CYP3A specifically catalyzes 1β-hydroxylation of deoxycholic acid: Characterization and enzymatic synthesis of a potential novel urinary biomarker for CYP3A activity

Article abstract of DOI:10.1124/dmd.116.070805

The endogenous bile acid metabolite 1β-hydroxy-deoxycholic acid (1β-OH-DCA) excreted in human urine may be used as a sensitive CYP3A biomarker in drug development reflecting in vivo CYP3A activity. An efficient and stereospecific enzymatic synthesis of 1β-OH-DCA was developed using a Bacillus megaterium (BM3) cytochrome P450 (P450) mutant, and its structure was confirmed by nuclear magnetic resonance (NMR) spectroscopy. A [²H₄]-labeled analog of 1β-OH-DCA was also prepared. The major hydroxylated metabolite of deoxycholic acid (DCA) in human liver microsomal incubations was identified as 1β-OH-DCA by comparison with the synthesized reference analyzed by UPLC-HRMS. Its formation was strongly inhibited by CYP3A inhibitor ketoconazole. Screening of 21 recombinant human cytochrome P450 (P450) enzymes showed that, with the exception of extrahepatic CYP46A1, the most abundant liver P450 subfamily CYP3A, including CYP3A4, 3A5, and 3A7, specifically catalyzed 1β-OH-DCA formation. This indicated that 1β-hydroxylation of DCA may be a useful marker reaction for CYP3A activity in vitro. The metabolic pathways of DCA and 1β-OHDCA in human hepatocytes were predominantly via glycine and, to a lesser extent, via taurine and sulfate conjugation. The potential utility of 1β-hydroxylation of DCA as a urinary CYP3A biomarker was illustrated by comparing the ratio of 1β-OH-DCA:DCA in a pooled spot urine sample from six healthy control subjects to a sample from one patient treated with carbamazepine, a potent CYP3A inducer; 1β-OH-DCA:DCA was considerably higher in the patient versus controls (ratio 2.8 vs. 0.4). Our results highlight the potential of 1β-OH-DCA as a urinary biomarker in clinical CYP3A DDI studies.

Full text of DOI:10.1124/dmd.116.070805

DMD Fast Forward. Published on July 11, 2016 as DOI: 10.1124/dmd.116.070805
This article has not been copyedited and formatted. The final version may differ from this version.
Title page
CYP3A specifically catalyzes 1 -hydroxylation of deoxycholic acid:
Characterization and enzymatic synthesis of a potential novel urinary
biomarker for CYP3A activity
Martin A. Hayes, Xue-Qing Li, Gunnar Grönberg, Ulf Diczfalusy
and Tommy B. Andersson
Cardiovascular and Metabolic Diseases, DMPK (M.A.H., X.-Q.L., T.B.A.) and Respiratory,
Inflammation and Autoimmune Disease, Medicinal Chemistry (G.G.) Innovative Medicines and Early
Development Biotech Unit, AstraZeneca, Pepparedsleden 1, Mölndal, 431 83, Sweden; Division of
Clinical Chemistry, Department of Laboratory Medicine, Karolinska Institutet, Karolinska University
Hospital, Huddinge, SE-141 86 Stockholm, Sweden (U.D.); Section of Pharmacogenetics, Department
of Physiology and Pharmacology, Karolinska Institutet, Stockholm, Sweden (T.B.A.)

DMD Fast Forward. Published on July 11, 2016 as DOI: 10.1124/dmd.116.070805
This article has not been copyedited and formatted. The final version may differ from this version.
Running Title Page
Characterization and synthesis of 1β-OH-DCA as a biomarker for CYP3A
activity
Address correspondence to: Tommy B. Andersson, PhD. Cardiovascular and Metabolic Diseases,
DMPK, Innovative Medicines and Early Development Biotech Unit, AstraZeneca, Pepparedsleden 1,
Mölndal, 431 83, Sweden
Telephone: +46 31 7761534
Fax: +46 31 7763760
Email: Tommy.B.Andersson@astrazeneca.com
Number of text pages:
Number of Tables: 2
Number of Figures: 8
Number of References: 26
Number of words in Abstract: 249
Number of words in Introduction: 658
Number of words in Discussion: 1392
Abbreviations: CA, cholic acid; CDCA, chenodeoxycholic acid; CDCA-24-G, chenodeoxycholic acid-
24-acyl- -D-glucuronide; DCA, deoxycholic acid; DCA-3-S, deoxycholic acid 3-sulfate; DDI, drug-drug
interaction; DHEA, dehydroepiandrosterone; G-CA, glycocholic acid; G-DCA, glycodeoxycholic acid;
G-CDCA, glycochenodeoxycholic acid; GDH, glucose dehydrogenase; G-1 -OH-DCA, 1β-hydroxy-
deoxycholic acid glycine conjugate; 1β-OH-DCA, 1β-hydroxy-deoxycholic acid, (1β,3α,5β,12α)-1,3,12-
2
2
trihydroxycholan-24-oic acid; [ H
4
4
]-1 -OH-DCA, (1β,3α,5β,12α)-1,3,12-trihydroxy-(2,2,4,4- H )-cholan-
24-oic acid; P450, cytochrome P450; T-CA, taurocholic acid; T-CDCA, taurochenodeoxycholic acid.

DMD Fast Forward. Published on July 11, 2016 as DOI: 10.1124/dmd.116.070805
This article has not been copyedited and formatted. The final version may differ from this version.
Abstract
The endogenous bile acid metabolite 1β-hydroxy-deoxycholic acid (1β-OH-DCA)
excreted in human urine may be used as a sensitive CYP3A biomarker in drug
development reflecting in vivo CYP3A activity. An efficient and stereospecific
enzymatic synthesis of 1β-OH-DCA was developed using a Bacillus megaterium
(BM3) cytochrome P450 (P450) mutant and its structure was confirmed by NMR
2
spectroscopy. A [ H
4
]-labelled analogue of 1β-OH-DCA was also prepared. The
major hydroxylated metabolite of deoxycholic acid (DCA) in human liver microsomal
incubations was identified as 1β-OH-DCA by comparison with the synthesized
reference analyzed by UPLC-HRMS. Its formation was strongly inhibited by CYP3A
inhibitor ketoconazole. Screening of 21 recombinant human cytochrome P450 (P450)
enzymes showed that with the exception of extrahepatic CYP46A1, the most
abundant liver P450 subfamily CYP3A, including CYP3A4, 3A5 and 3A7, specifically
catalyzed 1β-OH-DCA formation. This indicated that 1β-hydroxylation of DCA may be
a useful marker reaction for CYP3A activity in vitro. The metabolic pathways of DCA
and 1β-OH-DCA in human hepatocytes were shown to be predominantly via glycine
and to a lesser extent via taurine and sulfate conjugation. The potential utility of 1 -
hydroxylation of DCA as a urinary CYP3A biomarker was illustrated by comparing the
ratio of 1 -OH-DCA:DCA in a pooled spot urine sample from six healthy control
subjects to a sample from one patient treated with carbamazepine, a potent CYP3A
inducer; 1 -OH-DCA:DCA was considerably higher in the patient vs. controls (ratio
2.8 vs. 0.4). Our results highlight the potential of 1β-OH-DCA as a urinary biomarker
in clinical CYP3A DDI studies.

DMD Fast Forward. Published on July 11, 2016 as DOI: 10.1124/dmd.116.070805
This article has not been copyedited and formatted. The final version may differ from this version.
Introduction
Drug-drug interactions (DDIs) are a major concern in drug development and clinical
practice. The likelihood and magnitude of inhibition and/or induction of drug
metabolizing enzymes is an integral part of the assessment of new drug candidates.
When the potential for a DDI is identified in preclinical studies, further evaluation of
the risk of DDI is often required in clinical investigations. Such studies are most often
carried out using drug probe compounds, which are selectively metabolized by the
affected enzyme. The vast majority of clinical DDI studies investigate the effect on
cytochrome P450 (P450) 3A. A sensitive probe drug for CYP3A is midazolam which
is commonly used to assess inhibition or induction of this enzyme. Sensitive
endogenous biomarkers, especially urinary biomarkers if applicable, would have
great utility, avoiding unnecessary drug exposure and invasive sampling. The 6β-
hydroxycortisol to cortisol ratio (6β-hydroxycortisol ratio) in urine has been suggested
as an endogenous marker for CYP3A4 activity (Galteau and Shamsa, 2003). The
intra- and inter-individual variation of this ratio is large however and can only be used
when the subjects are their own controls. Its specificity for CYP3A activity is also
debated (Galteau and Shamsa, 2003). The urinary ratios of 6β-hydroxycortisone to
cortisone and 7β-hydroxy-dehydroepiandrosterone (7β-hydroxy DHEA) to DHEA
have also been suggested to be useful indicators of CYP3A4 activity, but this
observation needs to be validated in vivo (Shin et al., 2013). One newly developed
endogenous biomarker for CYP3A is the blood levels of 4β-hydroxycholesterol
(Kanebratt et al., 2008; Bjorkhem-Bergman et al., 2013). The elimination half-life of
4
-hydroxycholesterol is about 17 days, resulting in stable plasma concentrations
within subjects (Diczfalusy et al., 2009) but this excludes 4 -hydroxycholesterol as a
marker for rapid changes in CYP3A4 activity. A sensitive new biomarker, preferably

DMD Fast Forward. Published on July 11, 2016 as DOI: 10.1124/dmd.116.070805
This article has not been copyedited and formatted. The final version may differ from this version.
measureable in urine spot samples, with a fast response to change in CYP3A activity
would thus be of great benefit for the assessment of DDIs and phenotyping in
humans.
Concentrations of the bile acid metabolite 1β-hydroxy-deoxycholic acid (1 -OH-DCA)
were observed to increase significantly in the urine of patients taking the known
CYP3A4 inducer carbamazepine (Bodin et al., 2005). This study also showed that the
conversion of deoxycholic acid (DCA) to 1β-OH-DCA was catalysed in vitro by
recombinant CYP3A4 and by human liver microsomes though no data was presented
on the specificity of this transformation in terms of the contribution of additional
enzymes from the CYP3A family or other CYPs (Bodin et al., 2005). We reasoned
that it may be possible to use 1β-OH-DCA as a urinary biomarker to measure CYP3A
induction and inhibition in clinical drug–drug interaction studies rather than using a
conventional invasive plasma biomarker or probe drug.
The metabolite 1β-OH-DCA is not available commercially. A nine step chemical
synthesis (Tohma et al., 1986) and a microbial synthesis (Carlström et al., 1981) of
1β-OH-DCA have been reported previously. Bacterial P450s have shown promise in
the synthesis of drug metabolites (Gillam and Hayes, 2013) and we thought it may be
possible to synthesize 1β-OH-DCA enzymatically using a commercially available
bacterial mutant P450.
The aim of the present study was to investigate the 1 -hydroxylation of DCA by
human liver microsomes and to assess the contribution of individual CYPs using
human recombinant P450 enzymes and chemical inhibitors. To facilitate this work a
library of Bacillus megaterium P450 mutants was screened using DCA as a substrate
and a suitable enzyme was identified which produced the major hydroxylated DCA

DMD Fast Forward. Published on July 11, 2016 as DOI: 10.1124/dmd.116.070805
This article has not been copyedited and formatted. The final version may differ from this version.
metabolite formed in HLM. The aim was to generate sufficient quantities of this
hydroxylated metabolite for full structural characterization by NMR spectroscopy. A
2
stable isotope labelled analogue [ H
4
]-1 -OH-DCA, required for bioanalytical method
development was also synthesized using similar methodology. An ultra-high
performance liquid chromatography/high resolution mass spectrometry
(UPLC/HRMS) based assay was used for metabolite identification and quantification.
The potential application of 1 -hydroxylation of DCA as a urinary CYP3A biomarker
was illustrated by comparison of two spot human urine samples, one pooled sample
from healthy subjects as a control and one from a patient treated with
carbamazepine, a CYP3A inducer.

DMD Fast Forward. Published on July 11, 2016 as DOI: 10.1124/dmd.116.070805
This article has not been copyedited and formatted. The final version may differ from this version.
Materials and Methods
Chemicals and Reagents
Deoxycholic acid (DCA), cholic acid (CA), chenodeoxycholic acid (CDCA),
glycochenodeoxycholic acid (G-CDCA), glycocholic acid (G-CA),
taurochenodeoxycholic acid (T-CDCA), taurocholic acid (T-CA), NADPH, -
glucuronidase/arylsulfatase (Helix pomatia) and choloylglycine hydrolase from
Clostridium perfringens (C. welchii) were obtained from Sigma-Aldrich (Dorset, UK).
Deoxycholic -2,2,4,4-d acid (Lot No. M403P52, 99 atom % D) was purchased from
4
CDN isotopes, Quebec, Canada. Deoxycholic acid 3-sulfate (DCA-3-S) and
ketoconazole were obtained from Compound Management, AstraZeneca Gothenburg
(Sweden). Chenodeoxycholic acid-24-acyl- -D-glucuronide (CDCA-24-G) was
purchased from Toronto Research Chemicals (Toronto, Canada). Pooled human liver
microsomes (HLM) were purchased from BD Gentest (Woburn, MA) and recombinant
human CYPs 1A1, 1A2, 1B1, 2A6, 2A13, 2B6, 2C8, 2C9, 2C18, 2C19, 2D6, 2E1,
2J2, 3A4, 3A5, 3A7, 4A11, 4F2, 4F3, 17A1 and 46A1 were obtained from Cypex Ltd
(Dundee, Scotland, UK). Pooled human hepatocytes were purchased from Celsis-In
TM
Vitro Technologies (Baltimore, MD, USA). Microcyp screening plates (MCYP-
0343), purified P450 enzyme (MCYP0029) and glucose dehydrogenase (GDH) were
purchased from Codexis (Redwood City, CA, USA). Acetonitrile (LC–MS grade) was
obtained from Fisher Scientific (Fair Lawn, NJ, USA). Formic acid (purity 98-100%)
were purchased from Merck KGaA, Germany. All other chemicals and solvents were
of the highest quality available commercially.

DMD Fast Forward. Published on July 11, 2016 as DOI: 10.1124/dmd.116.070805
This article has not been copyedited and formatted. The final version may differ from this version.
Spot human urine samples
A spot human urine sample originated from a reported clinical study in which all
patients were on monotherapy with carbamazepine for at least 1 year and, apart from
epilepsy, were healthy (Bodin et al., 2005) (approved by the local Human Ethics
committee). Control urine was prepared by pooling 2 ml of spot urine samples
collected from six healthy subjects.
Screening for BM3 P450 catalysed conversion of deoxycholic acid
The screening of DCA against the Codexis Microcyp 96-well screening plate A
(MCYP-0343) was performed as follows: DCA (20 mg, 0.05 mmol) was dissolved in
acetonitrile (1.0 ml) and added to a solution of MCYP-RXN buffer (46.6 mL, 0.802 g)
in deionized water (28.1 ml). The MCYP-RXN buffer contained glucose (30 mM),
+
NADP (1.2 mM), glucose dehydrogenase (0.6 mg/ml) in potassium phosphate buffer
(120 mM). The solution was adjusted to pH 8.0 with NaOH (0.1 M). Aliquots (0.25 ml)
o
were pipetted into each enzyme well and the solution shaken at 30 C at 250 rpm in
an orbital shaker. An aliquot of reaction mixture (10 µL) was removed from each well
of the 96-well screening plate and diluted with 90 μL of mobile phase A (0.1% formic
acid in water). The mixture was centrifuged using a benchtop centrifuge at 1200 × g
then transferred to either a 96-well Nunc plate or to glass injection vials (Waters) prior
to UPLC/HRMS analysis. LC/MS data was processed using Metabolynx (Waters)
software. Hit enzymes were identified on the basis of forming a ‘+16 Da’
(monooxidised) metabolite of DCA with retention time 3.22 min.
2
Enzymatic synthesis of 1 -OH-DCA and [ H
4
]-1 -OH-DCA using MCYP0029
Preparation of 1β-OH-DCA; (1β,3α,5β,12α)-1,3,12-trihydroxycholan-24-oic acid.

DMD Fast Forward. Published on July 11, 2016 as DOI: 10.1124/dmd.116.070805
This article has not been copyedited and formatted. The final version may differ from this version.
To a 250 mL conical flask cooled in an ice/water bath was added phosphate buffer
+
(
0.12 M, 95 mL, pH 8), D-glucose (0.45 g, 2.5 mmol), NADP (74.3 mg, 0.1 mmol)
and glucose dehydrogenase (50 mg). Deoxycholic acid (39.2 mg, 0.1 mmol) was
dissolved in MeOH (100 µL) and added to the reaction flask. MicroCYP0029 (151
mg, 100 nmol) was dissolved in phosphate buffer (0.12 M, 5 mL, pH 8) and the
solution added to the reaction flask. The flask was shaken in an orbital shaker at 200
o
rpm, 30 C and the reaction monitored by UPLC/HRMS. After 16 hrs the reaction was
quenched with cold acetonitrile (100 mL) and the solution was allowed to stand in a
o
refrigerator at 4 C for 30 mins after which the contents transferred to Falcon tubes (4
o
×
50 mL) and centrifuged (4000 × g, 4 C) for 15 mins. The supernatents were
decanted off, pooled and concentrated on a rotary evaporator. After removal of the
organics the aqueous was split between two Oasis HLB SPE cartridges (2 × 6 cc,
2
200 mg). Each cartridge was washed with H O (1 × 5 mL) and eluted with acetonitrile
(
1 × 5 mL). The organic eluates were pooled, concentrated to dryness under an N
2
stream and the crude residue purified by mass directed preparative HPLC.
2
Preparation
of
[ H
4
]-1β-OH-deoxycholic
acid,
(1β,3α,5β,12α)-1,3,12-
2
trihydroxy(2,2,4,4- H )cholan-24-oic acid.
4
2
The [ H
4
]-1β-OH-deoxycholic acid was prepared using the same methodology as that
used for the unlabelled analogue. Deoxycholic-2,2,4,4-d
4
acid (39.3 mg, 0.1 mmol)
was used as substrate.
2
Mass directed preparative HPLC purification of 1 -OH-DCA and [ H
DCA
4
]-1 -OH-
The residue from the concentrated SPE eluate was dissolved in DMSO (1 mL) and
purified by mass directed preparative HPLC using a Waters Sunfire C18 ODB column

DMD Fast Forward. Published on July 11, 2016 as DOI: 10.1124/dmd.116.070805
This article has not been copyedited and formatted. The final version may differ from this version.
4
(30x150 mm, 5 μm). Mobile phases were 0.1 M NH OAc (Solvent A) and acetonitrile
(Solvent B). Gradient was 10%B increasing to 50% over 20 mins. Flow rate was 60
mL/min. A Waters ZQ mass spectrometer was used for mass detection. 1β-OH-DCA
eluted between 11.0-11.8 mins. Fractions containing the targeted mass of m/z 407,
corresponding to mono-oxidised DCA were pooled, the organics were removed on a
rotary evaporator and the aqueous residue was subjected to SPE on a Waters Oasis
HLB cartridge (500 mg). The cartridge was washed with water (3 mL), and allowed to
dry under vacuum and then eluted with MeOH (3 mL). The MeOH eluate was
concentrated to dryness yielding a white solid (10.5 mg) with mass of m/z 407.2799
(Figure 1).
2
The [ H
4
]-1β-OH-DCA was purified using identical methodology and yielded (13.4
mg) as a white solid with a mass of m/z 411.3060 (Figure 1).
2
NMR characterization of 1β-OH-DCA and [ H
4
]-1β-OH-DCA
NMR spectra were recorded on a Bruker 600 MHz AVANCE III system equipped with
a 5 mm QCI Cryoprobe using standard Bruker pulse sequences. Experiments were
3
run in deuterated methanol (CD OD) at 25°C. Chemical shifts are referenced relative
1
13
to the residual methyl signal in CD
3
OD set to 3.31 ppm ( H) or 47.8 ppm ( C). NMR
2
1
13
assignments of 1β-OH-DCA and [ H
4
]-1β-OH-DCA were based on H and C
1
1
1
13
1
chemical shifts and 2D correlations via H- H COSY, ROESY and H- C HSQC, H-
13
C HMBC experiments
Incubation of DCA in HLM
Incubation of DCA in HLM was performed in 96-well plates in a shaking incubator at
37°C. The incubation mixture contained DCA (at 2 and 20 µM, respectively), 1 mM
NADPH, and HLM at a protein concentration of 0.5 mg/mL in potassium phosphate

DMD Fast Forward. Published on July 11, 2016 as DOI: 10.1124/dmd.116.070805
This article has not been copyedited and formatted. The final version may differ from this version.
buffer (0.1 M, pH 7.4). The total incubation volume was 50 µL/well. All incubations
were performed in duplicate. Following 5 min sample preincubation at 37°C, the
reaction was initiated by the addition of NADPH and was stopped by adding 100 µL
of ice-cold acetontrile after 60 min. The reaction mixture was then centrifuged at
4,000 × g for 20 min, 4 °C, and an aliquot of 50 L of supernatant was diluted with an
equal volume of water. The resulting mixture was analysed by UPLC/HRMS.
To investigate the effect of the CYP3A inhibitor ketoconazole on the hydroxylation of
DCA, similar incubation conditions were conducted in HLM with or without co-
incubation of ketoconazole at concentrations of 0.1, 1 and 10 µM. The reaction was
initiated by the addition of NADPH and was stopped by the addition of ice-cold
acetontrile after 30 min. All samples were prepared for UPLC/HRMS analysis as
described previously.
Incubation of DCA in a panel of 21 recombinant human CYPs
A panel of 21 recombinant human CYPs (CYPs 1A1, 1A2, 1B1, 2A6, 2A13, 2B6,
2
C8, 2C9, 2C18, 2C19, 2D6, 2E1, 2J2, 3A4, 3A5, 3A7, 4A11, 4F2, 4F3, 17A1 and
6A1; 100 pmol/mL) were screened to investigate their catalytic activity on the 1 -
4
hydroxylation of DCA. The incubation conditions were similar to those in HLM. In
brief, the incubation mixture contained 100 pmol/mL of recombinant human CYP, 2 or
20 µM of DCA, and 1 mM NADPH in potassium phosphate buffer (0.1 M, pH 7.4). For
the purpose of metabolite characterization, an incubation time of 60 min was used.
Incubation of DCA and 1 -OH-DCA in human hepatocytes
DCA and 1 -OH-DCA were incubated in human hepatocytes separately. Each
incubation (50 µL/well) contained 1 million cells per mL and test compound, DCA or
1
-OH-DCA, respectively, at a concentration of 4 M, in 96-well flat bottomed

DMD Fast Forward. Published on July 11, 2016 as DOI: 10.1124/dmd.116.070805
This article has not been copyedited and formatted. The final version may differ from this version.
polyethylene plates at 37°C, 95% air, 5% CO
incubation at 37°C, the test compound was added to start the reaction. At 0, 40 and
20 minutes, the reactions were stopped by addition of 150 µL of ice-cold acetonitrile
2
in saturated humidity. After 10 min pre-
1
and kept on ice for at least 20 min before centrifugation at 4000 × g and 4°C for 20
min. The supernatants were analyzed by UPLC/HRMS. The effect of ketoconazole
on DCA metabolite formation was also examined with and without the addition of
ketoconazole (3 µM) in hepatocyte incubations with DCA for 4 hours.
Determination of 1 -OH-DCA and DCA in human urine
The analyses of 1 -OH -DCA and DCA in human urine were performed using a
modified method originally described by Bodin et al. (Bodin et al., 2005). To 200 µL of
urine, an aliquot of 200 µL of 0.1 M sodium phosphate buffer (pH 5) was added
followed by the addition of 5 µL of -glucuronidase/arylsulfatase and 5 µLof
choloylglycine hydrolase (2 U/µL), respectively. The mixture was incubated overnight
at °C. After the enzymatic hydrolysis of glucuronides and amino acid conjugates,
an aliquot of 50 µL of the reaction mixture was quenched with acetonitrile (150 µL,
2
containing 200 nM [ H
4
]-1 -OH-DCA as internal standard) and centrifuged at 10 000
×
g for 10 min. The supernatant was separated and diluted with an equal volume of
water for UPLC/HRMS analysis. A second aliquot of incubation mixture (150 μL) was
2
quenched with acetonitrile (150 μL containing 200 nM [ H
4
]-1 -OH-DCA as internal
standard) and centrifuged at 4 000 × g for 20 min. An aliquot of 150 µL of the
supernatant was separated and evaporated under nitrogen to a reduced volume of
less than 50 µL. The remaining liquid was acidified by adding formic acid (0.1% in
2
H O, 50 µL). The resulting mixture was loaded onto an Oasis MAX 96-well µElution
Plate (2 mg sorbent per well, 30 µm particle size, Waters) pre-conditioned with
methanol and water. Loaded cartridges were then successively washed with 0.2 mL

DMD Fast Forward. Published on July 11, 2016 as DOI: 10.1124/dmd.116.070805
This article has not been copyedited and formatted. The final version may differ from this version.
water and eluted with 150 µL isopropanol-acetonitrile (6:4, v/v). The eluate containing
bile acids was then solvolyzed. To the eluate, an aliquot of 100 µL acetone and 40 µL
of hydrochloric acid (1 M) was added and incubated for 1 hour at °C. After 1 hour,
the solvent was evaporated and the residue reconstituted in 0.5 mL NaOH (4 M) and
1.2 mL isopropanol and incubated overnight at 70 °C. The incubated mixtures were
then neutralized with the addition of formic acid (200 µL, 98%) at the end of the
solvolysis and the solvent evaporated to a volume of less than 300 µL. The remaining
aqueous phase was diluted with 200 µL of 0.1% fomic acid and a second extraction
was performed on an Oasis MAX 96-well µElution Plate using the same method
described previously. The eluate was then evaporated under nitrogen and the
residue was dissolved in 100 µL 50% acetonitrile solution prior to UPLC/HRMS
analysis. Calibration standards were constructed by spiking serial concentrations (16
to 5000 nM) of DCA and 1 -OH-DCA into 0.1 M sodium phosphate buffer (pH 8). The
standard solutions were then subjected to sample preparation in parallel with urine
samples including enzyme hydrolysis and solvolysis. Reference solutions in sodium
phosphate buffer (pH 5) containing 2 µM of G-CDCA, T-CDCA, G-CA, T-CA, CDCA-
24-G or DCA-3-S were subjected to enzymatic hydrolysis using a mixture of -
glucuronidase/arylsulfatase and choloylglycine hydrolase to monitor the hydrolysis of
bile acid amide, glucuronide and/or sulfate conjugates.
UPLC/HRMS analysis
To separate and identify the hydroxylated metabolites of DCA, a Synapt G2 Q-TOF
mass spectrometer equipped with an ACQUITY UPLC system (Waters Corp.) was
used. The UPLC/HRMS system was operated by MassLynx software (version 4.1,
Waters Corp.). UPLC separations were performed on an Acquity UPLC BEH C18
column (2.1×100 mm, 1.7 m; Waters, Milford, MA). Mobile phase A was 0.1% formic

DMD Fast Forward. Published on July 11, 2016 as DOI: 10.1124/dmd.116.070805
This article has not been copyedited and formatted. The final version may differ from this version.
acid aqueous solution and mobile phase B was acetonitrile. The initial mobile phase
was 90:10 A-B, and was transitioned via a linear gradient to 10:90 A-B over 6 min.
The flow rate was 0.5 mL/min and the total run time was 7 minutes. The column oven
and autosampler were set at 45 °C and 8 °C, respectively. The UPLC eluent was
introduced into the Q-TOF mass spectrometer with an electrospray ionization source
(ESI) in negative mode. Specific mass spectrometric source conditions were as
follows: capillary voltage of 0.5 kV, sample cone voltage of 40 V, desolvation
temperature and source temperature were of 550 and 150 °C, respectively. HRMS
E
and MS scan functions were programmed with two independent collision energies
(CE). The HRMS spectra were obtained with a low collision energy, where the
E
transfer CE and trap CE were 0 eV and 4 eV, respectively. The MS spectra were
acquired with a high collision energy, where the transfer CE and trap CE were 30 eV
and were ramped from 15 to 35 eV, respectively. Data were acquired over the range
m/z 100–1200. In the analysis, extracted ion chromatograms (XIC) at m/z 391.285
2
(
DCA), m/z 407.280 (1 -OH-DCA) and m/z 411.305 ([ H
4
]-1 -OH-DCA) were
generated with a mass tolerance of 10 mDa. Reference standards were analyzed in
parallel for definitive confirmation. Quantification of the 1 -OH-DCA and DCA was
performed using TargetLynx. The concentrations of bile acids were calculated based
on the peak area ratio of each compound to the internal standard and compared to
the calibration curves. Mass responses for standards of DCA, 1β-OH DCA, CA, G-
CA, T-CA, CDCA, G-CDCA and T-CDCA were similar by the UPLC-HRMS method
used. This allowed a semiquantitative assessment of the relative abundance of DCA
and 1β-OH-DCA and their conjugated metabolites in hepatocytes to be made by
comparison of peak areas. The metabolite identification of DCA and 1 -OH-DCA in
different samples was performed using Metabolynx (Waters) software.

DMD Fast Forward. Published on July 11, 2016 as DOI: 10.1124/dmd.116.070805
This article has not been copyedited and formatted. The final version may differ from this version.
Results
Enzyme synthesis and structural characterization of 1 -OH-DCA.
Deoxycholic acid was screened against a library of commercially available BM3 P450
mutants looking for formation of its major oxidized metabolite (M3) which had been
identified in HLM and rCYP3A4 incubations. Chromatographic retention time and
mass spectral data for M3 were known allowing a focussed screen to be performed.
Only BM3 mutants forming M3 in good conversion and ideally as a single metabolic
product were of interest. Of the 90 mutants screened only MCYP0027 and
MCYP0029 gave >35% conversion to a +16 Da metabolite which co-
chromatographed (retention time 3.22 mins) with the desired major HLM metabolite
based on semi-quantitative estimates of parent compound consumed and M3 formed
(Figure 2). The enzyme MCYP0029 was selected for scale up due to better
conversion (ca 50-60%) and a simpler metabolite profile. Interestingly a wide range of
mono-oxidized metabolites of DCA were also produced by the enzyme library with
retention times at 3.06, 3.59, 3.68, 3.73 and 3.98 mins. The identity of these
metabolites have not been investigated further.
A weighable quantity of M3 (which we reasoned was 1β-OH-DCA) was required for
2
structural confirmation by NMR. Larger quantities of both 1β-OH-DCA and [ H
4
]-1β-
OH-DCA were also required for use as bioanalytical standards in forthcoming clinical
DDI studies. Therefore scale up experiments using MCYP0029 and either DCA or
deoxycholic -2,2,4,4-d acid as substrate, were performed at 0.1 mmol scale and
4
monitored by UPLC/HRMS. After work up and purification by mass directed
preparative LC the colourless solids were characterized by NMR spectroscopy.

DMD Fast Forward. Published on July 11, 2016 as DOI: 10.1124/dmd.116.070805
This article has not been copyedited and formatted. The final version may differ from this version.
NMR spectroscopy confirmed that the identity of M3 was indeed 1β-OH-DCA. The
2
NMR assignments of 1β-OH-DCA and [ H
4
]-1β-OH-DCA were compared to the fully
1
13
assigned H and C spectra of DCA as shown in Table 1. Numbering below follows
the IUPAC approved atom numbering of steroids. Some marker groups of DCA are
1
13
the OH substituted positions CH(12); 3.96, 72.8 ( H, C chemical shift in ppm) and
CH(3); 3.53, 71.3, methyl groups CH (18); 0.71, 12.0, CH (19); 0.93, 22.5 and
(1) is found at 0.98(dt, J=14.2, 14.2, 3.4)/1.78, 35.2 ppm.
3
3
CH (21); 1.01, 16.4. CH
3 2
1
The equatorial (β) H signal from CH(12) appeared as a triplet with 2 small J-
1
couplings (1.8, 1.8 Hz) to the axial and equatorial CH (11) protons. The axial (β) H
2
signals from CH(3) appeared as a triplet of triplets with two large trans-couplings and
1
two smaller cis-couplings (11.1, 11.1, 4.6, 4.6 Hz). H NMR spectra of 1β-OH-DCA
2
13
and [ H
4
]-1β-OH-DCA are shown in Figure 3. C NMR spectra of 1β-OH-DCA and
2
[
4
H ]-1β-OH-DCA are shown in Supplemental Data Figure S1.
In 1β-OH-DCA the signals assigned to CH
signal appeared at δ 3.82, 72.9 ppm indicating hydroxylation at this position. The
assignment of this signal to CH(1) was supported by COSY cross-peaks to CH (2),
CH (19) (Supplemental Data Figure S2), ROESY cross peaks indicated its close
proximity to CH
H long-range correlations to C(10), CH
J~2.6, 2.6 Hz) to CH
2
(1) in DCA are missing and a new CH
2
3
2 3
(2), CH (19) and CH
2
(11eq (α)) at 1.43 ppm (Figure 4) as well as C-
1
2
(2), CH
3
(19). The H coupling constants (t,
2
(2) were consistent with an equatorial configuration of the
proton i.e. OH in the axial (β)-position.
1
It should be noted that the axial (β) H signal from CH(3) which appeared in DCA as
a well resolved triplet of triplets, were not as well resolved in 1β-OH-DCA though,
again, two large and two small couplings were observed. This difference was

DMD Fast Forward. Published on July 11, 2016 as DOI: 10.1124/dmd.116.070805
This article has not been copyedited and formatted. The final version may differ from this version.
explained by the fact that, in 1-β-OH-DCA, the resonances from the axial(α) CH(4)
and CH(5) appear at approximately the same shift resulting in 2nd order J-coupling
effects on CH(3).
2
The NMR data of [ H
4
]-1β-OH-DCA, deuterated at positions C(2) and C(4), confirmed
1
the findings of 1β-OH-DCA discussed above. As expected H signals from CH(1);
3.81, 72.8 and CH(3); 3.96, 65.6 now lacked J-couplings and appeared as singlets
13
(
Table 1). Notably the C signals of C(2) and C(4) broadened and disappeared in the
2
spectrum due to quadrupolar coupling to H and lack of Overhauser effect to the
1
2
directly bonded H. Also notable were the H induced isotope shifts, δ
i
, observed in
]-1β-OH-DCA compared to 1β-OH-DCA, which decreased
with the number of bonds from the deuterated carbons; δ, C(3);0.22, C(5);0.18,
C(6);0.04, C(1);0.09, C(10);0.06, C(9);0.01 ppm (Supplemental Data Figure S3).
13
2
the C spectrum of [ H
4
i
1
-Hydroxylation of DCA in human liver microsomes and the inhibitory effect
of CYP3A inhibitor ketoconazole.
Incubations of DCA with HLM in the presence of NADPH resulted in four
hydroxylated metabolites (M1 to M4, m/z 407.280) with M3 as the most abundant one
(Figure 2). A further oxidized metabolite (-2 Da) was also found in the incubation. The
profiles of the hydroxylated metabolites of DCA when incubated at 20 µM (data not
shown) were similar to those incubated with 2 µM of DCA (Figure 2). Metabolites M3
and M4 were identified as 1 -OH-DCA and CA, respectively, by comparison with the
synthesized and commercially available reference standards. In comparison with the
control samples, the formation of 1 -OH-DCA formation was significantly inhibited by
ketoconazole by 37%, 93% and 100% at ketoconazole concentrations of 0.1, 1, and

DMD Fast Forward. Published on July 11, 2016 as DOI: 10.1124/dmd.116.070805
This article has not been copyedited and formatted. The final version may differ from this version.
10 M, respectively, when incubated at 2 µM of DCA (Figure 5), and by 22%, 89%
and 98%, respectively, when incubated at 20 µM of DCA.
P450 enzymes involved in the 1 -hydroxylation of DCA.
To identify the specific P450s that catalyzed DCA 1 -hydroxylation, a panel of 21
recombinant human CYP isozymes was screened and compared to the profiles
obtained in human liver microsomes. The same conclusion was drawn from
incubations at both 2 and 20 µM of DCA. Of the 21 recombinant human CYPs tested,
CYP3A4 and 3A7 showed the highest catalytic activity of 1 -OH-DCA formation and
CYP3A5 and CYP46A1 showed much less activity towards this reaction. Only trace
amounts of 1 -OH-DCA were detected in CYP2C8 incubations and none was
detected in the incubation of the other CYPs investigated (Figure 6).
Metabolite profiles of DCA and 1 -OH-DCA in human hepatocytes.
Both DCA and 1 -OH-DCA were rapidly metabolized in human hepatocytes. No
parent DCA was detected after a 40-min incubation and the glycine conjugate (G-
DCA) was the predominant metabolite formed. Additional minor metabolic pathways
were also detected including direct taurine conjugation and the combination of G-
DCA with oxidation and sulfation (Figure 7). Similar metabolic pathways were also
detected in incubations of 1 -OH-DCA. After 40 min incubation, the remaining parent
1
-OH-DCA accounted for 44% of the 0 min sample and the direct glycine conjugate
(G-1 -OH-DCA) was the only metabolite detected. With a longer incubation time of
up to 2 hours, low levels of the direct taurine conjugate and the metabolites derived
from the combination of desaturation and sulfation of G-1 -OH-DCA were also
detected. It is noteworthy that the MS peak corresponding to the G-1 -OH-DCA in
1
-OH-DCA incubations was also detected in hepatocyte incubations of DCA, where

DMD Fast Forward. Published on July 11, 2016 as DOI: 10.1124/dmd.116.070805
This article has not been copyedited and formatted. The final version may differ from this version.
the MS peak area of the G-1 -OH-DCA was approximately 11% of the peak area of
G-DCA after 2 hours incubation of DCA. Furthermore, the formation of the G-1 -OH-
DCA was totally abolished in the co-incubation of DCA with ketoconazole in human
hepatocytes. However, no significant differences were observed on the formation of
amidated conjugates (glycine and taurine conjugates) of these two bile acids with or
without coincubation with ketoconazole.
Determination of 1 -OH-DCA and DCA in human urine samples.
Urine samples from healthy subjects and from a patient on monotherapy with
carbamazepine were prepared by enzymatic deconjugation followed by chemical
solvolysis. Table 2 shows the concentrations of 1 -OH-DCA, DCA and their ratios
expressed as 1 -OH-DCA/DCA in urine samples following enzyme hydrolysis with
and without solvolysis. After enzyme hydrolysis and chemical solvolysis the ratio of
1
-OH-DCA to DCA in the urine (ratio of 2.8) from a carbamazepine-treated patient
was 7-fold higher than that in the control urine (ratio of 0.4). Urine samples following
only enzymatic hydrolysis also showed that the ratio of 1 -OH-DCA/DCA was 7-fold
higher in patient urine (ratio of 1.3) than that in the control sample (ratio of 0.2)
(Figure 8). The efficiency of the enzyme catalyzed deconjugation was estimated by
monitoring the disappearance of the reference conjugates and the formation of their
corresponding free bile acids after incubation with the enzymes. The -
glucuronidase/arylsulfatase catalyzed hydrolysis of glucuronides and choloylglycine
hydrolase catalyzed hydrolysis of the amidated conjugates (glycine and taurine
conjugates) of bile acids was determined to be complete by UPLC analysis.
However, the desulfation of the reference compound DCA-3-S was poor with only low
levels of deconjugated DCA formed, indicating the poor catalytic activity of the

DMD Fast Forward. Published on July 11, 2016 as DOI: 10.1124/dmd.116.070805
This article has not been copyedited and formatted. The final version may differ from this version.
arylsulfatase to the desulfation reaction of these sulfated bile acids under the
experimental conditions.
Discussion
A prerequisite for using an endogenous metabolite as a sensitive biomarker for
probing a particular enzyme activity in humans is that the metabolite is formed
selectively by the enzyme of interest. The current study clearly shows that the
formation of 1 -OH-DCA from DCA is predominantly catalyzed by CYP3A enzymes.
The only other CYP enzyme investigated that produced a significant amount of this
metabolite was CYP46A1 which is a cholesterol metabolizing enzyme located in the
brain (Dutheil et al., 2009). Since the level of bile acids in the brain is very low (Bron
et al., 1977) the contribution of CYP46A1 to the overall production of 1 -OH-DCA in
humans should be minor thereby making the formation of 1β-OH-DCA from DCA a
specific biomarker for CYP3A activity.
The finding that CYP3A7 has a high capacity to produce 1 -OH-DCA is intriguing.
CYP3A7 is mainly a fetal form of CYP3A that is down regulated after birth (Kitada et
al., 1985). A high capacity of the fetal liver to metabolize DCA and specifically form its
1
-hydroxylated metabolite was described by Gustafsson et al. (Gustafsson et al.,
985). CYP3A7 is very active in the fetal liver, peaks during the first weeks after birth
1
and then progressively declines. At the same time CYP3A4 levels begin to rise
reaching 30-40% of adult levels after one month (Lacroix et al., 1997). There are
known substrate specificity differences between CYP3A7 and CYP3A4, e.g. the
catalytic efficiency of CYP3A7 to metabolize midazolam is much lower than for
CYP3A4 and CYP3A5 (Williams et al., 2002). The specificity of CYP3A to form 1 -
OH-DCA from DCA could also be developed into a useful in vitro tool for phenotyping

DMD Fast Forward. Published on July 11, 2016 as DOI: 10.1124/dmd.116.070805
This article has not been copyedited and formatted. The final version may differ from this version.
in human liver microsomes and human hepatocytes, keeping in mind the extensive
glycine conjugation of both DCA and 1β-OH-DCA observed in human hepatocytes.
Human P450s metabolise the majority of current small molecule therapeutics (Rendic
and Guengerich, 2015). Their use for preparative biotransformation is limited
primarily due to poor stability. Bacterial P450s continue to show promise for the
synthesis of drug metabolites (Gillam and Hayes, 2013) with mutant enzymes from
Bacillus megaterium being particularly thoroughly investigated (Sawayama et al.,
2009; Reinen et al., 2015). Both protein and process engineering approaches
continue to enhance the possibilities of using these diverse catalysts at scale (Kiss et
al., 2015; Venkataraman et al., 2015). The synthesis of rare and unusual
endogenous metabolites with engineered P450s is less well explored. In this study
2
we have successfully synthesized at mg scale both unlabelled 1β-OH-DCA and [ H
4
]-
1β-OH-DCA using a BM3 mutant P450, in one step with high stereo and regio-
selectivity.
Bile acids constitute a large family of steroids carrying a carboxyl group in the side
chain. DCA is a secondary bile acid and mainly excreted in urine as conjugates, the
sulfated and amidated metabolites and their combinations being the major ones
(Alme et al., 1977; Bathena et al., 2015a; Bathena et al., 2015b). In this study, we
have established that the 1 -hydroxylation of DCA is specifically catalyzed by CYP3A
and therefore has potential to be used in the clinic as a biomarker reflecting CYP3A
activity in DDI investigations. To be able to evaluate the applicability of this metabolic
reaction as a urinary CYP3A biomarker, establishing a robust and sensitive analytical
method to determine total 1 -OH-DCA and DCA concentrations was critical. One
approach to achieve this goal was to cleave the conjugates of these two bile acids,

DMD Fast Forward. Published on July 11, 2016 as DOI: 10.1124/dmd.116.070805
This article has not been copyedited and formatted. The final version may differ from this version.
enzymatically and chemically, releasing the free bile acids for quantification (Bodin et
al., 2005; Kakiyama et al., 2014). Using this approach, the quantification of the non-
conjugated 1 -OH-DCA and DCA by LC/MS is rather straightforward with access to
the reference standards. The disadvantage of this method is that the sample
preparation steps are very time consuming and the chemical cleavage reaction may
be incomplete or give side reactions. Another approach is to directly measure
concentrations of free bile acids and their corresponding amidated and sulfated
conjugates to obtain the total concentrations of bile acids. However, this approach
suffers from the lack of access to some of the reference standards for accurate
quantification. In this study, we modified a reported sample preparation method
(Bodin et al., 2005). In addition to the reported enzyme hydrolysis by -
glucuronidase, we have used choloylglycine hydrolase to deconjugate the
predominant amidated conjugates of 1 -OH-DCA and DCA in urine, followed by
solvolysis to complete desulfation. The urinary concentrations of DCA after enzymatic
hydrolysis and solvolysis were in the range of that reported for total DCA
concentrations in urine (Bathena et al., 2015a). The concentration ratio of 1 -OH-
DCA/DCA was used as a marker for CYP3A activity instead the individual
concentration. Using this method, the impact of variable DCA levels, as a substrate
for 1 -hydroxylation, can be minimized. Our results showed the ratio of 1 -OH-
DCA/DCA was 7-fold higher in the urine from a patient with elevated CYP3A activity
than that in the control sample. It is interesting to note that in performing only the mild
enzymatic hydrolysis without solvolysis, the ratio of 1 -OH-DCA/DCA in the patient
urine was again 7-fold higher than that in the control, though the concentrations of
DCA and 1 -OH-DCA obtained were lower than those in samples treated by an
additonal chemical solvolysis step. This observation indicated that the improved

DMD Fast Forward. Published on July 11, 2016 as DOI: 10.1124/dmd.116.070805
This article has not been copyedited and formatted. The final version may differ from this version.
enzymatic deconjugation method may be robust enough to provide data that
differentiates between ratios of 1 -OH-DCA/DCA observed with differing levels of
CYP3A activity.
In this study, the relative contribution of CYP3A4 versus CYP3A5 to DCA 1 -
hydroxylation has not been studied in detail and needs further investigation.
However, screening data from 21 rCYPs showed that 1 -OH-DCA formation by
rCYP3A4 was approximately 50- and 10-times higher than rCYP3A5 at DCA
concentrations of 2 and 20 M, respectively, and highlights the major contribution of
CYP3A4 compared with CYP3A5. The role of active transport of 1 -OH-DCA and its
conjugates into urine has not been fully explored. Future studies validating 1 -OH-
DCA as a CYP3A biomarker need to consider possible effects of inhibitors and
inducers on transporter functions as possible confounding factors.
The value of using 1 -OH-DCA as a urine biomarker for CYP3A activity needs to be
further studied in a large set of clinical samples to understand its variability over time
and intersubject differences. It is clear that if the analysis of spot urine samples
reveals human CYP3A activity, this would provide an obvious advantage in clinical
DDI studies. Urine sampling over an extended period of time could be avoided, as is
the case when using the biomarker 6β-hydroxycortisol. The 6β-
hydroxycortisol/cortisol ratios exhibit large diurnal variability in urinary levels thereby
necessitating urine collection over a long period of time to obtain samples
representing CYP3A activity (Shin et al., 2013).
The dynamic range of a biomarker should be large enough to clearly display effects
on CYP3A activity both by enzyme inhibitors and inducers. Ideally the change in 1 -
OH-DCA in spot urine should reflect the changes in CYP3A activity in subjects

DMD Fast Forward. Published on July 11, 2016 as DOI: 10.1124/dmd.116.070805
This article has not been copyedited and formatted. The final version may differ from this version.
treated with inhibitors or inducers. A carefully designed clinical study is needed to
optimize treatment and sampling schedules to evaluate this biomarker. Generally
potent inhibitors and inducers of cytochrome CYP3A are excluded from further
development based on preclinical evaluation. However, some compounds with the
potential for clinically relevant drug interactions do advance into human clinical
studies and some are eventually approved for use in patients. A useful biomarker
should be able to reliably detect and estimate the level of changes in CYP3A activity
by weak to moderate inhibitors and inducers. A disadvantage of using 4 -
hydroxycholesterol as a biomarker is the small dynamic range which is affected by
CYP3A inhibition. After 14 days treatment by the strong CYP3A inhibitor
ketoconazole, plasma 4 -hydroxycholesterol levels decreased by a maximum of 13%
while oral midazolam levels increased 11-fold (Kasichayanula et al., 2014)
highlighting the low dynamic range of this biomarker.
The plasma biomarker 4 -hydroxycholesterol was previously found to be sensitive to
the CYP3A5 genotype (Diczfalusy et al., 2008). Based on the results from expressed
CYP enzymes the contribution of CYP3A5 to the formation of 1 -OH-DCA is probably
minor but should be investigated in subjects representing the CYP3A5 slow and
extensive metabolising genotypes.
In conclusion we have shown that 1 -hydroxydeoxycholic acid is specifically formed
by CYP3A4/5/7 and may have potential as an in vivo CYP3A biomarker, replacing
drug probes currently used. A sensitive UPLC/HRMS method was established and an
efficient BM3 P450 mutant catalysed synthesis generated sufficient amounts of the
metabolite required for ongoing clinical studies.

DMD Fast Forward. Published on July 11, 2016 as DOI: 10.1124/dmd.116.070805
This article has not been copyedited and formatted. The final version may differ from this version.
Acknowledgements
We thank Dr. Ralf Nilsson for fruitful scientific discussions about the analysis and metabolism
of bile acids.

DMD Fast Forward. Published on July 11, 2016 as DOI: 10.1124/dmd.116.070805
This article has not been copyedited and formatted. The final version may differ from this version.
Authorship Contributions
Participated in research design: Hayes, Li, Diczfalusy and Andersson
Conducted experiment: Hayes, Grönberg and Li
Performed data analysis: Hayes, Grönberg and Li
Wrote or contributed to the writing of the manuscript: Hayes, Li, Grönberg and
Andersson

DMD Fast Forward. Published on July 11, 2016 as DOI: 10.1124/dmd.116.070805
This article has not been copyedited and formatted. The final version may differ from this version.
References
Alme B, Bremmelgaard A, Sjovall J, and Thomassen P (1977) Analysis of metabolic
profiles of bile acids in urine using a lipophilic anion exchanger and
computerized gas-liquid chromatorgaphy-mass spectrometry. J Lipid Res
1
8:339-362.
Bathena SP, Thakare R, Gautam N, Mukherjee S, Olivera M, Meza J, and Alnouti Y
2015a) Urinary bile acids as biomarkers for liver diseases I. Stability of the
(
baseline profile in healthy subjects. Toxicol Sci 143:296-307.
Bathena SP, Thakare R, Gautam N, Mukherjee S, Olivera M, Meza J, and Alnouti Y
(2015b) Urinary bile acids as biomarkers for liver diseases II. Signature
profiles in patients. Toxicol Sci 143:308-318.
Bjorkhem-Bergman L, Backstrom T, Nylen H, Ronquist-Nii Y, Bredberg E, Andersson
TB, Bertilsson L, and Diczfalusy U (2013) Comparison of endogenous 4beta-
hydroxycholesterol with midazolam as markers for CYP3A4 induction by
rifampicin. Drug Metab Dispos 41:1488-1493.
Bodin K, Lindbom U, and Diczfalusy U (2005) Novel pathways of bile acid
metabolism involving CYP3A4. Biochim Biophys Acta 1687:84-93.
Bron B, Waldram R, Silk DB, and Williams R (1977) Serum, cerebrospinal fluid, and
brain levels of bile acids in patients with fulminant hepatic failure. Gut 18:692-
696.
Carlström K, Kirk DN, and Sjovall J (1981) Microbial synthesis of 1 beta- and 15 beta-
hydroxylated bile acids. J Lipid Res 22:1225-1234.
Diczfalusy U, Kanebratt KP, Bredberg E, Andersson TB, Bottiger Y, and Bertilsson L
(2009) 4beta-hydroxycholesterol as an endogenous marker for CYP3A4/5

DMD Fast Forward. Published on July 11, 2016 as DOI: 10.1124/dmd.116.070805
This article has not been copyedited and formatted. The final version may differ from this version.
activity. Stability and half-life of elimination after induction with rifampicin. Br J
Clin Pharmacol 67:38-43.
Diczfalusy U, Miura J, Roh HK, Mirghani RA, Sayi J, Larsson H, Bodin KG, Allqvist A,
Jande M, Kim JW, Aklillu E, Gustafsson LL, and Bertilsson L (2008) 4Beta-
hydroxycholesterol is a new endogenous CYP3A marker: relationship to
CYP3A5 genotype, quinine 3-hydroxylation and sex in Koreans, Swedes and
Tanzanians. Pharmacogenet Genomics 18:201-208.
Dutheil F, Dauchy S, Diry M, Sazdovitch V, Cloarec O, Mellottee L, Bieche I,
Ingelman-Sundberg M, Flinois JP, de Waziers I, Beaune P, Decleves X,
Duyckaerts C, and Loriot MA (2009) Xenobiotic-metabolizing enzymes and
transporters in the normal human brain: regional and cellular mapping as a
basis for putative roles in cerebral function. Drug Metab Dispos 37:1528-1538.
Galteau MM and Shamsa F (2003) Urinary 6beta-hydroxycortisol: a validated test for
evaluating drug induction or drug inhibition mediated through CYP3A in
humans and in animals. Eur J Clin Pharmacol 59:713-733.
Gillam EM and Hayes MA (2013) The evolution of cytochrome P450 enzymes as
biocatalysts in drug discovery and development. Curr Top Med Chem
13:2254-2280.
Gustafsson J, Andersson S, and Sjövall J (1985) Bile acid metabolism during
development: Metabolism of taurodeoxycholic acid in human fetal liver.
Neonatology 47:26-31.
Kakiyama G, Muto A, Takei H, Nittono H, Murai T, Kurosawa T, Hofmann AF,
Pandak WM, and Bajaj JS (2014) A simple and accurate HPLC method for
fecal bile acid profile in healthy and cirrhotic subjects: validation by GC-MS
and LC-MS. J Lipid Res 55:978-990.

DMD Fast Forward. Published on July 11, 2016 as DOI: 10.1124/dmd.116.070805
This article has not been copyedited and formatted. The final version may differ from this version.
Kanebratt KP, Diczfalusy U, Backstrom T, Sparve E, Bredberg E, Bottiger Y,
Andersson TB, and Bertilsson L (2008) Cytochrome P450 induction by
rifampicin in healthy subjects: determination using the Karolinska cocktail and
the endogenous CYP3A4 marker 4beta-hydroxycholesterol. Clin Pharmacol
Ther 84:589-594.
Kasichayanula S, Boulton DW, Luo WL, Rodrigues AD, Yang Z, Goodenough A, Lee
M, Jemal M, and LaCreta F (2014) Validation of 4beta-hydroxycholesterol and
evaluation of other endogenous biomarkers for the assessment of CYP3A
activity in healthy subjects. Br J Clin Pharmacol 78:1122-1134.
Kiss FM, Lundemo MT, Zapp J, Woodley JM, and Bernhardt R (2015) Process
development for the production of 15beta-hydroxycyproterone acetate using
Bacillus megaterium expressing CYP106A2 as whole-cell biocatalyst. Microb
Cell Fact 14:28.
Kitada M, Kamataki T, Itahashi K, Rikihisa T, Kato R, and Kanakubo Y (1985)
Purification and properties of cytochrome P-450 from homogenates of human
fetal livers. Arch Biochem Biophys 241:275-280.
Lacroix D, Sonnier M, Moncion A, Cheron G, and Cresteil T (1997) Expression of
CYP3A in the human liver--evidence that the shift between CYP3A7 and
CYP3A4 occurs immediately after birth. Eur J Biochem 247:625-634.
Reinen J, Vredenburg G, Klaering K, Vermeulen NPE, Commandeur JNM, Honing M,
and Vos JC (2015) Selective whole-cell biosynthesis of the designer drug
metabolites 15- or 16-betahydroxynorethisterone by engineered Cytochrome
P450 BM3 mutants. J Mol Catal B-Enzym 121:64-74.

DMD Fast Forward. Published on July 11, 2016 as DOI: 10.1124/dmd.116.070805
This article has not been copyedited and formatted. The final version may differ from this version.
Rendic S and Guengerich FP (2015) Survey of human oxidoreductases and
cytochrome P450 enzymes involved in the metabolism of xenobiotic and
natural chemicals. Chem Res Toxicol 28:38-42.
Sawayama AM, Chen MM, Kulanthaivel P, Kuo MS, Hemmerle H, and Arnold FH
(2009) A panel of cytochrome P450 BM3 variants to produce drug metabolites
and diversify lead compounds. Chemistry 15:11723-11729.
Shin KH, Choi MH, Lim KS, Yu KS, Jang IJ, and Cho JY (2013) Evaluation of
endogenous metabolic markers of hepatic CYP3A activity using metabolic
profiling and midazolam clearance. Clin Pharmacol Ther 94:601-609.
Tohma M, Mahara R, Takeshita H, Kurosawa T, and Ikegawa S (1986) Synthesis of
the 1 beta-hydroxylated bile acids, unusual bile acids in human biological
fluids. Chem Pharm Bull (Tokyo) 34:2890-2899.
Venkataraman H, Te Poele EM, Rosloniec KZ, Vermeulen N, Commandeur JN, van
der Geize R, and Dijkhuizen L (2015) Biosynthesis of a steroid metabolite by
an engineered Rhodococcus erythropolis strain expressing a mutant
cytochrome P450 BM3 enzyme. Appl Microbiol Biotechnol 99:4713-4721.
Williams JA, Ring BJ, Cantrell VE, Jones DR, Eckstein J, Ruterbories K, Hamman
MA, Hall SD, and Wrighton SA (2002) Comparative metabolic capabilities of
CYP3A4, CYP3A5, and CYP3A7. Drug Metab Dispos 30:883-891.