Identification of Potential Off-target Toxicity Liabilities of Catechol-O-methyltransferase Inhibitors by Differential Competition Capture Compound Mass Spectrometry

Article abstract of DOI:10.1021/acs.jmedchem.5b01970

Structurally related inhibitors of a shared therapeutic target may differ regarding potential toxicity issues that are caused by different off-target bindings. We devised a differential competition capture compound mass spectrometry (dCCMS) strategy to ef

Full text of DOI:10.1021/acs.jmedchem.5b01970

Article
pubs.acs.org/jmc
Identification of Potential Off-target Toxicity Liabilities
of Catechol‑O‑methyltransferase Inhibitors by Differential
Competition Capture Compound Mass Spectrometry
Lisa von Kleist,† Simon Michaelis, Kathrin Bartho,‡ Olivia Graebner,§ Maren Schlief, Mathias Dreger,
́
Anna K. Schrey, Michael Sefkow,∥ Friedrich Kroll,⊥ Hubert Koester, and Yan Luo*
Caprotec Bioanalytics GmbH, Magnusstraße 11, 12489 Berlin, Germany
S
* Supporting Information
ABSTRACT: Structurally related inhibitors of a shared
therapeutic target may differ regarding potential toxicity issues
that are caused by different off-target bindings. We devised a
differential competition capture compound mass spectrometry
(dCCMS) strategy to effectively differentiate off-target profiles.
Tolcapone and entacapone are potent inhibitors of catechol-O-
methyl transferase (COMT) for the treatment of Parkinson’s
disease. Tolcapone is also known for its hepatotoxic side
effects even though it is therapeutically more potent than
entacapone. Here, we identified 3-hydroxyisobutyryl-CoA
hydrolase (HIBCH) as a possible toxicity-causing off-target
of tolcapone, and this protein is not bound by the less toxic
COMT inhibitor entacapone. Moreover, two novel compounds from a focused library synthesized in-house, N2,N2,N3,N3-
tetraethyl-6,7-dihydroxy-5-nitronaphthalene-2,3-dicarboxamide and 5-(3,4-dihydroxy-5-nitrobenzylidene)-3-ethylthiazolidine-2,4-
dione, were utilized to gain insight into the structure−activity relationships in binding to COMT and the novel off-target
HIBCH. These compounds, especially N2,N2,N3,N3-tetraethyl-6,7-dihydroxy-5-nitronaphthalene-2,3-dicarboxamide, could serve
as starting point for the development of improved and more specific COMT inhibitors.
competition capture compound mass spectrometry (dCCMS) to
elucidate the on- and off-target profiles of tolcapone in human
hepatoma-derived cell line HepG2 lysate. dCCMS is the newest
development of the CCMS technology, an effective chemical
proteomics approach based on a unique trifunctional chemical
structure combined with mass spectrometry.9,10
INTRODUCTION
■
Catechol-O-methyltransferase (COMT) is a magnesium dependent
enzyme responsible for the O-methylation of catecholamine
neurotransmitters in the brain.1 Over the last 50 years, COMT
has become an attractive target for the treatment of various
peripheral and central nervous system disorders. Clinically
relevant COMT inhibitors for the adjunctive treatment of
Parkinson’s disease are exemplified by tolcapone, which acts both
centrally and peripherally, and entacapone, which acts mainly
in the periphery. Both drugs, but entacapone in particular, are
commonly coadministered with levodopa, a precursor of
dopamine, to increase its half-life in vivo.2,3 Tolcapone is more
efficient, but due to cases of fulminant hepatic failure, it has been
withdrawn from the market in some countries.4−6 Ten years ago,
the suspension was lifted, and tolcapone was reintroduced with
requirements for careful liver function monitoring.7 Entacapone
has a more favorable toxicity profile but suffers from bio-
availability issues.8
We were interested in understanding the underlying mechanism
of the adverse side effects of tolcapone as well as in finding novel
potent COMT inhibitors with improved off-target profiles. While
this type of question is typically addressed by testing a panel of
candidate off-targets, we here aim at an unbiased approach,
which takes the entire proteome of a given biological sample
into account. Therefore, we decided to employ differential
The outcome of these studies led to the identification of
the protein 3-hydroxyisobutyryl-CoA hydrolase (HIBCH) as a
dominant off-target of tolcapone but not of entacapone. In
addition, two novel compounds, N2,N2,N3,N3-tetraethyl-6,7-
dihydroxy-5-nitronaphthalene-2,3-dicarboxamide and 5-(3,4-
dihydroxy-5-nitrobenzylidene)3-ethylthiazolidine-2,4-dione,
were identified as potent COMT inhibitors. The mitochondrial
toxicity profile of tolcapone, entacapone, N2,N2,N3,N3-tetraethyl-
6,7-dihydroxy-5-nitronaphthalene-2,3-dicarboxamide, and
5-(3,4-dihydroxy-5-nitrobenzylidene)3-ethylthiazolidine-2,4-
dione was consistent with the off-target interaction profile
of the compounds, supporting our hypothesis that HIBCH is
a candidate off-target underlying the observed mitotoxicity.
The application of dCCMS provides additional insight into the
on- and off-target profiles of this class of inhibitors.
Received: January 4, 2016
© XXXX American Chemical Society
A
DOI: 10.1021/acs.jmedchem.5b01970
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
one nitro group using 2,3,5,6-tetrabromo-4-methyl-4-nitrocyclo-
hexa-2,5-dienone.15 The structure of compound 12 could be
confirmed by two-dimensional NMR spectroscopy and mass
spectrometry.
RESULTS AND DISCUSSION
■
Synthesis and in Vitro Analysis of Two Novel COMT
Inhibitors. On the basis of our in-house directed structure-
based drug design of COMT inhibitors (data not shown), we
focused our work in part on heterocyclyl nitrocatechol moieties
to generate new COMT inhibitors. Two small molecules,
N2,N2,N3,N3-tetraethyl-6,7-dihydroxy-5-nitronaphthalene-2,3-
dicarboxamide (3) and 5-(3,4-dihydroxy-5-nitrobenzylidene)3-
ethylthiazolidine-2,4-dione (12) (Figure 1), were synthesized as
the lead COMT inhibitors.
In order to assess the potency of the novel COMT inhibitors,
in-house overexpressed recombinant human soluble COMT
(S-COMT) protein (see Experimental Section) was used in
the inhibition assay. The assay method described previously16
was applied, in which the methylation of the substrate
aesculetin to scopoletin in a reaction mixture in the presence
of S-(5-adenosyl)-L-methionine (SAM) was measured fluoro-
metrically. Figure 2 shows the COMT inhibition curves of the
compounds 12 and 3. Tolcapone and entacapone were included
in the inhibition assay as reference.
As it is reported that tolcapone is more potent than
entacapone,3 the same results were observed in our experiment
against the human recombinant COMT, with IC50 values of
127 nM for tolcapone and 386 nM for entacapone, respectively.
Our lead compound 3 showed an activity similar to that of
tolcapone with an IC50 of 179 nM, and 12 with an IC50 of
54 nM in the assay.
Identification of On-target and Off-target Proteins of
COMT Inhibitors by CCMS. In a previous study, the applica-
tion of capture compound mass spectrometry (CCMS) to
differentiate the protein interaction profiles of tolcapone and
entacapone17 was reported. In that study, proteins were isolated
and identified by tolcapone- or entacapone-capture compounds
(CCs) as compared to those which were isolated by the CC
scaffold structure alone. The list of proteins identified in the
previous study especially for tolcapone did possibly not all
represent individual specific tolcapone binders but more likely
interactors due to general protein-binding properties of
tolcapone. In the absence of competition experiments using
the small drug molecules, this aspect could not be clarified
further. Among the sets of experiments possible to control for
the specific capture of a protein, preincubation of the biological
sample with the underivatized drug molecule to which the sel-
ectivity function of the CC corresponds is the most meaningful
specificity control. In the full CC, upon derivatization of the
drug molecule, the affinity, selectivity or other aspects of the
binding mode may differ from the parent drug molecule. If only
the CC scaffold structure is used in control experiments,
no distinction can be made between proteins that are bound
by the selectivity function of the full CC because of properties
that correspond to the free drug or the selectivity function in a
derivatized form. These potentially altered properties have no
effect on the assignment of a protein as specifically captured
if the original underivatized drug molecule was used in the
competition experiment. Therefore, this type of experiment is
the one that is truly informative on drug binding specificity.
Our current study addresses specific binders of tolcapone in a
refined experimental setup, using differential competition experi-
ments across a much larger set of replicate samples, dose-
dependent CCMS experiments, and improved analysis of mass
spectrometric data. We conducted our CCMS experiments with
whole cell lysates from the human hepatoma-derived cell line
HepG2. As reported earlier, catechol moieties are expected to
target the COMT binding pocket, whereas the opposite benzylic
side of tolcapone protrudes outside the protein.
Figure 1. Chemical structures of tolcapone, entacapone, compound 3,
and compound 12.
a
Scheme 1. Synthesis of Compound 3
a
Reagents and conditions: (a) ethyl bromide, DIPEA, DMF, 23 °C, 4
days, 89%; (b) piperidine, MeOH, 80 °C, 17 h, 66%.
The synthesis of the two novel inhibitors 3 and 12 are
depicted in Schemes 1 and 2, respectively. Compound 3 was
accessible by piperidine promoted condensation reaction of alkyl
thiazolidine-2,4-dione derivative 2b with commercially available
3,4-dihydroxy-5-nitrobenzaldehyde 1 (Scheme 1).11 Bicyclic
nitro-catechol 12 was synthesized in eight synthetic steps
starting from commercially available 3,4-dimethoxybenzoic acid
4 (Scheme 2). Compound 4 was converted into bis-alcohol 6
by subsequent cyclization reaction with formaldehyde under acid
conditions and reductive ring-opening.12 An alternative route
starting from 1,2-bis(chloromethyl)-4,5-dimethoxybenzene and
subsequent double substitution with acetic acid and saponifica-
tion seems to be more practicable but was also lower in yields.
Whereas Swern oxidation of bis-alcohol 6 gave bis-aldehyde 7
in good yields, manganese dioxide promoted oxidation mainly
yielded lactone 5. Bis-aldehyde 7 was then converted into tri-
cyclic compound 9 via tetrabromide 8 and cyclization reaction
with maleic anhydride.13 Anhydride 9 was ring-opened with
dimethyl amine to form a monoamide intermediate and screening
of potential coupling reagents indicated HATU to be superior in
yielding bisamide 10 at 62% over both synthetic steps. Boron
tribromide promoted double ether cleavage delivered 11,14 which
was converted into nitro-catechol 12 by selective introduction of
Different parts of a drug may produce different interactions
with cellular proteins inducing different responses. Therefore,
two opposite points of attachment for tolcapone as selectivity
function to the capture compound scaffold were used in this
B
DOI: 10.1021/acs.jmedchem.5b01970
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
a
Scheme 2. Synthesis of Compound 12
a
Reagents and conditions: (a) formaldehyde, HCl (gas), 70 °C, 4 days, 93%; (b) LiAlH4, THF, 66 °C, 4 h, 77%; (c) oxalyl chloride, DMSO, TEA,
DCM, −78 °C, 105 min →23 °C, 60 min, 75%; (d)PBr5, benzene, 23 °C, 17 h, 41%; (e) maleic anhydride, NaI, DMF, 70 °C, 48 h, 69%; (f) diethyl
amine, HATU, DIPEA, DMF, 23 °C, 48 h, 62%; (g) BBr3, DCM, 0 °C, 20 min → 23 °C, 60 min, 98%; (h) 2,3,5,6-tetrabromo-4-methyl-4-
nitrocyclohexa-2,5-dienone, diethyl ether, 23 °C, 60 min, 55%.
To test the activities of CCs, the COMT inhibition assay
against the in-house expressed recombinant human COMT
protein was performed. As expected, the inactive-Tcp-CC (14)
in which tolcapone was linked via the catechol moiety had no
affinity to COMT (Figure 4). The active-Tcp-CC (13) attached
via the benzylic moiety of tolcapone, leaving the nitrocatechol
moiety free to interact with the target, displayed an IC50 of
648 nM to the recombinant human COMT, corresponding to
an ∼5-fold reduction of activity compared to that of the parent
molecule (IC50 of 127 nM).
We started our investigations of the protein interaction profile
of these small COMT inhibitors by performing the dCCMS
experiments in the HepG2 whole cell lysate. Enrichment and
isolation of specifically interacting proteins were carried out using
the tolcapone CC (compound 13 or 14) at a final concentration
of 1 μM. In the competition control experiments, the respective
free small COMT inhibitors (tolcapone, entacapone, 3 and 12
respectively) were added in excess to compete with the CC for
specific binding to COMT at 100 μM final concentration (see
Experimental Section). The captured and trypsinized proteins
were identified using a high resolution LTQ Orbitrap hybrid
■
Figure 2. COMT inhibition curves for tolcapone ( ), entacapone
(
◊
▽
○
), compound 12 ( ), and compound 3 ( ) using in-house
overexpressed recombinant human COMT protein. Data are the mean
SD (n = 3).
study: a COMT active tolcapone capture compound (active-
Tcp-CC, 13) and a COMT inactive tolcapone capture com-
pound (inactive-Tcp-CC, 14) (Figure 3). In the active orienta-
tion, the CC should be able to specifically interact with the
known target COMT; in the other orientation, there should
be no interaction with the target COMT, but there could still
be interactions with some off-target proteins.17
Figure 3. Chemical structures of tolcapone capture compounds in active (attachment via the catechol function, active-Tcp-CC, 13) and inactive-
Tcp-CC (attachment via benzylic group with free catechol function, inactive-Tcp-CC, 14).
C
DOI: 10.1021/acs.jmedchem.5b01970
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 4. COMT inhibition curves for the active-Tcp-CC (13) and
the inactive-Tcp-CC (14) using in-house overexpressed recombinant
human COMT protein. Data are the mean SD (n = 3).
mass spectrometer. Only proteins identified with more than 99%
probability and at least two unique peptides with 95% probability
were considered to be identified with sufficient confidence.
Profiling the parent compound, which displays toxic side
effects, by CCMS in a toxicity-relevant tissue could deliver
meaningful hypotheses on potential underlying unwanted
protein interactions of the compound. These interactions have
to be carefully tested in a meaningful biological assay for their
biological relevance. Once an interaction is validated to underlie
the toxicity phenomenon, a dCCMS study with a library of
analogues can be carried out to identify those compounds that
do not display this interaction. If no toxicity information is
available beforehand, each compound would need to be profiled
separately using individual compound-derived CCs to identify
the respective protein interaction profile, and testable hypotheses
have to be derived from these profiles.
As expected, COMT was not identified with the inactive-
Tcp-CC (14). However, three proteins fulfilled our significance
criteria for specific binding and competed by all four competitors.
These proteins are PDXK, AKR1B10, and NIPSNAP3A (Table 1).
When the active-Tcp-CC (13) was used, COMT was the most
robustly identified protein, along with three additional proteins
(CCBL2, ADI1, and GCDH; Table 1) that fulfilled the signifi-
cance criteria and competed by all four competitors (detailed
CCMS results for these specific proteins are in Supporting
Information Figure 1A and B)
We were particularly interested in proteins that showed
differential binding profiles across the competitors of the active-
Tcp-CC. In our experiments, the protein HIBCH showed a
unique differential pattern across the four competitors used.
HIBCH was specifically identified when tolcapone was used
as the competitor, while no competition was observed in the
presence of entacapone. Interestingly, this protein was specifi-
cally competed when compound 3 was used as the competitor,
and no competition was observed when compound 12 was used.
Figure 5A shows the capture pattern of normalized HIBCH
intensities and averaged MS/MS counts identified using the four
different COMT inhibitors, including the specificities of this
protein in the dCCMS experiments.
HIBCH is an enzyme primarily localized to the mitochon-
drial matrix and was first purified from rat liver.18 It reacts with
[S]-hydroxyisobutyryl-CoA and 3-hydroxypropionyl-CoA with
almost no reactivity toward other acyl-CoA and is expressed
mostly in the liver but also in the kidney, heart, muscle, and
brain.19,20 HIBCH deficiency is an autosomal recessive disorder
which has only been reported in four cases, the first of which was
described in 1982.21 Despite the high clinical relevance, biochemical
profiles in these patients have not yet been fully elucidated.22
The HIBCH enzyme is crucial to prevent the accumula-
tion of toxic methacrylyl-CoA by removing CoA from
D
DOI: 10.1021/acs.jmedchem.5b01970
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 5. (A) CCMS results for HIBCH protein using active tolcapone capture compound (active-Tcp-CC, 13). Left panel: average MS/MS count
of the HIBCH protein identified in dCCMS experiments. Right panel: normalized intensities of the HIBCH protein in the dCCMS experiments
identified by nanoLC-MS/MS. A, assay; C1, tolcapone as the competitor; C2, entacapone as the competitor; C3, compound 12 as the competitor;
and C4, compound 3 as the competitor. The table gives the FC and p-value of the HIBCH protein in the capturing experiments using active-Tcp-CC
13 with different competitors (C1, tolcapone; C2, entacapone; C3, compound 12; and C4, compound 3). Data are the mean SD (n = 4). Gene
name, gene symbol for the protein identified; Uniprot ID, Uniprot protein database aceession no.; full name, protein name; MW [kDA], molecular
weight of the protein; assay MS/MS, average number of MS/MS counts (spectrum−peptide matches) for the respective protein in capture assays;
Fc, intensity fold change of cumulated normalized intensity of the respective protein between capture assays and competition samples; and p-value,
p-value from two-sided Student’s t test for the significance of the fold change. (B) Dose-dependent capture compound mass spectrometry (CCMS)
experiment results using active-Tcp-CC (13) or inactive-Tcp-CC (14) at different concentrations from 5 μM to 40 nM. Left panel: CCMS results
for the COMT protein. Right panel: CCMS results for the HIBCH protein. Assay: capturing experiments using compound 13 or 14. C1/Tcp:
tolcapone as competitor in the presence of active-Tcp-CC (13) or inactive-Tcp-CC (14). C3/12: compound 12 as competitor in the presence of
active-Tcp-CC (13) or inactive-Tcp-CC (14). Data are the mean SD (n = 4).
3-hydroxyisobutyryl-CoA in the valine catabolic pathway.
Reduced enzyme activity in human livers has been correlated
to liver cirrhosis and hepatocellular carcinoma, suggesting that a
decrease in the capability to dispose methacrylyl-CoA could lead
to liver failure.23,24 Therefore, the identification of this protein
as a tolcapone-specific off-target provides a novel possible
explanation for tolcapone-induced toxic effects in liver.
Verification of Small Molecule Binding to HIBCH. In
order to further investigate the binding of tolcapone to HIBCH,
dose-dependent capturing using CCMS was performed. The
results from dose-dependent capture experiments using capture
compounds (13 and 14) at four different concentrations
(40 nM, 200 nM, 1 μM, and 5 μM), using tolcapone or 12 as
competitor at 100 μM are shown in Figure 5B. As expected, the
COMT protein was captured only with CC using the active-
Tcp-CC (13) except for a weak signal detected at the highest
concentration of the inactive Tcp-CC (14). The same intensity
pattern was observed for HIBCH protein in terms of the
different orientations in which tolcapone was attached to the CC
scaffold. HIBCH was not identified with the CC with tolcapone
attached to the CC-scaffold in the inactive orientation (14).
Moreover, the HIBCH interaction with capture compound 13
that carried tolcapone attached to the CC-scaffold in the active
orientation was competed only by tolcapone but not by the
small molecule 12. The pronounced dose dependency and
the consistent competition pattern in this experiment further
E
DOI: 10.1021/acs.jmedchem.5b01970
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 6. Active-Tcp-CC (13) captures recombinant HIBCH (left) and recombinant COMT (right). Anti-HIBCH Western blot demonstrates the
specific and direct capturing of recombinant HIBCH protein (A), and the streptavidin Western blot demonstrates the cross-linking of HIBCH by 13
(B). The signal of captured HIBCH protein was largely reduced in the presence of tolcapone and compound 3, whereas little reduction of the
HIBCH protein signal was observed in the presence of entacapone and compound 12. Anti-COMT Western blot demonstrates the specific and
direct capturing of recombinant COMT protein (C), and the streptavidin Western blot demonstrates the cross-linking of COMT by compound 12
(D). The reaction was competed in the presence of tolcapone, entacapone, 12, and 3.
supports the conclusion that tolcapone but not compound 12
specifically binds to HIBCH.
However, only tolcapone and compound 3 could penetrate
deep into the HIBCH binding pocket with the catechol inter-
acting with the catalytic Glu121 side chain, while entacapone
and 12 did not, even if the docking was restrained with respect
to the position of the catechol moiety (Figure 7).
The binding pocket is placed on the top of the Rossman-
fold like central beta-sheet (Gly145 and Gly146), surrounded
by two distorted helical walls (residues 168−172 and 96−118,
respectively), and covered by a helical flap (Val349 and Leu350).
The cocrystallized product 3-hydroxy-2-methylpropanoate
(3-HYDROXYISOBUTYRATE, HIU) is located on the bottom
of the binding pocket close to Glu121, while quercetin (QUE)
closes the gate. With the QUE plug, the local pH close to the
Glu121 side chain seems to be very low: both carboxylates, that
of Glu121 and that of HIU, have to be protonated to form the
shown geometry by interacting via two H-Bonds.
Similar to our compounds, QUE also contains a catechol
moiety. However, the binding to the protein is mainly mediated
by interactions with the bicyclic core. Further, a position similar
to the QUE catechol for the nitrocatechol moieties in our
compounds would not be in line with our SAR, which suggests
a free catechol and a modifiable “tail”. Thus, we hypothesized
that the nitrocatechol moiety in our compounds might bind to
the HIU rather than the QUE pocket.
The alignment of cocrystallized QUE and HIU and docked
tolcapone (Figure 7) supports this hypothesis: one of the
nitro and one of the catechol oxygens of tolcapone occupy
the positions of the HIU carboxylate close to Glu121, while the
second catechol oxygen lies close to the water molecule in the
X-ray structure that holds the loop formed by Ala96, Gly97,
and Gly98 in place. The tolyl moiety of tolcapone, however,
occupies the aromatic QUE binding pocket formed by Pro168,
Glu169, Ile172, and Leu174 in one of the walls and Val349 as
well as Leu350 in the helical flap. The same topology is found
for 3, while entacapone and 12 can be placed into the binding
pocket but cannot form the aromatic interaction with the QUE
binding pocket. Additionally, 12 and entacapone are missing
A unique feature of CCs is the formation of a covalent bond
between the CC and the target protein, and the engagement of
the target protein was competed in the presence of the small
molecules which bind to the same binding pocket of the target
proteins. We carried out capture experiments with recombinant
human HIBCH and COMT proteins, followed by anti-HIBCH
or anti-COMT, and streptavidin Western Blots to investigate
the binding profiles of the different competitors toward these
proteins by a mass spectrometry-independent readout method.
Figure 6A shows that HIBCH was captured by 13, and
capturing was largely reduced when tolcapone or 3 were used
as competitor, but not when entacapone or 12 were used as
competitors. In addition, the direct cross-link of active-Tcp-CC
13 to HIBCH was confirmed by streptavidin Western blot
(Figure 6B) that detects biotinylation of the target protein after
the cross-link of the CC. For comparison, the same type of
experiment was performed with human recombinant COMT
using an anti-COMT Western blot (Figure 6C) as readout.
Anti-COMT immunoreactivity was observed in the capture
assay sample, but the immunoreactivity signal was abolished in
the presence of any of the four small molecules tolcapone,
entacapone, 12, and 3. Consistent with this, the streptavidin-
reactive band in the Western blot corresponding to biotinylated
COMT was abolished in the presence of the four competitors
(Figure 6D). This observation correlates very well with the mass
spectrometry data (Figure 5A and B), which confirms that
tolcapone, entacapone, 12, and 3 are all potent COMT inhibi-
tors and suggests that the binding of 3 to HIBCH might take
place in the same binding pocket as the binding of tolcapone to
HIBCH.
To verify this hypothesis, docking experiments for all four
small molecules to COMT and HIBCH were conducted.
From this, it was obvious that the two novel COMT inhibitors,
compound 3 and 12, fit very well into the catechol binding
pocket of COMT (Supporting Information, Figure 4).
F
DOI: 10.1021/acs.jmedchem.5b01970
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 7. Overview (top panel) of the crystal structure HIBCH_HUMAN (pdb code 3BPT). Binding pocket of the crystal structure
HIBCH_HUMAN (pdb code 3BPT) with cocrystallzed ligands (yellow) and docked tolcapone (middle left, magenta), entacapone (middle right,
cyan), compound 3 (down left, orange), and compound 12 (down right, gray). The protein is shown in greenblue, and residues within a radius of
3 Å around the docked molecules are represented as capped sticks. Positions of ligand oxygen atoms important for protein binding are marked with
asterisks.
some crucial interactions in the HIU pocket. The catechol
moiety of 12 has a different orientation and that of entacapone
a slightly different position, so the full interactions with the
Glu121 chain and the Ala96, Gly97, and Gly98 loop cannot be
formed.
So, a possible explanation for our experimental findings
might be that while tolcapone and 3 seem to be able to enter
the binding pocket in a favorable fashion and additionally to
close the pocket in a way similar to QUE, the two other ligands
may find a way inside the binding pocket but hardly find good
interaction points. Moreover, with the aromatic lock missing,
they will be expelled from the binding pocket as the formed
HIU would be after catalysis in the absence of QUE.
Cytotoxicity Results of the COMT Inhibitors. Several
previous studies suggested that toxic side effects of tolcapone in
contrast to entacapone might be caused by the impairment of
mitochondrial function,25,26 and a pronounced general inter-
action of tolcapone with mitochondrial proteins was reported
previously.17 Nevertheless, the direct molecular target candidate
for tolcapone that may underlie this toxicity has not been reported.
Our study reports, for the first time, a distinct mitochondrial off-
target protein that binds to tolcapone, but not to entacapone,
and the binding is structurally rationalized. In order to investigate
a possible link between the differential binding profiles of the
COMT inhibitors (tolcapone, entacapone, 12 and 3) tested in
this study to possible mitochondrial toxicity, two assays were
performed: the Mitochondria ToxGlo assay and TMRE mito-
chondria membrane potential assay.
For the Mitochondria ToxGlo assay, HepG2 cells are grown
in galactose-containing or glucose-containing media and treated
with test compounds of interest. Mitochondrial toxicity is then
detected on the basis of measuring the ATP content of the
cells. In contrast to cultivation in glucose-containing medium,
ATP generation in cells cultivated in galactose-containing
medium largely relies on oxidative phosphorylation. If a
decrease in ATP content is only observed in cells that grow
in galactose-containing medium, but not in cells that grow in
glucose-containing medium, an inhibition of ATP generation
can be attributed to a mitochondrial dysfunction, indicating
mitotoxicity.27 Besides tolcapone and entacapone, our two
G
DOI: 10.1021/acs.jmedchem.5b01970
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 8. Dose−responses curves at 24 h for high-glucose-grown (22 mM) and galactose-grown (10 mM) HepG2 cells treated with tolcapone,
entacapone, compound 12, and compound 3. ATP levels were measured using the Mitochondria ToxGlo assay. Data are the mean SD (n = 6).
novel small molecule COMT inhibitors, compounds 3 and 12,
were included in the analysis. As reported above, they are
potent COMT binders but differ in terms of their off-target
profiles. The off-target profile of 3 resembled the off-target
profile of tolcapone, and the off-target profile of 12 resembled
the off-target profile of entacapone.
The results from the Mitochondrial ToxGlo assay (Figure 8)
show for tolcapone a more than 4-fold lower IC50 value in the
galactose media compared to that in the glucose media. This
observation indicates that tolcapone is a mitochondrial toxicant.
In contrast, a much smaller effect was observed for entacapone.
Interestingly, when we compared the mitochondrial toxicity
of compounds 3 and 12, the fold changes of IC50 values for
the inhibition of ATP generation in cells grown in galactose-
containing medium compared to cells grown in glucose-
containing medium for 12 are much less than that for 3. This
comparison indicates that the mitochondrial toxicity detected in
this assay is substantially less with compound 12 treatment than
treatment with compound 3.
Figure 9. Dose-dependent effect of tolcapone, entacapone, compound
3, or compound 12 on mitochondrial membrane potential (percentage
of control) in HepG2 cells using the TMRE mitochondria membrane
potential test. Data are the mean SD (n = 6).
For the TMRE mitochondria membrane potential assay,
HepG2 cells were treated with four COMT inhibitors at different
concentrations. TMRE (tetramethylrhodamine ethyl ester), a cell
permeable, positively charged, red-orange dye that readily
accumulates in active mitochondria due to their relative negative
charge, was used to label active mitochondria. Significant dif-
ference in mitochondrial membrane potential status between the
four COMT inhibitors was observed (Figure 9). At a concentra-
tion of 50 μM, the compounds tolcapone and 3 suppressed the
membrane potential to less than 20−30% of untreated HepG2
cells, while more than 50% of the membrane potential was still
observed using the compounds entacapone or 12. Compound 3
and tolcapone at a concentration of 100 μM, reduced the mem-
brane potential to 0−5% of the control, while for compound 12
and entacapone, 40−50% of the membrane potential remained
at this concentration. The data demonstrate a close resemblance
between entacapone and compound 12, as well as between
tolcapone and compound 3 in terms of the mitochondria
membrane potential at concentrations between 10 and 100 μM.
In previous studies, the toxicity of tolcapone was confirmed
in a whole cell based system, isolated mitochondria, rodent
models and clinical practice,25,26,28,29 whereas entacapone
did not demonstrate such toxicity at comparable concen-
trations. In this cytotoxicity assay, we demonstrate that our lead
H
DOI: 10.1021/acs.jmedchem.5b01970
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
5,6-Dimethoxyisobenzofuran-1(3H)-one (5). At 0 °C, HCl gas was
bubbled through a solution of formaldehyde (37% in water; 90 mL;
1.2 mol) for 30 min. 3,4-Dimethoxybenzoic acid 4 (10 g; 54.9 mmol)
was added, and the reaction mixture was heated to 70 °C, while HCl
gas was continued to bubble through the suspension. After 3 days, all
volatiles were removed under reduced pressure, and water (200 mL)
was added. The pH was adjusted to pH 7 by the addition of aqueous
NH3 solution. The solid product was separated by filtration and dried
under reduced pressure to yield lactone 5 (9.88 g; 50.9 mmol; 93%) as
an off-white solid. 1H NMR (d-DMSO): δ 7.26 (s, 1H), 7.23 (s, 1H),
5.27 (s, 2H), 3.87 (s, 3H), 3.84 (s, 3H).
compound 12 and entacapone are both less toxic than
compound 3 and tolcapone.
CONCLUSIONS
■
In summary, using the newly developed differential competition
CCMS method (dCCMS), dose-dependent CCMS experiments,
improved mass spectrometric measurement, and data analysis
allowed us to differentiate potential off-target toxicity liabilities
of COMT inhibitors and to identify HIBCH as the potential
off-target of tolcapone that causes mitotoxicity. Furthermore,
compound N2,N2,N3,N3-tetraethyl-6,7-dihydroxy5-nitronaphthalene-
2,3-dicarboxamide (12) was identified as a novel potent COMT
inhibitor that lacks this unwanted interaction. Our findings
clearly demonstrate how dCCMS can be employed not only
to identify a drug’s interactors and elucidate its mechanism of
action but also to find out how these results can be used to tune
a drug’s binding properties in a desired direction.
4,5-Dimethoxy-1,2-phenylene)dimethanol (6). A solution of
lactone 5 (7.0 g; 36.0 mmol) in dry THF (170 mL) was added
dropwise to a suspension of lithium aluminum hydride (3.67 g;
108 mmol) in dry THF (50 mL). The reaction mixture was stirred for
4 h at 66 °C. After cooling to 0 °C, the reaction was carefully
quenched by subsequent addition of water (7 mL) and 5 M NaOH
solution (7 mL). The solid was discarded after filtration, and all
solvents were removed under reduced pressure. Water was added, and
the product was extensively extracted with DCM. The combined
organic layers were dried over Mg2SO4, and DCM was removed under
reduced pressure to give alcohol 6 (5.5 g; 27.7 mmol; 77%). 1H NMR
(d-DMSO): δ 6.98 (s, 2H), 4.99 (br, s, 2H), 4.47 (s, 4H), 3.74 (s, 6H).
4,5-Dimethoxyphthalaldehyde (7). At −78 °C, a solution of
alcohol 6 (2.2 g; 11.1 mmol) in dry DCM (25 mL) was added to
a prestirred solution of oxalyl chloride (1.94 mL; 22.2 mmol) and
DMSO (3.47 mL; 48.8 mmol) in dry DCM (55 mL) and stirred for
45 min. Triethylamine (21.84 mL; 155 mmol) was added dropwise,
and the reaction mixture was stirred for an additional 60 min at −78 °C
and was then allowed to warm to 23 °C and stirred for an additional
60 min. The reaction was carefully quenched by the addition of water
(200 mL), and the product was extracted with DCM. The combined
organic layers were washed with ammonium chloride solution and
dried over Mg2SO4, and the solvent was removed under reduced
EXPERIMENTAL SECTION
■
Chemical Synthesis. General Procedures. Unless otherwise
noted, all reactions were performed in dried glassware under argon.
Commercially available high grade reagents and solvents were used as
received without further purification.
Column chromatography was carried out on a “CombiFlash”
Companion system from Teledyne Isco., using prepacked “GraceResolv
Silica” columns as stationary phase. MPLC purification was performed
on a Buechi system (Buechi control unit C-620; two Buechi pump
modules C-605; column, Omnifit (230 × 15 mm); stationary phase,
LiChroprep RP-select B 25−40 μm; Buechi UV photometer C-635;
Buechi fraction collector C-660; mobile phase A, 0,1% AcOH Millipore
grade; mobile phase B, acetonitrile prepsolv grade). LC-MS was carried
out on a Dionex Ultimate 3000 RS HPLC coupled to a Bruker amaZon
Speed ion trap via an ESI source. LC method: compounds were diluted
to ∼5 μM in 20% MeCN, of which 50 μL was injected. Samples were
separated on a Phenomenex Luna PFP column (50 × 2 mm, 3 μm,
100 Å column, with a linear gradient of 10% to 100% B in 5.5 min).
Mobile phase A: 0,1% HCOOH Millipore grade; mobile phase B, aceto-
nitrile LCMS grade. Proton nuclear magnetic resonance (1H NMR)
spectra and carbon nuclear magnetic resonance (13C NMR) spectra
were recorded on an Avance 400 (400 MHz) from Bruker. Chemical
shifts are reported in ppm relative to d-DMSO (δ 2.50) and CDCl3
1
pressure to yield aldehyde 7 (1.62 g; 8.36 mmol; 75%). H NMR
(d-DMSO): δ 10.52 (s, 2H), 7.54 (s, 2H), 3.94 (s, 6H).
1,2-Bis(dibromomethyl)4,5-dimethoxybenzene (8). A solution of
aldehyde 7 (1.6 g; 8.24 mmol) in dry benzene (55 mL) was treated
with pentabromophosphorane (8.87 g; 20.6 mmol) and stirred for
17 h at 23 °C. The reaction was carefully quenched by the addition of
1% KOH solution (50 mL). The layers were separated, and the water
layer was extracted twice with DCM. The combined organic layers
were washed with brine and dried over Mg2SO4, and all solvents were
removed under reduced pressure. The crude product was purified by
CombiFlash (cyclohexane/EtOAc, 9/1 to 1/1) to obtain bromide 8
1
1
(δ 7.27) for H NMR. Both H NMR and LC-MS were performed
for all synthesized compounds to confirm purity. Purity of all final
compounds was 95% or higher.
1
(1.61 g; 3.34 mmol; 41%). H NMR (CDCl3): δ 7.10 (br, s, 4H),
3.93 (s, 6H).
6,7-Dimethoxynaphtho[2,3-c]furan-1,3-dione (9). A solution of
bromide 8 (1.6 g; 3.32 mmol) and maleic anhydride (0.78 g;
7.97 mmol) in dry DMF (20 mL) was treated with sodium iodide
(2.34 g; 15.61 mmol) and stirred for 48 h at 70 °C. The reaction was
quenched by the addition of water (100 mL) and saturated Na2SO3
solution (20 mL). Solid product 9 (591 mg; 2.29 mmol; 69%) was
3-Ethylthiazolidine-2,4-dione (2b). A solution of thiazolidine-2,4-
dione 2a (500 mg; 4.27 mmol) and ethyl bromide (465 mg;
4.27 mmol) in dry DMF (11.4 mL) was treated with DIPEA (2.23 mL;
12.81 mmol) and stirred for 17 h at 23 °C. Since TLC indicated
incomplete conversion, another batch of ethyl bromide (465 mg;
4.27 mmol) was added, and the reaction mixture was stirred for
an additional 3 days at 23 °C. All volatiles were removed under
reduced pressure, and the crude product was purified by CombiFlash
(cyclohexane/EtOAc, 99/1 to 3/1) to obtain compound 2b (550 mg;
1
obtained after filtration and drying under reduced pressure. H NMR
(d-DMSO): δ 8.52 (s, 2H), 7.76 (s, 2H), 3.97 (s, 6H).
N2,N2,N3,N3-Tetraethyl-6,7-dimethoxynaphthalene-2,3-dicarbox-
amide (10). A solution of tricyclic compound 9 (171 mg;
0.662 mmol) in dry DMF (200 μL) was treated with diethyl amine
(152 μL; 1.46 mmol) and stirred for 1 h at 23 °C. HATU (302 mg;
0.80 mmol) and DIPEA (231 μL; 1.32 mmol) were added, and stirring
was continued for 48 h at 23 °C. All volatiles were removed under
reduced pressure, and the crude product was purified by MPLC
(water/acetonitrile, 4/1 to 1/1) to obtain bis-amide 10 (158 mg;
1
3.79 mmol; 89%) as a colorless liquid. H NMR (CDCl3): δ 3.94
(s, 2H), 3.69 (q, J = 7.1 Hz, 2H), 1.20 (t, J = 7.1 Hz, 3H).
5-(3,4-Dihydroxy-5-nitrobenzylidene)3-ethylthiazolidine-2,4-
dione (3). A solution of 3,4-dihydroxy-5-nitrobenzaldehyde 1 (185 mg;
1.01 mmol) and compound 2b (148 mg; 1.02 mmol) in dry MeOH
(2.5 mL) was treated with piperidine (155 μL; 1.57 mmol) and stirred
for 17 h at 80 °C. The reaction mixture was diluted with ethyl acetate
and washed twice with 0.5 M HCl solution. The organic layer was dried
over Mg2SO4, and all solvents were removed under reduced pressure.
The crude product was recrystallized from MeOH to obtain compound
1
0.409 mmol; 62%). H NMR (CDCl3): δ 7.59 (s, 2H), 7.10 (s, 2H),
4.01 (s, 6H), 3.52 (q, J = 7.1 Hz, 4H), 3.31 (q, J = 7.1 Hz, 4H), 1.24
(t, J = 7.1 Hz, 6H), 1.10 (t, J = 7.1 Hz, 6H).
1
N2,N2,N3,N3-Tetraethyl-6,7-dihydroxynaphthalene-2,3-dicarboxa-
mide (11). At 0 °C, a solution of bis-amide 10 (158 mg; 0.409 mmol)
in dry DCM (16 mL) was treated with BBr3 solution (1 M in DCM;
2.04 mL; 2.04 mmol). The reaction mixture was stirred for 20 min at
3 (207 mg; 0.67 mmol; 66%) as a yellow solid. H NMR (CDCl3):
δ 10.85 (s, 1H), 7.84 (s, 1H), 7.76 (s, 1H), 7.38 (s, 1H), 6.09 (s, 1H),
3.83 (q, J = 7.1 Hz, 2H), 1.28 (t, J = 7.1 Hz, 3H). LC-MS: rt =
3.27 min; 308.83 [M − H]−; 640.92 [2M + Na-2H]−.
I
DOI: 10.1021/acs.jmedchem.5b01970
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
0 °C, was allowed to warm to 23 °C, and was stirred for an additional
60 min. All volatiles were removed by nitrogen stream, and the crude
product was purified by MPLC (water/acetonitrile, 4/1 to 1/9) to
obtain bis-amide 11 (143 mg; 0.399 mmol; 98%). 1H NMR
(d-DMSO): δ 7.49 (s, 2H), 7.17 (s, 2H), 3.15 (q, J = 6.9 Hz, 4H),
1.08 (t, J = 6.9 Hz, 6H), 1.03 (t, J = 6.9 Hz, 6H). 1H NMR (CDCl3):
δ 8.11 (br, s, 2H), 6.83 (s, 2H), 3.54 (q, J = 6.4 Hz, 4H), 3.27 (q,
J = 6.4 Hz, 4H), 1.25 (t, J = 6.4 Hz, 6H), 1.07 (t, J = 6.4 Hz, 6H).
N2,N2,N3,N3-Tetraethyl-6,7-dihydroxy5-nitronaphthalene-2,3-di-
carboxamide (12). A solution of bis-amide 11 (15.0 mg; 0.042 mmol)
in dry diethyl ether (2 mL) was subsequently treated three times with
solutions of 2,3,5,6-tetrabromo-4-methyl-4-nitrocyclohexa-2,5-dienone
(9.81 mg; 0.021 mmol) in dry diethyl ether (1 mL) and stirred for
20 min at 23 °C. Since TLC indicated full conversion, all volatiles were
removed by nitrogen stream, and the crude product was purified by
MPLC (water/acetonitrile, 9/1 to 3/1) to obtain nitro-catechol 12
microplate at 37 °C, and the reaction was initiated by the addition of
AdoMet, at 37 °C, for a final concentration of 10 μM. Fluorescence was
measured at the start of the reaction and at 60 min using a CARY
Eclipse fluorescence spectrophotometer. The excitation and emission
wavelengths were 355 and 460 nm, respectively. Enzymatic activity was
estimated by subtraction of fluorescence at the beginning of the assay
from the fluorescence at 60 min, and inhibition curves were generated
using GraphPad Prism.
Capture Experiments. Capture experiments were carried out with
500 μg of HepG2 lysate (InVivo, Germany) or 1 μg of human
recombinant COMT or 1 μg human recombinant HIBCH protein,
supplemented with 5 μM SAM (Sigma-Aldrich), 20 μL of 5× capture
buffer (caprotec bioanalytics GmbH) (100 mM HEPES, 250 mM
potassium acetate, 50 mM magnesium acetate, and 50% glycerol),
and incubated with DMSO or 50 μM competitor in a total volume of
100 μL at 4 °C for 30 min. After incubation with CC at 4 °C for 1 h,
samples were irradiated for 10 min using the caprobox (caprotec
bioanalytics GmbH, Berlin, Germany). Subsequently, buffer conditions
were adjusted by adding 25 μL of 5× wash buffer (caprotec
bioanalytics GmbH) (250 mM Tris-HCl, pH 7.5, 5 mM EDTA,
5 M NaCl, and 0.5% octyl-β-D-glucopyranoside) to each reaction. The
samples were then incubated with 50 μL of Dynabeads MyOne
Streptavidin C1, (Invitrogen, Karlsruhe, Germany) for 1 h at 4 °C on a
rotation wheel. The beads, containing the CC−protein conjugates,
were then collected using caproMag (caprotec bioanalytics GmbH,
Berlin, Germany), a device for the convenient and efficient isolation
of biomolecules using magnetic beads, and washed with 200 μL of
1× wash buffer, then with MS grade water. The beads were stored at
4 °C in deionized water until further analysis by either SDS−PAGE or
protein digestion followed by nanoLC-MS/MS (nanoliter liquid
chromatography tandem mass spectrometry). The Western blots were
carried out according to standard laboratory procedures. Antibodies
were diluted as follows: goat anti-COMT 1:500 (abcam, Cambridge,
UK, ab51984) and antigoat secondary antibody 1:2,000 (Sigma,
Steinheim, Germany, S5512) conjugated to a horseradish peroxidase
or streptavidin conjugated to horseradish peroxidase were used at a
dilution of 1:1,000 (Sigma, Steinheim, Germany, S5512).
Protein Digestion and Mass Spectrometric Analysis. After protein
capturing, streptavidin magnetic beads were washed twice with 200 μL
of LC-MS grade water (Fluka, St. Louis, MO, USA). For tryptic
digestion, the proteins bound to the streptavidin magnetic beads
were incubated with 9 μL of 50 mM ammonium bicarbonate and 1 μL
of trypsin (0.5 μg/μL) (Roche, Germany) for 16 h at 37 °C on a
temperature-controlled shaker. The supernatant was removed and
evaporated to dryness in a miVac DNA vacuum centrifuge (Genevac,
UK) and used directly for mass spectrometric analysis. The tryptic
peptides were analyzed by online nanoliter LC-MS/MS on an UltiMate
3000 RS LCnano System coupled to a LTQ-Orbitrap Velos instrument
(Thermo Fisher Scientific, Germany) through a Proxeon nano-
electrospray ion source (Thermo Fisher Scientific, Germany). For
chromatographic separation, samples were first loaded on a reversed
phase (RP) precolumn (Acclaim PepMap100, 0.1 × 20 mm) and
separated on a RP analytical column (Acclaim PepMap RSLC C18
0.075 × 150 mm, Thermo Fisher Scientific, Germany). Mass spectro-
metric analysis was performed in the data-dependent mode allowing
automatic switching between Orbitrap-MS and LTQ-MS/MS acquisi-
tion for a full scan with subsequent collision-induced dissociation
(CID) fragmentation. Full scan MS spectra (from m/z 300−2000) were
acquired in the Orbitrap analyzer. The 20 most intense ions were
sequentially isolated and fragmented in the linear ion trap using CID.
All MS/MS data were analyzed using search engine Andromeda
implemented in MaxQuant (www.maxquant.org, release 1.4.0.8).
Automated database searching against the human UniProtKB/
Swiss-Prot database (release 2015_02 contains contains 20203 reviewed
sequence entries) was performed with 6 ppm precursor tolerance,
0.5 Da fragment ion tolerance, and full trypsin specificity allowing
for up to 2 missed cleavages, and methionine oxidation was allowed as
variable modification. The maximum false discovery rates were set to
0.01 at both protein and peptide levels, and 6 amino acids were
required as minimum peptide length. The label-free quantification
1
(9.2 mg; 0.023 mmol; 55%). H NMR (CDCl3): δ 12.55 (br, s, 1H),
8.76 (s, 1H), 7.58 (s, 1H), 7.50 (s, 1H), 3.55−3.32 (m, 4H), 3.29
(q, J = 7.2 Hz, 4H), 1.24 (t, J = 6.9 Hz, 6H), 1.16 (t, J = 7.2 Hz, 3H),
1.12 (t, J = 7.2 Hz, 3H). LC-MS: rt = 2.66 min; 330.98
[M − N(CH2CH3)2]+; 404.07 [M + H]+; 807.15 [2M + H]+;
829.27 [2M + Na]+.
Overexpression of COMT and HIBCH. The COMT enzyme exists
in two forms: a soluble form (S-COMT) and a membrane bound form
(MB-COMT) and is ubiquitously expressed.3,30 For the purpose of
generating recombinant COMT enzyme, a His-S-COMT expression
vector was purchased from Origene (NM_007310). Human HIBCH
was cloned in-house. For the cloning of HIBCH, RNA was isolated
from HepG2 cells using RNeasy Mini Kit (Qiagen) from which cDNA
was synthesized using SuperScript III Reverse Transcriptase (Invitrogen).
Using the HepG2 cDNA as a template, PCR reactions were run with
the following primers: 5′-gatcgaattcatggggcagcgcgagatg-3′ (HIBCH
forward), 5′-gatcctcgagtcaaaatttcaaatcactgcttcccaaa-3′ (HIBCH reverse),
inserting EcoRI and XhoI restriction sites. A G3PDH PCR reaction
with the following primers 5′-tgaaggtcggagtcaacggatttggt-3′ (G3PDH
forward) and 5′-catgtgggccatgaggtccaccac-3′ (G3PDH reverse) were run
in parallel as control. PCR and restriction products were purified using
peqGOLD Gel Extraction Kit (peqlab). Restriction of PCR fragments
and target vector (pET28a+, Merck Millipore) was performed at 37 °C
for 2 h in the presence of FastDigest restriction enzymes and FastDigest
Green Buffer (Thermo Fisher Scientific). After 1 h 40 min, CIP enzyme
was added to the vector reaction which was incubated at 37 °C for
another 20 min. Ligation reactions were incubated at 16 °C overnight
using T4 DNA Ligase (Roche). Five hundred nanograms of DNA
construct with S-COMT and HIBCH was used for the transformation of
competent BL21 E. coli cells and expression of recombinant proteins
were induced by IPTG. The cells were collected by centrifugation
(15 min, 4000g, 4 °C). Cells were resuspended in 5 mL of buffer
(COMT, 50 mM phosphate, 300 mM NaCl, and 5 mM MgCl2,
pH 7.4; HIBCH, 20 mM Tris, 500 mM NaCl, 10 mM imidazole, and
0.5% (w/v) Triton X-100, pH 8.0) and disrupted by treatment with
lysozyme and DNase I followed by sonication. The supernatant from the
subsequent centrifugation (the cytosolic fraction) was used as the protein
source in subsequent experiments.
COMT Activity Microplate Assay. In addition to the COMT
enzyme and magnesium ions, the cofactor S-(5-adenosyl)-L-methionine
(AdoMet or SAM) is required for the methylation of catecholic
substrates to occur. COMT activity measurement experiments were
essentially performed as described previously.16 In short, aesculetin
(reaction substrate) and competing catechol substrates were dissolved
in dimethyl sulfoxide (DMSO) and diluted with aqueous buffer
solution (100 mM phosphate, 5 mM MgCl2, and 20 mM L-cysteine,
pH 7.4) for a final DMSO concentration of 2% in 100 μL of the
reaction mixture. All reagents were dissolved in the same buffer
solution. For each sample, inhibitor and aesculetin (Sigma-Aldrich)
solutions were pipetted in triplicate into a black round-bottomed
96-well plate. Controls without inhibitor or SAM were included in each
microplate. The plate was placed on ice, and the cytosolic fraction
containing the enzyme was added to a final protein concentration of
15 μg/mL. A preincubation period of 5 min was started by placing the
J
DOI: 10.1021/acs.jmedchem.5b01970
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
option was selected for alignment between nanoLC-MS/MS runs.
The output of the MaxQuant analysis was postprocessed according to
a standardized procedure implemented at caprotec bioanalytics to
normalize protein intensities and subsequently calculate fold changes
and p-values.
Therefore, proteins that are detected by nanoLC-MS/MS in the
capture assay and by comparison are significantly reduced in competi-
tion control experiments and are defined to be specific. The specificity
was quantified by the fold change between assay and competition for
each protein. Moreover, experiments were carried out in quadruplicate,
allowing the determination of significance by calculation of p-values.
ATPlite Cytotoxicity Assay. Mitochondrial toxicity assessment
(Glu/Gal) of the compounds tolcapone, entacapone, 12, and 3 was
used as the basis for a catechol fragment. The docking was repeated
using weakly restricted fragment placing (cpen = 10).
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the ACS
Publications website at DOI: 10.1021/acs.jmedchem.5b01970.
■
S
Additional figures illustrating the tryptic peptide
sequence identified from human HIBCH protein, details
of CCMS capturing results of other specific proteins, and
scatter plots of the intensities between the assay and
competitions (PDF)
conducted by Pharmacelcus GmbH (Saarbrucken). Briefly, HepG2
̈
SMILES data (CSV)
cells cultured in glucose media (high glucose DMEM containing
22 mM glucose, 1 mM sodium pyruvate, 5 mM Hepes, 2 mM
L-glutamine, 10% FCS, and 1% penicillin/streptomycin) or galactose
media (high glucose DMEM containing 10 mM galactose, 1 mM
sodium pyruvate, 5 mM Hepes, 2 mM L-glutamine, 10% FCS, and
1% penicillin/streptomycin) were added to a 96 well plate (100 μL,
4 × 104 cells/well). For drug treatments, compound stock solutions
were prepared in DMSO and added into the wells to give the indicated
final drug concentrations. The final DMSO concentration in each well
after the addition of luminescent ATP detection reagent was 0.5%.
A total of 100 μL of the luminescent ATPlite (PerkinElmer) detection
reagent was added to each well. After a brief mix on a plate shaker, the
reactions were incubated at room temperature for 2 min, and lumines-
cence was measured after 2 min using a Victor X5 plate reader.
TMRE Mitochondria Membrane Potential Test. The mitochondrial
membrane potential (MMP) was measured using the positively charged
fluorescent dye TMRE, which readily accumulates in active mito-
chondria. FCCP, an ionophore uncoupler of oxidative phosphorylation,
served as the positive control. HepG2 cells were seeded into μClear
bottomed, black 96-well plates (ca. 4 × 104 cells/well). After reaching
confluency, the cells were treated with four COMT inhibitors
(tolcapone, entacapone, 12, or 3) at individual defined final con-
centrations overnight at 37 °C and 5% CO2. The culture media were
removed and replaced with fresh media, TMRE was added to the media
at 200 nM final concentration, and cells were incubated for 30 min at
37 °C and 5% CO2. After incubation, cells were washed with PBS with
0.2% BSA, and the fluorescence intensity was measured using a Cary
Eclipse fluorescence microplate reader equipped with at 549 and 575 nm
excitation and emission filters, respectively.
Computational Docking Methods. All used software tools are in
included in the SYBYLx1.2 package (Tripos Inc. St. Louis, MO). For
the preparation of molecular structures, atomic coordinates of human
β-hydroxyisobutyryl-CoA hydrolase (HIBCH_HUMAN) in complex
with quercetin (QUE) and (2R)-3-hydroxy-2-methylpropanoic acid
(HIU) was obtained from the PDB file 3BPT. The structure was
preprocessed using the Prepare Protein tool and minimized using the
MMFF force field, MMFF charges, and the Powell gradient algorithm
(rmsd 0.5). All heavy atoms were frozen during minimization. QUE
was removed to get access to the catalytic site. Ligand structures were
obtained from 2d sketches (ChemDraw) using the Ligand Tool. The
structures were standardized, and tautomers were generated (count
of charged atoms = 4). Catechol dianions were removed with the
substructure SMARTS filter (O[-1]C:CO[-1]), and one 3D structure
per tautomer was generated using Concord Standalone. Atom types on
the catechols and nitro groups were set as follows: O− → O.co2; N →
N.pl3, using an SPL script. The resulting structures were minimized
using the default settings (Tripos FF, rmsd = 0.05) and Gasteiger-
Marsili charges. Unrestrained flexible docking between HIBCH and
the putative inhibitors tolcapone, entacapone, 12, and 3 was performed
with Surflex-Dock. The cocrystallized product HIU was extracted, and
the protomol was generated based on HIU. Protein flexibility
(hydrogens and heavy atoms) was allowed and 10 start conformations
generated. Otherwise, default settings were used. One hundred poses
were sampled and investigated with respect to the docking scores,
catechol position, and protein. The best docking pose of tolcapone was
AUTHOR INFORMATION
Corresponding Author
*Phone: +49 30 6392 3998. Fax: +49 30 6392 3985. E-mail:
yan.luo@caprotec.com.
■
Present Addresses
†L.v.K.: Freie Universitaet Berlin, Thielallee 63, 14195 Berlin,
Germany.
‡K.B.: Thermo Fisher Scientific GmbH, Im Steingrund 4-6,
63303 Dreieich, Germany.
§O.G.: Metabolomic Discoveries GmbH, Am Muehlenberg 11,
14476 Potsdam, Germany.
∥M.S.: Celares GmbH, Robert-Rossle-Straße 10, 13125 Berlin,
̈
Germany.
⊥F.K.: YARA International, Hanninghof 35, 48249 Duelmen,
Germany.
Notes
The authors declare no competing financial interest.
ABBREVIATIONS USED
■
CC, capture compound; CCMS, capture compound mass
spectrometry; dCCMS, differential competition capture com-
pound mass spectrometry; COMT, catechol-O-methyltransfer-
ase; DMEM, Dulbecco’s modified Eagle’s medium; DIPEA,
N,N-diisopropylethylamine; Ecp, entacapone; FCCP, carbonyl
cyanide 4-(trifluoromethoxy)-phenylhydrazone; FCS, fetal calf
serum; HATU, O-(7-azabenzotriazol-1-yl)-N,N,N′-tetramethy-
luronium hexafluoro-phosphate; HIBCH, 3-hydroxyisobutyryl-
CoA hydrolase; IPTG, isopropyl β-D-1-thiogalactopyranoside;
nanoLC-MS/MS, nanoliter liquid chromatography tandem
mass spectrometry; SAM, S-(5-adenosyl)-L-methionine; Tcp,
tolcapone; TEA, triethylamine; TMRE, tetramethylrhodamine,
ethyl ester
REFERENCES
■
(1) Bonifacio, M. J.; Palma, P. N.; Almeida, L.; Soares-da-Silva, P.
Catechol-O-methyltransferase and its inhibitors in Parkinson’s disease.
CNS Drug Rev. 2007, 13, 352−379.
(2) Antonini, A.; Abbruzzese, G.; Barone, P.; Bonuccelli, U.; Lopiano,
L.; Onofrj, M.; Zappia, M.; Quattrone, A. COMT inhibition with
tolcapone in the treatment algorithm of patients with Parkinson’s
disease (PD): relevance for motor and non-motor features. Neuro-
psychiatr. Dis. Treat. 2008, 4, 1−9.
(3) Kiss, L. E.; Soares-da-Silva, P. Medicinal chemistry of catechol O-
methyltransferase (COMT) inhibitors and their therapeutic utility. J.
Med. Chem. 2014, 57, 8692−8717.
(4) Spahr, L.; Rubbia-Brandt, L.; Burkhard, P. R.; Assal, F.;
Hadengue, A. Tolcapone-related fulminant hepatitis: electron
microscopy shows mitochondrial alterations. Dig. Dis. Sci. 2000, 45,
1881−1884.
K
DOI: 10.1021/acs.jmedchem.5b01970
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
(5) Acuna, G.; Foernzler, D.; Leong, D.; Rabbia, M.; Smit, R.;
Dorflinger, E.; Gasser, R.; Hoh, J.; Ott, J.; Borroni, E.; To, Z.;
Thompson, A.; Li, J.; Hashimoto, L.; Lindpaintner, K. Pharmacoge-
netic analysis of adverse drug effect reveals genetic variant for
susceptibility to liver toxicity. Pharmacogenomics J. 2002, 2, 327−334.
(6) Olanow, C. W.; Watkins, P. B. Tolcapone: an efficacy and safety
review. Clin. Neuropharmacol. 2007, 30, 287−294.
(7) Marsala, S. Z.; Gioulis, M.; Ceravolo, R.; Tinazzi, M. A systematic
review of catechol-0-methyltransferase inhibitors: efficacy and safety in
clinical practice. Clin. Neuropharmacol. 2012, 35, 185−190.
(8) Harrison, S. T.; Poslusney, M. S.; Mulhearn, J. J.; Zhao, Z.; Kett,
N. R.; Schubert, J. W.; Melamed, J. Y.; Allison, T. J.; Patel, S. B.;
Sanders, J. M.; Sharma, S.; Smith, R. F.; Hall, D. L.; Robinson, R. G.;
Sachs, N. A.; Hutson, P. H.; Wolkenberg, S. E.; Barrow, J. C. Synthesis
and evaluation of heterocyclic catechol mimics as inhibitors of
catechol-O-methyltransferase (COMT). ACS Med. Chem. Lett. 2015,
6, 318−323.
(9) Koster, H.; Little, D. P.; Luan, P.; Muller, R.; Siddiqi, S. M.;
Marappan, S.; Yip, P. Capture compound mass spectrometry: a
technology for the investigation of small molecule protein interactions.
Assay Drug Dev. Technol. 2007, 5, 381−390.
(10) Luo, Y.; Blex, C.; Baessler, O.; Glinski, M.; Dreger, M.; Sefkow,
M.; Koster, H. The cAMP capture compound mass spectrometry as a
novel tool for targeting cAMP-binding proteins: from protein kinase A
to potassium/sodium hyperpolarization-activated cyclic nucleotide-
gated channels. Mol. Cell. Proteomics 2009, 8, 2843−2856.
(11) Pratap, R.; Satyanarayana, M.; Nath, C.; Raghubir, R.; Puri, A.;
Chander, R.; Tiwari, P.; Tripathi, B.; Srivastava, A. Oxy Substituted
Chalcones as Antihyperglycemic and Antidyslipidemic Agents. US
Patent 20060142303A1, 2006.
(22) Yamada, K.; Aiba, K.; Kitaura, Y.; Kondo, Y.; Nomura, N.;
Nakamura, Y.; Fukushi, D.; Murayama, K.; Shimomura, Y.; Pitt, J.;
Yamaguchi, S.; Yokochi, K.; Wakamatsu, N. Clinical, biochemical and
metabolic characterisation of a mild form of human short-chain enoyl-
CoA hydratase deficiency: significance of increased N-acetyl-S-(2-
carboxypropyl)cysteine excretion. J. Med. Genet. 2015, 52, 691−698.
(23) Shimomura, Y.; Honda, T.; Goto, H.; Nonami, T.; Kurokawa,
T.; Nagasaki, M.; Murakami, T. Effects of liver failure on the enzymes
in the branched-chain amino acid catabolic pathway. Biochem. Biophys.
Res. Commun. 2004, 313, 381−385.
(24) Loupatty, F. J.; Clayton, P. T.; Ruiter, J. P.; Ofman, R.; Ijlst, L.;
Brown, G. K.; Thorburn, D. R.; Harris, R. A.; Duran, M.; Desousa, C.;
Krywawych, S.; Heales, S. J.; Wanders, R. J. Mutations in the gene
encoding 3-hydroxyisobutyryl-CoA hydrolase results in progressive
infantile neurodegeneration. Am. J. Hum. Genet. 2007, 80, 195−199.
(25) Haasio, K.; Koponen, A.; Penttila, K. E.; Nissinen, E. Effects of
entacapone and tolcapone on mitochondrial membrane potential. Eur.
J. Pharmacol. 2002, 453, 21−26.
(26) Haasio, K.; Nissinen, E.; Sopanen, L.; Heinonen, E. H. Different
toxicological profile of two COMT inhibitors in vivo: the role of
uncoupling effects. J. Neural. Transm. 2002, 109, 1391−1401.
(27) Marroquin, L. D.; Hynes, J.; Dykens, J. A.; Jamieson, J. D.; Will,
Y. Circumventing the crabtree effect: replacing media glucose with
galactose increases susceptibility of HepG2 cells to mitochondrial
toxicants. Toxicol. Sci. 2007, 97, 539−547.
(28) Korlipara, L. V.; Cooper, J. M.; Schapira, A. H. Differences in
toxicity of the catechol-O-methyl transferase inhibitors, tolcapone and
entacapone to cultured human neuroblastoma cells. Neuropharmacol-
ogy 2004, 46, 562−569.
(29) Nissinen, E.; Kaheinen, P.; Penttila, K. E.; Kaivola, J.; Linden, I.
B. Entacapone, a novel catechol-O-methyltransferase inhibitor for
Parkinson’s disease, does not impair mitochondrial energy production.
Eur. J. Pharmacol. 1997, 340, 287−294.
(30) Myohanen, T. T.; Mannisto, P. T. Distribution and functions of
catechol-O-methyltransferase proteins: do recent findings change the
picture? Int. Rev. Neurobiol. 2010, 95, 29−47.
(12) Meziane, M. A. A.; Royer, S.; Bazureau, J. P. A practical ’one-pot’
synthesis of ethyl isoquinoline-3-carboxylate by domino reactions: a
potential entry to constrained nonproteogenic amino acid derivatives.
Tetrahedron Lett. 2001, 42, 1017−1020.
(13) Mcomie, J. F. W.; Perry, D. H. Aromatic annelation using ortho-
quinodimethanes as reactive intermediates. Synthesis 1973, 1973, 416−
417.
(14) Naito, Y.; Sugiura, M.; Yamaura, Y.; Fukaya, C.; Yokoyama, K.;
Nakagawa, Y.; Ikeda, T.; Senda, M.; Fujita, T. Quantitative structure
activity relationship of catechol derivatives inhibiting 5-lipoxygenase.
Chem. Pharm. Bull. 1991, 39, 1736−1745.
(15) Lemaire, M.; Guy, A.; Roussel, J.; Guette, J. P. Nitro-
cyclohexadienones - a new class of nitrating agents. Tetrahedron
1987, 43, 835−844.
(16) Kurkela, M.; Siiskonen, A.; Finel, M.; Tammela, P.; Taskinen, J.;
Vuorela, P. Microplate screening assay to identify inhibitors of human
catechol-O-methyltransferase. Anal. Biochem. 2004, 331, 198−200.
(17) Fischer, J. J.; Michaelis, S.; Schrey, A. K.; Graebner, O. G.;
Glinski, M.; Dreger, M.; Kroll, F.; Koester, H. Capture compound
mass spectrometry sheds light on the molecular mechanisms of liver
toxicity of two Parkinson drugs. Toxicol. Sci. 2010, 113, 243−253.
(18) Shimomura, Y.; Murakami, T.; Fujitsuka, N.; Nakai, N.; Sato, Y.;
Sugiyama, S.; Shimomura, N.; Irwin, J.; Hawes, J. W.; Harris, R. A.
Purification and partial characterization of 3-hydroxyisobutyryl-
coenzyme A hydrolase of rat liver. J. Biol. Chem. 1994, 269, 14248−
14253.
(19) Wanders, R. J.; Duran, M.; Loupatty, F. J. Enzymology of the
branched-chain amino acid oxidation disorders: the valine pathway. J.
Inherited Metab. Dis. 2012, 35, 5−12.
(20) Hawes, J. W.; Jaskiewicz, J.; Shimomura, Y.; Huang, B.; Bunting,
J.; Harper, E. T.; Harris, R. A. Primary structure and tissue-specific
expression of human beta-hydroxyisobutyryl-coenzyme A hydrolase. J.
Biol. Chem. 1996, 271, 26430−26434.
(21) Brown, G. K.; Hunt, S. M.; Scholem, R.; Fowler, K.; Grimes, A.;
Mercer, J. F.; Truscott, R. M.; Cotton, R. G.; Rogers, J. G.; Danks, D.
M. Beta-hydroxyisobutyryl coenzyme A deacylase deficiency: a defect
in valine metabolism associated with physical malformations. Pediatrics
1982, 70, 532−538.
L
DOI: 10.1021/acs.jmedchem.5b01970
J. Med. Chem. XXXX, XXX, XXX−XXX