An increase in side-group hydrophobicity largely improves the potency of ritonavir-like inhibitors of CYP3A4

Article abstract of DOI:10.1016/j.bmc.2020.115349

Identification of structural determinants required for potent inhibition of drug-metabolizing cytochrome P450 3A4 (CYP3A4) could help develop safer drugs and more effective pharmacoenhancers. We utilize a rational inhibitor design to decipher structure-activity relationships in analogues of ritonavir, a highly potent CYP3A4 inhibitor marketed as pharmacoenhancer. Analysis of compounds with the R₁ side-group as phenyl or naphthalene and R₂ as indole or naphthalene in different stereo configuration showed that (i) analogues with the R₂-naphthalene tend to bind tighter and inhibit CYP3A4 more potently than the R₂-phenyl/indole containing counterparts; (ii) stereochemistry becomes a more important contributing factor, as the bulky side-groups limit the ability to optimize protein-ligand interactions; (iii) the relationship between the R₁/R₂ configuration and preferential binding to CYP3A4 is complex and depends on the side-group functionality/interplay and backbone spacing; and (iv) three inhibitors, 5a-b and 7d, were superior to ritonavir (IC₅₀ of 0.055–0.085 μM vs. 0.130 μM, respectively).

Full text of DOI:10.1016/j.bmc.2020.115349

Bioorganic&MedicinalChemistry28(2020)115349
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
An increase in side-group hydrophobicity largely improves the potency of
ritonavir-like inhibitors of CYP3A4
⁎
Eric R. Samuelsa,1, Irina F. Sevrioukovab,
a Departments of Pharmaceutical Sciences, University of California, Irvine, CA 92697-3900, United States
b Departments of Molecular Biology and Biochemistry, University of California, Irvine, CA 92697-3900, United States
A R T I C L E I N F O
A B S T R A C T
Keywords:
Identification of structural determinants required for potent inhibition of drug-metabolizing cytochrome P450
3A4 (CYP3A4) could help develop safer drugs and more effective pharmacoenhancers. We utilize a rational
inhibitor design to decipher structure-activity relationships in analogues of ritonavir, a highly potent CYP3A4
inhibitor marketed as pharmacoenhancer. Analysis of compounds with the R1 side-group as phenyl or naph-
thalene and R2 as indole or naphthalene in different stereo configuration showed that (i) analogues with the R2-
naphthalene tend to bind tighter and inhibit CYP3A4 more potently than the R2-phenyl/indole containing
counterparts; (ii) stereochemistry becomes a more important contributing factor, as the bulky side-groups limit
the ability to optimize protein-ligand interactions; (iii) the relationship between the R1/R2 configuration and
preferential binding to CYP3A4 is complex and depends on the side-group functionality/interplay and backbone
spacing; and (iv) three inhibitors, 5a-b and 7d, were superior to ritonavir (IC50 of 0.055–0.085 μM vs. 0.130 μM,
respectively).
CYP3A4
Ligand binding
Inhibitor design
Crystal structure
Structure-activity relations
1. Introduction
ketoconazole in drug development drug-drug interaction studies.7,8
Therefore, elucidation of structure-activity relations for CYP3A4 in-
Human cytochrome P450 3A4 (CYP3A4) is a major hepatic and
intestinal isoform and a versatile catalyst that plays a central role in
drug metabolism.1 Drugs can serve not only as substrates but also as
inducers and inhibitors of CYP3A4.2 Inhibition of CYP3A4 is usually
undesired, because it could lead to clinically significant drug-drug in-
teractions, toxicity and therapeutic failures. Thus, unraveling the me-
chanisms that underline high promiscuity and adaptability of CYP3A4
to structurally diverse ligands could help predict and eliminate the in-
hibitory potential and other side effects in drug candidates early in the
discovery process.
hibitors could help develop safer, more potent and highly specific index
inhibitors and pharmacoenhancers.
Based on our previous studies on the interaction of CYP3A4 with
ritonavir and its analogues,9–12 we developed a pharmacophore model
for a CYP3A4-specific inhibitor13 and utilized a build-from-scratch ap-
proach to elucidate the relative importance of each pharmacophoric
determinant. Three groups of inhibitors (series I-III) with different
backbones and side-groups attached to the pyridine ring, serving as the
heme-ligating moiety, have been already characterized.14–16 These
studies demonstrated that the binding and inhibitory strength of rito-
navir-like compounds depends on the backbone length and composi-
tion, spacing between the functional groups, H-bonding to the active
site S119, hydrophobic interactions mediated by the R1/R2 side-groups
and, to a lesser degree, their stereo configuration.
Another area that could benefit from in-depth studies on the
CYP3A4 ligand binding and inhibitory mechanism is pharma-
coenhancement. Currently, three CYP3A4 inhibitors are used as phar-
macoenhancers
for
quickly-metabolized
antiviral
and
im-
munosuppressant drugs: ritonavir (originally designed as an HIV
protease inhibitor; Fig. 1), its derivative cobicistat, and the antifungal
agent ketoconazole.3–6 Neither booster is highly specific for CYP3A4 or
was designed based on its crystal structure. The same is true for itra-
conazole and clarythromycine, used as alternatives to highly toxic
The current study was designed to test one of the earlier predictions
that an increase in the R2 hydrophobicity could improve the inhibitory
strength.15,16 Eight (series IV) analogues were developed by modifying
the R2 functionality in two different scaffolds used for synthesis of 8f,
the most potent series II inhibitor,15 and 4e-h, the high-affinity
Abbreviations: CYP3A4, cytochrome P450 3A4; BFC, 7-benzyloxy-4-(trifluoromethyl)coumarin; Boc, tert-butyloxycarbonyl
⁎ Corresponding author.
E-mail address: sevrioui@uci.edu (I.F. Sevrioukova).
1 Present address: Allergan plc, 2525 Dupont Dr., Irvine, CA 92612, United States.
https://doi.org/10.1016/j.bmc.2020.115349
Received 8 December 2019; Received in revised form 22 January 2020; Accepted 27 January 2020
Availableonline31January2020
0968-0896/PublishedbyElsevierLtd.

E.R. Samuels and I.F. Sevrioukova
Bioorganic&MedicinalChemistry28(2020)115349
Fig. 1. Chemical structures of ritonavir and the most potent/high affinity CYP3A4 inhibitors, 8f and 4e-h, from the previously designed series II and III, respec-
tively.15,16 The R1 and R2 side-groups are indicated. In ritonavir, the phenyl side-groups would be in R,R configuration if the central hydroxyl group is removed.
subgroup from series III16 (Fig. 1). Here we report that, indeed, com-
pounds with the larger, more hydrophobic naphthalene ring at R2 po-
sition tend to bind tighter and inhibit CYP3A4 more potently than their
phenyl- or indole-containing counterparts. The backbone spacing and
side-group configuration were other factors that strongly influenced the
inhibitory strength and preferential binding to CYP3A4. Most im-
portantly, for the first time, this study identified three compounds that
were chemically simpler than ritonavir but inhibited CYP3A4 twice as
stronger.
vacuo, affording the crude methyl ester hydrochloride (1.34 g). The
crude salt was then dissolved in ethyl ether:methanol (7.5:1; 30 mL:4
mL) and cooled to −10 °C. Triethylamine (6 mL) was slowly added and
the reaction was stirred for one hour at −10 °C. After reaction com-
pletion, the solution was filtered through a pad of celite to remove Et3N
salt, and the filtrate was evaporated in vacuo affording the free base
methyl ester (1.14 g). The methyl ester was dissolved in methanol
(35 mL) and placed on an ice bath. Sodium borohydride (10 eq) was
added portion-wise over 20 min and the reaction was allowed to come
to room temperature overnight. The solvent was then evaporated, di-
luted with H2O (80 mL), and extracted with EtOAc (3 × 100 mL). The
combined organic layers were dried over MgSO4, filtrated and con-
centrated in vacuo affording 2a as a white powder (0.87 g, 94%), which
was used without further purification. 1H NMR (400 MHz, CDCl3T) δ
8.05 (d, J = 7.0 Hz, 1H), 7.88 (d, J = 9.1 Hz, 1H), 7.77 (d, J = 8.7 Hz,
1H), 7.53 (sext, J = 7.6 Hz, 2H), 7.43 (t, J = 7.7 Hz, 1H), 7.36 (d,
J = 6.2 Hz, 1H), 3.72 (dd, J = 3.7, 10.6 Hz, 1H), 3.49 (dd, J = 6.6,
10.6 Hz, 1H), 3.33 (m, 2H), 2.95 (dd, J = 9.7, 14.8 Hz, 1H). HRMS m/z
calculated for C13H16NO [M+H]+: 202.1232. Found: 202.1227.
Synthesis of Boc protected, tosylated, amino alcohols (compounds 3a-e)
– Compounds 3a-e were prepared using methods described pre-
viously,14 as outlined in Scheme 1.
2. Materials and methods
Chemistry General Methods – All reactions were performed with
commercially available reagents (Aldrich, Thermo-Fisher, Alfa Aesar,
Acros, Oakwood, Millipore) without further purification. Anhydrous
solvents were acquired through a solvent purification system (Inert
PureSolv and JC Meyer systems) or purified according to standard
procedures. 1H NMR spectra were recorded on Bruker DRX 400 MHz,
Bruker DRX 500 MHz, or Bruker Avance 600 MHz spectrometer and
processed using TopSpin 3.5 software. LRMS and HRMS data were
obtained via ESI LC-TOF on a Waters (Micromass) LCT Premier spec-
trometer (Waters), with PEG as the calibrant for HRMS. Optical rotation
was recorded on a Rudolph Autopol III Automatic Polarimeter at room
temperature in methanol. TLC was performed using EMD Millipore si-
lica gel 60 F254 aluminum plates. Separation by column chromato-
graphy was conducted using Fisher silica gel 60 (230–400 mesh). Purity
of final products was verified by 1H NMR with TMS as a standard and
by UPLC-MS (Waters Acquity UPLC H-class QDA with a FTN) with a
C18 column (2.1 × 50 mm; Acquity UPLC BEH C18; 1.7 μm particles).
All investigated compounds were > 95% pure as determined by UPLC-
MS. High resolution mass spectrometry data, NMR spectra, and UPLC-
MS chromatograms are included in the Supplementary Material.
General Procedure for Synthesis of Compounds 4a-b (Scheme 2) – To a
solution of N-phenylcysteine hydrobromide, prepared as described
previously,14 (0.093 g, 0.34 mmol, 1.5 eq) in DMF (4 mL), compound
3b (0.1 g, 0.22 mmol) was added. The reaction was allowed to stir for
30 min at 50 °C, after which 1 N NaOH (2 mL) was added dropwise. The
reaction was then stirred at 50 °C overnight. The flask was cooled to
room temperature and pH was examined/adjusted to 8.0. The solvent
was evaporated in vacuo to give the crude oily product 4a, which was
used in the next step without any further purification. LRMS m/z cal-
culated for C25H31N3O4S [M+Na]+: 492.2. Found: 492.1. 4b LRMS m/
z calculated for C27H32N2O4S [M+Na]+: 503.2. Found: 503.1.
General Procedure for Synthesis of Compounds 5a-b (Scheme 2) –
Crude 4a (0.37 g, 0.78 mmol (Theoretical: 0.10 g, 0.22 mmol)) was
dissolved in DMF (9 mL). To this solution, EDAC (0.23 g, 1.2 mmol,
1.5 eq) and HOBt (0.18 g, 1.2 mmol, 1.5 eq) were added, followed by
the addition of 3-(aminomethyl)pyridine (0.13 g, 1.2 mmol, 1.5 eq) and
DIPEA (0.31 g, 2.4 mmol, 3 eq). The reaction was stirred at room
temperature overnight. Upon completion, the solvent was evaporated
2.1. Synthesis of analogues
Typical procedure for amino alcohols from amino acids (compounds 1a-
e)17,18 – To a flask containing DL-1-Naphthylalanine (1a) (1.0 g,
4.6 mmol), TMSCl (1.51 g, 14 mmol, 3 eq) was added, followed by
anhydrous methanol (10 mL). The reaction was allowed to stir at room
temperature overnight. Upon completion, the volatiles were removed in
2

E.R. Samuels and I.F. Sevrioukova
Bioorganic&MedicinalChemistry28(2020)115349
Scheme 1.
and the reaction mixture was diluted with ethyl acetate (80 mL). The
organic layer was then washed with saturated NaHCO3 (50 mL), water
(2 × 50 mL), and brine (50 mL). The combined organic layers were
dried over MgSO4, filtrated and concentrated in vacuo to give the crude
product, which was purified via column chromatography (95:5
EtOAc:MeOH). The pure product 5a was obtained as a white fluffy solid
(0.03 g, 24%). TLC: EtOAc/MeOH 90:10 (Rf. 0.58). 1H NMR (400 MHz,
CDCl3) δ 8.46 (m, 2H), 8.33 (bs, 1H (NH)), 7.62 (d, J = 7.9 Hz, 1H),
7.50 (t, J = 6.3 Hz, 1H), 7.36 (m, 2H), 7.19 (m, 3H), 7.11 (m,
J = 3.8 Hz, 1H), 6.98 (s, 1H), 6.83 (t, J = 6.9 Hz, 1H), 6.64 (dd,
J = 7.8, 14.4 Hz, 2H), 4.78 (t, J = 10.2 Hz, 1H), 4.68 (bd, J = 27.6 Hz,
1H (NH)), 4.39 (m, 2H), 4.11 (bs, 1H), 3.94 (bd, J = 43.7 Hz, 1H (NH)),
3.15 (dt, J = 4.4, 13.7 Hz, 1H), 3.00 (m, 3H), 2.63 (m, 2H), 1.87 (bs,
1H (NH)), 1.42 (d, J = 12.1 Hz, 9H). HRMS m/z calculated for
6c LRMS m/z calculated for C27H31NO4S [M+Na]+: 488.2. Found:
488.1. 6d LRMS m/z calculated for C27H31NO4S [M+Na]+: 488.2.
Found: 488.1.
General Procedure for Synthesis of Compounds 7a-d (Scheme 3) –
Crude 6a (0.11 g, 0.24 mmol) was dissolved in DMF (4 mL). To this
solution, EDAC (0.07 g, 0.36 mmol, 1.5 eq) and HOBt (0.055 g,
0.36 mmol, 1.5 eq) were added, followed by the addition of 3-(2-ami-
noethyl)pyridine (0.04 g, 0.36 mmol, 1.5 eq) and DIPEA (0.09 g,
0.72 mmol, 3 eq). The reaction was stirred at room temperature over-
night. Upon completion, the solvent was evaporated and the reaction
mixture was diluted with ethyl acetate (80 mL). The organic layer was
then washed with saturated NaHCO3 (50 mL), water (2 × 50 mL), and
brine (50 mL). The combined organic layers were dried over MgSO4,
filtrated and concentrated in vacuo to give the crude product, which was
purified via column chromatography (95:5 EtOAc:MeOH). The pure
product 7a was obtained as a light yellow fluffy solid (0.029 g, 22%).
TLC: EtOAc/MeOH 90:10 (Rf. 0.5). 1H NMR (400 MHz, CDCl3) δ 8.43
(m, 2H), 8.29 (s, 2H), 8.25 (s, 1H (NH)), 7.56 (dt, J = 7.4, 23.3 Hz, 2H),
7.41–7.07 (m, 8H), 6.97 (s, 1H), 6.35 (t, J = 6.3 Hz, 1H), 4.66 (s, 1H
(Boc-NH)), 4.21 (m, 1H), 4.05 (m, 1H), 3.41 (q, J = 6.5 Hz, 1H), 3.35
(m, 1H), 3.23 (dd, J = 7.3, 13.5 Hz, 2H), 3.03 (m, 1H), 2.92 (dd,
J = 6.7, 13.3 Hz, 2H), 2.58 (m, 1H), 2.51 (dd, J = 6.1, 13.1 Hz, 1H),
C
31H37N5O3SNa [M+Na]+: 582.2515. Found: 593.2516. The pure
product 5b was acquired as a white fluffy solid (0.029 g, 23%). TLC:
EtOAc/MeOH 90:10 (Rf. 0.6). 1H NMR (400 MHz, CDCl3) δ 8.46 (s,
2H), 8.14 (d, J = 7.2 Hz, 1H), 7.85 (s, 1H) 7.75 (t, J = 7.4 Hz, 1H),
7.50 (m, 3H), 7.31 (d, J = 5.7 Hz, 2H), 7.18 (m, 3H), 6.83 (q,
J = 6.4 Hz, 1H), 6.63 (t, J = 10.2 Hz, 2H), 4.80 (dd, J = 8.0, 27.7 Hz,
1H), 4.61 (bd, J = 29.3 Hz, 1H (NH)), 4.40 (m, 2H), 4.18 (sext,
J = 6.8 Hz, 1H), 3.95 (bd, J = 36.6 Hz, 1H (NH)), 3.34 (dd, J = 7.2,
14.0 Hz, 1H), 3.23 (dd, J = 7.3, 13.5 Hz, 1H), 3.08 (m, 2H), 2.68 (t,
J = 5.8 Hz, 2H). 1.37 (d, J = 14.0 Hz, 9H). HRMS m/z calculated for
1.43 (s, 9H). HRMS m/z calculated for C32H38N4O3SNa [M+Na]+
:
581.2562. Found: 581.2573. The pure product 7b was acquired as a
light yellow fluffy solid (0.018 g, 12%). TLC: EtOAc/MeOH 90:10 (Rf.
0.5). 1H NMR (400 MHz, CDCl3) δ 8.42 (d, J = 4.64 Hz, 2H), 8.29 (s,
2H), 8.20 (s, 1H (NH)), 7.59 (dd, J = 7.9, 13.5 Hz, 2H), 7.41–7.08 (m,
8H), 6.97 (s, 1H), 6.35 (t, J = 5.7 Hz, 1H), 4.66 (s, 1H (Boc-NH)), 4.20
(m, 1H), 4.05 (m, 1H), 3.41 (q, J = 6.2 Hz, 1H), 3.35 (m, 1H), 3.23 (dd,
J = 7.5, 13.4 Hz, 2H), 3.03 (t, J = 6.3 Hz, 1H), 2.92 (dd, J = 7.0,
13.5 Hz, 2H), 2.58 (m, 1H), 2.51 (dd, J = 5.9, 13.1 Hz, 1H), 1.42 (s,
9H). HRMS m/z calculated for C32H38N4O3SNa [M+Na]+: 581.2562.
Found: 581.2543. The pure product 7c was obtained as an off-white
fluffy solid (0.04 g, 27%). TLC: EtOAc/MeOH 90:10 (Rf. 0.58). 1H NMR
(400 MHz, CDCl3) δ 8.41 (d, J = 4.4 Hz, 2H), 8.27 (s, 1H), 8.12 (d,
J = 8.4 Hz, 1H), 7.82 (d, J = 8.2 Hz, 1H), 7.73 (d, J = 8.1 Hz, 1H),
7.55 (t, J = 7.5 Hz, 1H), 7.48 (t, J = 7.4 Hz, 1H), 7.37 (t, J = 7.6 Hz,
1H), 7.32–7.19 (m, 6H), 7.12 (t, J = 6.4 Hz, 1H), 6.38 (t, J = 6.0 Hz,
C
33H38N4O3SNa [M+Na]+: 593.2562. Found: 593.2545.
General Procedure for Synthesis of Compounds 6a-d (Scheme 3) – To a
solution of L-α-thio-phenylalanine, prepared as described previously,18
(0.05 g, 0.28 mmol, 1.3 eq in 4 mL of DMF), compound 3b (0.1 g,
0.22 mmol) was added. The reaction was allowed to stir for 30 min at
50 °C, after which 1 N NaOH (2 mL) was added dropwise. The reaction
was then stirred at 50 °C overnight. The flask was cooled to room
temperature and pH was examined/adjusted to 8.0. The solvent was
evaporated, diluted with 10% HCl (50 mL), and extracted with Et2O
(3 × 50 mL). The combined organic layers were dried over MgSO4,
filtrated and concentrated in vacuo to give the crude oily product 6a,
which was used in the next step without any further purification. LRMS
m/z calculated for C25H30N2O4S [M+Na]+: 477.2. Found: 477.1. 6b
LRMS m/z calculated for C25H30N2O4S [M+Na]+: 477.2. Found: 477.1.
Scheme 2.
3

E.R. Samuels and I.F. Sevrioukova
Bioorganic&MedicinalChemistry28(2020)115349
Scheme 3.
1H), 4.72 (d, J = 7.3 Hz, 1H), 4.11 (sext, J = 6.8 Hz, 1H), 3.39 (t
J = 7.2 Hz, 2H), 3.25 (dd, J = 7.3, 13.8 Hz, 2H), 3.19 (m, 1H), 2.92
(dd, J = 6.8, 13.8 Hz, 1H), 2.61 (t, J = 7.0 Hz, 2H), 2.55 (d,
J = 7.0 Hz, 2H), 1.39 (s, 9H). HRMS m/z calculated for C34H39N3O3SNa
[M+Na]+: 592.2610. Found: 592.2607. The pure product 7d was ac-
quired as a light yellow fluffy solid (0.02 g, 13.5%). TLC: EtOAc/MeOH
90:10 (Rf. 0.58). 1H NMR (400 MHz, CDCl3) δ 8.41 (d, J = 3.9 Hz, 2H),
8.27 (s, 1H), 8.12 (d, J = 8.0 Hz, 1H), 7.83 (d, J = 8.2 Hz, 1H), 7.73 (d,
J = 7.9 Hz, 1H), 7.55 (t, J = 7.8 Hz, 1H), 7.48 (t, J = 6.9 Hz, 1H), 7.37
(t, J = 7.6 Hz, 1H), 7.31–7.17 (m, 6H), 7.12 (t, J = 6.5 Hz, 1H), 6.36 (t,
J = 6.0 Hz, 1H), 4.70 (d, J = 8.4 Hz, 1H), 4.12 (sext, J = 6.7 Hz, 1H),
3.39 (t J = 7.1 Hz, 2H), 3.25 (dd, J = 7.4, 13.9 Hz, 2H), 3.19 (m, 1H),
2.92 (dd, J = 7.0, 13.7 Hz, 1H), 2.61 (t, J = 7.0 Hz, 2H), 2.56 (d,
J = 6.4 Hz, 2H), 1.39 (s, 9H). HRMS m/z calculated for C34H39N3O3SNa
[M+Na]+: 592.2610. Found: 592.2599.
completion, the solvent was evaporated and the reaction mixture was
diluted with ethyl acetate (70 mL). The organic layer was then washed
with saturated NaHCO3 (30 mL), water (50 mL), and brine (50 mL). The
combined organic layers were dried over MgSO4, filtrated and con-
centrated in vacuo to give the crude product, which was purified via
column chromatography (95:5 EtOAc:MeOH). The pure product 9a was
acquired as a light yellow solid (0.053 g, 29%). TLC: EtOAc/MeOH
90:10 (Rf. 0.62). 1H NMR (400 MHz, CDCl3) δ 8.42 (s, 2H), 8.32 (s, 1H),
8.21 (d, J = 8.22 Hz, 2H), 8.06 (m, 6H), 7.94–7.29 (m, 8H), 5.11 (s, 1H
(Boc-NH)), 4.71 (q, J = 7.9 Hz, 2H), 4.39 (m, 1H), 4.35 (m, 1H), 4.24
(t, J = 4.9 Hz, 2H), 3.76 (m, 1H), 3.64 (m, 1H), 3.36–3.22 (m, 2H),
2.71 (m, 2H), 1.32 (s, 9H). HRMS m/z calculated for C37H39N3O3SNa
[M+Na]+: 628.2610. Found: 628.2632. (0.15 g, 0.7 mmol (Theore-
tical: 0.1 g, 0.22 mmol)) was dissolved in DMF (9 mL). 8 and 3-(ami-
noethyl)pyridine afforded the pure product 9b as a white solid (0.01 g,
5%). TLC: EtOAc/MeOH 90:10 (Rf. 0.6). 1H NMR (400 MHz, CDCl3) δ
8.40 (s, 1H), 8.32 (s, 1H), 8.06 (m, 2H), 7.86 (m, 6H), 7.74–7.36 (m,
8H), 4.65 (m, 2H), 4.14 (m, 2H), 3.76 (m, 2H), 3.41 (m, 2H), 3.23 (m,
2H), 2.86 (m, 1H), 2.83 (m, 1H), 1.30 (m, 9H). HRMS m/z calculated
for C38H41N3O3SNa [M+Na]+: 642.2766. Found: 642.2747.
Synthesis of Compound 8 (Scheme 4) – To a solution of α-thio-nap-
thylalanine, prepared as described previously,18 (0.2 g, 0.86 mmol,
1.3 eq) in DMF (5 mL), compound 3a (0.3 g, 0.66 mmol) was added.
The reaction was allowed to stir for 30 min at 50 °C, after which 1 N
NaOH (2 mL) was added dropwise. The reaction was then stirred at
50 °C overnight. The flask was cooled to room temperature and pH was
examined/adjusted to 8.0. The solvent was evaporated, diluted with
10% HCl (50 mL), and extracted with Et2O (3 × 50 mL). The combined
organic layers were dried over MgSO4, filtrated and concentrated in
vacuo to give the crude oily product 8, which was used in the next step
without any further purification. LRMS m/z calculated for C31H33NO4S
[M+Na]+: 538.2. Found: 538.1.
Protein Expression and Purification - Codon-optimized full-length and
Δ3-22 human CYP3A4 were produced as reported previously19 and
used for assays and crystallization, respectively.
Spectral Binding Titrations – Equilibrium ligand binding to CYP3A4
was monitored in a Cary 300 spectrophotometer at ambient tempera-
ture in 0.1 M phosphate buffer, pH 7.4, supplemented with 20% gly-
cerol and 1 mM dithiothreitol. Inhibitors were dissolved in dimethyl
sulfoxide (DMSO) and added to a 2 μM protein solution in small ali-
quots, with the final solvent concentration < 2%. Spectral dissociation
constants (Ks) were determined from quadratic fits to titration plots.
Thermal Denaturation – Thermal denaturation curves were recorded in
0.1 M phosphate buffer, pH 7.4, in a Cary 300 spectrophotometer. Protein
(1 μM) was mixed with a ligand (20 μM) and incubated for 15 min at
General Procedure for Synthesis of Compounds 9a-b (Scheme 4) –
Crude 8 (0.15 g, 0.3 mmol) was dissolved in DMF (4 mL). To this so-
lution, EDAC (0.09 g, 0.45 mmol, 1.5 eq) and HOBt (0.07 g, 0.45 mmol,
1.5 eq) were added, followed by the addition of 3-(aminomethyl)pyr-
idine (0.05 g, 0.45 mmol, 1.5 eq) and DIPEA (0.12 g, 0.9 mmol, 3 eq).
The reaction was stirred at room temperature overnight. Upon
Scheme 4.
4

E.R. Samuels and I.F. Sevrioukova
Bioorganic&MedicinalChemistry28(2020)115349
calculate the remaining activity, with the DMSO-containing sample
used as a control (100% activity). The IC50 values were derived from
the [% activity] vs. [inhibitor] plots as described previously.16
Kinetics of Ligand Binding and Heme Reduction – Kinetics of ligand
binding to CYP3A4 and its reduction with sodium dithionite were
measured at 426 nm and 443 nm, respectively, in a SX.18MV stopped
flow apparatus (Applied Photophysics, UK), as described earlier.16
Determination of the X-ray Structures – Δ3-22 CYP3A4 was co-crys-
tallized with the inhibitors at room temperature by a microbatch
method under oil. Protein (70–80 mg/mL) was incubated with a 5–10-
fold ligand excess and centrifuged to remove the precipitate. The su-
pernatant (0.4–0.6 μl) was mixed with 0.4–0.6 μl of the crystallization
solution, containing 10–12% polyethylene glycol 3350 and either
60–80 mM sodium malonate, pH 6.0–7.0 (for 5a, 7b-d and 9b), 80 mM
succinate pH 7.0 (for 5b and 7a) or 80 mM malate, pH 7.0 (for 9a), and
covered with paraffin oil. After harvesting, crystals were cryoprotected
with Paratone-N and frozen in liquid nitrogen. X-ray diffraction data
were collected at the Stanford Synchrotron Radiation Lightsource
beamlines 9–2, 12–2 and 14–1, and the Advanced Light Source beam-
line 5.0.2. The high resolution cutoffs were chosen based on the CC1/2
value as recommended.20 Crystal structures were solved by molecular
replacement with PHASER21 and the 5VCC structure as a search model.
Ligands were built with eLBOW22 and manually fit into the density with
COOT.23 The initial models were rebuilt and refined with COOT and
PHENIX.22 For racemic compounds, the side-group configuration was
automatically assigned during the first refinement cycle. The following
refinement was conducted with all possible stereoisomer combinations
to confirm that the assigned chirality was most optimal. Polder omit
electron density maps were calculated with PHENIX. Data collection
and refinement statistics are summarized in Tables S1 and S2. The
atomic coordinates and structure factors for the 5a-, 5b-, 7a-, 7b-, 7c-,
7d-, 9a- and 9b-bound CYP3A4 were deposited in the Protein Data
Bank with the ID codes 6UNE, 6UNG, 6UNH, 6UNI, 6UNJ, 6UNK, 6UNL
and 6UNM, respectively.
3. Results and discussion
3.1. Rationale for series IV analogues
Structure-activity studies on series II and III analogues suggested
that an increase in hydrophobicity of the R2 side-group, which pre-
ferably binds near the heme-ligating pyridine (designated as P2 site),
could increase the inhibitory potency for CYP3A4.15,16 To test this
prediction, series IV analogues were produced by modifying two dif-
ferent scaffolds/parent compounds: 8f, the most potent series II in-
hibitor,15 and 4e-h, the high-affinity subgroup from series III16 (Fig. 1).
Compared to the latter compounds, 8f has a shorter pyridyl linker,
racemic N-phenyl at the R1 position, longer R1-R2 spacer, and indole
instead of phenyl as R2. The 8f scaffold was modified by either changing
the R2 stereochemistry or replacing indole with naphthalene: com-
pounds 5a (rac,R) and 5b (rac,rac), respectively (Fig. 2). In the 4e-h
scaffold, the R1-phenyl was kept unchanged and R2 was substituted
with indole or naphthalene in opposite configuration: 7a (S,R)/7b (R,S)
and 7c (R,S)/7d (S,R) pairs, respectively (Fig. 2). The opposing R1/R2
configuration is less favorable for the inhibition of CYP3A416 and,
hence, we reasoned that a pairwise comparison of the S/R and R/S
stereoisomers would help better detect the impacts of R2 substitution.
Two additional analogues with racemic naphthalene at R1 and R2 po-
sitions, and pyridyl-methyl or pyridyl-ethyl linker (9a and 9b, respec-
tively) were synthesized to determine how an increase in bulkiness/
hydrophobicity of both side-groups affects the binding mode and se-
lectivity of ligand binding. Interaction of analogues with the full-length
CYP3A4 was spectrally, kinetically and functionally analyzed, and their
co-crystal structures with Δ3-22 CYP3A4 were determined.
Fig. 2. Chemical structures of the investigated compounds.
23 °C. Melting curves were recorded at 260 nm using a 0.2 °C measure-
ment step, 0.9 °C/min ramp rate, and 50–75 °C temperature range. A
denaturation midpoint (melting temperature; Tm) was determined from
non-linear fittings to the melting curve as described earlier.16
Inhibitory Potency Assays – Inhibitory potency for the 7-benzyloxy-4-
(trifluoromethyl)coumarin (BFC) O-debenzylation activity of CYP3A4
was evaluated fluorometrically in a soluble reconstituted system. The
full-length CYP3A4 and rat cytochrome P450 reductase (40 μM and
60 μM, respectively) were preincubated at room temperature for 1 hr
before 20-fold dilution with the reaction buffer consisting of 0.1 M
potassium phosphate, pH 7.4, catalase and superoxide dismutase (2
Units/mL each), and 0.0025% CHAPS (3-[(3-cholamidopropyl)di-
methyl-ammonio]-1-propanesulfonate). Prior to measurements, 85 μl of
the reaction buffer was mixed with 10 μl of the NADPH-regenerating
system (10 mM glucose, 0.2 mM NADP+, and 2 Units/mL glucose-6-
phosphate dehydrogenase), 5 μl of the protein mixture (0.1 μM final
CYP3A4 concentration), and 2 μl of the inhibitor solution or DMSO. The
mixture was incubated for 2 min, after which 1 μl of 2 mM BFC and 1 μl
of 7 mM NADPH were added to initiate the reaction. Accumulation of
the fluorescent product, 7-hydroxy-4-(trifluoromethyl)coumarin, was
monitored for 2 min at room temperature in a Hitachi F400 fluorimeter
(λex = 404 nm; λem = 500 nm). Within this time interval, fluorescence
changes were linear. The average of three measurements was used to
Earlier, we identified parameters that, collectively, tend to correlate
with the inhibitory potency of ritonavir-like compounds.16 These
5

E.R. Samuels and I.F. Sevrioukova
Bioorganic&MedicinalChemistry28(2020)115349
include: (a) an absorbance maximum for the ligand-bound CYP3A4
(λmax) and a peak/trough amplitude in the difference spectra for ferric
ligand-free/ligand-bound forms (ΔAmax), which reflect electronic
properties of the ligating group and steric properties/environment
distant from the heme, respectively; (b) spectral dissociation constant
(Ks), a measure of the binding affinity; (c) changes in the melting
temperature (ΔTm), reflecting protein stability; and (d) CYP3A4 re-
duction rate (kET) that depends on accessibility and redox potential of
the heme cofactor. These parameters, as well as the ligand binding rate,
were measured for all new analogues and 8f, whose interaction with the
full-length CYP3A4 has not been previously investigated. Although ri-
tonavir is generally considered as a mechanism-based inhibitor,24–27
our prior work on series III inhibitors showed that preincubation of
ritonavir or its analogues with NADPH in a lipid-free reconstituted
system leads to a small increase rather than decrease in IC50, likely to
due to partial metabolism.16 Therefore, in this study, inhibitory assays
with NADPH preincubation were not conducted.
decreased accessibility of the reductant, sodium dithionite, as no con-
version to the inactivated P420 form was observed during lengthy ti-
trations (spectra of the ferrous CO-adduct are shown in blue in
Fig. 3A–C).
Comparable dissociation constants derived for 8f and 5a-b
(0.070–0.086 μM; Table 1) suggest that the binding affinity is only
weakly affected by changes in R2 hydrophobicity and configuration. In
contrast, the IC50 values for BFC debenzylase activity of CYP3A4 varied
by 4-to-6-fold: from 0.320 μM for 8f to 0.085 and 0.055 μM for 5a and
5b, respectively. To better understand this result, the ligand binding
and heme reduction rates were compared. Within the studied time in-
terval, both reactions had a fast phase and one or two slow phases. The
ET
rate constants for the fast phase (kfast and k , respectively) and their
fast
relative percentage are listed in Table 1, whereas kinetic traces and the
rate constants for the slow phase(s) can be seen in Figs. S2 and S3.
Ligation of 5a-b to CYP3A4 proceeded faster than the parent compound
(kfast of 6.5–7.5 s−1 vs. 2.7 s−1 for 8f), and the resulting complexes
ET
were reduced with a slower rate (~30% decrease in k ). This suggests
fast
that 5a-b lower the heme accessibility to the higher degree and, as a
consequence, inhibit the substrate access and metabolism more po-
tently than 8f. It should be noted also that a slow decay of the 443 nm
absorbance was observed for the 5b-bound CYP3A4 (Fig. S3), likely due
to structural rearrangements in the active site. The ability of 5b to
destabilize the ferrous form could contribute to the inhibitory potency,
which was the highest in the subgroup.
3.2. Series II modification (5a-b)
Spectral and biochemical properties – As expected, upon binding to
CYP3A4, 8f and its derivatives, 5a and 5b, induced a red shift in the
Soret band (red spectra in Fig. 3A–C), indicative of the pyridine ni-
trogen ligation to the heme iron. While λmax remained the same
(422 nm; Table 1), the ΔAmax value, expressed as a percentage of that
for ritonavir (Fig. S1), decreased from 111% for 8f to 104 and 107% for
5a and 5b, respectively. One striking difference was a considerably
lower amplitude of the 443 nm peak for the ferrous CYP3A4-5b com-
plex (compare green spectra in Fig. 3A–C). This could result from a
Although there was no clear correlation between IC50, ΔAmax, Ks and
Tm (Table 1), a relationship between the side-group stereochemistry
and IC50 was noticeable for 8f and 5a. Therefore, structural information
on the inhibitory complexes was obtained to identify stereo
Fig. 3. A–C, Spectral changes in CYP3A4 induced by 8f and 5a-b, respectively. Absorbance spectra of oxidized ligand-free and ligand-bound CYP3A4 are in black and
red, respectively. Spectra of ferrous ligand-bound CYP3A4 and its CO-adduct are in green and blue, respectively. Left insets are difference spectra recorded during
equilibrium titrations. Right insets are titration plots with quadratic fittings. Spectral dissociation constants (Ks) derived from titration plots are listed in Table 1. D
and E, The S/R and S/S conformers of 5a and 5b, respectively, bound to CYP3A4 in the crystal structures (PDB ID 6UNE and 6UNG, respectively). The adjacent I-helix
and F304 in the inhibitory complexes and water-bound CYP3A4 (4I3Q structure) are shown in gray and black, respectively. Polder omit maps contoured at 3σ level
are shown as green mesh. F and G, Structural overlay of the CYP3A4-8f (S,S) complex (PDB ID 6BDM) with the 5a- and 5b-bound models, respectively. H and I, Side
and top views at superimposed 5a/b-CYP3A4 structures.
6

E.R. Samuels and I.F. Sevrioukova
Bioorganic&MedicinalChemistry28(2020)115349
Table 1
Properties of series IV inhibitors.
ET f
Compound
λmax (nm) ligand-bound
ΔAmax (%)a
KsbμM
IC50cμM
IC50/Ks
ΔTmd°C
kfastes−1
k
s−1
fast
5-atom R1-R2 spacer, pyridyl-methyl linker
R1-N-phenyl/R2-indole
8f (rac,S)
5a (rac,R)
422
422
111
104
0.074
0.086
0.003
0.003
0.320
0.085
0.018
0.005
4.3
1.0
+4.0
+2.5
2.7 (55%)
7.5 (39%)
0.061 (23%)
0.047 (35%)
R1-N-phenyl/R2 -naphthalene
5b (rac,rac)
422
107
0.070
0.002
0.055
0.005
0.8
+3.3
6.5 (40%)
0.038 (36%)
4-atom R1-R2 spacer
pyridyl-ethyl linker
R1-phenyl/R2-indole
7a (S, R)
422
422
108
110
0.062
0.085
0.002
0.002
0.173
0.102
0.016
0.009
2.8
1.2
+3.8
+4.1
4.1 (49%)
6.8 (43%)
0.015 (46%)
0.020 (28%)
7b (R, S)
pyridyl-ethyl linker
R
1-phenyl/R2-naphthalene
7c (S, R)
7d (R, S)
422
422
110
108
0.052
0.017
0.002
0.001
0.356
0.075
0.031
0.006
6.8
4.4
+4.5
+4.5
9.0 (41%)
0.016 (15%)
0.027 (25%)
10.2 (43%)
pyridyl-methyl linker
R1-naphthalene/R2-naphthalene
9a (rac,rac)
421
98
0.190
0.037
0.010
0.002
0.244
0.218
0.012
1.3
5.9
+4.0
+4.1
1.3 (33%)
3.9 (33%)
0.035 (32%)
0.027 (41%)
pyridyl-ethyl linker
R1-naphthalene/R2-naphthalene
9b (rac,rac)
422
105
0.0235
a
Maximal absorbance change in the ferric ligand-bound CYP3A4 was calculated from the difference spectra (Fig. S1) and expressed as a percentage relative to that
induced by ritonavir.
b
c
d
e,f
Spectral dissociation constant for the CYP3A4-inhibitor complex determined from titration plots.
The half maximal inhibitory concentration for the BFC debenzylase activity of CYP3A4.
Ligand-dependent change in the melting temperature of CYP3A4.
Rate constants for the fast phase of the ligand binding reaction and reduction of the ligand-bound CYP3A4 with sodium dithionite, respectively.
configuration preferably selected by CYP3A4.
b have comparable binding affinity (0.062–0.085 μM) and association
5a-b binding modes and configuration – Crystal structures of CYP3A4
bound to 5a and 5b were solved to 2.55 and 2.30 Å resolution, re-
spectively (Table S1). Assignment of stereo configuration was based on
visual inspection and refinement statistics, where lower R/Rfree-factors
indicate better agreement between the model and observed X-ray dif-
fraction data. Similar to 8f (rac,S), whose S/S stereoisomer co-crystal-
lized with CYP3A4,15 the S-configuration of R1-N-phenyl was found to
provide a better fit for both 5a and 5b. The 5b R2-naphthalene was also
in S-configuration. Thus, for this particular scaffold, the S/S side-group
conformation appears to promote association to CYP3A4.
rate (4.1–6.8 s−1), and their complexes with CYP3A4 are equally stable
(ΔTm of 3.8–4.1 °C) and reduced with a similar rate (0.017–0.020 s−1
;
Table 1). This correlates well with a small difference in the IC50 values:
0.173 and 0.102 μM for 7a and 7b, respectively.
For the naphthalene-containing 7c-d, kfast and ΔTm were also similar
and the highest overall (9.0–10.2 s−1 and 4.5 °C, respectively).
Conversely, the heme reduction rate was one of the lowest
ET
(0.016–0.027 s−1). For comparison, k
for the ligand-free enzyme
fast
is > 100-fold higher (4.6 s−1; Fig. S3). Thus, ligation of 7c-d and other
analogues drastically decreases accessibility of the active site, even for
small molecules such as sodium dithionite. Importantly, in the sub-
series, 7d had the highest binding affinity and inhibitory potency
(0.017 μM and 0.075 μM, respectively), which were 3-to-5-fold lower
for 7c (Table 1). This further confirms that, for the series III scaffold,
the S/R configuration is less favorable for association and inhibition of
CYP3A4.
Positioning of 5a-b relative to the central I-helix and electron den-
sity maps are shown in Fig. 3D,E. Both compounds bind in a traditional
orientation, with R1 inserted into a hydrophobic pocket adjacent to the
I-helix (P1 site) and R2 near the heme-ligating pyridine (P2 site). 5a-b
also establish an H-bond with the active site S119 and largely displace
the I-helix (by 1.5 and 1.9 Å, respectively). As shown in Fig. 3F,G, 5a
(S,R) orients similarly to 8f (S,S) (6BDM structure), whereas 5b (S,S)
adopts a different conformation to accommodate and optimally place
the bulky R2-naphthalene. As a result, the terminal tert-butylox-
ycarbonyl (Boc)-group of 5b points in opposite direction.
Crystal structures of CYP3A4 complexed with 7a-d were solved to
2.60–2.75 Å resolution (Tables S1 and S2) and will also be compared in
a pairwise fashion.
7a-b binding modes – Both 7a and 7b bind to CYP3A4 in a traditional
orientation and trigger a large distortion in the I-helix (1.8–2.0 Å;
Fig. 5A,B). Because of distinct backbone curvatures, the Boc-groups
point in opposite directions (Fig. 5C). Structure comparison shows that
7a forms shorter Fe-N and O21-S119 bonds, has larger aromatic overlap
between the parallel stacked R1-phenyl and F304 rings, and its F’-G’
fragment and end-group are more ordered, enabling formation of
multiple van der Waals contacts (Table 2). Nonetheless, 7b has a higher
inhibitory potency for CYP3A4 (Table 1), which could be explained by
its distinct, less compact fold. Unlike the S/R conformer, 7b partially
occupies the substrate binding channel and, hence, could more effi-
ciently block the substrate access to the catalytic center.
The outlined differences are better seen in Fig. 3H,I, where or-
ientations of 5a and 5b are directly compared. One striking feature is
the coinciding location of the aromatic side-groups, leading to an equal
overlap between the R1/F304 and R2/pyridine rings (Table 2). Even so,
5b ligates to the heme stronger (forms the shortest Fe-N bond), estab-
lishes more extensive hydrophobic interactions at P2 site through the
bulkier R2-naphtalene, and inserts R1-phenyl deeper into the P1 pocket.
This explains why this analogue inhibits CYP3A4 more potently than
5a.
3.3. Series III modification (7a-d)
7c-d binding modes – The CYP3A4-7c complex was crystallized in the
C2 space group, with two molecules per asymmetric unit. The better
defined molecule A was used for comparative analysis outlined below.
One characteristic feature of 7c is the reverse side-group orientation:
R1-phenyl is placed at P2 site perpendicular to the heme-ligating
Spectral and biochemical properties – The λmax and ΔAmax values
derived for the 7a-d subseries were similar (Fig. 4A–D; Table 1). The
only spectral difference was the lower amplitude of the 443 nm peak for
the ferrous CYP3A4-7c complex. A pairwise comparison shows that 7a-
7

E.R. Samuels and I.F. Sevrioukova
Bioorganic&MedicinalChemistry28(2020)115349
Table 2
Structural features of the CYP3A4-inhibitor complexes.
Compound
Fe-N bond
Pyridine ring
I-helix
H-bond with Pyridine-R2 ring Phe304-R1 ring
F-G Disorder Boc-group
c
rotation (°)b
displacement (Å)
Ser119 (Å)d
angle and
overlap
angle and
overlap
conformation/
Distance (Å) Angle (°)a
contacts
5a (S,R)e
5b (S,S)e
7a (S,R)
7b (R,S)
2.34
2.20
2.25
2.40
12
2
37
30
25
28
1.54–1.49
1.89–1.85
1.77–1.52
1.77–1.97
2.68
2.85
2.85
3.31
55°; half
Parallel; half
20°; half
210–212
210–212
198–217
210–212
Disordered
56°; partial
−38°; half
44°; half
Disordered
8
Parallel; full
42°; half
Disordered
3
Traceable; 105–108,
120
7c (S,R) molecule 2.82
0
3
15
35
2.18–2.25
1.55–1.25
–
−85°; nonef
52°; partialf
None
ordered; 213, 215,
371, 482
A
7d (R,S)
9a (S,R)e
2.37
2.40
2.32
35°; partial
45°; partial
210–213
198–217
Traceable; 106, 108,
215, 374
10
10
32
35
1.20–1.93
3.28
25°; half
20°; half
204–217
Ordered; 108, 220
Traceable, 105–108,
374
9b (S,S)e molecule 2.30
2.03–1.83 2.57
25°; full
25°; partial
A
a
Deviation from perpendicularity.
b
c
d
e
f
Angle between the planes passing through the pyridine ring and the NB-ND heme atoms.
Distance between Cα-atoms of Phe304 and Ala305 in the inhibitor-bound and ligand-free CYP3A4 (PDB ID 5VCC).
Hydrogen bond length between inhibitor’s carbonyl oxygen atom and Ser119 hydroxyl group.
These setereoisomers selectively co-crystallized with CYP3A4.
In 7c, R1 and R2 side groups are in reverse orientation and placed near the heme-ligating pyridine and Phe304, respectively.
pyridine and parallel to the cofactor, whereas R2-naphthalene is in the
P1 pocket (Fig. 6A). This is the third incident when the ritonavir-like
compound ligates to CYP3A4 in such a distinct conformation. Two
previous cases involved series III analogs, 4a and 4e (6DA2 and 6DAB
structures, respectively), whose R1/R2-phenyls were also in S/R con-
figuration,16 Evidently, the 7c conformation is unfavorable because it
markedly weakens the FeeN bond, leads to an extreme I-helix distor-
tion, prevents H-bonding to S119, and minimizes/eliminates aromatic
overlap mediated by the side-groups (Table 2).
virtually the same spot at P2 site, which is achieved through adjust-
ments in the backbone curvature and R1-F304 interplay. Despite this
conformational similarity, modification of R2 had a larger impact on the
binding and inhibitory strength than for respective S/R analogues. The
phenyl-to-indole substitution in R2 led to a 2-fold increase in Ks (0.045
vs. 0.085 μM for 4f and 7b, respectively), whereas the binding affinity
of naphthalene-containing 7d was 3-to-5-fold lower (0.017 μM). A more
drastic improvement was observed in IC50, which decreased from
0.68 μM for 4f to 0.102 and 0.075 μM for 7b and 7d, respectively. This
further supports the notion that the potency of ritonavir-like inhibitors
largely depends on the R2 side-group functionality.
To better understand the impact of R2 substitution, we compared
the binding modes of 7c, 7a and 4e stereoisomers, whose R2 func-
tionality is represented by naphthalene, indole and phenyl, respec-
tively. As seen in Fig. 6C, 7a changes the side-group tilt angle and depth
of immersion to overcome the disadvantage of the S/R configuration. It
is possible that the higher polarity of the indole ring precludes its em-
bedment into the highly hydrophobic P1 pocket. However, that all
three S/R analogues have similar Ks (0.052, 0.062 and 0.042 μM for 7c,
7a and 4e,16 respectively) suggests that the binding affinity is not sig-
nificantly affected by changes in R2 hydrophobicity and/or side-group
placement. The IC50 value, however, varied more dramatically and
decreased by 2-fold upon R2-phenyl-to-naphthalene substitution (0.83
vs. 0.36 μM for 4e and 7c, respectively), and by another 2-fold for the
R2-indole containing 7a (0.17 μM). Thus, hydrophobic and aromatic
interactions mediated by R2 at P2 site are important and largely affect
the inhibitory potency for CYP3A4.
3.4. Analogues with naphthalene at R1 and R2 positions
Since placement of naphthalene at R2 position markedly lowered
IC50 for 5b and 7d, we next tested if substitution of both side-groups
with naphthalene could further improve the inhibitory strength. For
this purpose, two analogues were synthesized with either pyridyl-me-
thyl or pyridyl-ethyl linker, for length comparison, and racemic R1/R2-
naphthalene, for group size investigation: 9a and 9b, respectively
(Fig. 2).
Spectral and biochemical properties – Absorbance changes induced by
9a-b and titration plots are shown in Fig. 7A,B. As evident from the
lowest λmax and ΔAmax values (Table 1), 9a triggers the least prominent
perturbations in the Soret band upon ligation. The common features of
9a-b were the lowest 443 nm absorption of the ferrous ligand-bound
forms (compare green spectra in Figs. 3A–C, 4, and 7A-B) and slow
association rates (1.3–3.9 s−1). Both analogues had a similar stabilizing
effect (ΔTm of 4.0–4.1 °C) and inhibitory potency (IC50 of
~0.22–0.24 μM) but largely differed in the binding affinity for CYP3A4:
Ks of 0.190 and 0.037 μM for 9a and 9b, respectively. Thus, simulta-
neous placement of naphthalene at R1 and R2 positions does not im-
prove the inhibitory potency, whereas elongation of the pyridyl spacer
enhances the binding affinity regardless of the side-group chemical
nature.
Unlike 7c, its enantiomer 7d binds to CYP3A4 in a traditional or-
ientation that enables strong heme ligation and H-bonding to S119
(Fig. 6B; Table 2). The 7d R2-naphthalene is placed considerably closer
to the heme than R1-phenyl of 7c (3.5 Å vs. 5.0 Å, respectively) and
with a larger tilt angle, which not only increases an area of aromatic
overlap but also enables π-cation interactions with the nearby R105
guanidine (Fig. 6B). One advantage of 7c is the lower R2 position,
preventing steric clashing with the adjoining F’-G’-loop (residues
212–215). As a result, the latter fragment stays in place and provides
stabilizing hydrophobic and van der Waals contacts in the CYP3A4-7c
complex but becomes partially disordered upon binding of 7d (Fig. 6D).
It is possible though that this conformational distinction arises from
different crystal packing.
9a-b binding modes – Crystal structures of the 9a/b-CYP3A4 com-
plexes were determined to 2.55–2.83 Å resolution (Table S2). Similar to
the 7c-bound form, CYP3A4-9b crystallized in the C2 space group, with
two molecules per asymmetric unit. The better defined molecule A was
used for comparative analysis, which revealed that 9a and 9b were in
S/R and S/S configuration, respectively. The ligand orientation and
It was also of interest to compare orientations of 7d, 7b and 4f
stereoisomers, differing only in the R2 group (naphthalene, indole and
phenyl, respectively). As seen in Fig. 6E, the R2 moieties occupy
8

E.R. Samuels and I.F. Sevrioukova
Bioorganic&MedicinalChemistry28(2020)115349
Fig. 5. A and B, The binding mode of 7a and 7b. The adjacent I-helix and F304
in the inhibitory complexes and water-bound CYP3A4 (4I3Q structure) are
shown in black and gray, respectively. Polder omit maps contoured at 3σ level
are shown as green mesh. C, Superposition of the 7a/b-bound CYP3A4 (6UNH
and 6UNI structures, respectively).
3.5. Comparison of 7d and ritonavir
Series IV had three lead compounds, 5a-b and 7d, whose IC50 was in
the low submicromolar range (< 0.1 μM). For comparison, under
identical experimental conditions, ritonavir inhibits CYP3A4 with IC50
of 0.13 μM.16 Besides being a better inhibitor, 7d was also a high af-
finity ligand for CYP3A4: Ks 0.017 vs. 0.019 μM for ritonavir.16 More-
over, among the high affinity/potency inhibitors that were designed so
far, only 7d orients most similarly to ritonavir (Fig. 8). The R1-phenyls
of both compounds penetrate the P1 pocket equally as deep and nearly
coincide, whereas the 7d R2-naphthalene trails ritonavir’s R2-phenyl
but has a larger tilt angle (25° difference vs. 40–80° for other analo-
gues). Having two conjoined phenyl rings, the naphthalene is poised to
form stronger aromatic interactions with the heme and pyridine rings
and, unlike ritonavir’s R2-phenyl, can establish π-cation interactions
with the R105 guanidine. These additional protein-ligand contacts are
important because they not only strengthen and stabilize the inhibitory
complex but also decrease the ligand mobility and allow a more effi-
cient blockage of the active site. That 7d is structurally simpler and
easier to synthesize than ritonavir but inhibits CYP3A4 twice as
strongly gives grounds for optimism that more potent and less ex-
pensive pharmacoenhancers could be developed.
Fig. 4. A–D, Spectral changes in CYP3A4 induced by 7a-d, respectively.
Absorbance spectra of oxidized ligand-free and ligand-bound CYP3A4 are in
black and red, respectively. Spectra of ferrous ligand-bound CYP3A4 and its CO-
adduct are in green and blue, respectively. Left insets are difference spectra
recorded during equilibrium titrations. Right insets are titration plots with
quadratic fittings. Spectral dissociation constants (Ks) derived from titration
plots are listed in Table 1.
electron density maps can be seen in Fig. 7C,D. Both analogues ligate to
the heme in a traditional orientation, with the R2-naphthalenes occu-
pying the same spot near the pyridine ring but with a different tilt
angle. In 9a, the shorter pyridyl-linker disallows insertion of the R1-
naphthalene into the P1 pocket, owing to which this moiety points
upward toward the roof of the active site, pushing F241 aside. In con-
trast, the elongated 9b adopts more relaxed conformation, with R1 fully
immersed into the P1 site. Nonetheless, both R1-naphthalenes are near
parallel to the F304 ring, which flips to maximize aromatic overlap
(Fig. 7E). Structure comparison shows that 9b ligates to the heme
stronger, forms a shorter H-bond with S119, and has a larger R2/pyr-
idine overlap which, collectively, could improve the binding affinity
and inhibitory strength. That 9a-b had similar IC50 (Table 1) could be
due to utilization of racemic mixtures in the functional assays.
4. Concluding summary
This work was a continuation of our studies on ritonavir-like in-
hibitors of CYP3A4, aimed to determine the relative impact of the side-
group size and hydrophobicity on the binding and inhibitory strength.
Two different scaffolds were used to design eight series IV inhibitors
with phenyl or naphthalene at R1 position and indole or naphthalene as
R2. Interaction of the analogues with CYP3A4 was spectrally,
9

E.R. Samuels and I.F. Sevrioukova
Bioorganic&MedicinalChemistry28(2020)115349
Fig. 6. A and B, The binding mode of 7c and 7d, respectively. The adjacent I-helix and F304 in the inhibitory complexes and water-bound CYP3A4 (4I3Q structure)
are shown in black and gray, respectively. Polder omit maps contoured at 3σ level are shown as green mesh. The CYP3A4-7c complex crystallizes in a distinct space
group (Table S2) with two molecules per asymmetric unit; the better-defined molecule A was used for structural comparison. C, Superposition of the 7a-, 7c- and 4e-
bound structures of CYP3A4 (6UNH, 6UNJ and 6DAB, respectively). 4e is a series III inhibitor with a similar backbone and R1/R2-phenyls in S/R configuration. D,
Superposition of the 7c/d-bound CYP3A4. E, Superposition of the 7b-, 7d- and 4f-bound structures of CYP3A4 (6UNI, 6UNK and 6DAC, respectively). 4f is a series III
inhibitor with a similar backbone and R1/R2-phenyls in R/S configuration.
biochemically and structurally characterized and, when appropriate,
comparisons with the previously reported compounds were conducted
to better understand structure-activity relationships.
that an increase in size and hydrophobicity/aromaticity of the R2-sub-
stituent markedly increases the inhibitory potential of ritonavir-like
molecules, with IC50 decreasing from mid- to low-submicromolar range
In agreement with the prior suggestions, our results demonstrate
upon phenyl > indole > naphthalene substitution. In contrast,
Fig. 7. A and B, Spectral changes in CYP3A4 induced by 9a–d, respectively. Absorbance spectra of oxidized ligand-free and ligand-bound CYP3A4 are in black and
red, respectively. Spectra of ferrous ligand-bound CYP3A4 and its CO-adduct are in green and blue, respectively. Left insets are difference spectra recorded during
equilibrium titrations. Right insets are titration plots with quadratic fittings. Spectral dissociation constants (Ks) derived from titration plots are listed in Table 1. C
and D, The binding mode of 9a and 9b (PDB ID 6UNL and 6UNM, respectively). The adjacent I-helix and F304 in the inhibitory complexes and water-bound CYP3A4
(4I3Q structure) are shown in gray and black, respectively. Polder omit maps contoured at 3σ level are shown as green mesh. The CYP3A4-9b complex was
crystallized in a distinct space group (Table S2) with two molecules per asymmetric unit; the better-defined molecule A was used for structural comparison. E,
Superposition of the 9a/b-bound CYP3A4.
10

E.R. Samuels and I.F. Sevrioukova
Bioorganic&MedicinalChemistry28(2020)115349
Funding
This work was supported by the National Institutes of Health Grant
ES025767.
Appendix A. Supplementary material
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.bmc.2020.115349.
References
1. Zanger UM, Schwab M. Cytochrome P450 enzymes in drug metabolism: regulation of
gene expression, enzyme activities, and impact of genetic variation. Pharmacol Ther.
2013;138:103–141.
2. Zhou SF. Drugs behave as substrates, inhibitors and inducers of human cytochrome
P450 3A4. Curr Drug Metab. 2008;9:310–322.
3. Xu L, Desai MC. Pharmacokinetic enhancers for HIV drugs. Curr Opin Investig Drugs.
2009;10:775–786.
Fig. 8. Relative positioning of 7d (in magenta), ritonavir (in cyan; 5VCO
structure) and other high affinity/inhibitory potency analogues (8f, 5a-b, 7b
and 4f-h; in black), whose end-groups were removed for clarity of presentation.
Location of the P1 and P2 sites and the entrance to the substrate binding
channel are indicated.
4. Chapman SA, Lake KD, Solbrack DF, Milfred SK, Marshall PS, Kamps MA.
Considerations for using ketoconazole in solid organ transplant recipients receiving
cyclosporine immunosuppression. J Transpl Coord. 1996;6:148–154.
5. Kempf DJ, Marsh KC, Denissen JF, et al. ABT-538 is a potent inhibitor of human
immunodeficiency virus protease and has high oral bioavailability in humans. Proc
Natl Acad Sci USA. 1995;92:2484–2488.
6. Brayer SW, Reddy KR. Ritonavir-boosted protease inhibitor based therapy: a new
strategy in chronic hepatitis C therapy. Expert Rev Gastroenterol Hepatol.
2015;9:547–558.
simultaneous placement of naphthalene at R1 and R2 position improves
the binding but not inhibitory strength.
7. Ke AB, Zamek-Gliszczynski MJ, Higgins JW, Hall SD. Itraconazole and clarithromycin
as ketoconazole alternatives for clinical CYP3A inhibition studies. Clin Pharmacol
Ther. 2014;95:473–476.
For the series IV compounds, stereochemistry became a more im-
portant contributing factor, as the bulky side-groups limit the ability to
adjust and optimize interactions within the active site. Usage of racemic
mixtures complicated experimental data interpretation but, at the same
time, provided valuable insights into stereoselectivity of ligand binding.
The relationship between R1/R2-configuration and preferential binding
to CYP3A4 was not straightforward and depended on both the back-
bone spacing and side-group functionalities and interplay. The S/S and
R/S configurations seem to be more preferable for the association to
CYP3A4, as the S/R conformers tend to ligate with the unfavorable
reverse side-group orientation.
8. Greenblatt DJ, Harmatz JS. Ritonavir is the best alternative to ketoconazole as an
index inhibitor of cytochrome P450–3A in drug-drug interaction studies. Br J Clin
Pharmacol. 2015;80:342–350.
9. Sevrioukova IF, Poulos TL. Interaction of human cytochrome P4503A4 with ritonavir
analogs. Arch Biochem Biophys. 2012;520:108–116.
10. Sevrioukova IF, Poulos TL. Pyridine-substituted desoxyritonavir is a more potent
cytochrome P450 3A4 inhibitor than ritonavir. J Med Chem. 2013;56:3733–3741.
11. Sevrioukova IF, Poulos TL. Dissecting cytochrome P450 3A4-ligand interactions
using ritonavir analogues. Biochemistry. 2013;52:4474–4481.
12. Sevrioukova IF, Poulos TL. Structure and mechanism of the complex between cyto-
chrome P4503A4 and ritonavir. Proc Natl Acad Sci USA. 2010;107:18422–18427.
13. Sevrioukova IF, Poulos TL. Ritonavir analogues as a probe for deciphering the cy-
tochrome P450 3A4 inhibitory mechanism. Curr Top Med Chem. 2014;14:1348–1355.
14. Kaur P, Chamberlin AR, Poulos TL, Sevrioukova IF. Structure-based inhibitor design
for evaluation of a CYP3A4 pharmacophore model. J Med Chem.
2016;59:4210–4220.
Finally, for the first time, this study demonstrated that less struc-
turally complex inhibitors that are more potent than ritonavir could be
rationally developed. Among three lead compounds, 5a-b and 7d, the
latter had the highest affinity for CYP3A4 and, thus, it could be used as
a template for further optimization.
15. Samuels ER, Sevrioukova IF. Inhibition of human CYP3A4 by rationally designed
ritonavir-like compounds: Impact and interplay of the side group functionalities. Mol
Pharm. 2018;15:279–288.
16. Samuels ER, Sevrioukova I. Structure-activity relationships of rationally designed
ritonavir analogs: Impact of side-group stereochemistry, head-group spacing, and
backbone composition on the interaction with CYP3A4. Biochemistry.
2019;58:2077–2087.
Declaration of Competing Interest
17. Chong HS, Song HA, Dadwal M, Sun X, Sin I, Chen Y. Efficient synthesis of func-
tionalized aziridinium salts. J Org Chem. 2010;75:219–221.
18. Samuels E, Sevrioukova I. Direct synthesis of alpha-thio aromatic acids from aro-
matic amino acids. Tetrahedronlett. 2018;59:1140–1142.
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influ-
ence the work reported in this paper.
19. Sevrioukova IF. High-level production and properties of the cysteine-depleted cyto-
chrome P450 3A4. Biochemistry. 2017;56:3058–3067.
20. Karplus PA, Diederichs K. Linking crystallographic model and data quality. Science.
2012;336:1030–1033.
21. McCoy AJ, Grosse-Kunstleve RW, Adams PD, Winn MD, Storoni LC, Read RJ. Phaser
crystallographic software. J Appl Crystallogr. 2007;40:658–674.
22. Adams PD, Afonine PV, Bunkoczi G, et al. PHENIX: a comprehensive Python-based
system for macromolecular structure solution. Acta Crystallogr Section D.
2010;66:213–321.
Acknowledgements
This work involves research carried out at the Stanford Synchrotron
Radiation Lightsource and the Advanced Light Source. Use of the
Stanford Synchrotron Radiation Lightsource, SLAC National Accelerator
Laboratory, is supported by the U.S. Department of Energy, Office of
Science, Office of Basic Energy Sciences under Contract No. DE-AC02-
76SF00515. The SSRL Structural Molecular Biology Program is sup-
ported by the DOE Office of Biological and Environmental Research,
and by the National Institutes of Health, National Institute of General
Medical Sciences (including P41GM103393). The Advanced Light
Source is supported by the Director, Office of Science, Office of Basic
Energy Sciences, of the U.S. Department of Energy under Contract No.
DE-AC02-05CH11231.
23. Emsley P, Lohkamp B, Scott WG, Cowtan K. Features and development of Coot. Acta
Crystallogr Section D. 2010;66:486–501.
24. Koudriakova T, Iatsimirskaia E, Utkin I, et al. Metabolism of the human im-
munodeficiency virus protease inhibitors indinavir and ritonavir by human intestinal
microsomes and expressed cytochrome P4503A4/3A5: mechanism-based inactiva-
tion of cytochrome P4503A by ritonavir. Drug Metab Dispos. 1998;26:552–561.
25. von Moltke LL, Durol AL, Duan SX, Greenblatt DJ. Potent mechanism-based inhibi-
tion of human CYP3A in vitro by amprenavir and ritonavir: comparison with keto-
conazole. Eur J Clin Pharmacol. 2000;56:259–261.
26. Obach RS, Walsky RL, Venkatakrishnan K. Mechanism-based inactivation of human
cytochrome p450 enzymes and the prediction of drug-drug interactions. Drug Metab
Dispos. 2007;35:246–255.
27. Rock BM, Hengel SM, Rock DA, Wienkers LC, Kunze KL. Characterization of rito-
navir-mediated inactivation of cytochrome P450 3A4. Mol Pharmacol.
2014;86:665–674.
11