The relevance of Ki calculation for bi-substrate enzymes illustrated by kinetic evaluation of a novel lysine (K) acetyltransferase 8 inhibitor

Article abstract of DOI:10.1016/j.ejmech.2017.05.015

Histone acetyltransferases (HATs) are important mediators of epigenetic post-translational modifications of histones that play important roles in health and disease. A disturbance of these modifications can result in disease states, such as cancer or inflammatory diseases. Inhibitors of HATs (HATi) such as lysine (K) acetyltransferase 8 (KAT8), could be used to study the epigenetic processes in diseases related to these enzymes or to investigate HATs as therapeutic targets. However, the development of HATi is challenged by the difficulties in kinetic characterization of HAT enzymes and their inhibitors to enable calculation of a reproducible inhibitory potency. In this study, a fragment screening approach was used, enabling identification of 4-amino-1-naphthol, which potently inhibited KAT8. The inhibitor was investigated for enzyme inhibition using kinetic and calorimetric binding studies. This allowed for calculation of the K_i values for both the free enzyme as well as the acetylated intermediate. Importantly, it revealed a striking difference in binding affinity between the acetylated enzyme and the free enzyme, which could not be revealed by the IC₅₀ value. This shows that kinetic characterization of inhibitors and calculation of K_i values is crucial for determining the binding constants of HAT inhibitors. We anticipate that more comprehensive characterization of enzyme inhibition, as described here, is needed to advance the field of HAT inhibitors.

Full text of DOI:10.1016/j.ejmech.2017.05.015

Accepted Manuscript
The relevance of K calculation for bi-substrate enzymes illustrated by kinetic
i
evaluation of a novel lysine (K) acetyltransferase 8 inhibitor
Hannah Wapenaar, Thea van den Bosch, Niek G.J. Leus, Petra E. van der Wouden,
Nikolaos Eleftheriadis, Jos Hermans, Gebremedhin Solomon Hailu, Dante Rotili,
Antonello Mai, Alexander Dömling, Rainer Bischoff, Hidde J. Haisma, Frank J. Dekker
PII:
S0223-5234(17)30376-8
10.1016/j.ejmech.2017.05.015
EJMECH 9445
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 17 February 2017
Revised Date: 11 April 2017
Accepted Date: 4 May 2017
Please cite this article as: H. Wapenaar, T. van den Bosch, N.G.J. Leus, P.E. van der Wouden, N.
Eleftheriadis, J. Hermans, G.S. Hailu, D. Rotili, A. Mai, A. Dömling, R. Bischoff, H.J. Haisma, F.J.
Dekker, The relevance of K calculation for bi-substrate enzymes illustrated by kinetic evaluation of a
i
novel lysine (K) acetyltransferase 8 inhibitor, European Journal of Medicinal Chemistry (2017), doi:
10.1016/j.ejmech.2017.05.015.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT
The relevance of K calculation for bi-substrate enzymes illustrated
i
by kinetic evaluation of a novel lysine (K) acetyltransferase 8
inhibitor
a1
a1
a
Hannah Wapenaar , Thea van den Bosch , Niek G. J. Leus , Petra E. van der
a
a
b
c
Wouden , Nikolaos Eleftheriadis , Jos Hermans , Gebremedhin Solomon Hailu , Dante
c
c
d
b
a
Rotili , Antonello Mai , Alexander Dömling , Rainer Bischoff , Hidde J. Haisma , Frank
a *
J. Dekker
1
These authors contributed equally to this work
*
Corresponding author
Frank J. Dekker, Department of Chemical and Pharmaceutical Biology, Groningen Research
Institute of Pharmacy (GRIP), University of Groningen. Antonius Deusinglaan 1, 9713 AV
Groningen, The Netherlands. Email f.j.dekker@rug.nl, tel. +31 503638030.
Author affiliations:
a. Department of Chemical and Pharmaceutical Biology, Groningen Research Institute
of Pharmacy (GRIP), University of Groningen, Antonius Deusinglaan 1, 9713 AV
Groningen, The Netherlands.
b. Department of Analytical Biochemistry, Groningen Research Institute of Pharmacy
(
GRIP), University of Groningen, Antonius Deusinglaan 1, 9713 AV Groningen, The
Netherlands.
c. Department of Drug Chemistry and Technologies, Sapienza University of Rome, P.le
A. Moro 5, 00185 Rome, Italy
d. Department of Drug Design, Groningen Research Institute of Pharmacy (GRIP),
University of Groningen, Antonius Deusinglaan 1, 9713 AV Groningen, The
Netherlands.
1

ACCEPTED MANUSCRIPT
Abstract
Histone acetyltransferases (HATs) are important mediators of epigenetic post-
translational modifications of histones that play important roles in health and disease. A
disturbance of these modifications can result in disease states, such as cancer or
inflammatory diseases. Inhibitors of HATs (HATi) such as lysine (K) acetyltransferase 8
(
KAT8), could be used to study the epigenetic processes in diseases related to these
enzymes or to investigate HATs as therapeutic targets. However, the development of HATi is
challenged by the difficulties in kinetic characterization of HAT enzymes and their inhibitors
to enable calculation of a reproducible inhibitory potency. In this study, a fragment screening
approach was used, enabling identification of 4-amino-1-naphthol, which potently inhibited
KAT8. The inhibitor was investigated for enzyme inhibition using kinetic and calorimetric
binding studies. This allowed for calculation of the K values for both the free enzyme as well
i
as the acetylated intermediate. Importantly, it revealed a striking difference in binding affinity
between the acetylated enzyme and the free enzyme, which could not be revealed by the
IC value. This shows that kinetic characterization of inhibitors and calculation of K values is
50
i
crucial for determining the binding constants of HAT inhibitors. We anticipate that more
comprehensive characterization of enzyme inhibition, as described here, is needed to
advance the field of HAT inhibitors.
Keywords
Histone acetyltransferases, KAT8, fragment screening, histone acetylation, inhibitor, enzyme
kinetics.
Abbreviations
HAT, histone acetyltransferase; HATi, histone acetyltransferase inhibitor; KAT8, Lysine (K)
acetyltransferase 8; Ac-CoA, acetyl coenzyme A.
2

ACCEPTED MANUSCRIPT
Introduction
Epigenetics is a field of study in which novel therapeutic targets are found for various
diseases, such as inflammatory diseases and cancer (1). It is defined as the structural
adaptation of chromosomal regions so as to register, signal or perpetuate altered
transcriptional activity (2). Epigenetic processes include post-translational modifications of
histones such as lysine acetylations, which play an important role in the regulation of gene
transcription by controlling the chromatin structure of DNA (3). Lysine acetylations are
installed by histone acetyltransferases (HATs) and removed by histone deacetylases
(
HDACs). These enzymes balance lysine acetylation, resulting in a controlled expression of
genes. A disturbance of this balance can result in disease states, such as cancer or
inflammatory diseases (4). Therefore, restoring the balance between HAT and HDAC activity
using small molecule HAT inhibitors (HATi) could be a therapeutic strategy for several
diseases.
Development of HATi is an important challenge that has been addressed with limited
success so far. Important drawbacks of current HATi include intrinsic chemical reactivity,
instability, low potency, lack of selectivity as well as the limited kinetic characterization of the
inhibitors (5). Kinetic characterization of inhibitors in an early stage of lead identification is
essential. Since enzyme activity assays are often used for the discovery of novel inhibitors,
the 50% inhibitory concentration (IC ) is a common measure for the potency of inhibitors.
50
However, the IC depends on the conditions used in the assay, such as the concentration of
50
the enzyme substrates and their respective Km values. This value is therefore not
reproducible unless exactly the same assay conditions are used. The inhibitory potency (Ki)
value is a potency value independent of the assay conditions and is therefore much more
representative as potency of the inhibitor (6). Calculation of the K value from the IC is
i
50
crucial for the comparison of the potency with known inhibitors or between different assays
and for the determination of selectivity. In case of competitive inhibitors of enzymes
converting only one substrate, the K can be calculated using methods like the Cheng-Prusoff
i
equation, Dixon plot or using double reciprocal plots (7-9). However, since HATs use a
3

ACCEPTED MANUSCRIPT
cofactor, acetyl coenzyme A (Ac-CoA), for the acetylation of the lysine, they are bisubstrate
enzymes: they convert two substrates into two products. In this case, it is not possible to use
standard methods for calculation of the K and more elaborate kinetic evaluations are
i
necessary (10). It is therefore important to investigate the kinetic behavior of HAT inhibitors
toward determination of reproducible inhibitory constants.
The HATs are a disparate group of enzymes from which most isoenzymes can be
assigned to five main families based on primary structure homology. Three families that have
been studied extensively are the GNAT (GCN5-related N-acetyltransferase) family, the
p300/CBP (p300/CREB binding protein) family and the MYST (acronym for MOZ, Ybf2,
Sas2, and Tip60) family (11). In this study, we focused on Lysine (K) acetyltransferase 8
(
KAT8), a member of the MYST HAT family. KAT8 (also: males absent on the first, MOF, or
MYST 1) is part of the male-specific-lethal (MSL) complex, which specifically acetylates
histone H4 lysine 16 (12). KAT8 has additionally been shown to form different complexes
containing WD repeat domain 5 (WDR5, MSL1v1 or NSL complex) and MLL, broadening the
substrate specificity to histone H4 lysines 5 and 8 and non-histone targets, such as lysine
120 on the tumor-suppressor protein p53 (13-15). As part of these complexes, KAT8 has
been shown to play a role in stem cell pluripotency, cell proliferation and DNA damage
response (16). Using KAT8 conditional deletion and a small molecule inhibitor for MYST
family HATs in cell lines and a mouse model, it was recently shown to be important for
sustaining MLL-AF9-driven leukemia and was suggested as a potential therapeutic target for
MLL-rearranged leukemia (17). Therefore, small molecule KAT8 inhibitors could facilitate
investigation of its function in disease or may be used as potential therapeutic agents.
Currently, one class of HATi have been described to inhibit KAT8, which are
anacardic acid and a number of its derivatives (18). These inhibitors were shown to inhibit
KAT8 by interacting with an acetylated form of the enzyme, required the binding of Ac-CoA
and competed with the lysine substrate. Calculation of the K values revealed that although
i
the determined IC values were above 200 µM, the K values were in the range of 37 - 64
50
i
µM. This shows that the k value can differ significantly from tested IC values and that this is
i
50
4

ACCEPTED MANUSCRIPT
dependent on the mechanism of inhibition of the inhibitors. Therefore, we aimed at
discovering novel, structurally unrelated KAT8 inhibitors and determining their kinetic profile.
Using a fragment screening approach, 4-amino-1-naphthol (compound 13) was
identified as a potent KAT8 inhibitor. The mechanism of KAT8 inhibition and structure-activity
relationship (SAR) were investigated. Enzyme kinetic measurements as well as calorimetric
binding studies suggested a reversible inhibition mode and a direct interaction with KAT8 for
compound 13. Kinetic studies allowed calculation of the K value for both the free enzyme
i
form of KAT8 (K = 2.6 µM), and the acetylated form (K = 0.017 µM), which indicated very
i1
i2
high potency for the acetylated enzyme intermediate. Taken together, our approach to link
fragment screening with enzyme kinetic analysis demonstrated large affinity differences for
the different enzyme species involved in catalysis, which is not obvious from the IC values.
50
We anticipate that unravelling the inhibitory potencies of inhibitors for individual enzyme
species is key to inhibitor discovery for this type of enzymes.
Results and Discussion
Fragment screening
Towards discovery of a novel inhibitor of KAT8, an in-house library of fragments with
a broad range of structures and low molecular weight (MW < 250 Da), was screened for
inhibition of the KAT8 HAT. An assay based on fluorescence-detection of CoA was used to
screen all fragments. Currently known KAT8 inhibitors showed an IC of higher than 200 µM
50
under the same assay conditions (18). Taking this potency as a reference, a concentration of
00 µM was chosen for screening the fragments. A control was included for potential
2
fluorescence quenching by the fragments to identify and rule out any hits directly interfering
with the assay. Structures containing maleimide or thiol moieties were excluded due to
reactivity with the assay product or fluorophore. The resulting hits (1 and 10) and a small
number of similar fragments were tested for their 50% inhibitory concentration (IC ) (Table 1,
50
figure S1).
5

ACCEPTED MANUSCRIPT
4
- fluoro phenyl hydrazine (1) showed an IC of 310 µM. Several phenyl hydrazines
50
similar to 1 were investigated. 4- methoxy phenyl hydrazine (2) lost activity compared to 1.
Other substitution patterns like hydrogen or chlorine were not active (3-5). To investigate the
hydrazine moiety further, 4- fluoroaniline or benzylamine (6, 7) were tested. These were not
active, and neither were two other aniline compounds (8, 9), showing that the inhibition was
specific for the hydrazine moiety. Taken together, no phenyl hydrazine structures were found
to be more potent inhibitors than 1. Therefore, this structural entity was not further
investigated.
The hit 1-aminonaphthalene (10), showed an IC of 180 µM; the most potent hit from
50
the screening. Testing similar structures showed that 1-nitro naphthalene (11) or a sulphonic
acid substituted aminonaphthalene (12) were not active. The hydroxyl substituted fragment,
4-amino-1-naphthol (13), showed an excellent IC50 of 9.7 ± 3.0 µM. Strikingly, 9, which is
very similar to 13, was not active, showing that the naphthalene moiety is also important for
KAT8 activity. Therefore, compound 13 was investigated in more detail.
Structure-activity relationship
A SAR study was done around the structure of 13 to investigate the importance of the
different parts of the scaffold. Derivatives of 13 were synthesized and tested for their
inhibitory potency on KAT8 (Table 2, figure S2). The naphthalene ring was replaced by a
isoquinoline (14). This moiety has a slightly different logP and is weakly basic, but it did not
1
significantly influence the activity. To investigate the importance of the free amine on R
position, the amine was included in the ring (16) or a carbon or carboxyl spacer was
introduced (17, 18). These compounds did not show inhibitory activity on KAT8. Additionally,
the amine was replaced by a nitro (19). This compounds had a reduced potency compared to
13, but still showed activity. This suggests that the position of the amine is important, but it
can be slightly modified. To investigate an ortho substitution to the amine, a methoxy was
inserted (20). This reduced activity compared to 13, but was still reasonably active. Next, the
hydroxyl group was investigated. Both replacing it with a nitrile (15) as well as introducing a
6

ACCEPTED MANUSCRIPT
carbon spacer (21), completely abolished activity. When the hydroxyl moiety was substituted
with a methyl (22), it retained reasonable activity, but the acetyl and benzyl (23-24)
substituted compounds lost activity, suggesting that the free hydroxyl is important for activity
and there is little space for substituents here. Subsequently, substitutions to the amine were
investigated. The dimethylated amine (25) showed good potency, but the acetylated (26) and
propionylated (27) amine lost activity. Strikingly, substituting with a long aliphatic tail (28)
recovered activity again, which is probably due to lipophilic interactions. The preference of
KAT8 for lipophilic substitutions has been previously observed with the anacardic acid
derived inhibitors as well (18). A phenyl instead of a aliphatic tail (29) also showed better
activity than the propionyl, suggesting that a lipophilic or aromatic interaction is gained close
to the aminonaphthol scaffold. This was observed with two sulphonamide derivatives as well
(
30, 31). These derivatives showed good activity for KAT8 and the toluene sulphonyl
derivative was approximately twice as active as the methyl sulphonyl, suggesting a gained
lipophilic or aromatic interaction. However, none of the derivatives was significantly more
active than compound 13. Therefore, compound 13 was chosen for further investigation.
Inhibitor properties
An often encountered problem with high-throughput screening is the discovery of hits
that turn out to be pan assay-interfering compounds (PAINS) (19). Computational tools are
available to easily test structural entities for PAIN properties (20), but recent criticism warns
agains blind use of these tools, since it was observed that the tools will not recognize all
PAIN structures (21). It is therefore necessary to experimentally test the properties of the
high throughput hit to test for PAINS behavior. Therefore, compound 13 was investigated for
thiolreactivity, stability, reversibility, anti-oxidant properties and selectivity.
Compounds similar to compound 13 were discovered to be assay interfering
compounds (PAINS), seeming to inhibit the HAT Rtt109 in an assay based on detection of
the product CoA, but instead reacting with the thiol group of the product, which prevented
detection of CoA (22). The proposed mechanism was a Michael addition-elimination reaction
7

ACCEPTED MANUSCRIPT
in which the thiol, gluthathion, replaced a thiol or halogen substituent on position 2 from the
hydroxyl moiety. Therefore, the thiolreactivity of compound 13 was investigated in an assay
containing all elements of the IC assay, but where Ac-CoA was replaced with CoA (Figure
50
S3). No reduction in fluorescence was found in this assay, suggesting that compound 13
does not react with thiols under these conditions. Additionally, the stability of compound 13
was investigated using HPLC and no degradation of the compound was observed for at least
4
hours of incubation with the assay buffer (Figure S4). Longer incubation times were not
tested.
To investigate whether 13 shows irreversible inhibition, a pre-incubation assay was
performed (Figure 1A). Three different concentrations (0.75, 2 and 10 x the IC ) of
5
0
compound 13 were pre-incubated with KAT8 for 2, 5 or 10 minutes before adding the
enzyme substrates and measuring enzyme activity. An irreversible inhibitor will show time-
dependent inhibition and therefore give more inhibition when pre-incubated with the enzyme
for a longer time. However, no difference was observed between 2, 5 or 10 minutes pre-
incubation with any of the concentrations, indicating that 13 is a reversible inhibitor of KAT8.
This was confirmed using a dilution experiment (Figure 1A) in which KAT8 was pre-incubated
with 13 at a concentration of > 5 times the IC value and subsequently diluted 100 times in a
50
solution containing the substrates. The enzyme activity was measured over time and was
shown to recover linearly, indicating a fully reversible inhibitor (23). These data are
consistent with a model where 13 is a reversible inhibitor of KAT8.
Due to their phenolic structure, 13 and several other derivatives from the SAR have
anti-oxidant activity. Although KAT8 does not depend on redox cycling mechansims in its
enzymatic reaction, it may interfere with the assay, which could lead to observation of
inhibition, where no true inhibition of KAT8 takes place. Therefore, the anti-oxidant activity of
13 and all derivatives was determined in a DPPH (2,2-diphenyl-1-picrylhydrazyl) assay, a
widely used assay for determination of anti-oxidant activity of single molecules (24). The
concentration that gave 50% reduction in DPPH absorbance (EC ) was determined for all
50
compounds (Figure S5A). As expected, many derivatives containing the phenolic structure
8

ACCEPTED MANUSCRIPT
were anti-oxidants, showing very similar EC ’s between 7 and 26 µM. However, their anti-
50
oxidant activity did not show a relationship with their inhibitory activity on KAT8. For example,
many compounds showing anti-oxidant activity, did not show KAT8 inhibition (9, 23, 26, 27).
Compound 19 showed no anti-oxidant activity due to the strong deactivating properties of the
nitro substituent, but did show inhibitory activity on KAT8. Additionally, the IC of compound
50
13 was determined in the presence of non-thiol containing redox active substances (tris(2-
+
carboxyethyl)phosphine (TCEP, 1 µM), NAD (nicotinamide adeninedinucleotide, 100 µM)
and NADH (100 µM)) (Figure S5B). The presence of these redox active substances in the
assay did not significantly influence the IC of 13. Taken toghether, these data indicate that
50
the anti-oxidant activity of compound 13 does not influence the inhibition of KAT8.
The selectivity of 13 and the active derivatives was investigated using KAT3B and
KAT2B, representing two main other HAT families; p300/CBP and GNAT. An assay similar to
that of KAT8 based on fluorescence detection of CoA was set up for KAT2B and KAT3B and
the IC50 values of 13 and all derivatives showing activity on KAT8, were determined for
KAT2B and KAT3B (Table 2, Figure S6 and S7). Compound 13 showed inhibitory activity for
both KAT2B (IC = 3.6 ± 1.1) and KAT3B (IC = 1.4 ± 0.1). Although the derivatives showed
50
50
in general no selectivity over KAT2B and KAT3B, based on IC50 values, some differences
were observed. For example, compound 22 shows activity for KAT8 and KAT3B, but seems
inactive on KAT2B, suggesting that substitutions on the hydroxyl group are even more
constrained for this enzyme. Additionally, compound 19 shows moderate activity on KAT8
and is preferred above for example 20 and 28, but on KAT2B and KAT3B, it is less active
compared to these others. This suggests that the nitro group is not preferred on KAT2B and
KAT3B as on KAT8. Therefore, although 13 and derivatives are generally non-selective
between KAT8, KAT2B and KAT3B based on IC50 values, the different enzymes were
inhibited to different extents by the derivatives.
However, due to differences in catalytic mechansims, substrates and substrate
affinities of the different HATs, which all influence the IC values, for accurate comparison of
50
the potency for different enzymes, the K values should be calculated. For the calculation of
i
9

ACCEPTED MANUSCRIPT
the K values it is essential to determine the K values of the substrates for KAT2B and
i
m
KAT3B. Therefore, both KAT2B and KAT3B were kinetically investigated for their catalytic
mechanisms as initially done for KAT8 (Figure S8A/B). KAT2B showed sigmoidal kinetics,
suggesting a catalytic process with multiple steps of different velocities, for example
cooperative subunits or two preferred catalytic mechanisms which depend on the
concentration of Ac-CoA present (10) (Figure S8A). This deviation from Michaelis-Menten
kinetics is clearly observed when transforming the data to a Lineweaver-Burke plot. The
normally linear regression of the double reciprocal is not linear, but curves upward (Figure
S8A). In case of KAT3B, it was not possible to fully reach saturation of the velocity (Figure
S8B). The histone 3 peptide substrate (H3 substrate), for which p300 has affinity as well,
showed the same behavior. Unfortunately, since the determination of K values is based on
m
model enzymes that follow Michaelis-Menten kinetics, in both of these cases it is not
currently described how to determine K values of the substrates. Therefore, calculation of
m
the K values, and an accurate determination of the selectivity of the inhibitors, could not be
m
done and remains to be investigated.
Additionally, the selectivity of compound 13 was investigated towards two unrelated
enzymes, human arginase 1 and histone deacetylase 3 (HDAC3) (Figure S9). Compound 13
was tested for activity on HDAC3 in a biochemical enzyme activity assay and on human
arginase 1 in a binding assay based on differential scanning fluorimetry (DSF). Compound
1
3 showed inhibition of HDAC3 at higher concentrations (estimated IC ≈ 80 - 90 µM). In the
50
DSF assay, compound 13 showed no change in the melting temperature (T ) of human
m
arginase 1, suggesting that compound 13 did not bind this enzyme.
Taken together, considering the lack of selectivity for other HATs and HDAC3 and its
anti-oxidant properties, which could cause effects unrelated to KAT8 inhibition in more
advanced systems such as cells, we would not suggest using compound 13 for development
of a selective drug targeting KAT8. However, control experiments suggest compound 13
does not interfere in the KAT8 assay due to thiolreactivity or its anti-oxidant properties. Since
the aim of this study was to investigate the kinetic behavior of an inhibitor structurally
1
0

ACCEPTED MANUSCRIPT
unrelated to the current inhibitors on KAT8, compound 13 was considered suitable for further
investigation of the mechanism of inhibition.
Mechanism of inhibition
Figure 1. A) A preincubation assay was performed to investigate whether 13 is an irreversible inhibitor.
Compound 13 (0.75, 2 and 10 x IC50) was preincubated for 2, 5 or 10 minutes with KAT8. Subsequently, the
substrates were added and the enzyme reaction was performed. No difference in percentage inhibition was
observed between 2, 5 or 10 minutes preincubation with any of the concentrations. This indicates that 13 is a
reversible inhibitor. A dilution experiment was done to further confirm that 13 is a reversible inhibitor. KAT8 was
preincubated with 13 at a concentration of 50 µM (> 5 times the IC50) and subsequently diluted 100 times in a
solution containing the substrates. The enzyme activity was recovered linearly over time, indicating fully reversible
inhibition. B) The velocity of the substrate conversion of KAT8 was measured at increasing concentrations of Ac-
CoA in the presence of different concentrations of 13. An increase in 13 concentration resulted in a decrease in
V
max, but no change in the K
m
of Ac-CoA. This indicates non-competitive behavior with Ac-CoA. The velocity of
the substrate conversion of KAT8 was measured at increasing concentrations of histone substrate in the
1
1

ACCEPTED MANUSCRIPT
presence of different concentrations of 13. An increase in 13 concentration did not influence the Vmax, but resulted
in an increase in K
The equilibrium dissociation constant (K
ITC). Compound 13 showed a K of 2.6 µM for KAT8. N = stoichiometry, K = association constant (1/K
equilibrium dissociation constant (1/K). D) The equilibrium dissociation constant (K ) of 13 to KAT8 in the
presence of Ac-CoA was determined using isothermal titration calorimetry (ITC). Compound 13 showed a K of
.7 µM. N = stoichiometry, K = association constant (1/K ), K = equilibrium dissociation constant (1/K).
m
of the histone substrate. This indicates competitive behavior with the histone substrate. C)
) of 13 to KAT8 was determined using isothermal titration calorimetry
), K
d
(
d
d
d
=
d
d
0
d
d
To investigate the mechanism of inhibition by 13, kinetic studies were done using
Michaelis-Menten enzyme kinetics. The velocity of substrate conversion by KAT8 was
measured at increasing concentrations of Ac-CoA or histone substrate in the presence of
different concentrations of 13. The apparent maximal velocity (Vmax app.) and Michaelis
constants (Km app.) of Ac-CoA and the histone substrate were determined. In case of Ac-CoA,
an increase in 13 concentration resulted in a decrease in Vmax app., but no change in the Km app.
of Ac-CoA (Figure 1B). This indicates non-competitive behavior with Ac-CoA, where binding
of 13 does not influence binding of Ac-CoA. In case of the histone substrate, an increase in
13 concentration did not influence the Vmax app., but resulted in an increase in Km app. of the
histone substrate. This indicates competitive behavior with the histone substrate. KAT8 has
been shown to operate via a ping-pong mechanism (18) in which Ac-CoA has to bind first to
KAT8, followed by an acetylation of the enzyme, which creates a second intermediate form
of the enzyme. Subsequently the histone substrate binds and is acetylated, generating the
acetylated lysine and returning KAT8 back to its free form. Competitive behavior of 13 with
the histone substrate therefore suggests that 13 binds not only to the free KAT8 enzyme, but
also to the acetylated intermediate.
To further investigate this, binding studies of 13 to KAT8 were done using isothermal
titration calorimetry (ITC). Compound 13 was titrated to KAT8 in the absence of Ac-CoA (the
free enzyme) and the equilibrium dissociation constant (K ) of 13 was determined (2.6 ± 0.7
d
µM, Figure 1C, Figure S10A). The stoichiometry of the interaction was close to one,
suggesting that one molecule of 13 binds to one molecule of KAT8. Compound 13 thus
1
2

ACCEPTED MANUSCRIPT
directly interacts with KAT8. Additionally, 13 was titrated to KAT8 in the presence of Ac-CoA
to generate a certain amount of the acetylated/Ac-CoA bound enzyme form. The binding
data showed a K of 0.7 ± 0.06 µM (Figure 1D, Figure S10B). The K for Ac-CoA bound
d
d
KAT8 is lower than for free KAT8, indicating that 13 binds with a higher affinity to the
acetylated intermediate than to the free enzyme. These data suggest that compound 13
inhibits KAT8 by interacting both with the free enzyme form as well as with the Ac-CoA
bound form, but has higher affinity for the acetylated intermediate.
Figure 2. KAT8 uses a ping-pong mechanism in which the cofactor, acetyl coenzyme A (Ac-CoA), binds first to
the free enzyme form (E) and acetylates a residue on the enzyme (AcE). Coenzyme A (CoA) leaves the enzyme
as product. Subsequently, the histone substrate can bind and is acetylated by the enzyme. The enzyme returns to
its free form (E). Compound 13 (I) can interact with the free enzyme (E), with an inhibitory potency (Ki1). It can
also interact with the acetylated enzyme form (AcE) with an inhibitory potency (Ki2).
Determination of the K values
i
The IC50 value determined in an enzyme inhibition assay is dependent on the
conditions used in the assay, such as the concentration of the enzyme substrates and their
respective K values. This value is therefore not reproducible unless exactly the same assay
m
conditions are used. The K value is a potency value independent of the substrate
i
concentration and K values and is therefore much more representative as potency of the
m
inhibitor (6). Calculation of the K value from the IC is crucial for the comparison of the
i
50
potency with known inhibitors or between different assays and for the determination of
selectivity. KAT8 is a bi-substrate enzyme that converts two substrates (Ac-CoA and the
histone substrate) to two products (CoA and the acetylated histone substrate). Therefore,
knowledge on the catalytic mechanism of the enzyme, as well as the mechanism of inhibition
1
3

ACCEPTED MANUSCRIPT
by the inhibitor is needed to calculate the K value (9). It was previously reported by us that
i
KAT8 follows a ping-pong mechanism in which Ac-CoA binds first and acetylates a residue
on the enzyme. Subsequently the histone substrate binds and is acetylated (18). Combining
the knowledge that KAT8 follows a ping-pong mechanism and that 13 inhibits KAT8 by
interacting both with the free enzyme form as with the Ac-CoA bound form as determined
from the kinetic experiments, equation 1 can be used to calculate the K values (9). It must be
i
noticed that this equation yields two K values, one for the free KAT8 enzyme (K ) and one
i
i1
for the acetylated intermediate (K ). Figure 2 schematically shows the ping-pong mechanism
i2
of KAT8 and the inhibition of 13 of the free enzyme (E) and the acetylated enzyme form (AcE)
and its respective K values. The K values of Ac-CoA and the histone substrate were
i
m
derived from a kinetic assay with both substrates as described by Segel (10) and were
consistent with previously reported values (Figure S10C/D) (18). We assumed that the Ki1
was equal to the K of 13 to KAT8 as determined by ITC, since for the process of inhibiting
d
the free enzyme (E), only binding of 13 is of influence. The K of 13 was therefore assumed
i1
to be 2.6 µM and the K was 0.017 µM as derived from the equation (see SI for calculation).
i2
This suggests that the inhibitory potency was much better for the acetylated intermediate
than for the free enzyme. This stronger interaction was also observed in the K value of 13
d
for KAT8 in presence Ac-CoA as determined by ITC, which was lower than the K for KAT8
d
in absence of Ac-CoA. Additionally, the kinetic behavior of 13 is different from the behavior of
the previously reported anacardic acid derivatives, which did not interact with the free
enzyme (18). This kinetic behavior has a large influence on the calculation of the resulting Ki
values, suggesting that kinetic evaluations of inhibitors for this enzyme are crucial for
determining inhibitory potency. Taken together, this suggests that the inhibition of compound
13 is mostly due to inhibition of the acetylated enzyme intermediate by competition with the
histone substrate and that these kinetic evaluations are crucial for determining the inhibitory
potency of KAT8 inhibitors.
1
4

ACCEPTED MANUSCRIPT
Eq.1
IC50 = the IC50 determined in the enzyme inhibition assay
K
K
a
b
= K
= K
m
m
of Ac-CoA
of the histone substrate
A = the concentration of Ac-CoA used in the IC50 assay
B = the concentration of the histone substrate used in the IC50 assay
K
K
i1 = the K
i2 = the K
i
of 13 for the free enzyme
i
of 13 for the acetylated intermediate
Conclusion
In conclusion, this study describes the discovery of a potent fragment inhibitor of
KAT8, compound 13, via a fragment screening approach. A SAR study was done, which
showed that 13 could be modified, although no derivatives were found that were significantly
more potent. Investigation of the compound properties suggested that compound 13 was not
selective for other HATs or HDAC3 and had anti-oxidant properties. However, experiments
with CoA and redox active substances suggest compound 13 was a reversible inhibitor that
did not interfere in the KAT8 assay due to thiolreactivity or its anti-oxidant properties.
Therefore it was used for investigation of its mechanism of inhibiton of KAT8. The results of
kinetic studies and ITC are consistent with a model where the fragment interacts with both
the free enzyme and the acetylated intermediate form. This enabled the calculation of the
assay-independent K values of 13 for both the free enzyme form of KAT8 (K = 2.6 µM), and
i
i1
the acetylated form (K = 0.017 µM), which suggested that its inhibition is mostly due to
i2
interaction with the acetylated form of the enzyme. Taken together, in this study a fragment
screening is presented that provides a fragment inhibitor of KAT8 that is further characterized
by enzyme kinetics and biophysics. This combination reveals a striking difference in binding
affinity between the acetylated enzyme and the free enzyme that is not revealed by the IC50
determinations. This shows that kinetic characterization of inhibitors and calculation of Ki
1
5

ACCEPTED MANUSCRIPT
values is crucial for determining the binding constants of HAT inhibitors. We anticipate that
more comprehensive characterization of enzyme inhibition, as described here, is needed to
advance the field of HAT inhibitors.
Experimental Procedures
Detailed description of the experimental procedures can be found in the supplementary
material.
Acknowledgements
We acknowledge the European Research Council for providing an ERC starting grant
(
309782) and the NWO for providing a VIDI grant (723.012.005) to F. J. Dekker. This work
was supported by PRIN 2012 (prot. 2012CTAYSY) to D. Rotili and PRIN 2016 (prot.
0152TE5PK) to A. Mai.
2
1
6

ACCEPTED MANUSCRIPT
Tables
Table 1: IC50 values of the fragment screening hits and a small number of related structures.
1
2
3
Compound
R
R
R
IC (µM)
50
+
-
-
1
2
3
4
5
6
7
8
9
-NHNH Cl
4 - F
4 - OCH3
H
-
310
680
3
+
-NHNH Cl
-
3
-NHNH2
-
>1000
>1000
>1000
>1000
>1000
>1000
>1000
181
+
-
-
-NHNH Cl
H
-
3
+
-NHNH Cl
2, 4 - Cl
4- F
4-F
-
3
-NH2
-
-CH NH
-
2
2
-NH2
4- Cl
4- OH
H
-
-NH2
-NH2
-NO2
-NH2
-NH2
-
10
11
12
13
H
H
H
>1000
940
-
H
SO3
OH
H
9.7 ± 3.0
1
7

ACCEPTED MANUSCRIPT
Table 2: IC50 values of compound 13 derivatives and the selectivity of compounds 13, 9 and the
derivatives showing activity on KAT8. Data presented are IC50 values. Me = methyl, Ac = acetyl, Bn =
benzyl, n.d = not determined
1
2
3
IC50 KAT8
µM)
IC50 KAT2B
(µM)
IC50 KAT3B
(µM)
Compound
R
R
R
X
Y
(
1
3
NH2
OH
H
C
C
9.7 ± 3.0
> 500
3.6 ± 1.1
> 500
7.1 ± 0.8
n.d.
1.4 ± 0.1
> 500
1.1 ± 0.3
n.d.
9
14
15
16
17
18
19
20
21
22
23
24
25
26
NH2
NH2
-
OH
CN
OH
OH
OH
OH
OH
H
H
C
C
N
C
C
C
C
C
C
C
C
C
C
N
C
C
C
C
C
C
C
C
C
C
C
C
7.6 ± 0.8
> 500
H
> 500
n.d.
n.d.
CH NH2
H
> 500
n.d.
n.d.
2
CONH2
NO2
H
> 500
n.d.
n.d.
H
141 ± 43
177 ± 22
> 500
221 ± 115
200 ± 28
n.d.
106 ± 22
52 ± 28
n.d.
NH2
OMe
H
CH O
2
NH2
H
NH2
OMe
H
89 ± 11
> 500
> 500
n.d.
95 ± 60
n.d.
NH2
OAc
OBn
OH
H
NH2
H
> 500
n.d.
n.d.
NMe2
NHAc
H
23 ± 4
> 500
4.2 ± 1.9
n.d.
3.7 ± 2.1
n.d.
OH
H
2
7
8
OH
OH
H
H
C
C
C
C
> 500
n.d.
n.d.
2
250 ± 96
160 ± 25
60 ± 22
29
OH
H
C
C
365 ± 110
> 500
250 ± 137
NHSO2
Me
NHSO2t
oluene
3
0
1
OH
OH
H
H
C
C
C
C
66 ± 12
37 ± 5.3
3.8 ± 2.8
12 ± 1.4
3.0 ± 1.9
2.6 ± 1.4
3
Me = methyl, Ac = acetyl, Bn = benzyl, n.d = not determined
1
8

ACCEPTED MANUSCRIPT
References
(
2
1) Ganesan A. Multitarget Drugs: an Epigenetic Epiphany. ChemMedChem 2016 Jun
0;11(12):1227-1241.
(
(
2) Bird A. Perceptions of epigenetics. Nature 2007;447(7143):396-398.
3) Strahl BD, Allis CD. The language of covalent histone modifications. Nature
2000;403(6765):41-45.
(
4) Kaypee S, Sudarshan D, Shanmugam MK, Mukherjee D, Sethi G, Kundu TK. Aberrant
lysine acetylation in tumorigenesis: Implications in the development of therapeutics.
Pharmacol Ther 2016 Jun;162:98-119.
(
5) Simon RP, Robaa D, Alhalabi Z, Sippl W, Jung M. KATching-Up on Small Molecule
Modulators of Lysine Acetyltransferases. J Med Chem 2016 Feb 25;59(4):1249-1270.
(
6) Wapenaar H, Dekker FJ. Histone acetyltransferases: challenges in targeting bi-substrate
enzymes. Clin Epigenetics 2016 May 26;8:59-016-0225-2.
(
7) Lineweaver H, Burk D. The Determination of Enzyme Dissociation Constants. J Am
Chem Soc 1934 03/01;56(3):658-666.
(
8) Dixon M. The determination of enzyme inhibitor constants. Biochem J 1953
Aug;55(1):170-171.
(
9) Cheng Y, Prusoff WH. Relationship between the inhibition constant (K1) and the
concentration of inhibitor which causes 50 per cent inhibition (I50) of an enzymatic reaction.
Biochem Pharmacol 1973 Dec 1;22(23):3099-3108.
(
10) Segel I, H. Enzyme Kinetics: Behavior and Analysis of Rapid Equilibrium and Steady-
State Enzyme Systems. : Wiley Classics Library; 1993.
(
2
11) Marmorstein R. Structure and function of histone acetyltransferases. Cell Mol Life Sci
001 May;58(5-6):693-703.
(
12) Smith ER, Cayrou C, Huang R, Lane WS, Cote J, Lucchesi JC. A human protein
complex homologous to the Drosophila MSL complex is responsible for the majority of
histone H4 acetylation at lysine 16. Mol Cell Biol 2005 Nov;25(21):9175-9188.
(
13) Dou Y, Milne TA, Tackett AJ, Smith ER, Fukuda A, Wysocka J, et al. Physical
association and coordinate function of the H3 K4 methyltransferase MLL1 and the H4 K16
acetyltransferase MOF. Cell 2005 Jun 17;121(6):873-885.
(
14) Li X, Wu L, Corsa CA, Kunkel S, Dou Y. Two mammalian MOF complexes regulate
transcription activation by distinct mechanisms. Mol Cell 2009 Oct 23;36(2):290-301.
(
15) Cai Y, Jin J, Swanson SK, Cole MD, Choi SH, Florens L, et al. Subunit composition and
substrate specificity of a MOF-containing histone acetyltransferase distinct from the male-
specific lethal (MSL) complex. J Biol Chem 2010 Feb 12;285(7):4268-4272.
1
9

ACCEPTED MANUSCRIPT
(
16) Rea S, Xouri G, Akhtar A. Males absent on the first (MOF): from flies to humans.
Oncogene 2007 Aug 13;26(37):5385-5394.
(
17) Valerio DG, Xu H, Chen CW, Hoshii T, Eisold ME, Delaney C, et al. Histone
Acetyltransferase Activity of MOF Is Required for MLL-AF9 Leukemogenesis. Cancer Res
017 Apr 1;77(7):1753-1762.
2
(
18) Wapenaar H, van der Wouden PE, Groves MR, Rotili D, Mai A, Dekker FJ. Enzyme
kinetics and inhibition of histone acetyltransferase KAT8. Eur J Med Chem 2015 Nov
3;105:289-296.
1
(
19) Baell JB. Observations on screening-based research and some concerning trends in the
literature. Future Med Chem 2010 Oct;2(10):1529-1546.
(
20) Aldrich C, Bertozzi C, Georg GI, Kiessling L, Lindsley C, Liotta D, et al. The Ecstasy and
Agony of Assay Interference Compounds. ACS Chem Biol 2017 Mar 17;12(3):575-578.
(
21) Capuzzi SJ, Muratov EN, Tropsha A. Phantom PAINS: Problems with the Utility of Alerts
for Pan-Assay INterference CompoundS. J Chem Inf Model 2017 Mar 27;57(3):417-427.
(
22) Dahlin JL, Nissink JW, Strasser JM, Francis S, Higgins L, Zhou H, et al. PAINS in the
assay: chemical mechanisms of assay interference and promiscuous enzymatic inhibition
observed during a sulfhydryl-scavenging HTS. J Med Chem 2015 Mar 12;58(5):2091-2113.
(
23) Copeland R. Evaluation of Enzyme Inhibitors in Drug Discovery: A Guide for Medicinal
Chemists and Pharmacologists. 2nd ed.: Wiley; 2013.
(
24) Kedare SB, Singh RP. Genesis and development of DPPH method of antioxidant assay.
J Food Sci Technol 2011 Aug;48(4):412-422.
2
0

ACCEPTED MANUSCRIPT
Supplementary material
Fragment screening
Figure S1. IC curves of compound 1 - 13 from the fragment screening
50
2
1

ACCEPTED MANUSCRIPT
Structure activity relationship
Figure S2. The derivatives of 13 were tested for inhibition of KAT8. A single point screening was done at a
concentration of 250 µM. Derivatives showing inhibition, were tested for their IC values.
50
2
2

ACCEPTED MANUSCRIPT
Inhibitor properties
Figure S3. The thiolreactivity of compound 13 was investigated. The fluorescence intensity was measured in the
resence of CoA (4 µM), histone substrate (60 µM), KAT8 (250 nM) and the compound 13 (0 – 100 µM in 2%
DMSO). No decrease in fluorescence was observed.
uV
5
000 Data1:19772_170331_4A1N_1.2_001.lcdPDACh2 340nm,4nm
Buffer only
0
0
,0
2,5
5,0
7,5
10,0
10,0
10,0
10,0
12,5
12,5
12,5
12,5
15,0
15,0
15,0
15,0
17,5
17,5
17,5
17,5
min
min
min
min
uV
Data1:19773_170331_4A1N_1.3_001.lcdPDACh2 340nm,4nm
2
5000
Immediately
0
0,0
2,5
5,0
7,5
uV
Data1:19774_170331_4A1N_1.4_001.lcdPDACh2 340nm,4nm
50000
1
4
h
h
25000
0
0,0
2,5
5,0
7,5
uV
Data1:19775_170331_4A1N_1.5_001.lcdPDACh2 340nm,4nm
50000
0
0,0
2,5
5,0
7,5
Figure S4. Stability of compound 13 in assay buffer. HPLC of compound 13 in assay buffer after 0, 1 and 4
hours of incubation.
2
3

ACCEPTED MANUSCRIPT
Figure S5. A) The antioxidant activity (EC ) of 9, 13 and all derivatives was measured in a DPPH assay. B)
5
0
+
The IC of 13 was measured in the presence of TCEP (1 µM), NAD (100 µM) and NADH (100 µM). *
50
Different conditions were used for this assay: Ac-CoA (1 µM), histone substrate (60 µM) and KAT8 (14.5 nM).
2
4

ACCEPTED MANUSCRIPT
Selectivity
KAT2B
Figure S6. Compounds 9, 13 and all derivatives active on KAT8, were tested for activity on KAT2B. IC50
curves of these compounds on KAT2B under the following conditions: 25 µM histone H3 substrate, 10 µM Ac-
CoA and 100 nM KAT2B.
2
5

ACCEPTED MANUSCRIPT
KAT3B
Figure S7. Compounds 9, 13 and all derivatives active on KAT8, were tested for activity on KAT3B. IC50
curves of these compounds on KAT2B under the following conditions: 100 µM histone H4 substrate, 50 µM Ac-
CoA and 100 nM KAT3B.
2
6

ACCEPTED MANUSCRIPT
Kinetic evaluation of KAT2B and KAT3B
Figure S8. The catalytic mechanism of KAT2B and KAT3B was investigated with the aim of determining the
K values of their substrates. A) The velocity of KAT2B was measured at increasing concentrations of Ac-CoA
m
(
0-20 µM) in the presence of different concentrations of the histone H3 substrate (15, 25, 50 and 100 µM). The
enzyme showed sigmoidal kinetics. The Lineweaver-Burke plot was created by transforming the data to the
double reciprocal and was shown to be not linear. B) The velocity of KAT3B was measured at increasing
concentrations of Ac-CoA (0-500 µM) in the presence of different concentrations of histone H3 substrate or
histone H4 substrate (5, 50 and 500 µM). Saturation could not be reached under these conditions.
HDAC3 and arginase 1
HDAC3
Arginase 1
150
100
80
60
40
20
0
1
00
5
0
0
-8
-6
-4
-2
0
2
4
25
2.5
0.25
0
log[13] (µM)
[13] (µM)
Figure S9. The selectivity of compound 13 was tested for HDAC3 in a biochemical enzyme inhibition assay and
for arginase 1 in a differential scanning fluorimetry based assay.
2
7

ACCEPTED MANUSCRIPT
Mechanism of inhibition
Figure S10. A) Compound 13 (400 µM) was titrated to KAT8 (top), raw data. A baseline was generated by
titrating 13 to buffer using the same settings (bottom). B) Compound 13 (400 µM) was titrated to KAT8 (40 µM)
in the presence of Ac-CoA (80 µM) (top) , raw data. A baseline was generated by titrating 13 to buffer with Ac-
CoA (80 µM) using the same settings (bottom). C) Determination of Km app. values of Ac-CoA for KAT8. The
velocity was measured at increasing concentrations of Ac-CoA (0-25 µM) in the presence of different
concentrations of the histone substrate (30, 60 and 90 µM). Representative graph (triplicate) of two individual
experiments.This yielded three apparent K values for Ac-CoA, which are dependent on the concentration of
m
histone substrate. D) Determination of the real K values of Ac-CoA and the histone substrate for KAT8. Replot
m
of the reciprocal of the Km app. values of Ac-CoA against the reciprocal of the histone substrate concentration
used in the kinetic assay. A linear regression of these data points gave the K of Ac-CoA as reciprocal of the
m
intercept of the X axis (4.3 µM) and the K of the histone substrate as negative reciprocal of the intercept with
m
the X axis (89 µM).
2
8