their
precursors in the skin. Dermatoendocrinol 5:1–13.
Szyszka P, Zmijewski MA, and Slominski AT (2012) New vitamin D analogs as potential
therapeutics in melanoma. Expert Rev Anticancer Ther 12:585–599.
Takahashi T and Morikawa K (2006) Vitamin D receptor agonists: opportunities and challenges
in drug discovery. Curr Top Med Chem 6:1303–1316.
Tang EKY, Li W, Janjetovic Z, Nguyen MN, Wang Z, Slominski A, and Tuckey RC (2010a)
Purified mouse CYP27B1 can hydroxylate 20,23-dihydroxyvitamin D3, producing 1a,20,23-
trihydroxyvitamin D3, which has altered biological activity. Drug Metab Dispos 38:
1553–1559.
Tang EKY, Tieu EW, and Tuckey RC (2012) Expression of human CYP27B1 in Escherichia coli
and characterization in phospholipid vesicles. FEBS J 279:3749–3761.
Tang EKY, Voo KJQ, Nguyen MN, and Tuckey RC (2010b) Metabolism of substrates in-
corporated into phospholipid vesicles by mouse 25-hydroxyvitamin D3 1a-hydroxylase
(CYP27B1). J Steroid Biochem Mol Biol 119:171–179.
Tieu EW, Li W, Chen J, Baldisseri DM, Slominski AT, and Tuckey RC (2012a) Metabolism of
cholesterol, vitamin D3 and 20-hydroxyvitamin D3 incorporated into phospholipid vesicles by
human CYP27A1. J Steroid Biochem Mol Biol 129:163–171.
Tieu EW, Tang EKY, Chen J, Li W, Nguyen MN, Janjetovic Z, Slominski A, and Tuckey RC
(2012b) Rat CYP24A1 acts on 20-hydroxyvitamin D(3) producing hydroxylated products with
increased biological activity. Biochem Pharmacol 84:1696–1704.
Tuckey RC and Kamin H (1982) Kinetics of the incorporation of adrenal cytochrome P-450scc
into phosphatidylcholine vesicles. J Biol Chem 257:2887–2893.
Tuckey RC, Li W, Shehabi HZ, Janjetovic Z, Nguyen MN, Kim TK, Chen J, Howell DE, Benson
HAE, and Sweatman T, et al. (2011) Production of 22-hydroxy metabolites of vitamin d3 by
cytochrome p450scc (CYP11A1) and analysis of their biological activities on skin cells. Drug
Metab Dispos 39:1577–1588.
Tuckey RC, Li W, Zjawiony JK, Zmijewski MA, Nguyen MN, Sweatman T, Miller D,
and Slominski A (2008a) Pathways and products for the metabolism of vitamin D3 by cyto-
chrome P450scc. FEBS J 275:2585–2596.
Tuckey RC, Nguyen MN, and Slominski A (2008b) Kinetics of vitamin D3 metabolism by
cytochrome P450scc (CYP11A1) in phospholipid vesicles and cyclodextrin. Int J Biochem Cell
Biol 40:2619–2626.
Tuckey RC and Sadleir J (1999) The concentration of adrenodoxin reductase limits cytochrome
p450scc activity in the human placenta. Eur J Biochem 263:319–325.
Uchida E, Kagawa N, Sakaki T, Urushino N, Sawada N, Kamakura M, Ohta M, Kato S,
and Inouye K (2004) Purification and characterization of mouse CYP27B1 overproduced by an
Escherichia coli system coexpressing molecular chaperonins GroEL/ES. Biochem Biophys Res
Commun 323:505–511.
Fig. 11. The new secosteroids, 1,20,25(OH)3D3 and 1,20,26(OH)3D3, inhibit
colony formation by SKMEL-188 cells. Colony formation was determined using the
soft agar assay, and colonies of greater than 0.2 mm were counted. Data represent
means 6 S.E.M. (n = 4), and were analyzed using the Student’s t test on GraphPad
Prism. *P , 0.05; **P , 0.01; ***P , 0.001.
Wang J, Slominski A, Tuckey RC, Janjetovic Z, Kulkarni A, Chen J, Postlethwaite AE, Miller D,
and Li W (2012) 20-hydroxyvitamin D₃ inhibits proliferation of cancer cells with high efficacy
while being non-toxic. Anticancer Res 32:739–746.
Woods ST, Sadleir J, Downs T, Triantopoulos T, Headlam MJ, and Tuckey RC (1998) Ex-
pression of catalytically active human cytochrome p450scc in Escherichia coli and mutagenesis
of isoleucine-462. Arch Biochem Biophys 353:109–115.
Yamamoto K, Masuno H, Sawada N, Sakaki T, Inouye K, Ishiguro M, and Yamada S (2004)
Homology modeling of human 25-hydroxyvitamin D3 1alpha-hydroxylase (CYP27B1) based
on the crystal structure of rabbit CYP2C5. J Steroid Biochem Mol Biol 89-90:167–171.
Yamamoto K, Uchida E, Urushino N, Sakaki T, Kagawa N, Sawada N, Kamakura M, Kato S,
Inouye K, and Yamada S (2005) Identification of the amino acid residue of CYP27B1 re-
sponsible for binding of 25-hydroxyvitamin D3 whose mutation causes vitamin D-dependent
rickets type 1. J Biol Chem 280:30511–30516.
Zbytek B, Janjetovic Z, Tuckey RC, Zmijewski MA, Sweatman TW, Jones E, Nguyen MN,
and Slominski AT (2008) 20-Hydroxyvitamin D3, a product of vitamin D3 hydroxylation
by cytochrome P450scc, stimulates keratinocyte differentiation. J Invest Dermatol 128:
2271–2280.
Lu Y, Chen J, Janjetovic Z, Michaels P, Tang EKY, Wang J, Tuckey RC, Slominski AT, Li W,
and Miller DD (2012) Design, synthesis, and biological action of 20R-hydroxyvitamin D3.
J Med Chem 55:3573–3577.
Masuda S and Jones G (2006) Promise of vitamin D analogues in the treatment of hyper-
proliferative conditions. Mol Cancer Ther 5:797–808.
Nguyen MN, Slominski A, Li W, Ng YR, and Tuckey RC (2009) Metabolism of vitamin d2 to
17,20,24-trihydroxyvitamin d2 by cytochrome p450scc (CYP11A1). Drug Metab Dispos 37:
761–767.
Omura T and Sato R (1964) The carbon monoxide-binding pigment of liver microsomes. I.
Evidence for its hemoprotein nature. J Biol Chem 239:2370–2378.
Pinczewski J and Slominski A (2010) The potential role of vitamin D in the progression of benign
and malignant melanocytic neoplasms. Exp Dermatol 19:860–864.
Prosser DE and Jones G (2004) Enzymes involved in the activation and inactivation of vitamin D.
Trends Biochem Sci 29:664–673.
Slominski A, Janjetovic Z, Tuckey RC, Nguyen MN, Bhattacharya KG, Wang J, Li W, Jiao Y,
Gu W, and Brown M, et al. (2013a) 20S-Hydroxyvitamin D3, noncalcemic product of
CYP11A1 action on vitamin D3, exhibits potent antifibrotic activity in vivo. J Clin Endocrinol
Metab 98:E298–E303.
Slominski A, Semak I, Wortsman J, Zjawiony J, Li W, Zbytek B, and Tuckey RC (2006) An
alternative pathway of vitamin D metabolism. Cytochrome P450scc (CYP11A1)-mediated
conversion to 20-hydroxyvitamin D2 and 17,20-dihydroxyvitamin D2. FEBS J 273:
2891–2901.
Address correspondence to: Dr. Robert C. Tuckey, School of Chemistry and
Biochemistry, M310, The University of Western Australia, Crawley, WA, 6009,
Australia. E-mail: robert.tuckey@uwa.edu.au