Proluciferin acetals as bioluminogenic substrates for cytochrome P450 activity and probes for CYP3A inhibition

Article abstract of DOI:10.1124/dmd.111.041541

Cytochrome P450 (P450) assays use probe substrates to interrogate the influence of new chemical entities toward P450 enzymes. We report the synthesis and study of a family of bioluminogenic luciferin acetal substrates that are oxidized by P450 enzymes to form luciferase substrates. The luciferin acetals were screened against a panel of purified P450 enzymes. In particular, one proluciferin acetal has demonstrated sensitive and selective CYP3A4-catalyzed oxidation to a luciferin ester - K_m and k_cat are 2.88 μM and 5.87 pmol metabolite·min^-1·pmol enzyme ^-1, respectively. The proluciferin acetal was used as a probe substrate to measure IC₅₀ values of known inhibitors against recombinant CYP3A4 or human liver microsomes. IC₅₀ values for the known inhibitors correlate strongly with IC₅₀ values calculated from the traditional high-performance liquid chromatography-based probe substrate testosterone. Luciferin acetals are rapidly oxidized to unstable hemi-orthoesters by CYP3A resulting in luciferin esters and, therefore, are conducive to simple rapid CYP3A bioluminescent assays. Copyright

Full text of DOI:10.1124/dmd.111.041541

Supplemental material to this article can be found at:
http://dmd.aspetjournals.org/content/suppl/2011/09/02/dmd.111.041541.DC1.html
0090-9556/11/3912-2403–2410$25.00
DRUG METABOLISM AND DISPOSITION
Copyright © 2011 by The American Society for Pharmacology and Experimental Therapeutics
DMD 39:2403–2410, 2011
Vol. 39, No. 12
41541/3729823
Printed in U.S.A.
Proluciferin Acetals as Bioluminogenic Substrates for Cytochrome
□
S
P450 Activity and Probes for CYP3A Inhibition
Poncho L. Meisenheimer, H. Tetsuo Uyeda, Dongping Ma, Mary Sobol, Mark G. McDougall,
Cesear Corona, Dan Simpson, Dieter H. Klaubert, and James J. Cali
Promega Biosciences, San Luis Obispo, California (P.L.M., H.T.U., M.G.M., C.C., D.H.K.); and Promega Corporation, Madison,
Wisconsin (D.M., M.S., J.J.C., D.S.)
Received July 11, 2011; accepted September 2, 2011
ABSTRACT:
Cytochrome P450 (P450) assays use probe substrates to interro- The proluciferin acetal was used as a probe substrate to measure
gate the influence of new chemical entities toward P450 enzymes. IC50 values of known inhibitors against recombinant CYP3A4 or
We report the synthesis and study of a family of bioluminogenic human liver microsomes. IC50 values for the known inhibitors cor-
luciferin acetal substrates that are oxidized by P450 enzymes to relate strongly with IC50 values calculated from the traditional
form luciferase substrates. The luciferin acetals were screened high-performance liquid chromatography-based probe substrate
against a panel of purified P450 enzymes. In particular, one prolu- testosterone. Luciferin acetals are rapidly oxidized to unstable
ciferin acetal has demonstrated sensitive and selective CYP3A4- hemi-orthoesters by CYP3A resulting in luciferin esters and, there-
catalyzed oxidation to a luciferin ester—Km and kcat are 2.88 ␮M fore, are conducive to simple rapid CYP3A bioluminescent assays.
and 5.87 pmol metabolite ꢀ min؊1 ꢀ pmol enzyme؊1, respectively.
Introduction
demonstrates substrate-dependent responses to certain analytes (Wang
et al., 2000). Even traditional probes such as nifedipine, midazolam,
and testosterone interrogate NCEs in a substrate-dependent manner.
Bioluminescent luciferase substrates produce photons in a lu-
ciferase-catalyzed reaction and have been used extensively as tools in
structural biology and drug development (Fan and Wood, 2007).
Numerous structure/activity articles (White et al., 1965; White and
Woerther, 1966; Branchini et al., 1989; Woodroofe et al., 2008) have
outlined the structural requirements of bioluminescent luciferase sub-
strates. In contrast to a bioluminescent substrate, a bioluminogenic
substrate is a luciferase prosubstrate that does not produce light with
luciferase but can be converted to a light-generating bioluminescent
substrate by an enzyme or system of interest. The bioluminescent
product of the reaction of interest can then be detected as a lumines-
cent signal in a second reaction with firefly luciferase. Assays based
on bioluminogenic substrates have been reported for proteases
(O’Brien et al., 2005), oxidases (Zhou et al., 2006b), phosphatases
(Zhou et al., 2008), glutathione transferases (Zhou et al., 2006a), and
glycosidases (Cali et al., 2008). In the majority of these cases, the
bioluminogenic substrates have been modified at the 6Ј-carbon atom
of luciferin with a recognition moiety for the enzyme of interest (see
Scheme 1 for luciferin numbering).
Cytochrome P450 (P450) enzymes oxidize endogenous com-
pounds, drugs, new chemical entities (NCEs), and other xenobiotics
(Guengerich, 2007). P450 enzyme inhibition or CYP gene induction
by drugs can affect the elimination of concomitantly administered
drugs thereby resulting in adverse drug-drug interactions (Wienkers
and Heath, 2005). Fatalities due to profound drug-drug interactions
are not uncommon (Lazarou et al., 1998; Ko¨hler et al., 2000), and
efforts to identify P450 active NCEs early in the drug discovery
process continue (Zientek et al., 2008; Grime et al., 2009).
In general, P450 assays use probe substrates to interrogate the
influence of NCE’s toward P450 enzymes. Probes based on known
drugs are traditionally detected by radiometric assays (Stresser et al.,
1996; Moody et al., 1999) and mass spectrometry, yet these assays are
disadvantaged by concerns over the safety and disposal of radioactive
isotopes or by low throughput, respectively. Chromogenic, fluoro-
genic (Friden et al., 2006), and bioluminogenic probes (Cali et al.,
2006) are rapid and sensitive, yet concerns regarding their analogy to
traditional drug probes (Bell et al., 2008) must be addressed by
empirical correlation to traditional probes. A complexity in interpret-
ing correlative studies is associated in particular with CYP3A4, which
The electronic characteristics of substrate molecules play a decisive
role in the site of oxidation in P450-catalyzed reactions. A strong
correlation has been demonstrated between the oxidation site in a
P450 reaction and the most stable radical site in the organic substrates
(Jones et al., 2002). The relative stabilities of the possible radicals
formed on an organic molecule relate to the homolytic bond dissoci-
ation enthalpy of the carbon-hydrogen bond. Oxygens impart radical
This work was financially supported by Promega Corporation and Promega
Biosciences.
Article, publication date, and citation information can be found at
http://dmd.aspetjournals.org.
doi:10.1124/dmd.111.041541.
□S The online version of this article (available at http://dmd.aspetjournals.org)
contains supplemental material.
ABBREVIATIONS: P450, cytochrome P450; NCE, new chemical entity; HPLC, high-performance liquid chromatography; RLU, relative light units;
HLMs, human liver microsomes; KPO4, potassium phosphate buffered to the stated pH; LOD, limit of detection.
2403

2404
MEISENHEIMER ET AL.
SCHEME 3. Synthesis of D-luciferin isopropyl acetal (compound 1) from D-cysteine.
formulation supplemented with a carboxyl esterase (Cali et al., 2006).
In such a formulation, any luciferin acetal oxidized by a P450 accord-
ing to the proposed scheme would convert to the luminogenic form
and produce light with luciferase.
We report the synthesis and study of a family of bioluminogenic
luciferin acetals that are converted to luciferase substrates selec-
tively by the CYP3A enzymes. This study examines the activity of
the acetals toward a panel of P450 enzymes, and we examine a
subset of acetals as bioluminescent probes for CYP3A4 inhibition
measurements.
Materials and Methods
Synthesis. A series of acetals were synthesized to examine the structure/
bioluminogenic profile with P450 enzymes (Scheme 2). D-cysteine was con-
verted to compound 1 in six steps (Scheme 3). D-cysteine was allowed to react
with acetone to form the thiazolidine, which was subsequently protected with
di-t-butyldicarbonate. The carboxylic acid was converted to the Weinreb amide
and subsequently reduced to the aldehyde with lithium aluminum hydride to
yield 20. The aldehyde was refluxed in 2-propanol and HCl. The intermediate
was allowed to react with the nitrile of 5b to yield 1. Tris(2-carboxyethyl-
)phosphine was added to prevent any disulfide formation during the reaction.
As a control, DL-cysteine was used in a parallel synthesis to generate the
racemized version of 1. Chiral high-performance liquid chromatography
(HPLC) was used to show that the reaction starting with D-cysteine results in
enantiomerically pure 1 (data not shown).
SCHEME 1. Proposed mechanism for bioluminogenic P450 activity of proluciferin
acetals.
A series of dimethyl acetals based on alternative luciferase substrates were
synthesized from the thiazolidine aldehyde 20 (Scheme 4). The dimethyl acetal
of thiazolidine aldehyde 22 was condensed with various benzothiazole nitriles
(2b–5b). Under aqueous conditions, compound 22 was proposed to establish
equilibrium between the thiazolidine ring and a ring-opened ␤-mercaptoamine;
stability to neighboring carbons, and P450-catalyzed O-dealkylations
result from the oxidation of these carbons. However, the oxygens of
an acetal group impart significant radical stability on the carbon atom
sandwiched between them. Bond enthalpies for the C-H bond of a the ␤-mercaptoamine rapidly reacts with the nitriles 2b to 5b to form the
thiazoline products (2–5).
carbon between the oxygens in an acetal are some 10 kcal/mol lower
than the corresponding C-H bonds of an ether (Blanksby and Ellison,
2003). Correspondingly, P450 enzymes should oxidize acetals at the
carbon between the two oxygens thereby forming an unstable hemi-
orthoester, which rapidly decomposes to an ester. This expectation
suggested that acetal derivatives of D-luciferin could serve as biolu-
minogenic P450 substrates according to the hypothetical scheme
shown (Scheme 1). The scheme recognizes that luciferin esters do not
react with luciferase to produce light but are rapidly and quantitatively
Aryl containing acetals (8–10), dioxanes, and dithianes (11–13) were con-
structed in a similar manner (Scheme 5). The acetals containing basic amines
(6, 7, 14) were synthesized from a trans-acetalization of 5 with the appropriate
halogenated alcohol followed by substitution with dimethylamine (Scheme 6).
Detailed experimental protocols are included in supplemental materials.
Screening of P450s. Recombinant P450 enzymes in microsomal fractions
(Supersomes) were purchased from BD Biosciences Discovery Labware (Bed-
ford, MA). Each P450 enzyme is coexpressed with cytochrome P450 reduc-
tase; cytochrome b5 is also coexpressed with some of the P450s (CYP2A6,
de-esterified to the luminogenic carboxylic acid form in a luciferase -2B6, -2C8, -2C9, -2C19, -2E1, -2J2, -3A4, -3A5, -3A7, -4F2, -4F3A, -4F3B,
SCHEME 2. Series of proluciferin acetals tested against P450 activity.

PROLUCIFERIN ACETALS AS CYP3A PROBES
2405
SCHEME 4. Dimethyl acetal derivatives of N-alkyl aminoluciferins were synthe-
sized by converting 20 (Scheme 3) to the isolated intermediate 22 followed by the
condensation with previously reported benzothiazole nitriles 2b to 5b.
and -4F12). The P450-negative controls are Supersomes devoid of P450
enzyme expression. To screen the proluciferin acetal compounds as P450
substrates, they were first prepared as 50 mM stock solutions in dimethyl
sulfoxide and then combined at 50 ␮M with P450 enzymes at 20 nM recom-
SCHEME 6. Mixed acetals and amine containing acetals were synthesized from 5.
binant P450 enzyme and an NADPH-generating system (1.3 mM NADPϩ, 3.3 calculate IC50 values. Inhibition reactions were performed in 100 mM KPO4
mM glucose 6-phosphate, 3.3 mM MgCl2, and 0.4 U/ml glucose-6-phosphate
dehydrogenase) in an appropriate buffer as recommended by the P450 enzyme
buffer, pH 7.4 with 20 nM recombinant CYP3A4 and 10 ␮g/ml HLMs using
50 ␮l and 100 ␮l volumes, respectively, for the compound 1 and testosterone
manufacturer (200 mM KPO4, pH 7.4 for CYP3A4, 100 mM KPO4, pH 7.4 for assays. Testosterone reactions were performed as described previously (Ken-
CYP1A1, -1A2, -1B1, -2D6, -2E1, -3A5, -3A7, -2J2, -4F12, and -19; 50 mM worthy et al., 1999). Compound 1 concentrations of 3 ␮M for recombinant
KPO4, pH 7.4 for CYP2B5, -2C8, -2C19, -4F2, -4F3A, and -4F3B; 25 mM
P450 or 8 ␮M for HLMs were used. Testosterone concentrations of 50 ␮M for
KPO4, pH 7.4 for CYP2C9, 100 mM Tris-HCl, pH 7.5 for CYP2A6, -2C18, recombinant P450 or 10 ␮M for HLMs were used. Substrates and inhibitors
and -4A11). The 50 mM luciferin acetal stock solutions contributed 0.1%
dimethyl sulfoxide to each reaction mixture. Reactions were performed in a
volume of 50 ␮l in opaque white 96-well plates that were incubated in a 37°C
H2O bath for 30 min. Reactions were initiated by addition of the NADPH-
generating system. Reactions were stopped, and luminescence was initiated by
adding 50 ␮l of P450-Glo Luciferin Detection Reagent (Promega Corporation,
were allowed to incubate for 10 min at 37°C with CYP3A4 or HLMs before
initiating reactions by adding the NADPH-generating system. The reactions
with compound 1 were stopped, and luciferin product was detected by lumi-
nescence in the same manner as described for P450 screening. The reactions
with testosterone were stopped by addition of 50 ␮l of acetonitrile and then
centrifuged at 10,000g for 3 min. The supernatant (50 ␮l) was injected into a
Madison, WI) supplemented with 20 units/ml porcine esterase (Sigma-Aldrich, 3 ϫ 100 mm, 3u C18 HPLC column and separated with a mobile phase
St. Louis, MO). After 20 min at room temperature, luminescence was read as
initially of 15% acetonitrile increasing to 100% acetonitrile over 12 min, at a
relative light units (RLU) on a plate reading luminometer [Polarstar Optima (BMG flow rate of 0.75 ml/min. The CYP3A4 reaction product 6␤-hydroxytestoster-
Labtech, Durham, NC) or Veritas (Promega Biosystems, Sunnyvale, CA)].
one was detected by its absorbance at 240 nm (HPLC retention time ϭ 5 min)
HPLC Analysis of CYP3A4-Catalyzed Oxidation of 1. Fluorescence and quantitated by comparison to absorbance of authentic 6␤-hydroxytestos-
chromatograms (Supplemental Fig. 1) were used for the analysis of compound terone standards.
1 with a control Sf9 membrane preparation, with CYP3A4 membranes, and
with CYP3A4 membrane preparation with esterase treatment. Luciferin at 2.2
min was identified with an authentic standard. A peak at 7.5 min disappeared
Results
Screening Cytochrome P450s for Activity with Proluciferin
on treatment with esterase and was therefore assigned as the putative luciferin
ester intermediate. Separation was obtained on an Adsorbosphere HS C18 Acetals. A series of proluciferin acetals (Scheme 2) were screened for
column (5u, 150 ϫ 4.6 mm; Grace Davison Discovery Sciences, Deerfield, IL)
with a 50 mM ammonium acetate buffer (pH6) and linearly modified aceto-
nitrile gradient. Component elution was monitored with fluorescence by means
of 330 nm excitation and 520 nm emission.
Measuring P450 Inhibition. Several known CYP3A4 inhibitors were ap-
plied in a dose-response mode to reactions of recombinant CYP3A4 or human
liver microsomes (HLMs) with compound 1 or with the CYP3A4 substrate
testosterone. The HLMs were a mixed gender pool from nine male and six
female donors (BD Biosciences Discovery Labware). Reaction conditions for
interrogation of inhibitors were established as described under Results, and
bioluminogenic activity against a panel of recombinant human P450
enzymes in microsomal fractions (Fig. 1). A moderate to high sub-
strate concentration of 50 ␮M was used with 20 nM P450 enzymes
and a liberal reaction time of 30 min at 37°C. A control microsome
fraction with no recombinant P450 expression was used to monitor
intrinsic instability or non-P450-catalyzed conversion of the prolu-
ciferin acetals to luciferin. The bioluminogenic reactions are per-
formed by incubating the P450 enzymes with proluciferin P450 sub-
strates and NADPH in a first reaction where an active enzyme/
GraphPad Prism (GraphPad Software Inc., San Diego, CA) was used to substrate combination generates luciferin ester. In a second reaction,
SCHEME 5. Aryl containing acetal derivatives, dioxanes, and dithi-
anes were synthesized from intermediate 20.

2406
MEISENHEIMER ET AL.
FIG. 1. Recombinant P450 (CYP) enzyme screen against proluciferin acetals. Compounds 1 to 18 are the respective acetals illustrated in Scheme 2 exposed at 50 ␮M to
20 nM recombinant P450 enzymes for 30 min at 37°C. 1a is compound 1 exposed at 3 ␮M to 2 nM recombinant P450 enzymes for 10 min at 37°C.
carboxyl esterase and luciferase convert any luciferin ester to a acetal and its CYP3A4 reaction products with and without addition of
proportional number of photons. P450 activity is detected on a lumi- carboxyl esterase. Incubation of 1 with CYP3A4 gave rise to two
nometer as RLU. Other than 13, compounds 1 to 18 showed only a apparent reaction product peaks (supplemental data). The smaller of
minor luminescence signal in the control, indicating that non-P450 the two peaks comigrated with an authentic D-luciferin standard.
enzymes and/or uncatalyzed conversions of the proluciferin do not Esterase treatment shifted the second intermediate peak to the position
contribute to the signals. The diisopropyl acetal 1 displayed the largest of D-luciferin. This analysis is consistent with the model of Scheme 1
CYP3A4 signal. CYP3A5 activity was highest for 16, 17, and 18. where the P450 enzyme initiates a reaction sequence that yields a
CYP3A7 activity was highest for 8. Compounds 2, 7, and 14, gave no luciferin ester that is processed to luciferin by added esterase. The
light in excess of the P450-negatvie control. The 1,3 dioxane 11 appearance of a small D-luciferin peak in the absence of added
showed some CYP3A activity along with detectable CYP1A and -1B esterase is likely due to esterase activity already present in the
activity. The dithiane 13 displayed CYP3A5 and CYP3A7 activity CYP3A4 microsome preparation, and this was confirmed by observ-
only with substantial light in the control, indicating some level of ing elimination of the peak by pretreatment of the microsomes with
luciferin contamination in the sample. Fourteen of 18 proluciferin the carboxylesterase inhibitor phenylmethylsulfonyl fluoride (data not
acetals were biolumingenic primarily against CYP3A. Detailed shown). The absence of luciferin and luciferin ester peaks after
CYP3A net luminescent signals were extracted from Fig. 1 and incubation with control membranes lacking P450 enzyme confirms
displayed with error bars in 2 dimensions for better visualization that conversion of 1 to these products is P450-dependent. Noncatalytic
(Fig. 2).
or non-P450 enzymatic conversions of 1 do not occur to a significant
Characterization of P450 Activity with Compound 1. The diiso- extent in the Supersome system.
propyl acetal 1 was studied in further detail because it displayed the
Kinetic Characterization of CYP3A4-Catalyzed Oxidation of
largest signal with CYP3A4, an enzyme of particular interest because Compound 1. An initial test that varied the concentration of the
of its prominent role in drug metabolism. Preferential activity with common P450 reaction buffer KPO4 (pH 7.4) showed the CYP3A4
CYP3A4, -3A5, and -3A7 suggested this compound could be used as reaction with compound 1 was optimal at 100 mM KPO4 (data not
a selective CYP3A probe. The recombinant CYP3A4 reaction with 1 shown), and this concentration was used for subsequent characteriza-
was performed using fluorescence HPLC to detect the proluciferin tions. The reaction was characterized by varying CYP3A4 enzyme

PROLUCIFERIN ACETALS AS CYP3A PROBES
2407
FIG. 2. Recombinant CYP3A enzyme screen against proluciferin acetals. Net luminescence values are signals from the P450 reactions shown in Fig. 1, minus the respective
P450-independent controls (n ϭ 3, Ϯ S.D.).
concentration, the initial concentration of 1, and incubation time. Plots of detection (LOD) for the CYP3A4 enzyme was estimated at 0.04
of time-dependent (Fig. 3, A–C) and enzyme concentration-dependent nM. LOD is defined as the concentration in which the luminescent
(Fig. 3, E–H) signal increases showed initial linear phases with value falls within three S.D.s of the background.
decreases in linearity at longer time points, increased enzyme con-
These steady-state CYP3A4 conditions were also used in a repeat
centration, and decreased initial concentrations of 1, indicating that 1 screen of compound 1 against the recombinant human P450 enzyme
might become limiting in the nonlinear phases. The dose-dependent panel, where a substantial increase in selectivity for CYP3A4 over
activity curves of compound 1 were fit to a hyperbola for each of the CYP3A5 and -3A7 was observed compared with the conditions of the
various reaction times and enzyme concentrations (Fig. 3, I–K). The initial screen with 20 nM enzyme, 50 ␮M substrate, and 30-min
average of the Km values derived from the curve fits for 12 sets of incubations (Fig. 1). This preference of compound 1 for CYP3A4 over
conditions was 2.78 Ϯ 0.80 ␮M. The substrate concentration curve CYP3A5 and -3A7 is explained in terms of the Km, kcat, and enzy-
was repeated under linear conditions of reaction time and enzyme matic efficiencies (kcat/Km) of the three enzymes (Fig. 3, D and L;
concentration (10 min, 2 nM CYP3A4) (Fig. 3L). In this case, Table 1). Ratios of the enzymatic efficiencies, CYP3A4/CYP3A5 or
CYP3A4 reaction rates were calculated after converting RLU values CYP3A4/CYP3A7, indicate a 9-fold and 102-fold selective prefer-
to luciferin product concentrations by interpolation from a luciferin ence for CYP3A4 over the other two enzymes.
standard curve (data not shown). Under these conditions the curve fit
Inhibition Assays. Compound 1 (Luciferin-IPA; Promega) was
found Km and kcat values of 2.88 Ϯ 0.18 ␮M and 5.87 Ϯ 0.14 pmol used as a probe substrate to measure IC50 values of known inhibitors
metabolite ꢀ minϪ1 ꢀ pmol enzymeϪ1, respectively, for CYP3A4 against recombinant CYP3A4 or HLMs. The linear reaction condi-
(Table 1). A contribution from substrate depletion to the loss of tions for recombinant CYP3A4 at substrate Km were used (3 ␮M
linearity observed at lower substrate concentration, higher enzyme compound 1, 2 nM CYP3A4, 10-min reactions). Recombinant
concentration, and greater incubation times is consistent with the CYP3A4 inhibition was also measured against the commonly used
reaction rates calculated for Fig. 3L. In contrast, steady-state reaction testosterone 6␤-hydroxylation reaction (Kenworthy et al., 1999;
conditions can be assumed in 10 min reactions with 2 nM CYP3A4 at Wang et al., 2000), and IC50 values from this study were compared
Km substrate concentration where we estimate Ͻ2% of the substrate is with those determined with compound 1. Conditions for the testos-
consumed, and these conditions are reasonable for interrogation terone assays were first established so that reactions were performed
against CYP3A4 inhibitors. Figure 3F includes a CYP3A4 enzyme at substrate Km (50 ␮M) under conditions of linearity with respect to
titration curve performed under nearly identical conditions: 10 min time and CYP3A4 concentration (10 min, 20 nM). The IC50 values for
with 2.5 ␮M compound 1. From a linear regression analysis, the limit known CYP3A4 inhibitors correlated closely between the compound

2408
MEISENHEIMER ET AL.
FIG. 3. Characterization of the CYP3A4 reaction with compound 1. Fifty-microliter reactions in duplicate or triplicate were performed at 37°C. Recombinant CYP3A4
activity was scored as RLU after adding a luciferase reaction mixture for detection of the compound 1/CYP3A4 reaction product. A–C, time-dependent luminescence; E–H,
CYP3A4 concentration-dependent luminescence; I–K, compound 1 dose-dependent luminescence. For D, recombinant CYP3A5 or -3A7 were substituted for CYP3A4. For
D and L, rates of D-luciferin production were calculated after conversion of RLU to D-luciferin concentration by comparison to a D-luciferin standard generated with the
luciferase reaction mixture.
1 assay and testosterone assays (Fig. 4A). A similar comparison was
performed using HLMs in place of recombinant CYP3A4 enzyme.
Reaction conditions for interrogation of inhibitors were first estab-
lished and found to be somewhat different from those observed with
recombinant CYP3A4. For compound 1, an apparent Km concentra-
tion of 8 ␮M was observed and used in 10-min reactions with 20
␮g/ml HLMs. It was also noted that significant luminescent signals
were not observed in HLMs when NADPH was withheld from the
reactions (data not shown). This was consistent with a P450-depen-
dent mechanism and the absence of noncatalytic or non-P450 enzyme-
dependent substrate conversion. Using testosterone as a substrate, an
apparent Km of 50 ␮M was observed and used in 10-min reactions
with microsomes at 20 ␮g/ml. As with recombinant CYP3A4, a close
correlation of IC50 values between the testosterone and compound 1
assay was observed for the eight inhibitors tested (Fig. 4B).
Discussion
The P450-catalyzed oxidation of proluciferin acetals results in
bioluminescent luciferase substrates. Scheme 1 describes the proposed
mechanism by which light is produced from P450 activity. The
oxidation of the acetal yields an unstable hemi-orthoester, which is
proposed to decompose to the luciferin ester. The resulting luciferin
ester is converted to D-luciferin by esterase, and its concentration is
measured using the bioluminescent output from a luciferase reaction.
Light output is proportional to P450 activity.
The acetals were designed to explore the structural effects of
hydrophobicity (1, 5, 9, 11, 12, 15) and charge (6, 7, 14) on the P450
activity. Mixed acetals (10, 16–18) and a dithiane (13) were also
produced. Furthermore, alkyl aminoluciferin luciferase substrates
(Woodroofe et al., 2008) have been reported, and a series of related
acetals (2, 3, 4) were generated. The screen of P450 enzymes was
designed to push the system to detect any reasonable activity.
CYP3A converts 1 to a compound that is easily measured in a
bioluminescent reaction. A direct relationship between the P450 ac-
tivity and light output exists because the concentration of luciferin
produced by P450 is within the linear range of a standard curve of
luciferin bioluminescence. A high degree of selectivity by this com-
pound 1 for the CYP3A subfamily was expected based on screening
a panel of recombinant P450 enzymes (Fig. 1), and greater selectivity
for CYP3A4 could be dialed in with reaction conditions that favor
TABLE 1
Kinetic data for CYP3A-catalyzed oxidation of proluciferin acetal 1 based on
analysis of Fig. 3, D and L
Reaction
Km
kcat
k
cat/Km
␮M
CYP3A4
CYP3A5
CYP3A7
2.88 Ϯ 0.18
4.21 Ϯ 0.31
8.85 Ϯ 0.53
5.87 Ϯ 0.14
0.94 Ϯ 0.02
0.17 Ϯ 0.004
2.04
0.22
0.02

PROLUCIFERIN ACETALS AS CYP3A PROBES
2409
IC50 comparisons for known CYP3A4 inhibitors between the com-
pound 1 assay and the testosterone 6␤-hydroxylation assay showed
nearly identical rank ordering for both recombinant CYP3A4 and
HLM reactions, indicating that 1 can be used as a surrogate for the
testosterone CYP3A4 assay (Table 2). The HLM correlation is also
consistent with the high degree of selectivity of compound 1 for
CYP3A4 enzymes. The IC50 values were closely correlated for most
of the inhibitors showing fold differences between unity and approx-
imately 4-fold. A few compounds showed larger fold differences,
which is not unexpected given the substrate-dependent nature of
CYP3A4 inhibition (Wang et al., 2000). Differences between HLMs
and recombinant CYP3A4 inhibition are known from the literature,
although they are not explained within the scope of this study (Nomeir
et al., 2001).
In addition to cell-free enzyme inhibition studies, the utility of
compound 1 for nonlytic cell-based assays of CYP3A induction in
cultured hepatocytes has also been described previously (Li, 2009;
Doshi and Li, 2011). The CYP3A compound 1 was also applied to
hepatocytes in a three-part multiplex format that provided luminescent
CYP3A, a fluorescent CYP1A, and a luminescent viability measure-
ment from a single culture well (Larson et al., 2011).
The simple nature of the luminogenic proluciferin acetal assays
likely provides an advantage to P450 activity screens of large chem-
ical libraries and other applications that benefit from a high-through-
put approach. P450 assays based on HPLC or liquid chromatography
mass spectrometry analysis are widely accepted, but they are often too
slow for high-throughput applications and are limiting when rapid
turnaround times are needed. Furthermore, the ease of implementing
a bioluminescent P450 assay facilitates expansion from assays of a
single concentration of test compound to more informative full-dose
response screens. Data from such quantitative high-throughput P450
screens reveal structure/activity relationships and inhibition trends in
large compound sets, and enable the annotation of compound libraries
with P450 data (MacArthur et al., 2009; Veith et al., 2009). The strong
correlation between inhibitory data from the CYP3A4 luciferin acetal
and the testosterone assays supports the expectation that the biolumi-
nescent assays will provide relevant data while meeting throughput
requirements in large P450 screens.
FIG. 4. Correlation of testosterone and compound 1 as probe substrates for P450
inhibition assays. IC50 comparisons of recombinant CYP3A4 assays (r2 ϭ 0.96, P Ͻ
0.0001) (A) and human liver microsome assays (r2 ϭ 0.91, P ϭ 0.0003) (B).
CYP3A4 over CYP3A5 and CYP3A7 (Fig. 2; Table 1). The selec-
tivity of compound 1 was confirmed by inhibition studies in HLMs
that contain a mixture of potentially cross-reacting P450 enzymes.
Known CYP3A4 inhibitors inhibited the HLM activity at IC50 values
consistent with selective CYP3A4 inhibition, as indicated by close
correlation to the IC50 values determined by the well established
CYP3A4-selective testosterone 6␤-hydroxylation reaction (Fig. 4A).
The 0.04 nM LOD for CYP3A4 activity with substrate 1 is very
low. This allows for an assay of exquisite sensitivity. By reducing the
amount of P450 containing microsomes in a reaction, the extent of
nonspecific substrate or analyte binding is kept to a minimum or even
nondetectable level. A very sensitive assay also enables detection of
activity in samples with low P450 expression. The limit of CYP3A4
detection is potentially influenced by three scenarios: 1) contamina-
tion of 1 with D-luciferin, 2) non-P450-catalyzed conversion of 1 to
D-luciferin, or 3) uncatalyzed conversion of 1 to D-luciferin. Due to the
drastic difference in functional groups and intrinsic stability of the
acetal functionality, contaminating luciferin was typically reduced to
less than 0.0001% of proluciferin acetals using either silica or re-
versed-phase chromatography. Non-P450-catalyzed and uncatalyzed
conversions apparently do not contribute significantly to background
signals because there was no evidence of conversion in the P450-
negative control of the recombinant P450 system (Fig. 1; Supplemen-
tal Fig. S1), and conversion in HLMs was NADPH-dependent, a
hallmark of P450-dependent catalysis.
TABLE 2
IC50 comparisons
Data compare inhibition with the compound
1 assay and testosterone 6␤-hydroxylase
assay using recombinant CYP3A4 or human liver microsomes.
rCYP3A4
HLMs
c
c
CYP3A4 Inhibitors
1a
TSb
⌬
1a
TSb
⌬
IC50 (␮M)
IC50 (␮M)
Clotrimazole
Ketoconazole
Troleandomycin
Midazolam
Nifedipine
Verapamil
Erythromycin
Omeprazole
␣-Naphthoflavone
Fluvoxamine
Disopyramide
0.006
0.04
0.52
2.06
3.24
3.45
7.33
11.21
10.00
NDf
0.05
0.10
1.13
8
0.003
0.074
1.96
16.82
5.99
0.052
0.10
1.11
3.64
6.90
16.7
1.3
1.8
4.6
1.2
7.7
2.5
2.2
3.3
4.2
3.8
1.7
1.1
3.7
6.80
13.55
12.95
12.27
10.00
36.64
NDf
115
14.86
NCd
68.27
NCe
NCd
13.52
NCe
48.31
48.80
5.0
46.94
71.09
1.0
1.5
53.29
65.88
1.2
r, recombinant; TS, testosterone; ND, not determined; NC, not calculated.
a IC50s determined using bioluminescent assay with compound 1.
b IC50 determined using testosterone 6␤-hydroxylase assay.
c ⌬ is the fold difference in IC50 values between the two assays.
d IC50 was not calculated due to only modest inhibition of both assays.
e IC50 was not calculated due to a modest stimulation.
f Fluvoxamine was not applied to recombinant CYP3A4.

2410
MEISENHEIMER ET AL.
hepatocytes: effects of organic solvents, cytochrome P450 (P450) inhibitors, and P450 induc-
ers. Drug Metab Dispos 37:1598–1603.
MacArthur R, Leister W, Veith H, Shinn P, Southall N, Austin CP, Inglese J, and Auld DS (2009)
Monitoring compound integrity with cytochrome P450 assays and qHTS. J Biomol Screen
14:538–546.
Authorship Contributions
Participated in research design: Meisenheimer, Ma, Sobol, Klaubert, and
Cali.
Conducted experiments: Meisenheimer, Uyeda, Ma, Sobol, McDougall,
Corona, Simpson, and Cali.
Contributed new reagents or analytic tools: Meisenheimer and Simpson.
Performed data analysis: Meisenheimer, Ma, Sobol, Simpson, and Cali.
Wrote or contributed to the writing of the manuscript: Meisenheimer,
Uyeda, Ma, Sobol, McDougall, Corona, Simpson, Klaubert, and Cali.
Moody GC, Griffin SJ, Mather AN, McGinnity DF, and Riley RJ (1999) Fully automated
analysis of activities catalysed by the major human liver cytochrome P450 (CYP) enzymes:
assessment of human CYP inhibition potential. Xenobiotica 29:53–75.
Nomeir AA, Ruegg C, Shoemaker M, Favreau LV, Palamanda JR, Silber P, and Lin CC (2001)
Inhibition of CYP3A4 in a rapid microtiter plate assay using recombinant enzyme and in
human liver microsomes using conventional substrates. Drug Metab Dispos 29:748–753.
O’Brien MA, Daily WJ, Hesselberth PE, Moravec RA, Scurria MA, Klaubert DH, Bulleit RF,
and Wood KV (2005) Homogeneous, bioluminescent protease assays: caspase-3 as a model.
J Biomol Screen 10:137–148.
Stresser DM, Dehal SS, and Kupfer D (1996) Ring hydroxylation of [o-3H]methoxychlor as a
probe for liver microsomal CYP2B activity: potential for in vivo CYP2B assay. Anal Biochem
233:100–107.
Veith H, Southall N, Huang R, James T, Fayne D, Artemenko N, Shen M, Inglese J, Austin CP,
Lloyd DG, et al. (2009) Comprehensive characterization of cytochrome P450 isozyme selec-
tivity across chemical libraries. Nat Biotechnol 27:1050–1055.
References
Bell L, Bickford S, Nguyen PH, Wang J, He T, Zhang B, Friche Y, Zimmerlin A, Urban L, and
Bojanic D (2008) Evaluation of fluorescence- and mass spectrometry-based CYP inhibition
assays for use in drug discovery. J Biomol Screen 13:343–353.
Blanksby SJ and Ellison GB (2003) Bond dissociation energies of organic molecules. Acc Chem
Res 36:255–263.
Branchini BR, Hayward MM, Bamford S, Brennan PM, and Lajiness EJ (1989) Naphthyl- and
quinolylluciferin: green and red light emitting firefly luciferin analogues. Photochem Photo-
biol 49:689–695.
Wang RW, Newton DJ, Liu N, Atkins WM, and Lu AY (2000) Human cytochrome P-450 3A4:
in vitro drug-drug interaction patterns are substrate-dependent. Drug Metab Dispos 28:360–
366.
Cali JJ, Ma D, Sobol M, Simpson DJ, Frackman S, Good TD, Daily WJ, and Liu D (2006)
Luminogenic cytochrome P450 assays. Expert Opin Drug Metab Toxicol 2:629–645.
Cali JJ, Niles A, Valley MP, O’Brien MA, Riss TL, and Shultz J (2008) Bioluminescent assays
for ADMET. Expert Opin Drug Metab Toxicol 4:103–120.
White EH and Wo¨rther H (1966) Analogs of firefly luciferin. 3. J Org Chem 31:1484–1488.
White EH, Worther H, Field GF, and McElroy WD (1965) Analogs of firefly luciferin. J Org
Chem 30:2344–2348.
Wienkers LC and Heath TG (2005) Predicting in vivo drug interactions from in vitro drug
Doshi U and Li AP (2011) Luciferin IPA-based higher throughput human hepatocyte screening
assays for CYP3A4 inhibition and induction. J Biomol Screen 16:903–909.
Fan F and Wood KV (2007) Bioluminescent assays for high-throughput screening. Assay Drug
Dev Technol 5:127–136.
Friden M, Vanaja K, and Nandi VN (2006) Drug-drug interactions of anti-infective drugs: utility
of fluorescence CYP inhibition assays in drug discovery. Drug Metabol Drug Interact
21:163–185.
Grime KH, Bird J, Ferguson D, and Riley RJ (2009) Mechanism-based inhibition of cytochrome
P450 enzymes: an evaluation of early decision making in vitro approaches and drug-drug
interaction prediction methods. Eur J Pharm Sci 36:175–191.
discovery data. Nat Rev Drug Discov 4:825–833.
Woodroofe CC, Shultz JW, Wood MG, Osterman J, Cali JJ, Daily WJ, Meisenheimer PL, and
Klaubert DH (2008) N-alkylated 6Ј-aminoluciferins are bioluminescent substrates for Ultra-
Glo and QuantiLum luciferase: new potential scaffolds for bioluminescent assays. Biochem-
istry 47:10383–10393.
Zhou W, Andrews C, Liu J, Shultz JW, Valley MP, Cali JJ, Hawkins EM, Klaubert DH, Bulleit
RF, and Wood KV (2008) Self-cleavable bioluminogenic luciferin phosphates as alkaline
phosphatase reporters. Chembiochem 9:714–718.
Zhou W, Shultz JW, Murphy N, Hawkins EM, Bernad L, Good T, Moothart L, Frackman S,
Klaubert DH, Bulleit RF, et al. (2006a) Electrophilic aromatic substituted luciferins as
bioluminescent probes for glutathione S-transferase assays. Chem Commun (Camb) 4620–
4622.
Guengerich FP (2007) Mechanisms of cytochrome P450 substrate oxidation: MiniReview.
J Biochem Mol Toxicol 21:163–168.
Jones JP, Mysinger M, and Korzekwa KR (2002) Computational models for cytochrome P450:
a predictive electronic model for aromatic oxidation and hydrogen atom abstraction. Drug
Metab Dispos 30:7–12.
Zhou W, Valley MP, Shultz J, Hawkins EM, Bernad L, Good T, Good D, Riss TL, Klaubert DH,
and Wood KV (2006b) New bioluminogenic substrates for monoamine oxidase assays. J Am
Chem Soc 128:3122–3123.
Kenworthy KE, Bloomer JC, Clarke SE, and Houston JB (1999) CYP3A4 drug interactions:
correlation of 10 in vitro probe substrates. Br J Clin Pharmacol 48:716–727.
Ko¨hler GI, Bode-Bo¨ger SM, Busse R, Hoopmann M, Welte T, and Bo¨ger RH (2000) Drug-drug
interactions in medical patients: effects of in-hospital treatment and relation to multiple drug
use. Int J Clin Pharmacol Ther 38:504–513.
Zientek M, Miller H, Smith D, Dunklee MB, Heinle L, Thurston A, Lee C, Hyland R, Fahmi O,
and Burdette D (2008) Development of an in vitro drug-drug interaction assay to simultane-
ously monitor five cytochrome P450 isoforms and performance assessment using drug library
compounds. J Pharmacol Toxicol Methods 58:206–214.
Larson B, Moeller T, Banks P, and Cali JJ (2011) Automated triplexed hepatocyte-based viability
and CYP1A and -3A induction assays. J Biomol Screen 16:895–902.
Address correspondence to: Poncho L. Meisenheimer, R&D, Promega Bio-
sciences, San Luis Obispo, CA 93401. E-mail: pmeisenh@promega.com
Lazarou J, Pomeranz BH, and Corey PN (1998) Incidence of adverse drug reactions in hospi-
talized patients: a meta-analysis of prospective studies. JAMA 279:1200–1205.
Li AP (2009) Evaluation of luciferin-isopropyl acetal as a CYP3A4 substrate for human