Cofactor Analogues as Active Site Probes in Lysine Acetyltransferases

Article abstract of DOI:10.1021/acs.jmedchem.8b01887

Lysine acetyltransferases (KATs, also termed histone acetyltransferases, HATs) catalyze the acetylation of substrate lysine residues by employing the cofactor acetyl-coenzyme A (AcCoA), thereby providing a dynamic control mechanism of protein function. Because of their major involvement in cell development and homeostasis, small-molecule modulators of KAT activity are urgently needed to assess their therapeutic potential and for probing their underlying biology. Recent advances in the field suggest that targeting the cofactor binding site represents a promising strategy for identifying potent and selective ligands. Here, we present the synthesis of two functional cofactor-based chemical probes and their usage as mechanistic tools in a broadly applicable assay platform. A fluorescence polarization (FP)-based binding assay was combined with biolayer interferometry competition analysis and a FP competition activity immunoassay to enable easy, reliable, and profound evaluation of ligands that target the KAT cofactor binding site.

Full text of DOI:10.1021/acs.jmedchem.8b01887

Subscriber access provided by MIDWESTERN UNIVERSITY
Article
Cofactor Analogues as Active Site Probes in Lysine Acetyltransferases
Roman P. Simon, Tobias Rumpf, Vaida Linkuvien#, Daumantas Matulis, Asifa Akhtar, and Manfred Jung
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.8b01887 • Publication Date (Web): 20 Feb 2019
Downloaded from http://pubs.acs.org on February 21, 2019
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 49
Journal of Medicinal Chemistry
1
2
3
4
Cofactor Analogues as Active Site Probes in Lysine Acetyltransferases
5
6
7
8
9
Roman P. Simon,† Tobias Rumpf,‡ Vaida Linkuviene,§ Daumantas Matulis,§ Asifa Akhtar,‡ and
Manfred Jung†,*
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
† Institute of Pharmaceutical Sciences, University of Freiburg, Albertstraße 25, 79104 Freiburg im
Breisgau, Germany
‡ Department of Chromatin Regulation, Max-Planck-Institute of Immunobiology and Epigenetics,
Stuebeweg 51, 79108 Freiburg, Germany
§ Department of Biothermodynamics and Drug Design, Institute of Biotechnology, Life Sciences Center,
Vilnius University, Saulėtekio 7, 10257 Vilnius, Lithuania
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 49
1
2
3
4
5
6
7
8
9
ABSTRACT: Lysine acetyltransferases (KATs, also termed histone acetyltransferases, HATs) catalyze the
acetylation of substrate lysine residues by employing the cofactor acetyl-coenzyme A (AcCoA), thereby
providing a dynamic control mechanism of protein function. Because of their major involvement in cell
development and homeostasis, small molecule modulators of KAT activity are urgently needed to assess
their therapeutic potential and for probing their underlying biology. Recent advances in the field suggest
that targeting the cofactor binding site represents a promising strategy for identifying potent and selective
ligands. Here we present the synthesis of two functional cofactor-based chemical probes and their usage as
mechanistic tools in a broadly applicable assay platform. A fluorescence polarization (FP) based binding
assay was combined with biolayer interferometry (BLI) competition analysis and a FP competition
activity immunoassay to enable easy, reliable, and profound evaluation of ligands that target the KAT
cofactor binding site.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
ACS Paragon Plus Environment

Page 3 of 49
Journal of Medicinal Chemistry
1
2
3
4
INTRODUCTION
5
6
7
8
The reversible acetylation of lysine residues has been identified as a means of fundamental regulation of
protein and genome function, which is conserved from prokaryotes to mammals.1-2 The dynamic
acetylation equilibrium within a cell is balanced by lysine acetyltransferases (KATs) and lysine
deacetylases (KDACs).3 Although originally identified as modifiers of histone proteins (HATs and
HDACs), enzymatic activity of these enzymes has also been demonstrated for a variety of non-histone
substrates, underlining their important role in cell physiology.4 All KATs employ the cofactor acetyl-
coenzyme A (AcCoA (Figure 1)) to transfer an acetyl group to the ε-N moiety of lysine residues of
substrate enzymes, releasing the formed acetamide and CoA. Although AcCoA is the general KAT
cosubstrate, acylation activity employing other short-chain acyl-CoAs has been demonstrated for some
KAT enzymes.5-7 Based on sequence similarity and catalytic mechanism, the major KAT activity
containing proteins identified in humans are grouped into three distinct families, named the GNAT (Gcn5
related N-acetyltransferase) family (including Gcn58, pCAF9, HAT110, ATAT-111), the metazoan specific
p300/CBP family (including p30012 and CBP13), and the MYST family (including MOF14, MOZ, MORF15,
Tip6016 and HBO117). Multiple other enzymes have been reported to catalyze protein lysine acetylation
among which are yeast homologues of mammalian KATs, non-structurally related proteins, and poorly
characterized KAT candidates.18-21 The catalytic domains of KAT enzymes share high sequence homology
within a distinct family and little to no sequence similarity to other families. These core domains are
generally flanked by multiple other protein modules, which directly or indirectly contribute to substrate
specificity and efficacy of the acetylation reaction. The functional consequences of enzymatic lysine
acetylation as a post-translational protein modification are attributable to multiple mechanisms including
recruitment of acetyl lysine binding proteins, protein stabilization, autoacetylation, and chromatin
relaxation.22 Dysregulation of these processes as a consequence of KAT aberration have been causally
correlated with a number of malignancy phenotypes such as neurodegenerative disease23, viral and
parasitic infections24-25, inflammation26, cancer27-29, and metabolic disorders30. Mutations in gene
sequence, changes in expression levels and fusion proteins as a consequence of chromosomal
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 49
1
2
3
4
5
6
7
8
9
translocation represent possible causes for dysregulated KAT activity in such diseases. Because many
KATs associate with effector proteins under physiological conditions, it has been challenging to dissect
whether a certain biological effect is attributable to their enzymatic activity or to their scaffolding function
in a cellular context.31 An appropriate way to encounter this problem is the use of potent and specific
small molecule inhibitors to assess the biological consequences and therapeutic potential of KAT
inhibition. Despite major efforts towards the identification and development of such compounds, only two
drug-like inhibitors for p300/CBP32 (KAT3a/b) and MOZ/MORF33 (KAT6a/b), respectively, were made
available until to date.34-35 Early approaches to this topic focused on bisubstrate mimics via covalent
connection of the physiological cofactor to substrate peptides (acetonyl-CoA, Lys-CoA, and H3-CoA-20;
Figure 1).36-37 While these compounds suffer from poor cell permeability and complex structure, they
exhibit potent and selective KAT inhibition. The concept was adopted for the generation of functionalized
chemical probes mainly for the profiling of KAT substrates and for chemoproteomic applications.38
Selected examples are depicted in Figure 1: the reactive compound desthiobiotin-sulfone-CoA was used
for active site cysteine labeling of several acetyltransferases39, the cofactor surrogates 1a and 1b led to
functional labeling of KAT substrates40-41, and the bisubstrate inhibitor based probe 1c equipped with an
biotin affinity handle was used for selective enrichment of KATs from crude cell lysates.42 In the present
article, we describe our adaptation of the bisubstrate strategy to generate and characterize functionalized
cofactor-based probes of reduced structural complexity to assemble a versatile and broadly applicable
assay platform. The catalytic domain of pCAF (KAT2B) was used to develop a fluorescence polarization
(FP)-based displacement assay supplemented with a biolayer interferometry (BLI) assay that allows for
the measurement of real time binding data. Both assays were then transferred to p300 (KAT3B) and MOF
(KAT8) to show the versatility of the platform. These test systems were combined with a FP competition
activity immunoassay to enable easy, reliable, and profound screening, characterization, and profiling of
candidate compounds that target the KAT catalytic domain.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
4
ACS Paragon Plus Environment

Page 5 of 49
Journal of Medicinal Chemistry
1
2
3
4
RESULTS
5
6
7
8
Probe Design and Synthesis. Although there is a lot of data published on the inhibitory properties of
cofactor analogues and bisubstrate inhibitors, comparably little is known about the thermodynamic
binding affinities of respective compounds. KD values of 0.64 ± 0.12 µM and 8.7 ± 1.2 µM have been
published for AcCoA binding to pCAF and MOF, respectively.43-44 For p300, an affinity constant of 0.74
± 0.06 µM for the interaction with the non-hydrolyzable reference inhibitor acetonyl-CoA has been
reported.45 While it has been shown that extensive modification of the peptide part of bisubstrate inhibitors
can result in better inhibitory potency and KAT specificity, we sought for a way to generate CoA
analogues with a broader scope of KAT interactions to expand the applicability of these probes. We
therefore conceptually reduced the peptide part to a single lysine mimicking aminohexanoic acid residue
that bridges the CoA part of the probe with a functional label (Figure 2). Immobilized acetonyl-CoA has
previously been used for affinity enrichment of KATs originating from different enzyme families.46 We
hypothesized that such functionally labeled CoA analogues would exhibit sufficiently high binding
affinities to be used as chemical tool compounds in screening and profiling of cofactor competitive
ligands. The synthesis of our probes was started with the assembly of a bifunctionalized linker, which was
subsequently coupled to CoA and the functional label (Scheme 1). Therefore, Boc-protected 6-
aminohexanoic acid (2) was condensed with propargylamine to the alkynylated intermediate 3 using
standard amide bond formation. Deprotection followed by chloroacylation in DMF yielded the
bifunctionalized linker 4. We next assembled the fluorescence tag by submitting amine 5 to a 1H-
imidazole-1-sulfonyl azide promoted diazotransfer to generate the azide 6, which was treated with oxalyl
chloride in dichloromethane in presence of catalytic amounts of DMF. The resulting acyl chloride was
coupled to stoichiometric amounts of 6-aminofluorescein to yield the fluorescence tag 7. Finally, alkyl
linker 4 was converted to the respective probes. For the fluorescence labeled derivative, 4 and 7 were
combined in a Cu-catalyzed Huisgen cycloaddition to furnish compound 8, which was further reacted with
CoA in a SN2 reaction to generate the probe 11. In the case of the affinity labeled probe, nucleophilic
substitution of the alkyl halide was carried out first (9), followed by the cycloaddition reaction to yield the
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 49
1
2
3
4
5
6
7
8
9
desired compound 12 (Figure 2). In the next step, we selected pCAF as a prototypical KAT for probe
validation and assay development, because of its high cofactor binding affinity and well-defined binding
mode. The catalytic domain of pCAF fused to a N-terminal His-tag (pCAF(493-658) was expressed and
purified from E. coli cells to provide sufficient amounts of protein. To test whether our compounds are
capable to bind to the investigated KAT proteins, a pull-down experiment was carried out. Streptavidin
coated magnetic beads were used in combination with 12 to individually capture recombinant proteins
from solutions in the presence of BSA (Supporting Information, Figure S1). Control experiments
containing either 100 µM acetonyl-CoA or 50 µM biotin were included for every KAT. SDS-gel analysis
of the isolated fractions showed successful pull-down of all three KAT proteins respectively, which is
indicative for effective binding of 12 to the tested enzymes. Next, we aimed at quantifying the binding
affinities of the probes. ITC measurements of pCAF(493-658) interacting with either the physiological
cofactor AcCoA, the reference inhibitor acetonyl-CoA, fluorescence labeled probe 11, or the affinity
labeled probe 12, yielded KD values of 0.77 ± 0.17 µM, 0.71 ± 0.31 µM, 0.75 ± 0.36 µM, and 1.39 ± 0.23
µM, respectively (Supporting Information, Figure S2). These results are in good agreement with the
previously published binding affinity for AcCoA (0.64 ± 0.12 µM) and demonstrate the possibility to
functionally label at the selected derivatization site whilst maintaining binding affinities comparable to the
physiological ligand. With the probes 11 and 12 validated as efficient binders of three different KAT
proteins, we then turned our attention to the applications of the tool compounds.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Fluorescence Polarization Competition Assay. The labelled CoA analogue 11 was used as a fluorescent
tracer to set up a competition binding fluorescence polarization assay. As the degree of polarization is
inversely proportional to the molecular mobility and therefore correlates with the size of the fluorescent
species, the polarization signal increases with the fraction of 11 bound to the receptor (here KAT protein).
In a setup with fixed concentrations of 11 and KAT protein, competing test ligands reduce the fraction of
tracer that is bound to the receptor, which results in a decrease of measured mP values (Figure 3a). To
assess the optimal probe concentration, we prepared a serial dilution of 11 in assay buffer covering a
6
ACS Paragon Plus Environment

Page 7 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
concentration range of 3.8-30 nM in 20 µl volume and measured the fluorescence intensities (FI) in the p-
and s-plane. Resulting values were plotted against probe concentrations to show a linear increase of
fluorescence intensity for both planes (R2=0.999 and 0.999) and a stable ratio of FI values over the entire
concentration range (Supporting Information, Figure S3a). A probe concentration of 8 nM was selected
for further experiments as this amount of 11 gave FI values that exceeded the values measured for buffer
controls five times, which resulted in a stable and robust FP signal. We next performed a saturation
binding experiment by titrating a dilution series of pCAF(493-658) protein to 11 in presence or without a
saturating concentration of reference inhibitor (200 µM acetonyl-CoA). To control for intrinsic protein
fluorescence, the protein dilution series was also incubated without fluorescent tracer. The fluorescence
intensities for p-plane and s-plane were then measured from all wells and values from the protein
fluorescence controls were subtracted from tracer containing wells for every pCAF(493-658) protein
concentration. From the resulting fluorescence intensities, the mP values were calculated. From reference
inhibitor containing wells, the non-specific portion of measured fluorescence polarization signals was
determined and values were corrected accordingly. With the resulting signals for specific binding of 11 to
pCAF(493-658) we were able to calculate the affinity constant of the interaction, which was determined to
be KD = 824 ± 40 nM by non-linear regression using a one-site binding fit (Figure 3b). This value is in
excellent agreement with our results from ITC measurements and sufficiently low to use 11 as a
fluorescent tracer in FP applications. Further, by calculating total fluorescence intensities, we were able to
show that there is no substantial fluorescence quenching event evident upon binding of tracer to the
protein (Supporting Information, Figure S3b). To test the competitive displacement of 11 from binding to
pCAF(493-658), we then titrated acetonyl-CoA to mixtures containing fixed concentrations of protein and
tracer (400 nM and 8 nM, respectively) and measured the FP signal at different time points. The relative
competition compared to control wells was calculated for every ligand concentration and resulting values
were plotted against acetonyl-CoA concentrations. Figure 3b shows the results from this experiment after
non-linear curve fitting, demonstrating a stable dose-response at all measured time-points. From obtained
IC50 values, an affinity constant for the interaction of acetonyl-CoA to pCAF(493-658) of KD = 592 ± 66
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 49
1
2
3
4
5
6
7
8
9
nM was calculated, which again is in good agreement with the value determined by ITC. With optimized
assay parameters of 400 nM pCAF(493-658), 8 nM 11, and 30 minutes incubation time, we investigated
the influence of increasing DMSO content on our system. Therefore, we measured positive controls,
containing all assay components and increasing amounts of DMSO (v/v), and compared them to negative
controls containing 50 µM acetonyl-CoA additionally (Supporting Information, Figure S3c). The signal
window of the assay did not decrease by more than 10% for concentrations up to 8% (v/v) of organic
solvent. We selected a standard DMSO concentration of 5% (v/v) for the screening of ligands targeting the
CoA binding site and evaluated the reproducibility and robustness of the assay by preparing 32 positive
and 32 negative (containing 50 µM acetonyl-CoA) controls on a single 384-well plate. From the FP
signals of these measurements, the Z’-factor for the assay was calculated as described by Zhang et al. as a
surrogate for assay quality (Supporting Information, Figure S3d).47 With a Z’-factor of 0.81 our assay
meets the requirements for a stable and robust assay (Z’ > 0.5) and facilitates the screening for
competitive ligands targeting the cofactor binding site.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Biolayer Interferometry Binding Assay. As a confirmatory biophysical assay, we chose biolayer
interferometry as a method to revalidate candidate ligands identified as competitors by the fluorescence
polarization competition assay. BLI measures the interference pattern of white light that is reflected on a
biocompatible surface immobilized on a sensor tip compared to a reference beam. Upon binding of
biomolecules to the functionalized surface the optical thickness of this layer increases, which results in a
shift of the interference pattern that is expressed in nm. The increase of the BLI signal is proportional to
the number of biomolecules bound to the functionalized surface and can be measured in real time proving
a means for the determination of kinetic data of the molecular interaction. In our approach, we used
streptavidin coated sensor tips to immobilize the biotinylated CoA analogue 12 onto the biosurface and
investigate the binding of acetyltransferase protein to the immobilized probe (Figure 3c). First, we
determined the optimal loading density of 12 on the biosensors. Oversaturation of biosensors can cause
steric hindrance and aggregation, whereas too little ligand immobilized can result in insufficient signal
8
ACS Paragon Plus Environment

Page 9 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
intensities. To assess a suitable loading concentration, a dilution series of 12 in assay buffer was prepared
and the BLI signal increase during the loading step and subsequently for the binding phase, during which
the loaded sensors were exposed to a 1 µM solution of pCAF(493-658), were monitored (Supporting
Information, Figure S4a and S4b). A probe concentration of 125 nM was found to give a linear signal
increase in the loading phase without reaching saturation while the signal increase during the binding
phase remained sufficiently high. Next we decided to measure a KD value for the interaction of
pCAF(493-658) to immobilized 12. Because global fitting of sensorgrams to a one site binding model for
different protein concentrations was unprecise in the dissociation phase, we decided to use a saturation
binding approach for the determination of the affinity constant (Supporting Information, Figure S4c).
Therefore, we incubated preloaded biosensors with a dilution series of protein concentrations spanning
from 0.04 µM to 30 µM, which was supposed to cover a concentration range of approximately 0.05 to 20
times KD. Exposure of a biotin blocked biosensor to 30 µM pCAF(493-658) in assay buffer showed no
significant increase in binding signal, suggesting the absence of unspecific interactions of the protein to
the biosensor. BLI signals for the association- and dissociation phase were recorded as described in the
experimental section (Figure 3d). Report points at the end of the association phase were then used to
extract binding signals from the recorded sensorgrams for every protein concentration used. Plotting of
these values against protein concentration followed by non-linear curve fitting to a one site binding model
gave a KD value of 1.00 ± 0.28 µM, which is in good agreement with values determined by ITC and FP
(1.39 ± 0.23 µM and 0.71 ± 0.08 µM, respectively). To test the feasibility of the system to evaluate
compounds as competitive ligands, we tested the influence of increasing concentrations of reference
inhibitor in protein solutions on the sensorgrams of the association step. A fixed concentration of 250 nM
pCAF(493-658) was incubated with different amounts of acetonyl-CoA and the association curves were
recorded on the BLI instrument (Figure 3e). We observed a concentration dependent decrease in binding
signal with a maximum that almost reached blank (loaded sensors in assay buffer) level. Comparing the
BLI signals at report points set at the end of the association phase allowed us to quantify the degree of
competition for every acetonyl-CoA concentration and plotting of respective nm values against ligand
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 49
1
2
3
4
5
6
7
8
concentrations gave an IC50 value of 0.78 ± 0.10 µM (Figure 3e). This value is very close to the KD values
for the interaction of pCAF(493-658) with acetonyl-CoA determined by other methods and shows the
utility of the assay for the validation of candidate ligands.
9
Applicability of the Assays to Other Family KATs. To test the applicability of the FP assay and the BLI
assay to other KAT family members, we chose two representative KATs, p300 (p300(1284-1673)) of the
p300/CBP family and MOF (MOF(174-449)) of the MYST family, and subjected them to the same cycles
of assay optimization as described for pCAF (Figure 4). For p300(1284-1673), slow but tight binding to
the fluorescent probe was observed and in comparison to pCAF, longer incubation times were needed to
reach equilibrium. The FP saturation binding experiment yielded a KD value of 396 ± 31 nM for binding to
11, which allowed to calculate an affinity constant of KD = 653 ± 56 nM for the reference inhibitor from
the dose-response curve. This value is close to the published value of 780 ± 60 nM. The slow binding of
p300(1284-1673) was also observable in the BLI assay and the protocol had to be adjusted accordingly.
Instead of loading the biosensor with 12 first, we incubated the protein with 125 nM probe alone or in the
presence of competing ligands for 30 minutes and measured the association to the streptavidin coated
sensor from these solutions. While this does not allow for deducing any kinetic data from the association
curves, the new protocol proved to be usable to evaluate ligands competitive to 12. Titrating acetonyl-CoA
in this assay resulted in an IC50 of 779 ± 30 nM. Notably, the tight binding of p300(1284-1673) to 12 was
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
reflected in only very little dissociation of the protein from the sensortip during the dissociation step (koff
=
2.1∙10-4 1∙s-1). For MOF(174-449), the saturation binding experiment yielded a KD value of 1.39 ± 0.11
µM for binding to 11. The increase in FP signal in this experiment was less pronounced compared to the
other tested KAT proteins. The affinity constant for MOF(174-449) binding to acetonyl-CoA was
determined from this experiment and resulted in KD = 1.58 ± 0.14 µM, which is distinctly lower than the
published value for the protein cofactor interaction (8.7 ± 1.2 µM). Since we were not able to measure BLI
signals at sufficiently high protein concentrations to reach saturation binding, we determined a KD = 4.13
± 0.4 µM for MOF(174-449) binding to 12 by applying a global one site binding fit to the sensorgrams of
a protein dilution series (Supporting Information, Figure S4d). Titrating the reference inhibitor against 4
10
ACS Paragon Plus Environment

Page 11 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
µM protein in this assay resulted in an IC50 of 1.35 ± 0.18 µM for acetonyl-CoA. Together, these
experiments demonstrate the broad applicability of 11 and 12 as chemical tool compounds for KAT
enzymes and the versatility of our assays platform.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
FP Competition Activity Immunoassay. To complement the biophysical test systems, we sought for a
non-radioactive assay to measure the activity of the investigated acetyltransferases under in vitro
conditions. We also aimed at an approach that does not require the use of labeled substrates or reactive
detection reagents, which can lead to data artefacts like high background signals or false positive
screening hits. Because of our good experience with fluorescence polarization, we hypothesized that it
would be possible to design an assay that measures the enzyme activity dependent competition of reaction
product to a fluorescence labelled tracer for binding to a specific antibody, as it has been reported for
kinases (Scheme 2).48 This would allow for free scaling of substrate and cofactor concentrations and the
use of unlabeled and more physiological peptide or protein substrates. Although the preferred H3
acetylation site for pCAF has been reported to be H3K14, we had obtained good signal to background
ratios using a polyclonal anti-acH3K9 antibody (#39137, Active Motif) in our biochemical assays (data
not shown). As a fluorescent tracer, we used a peptide that contains the first 14 amino acids of the histone
H3 N-terminal sequence with a rhodamine label attached to the side chain of K 14 and an acetylated K9
moiety (acH3K9-Rh). For MOF(174-449) and p300(1284-1673), we expected a broader spectrum of
acetylation products. Therefore, a tetra-acetylated H4(1-23) peptide with a C-terminal fluorescein label
was used as a tracer (acH4-Fl). The exact sequence of peptides can be found in the Supporting
Information. The fluorescent properties of acH3K9-Rh and acH4-Fl were investigated as described for 11
and tracer concentrations of 8 nM and 3 nM were selected, respectively (Supporting Information, Figure
S5a). In the next step, different antibodies were titrated to solutions containing the indicated concentration
of acH3K9-Rh or acH4-Fl and 10 µM of H3(1-21) peptide or H4(1-23) peptide, respectively, in 25 µl
volume on 384-well microtiter plates and the changes in FP signals were recorded. Satisfactorily, we
identified two antibodies that were capable to bind their epitope in solution thereby inducing a
11
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 49
1
2
3
4
5
6
7
8
9
concentration dependent FP signal increase (Figure 5a). Our selected anti-acH3K9 antibody caused an
almost linear increase in mP values up to the highest concentration tested (5 µg/ml), which corresponds to
a 1:200 dilution in a final assay volume of 25 µl. For acH4-Fl, the only antibody with a good response in
this experiment was a monoclonal anti-acH4 antibody (#177790, abcam) directed against tetra-acetylated
H4 (K5, 8, 12, 16). Addition of anti-acH4 to the tracer peptide resulted in a steep increase of FP signal that
leveled around 160 mP for antibody concentrations exceeding 0.5 µg/ml (1:700 in 25 µl volume). The
feasibility of the setup to measure KAT activity was then investigated for established tracer/receptor pares
(8 nM acH3K9-Rh/5 µg/ml anti-acH3K9 and 3 nM acH4-Fl/0.4 µg/ml anti-acH4, respectively). For
successful activity measurements, the product of the enzymatic reaction has to be competitive to the tracer
peptide for binding to the antibody. If the system fulfills this requirement, the degree of competition can
be calculated from recorded FP values and is proportional to enzyme activity. Concentration series of
enzymes were incubated with substrate peptide or protein (50 µM H3(1-21) for pCAF(493-658), 5 µM full
length H4-His for MOF(174-449) and 5 µM H4(1-23) for p300(1284-1673)) and 20 µM AcCoA for 60
minutes. Enzymatic reactions were then stopped by adding 2.5 µl of tracer solution containing 500 µM
acetonyl-CoA followed by a short incubation step. Finally, 2.5 µl antibody solution in assay buffer were
dispensed into every well and after an additional 30 minutes incubation step, FP signals were recorded and
competition values calculated. Figure 5b shows a clear increase in competition with increasing enzyme
concentrations, which strongly suggests a concentration dependent displacement of tracer from the
receptor by the product of enzymatic turnover. Further, the use of full length H4 histone protein as a
substrate demonstrates the capability of the assay to use physiological protein substrates. Enzyme
concentrations from the ends of the linear portions of the enzyme progression curves were selected for
further experiments (30 nM pCAF(493-658), 50 nM MOF(174-449), and 15 nM p300(1284-1673)). We
then determined the assay specific Km values for AcCoA and selected screening concentrations that, again,
were at the upper end of the linear portion of the Michaelis-Menten plots to preserve a sufficient assay
window (Supporting Information, Figure S5b). With the defined assay parameters, dose-response curves
for the reference inhibitor were recorded (Supporting Information, Figure S5c) and Ki values were
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
12
ACS Paragon Plus Environment

Page 13 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
calculated according to the method of Cheng and Prusoff.49 We also determined the assay specific Z’-
factors (all > 0.5) as a measure of assay robustness (Supporting Information, Figure S5d). The exact
results from our activity assay development and reference inhibitor validation are summarized in Table 2.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Versatility of the Assay Platform. With the evaluated tool compounds and optimized assay parameters,
the utility of the assay platform for different applications was investigated.
Profiling of acyl-CoA variants: First, we followed observations by Montgomery et al. 42 and Ringel et al.50
who have reported inhibitory properties of metabolic acyl-CoAs (e.g. butyryl-CoA and several long-chain
acyl-CoAs(Figure 1)) on the GNAT family KAT hGcn5. To validate the role of the chain length and
quantify the degree of interaction, a small set of acyl-CoA cofactor analogues were profiled for their
binding properties to the closely related Gcn5 homologue pCAF. We included the enzymatic reaction
product CoA (C-0), the physiological cofactor AcCoA (C-2), butyryl-CoA (C-4), lauroyl-CoA (C-12),
aromatic benzoyl-CoA, and the more hydrophilic acyl-CoA variant hydroxybutyryl-CoA. A comparison
of the competition profiles from FP and BLI assays showed a strong increase of interaction for the
physiological cofactor compared to the enzymatic reaction product CoA (Figure 6a). Elongation of the
carbon chain to a butyryl residue then resulted in a drop of competition to CoA level. The same effect was
observed for structurally more demanding acyl-CoA variants benzoyl-CoA and hydroxybutyryl-CoA.
Further elongation to the C-12 derivative lauroyl-CoA then led to a strong increase of potency that
exceeded the level of AcCoA. Recording the dose-response curves of acyl-CoA variants with increasing
chain length in the FP assay allowed us to calculate the affinity constants of respective interactions from
IC50 values (Figure 6c). We then revalidated our findings by comparing the results from FP and BLI
assays to the stabilizing effects (ΔTm) of tested CoA analogues in a fluorescent thermal stability assay
(Figure 6b). Table 3 shows the results from these experiments and assigns lauroyl-CoA as a potent binder
(KD = 0.18 ± 0.01 µM) of pCAF(493-658) with a five times higher affinity than the physiological cofactor
(KD = 1.03 ± 0.09 µM). To test whether the high affinity also translates to inhibitory potency, lauroyl-CoA
was evaluated in the FP competition activity immunoassay (Figure 6d). The compound inhibits
13
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 49
1
2
3
4
5
6
7
8
9
pCAF(493-658) more potently (Ki = 0.78 ± 0.18 µM) than the reference inhibitor acetonyl-CoA (Ki = 1.32
± 0.24 µM). Molecular docking to hGcn5 and crystallographic data of p300 has shown that the alkyl chain
of CoA variants is positioned in the lysine binding site of the enzymes.42, 45 Structural analysis of p300
further revealed the presence of a hydrophobic binding pocket located at the catalytic center of the protein,
a structural feature absent in pCAF and MOF. This pocket was proposed to accommodate the alkyl portion
of structurally more demanding acyl-CoAs after substrate induced ligand rearrangement during the
catalysis of the acyl transfer by p300.45 To compare the effects of acyl-CoAs on pCAF(493-658) with
effects on p300(1284-1673), the ability of both enzymes to use acyl-CoA variants as cofactor for
enzymatic turnover of a peptide substrate was tested in a thiol scavenging assay that detects the acyl-CoA
hydrolysis product CoA (Supporting Information, Figure S6a). As expected, p300, but not pCAF, is
capable to use butyryl-CoA and hydroxybutyryl-CoA for enzymatic turnover, although the efficiency of
turnover is decreased compared to AcCoA. In the presence of lauroyl-CoA, no activity could be detected
within this assay. The profile of acyl-CoAs in the FP assay was then recorded for p300(1284-1673).
Interestingly, in contrast to pCAF, the degree of competition for binding to p300 did not show an initial
decrease with increasing chain length but increased continuously from C-2 to C-12 (Supporting
Information, Figure S6b). Dose-response analysis revealed that, despite being less effective cofactors,
p300(1284-1673) binds butyryl-CoA and lauroyl-CoA with significantly higher affinities than the
predominant cofactor AcCoA (Figure 6c). This observation is in agreement with inhibition data from the
FP competition activity immunoassay (Figure 6d). We then tested whether the increased affinity of
lauroyl-CoA could be attributable to additional interactions of the ligand with the peptide binding site of
the enzymes. Therefore, a peptide binding assay (Alpha) was performed, which showed considerable
competition of lauroyl-CoA, but not of the reference inhibitor or probe 11, for binding of His-tagged
pCAF to biotinylated H3(1-21) peptide (Supporting Information, Figure S6c). Together, our results
propose that the comparatively high affinity of lauroyl-CoA for pCAF(493-658) and p300(1284-1673) is
the result of additional hydrophobic interactions of the ligand with the substrate binding site of the
respective protein. In the case of pCAF, this site seems to be located distally to the lysine binding groove
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
14
ACS Paragon Plus Environment

Page 15 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
where the N- and C-terminal regions have been shown to form strong van der Waals contacts with the
physiological substrates.
7
8
9
Probing the cofactor binding site: We were interested in the possibility of using 11 to identify structural
features of the cofactor that are crucial for binding to the respective KAT enzyme. A comparison of such
features for all our tested KATs could possibly reveal differences in protein cofactor interactions, which
could prove useful in the design of more specific ligands targeting the CoA binding site. To do so, a set of
chemical entities containing different structural motifs of the cofactor was assembled and each compound
was tested at a fixed concentration of 1 mM against every KAT in the FP assay (Figure 6e). By this, we
were also able to validate the assay for its applicability to fragment-based approaches. The results from
this experiment are depicted in Figure 6. The strong interaction of CoA´s phosphate backbone with the
cofactor binding site has previously been shown for pCAF51, p30052, and MOF53 by crystallographic
means and our findings accordingly show a pronounced decrease in competition for fragments lacking one
or more phosphate moieties. For all tested KATs, the segmentation of the cofactor into two individual
fragment portions, the pantetheine part and the adenosine phosphate part, led to an almost complete loss of
interaction. Only the adenosine 3’,5’-diphosphate fragment exhibited modest competition with the probe
for binding to KAT proteins. For MOF, competition exceeding 20% was also observed for ADP, which
was completely inactive against pCAF and p300. Unphosphorylated pantethein did not exhibit any
significant competitive displacement of the fluorescent tracer, as it was proposed based on the findings of
previous studies on the interaction of pantethein analogues with acetyltransferases.54-55 A comparison of
competition values for the three KAT enzymes shows an overall better amenability of MOF to CoA-
fragment competition, which likely is due to the comparatively lower affinity of MOF(174-449) to 11. In
the case of p300(1284-1673), a strong effect on binding was observable for the 3’-phosphate residue, as
the cofactor fragment lacking only this moiety (dephospho-CoA) was not able to promote full competition
to 11, even at a concentration as high as 1 mM (CoA FP-assay IC50 = 3.0 ± 0.2 µM). Together, our
comparison of fragment competition enables an interaction mapping of the cofactor binding site that is in
good accordance with findings from structural biology approaches. These results suggest that the tested
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
15
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 49
1
2
3
4
5
6
7
8
9
KAT enzymes share a strong dependence of binding potency on the presence and precise orientation of
phosphate residues within the cofactor structure. Still, differences in the overall competition profile of
cofactor fragments between the different KAT enzymes could be observed in our FP approach,
demonstrating the applicability and versatility of 11 for this method.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Fragment-based ligand screening: To show the feasibility of the assay platform to screen larger quantities
of candidate ligands for binding to KAT protein, we performed a fragment-based approach. 1000 small
molecule fragments from a commercially available library (rule of 3 compliant) were evaluated in the FP
assay against pCAF(493-658). Compounds were first screened at 1 mM in duplicates and a statistic cut-off
of two times the mean standard deviation over all measurement was applied (Figure 7a). In a second
round, 62 initial hits were tested at three different concentrations (3 mM, 1 mM, and 0.3 mM). Fragments
that gave at least 30% inhibition at 3 mM and a two-fold increase in inhibition over the concentration
range were considered as hits. Autofluorescent compounds were discarded. The remaining 27 compounds
were then subjected to a fluorescent thermal stability assay and destabilizing fragments (ΔTm < -1 °C)
were sorted out (Figure 7b). Eventually, 14 fragments that showed a good response in the FP assay and no
protein destabilizing effects were evaluated for their competitive effects at 1 mM in the BLI assay (Figure
7c). Three fragments prevented binding of pCAF in this assay by more than 20% and their structures are
depicted in Figure 7d. These fragments share the structural motif of a five- or six-membered aromatic
heterocycle with a carboxylic acid residue in position 2, which indicates that we successfully enriched a
certain chemotype that showed effective competition to our probes for binding to the cofactor binding site
of pCAF(493-658) in two orthogonal assay systems.
16
ACS Paragon Plus Environment

Page 17 of 49
Journal of Medicinal Chemistry
1
2
3
4
DISCUSSION AND CONCLUSION
5
6
7
8
In the present study, we report the development of two cofactor-based chemical tool compounds,
fluorescently labeled compound 11 and biotin labeled compound 12, and their utilization in development
of a biophysical assay platform for the evaluation of candidate ligands targeting the KAT cofactor binding
site. The structures of these probes were deduced from the bisubstrate concept of covalently linking the
products of enzymatic turnover, CoA and an acetylated lysine mimic, via a non-hydrolysable thioether
linkage. By reducing the substrate mimicking moiety to a simple aminohexanoic acid residue and coupling
of the functional label to this site, the resulting compounds exhibit a broad applicability across KAT
enzymes originating from different enzyme families, while maintaining sufficiently high binding affinities
to be used in binding assays. While such tool compounds lack the cell permeability and selectivity to be
used in biological systems they provide a valuable means for in vitro evaluations. Recent reports of the
first potent and selective small molecule inhibitors of p300/CBP32 and MOZ/MORF33 activity, both
targeting the cofactor binging site of the respective enzyme, underline the potential of our cofactor
competition binding approach. The target compounds were synthesized via a convergent route with an
intermediate clickable handle that enables easy functionalization with a label of choice. We used the
fluorescence labeled probe 11 to set up a FP-based competition assay for the evaluation of active site
ligands. The relative increase in molecular volume upon binding and the conformational constraint of the
fluorophore are the main drivers for the generation of a sufficient dynamic assay window. While larger
peptide-based bisubstrate ligands could provide higher affinity and subtype selectivity, they are limited by
their structural flexibility at possible label attachment sites and high molecular weights. The
comparatively smaller size of 11 and the labelling in a position that places the fluorophore close to the
lysine binding site facilitated its use as a tracer in FP applications with sufficient signal intensities. After
successful implementation of the FP-assay, we optimized assay parameters for representative KAT
enzymes from each of the three major families The same was achieved with the affinity labeled cofactor
derivative 12 in a BLI-based setup. Utilizing the reference inhibitor acetonyl-CoA, these assays were then
used to record thermodynamic and kinetic binding data, which is in excellent agreement with previously
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
17
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 49
1
2
3
4
5
6
7
8
9
published values and data from ITC measurements, thus validating our probes and assays as sensitive and
versatile tools for the identification and characterization of cofactor competitive ligands. The combination
of binding assays with a system that measures enzyme activity provides an orthogonal means to assess the
functional consequences of cofactor competition by candidate ligands. For this approach, we reported a
non-radioactive FP competition activity immunoassay that does not rely on substrate labels, reactive
detection reagents, or functional surfaces, but allows free scaling of substrate and cofactor concentrations
and the use of whole protein substrates, which resemble a more physiological background. Although such
assay formats have been reported for kinases, to our knowledge, this is the first report of a FP competition
immunoassay measuring KAT activity to date. The application of the entire assay platform was further
validated by three different approaches. First, profiling of acyl-CoA variants revealed distinct binding
affinity patterns dependent on the alkyl chain length of the acyl moiety for pCAF and p300. The potent
binding of lauroyl-CoA to these enzymes was also shown to translate into inhibitory potency and for
pCAF we were able to rationalize our findings by demonstrating competition of the long chain acyl-CoA
to the peptide substrate, suggesting additional interactions with the protein at the substrate binding site.
Further, we used cofactor fragments to identify critical interaction sites and systematically tested structural
motifs of the physiological ligand to reveal subtype selective characteristics. 11 and 12 were finally used
to screen a commercially available fragment library in consecutive screening steps and we identified three
structurally analogous compounds that show profound and dose-dependent competition to the probes for
binding to pCAF, whilst not showing any disruption of protein integrity or assay interference. Such
screening hits provide excellent candidates for future cocrystallization and crystal soaking experiments to
develop the initial hits into new lead-like candidates by structure guided medicinal chemistry
approaches.56 Since the here reported tool compounds are likely promiscuous to other CoA utilizing
enzymes, our biophysical approach is presumably transferable to other in vitro setups. In summary, we
propose that our probes and assays are versatile and valuable tools for the in vitro evalutation of KAT
characteristics, which will prove useful for the identification and profiling of new drug candidates
targeting the cofactor binding site.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
18
ACS Paragon Plus Environment

Page 19 of 49
Journal of Medicinal Chemistry
1
2
3
4
EXPERIMENTAL SECTION
5
6
7
8
General: Starting materials (chemicals) were purchased from commercial suppliers and used without any
further purification. Solvents were used in p.a. quality and dried according to common procedures, if
necessary. Thin-layer chromatography (TLC) for reaction monitoring was performed with alumina plates
coated with Merck silica gel 60 F254 (layer thickness: 0.2 mm) or Merck silica gel 60 RP-18 F254 (layer
thickness: 0.2 mm) and analyzed under UV-light (254 nm). Flash column chromatography was performed
with TELOS Flash-LL Silica Columns 60M (0.040-0.063 mm, 230-400 mesh) as a stationary phase on a
Biotage Isolera One automated flash purification system with UV-Vis detector. Size exclusion
chromatography of compounds was conducted using Sephadex-G10 media and dest. H2O as the eluent.
All yields were not optimized. NMR spectra were recorded using a Bruker Avance III HD Spectrometer
(1H: 400 MHz, 13C: 100 MHz) instrument. The spectra are referenced against the NMR solvent and are
reported as follows: chemical shift δ (ppm), integration, multiplicity (s = singlet, d = doublet, dd = doublet
of doublets, t = triplet, q = quartet, p = quintet, m = multiplet), coupling constant (J in Hz), and
assignment. Mass spectra were recorded on a LCQ Advantage or Exactive device (Thermo Fisher
Scientific, Waltham, MA) or an Expression CMS device (Advion, Ithaca, NY). Purity was determined by
HPLC analysis using an Agilent 1260 Infinity series system with a Synergi Kinetex XB-C18 100 Å, 250 x
4.6 mm as the analytical column and UV detection (λ = 210 nm (HPLC-Method A) or 260 nm (HPLC-
Method B). A linear gradient of eluent A (water + 0.05 % (v/v) TFA) and eluent B (ACN + 0.05 % (v/v)
TFA) of 0–4 min, A=90%, B=10%; 4–29 min, linear increase to B=100%; 29–31 min, B=100%, 31–40
min A=90%, B=10% was applied at a constant flow rate of 1 ml·min-1. All synthesized compounds were
found to possess purities ≥ 95% by HPLC analysis. Peptides were purchased from PSL GmbH,
Heidelberg, Germany. p300(1284-1673) enzyme was purchased from Enzo Life Science, Farmingdale,
USA.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
PAINS analysis: Active compounds were tested for known classes of assay interference compounds with
the publicly available online tool “False Positive Remover” (www.cbligand.org). None of the tested
compounds were flagged as PAINS. In addition, all compounds reported in this study were characterized
19
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 49
1
2
3
4
5
6
in at least two individual assays with different readout techniques to provide orthogonal prove of their
specific activities.
7
8
Chemistry:
9
Method A: General procedure for the Cu-catalyzed Huisgen cycloaddition: Alkyne (1 equiv.) and azide
(1.1 equiv.) were dissolved in a water/tert-butanol mixture (2 ml, 1:1). TBTA (0.1 equiv.) was dissolved
in 1 ml DMF and added to the reaction mixture. An aqueous CuSO4 solution (0.1 M, 0.1 equiv.) and an
aqueous solution of sodium ascorbate (0.1 M, 0.2 equiv.) were added in that order. The resulting reaction
mixture was stirred for 12 h at room temperature under nitrogen atmosphere. After completion, the
reaction was quenched by the addition of 5 ml of H2O and the aqueous phase was extracted with ethyl
acetate (3x 15 ml). The product containing layer was collected and solvents were evaporated under
reduced pressure. The crude product was purified using automated column chromatography or semi
preparative HPLC.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Method B: General procedure for the nucleophilic halide substitution: To a solution of CoA trilithium salt
(1 eq.) in 0.1 M TEAB buffer pH 8.5/EtOH (4 ml, 1:1) at 0 °C was added dropwise chloroacetone (1.2
equiv.) in 2 ml EtOH. The solution was stirred under nitrogen atmosphere at 0 °C for 2 h and then allowed
to warm to room temperature and stirred overnight. The reaction was quenched by the addition of 10 ml
H2O and the aqueous phase was washed with ethyl acetate (3x 20 ml). The aqueous layer was separated
and solvents were evaporated under reduced pressure. The crude product was then purified using semi
preparative HPLC.
acetonyl-CoA: Method B using CoA trilithium salt (300 mg, 0.37 mmol) and chloroacetone (35 µl, 0.44
mmol). 255 mg (0.4 equiv. TEA), 80 % yield; 1H NMR (400 MHz, D2O) δ 8.65 (1H, s, H-17), 8.43 (1H, s,
H-20), 6.21 (1H, d, J = 5.4 Hz, H-16), 4.89 (2H, d, J = 5.3 Hz, H-14 and H-15), 4.60 (1H, s, H-13), 4.26
(2H, s, H-12), 4.01 (1H, s, H-7), 3.86 (1H, dd, J = 9.5, 4.9 Hz, H-11), 3.61 (1H, dd, J = 10.0, 4.9 Hz, H-
11), 3.52 (2H, s, CO-CH2-S), 3.46 (2H, t, J = 6.6 Hz, H-5), 3.31 (2H, t, J = 6.6 Hz, H-2), 2.60 (2H, t, J =
6.6 Hz, H-1), 2.46 (2H, t, J = 6.5 Hz, H-4), 2.28 (3H, s, CH3-CO), 0.94 (3H, d, J = 7.7 Hz, H-9 or 10),
0.81 (3H, s, H-9 or 10). HPLC purity (HPLC-Method A): 97.5% (tR = 4.72 min.);HRMS (ESI-): m/z calc.
20
ACS Paragon Plus Environment

Page 21 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
for C24H40N7O17P3S [M-H]- = 822.1341; [M+Li-2H]- = 828.0678; found: [M-H]- = 822.1332; [M+Li-2H]-
= 828.1414
6
7
8
9
tert-butyl(6-oxo-6-(prop-2-yn-1-ylamino)hexyl)carbamate (3): 6-tert-butylaminocaproic acid (2)
(2.169 g; 9.28 mmol) and HBTU (3.520 g, 9.28 mmol) were mixed in 50 ml dichloromethane and DIPEA
(4.85 ml, 27.8 mmol) was added. After stirring for 5 minutes, propargylamine (646 µl, 9.38 mmol) was
added dropwise to the mixture. The reaction was stirred overnight and the organic phase was washed with
saturated NaHCO3 solution, 1 M HCl and brine (3x 50 ml, each). The organic layer was then dried over
Na2SO4 and the solvent was removed by evaporation. Purification of the crude material was achieved by
automated column chromatography (cyclohexane/ ethyl acetate, 20-100 %) to obtain compound 5 as a
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
white solid. 1.668 g, 67 % yield; H NMR (400 MHz, methanol-d4): δ 3.94 (2H, d, J = 2.6 Hz, NH-CH2-
C≡CH), 3.02 (2H, t, J = 7.0 Hz, H-6’), 2.57 (1H, t, J = 2.6 Hz, C≡CH), 2.19 (2H, t, J = 7.5 Hz, H-2’), 1.62
(2H, p, J = 7.5 Hz, H-5’), 1.52 – 1.45 (2H, m, H-3’), 1.43 (9H, s, (C(CH3)3), 1.37 – 1.28 (2H, m, H-4’).
MS (ESI+): m/z calc. for C14H24N2O3 [M+Na]+ = 291.2; found [M+Na]+ = 291.4.
6-(2-chloroacetamido)-N-(prop-2-yn-1-yl)hexanamide (4): The Boc-group of propargylation product 3
(1.208 g, 4.50 mmol) was cleaved by addition of 15 ml TFA/dichloromethane and stirring for 2 hours.
Solvents were then evaporated and remaining TFA was coevaporated with methanol (5x). The residue was
then dissolved in 11 ml DMF and DIPEA (3.91 ml, 22.5 mmol) was added to the solution. After stirring
for 10 minutes, the mixture was cooled to 0 °C and chloroacetyl chloride (717 µl, 9.01 mmol) in 2 ml
DMF was added dropwise over a period of 30 minutes. The reaction was quenched after 2 hours by adding
50 ml of saturated NaHCO3 solution. Crude product 4 was extracted from the aqueous phase with DCM
(3x 50 ml) and the organic layers were combined and dried over Na2SO4 to give a black solid after
evaporation of the solvent. The residue was purified by means of column chromatography using a gradient
of methanol in dichloromethane (3-12 %). Product 4 was isolated as an off-white solid. 0.848 g, 77 %
1
yield; H NMR (400 MHz, methanol-d4): δ 4.02 (2H, s, NHCO-CH2-Cl), 3.94 (2H, d, J = 2.6 Hz, NH-
CH2-C≡CH), 3.22 (2H, t, J = 7.1 Hz, H-6’), 2.57 (1H, t, J = 2.6 Hz, C≡CH), 2.20 (2H, t, J = 7.5 Hz, H-2’),
21
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 49
1
2
3
4
5
6
1.63 (2H, p, J = 15.2, 7.5 Hz, H-5’), 1.54 (2H, p, J = 7.5 Hz, H-3’), 1.40 – 1.30 (2H, m, H-4’). MS (ESI+):
m/z calc. for C11H17ClN2O2 [M+Na]+ = 267.1; found [M+Na]+ = 267.0.
7
8
9
4-azidobutanoic acid (6): γ-aminobutanoic acid (5) (1.050 g, 10.18 mmol), imidazole-1-sulfonylazide *
HCl (2.565 g, 12.26 mmol), potassium carbonate (2.084 g, 15.08 mmol) and copper sulfate * 5 H2O (29.3
mg, 0.12 mmol) were dissolved in methanol and stirred overnight under nitrogen atmosphere. The solvent
was then removed by evaporation and the residue was taken up in demin. H2O and acidified to pH 2 using
HCl (5 M). The resulting suspension was extracted with ethyl acetate (3x 100 ml) and the organic layers
were combined and dried over Na2SO4. After removing the organic solvent by evaporation, the resulting
crude material was purified by automated column chromatography using a gradient of 1-5 % methanol in
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
dichloromethane to yield product 6 as colorless oil. 1.135 g, 86 % yield; H NMR (400 MHz, methanol-
d4): δ 3.36 (2H, t, J = 6.8 Hz, H-2’’), 2.39 (2H, t, J = 7.3 Hz, H-4’’), 1.86 (2H, p, J = 6.9 Hz, H-3’’). MS
(ESI+): m/z calc. for C4H7N3O2 [M+H]+ = 130.1; found [M+H]+ = 130.1.
4-azido-N-(3',6'-dihydroxy-3-oxo-3H-spiro[isobenzofuran-1,9'-xanthen]-6-yl)butanamide (7): 4-
azidobutanoic acid (6) (1.33 g, 8.0 mmol) was dissolved in 20 ml dichloromethane and 2.07 ml oxalyl
chloride (24.0 mmol) were added to this solution. After addition of one drop of DMF, the mixture was
stirred for 5 h. Evaporation of the solvents yielded the acyl chloride, which was directly used in the next
step without further purification. 42 mg (crude) of 4-azidobutanoyl chloride were dissolved in 10 ml THF
and dropwise added to the same amount of cooled solvent containing 0.9 equivalents of fluoresceinamine
(6-isomer, 103 mg, 0.3 mmol) and 10 equivalents of NaHCO3 (252 mg, 3 mmol). After complete addition
of the acyl chloride, the ice bath was removed. The mixture was stirred overnight and the reaction was
subsequently quenched by adding 15 ml of 5 % NaHCO3 aqueous solution. The organic solvent was
evaporated and the remaining solution was acidified to pH 2 by adding 5 M HCl. Ethyl acetate (3x 25 ml)
was added to the suspension to extract the product from the aqueous layer. After evaporation of organic
solvent, the product containing residue was purified by column chromatography using a gradient of 80-
100 % ethyl acetate in cyclohexane. The desired product 7 was isolated as an orange solid. 113 mg, 82 %
yield; 1H NMR (400 MHz, methanol-d4): δ 7.92 (1H, d, J = 8.5 Hz, H-8’’), 7.76 (1H, dd, J = 8.4, 1.8 Hz,
22
ACS Paragon Plus Environment

Page 23 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
H-9’’), 7.58 (1H, d, J = 1.4 Hz, H-11’’), 6.67 (2H, d, J = 2.4 Hz, H-17’’and H-20’’), 6.64 – 6.60 (2H, m,
H-14’’ and H-23’’), 6.54 (2H, dd, J = 8.7, 2.4 Hz, H-15’’ and H-22’’), 3.36 – 3.32 (2H, m, H-2’’), 2.45
(2H, t, J = 7.3 Hz, H-4’’), 1.87 (2H, p, J = 6.9 Hz, H-3’’). MS (ESI+): m/z calc. for C24H18N4O6 [M+H]+ =
459.1; found [M+H]+ = 459.2.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
6-(2-chloroacetamido)-N-((1-(4-((3',6'-dihydroxy-3-oxo-3H-spiro[isobenzofuran-1,9'-xanthen]-6-
yl)amino)-4-oxobutyl)-1H-1,2,3-triazol-4-yl)methyl)hexanamide (8): Method A using 4 (40 mg, 0.16
mmol) and 7 (80 mg, 0.18 mmol). 75 mg, 66 % yield; 1H NMR (400 MHz, methanol-d4): δ 7.90 (1H, d, J
= 8.4 Hz, H-8’’), 7.81 (1H, s, H-9’), 7.70 (1H, dd, J = 8.4, 1.7 Hz, H-9’’), 7.58 (1H, d, J = 1.5 Hz, H-
11’’), 6.67 (2H, d, J = 2.4 Hz, H17-‘’ and H-20’’), 6.62 (2H, d, J = 8.7 Hz, H-14’’ and H-23’’), 6.54 (2H,
dd, J = 8.7, 2.4 Hz, H-15’’ and H-22’’), 4.42 (2H, t, J = 6.8 Hz, H-4’’), 4.35 (2H, s, H-7’), 4.00 (2H, s,
NHCO-CH2-Cl), 3.18 (2H, t, J = 7.1 Hz, H-6’), 2.38 (2H, t, J = 7.1 Hz, H-2’’), 2.25 – 2.12 (4H, m, H-3’’
and H2’), 1.68 – 1.43 (4H, m, H-3’ and H-5’), 1.31 (2H, m, H-4’). MS (ESI+): m/z calc. for
C35H35ClN6O8 [M+Na]+ = 725.2105; found [M+Na]+ = 725.2097.
Fl-hex-CoA (11): Method B using CoA trilithium salt (40 mg, 0.05 mmol) and 8 (42 mg, 0.06 mmol). 34
mg, 48 % yield; 1H NMR (400 MHz, D2O): δ 8.46 (1H, s, H-17), 8.13 (1H, s, H-20), 7.79 – 7.73 (2H, m,
H-9’ and H-8’’), 7.48 (1H, dd, H-9’’), 7.22 (1H, d, J = 2.1 Hz, H-11’’), 7.07 (2H, d, J = 9.1 Hz, H14’’ and
H-23’’), 6.56 – 6.49 (4H, m, H-15’’, H-17’’, H-20’’ and H-22’’), 6.08 (1H, d, J = 6.9 Hz, H-16), 4.97 –
4.67 (2H, m, H-14 and H-15), 4.53 – 4.47 (1H, m, H-13), 4.39 (2H, t, J = 6.5 Hz, H-4’’), 4.22 – 4.11 (4H,
m, H-12 and H-7’), 3.91 (1H, s, H7), 3.76 – 3.70 (1H, m, H-11), 3.48 – 3.42 (1H, m, H-11), 3.39 – 3.30
(2H, m, H-2), 3.18 (2H, t, J = 6.6 Hz, H-5), 3.06 (2H, s, H-11’), 2.98 (2H, t, J = 6.9 Hz, H-6’), 2.50 (2H, t,
J = 6.7 Hz, H-1), 2.38 – 2.29 (4H, m, H-4 and H-2’’), 2.20 – 2.12 (2H, m, H3’’), 2.06 (2H, t, J = 7.3 Hz,
H-2’), 1.40 (2H,, dt, J = 14.4, 7.2 Hz, H-5’), 1.30 (2H, dt, J = 15.0, 7.1 Hz, H-3’), 1.12 – 1.02 (2H, m,
H4’), 0.77 (3H, s, H-9 or 10), 0.62 (3H, s, H-9 or 10).HPLC purity (HPLC-Method-B): 95.0% (tR = 12.30
min.); HRMS (ESI-): m/z calc. for C56H70N13O24P3S [M-3H]3- = 476.7791, [M+Na-3H]2- = 726.6632;
found [M-3H]3- = 476.7790; [M+Na-3H]2- = 726.6630
23
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 49
1
2
3
4
5
6
7
8
9
Biotin-PEGhex-CoA (12): Method B using CoA trilithium salt (41 mg, 0.05 mmol) and 4 (15 mg, 0.06
mmol). followed by Method A using the reaction product and 10 (45 mg, 0.06 mmol). 16 mg, 18 % yield
over 2 steps; 1H NMR (400 MHz, D2O) δ 8.62 (1H, s, H-17), 8.39 (1H, s, H-20), 8.07 (1H, s, H-9’), 6.17
(1H, d, J = 4.7 Hz, H-16), 4.97 – 4.84 (2H, m, H-14 and H-15), 4.68 – 4.49 (4H, m, H-13, H-7’’, and H-
11’’), 4.49 – 4.31 (3H, m, H-7’ and H-9’’), 4.24 (2H, s, H-12), 4.01 – 3.77 (4H, m, H-7, H-11, and H-
12’’), 3.67 – 3.50 (43H, m, H-11 and triazole-CH2-CH2-[O-CH2-CH2]10-O-CH2-CH2-NHCO), 3.42 (2H, t,
H-5), 3.36 – 3.21 (5H, m, H-2, H-6’’, and H-14’’), 3.19 (2H, s, H-11’), 3.11 (2H, t, J = 6.5 Hz, H-6’), 2.92
(1H, dd, J = 4.9, 13.1 Hz, H-10’’), 2.70 (1H, d, J = 13.2 Hz, H-10’’), 2.62 (2H, t, H-1), 2.42 (2H, t, H-4),
2.31 – 2.07 (4H, m, H-2’ and H-2’’), 1.71 – 1.16 (12H, m, H-3’, H-4’, H-5’, H-3’’, H-4’’, and H-5’’), 0.89
(3H, s, H-9 or 10), 0.78 (3H, s, H-9 or 10). HPLC purity (HPLC-Method B): 95.3% (tR = 11.40 min.);
HRMS (ESI-): m/z calc. for C66H116N15O31P3S2 [M+K-2H]- = 1808.6102; found [M+K-2H]- = 1808.5994
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Protein expression and purification:
E. coli BL21 (DE3) cells were transformed with a pQE-80L plasmid containing a 0.5 kb insert coding for
the catalytic domain of human pCAF protein (bp 1923-2420 with an N-terminal His-tag). Bacteria were
cultured in LB medium complemented with 100 mg/L ampicillin at 37 °C until an OD600 of 0.6-0.8. The
recombinant expression of His-tagged pCAF catalytic domain (aa 493-658) was induced with 1 mM
isopropyl-α-D-thiogalactopyranoside (IPTG) and incubated overnight at 18 °C. The cell suspension was
then centrifuged and the cell pellet was resuspended in 40 ml of lysis buffer (50 mM Tris (pH 8.0), 150
mM NaCl, 1 mM β-mercaptoethanol, and 10% glycerol (v/v) with protease inhibitors (cOmplete, Mini
Tablets, EDTA-free; Roche, Basel, CHE)). Cell lysis was carried out with the aid of sonification ( 5x 15
sec. with 1 min. pauses in between (Branson Sonifier II W-250, Heinemann, Schwäbisch Gmünd, GER))
and the resulting suspension was centrifuged (9000 rcf, 1h, 4 °C) to separate the cell debris from the
soluble protein fraction. Purification of the His-tagged protein was carried out on an automated
chromatography system (NGC Chromatography System, Bio-Rad Laboratories, Hercules, USA). The
clear supernatant after centrifugation was filtered and loaded on a 5 ml nickel-charged IMAC column
24
ACS Paragon Plus Environment

Page 25 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(HisTrap HP; GE Healthcare, Chalfont St Giles, UK). After washing off non-specifically bound proteins
with washing buffer (50 mM Tris (pH 8.0), 150 mM NaCl, 1 mM β-mercaptoethanol, 20 mM imidazole),
a second washing step at high salt concentration (washing buffer containing 2 M NaCl and no imidazole)
was included to remove residual proteins and nucleic acid contaminations. The desired protein was then
eluted using 300 mM imidazole in washing buffer and the protein containing fractions were collected,
pooled, and concentrated by ultrafiltration (Vivaspin® Turbo 4 concentrators, Sartorius, Göttingen, GER)
to a final volume of 2 ml. Further purification of the protein solution was achieved by gel filtration on a
HiLoad 26/60 Superdex 200 prep grade column (Pharmacia, New Jersey, USA) equilibrated with gel
filtration buffer (22 mM Na-citrate (pH 6.5), 330 mM NaCl). Collected fractions containing the
monodisperse enzyme were again pooled, concentrated by ultrafiltration, and glycerol (10% (v/v)) was
added. The purity of the preparation was checked by SDS-page gel analysis and the protein concentration
was determined by measuring the UV absorbance at 280 nm (Nanodrop 1000 Spectrometer, Thermo
Fisher Scientific, Waltham, USA). The enzyme solution was then aliquoted, flash frozen in liquid
nitrogen, and stored at -80 °C until used. E. coli Rosetta 2 (DE3) pLysS cells were transformed with a
modified pET15b vector coding for MOF(174-449) with an N-terminal His10-tag followed by a TEV
protease cleavage site. Overexpression was induced in LB media at an OD of ~0.5 by the addition of IPTG
(0.1 mM) and carried out at 18 °C overnight. After harvest, the bacterial pellets were resuspended in lysis
buffer (50 mM HEPES, 500 mM NaCl, 20 mM imidazole, 5 mM β-mercaptoethanol, 0.01 % (w/v)
taurodeoxycholate, pH 7.5) supplemented with cOmplete Mini EDTA-free protease inhibitor cocktail
tablets (Roche) and lysed via sonication. After removal of cell debris, the supernatant was applied to a
HisTrap HP column (1 mL, GE Healthcare, Freiburg), washed extensively with lysis buffer and then
eluted with an imidazole gradient (20-300 mM). Fractions containing MOF were pooled, concentrated and
purified using a HiLoad 16/60 Superdex 75 gel filtration column (GE Healthcare, Freiburg; 25 mM
HEPES, 500 mM NaCl, pH 7.5). MOF-containing fractions were pooled, concentrated and flash-frozen in
liquid nitrogen. All chromatography steps were performed on Äkta prime chromatography systems.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
25
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 49
1
2
3
4
5
6
Identity of MOF was verified by mass spectrometry and protein concentration was determined using the
Bradford assay.
7
8
9
Fluorescence Polarization Assay:
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Saturation binding experiment and KD determination: Recombinant acetyltransferase protein was serially
diluted in assay buffer (50 mM HEPES (pH 7.5), 150 mM NaCl, 0.1 mM EDTA, 0.01 % BSA (w/v), 0.01
% Tween-20 (v/v) (plus 1 mM DTT in the case of p300)) to a 2x intermediate concentration. From these
dilutions, 10 µl were mixed with an equal volume of a 16 nM Fl-hex-CoA solution in assay buffer on a
384-well assay plate (cat. no. 11866091, Greiner Bio-One, Kremsmünster, AUT) (20 µl final volume
containing 8 nM Fl-hex-CoA and appropriate concentrations of corresponding acetyltransferase protein
(0-20 µM for His-pCAF(493-658), 0-3 µM for MOF(174-449) and 0-0.75 µM for p300(1284-1673))). The
plate was spun down, incubated, and measured using a 2102 EnVision® Multilabel Reader (PerkinElmer,
Waltham, USA) operating in FP-mode (ex. 480 nm, em. p-plane 535 nm, em. s-plane 535 nm). All tested
protein concentrations were also measured after incubation with 8 nM fluorescent probe Fl-hex-CoA in
the presence of competing ligand acetonyl-CoA (200 µM final concentration), to simulate a situation,
where the probe is completely displaced from the protein. In addition, the fluorescence intensities for p-
plane and s-plane were measured for protein dilutions in assay buffer without Fl-hex-CoA to detect any
intrinsic fluorescence not attributable to the fluorescent probe. The measured I∥ and I∟ values were
corrected for intrinsic protein fluorescence and the fluorescence polarization (P) was calculated for every
protein concentration incubated with 8 nM Fl-hex-CoA in the presence (PI) or without (PM) competing
ligand (100 µM acetonyl-CoA) using the following equation:
퐼 ∥ ― 퐺 ∗ 퐼∟
푃 =
∗ 1000 [푚푃]
퐼 ∥ + 퐺 ∗ 퐼∟
In this equation, G is an instrument specific constant. The fluorescence polarization values for samples
containing no protein (PL) or the highest protein concentration (PL*R), respectively resemble situations
where the fluorescent probe is either free in solution or fully bound to the receptor. From these values, the
fraction of bound Fl-hex-CoA (FB) can be calculated for every protein concentration with:
26
ACS Paragon Plus Environment

Page 27 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
푃푀 ― 푃퐿
퐹퐵 =
푃퐿 ∗ 푅 ― 푃퐿
6
7
8
9
The contribution of non-specific binding (PNS) was then determined for every protein concentration using
the following equation:
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
푃푁푆 = (푃퐼 ― 푃퐿) ∗ (1 ― 퐹퐵)
The specific fluorescence polarization signals (PS) were calculated by subtracting the non-specific portions
from PM values. The saturation binding curve was then generated by plotting PS values against protein
concentrations and the KD value for the interaction of Fl-hex-CoA with acetyltransferase protein was
calculated by fitting the corresponding results to a Hill equation using OriginPro 9 software.
Ligand evaluation: Test compounds were appropriately diluted in DMSO or H2O to a 20x intermediate
concentration. Resulting solutions were then mixed with an equal volume of assay buffer and 2 µl of the
resulting dilutions were spotted into the wells of a black 384-well non-binding microplate. For positive
controls (PP), 2 µl of the solvent/assay buffer mixture without test compound were used. Next, 8 µl of a 10
nM Fl-hex-CoA solution were dispensed into each well, followed by 10 µl of acetyltransferase protein in
assay buffer (PC). Negative controls (PN) included 10 µl of assay buffer instead of protein solution and
acetonyl-CoA (50 µM final concentration) was used as a reference inhibitor. Plates were then sealed with
adhesive foil (TopSeal®-A, PerkinElmer, Waltham, USA), spun down for 1 min. at 300 rcf and incubated
at room temperature. Upon completion, the fluorescence intensity signals for p-plane and s-plane (I∥ and
I∟) were read using a 2102 EnVision® Multilabel Reader (PerkinElmer, Waltham, USA) operating in FP-
mode (ex. 480 nm, em. p-plane 535 nm, em. s-plane 535 nm). Results were blank corrected and the
fluorescence polarization for every well was calculated as described above. Competition values (C) were
calculated from the averages of controls (PP and PN) and determined PC values with the use of the
following equation:
(푃퐶 ― 푃푃)
퐶 =
∗ 100 [%]
(푃푁 ― 푃푃)
27
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 49
1
2
3
4
5
6
7
8
Calculation of IC50 values and optical visualization were done with OriginPro 9 software by plotting
competition values against logarithmic compound concentrations and applying a logistic 4 parameter
sigmoidal fit with variable slope.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Biolayer Interferometry Assay:
KD value determination for the interaction of His-pCAF(493-658) and Biotin-PEGhex-CoA: Streptavidin
coated biosensors (cat. no. 18-5019, Pall FortéBio, Menlo Park, USA) were hydrated in assay buffer (50
mM HEPES (pH 7.5), 300 mM NaCl, 0.1 mM EDTA, 0.1% BSA (w/v), 0.01% Tween-20 (v/v) (plus 1
mM DTT in the case of p300)) and loaded with biotinylated probe from a 125 nM Biotin-PEGhex-CoA
solution placed in the drop holder. The loaded sensor tips were then dipped into solutions of serially
diluted His-pCAF(493-658) (0.04-30 µM) and the association of protein to the sensor tip was recorded for
120 sec. from the dropholder position on a BLItz® system (Pall FortéBio, Menlo Park, USA).
Subsequently, the sensor tips were exposed to assay buffer and the dissociation of protein from the
sensortip was recorded for 120 sec. from the tube position. A control run without enzyme was recorded for
association and dissociation and the measured sample curves were referenced against this control. Global
fitting of this data (for His-pCAF(493-658) concentrations spanning 0.04-3.33 µM) to a 1:1 binding model
using the BLItz Pro 1.2 software gave the rate constants of association (ka) and dissociation (kd), as well as
the affinity constant (KD) for the interaction of His-pCAF(493-658) and Biotin-PEGhex-CoA. In a
different attempt, report points of the association signal after 140 sec. total measurement time were plotted
against His-pCAF(493-658) concentration and the KD value was calculated by fitting the data to a Hill
equation using OriginPro 9 software.
Ligand evaluation: To evaluate candidate ligands for competition to Biotin-PEGhex-CoA for binding to
His-pCAF(493-658), compounds were appropriately diluted in DMSO or H2O and 1 µl of each compound
dilution was mixed with 19 µl acetyltransferase protein in assay buffer (final concentrations: 250 nM for
His-pCAF(493-658) , 5 µM for MOF(174-449)). Positive controls received 1 µl of pure solvent (DMSO
or H2O). The mixtures were incubated for 30 min. at room temperature. At the same time, streptavidin
28
ACS Paragon Plus Environment

Page 29 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
coated biosensors were hydrated in assay buffer and subsequently loaded with biotinylated capture ligand
by exposing the sensor tips to a 125 nM solution of Biotin-PEGhex-CoA for 120 sec. in the drop holder
position. After loading, a 30 sec. baseline was generated in assay buffer from the tube position.
Preincubated protein solutions were then transferred to the drop holder in 4 µl volume and the association
of protein to the immobilized capture ligand was measured for 120 sec. from these solutions, followed by
a 120 sec. dissociation step in a tube filled with assay buffer. Recorded binding curves were referenced
against a control run in assay buffer in the absence of protein. The binding signals at report points after
110 sec. of association (BM) were used to calculate the degree of competition (C) for every ligand
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
concentration compared to positive controls (BP) by using the following equation:
퐵푃 ― 퐵푀
퐶 = 100 ∗
[%]
퐵푃
The calculated competition values were plotted against logarithmic compound concentrations and the data
was fit to a four parameter logistic fit with variable slope using OriginPro 9 software.
For experiments on p300(1284-1673), 500 nM enzyme in assay buffer were incubated with 125 nM
Biotin-PEGhex-CoA and appropriately diluted candidate ligands for 60 min. at room temperature in 20 µl
volume. Fresh biosensor tips were hydrated and a 30 sec. baseline was recorded from assay buffer.
Biosensors were then dipped into 4 µl of these mixtures and the binding signal was measured for 120 sec.
from the drop holder position. Competition values were determined as described above.
Fluorescence Polarization Competition Activity Immunoassay:
Experiments were performed in assay buffer (50 mM HEPES (pH 7.5), 150 mM NaCl, 0.1 mM EDTA,
0.01 % BSA (w/v), 0.01 % Tween-20 (v/v) (+ 1 mM DTT for experiments with p300) in 25 µl final
volume on black 384-well non-binding microplates (cat. no. 11866091, Greiner Bio-One, Kremsmünster,
AUT). After dispensing of assay components, plates were spun down for 1 min. at 300 rcf, sealed with
adhesive foil (TopSeal®-A, PerkinElmer, Waltham, USA), and incubated at room temperature. Blank
controls containing 25 µl assay buffer were included in all measurements. Combinations of specific
antibodies and acetylated fluorescent tracer peptides were utilized as detection reagents and the
29
ACS Paragon Plus Environment