Bromodomains (BD) are readers of lysine acetylation marks present in numerous proteins associated with chromatin. Here we describe a dual inhibitor of the bromodomain and PHD finger (BRPF) family member BRPF2 and the TATA box binding protein-associated factors TAF1 and TAF1L. These proteins are found in large chromatin complexes and play important roles in transcription regulation. The substituted benzoisoquinolinedione series was identified by high-throughput screening, and subsequent structure-activity relationship optimization allowed generation of low nanomolar BRPF2 BD inhibitors with strong selectivity against BRPF1 and BRPF3 BDs. In addition, a strong inhibition of TAF1/TAF1L BD2 was measured for most derivatives. The best compound of the series was BAY-299, which is a very potent, dual inhibitor with an IC50 of 67 nM for BRPF2 BD, 8 nM for TAF1 BD2, and 106 nM for TAF1L BD2. Importantly, no activity was measured for BRD4 BDs. Furthermore, cellular activity was evidenced using a BRPF2- or TAF1-histone H3.3 or H4 interaction assay.
Bromodomains (BD) are readers of lysine acetylation marks present in numerous proteins associated with chromatin. Here we describe a dual inhibitor of the bromodomain and PHD finger (BRPF) family member BRPF2 and the TATA box binding protein-associated factors TAF1 and TAF1L. These proteins are found in large chromatin complexes and play important roles in transcription regulation. The substituted benzoisoquinolinedione series was identified by high-throughput screening, and subsequent structure-activity relationship optimization allowed generation of low nanomolar BRPF2 BD inhibitors with strong selectivity against BRPF1 and BRPF3 BDs. In addition, a strong inhibition of TAF1/TAF1LBD2 was measured for most derivatives. The best compound of the series was BAY-299, which is a very potent, dual inhibitor with an IC50 of 67 nM for BRPF2 BD, 8 nM for TAF1BD2, and 106 nM for TAF1LBD2. Importantly, no activity was measured for BRD4 BDs. Furthermore, cellular activity was evidenced using a BRPF2- or TAF1-histone H3.3 or H4 interaction assay.
Proteins involved in
chromatin modulation
usually contain several
functional regions that serve to recruit and tether binding partners,
ultimately leading to the formation of large multiprotein complexes.
A clear understanding of the function of these complexes will only
be possible through the dissection of the interactions between the
individual protein partners and the analysis of the exact role of
their interactive modules. Different approaches are currently being
used to address this. One method is the selective deletion of gene
exons by CRISPR-Cas9 followed by phenotypic analyses.[1] Another, complementary approach is the identification of
specific domain inhibitors and evaluation of their functional impact
in cellular, and in some cases, animal models.[2,3] Here,
great advances have recently been reported in the area of bromodomains
(BDs), the best-characterized readers of acetyl-lysine (KAc) marks
which are found on histones and other proteins.[2−4] BDs are potentially
amenable to small molecule inhibition, as impressively shown in the
case of bromo and extra-terminal (BET) BD inhibitors which have been
extensively used in recent years to understand the role of this family
in normal physiology and pathology, ultimately paving the way for
their clinical evaluation in cancer indications.[3−6] In the wake of this success, several
other BDs have been successfully targeted by inhibitors with very
different scaffolds, further highlighting the druggability of this
target family.[7−14]The bromodomain and PHD finger (BRPF) family encompasses three
paralogs, BRPF1, BRPF2, and BRPF3, which are found in histone acetyltransferase
(HAT) complexes.[15] BRPFs possess different
functional modules, including C2H2, plant homeodomain-linked fingers,
bromo, and PWWP domains. Mouse studies show that loss of either BRPF1
or BRPF2 leads to embryonic lethality at E9.5 and E15.5, respectively,
due to different developmental defects.[16,17] In sharpcontrast,
BRPF3 is not needed for mouse development or survival, and no compensatory
expressionchange is observed for BRPF1 or BRPF2, in line with nonoverlapping
functions.[18] BRPF1 is involved in hematopoietic
stem cell development and BRPF2 in embryonic stem cell differentiation.[16,19] Little is known about the implication of BRPFs in disease, but a
critical role of BRPF1 in the leukemia subgroup harboring the MOZ-TIF
fusion has been reported.[20] First BRPF
BD inhibitors have been published (Figure ). Compounds 1 (OF-1)[21] and 2 (NI-57)[10] are pan-inhibitors recognizing all three BRPF paralogs
and derived from two distinct structure classes, benzimidazolones
and quinolinones, which make them valuable complementary chemical
probes. Two additional inhibitors, 3 (PFI-4) and 4 (close analog of IACS-9571), as well as (R)-N-[1,3-dimethyl-6-(2-methylpiperazin-1-yl)-2-oxo-2,3-dihydro-1H-benzo[d]imidazol-5-yl]-2-methoxybenzamide
(GSK6853) and N-[2,3-dihydro-1,3-dimethyl-2-oxo-6-(1-piperidinyl)-1H-benzimidazol-5-yl]-2-methoxybenzamide (GSK5959), are derived
from the 1,3-dimethyl-benzimidazolone scaffold and exhibit potent
and selective binding to BRPF1.[22−24] Compound 4 shows
additional strong activity on TRIM24, a nuclear receptor cofactor
associated with different cancer types,[25] being therefore a dual BRPF1/TRIM24 inhibitor.[21,23] Altogether, these findings demonstrate the high usefulness of the
1,3-dimethylbenzimidazolone headgroup as a KAc mimetic binding motif
and also suggest that the addition of novel decorations to this scaffold
may result in a different selectivity profile.[21,22,24] To the best of our knowledge, however, no
compounds showing selectivity for BRPF2 BD over its two paralogs have
yet been reported.
Figure 1
Comparison of 5 with previously described
BRPF inhibitors.
Comparison of 5 with previously described
BRPF inhibitors.Several TATA-binding
protein-associated factors (TAFs) have been
identified in the basal transcription factor TFIID. TAF1 is the largest
subunit of the complex and represents the essential scaffold to which
the other TAFs and the TATA-binding protein attach.[26] Recent structural data indicate that TAF1 makes essential
interactions with the core promoter DNA.[27] It contains multiple modules, including a kinase motif, a ubiquitin-conjugating
activity, a HAT region, and two related BDs. This modular organization
is also found in the highly related paralog TAF1L, which functionally
replaces TAF1 in male germ cells.[28] Mutations
in the TAF1 gene are associated with neurological pathologies in affected
males.[29,30] However, little is known about the implication
of individual TAF1 domains in these diseases or about a role of TAF1
in cancer. The recent progress reported in the area of BD inhibitors
suggests that addressing this module may be a good starting point
for functional studies. Indeed, a first paper describing a BET inhibitor
with additional submicromolar TAF1BD2 activity, named UMB-32, has
recently been presented.[31] Thiscompound
is derived from an imidazo-pyrazine scaffold bearing a dimethyl-isoxazole;[32−36] however, its usefulness is restricted by its limited on-target activity
and cross-reactivity with BET BDs, which are known to exert multiple
cellular effects.[2,13,37,38] Very recently, TAF1 inhibitors derived from
a methyl pyrrolopyridone fragment were described.[39,40] The most selective example interacts mainly with TAF1 and TAF1LBD1, but also with BRD9, BRD7, and BET BDs.Here, we present
the results of a screening and optimization strategy
which allowed discovery of the 1,3-dimethylbenzimidazolone-substituted
benzoisoquinolinedione 5 (BAY-299),[41] which exhibits potent dual inhibitory activity for both
BRPF2 BD and TAF1/TAF1LBD2 in biochemical and cellular assays. Thiscompound is the first reported BRPF2 inhibitor with high selectivity
against the BRPF1 and BRPF3 paralogs. It represents a powerful chemical
tool suited for investigation of the role of two little-explored BD-containing
proteins. Importantly, 5 is inactive against BET BDs,
so that the generated cellular results will not be blurred by residual
interference with BRD4 or its paralogs, which are known to possess
numerous biological functions.[38]
Results
and Discussion
Screening and Hit Identification
Our original aim was
to identify inhibitors of BRPF BDs with cellular activity and without
any activity on BET BDs. High-throughput screening (HTS) of a 3.5-million
compound library was performed using a BRPF1 time-resolved fluorescence
resonance energy transfer (TR-FRET) assay (Figure A). Confirmed hits were evaluated by dose–response
analysis (IC50 determination) in BRPF1, BRPF2, and BRPF3
TR-FRET (Figure A).
BRD4 TR-FRET was used as counter screen. This exercise yielded several
overlapping hit clusters with nanomolar pan-BRPF BD activities and,
surprisingly, allowed the identification of a single hit (6, Figure A) containing
a 1,3-dimethylbenzimidazolone scaffold[22] with nanomolar activity for BRPF2 while being inactive up to concentrations
of 20 μM for BRPF1 (Figure B), BRPF3 (Figure C), and BRD4 (Figure D). The binding specificity and selectivity profile
of the compound were confirmed by thermal shift assays (TSA), in which
hit 6 stabilized BRPF2 with a ΔTm of 0.8 K (Figure E), whereas only a ΔTm of
0.2 K was measured for BRPF1 (not shown). After confirmation of sample
identity and based on its favorable binding efficiency index and lipophilic
ligand efficiency (BEI: 17.3; LLE: 3.7), hit 6 was selected
as a starting point for further optimization cycles using additional
assays (Figure A).
Figure 2
Screening
strategy for the discovery of 6 and identification
of 5. (A) Overview of the HTS campaign leading to 6, and of the optimization process which culminated with the
identification of the probe. (B–D) Potency correlation plots
of HTS hits showing the BRPF2 selectivity of 5: TR-FRET
IC50 values for BRPF2 (x-axis) compared
with BRPF1 (B), BRPF3 (C), and BRD4 (D) IC50 values. Symbols
corresponding to 6 are shown in black. (E) TSA melting
curve of BRPF2 in the presence of 6 compared to vehicle.
Screening
strategy for the discovery of 6 and identification
of 5. (A) Overview of the HTS campaign leading to 6, and of the optimization process which culminated with the
identification of the probe. (B–D) Potency correlation plots
of HTS hits showing the BRPF2 selectivity of 5: TR-FRET
IC50 values for BRPF2 (x-axis) compared
with BRPF1 (B), BRPF3 (C), and BRD4 (D) IC50 values. Symbols
corresponding to 6 are shown in black. (E) TSA melting
curve of BRPF2 in the presence of 6 compared to vehicle.
Chemistry
One
structural feature of 6 is
the tricyclic decoration of the benzimidazolone at position 5 (naphthalimide
substituent). Surprisingly, the first analyses revealed that 6 had a logD of 2.1 and an excellent permeability profile
in Caco2cells; furthermore, a low solubility (3.7 mg/L) was measured,
which was not unexpected. Docking studies revealed that the naphthalimide
moiety was responsible for the strong BRPF2 selectivity, as it forms
key interactions which are not possible with BRPF1 (see Figure cocrystal structure discussion).
That this bulky substituent is indeed essential was confirmed by our
early attempts to reduce its molecular size or aromatic architecture,
which always led to less potent and less selective compounds (data
not shown). Based on this binding mode hypothesis, we initiated an
optimization program around the BRPF2-selective HTS hit 6.
Figure 3
Binding mode of 7 in BRPF2 BD and basis for selectivity.
(A) Cartoon representation of BRPF2 BD domain (white) with key interacting
side-chains shown in stick representation with C, N, and O atoms colored
white, blue, and red, respectively. Compound 7 is also
shown in stick representation with C, N, and O atoms colored green,
blue, and red, respectively. Hydrogen-bonding interactions between 7 and the protein are shown as red dotted lines. (B) Overlay
of the BRPF2–7 complex structure with selected
residues from the BRPF1–apostructure (PDB 4LC2) (C, N, and O atoms
colored dark gray, blue, and red, respectively). BRPF1 BD Pro658 is unable to form a key hydrogen bond with 7. (C) Overlay of the BRPF2–7 complex structure
with selected residues from the BRD4–apostructure (PDB 2OSS) (C, N, and O atoms
colored dark gray, blue, and red, respectively).
Binding mode of 7 in BRPF2 BD and basis for selectivity.
(A) Cartoon representation of BRPF2 BD domain (white) with key interacting
side-chains shown in stick representation with C, N, and O atoms colored
white, blue, and red, respectively. Compound 7 is also
shown in stick representation with C, N, and O atoms colored green,
blue, and red, respectively. Hydrogen-bonding interactions between 7 and the protein are shown as red dotted lines. (B) Overlay
of the BRPF2–7 complex structure with selected
residues from the BRPF1–apostructure (PDB 4LC2) (C, N, and O atoms
colored dark gray, blue, and red, respectively). BRPF1 BD Pro658 is unable to form a key hydrogen bond with 7. (C) Overlay of the BRPF2–7 complex structure
with selected residues from the BRD4–apostructure (PDB 2OSS) (C, N, and O atoms
colored dark gray, blue, and red, respectively).
Structure–Activity Relationships
We first focused
on improving in vitro BRPF2 activity. Therefore,
the influence of substituents at position 6 of the 1,3-dimethylbenzimidazolonecore was studied (Table ). A significant, 4-fold improvement of the BRPF2 activity was achieved
by introducing small groups which lock the central dihedral angle
in its bioactive conformation; however, this was also accompanied
by an increased interaction with BRPF1. The best results were obtained
for the methyl- and methoxy-derivatives (compounds 7 and 8) which offered the highest selectivity for BRPF2 over BRPF1.
Compounds 9 and 10 were also potent BRPF2
inhibitors but less selective toward BRPF1. Addition of the larger
methylpiperazine moiety (compound 11) did not lead to
an activity improvement. Importantly, all compounds demonstrated a
high selectivity against BRD4 BDs. Furthermore, a strong activity
was also observed for TAF1BD2.
Table 1
SAR of Position 6
for BRPF2 BD, BRPF1
BD, TAF1 BD2, and BRD4 BD1 Activity
Determined using
a TR-FRET assay.
All compounds
show an IC50 > 20000 nM for BRD4 BD2; n.a. = not available.
Compounds are assumed to be
racemates.
Determined using
a TR-FRET assay.All compounds
show an IC50 > 20000 nM for BRD4BD2; n.a. = not available.Compounds are assumed to be
racemates.The cellular
activity of selected compounds was then determined
using the NanoBRET assay[42] to measure the
interaction between the BRPF BD fused to the NanoLuc luciferase and
full-length histone H4 fused to the HaloTag (Supporting
Information Table S1). Compound 7 inhibited the
interaction of BRPF2 BD and histone H4 with an IC50 of
470 nM, whereas no effect on the BRPF1 BD–histone H4 interaction
was observed (IC50 > 10000 nM). Compound 8 was also highly selective for BRPF2 BD (IC50 = 1500 nM),
compared to BRPF1 BD (IC50 > 10000 nM). In addition,
inhibition
of the TAF1BD2–histone H4 interaction was also measured (Supporting Information Table S1).The cocrystal
structure of 7 with BRPF2 BD revealed
the binding mode within the KAc binding pocket (Figure A). The expected hydrogen bond is formed
between the carbonyl moiety of the benzimidazolone backbone of 7 and residue Asn642 as well as the π-stacking
with the Phe648. As predicted by our in silico analyses, a second hydrogen bond of the carbonyl moiety of the naphthalimide
moiety to the backbone nitrogen of Ser592 is formed, which
is key for the BRPF1 vs BRPF2 selectivity profile. A proline residue
(Pro658) at the equivalent position in BRPF1 is unable
to form thishydrogen bond (Figure B). The crystal structure also revealed that the Gln589 side-chain in BRPF2 unexpectedly adopted an alternative
conformation in the presence of 7, forming an additional
Van-der-Waals-type interaction by stacking underneath the tricyclic
ring system (Figure A). Overlay with the BRD4 apostructure revealed a number of key residue
differences that contribute to the strong selectivity of our compounds
(Figure C).In order to enrich our structure–activity relationship (SAR)
knowledge, we then focused on the naphthalimide moiety and explored
two different positions (5′ and 6′) (Table ). For compounds bearing an
electron-withdrawing group at position 6′ (compounds 12 and 13), a low, two-digit nanomolar biochemical
potency for BRPF2 was reached. Aside from 15, which displayed
an IC50 of 63 nM for BRPF2, analogues with a modification
at position 5′ (compounds 14, 16–18) displayed no significant improvement over 7. Except for 14, all modified compounds possessed strong
additional TAF1BD2 in vitro activity and unchanged
selectivity against BRPF1. The best TAF1BD2 IC50 values
were obtained for 12 and 13, which possess
an electron-withdrawing group at position 6′.
Table 2
SAR of Positions 5′ and 6′
for BRPF2 BD, BRPF1 BD, TAF1 BD2, and BRD4 BD1 Activity
Determined using a TR-FRET assay.
All compounds show an IC50 > 20000
nM for BRD4 BD2.
Compounds
are racemates.
Determined using a TR-FRET assay.All compounds show an IC50 > 20000
nM for BRD4BD2.Compounds
are racemates.Encouraged
by these data, we characterized 13 in more
detail. While cellular activity in the NanoBRET assay was determined
as IC50 675 nM for BRPF2, there was no activity for BRPF1
(Supporting Information Table S1). Inhibition
of the TAF1BD2–histone H4 interaction was furthermore observed
(Supporting Information Table S1). In addition,
the selectivity of 13 was assessed using TSA with a panel
of 48 BDs (Supporting Information Table
S2). The strongest stabilization was observed for BRPF2 (ΔTm = 6.2 K), and weaker stabilization was seen
for TAF1BD2, TAF1LBD2, CECR2 BD, and adenoviral E1A binding protein
(EP300) BD. As the low solubility of 13 limited further
characterization, we elected to optimize this parameter by tuning
the polarity via modifications at position 6′ of the naphthalimide
moiety (Table ). A
range of amines with diverse degrees of basicity were introduced by
nucleophilic aromatic substitution (SNAr) or Buchwald–Hartwig
coupling to provide 19–28 (Table ). A significant increase
in the thermodynamic solubility from DMSO solution with maintenance
of the selectivity was indeed observed for 21, 22, and 26–28. One acyclicamine displaying high selectivity for BRPF2 over BRPF1 BD (118-fold)
and strong TAF1BD2 activity could also be synthesized (compound 26). Interestingly, synthesis of its corresponding C-substituted
terminal alkoxy analogues (5 and 29) led
to more selective compounds. We opted for further investigation of 5 (BEI: 16.72, LLE: 4.72), since it showed the strongest selectivity
and cellular activity for BRPF2 with a reasonable solubility and a
logD of 2.0. The physicochemical profile and characteristics of 5 are summarized in the Supporting Information (Table S3). Overall, these SAR investigations revealed a high tolerability
to amines with respect to the BRPF2 BD or TAF1BD2 in vitro activity.
Table 3
Selected SAR: Optimizing Solubility
Determined using a TR-FRET assay.
Solubility from DMSO solution.
Determined using NanoBRET.
All compounds are racemates.
All compounds show an IC50 > 15000
nM for BRD4 BD2; n.a. = not available; n.c. = no calibration
possible.
Determined using a TR-FRET assay.Solubility from DMSO solution.Determined using NanoBRET.All compounds are racemates.All compounds show an IC50 > 15000
nM for BRD4BD2; n.a. = not available; n.c. = no calibration
possible.
General Synthesis of 1,3,6-Trimethylbenzimidazolone-Substituted
Naphthalimides
The 1,3,6-trimethylbenzimidazolone-substituted
naphthalimides were synthesized via a linear and straightforward reaction
pathway, as outlined in Scheme . Condensation of amine 30 with various commercially
available 1,8-naphthalic anhydrides resulted in 7, 12–14, and 16–18. Subsequent reaction of aryl bromide 12 with
the appropriate amines via nucleophilic aromatic substitution provided 19–25, 27, and 28. Compound 26 was synthesized from aryl bromide 12 and propylamine via Buchwald–Hartwig amination.
Sonogashiracoupling of aryl bromide 12 with propargyl
alcohol or but-3-yn-1-ol, followed by hydrogenation of the resulting
coupled alkynes with palladium on carbon, resulted in 5 and its analogue 29.
Scheme 1
General Route for
the Synthesis of 1,3,6-Trimethylbenzimidazolone-Substituted
Naphthalimides
a) 1,8-naphtalic substituted anhydrids, CH3COOH, rfx; b) amine, potassium carbonate, DMSO, 110 °C; c) amine, cesium carbonate, tris(dibenzylideneacetone)dipalladium, Xantphos, toluene, 60 °C d) alkyl-alcohol, Et3N, bis(triphenylphosphine)palladium(II) chloride, CuI, THF, 65 °C; e) H2, Pd/C, EtOAc, rt.
General Route for
the Synthesis of 1,3,6-Trimethylbenzimidazolone-Substituted
Naphthalimides
a) 1,8-naphtalic substituted anhydrids, CH3COOH, rfx; b) amine, potassium carbonate, DMSO, 110 °C; c) amine, cesium carbonate, tris(dibenzylideneacetone)dipalladium, Xantphos, toluene, 60 °C d) alkyl-alcohol, Et3N, bis(triphenylphosphine)palladium(II) chloride, CuI, THF, 65 °C; e) H2, Pd/C, EtOAc, rt.
Atropisomerism
Introduction of the methyl group at
position 6 of the benzimidazolonecore of the substituted naphthalimides
hinders the rotation around the central torsion (axial chirality).[43] Compound 5 is therefore a racemic
mixture of the two atropisomers shown in Figure A. Several compounds containing C–N-coupled
heterobiaryls with related atropisomerism have already been described.[44−46] The calculated rotational energy barrier of 25 kcal/mol for a representative
fragment (Figure B)
classifies the atropisomers of 5 as “class two”.[47] For thisclass of atropisomers, interconversion
might occur depending on the medium the compound is exposed to.[47] In the case of 5; the atropisomers 5a and 5b could be separated using chiral high-performance
liquid chromatography (HPLC) and individually characterized (Table ). Compound 5 and its individual atropisomers showed similar biochemical
potency, selectivity, and cellular NanoBRET activity for BRPF BD,
TAF1BD2, and BRD4 BDs. Furthermore, the isothermal titration calorimetry
(ITC) and Caco2cell permeability profiles were comparable. Stirring
experiments with the individual atropisomers revealed slight racemization
in acetonitrile and DMSO after 1 week at 37 °C, and 30% racemization
after 2 days at 80 °C (data not shown). The racemate 5 was therefore selected as the chemical probe.
Figure 4
Atropisomerism. (A) Atropisomers 5a and 5b of 5. (B) Energy profile
of the central torsion for
a 6′-methyl- (solid lines) and unsubstituted (dashed lines)
phenylnaphthalimide representative fragment. Optimization was performed
at the B3LYP/6-31G* level followed by single-point LMP2/cc-pVTZ(-f)
calculations.[54]
Table 4
Comparison of 5 with
Its Atropisomersa
Compound 5
atropisomer 1
atropisomer 2
IC50 BRPF1 (TR-FRET) [nM]
3150
5860
5490
IC50 BRPF2 (TR-FRET) [nM]
67
62
74
IC50 TAF1 BD2 (TR-FRET) [nM]
8
17
11
IC50 BRD4 BD2 (TR-FRET) [nM]
>20000
>20000
>20000
KD BRPF2 (ITC) [nM]
45
36
37
KD TAF1 BD2 (ITC) [nM]
17
n.a.
n.a.
IC50 BRPF2 cellular [nM]
575
354
565
BRPF2 cellular
inhibition [%]
82
84
84
TAF1 BD2 cellular inhibition [%]
64
63
63
Caco2 Papp(A-B) [nm/s]/Papp (B-A) [nm/s]/ratio
163/191/1.2
181/183/1.0
151/206/1.4
n.a. = not available.
Atropisomerism. (A) Atropisomers 5a and 5b of 5. (B) Energy profile
of the central torsion for
a 6′-methyl- (solid lines) and unsubstituted (dashed lines)
phenylnaphthalimide representative fragment. Optimization was performed
at the B3LYP/6-31G* level followed by single-point LMP2/cc-pVTZ(-f)
calculations.[54]n.a. = not available.
Biochemical Potency, Binding, and Selectivity
TR-FRET
assays showed that 5 was a potent inhibitor of BRPF2
BD with an IC50 of 67 nM, and a selectivity of 47- and
83-fold over BRPF1 and BRPF3 BDs (Table ). The profile of 5 was further
confirmed by AlphaScreen assay, where an IC50 of 97 nM
and a selectivity of 23- and 25-fold over BRPF1 and BRPF3 BDs were
measured (Table ).
Expanded selectivity studies using TSA with a panel of 48 BDs (Figure A) confirmed the
selectivity of 5 within the BRPF family with a thermal
shift for BRPF2 BD more than five degrees higher than for BRPF1 BD
and BRPF3 BD. Results from the BROMOscan panel further confirmed the
high selectivity with the BD family (Figure B). TR-FRET assays also disclosed a potent
inhibition of TAF1BD2 and the highly related TAF1LBD2 (Table ). Finally, ITC was
performed and KD values of 45 nM and 17
nM were determined for BRPF2 and TAF1BD2, respectively (Figure A–B). A cocrystal
structure of 5 in complex with TAF1BD2 revealed a highly
conserved binding mode reminiscent of the one obtained for the precursor 7 in BRPF2 BD (Supporting Information Figure S1). In addition, interactions of 5 with the
BDs of cyclic-AMP response element binding protein (CREBBP) and the
related EP300 were also detected by ITC (Figure A–B). However, with the BROMOscanTM
panel and in the AlphaScreen assays, only a weak activity was recorded
(30% and 41% inhibition in the presence of 100 nM compound; IC50 (CREBBP)
= 1710 nM) (Supporting Information Table
S4 and Table ). ITC
also revealed only modest interaction of 5 with CREBBP
BD (KD = 1390 nM) (Figure B). Compound 5 was also tested
at 10 μM in a panel containing over 300 kinases, and in all
cases, inhibition was less than 50% (Table ). Finally, 5 was evaluated
in the LeadProfilingScreen, which contains 68 potential targets, and
no inhibition or stimulation exceeding 25% was noted in the presence
of 10 μM compound (Table ).
Table 5
Biochemical Characterization of 5
Compound 5
IC50 BRPF2 (TR-FRET/Alpha Screen) [nM]
67/97
IC50 TAF1 BD2 (TR-FRET) [nM]
8
IC50 TAF1L BD2 (TR-FRET)
[nM]
106
IC50 BRPF1
(TR-FRET/Alpha Screen) [nM]
3150/2280
IC50 BRPF3 (TR-FRET/Alpha Screen) [nM]
5550/2450
IC50 BRD4 BD1/BD2 (TR-FRET)
[nM]
>15650/>20000
IC50 CREBBP (Alpha Screen) [nM]
1710
Kd BRPF2
(ITC) [nM]
45
Kd TAF1 BD2 (ITC) [nM]
17
IC50 BRPF2 BD/H4 cellular [nM]a
575
IC50 TAF1
BD2/H4 cellular [nM]a
970
Selectivity Kinases (300 entries)
>10000
Selectivity LeadProfilingScreen
(68 entries)
>10000
Determined
using NanoBRET.
Figure 5
Biophysical selectivity profile of 5. (A) TSA results
for 48 BDs at 10 μM compound concentration. (B) BROMOscan panel
results from DiscoverX at 100 nM compound concentration.
Figure 6
ITC results for 5. (A) ITC binding study:
data show
raw injection heats for titration of protein into compound. The inset
shows the normalized binding enthalpies corrected for the heat of
protein dilution as a function of binding site saturation. Solid lines
represent a nonlinear least-squares fit using a single-site binding
model. (a) BRPF2 BD–5; (b) TAF1 BD2–5; (c) TAF1L BD2–5; (d) CREBBP BD–5. (B) Summary of ITC data.
Biophysical selectivity profile of 5. (A) TSA results
for 48 BDs at 10 μM compound concentration. (B) BROMOscan panel
results from DiscoverX at 100 nM compound concentration.ITC results for 5. (A) ITC binding study:
data show
raw injection heats for titration of protein into compound. The inset
shows the normalized binding enthalpies corrected for the heat of
protein dilution as a function of binding site saturation. Solid lines
represent a nonlinear least-squares fit using a single-site binding
model. (a) BRPF2 BD–5; (b) TAF1BD2–5; (c) TAF1LBD2–5; (d) CREBBP BD–5. (B) Summary of ITC data.Determined
using NanoBRET.
Cellular Activity
NanoBRET experiments (Figure A–C) showed that the
interaction of BRPF2 BD with histones H4 and H3.3 was blocked by 5 with IC50 values of 575 and 825 nM, respectively.
For TAF1BD2, the IC50 values were 970 and 1400 nM, respectively.
No inhibitory effect was observed for the interaction between BRPF1
or BRD4 and histone H4 up to 10 μM for 5.
Figure 7
Cellular activity
of 5 and 31 in NanoBRET
assays. (A) Dose-dependent activity of 5. (B) Dose-dependent
activity of 31. (C) Overview of IC50 values
determined for 5 and 31.
Cellular activity
of 5 and 31 in NanoBRET
assays. (A) Dose-dependent activity of 5. (B) Dose-dependent
activity of 31. (C) Overview of IC50 values
determined for 5 and 31.
Antiproliferative Activity
Cell
viability experiments
were performed with 5 to determine a potential link between
BRPF2 or TAF1 with proliferative diseases. Altogether little or no
effect was observed. A weak inhibition was seen in cell lines originating
from hematological tumors (Supporting Information Table S5). Interestingly, a newly described quinoline-2-one inhibitor
targeting mainly BRPF1 and BRPF2 BDs also shows low micromolar activity
in leukemiacell lines.[48]
In
Vitro and in Vivo Pharmacokinetics
The metabolic stability of 5 was high after incubation
with mouse, rat, dog, or human liver microsomes (Table A). In contrast, metabolic stability
was low after incubation with rat or dog hepatocytes. This difference
may result from phase II metabolism, and indeed, corresponding metabolites
have been identified in vitro. Permeability through
Caco2cell monolayers was high, with no hint of active transport.
Plasma protein binding was low in mouse and moderate in human.
Table 6
Pharmacokinetic Properties of 5
Part A
Compound 5
Metabolic
Stability after Incubation with Liver Microsomes
Mouse CLblood [L/h/kg]/Fmax [%]
0.19/96
Rat CLblood [L/h/kg]/Fmax [%]
0.15/96
Dog CLblood [L/h/kg]/Fmax [%]
<0.001/100
Human CLblood [L/h/kg]/Fmax [%]
<0.001/100
Metabolic
Stability after Incubation with Hepatocytes
Rat CLblood [L/h/kg]/Fmax [%]
3.8/9.4
Dog CLblood [L/h/kg]/Fmax [%]
1.5/28
Permeability
in Caco2 Monolayer Cells
Caco2 Papp(A-B) [nm/s]/Papp (B-A) [nm/s]/ratio
106/204/1.9
Plasma Protein
Binding (Fraction Unbound)
NMRI mouse
female [%]
0.58
Caucasian human
(mixed gender) [%]
4.3
Studies of the in vivo pharmacokinetic properties
of 5 in rat revealed that blood clearance was low (ca.
17% of hepatic blood flow), volume of distribution in steady-state
high, terminal half-life long to very long, and bioavailability high
(F = 73%) (Table B). In vivo blood clearance was as
anticipated based on rat liver microsome values but lower than expected
based on hepatocyte data. Bioavailability matched blood clearance.
Negative Control
In order to deliver a valuable tool,
synthesis of an inactive, structurally close analogue of a chemical
probe, as a direct comparator from the same chemotype in biochemical
and cellular assays, is recommended.[13] For
this purpose, the key potency drivers of candidate 5 were
disrupted. As a first strategy, we opted for inducing steric hindrance
at the 1,3-dimethyl-part of the benzimidazolone moiety. The corresponding
diverse set of compounds had, however, remaining BRPF2 or TAF1 activity.
Then a different strategy aiming at reducing the size of the molecule
was followed. Previous results suggested that a significant loss of
potency could be achieved upon removing the methyl group at position
6 of the benzimidazolone. Additional removal of one or both methyl
groups from the 1,3-dimethylimidazolone led to a complete loss of
activity on BRPF2 BD, low residual activity on TAF1BD2, and no activity
on TAF1LBD2. 31 (BAY-364)[41] was selected among several candidates due to its favorable permeability
profile (Figure ).
Compound 31 showed no activity for BRPF3 or BRD4 BDs
in the AlphaScreen assay. Also, no cellular effect was seen in the
NanoBRET assay for BRPF2 BD or TAF1BD2, or for BRPF1 or BRD4 BDs
(Figure A–C).
Furthermore, 31 did not show any activity up to 10 μM
in a panel of 22 kinases (data not shown). Thus, 31 fulfills
the requirements of an inactive, structurally close analogue of probe 5.
Figure 8
Biochemical and cellular data for the negative control 31.
Biochemical and cellular data for the negative control 31.
Conclusion
An
HTS approach followed by medicinal chemistry optimization led
to the identification of 5, a highly potent and first
selective dual inhibitor of BRPF2 BD and TAF1/TAF1LBD2, which differs
from previously reported BRPF inhibitors. Our straightforward three-step
synthesis ensures 5 is an easily accessible tool. A cocrystal
structure of the close congener 7 in complex with BRPF2
BD revealed four key interactions, including a hydrogen bond between
a carbonyl group and Ser592 which was essential for selectivity
over BRPF1 BD. Importantly 5 was demonstrated to engage
BRPF2 and TAF1 in cells at submicromolar concentrations. Antiproliferative
activity was generally low and mainly observed in leukemiacell lines.
The strong biochemical and cellular activities of 5 suggest
it to be a high-quality chemical probe for the in-depth analysis of
the physiological function and pathological relevance of BRPF2 and
TAF1/TAF1L, which has not been possible to date. This novel tool enlarges
the repertoire of available chemical probes for epigenetic players
such that new knowledge about an implication of these two BD-containing
proteins in diverse pathologies can be expected forthwith.[49−53]
Experimental Procedures
General
Comments
Chemical names and numbers of the
compounds are given in italics and in bold, respectively. Chemical
names were generated using ACD/Name Batch or Autonom 2000, following
IUPAC nomenclature.
Stereochemistry
In certain cases
(e.g., where the substituent
at position 6 is not a hydrogen atom) atropisomerism could be induced,
and the compounds are assumed to be obtained as racemic mixtures of
atropisomers. In the cases where the atropisomericratio was investigated,
the ratio and the retention time of the corresponding atropisomers
are given in the analytical data part of the corresponding racemates.
In some cases, the atropisomers were separated using chiral HPLC.
The HPLC methods used for separation are described in the individual
cases, along with the corresponding analytical data. The atropisomers
with the shorter retention time (LCMS) were defined as atropisomer
1 (atrop1) and those with the longer retention time as atropisomer
2 (atrop2). The exact configuration of the obtained atropisomers was
not determined.
Analytical Methods
All NMR spectroscopy
results were
recorded on Bruker Avance III HD spectrometers. 1H spectra
were obtained at 400 MHz and referenced to residual solvent signals
(2.50 ppm for [D]6DMSO). All spectra were obtained at ambient
temperature (22 ± 1 °C). Peak forms and multiplicities are
specified as apparent in the spectra; potential higher-order effects
have not been considered. Chemical shifts (δ) are reported in
parts per million (ppm), and coupling constants (J) in hertz (Hz). Spin multiplicities are reported as s = singlet,
d = doublet, t = triplet, q = quartet, m = multiplet, bs = broad singlet,
and mc = centered multiplet. Optical rotations were measured
on a P2000 JASCO polarimeter using a 1 mL microcell (10 cm thickness,
3 mm diameter) and the alpha-D-line of sodium at 20 °C.
Chromatographic
Methods
Analytical liquid chromatography–mass
spectrometry (LC-MS) methods were performed using UPLCMS on a Waters
Acquity UPLCMS SingleQuad instrument (with an Acquity UPLC BEH C18 1.7 μm, 50 × 2.1 mm column or a Kinetex 2.6 μm
50 × 2.1 mm column) or on an Agilent 1290 UPLCMS 6230 time-of-flight
instrument with a BEH C18 1.7 μm, 50 × 2.1 mm
column. The different methods are described below.Analytical
methods with the Waters instruments:Method I: column: Acquity UPLC BEH C18 1.7
μm, 50 × 2.1 mm; eluent A: water + 0.1 vol % formic acid
(99%), eluent B: acetonitrile (ACN); gradient: 0–1.6 min 1–99%
B, 1.6–2.0 min 99% B; flow: 0.8 mL/min; temperature: 60 °C;
diode array detector (DAD) scan: 210–400 nm.Method II: column: Acquity UPLC BEH C18 1.7
μm, 50 × 2.1 mm; eluent A: water + 0.2 vol % aqueous ammonia
(32%), eluent B: ACN; gradient: 0–1.6 min 1–99% B, 1.6–2.0
min 99% B; flow: 0.8 mL/min; temperature: 60 °C; DAD scan: 210–400
nm.Method III: column: BEH C18 1.7 μm
50 × 2.1 mm; eluent A: water + 0.05% formic acid (99%), eluent
B: ACN + 0.05% formic acid (99%); gradient: 0–0.2 min 2% B,
0.2–1.7 min 2–90% B, 1.7–1.9 min 90% B; 1.9–2.0
min 90–2% B, 2.0–2.5 min 2% B, flow: 1.3 mL/min; temperature:
60 °C; DAD scan: 200–400 nm.Method IV: column: Kinetex 2.6 μm, 50 × 2.1
mm; eluent A: water + 0.05% formic acid (99%), eluent B: ACN + 0.05%
formic acid (99%); gradient: 0–1.9 min 1–99% B, 1.9–2.1
min 99% B; flow: 1.3 mL/min; temperature: 60 °C; DAD scan: 200–400
nm.Analytical methods with the Agilent
instruments:Method V: eluent
A: water + 0.05% formic acid (99%),
eluent B: ACN + 0.05% formic acid (99%); gradient: 0–1.7 min
2–90% B, 1.7–2.0 min 90% B; flow: 1.2 mL/min; temperature:
60 °C; DAD scan: 190–400 nm.
Reagents
All reagents for which the synthesis is not
described below are either commercially available or were synthesized
according to literature procedures. All final products were at least
95% pure, as determined by analytical HPLC, apart from the two cases
otherwise indicated.
General Procedure for the Condensation of
Naphthalic Anhydrides
under Acidic Conditions (GP1)
In a pressure vessel, or a
round-bottom flask, the aniline (1.0 equiv) was suspended in acetic
acid (3.8 mL/mmol) under argon atmosphere. Then, the commercially
available anhydride partner (0.9–1.5 equiv) was added. The
reaction mixture was heated to reflux for 3–16 h. After cooling,
the mixture was concentrated under reduced pressure. The residue was
suspended in methanol (10 mL/mmol), and 4 N sodium hydroxide solution
(177 μL/mmol) was added. After stirring at room temperature
(RT) for 1 h, water (74 mL/mmol) was added and the mixture was further
stirred for 10 min. The precipitate was separated by filtration, washed
three times with water and methanol, and dried overnight in a vacuum
oven at 40 °C. If necessary, the precipitate was purified by
HPLC or flash column chromatography to give the desired product.
General Procedure for the Nucleophilic Substitution of Amines
(GP2)
In a pressure vessel, the aryl bromide (1.0 equiv)
was dissolved in DMSO (7.5 mL/mmol) under argon atmosphere. The commercially
available amine partner (1.1 equiv) and potassium carbonate (2.5 equiv)
were added ,and the mixture was heated to 110 °C overnight. After
cooling, the precipitate was separated by filtration and purified
by HPLC.
General Procedure for the Sonogashira Coupling
(GP3)
In a pressure vessel, the aryl bromide (1.0 equiv)
was dissolved
in tetrahydrofuran (0.4 mL/mmol). The commercially available catalyst
bis(triphenylphosphine)palladium(II) dichloride (0.02 equiv), copper(I)
iodide (0.02 equiv), triethylamine (7.4 equiv), and the coupling partner
(2.0 equiv) were added to the solution, and the mixture was degassed
with argon for 5 min. The mixture was heated to 65 °C for 5–16
h. After cooling, the mixture was diluted with aqueous ammonia (6
mL/mmol) and was stirred again for 5 min. Dichloromethane (60 mL/mmol)
and water (15 mL/mmol) were added, and the layers were separated.
The aqueous layer was extracted with dichloromethane (2 × 40
mL/mmol). The combined organic layers were washed with brine (60 mL/mmol)
and dried over sodium sulfate. After filtration, the solvent was removed
under reduced pressure. The residue was purified by HPLC to give the
desired product. In some cases, the crude product was directly used
in the next step without any further purification.
General Procedure
for Hydrogenation with Palladium on Carbon
(GP4)
In a hydrogenation flask, the alkyne (1.0 equiv) was
dissolved in ethyl acetate (450–1000 mL/mmol) and Pd/C (0.02
equiv) was added. The suspension was flushed with argon and vigorously
shaken under a 1 bar hydrogen atmosphere for 2–4 h. After reaction
completion, the suspension was flushed with argon again, the catalyst
was removed by filtration through a pad of Celite, and the solvent
was removed under reduced pressure. The residue was purified by HPLC
or flash column chromatography to give the desired product.
According to GP1, 30 (150 mg,
0.784 mmol) and 1,8-naphthalic anhydride (233 mg, 1.18 mmol) in acetic
acid (3 mL) gave 7 as a colorless solid (40 mg, 13%). 1H NMR (400 MHz, [D]6DMSO): δ = 2.09, 3.29,
3.38 (3 s, 3 H each), 7.18 (mc, 2 H), 7.93 (dd, J = 7.8 Hz, 2 H), 8.53–8.56 (m, 4 H) ppm; LCMS (method
III): R = 1.04 min; m/z: [M + H]+ = 372. Compound 7 was assumed to be obtained as a racemic mixture of atropisomers.
The atropisomericratio was not investigated at this stage.
According to GP1, 5-amino-6-methoxy-1,3-dimethyl-1,3-dihydro-2H-benzimidazol-2-one (159 mg, 0.769 mmol) and 1,8-naphthalic
anhydride (229 mg, 1.15 mmol) in acetic acid (3 mL) gave 8 as a colorless solid (4.2 mg, 1%). 1H NMR (400 MHz, [D]6DMSO): δ = 3.29, 3.41, 3.72 (3 s, 3 H each), 7.13, 7.21
(2 mc, 1 H each), 7.92 (t, J ≈
7.8 Hz, 2 H), 8.53 (t, J ≈ 8.4 Hz, 4 H) ppm;
LCMS (method III): R = 0.98 min; m/z: [M + H]+ = 388. Compound 8 was assumed to be obtained as a racemic mixture of atropisomers.
The atropisomericratio was not investigated at this stage.
According to GP1, 5-amino-6-(dimethylamino)-1,3-dimethyl-1,3-dihydro-2H-benzimidazol-2-one dihydrochloride (244 mg, 0.833 mmol)
and 1,8-naphthalic anhydride (248 mg, 1.15 mmol) in acetic acid (3
mL) gave 9 as a colorless solid (20 mg, 6%). 1H NMR (400 MHz, [D]6DMSO): δ = 2.47, 3.29, 3.40
(3 s, 6 H, 3 H, 3 H), 7.18, 7.26 (2 mc, 1 H each), 7.92
(t, J ≈ 7.7 Hz, 2 H), 8.53, 8.55 (2 mc, 2 H each) ppm; LCMS (method V): R = 0.78 min; m/z: [M + H]+ = 401. Compound 9 was assumed to be obtained
as a racemic mixture of atropisomers. The atropisomericratio was
not investigated at this stage.
Commercially available 5-chloro-1,3-dimethyl-6-nitro-1,3-dihydro-2H-benzimidazol-2-one (250 mg, 1.03 mmol) was suspended under
an atmosphere of argon in N,N-dimethylformamide
(10 mL), and sodium carbonate (286 mg, 2.07 mmol) and 1-methylpiperazine
(170 μL, 1.6 mmol) were added. The reaction mixture was heated
to 120 °C overnight. After cooling, the reaction mixture was
filtered and the solvent was removed under reduced pressure. The obtained
residue was then suspended in ethanol (19.2 mL), and Pd/C (57.9 mg,
0.054 mmol) was added under argon. The suspension was flushed with
argon and vigorously shaken under an atmosphere of hydrogen (1 bar)
for 4 h. After reaction completion, the suspension was flushed with
argon again, the catalyst was removed by filtration through a pad
of Celite, and the solvent was removed under reduced pressure. The
crude product was directly used in the next step without any further
purification. LCMS (method I): R = 0.48
min; m/z: [M + H]+ =
275. According to GP1, crude product (200 mg, 0.726 mmol) and 1,8-naphthalic
anhydride (144 mg, 0.269 mmol) in acetic acid (2.6 mL) gave 11 was obtained as a colorless solid (20 mg, 6%). 1H NMR (400 MHz, [D]6DMSO): δ = 1.87 (s, 3 H), 2.72
(mc, 4 H), 3.30, 3.39 (2 s, 3 H each), 7.22, 7.27 (2 mc, 1 H each), 7.93 (t, J ≈ 7.7 Hz,
2 H), 8.53, 8.55 (2 mc, 2 H each) ppm; four proton signals
could not be detected; LCMS (method V): R = 0.61 min; m/z: [M + H]+ = 456. Compound 11 was assumed to be a racemic
mixture of atropisomers. The atropisomericratio was not investigated
at this stage.
Compound 14 (78 mg, 0.187
mmol) was suspended in methanol (110 mL), and Pd/C (40 mg, 0.037 mmol)
was added. The suspension was flushed with argon and vigorously shaken
under an atmosphere of hydrogen (1 bar) for 3 h. After reaction completion,
the suspension was flushed with argon again, the catalyst was removed
by filtration through a pad of Celite, and the solvent was removed
carefully under reduced pressure. The residue (23 mg) was purified
by HPLC (basic) and lyophilized to give the desired aniline 15 as a yellow solid (10 mg, 12%), which was stored under
argon in the fridge. 1H NMR (400 MHz, [D]6DMSO):
δ = 2.06, 3.29, 3.37 (3 s, 3 H each), 6.05 (bs, 2 H), 7.16*
(mc, 2 H), 7.35 (d, J = 2.3 Hz, 1 H),
7.66 (dd, J = 7.4, 8.1 Hz, 1 H), 8.00 (d, J = 2.3 Hz, 1 H), 8.09 (mc, 2 H) ppm; *overlapping
signals; LCMS (method V): R = 0.86 min; m/z: [M + H]+ = 387. Compound 15 was assumed to be obtained as a racemic mixture of atropisomers.
The atropisomericratio was not investigated at this stage.
In a pressure vessel, aryl bromide 12 (100 mg, 0.222 mmol, 1.0 equiv) was suspended in toluene
(2 mL) under argon atmosphere. The commercially available catalyst
tris(dibenzylideneacetone)dipalladium (8.1 mg, 0.009 mmol, 0.040 equiv),
9,9-dimethyl-4,5-bis(diphenylphosphino)xanthene (Xantphos, 5.1 mg,
0.009 mmol, 0.040 equiv), cesium carbonate (218 mg, 0.671 mmol, 3.0
equiv), and propylamine (56 μL, 0.671 mmol, 2.1 equiv) were
added, and the mixture was degassed with argon for 5 min. The reaction
mixture was heated to 90 °C for 18 h. After cooling, the solvent
was removed under reduced pressure. The residue was purified by HPLC,
giving product 26 as a green solid (10 mg, 10%). 1H NMR (400 MHz, [D]6DMSO): δ = 1.00 (t, J = 7.4 Hz, 3 H), 1.74 (mc, 2 H), 2.04, 3.29,
3.37* (3 s, 3 H each), 3.14–3.37* (m, 2 H), 6.83 (d, J = 8.9 Hz, 1 H), 7.09, 7.15 (2 mc, 1 H each),
7.72 (t, J = 7.9 Hz, 1 H), 7.89 (mc, 1
H), 8.28 (d, J = 8.6 Hz, 1 H), 8.45 (d, J = 6.8 Hz, 1 H), 8.79 (d, J = 8.1 Hz, 1 H) ppm;
*overlapping signals; LCMS (method III): R = 1.18 min; m/z: [M + H]+ = 429. Compound 26 was assumed to be obtained
as a racemic mixture of atropisomers. The atropisomericratio was
not investigated at this stage.
According to GP1, 6-bromo-2-(1-methyl-2-oxo-2,3-dihydro-1H-benzimidazol-5-yl)-1H-benzo[de]isoquinoline-1,3(2H)-dione was prepared from 5-amino-1-methyl-1,3-dihydro-2H-benzimidazol-2-one (3.34 g, 20.5 mmol) and 4-bromo-1,8-naphthalic
anhydride (5.96 g, 21.5 mmol) in acetic acid (82 mL). The corresponding
aryl bromide was obtained as a yellow solid (6.22 g), part of which
(500 mg, 1.18 mmol) was used in the next step, according to GP3, with
propargyl alcohol (138 μL, 2.39 mmol), triethylamine (1.3 mL,
8.81 mmol), CuI (4.5 mg, 0.024 mmol), and bis(triphenylphosphine)palladium(II)
dichloride (17 mg, 2.37 mmol) in tetrahydrofuran (40 mL). After workup,
the residue [200 mg, LCMS (method II): R = 0.87 min; m/z: [M + H]+ = 397] was used directly in the next step. According to GP4,
intermediate alkyne and Pd/C (214 mg, 0.101 mmol) in EtOAc (167 mL)
gave 31 as a yellow solid (7 mg, 3%). 1H NMR
(400 MHz, [D]6DMSO): δ = 1.87, 3.28 (2 mc, 2 H each), 3.34* (s, 3 H), 3.52, 4.66 (2 mc, 2 H, 1
H), 6.98–7.03 (m, 2 H), 7.18 (d, J = 8.1 Hz,
1 H), 7.76 (d, J = 7.6 Hz, 1 H), 7.92 (dd, J = 7.2, 8.5 Hz, 1 H), 8.42 (d, J = 7.4
Hz, 1 H), 8.51 (dd, J = 1.0, 7.1 Hz, 1 H), 8.65 (dd, J = 1.0, 8.5 Hz, 1 H), 11.00 (bs, 1 H) ppm; *overlapping
with water solvent peak of [D]6DMSO; LCMS (method IV): R = 0.75 min; m/z: [M + H]+ = 402.
Protein Expression
and Purification
For screening and
biophysical testing cDNAs encoding humanBRPF2 BD (residues 563–687,
UniProt O95696), humanTAF1BD2 (residues 1501–1635, UniProt
P21675), TAF1L (residues 1523–1654, UniProt Q8IZX4), and humanCREBBP BD (residues 1081–1197, UniProt Q92793) were cloned
into a modified pET28-derived expression vector (pNIC28-Bsa4) using
ligation-independent cloning. Alternatively for TR-FRET assays, GST-TAF1 and GST-TAF1L proteins were purchased from BPS Biosciences. Expressionconstructs encoding the respective
target proteins were then transformed into competent E. coli BL21 (DE3)-R3-pRARE2 bacteria (phage-resistant derivative with a
pRARE plasmid encoding rare codon tRNAs). Bacteria were then grown
at 37 °C in Terrific Broth from overnight cultures until the
OD600 reached between 0.6 and 0.8, followed by cooling
and addition of 0.2 mM isopropyl-β-d-1-thiogalactopyranoside
(IPTG) to allow protein expression at 18 °C for 16 h. Cells were
lysed in loading buffer (20 mM Hepes pH 7.5, 500 mM NaCl, 20 mM imidazole,
1 mM TCEP, 5% glycerol) in the presence of Protease Inhibitor Cocktail
EDTA-free using an EmulsiFlex-C5 high-pressure homogenizer. Lysates
were cleared by centrifugation at 14000g for 45 min
at 4 °C, after which the supernatant was loaded onto a nickel
affinity column (HisTrap Chelating FF 5 mL), followed by washing and
elution with a linear imidazole gradient. The eluted proteins were
collected and treated overnight with Tobacco Etch Virus protease at
4 °C, followed by a second affinity purification step. Proteins
were further purified by size exclusion chromatography (HiLoad 16/600
Superdex 75 GE Healthcare Life Sciences). Protein identity was confirmed
by intact mass analysis.The construct for cocrystallization
of humanBRPF2 encoded the residues 556–688 with two mutations
P566E and V569R integrated in the pNIC28-Bsa4 vector (Addgene, Teddington,
U.K.). The expressionconstruct was introduced into competent E. coli BL21 (DE3) bacteria, which were then grown in buffered
LB medium until the OD550 reached 1.0. This was followed
by cooling and addition of 0.1 mM IPTG to allow protein expression
at 17 °C for 24 h. Cells were resuspended in base buffer (10
mM Hepes pH 7.5, 500 mM NaCl, 10 mM imidazole, 0.5 mM TCEP, 5% glycerol)
supplemented with 0.1% Triton X100 and 0.025% NP40 in the presence
of Protease Inhibitor Cocktail EDTA-free, and passed through a Microfluidizer
(3 × 900 bar) for lysis. Lysates were cleared by centrifugation
at 30000g at 4 °C for 60 min, after which the
supernatant was loaded onto a nickel affinity column (HisTrap Chelating
HF 5 × 5 mL), followed by a washing step with 25 mM and subsequent
elution step with 250 mM imidazole in base buffer. The eluted proteins
were collected, dialyzed against base buffer, and treated with Tobacco
Etch Virus protease at 4 °C overnight, followed by a second affinity
purification step. Proteins were further purified by size exclusion
chromatography using base buffer without imidazole (HiLoad 35/600
Superdex 75 GE Healthcare Life Sciences). Monomeric peak fractions
were pooled and concentrated to 13 mg/mL. These samples were directly
used for crystallization trials. Protein identity was confirmed by
intact mass analysis. For cocrystallization of humanTAF1BD2, the
same plasmid construct, expression, and purification conditions as
above were used.
Computational Methods
Coordinate Scans
Optimization was performed at the
B3LYP/6-31G* level followed by single-point LMP2/cc-pVTZ(-f) calculations.[54]
Predictions and Calculation Method
BEI and LLE were
determined by a proprietary software tool using the following equations
at physiological pH: LLE = −log(IC50) – clogD and BEI = [−log(IC50)/MWcorr]*1000. clogD (pH 7.5) as estimated from HPLC retention times was
calculated by using a highly predictive method.[55−57]
Crystallography
BRPF2
Apo BRPF2crystals were grown at 22 °C using
the sitting-drop method by mixing 200 nL of protein (13 mg/mL/0.825
mM) with 100 nL of well solution (100 mM bis-tris pH 6.5, 30% (w/v)
PEG 3350). For complex formation, a single crystal was transferred
into a soaked drop containing mother liquor supplemented with 10 mM
compound (prepared as a 100 mM stock solution in 100% ethylene glycol).
After soaking for 48 h, the crystal was briefly immersed in cryoprotection
solution consisting of mother liquor supplemented with 10 mM compound
and 30% ethylene glycol and then flash-frozen in liquid nitrogen.
X-ray data were collected on the Helmholtz-Zentrum Berlin beamline
14–1 at a wavelength of 0.91814 Å and using a PILATUS
detector. Data were integrated, scaled, and merged using the programs
XDS and SCALA.[58,59] The structure was solved by molecular
replacement using the program Phaser.[60] Initial electron density maps clearly indicated binding of the compound.
The model was refined through iterative manual and maximum-likelihood
refined using the programs COOT and REFMAC5.[58,61] Statistics for the final model are given in the Supporting Information Table S6. Coordinates and structure
factors have been submitted to the PDB database and are accessible
with the code 5N49.
TAF1
Aliquots of purified TAF1BD2 were setup for crystallization
using a mosquito crystallization robot (TTP Labtech, Royston U.K.).
Coarse screens were typically set up in Greiner 3-well plates using
three different drop ratios of precipitant to protein per condition
(200 + 100 nL, 150 + 150 nL, and 100 + 200 nL). All crystallizations
were carried out using the sitting-drop vapor diffusion method at
4 °C. TAF1crystals with 5 were grown by mixing
100 nL of protein (11.3 mg/mL mixed with 1:3 molar ratio of ligand)
with 200 nL of reservoir solution containing 20% PEG 10K, 0.20 M NaCl,
and 0.1 M citrate/phosphate pH 4.2. TAF1complex crystals were cryoprotected
using well solution supplemented with additional 20% ethylene glycol
for flash-freezing in liquid nitrogen. Data were collected at Diamond
beamline I02 at a wavelength of 0.9794 Å. Indexing, integration,
and scaling were carried out using autoPROC[62] (Global Phasing Limited - GPhL); this package uses XDS,[59] POINTLESS,[63] AIMLESS,[64] and CCP4.[65] Initial
phases were calculated by molecular replacement with PHASER[60] using the apo TAF1 structure (PDB ID: 3UV4). Initial models
were built by ARP/wARP[66] followed by manual
building in COOT.[67] Refinement was carried
out in REFMAC5.[68] GRADE[69] (GPhL) was used to generate compound coordinates and cif
files. All model validations were carried out using MolProbity.[70] Data collection and refinement statistics can
be found in Supporting Information Table
S6. The model and structure factors have been deposited with PDB accession
code 5MG2.
Time-Resolved Fluorescence Resonance Energy Transfer Assays
TR-FRET binding competition assays were performed at RT in 1536-
or 384-well black microliter plates in a final volume of 5 μL.
Compounds were diluted in DMSO at a single concentration (10 μM
for primary screening, 5 and 25 μM for confirmation tests) or
in a titration series (3.5-fold, 11-points, maximal test concentration:
20 μM) and dispensed 100× concentrated onto the plates.
Then, 25 nM His-tagged BRPF1 (a.a. 626–740), BRPF2 (a.a. 563–688),
or BRPF3 (a.a. 592–711), 5 nM GST-tagged TAF1 (a.a. 1519–1651), 5 nM GST-tagged TAF1L (a.a. 1517–1649) or 10 nM BRD4BD1 (a.a. 67–152)
in assay buffer [50 mM Hepes pH 7.5, 50 mM NaCl, 50 mM KF, 0.01% NP40,
0.05% bovineserum albumin (BSA) and 1 mM dithiothreitol (DTT)] was
added, and the plates were incubated for 15 min. Finally, 50 nM (25 nM for TAF assays) C-terminal
biotinylated synthetic acetylated peptides derived from histone H4
a.a.1–20 (K5,12 diacetylated for BRPF1, 2, 3 or K5,8,12,16
tetra-acetylated for TAF1, TAF1L and BRD4) and detection reagents (10 nM anti-6His-XL665
(Cisbio), 2.5 nM streptavidin Eu (PerkinElmer) or 2.5 nM anti-GST-Tb Cryptate and 3.12 nM streptavidin-XL665 (Cisbio), both in assay buffer)
were dispensed and further incubated for at least 1 h. TR-FRET signals
corresponding to the number of protein–peptide complexes in
equilibrium were acquired either with Viewlux (PerkinElmer) or Pherastar
(BMG Lab Technologies) microtiter plate readers. The normalized ratios
of the fluorescence emission at 665 nm and at 620–622 nm after
excitation at 330–350 nm, at increasing compound concentrations
were used to calculate IC50 values by regression analysis
based on a four-parameter equation [minimum, maximum, IC50, Hill; Y = Max + (Min – Max)/(1 + (X/IC50)Hill)].
AlphaScreen
Assay conditions: BRD4BD1, BRPF1B, BRPF2,
BRPF3, CECR2, CREBBP, and FALZ/BPTF assays were performed as described
previously,[71] with minor modifications
of the manufacturer’s protocol (PerkinElmer) using the same
acetylated H4 peptide. All reagents were diluted in 25 mM Hepes, 100
mM NaCl, 0.1% BSA, pH 7.4 supplemented with 0.05% CHAPS. Plates were
filled with 5 μL of the assay buffer followed by 7 μL
of biotinylated peptide and His-tagged protein to achieve final assay
concentrations of typically 50 nM. Plates were sealed and incubated
for a further 30 min, before the addition of 8 μL of the mixture
of streptavidin-coated donor beads (12.5 μg/mL) and nickelchelate
acceptor beads (12.5 μg/mL) under low light conditions. Plates
were foil-sealed for protection from light, and incubated at RT for
60 min before reading.
Isothermal Titration Calorimetry
For BRPF2 and TAF1,
experiments were performed in buffer (20 mM Hepes pH 7.5, 100 mM NaCl)
on a VP-ITC microcalorimeter (MicroCal) at 20 °C. For CREBBP
and TAF1L, experiments were done at 15 °C using a buffer containing
20 mM Hepes pH 7.5, 150 mM NaCl, and 0.5 mM TCEP. Protein concentrations
were between 200 and 240 μM, and the compound concentration
was set between 18 and 25 μM. The titrations were conducted
using an initial injection of 2 μL followed by identical injections
of 8 μL (between 24 and 35 injections). The dilution heats were
determined in separate experiments and subtracted from the titration
data. Thermodynamic parameters were calculated using ΔG = ΔH – TΔS = −RT ln KB, where ΔG, ΔH, and ΔS are the changes in free energy, enthalpy,
and entropy of binding, respectively. In all cases, a single binding
site model was employed.
Thermal Shift Assay
Thermal melting
experiments were
carried out using an Mx3005p Real-Time PCR machine (Stratagene), as
described.[72] Proteins were prepared in
10 mM Hepes pH 7.5, 500 mM NaCl, and assayed in a 96-well plate at
a final concentration of 2 μM in 20 μL volume. Compounds
were added at a final concentration of 10 μM. SYPRO Orange (Molecular
Probes) was added as a fluorescence probe at a dilution of 1:1000.
The temperature gradient was 3 °C per min. The melting temperature
was calculated by fitting the Boltzmann equation using Prism 5. Thermal
shift assays for hit validation were performed with a ViiA Real-Time
PCR machine (Thermo Fisher Scientific) in a 384-well plate format.
The thermal ramping was 4 °C per minute from 25 to 80 °C.
BRPF1 protein was diluted to 2.5 μM in 10 mM HEPES pH 7.5, 150
mM NaCl, and BRPF2 with a protein concentration of 2.7 μM in
HEPES pH 7.5 and 300 mM NaCl. Both proteins were measured in the presence
or absence of 100 μM of ligand (1% final DMSOconcentration)
and a 6-fold SYPRO-Orange dye concentration. TSA data were analyzed
using a Genedata Assay Analyzer.
NanoBRET Assay
The assay principle has been described.[42] Expression vectors were purchased from Promega
(Madison, Wisconsin, U.S.A). The NanoLuc luciferase was C-terminally
fused to BRPF2 BD (a.a. 560–666), BRPF1 BD (a.a. 626–732),
or TAF1BD2 (a.a. 1528–1640). A BRD4-NanoLuc fusion was used
as control. The HaloTag was C-terminally fused to histone H3.3 or
H4. The resulting plasmids were transfected into HCT116cells using
standard procedures. Following addition of the HaloTag ligand, the
cells were treated with compound (10 nM to 10 μM) for 4 or 24
h, and the IC50 values were calculated.
LeadProfilingScreen
and Kinase Panel
The LeadProfilingScreen
containing 68 different targets and the kinase panel experiment (300
kinases) were performed by Eurofins, Cerep Panlabs (Taipei, Taiwan).
Cell Proliferation Assay
Cell lines were obtained from
ATCC (Manassas, Virginia, USA) or the DSMZ (Braunschweig, Germany),
authenticated, and grown following standard procedures. The cells
were treated with compound while in the logarithmic growth phase,
and their viability was determined by AlamarBlue staining (ThermoFisher,
Waltham, Massachusetts, U.S.A.).
Permeability Assay
Caco2cells purchased from the DSMZ
were seeded at a density of 4.5 × 104 cells per well
and grown for 15 days in DMEM with typical supplements. Cells were
kept at 37 °C in a humidified 5% CO2 atmosphere. Before
the permeation assay was run, the culture medium was replaced by a
FCS-free Hepes carbonate transport buffer (pH 7.2). For assessment
of monolayer integrity, the transepithelial electrical resistance
was measured. Test compounds were predissolved in DMSO and added to
either the apical or basolateral compartment at a final concentration
of 2 μM. Before and after 2 h incubation at 37 °C, samples
were taken from both compartments. Analysis of compound content was
conducted after precipitation with methanol by LC/MS/MS analysis.
Permeability (Papp) was calculated in the apical to basolateral
(A → B) and basolateral to apical (B → A) directions.
The efflux ratio basolateral (B) to apical (A) was calculated by dividing
Papp(B-A) by Papp(A-B). Reference compounds
were analyzed in parallel as assay control.
Pharmacokinetics
Metabolic
Stability in Hepatocytes
A hepatocyte cell
suspension of the respective species was filtered through sterile
gauze in 50 mL Falcon tubes and centrifuged at 50g for 3 min at RT. The cell pellet was resuspended in 30 mL of WME
and centrifuged through a Percoll gradient twice at 100g. The hepatocytes were washed again with WME and resuspended in medium
containing 5% FCS. Liver cells were distributed in WME containing
5% FCS to glass vials at a density of 1.0 × 106 vital
cells/mL. Compound 5 was added at a final concentration
of 1 μM. During incubation, the hepatocyte suspensions were
continuously shaken at 580 rpm and aliquots were taken at 2, 8, 16,
30, 45, and 90 min, to which equal volumes of cold methanol were immediately
added. Samples were frozen at −20 °C overnight, and then
centrifuged for 15 min at 3000 rpm, and the supernatant was analyzed
by LC/MS/MS. The half-life of 5 was determined from the
concentration–time plot. The intrinsicclearances were calculated
from the half-life.
Metabolic Stability in Liver Microsomes
The in vitro metabolic stability was determined
by incubating
compounds at 1 μM concentration in a suspension of liver microsomes
in 100 mM phosphate buffer, pH 7.4 (NaH2PO4·H2O + Na2HPO4·2H2O) and
at a protein concentration of 0.5 mg/mL at 37 °C. The microsomes
were activated by adding a cofactor mix containing 8 mM glucose-6-phosphate,
4 mM MgCl2, 0.5 mM NADP, and 1 IU/mL G-6-P-dehydrogenase
in phosphate buffer, pH 7.4. The metabolic assay was started shortly
afterward by adding 5 to the incubation in a final volume
of 1 mL. During incubation, the microsomal suspensions were continuously
shaken at 580 rpm and aliquots were taken at 2, 8, 16, 30, 45, and
60 min, to which equal volumes of cold methanol were immediately added.
Samples were frozen at −20 °C overnight and then centrifuged
for 15 min at 3000 rpm, and the supernatant was analyzed by LC/MS/MS.
The half-life of 5 was determined from the concentration–time
plot. The intrinsicclearances were calculated from the half-life.
In Vivo Rat Pharmacokinetics
Compound 5 was administered to three male Wistar rats per arm, either
intravenously or intragastrally formulated as solutions. 5 was given as i.v. bolus, and blood samples were taken at 2 min,
8 min, 15 min, 30 min, 45 min, 1 h, 2 h, 4 h, 6 h, 8 h, and 24 h after
dosing. For pharmacokinetics after intragastral administration, 5 was given intragastrally to fasted rats and blood samples
were taken at 5 min, 15 min, 30 min, 45 min, 1 h, 2 h, 4 h, 6 h, 8
h, and 24 h after dosing. Blood was collected into lithium-heparin
tubes (Monovettes, Sarstedt) and centrifuged for 15 min at 3000 rpm.
An aliquot of 100 μL from the supernatant (plasma) was taken
and precipitated by addition of 400 μL of cold ACN and frozen
at −20 °C overnight. Samples were thawed and centrifuged
at 3000 rpm, 4 °C for 20 min. Aliquots of the supernatants were
submitted to LC/MS/MS analysis employing a validated method on an
Agilent 1200 HPLCcoupled to a Sciex API 5500 instrument. Standard
pharmacokinetic parameters were calculated by noncompartmental analysis
using dedicated software.
Physicochemistry
Reversed-Phase
HPLC Method for Lipophilicity (logD)
logD (pH 7.5) was determined
by reversed-phase HPLC at physiological
pH, as described.[73] A homologous series
of n-alkan-2-ones (C3–C16, 0.02 mol in ACN) was used for calibration. Test compounds were
applied as 0.67 mM DMSO stock solutions in ACN/H2O 1:1.
The lipophilicity of compounds was then assessed by comparison to
the calibration curve.
High-Throughput Determination of Aqueous
Drug Solubility (100
mM in DMSO)
The high-throughput screening method to determine
aqueous drug solubility was performed as described previously.[74] Test compounds were applied as 1 mM solutions
in DMSO. After the addition of a pH 6.5 buffer, the solution was shaken
for 24 h at RT. Undissolved material was separated by filtration.
The compound dissolved in the supernatant was diluted accordingly
and quantified by HPLC-MS/MS.
Authors: Niall Igoe; Elliott D Bayle; Oleg Fedorov; Cynthia Tallant; Pavel Savitsky; Catherine Rogers; Dafydd R Owen; Gauri Deb; Tim C P Somervaille; David M Andrews; Neil Jones; Anne Cheasty; Hamish Ryder; Paul E Brennan; Susanne Müller; Stefan Knapp; Paul V Fish Journal: J Med Chem Date: 2017-01-09 Impact factor: 7.446
Authors: Paul Bamborough; Hawa Diallo; Jonathan D Goodacre; Laurie Gordon; Antonia Lewis; Jonathan T Seal; David M Wilson; Michael D Woodrow; Chun-Wa Chung Journal: J Med Chem Date: 2012-01-11 Impact factor: 7.446
Authors: Paul Bamborough; Heather A Barnett; Isabelle Becher; Mark J Bird; Chun-Wa Chung; Peter D Craggs; Emmanuel H Demont; Hawa Diallo; David J Fallon; Laurie J Gordon; Paola Grandi; Clare I Hobbs; Edward Hooper-Greenhill; Emma J Jones; Robert P Law; Armelle Le Gall; David Lugo; Anne-Marie Michon; Darren J Mitchell; Rab K Prinjha; Robert J Sheppard; Allan J B Watson; Robert J Watson Journal: ACS Med Chem Lett Date: 2016-05-09 Impact factor: 4.345
Authors: David Remillard; Dennis L Buckley; Hyuk-Soo Seo; Fleur M Ferguson; Sirano Dhe-Paganon; James E Bradner; Nathanael S Gray Journal: ACS Med Chem Lett Date: 2019-09-13 Impact factor: 4.345
Authors: Jeffrey Mowat; Alexander H M Ehrmann; Sven Christian; Carolyn Sperl; Stephan Menz; Judith Günther; Roman C Hillig; Marcus Bauser; Wolfgang Schwede Journal: ACS Med Chem Lett Date: 2022-02-18 Impact factor: 4.345
Authors: Michael A Clegg; Natalie H Theodoulou; Paul Bamborough; Chun-Wa Chung; Peter D Craggs; Emmanuel H Demont; Laurie J Gordon; Gemma M Liwicki; Alex Phillipou; Nicholas C O Tomkinson; Rab K Prinjha; Philip G Humphreys Journal: ACS Med Chem Lett Date: 2021-08-03 Impact factor: 4.632
Authors: Changxin Wan; Matthew P Keany; Han Dong; Linah F Al-Alem; Unnati M Pandya; Suzan Lazo; Karsten Boehnke; Katherine N Lynch; Rui Xu; Dominique T Zarrella; Shengqing Gu; Paloma Cejas; Klothilda Lim; Henry W Long; Kevin M Elias; Neil S Horowitz; Colleen M Feltmate; Michael G Muto; Michael J Worley; Ross S Berkowitz; Ursula A Matulonis; Marisa R Nucci; Christopher P Crum; Bo R Rueda; Myles Brown; Xiaole Shirley Liu; Sarah J Hill Journal: Cancer Res Date: 2020-11-06 Impact factor: 13.312
Authors: Ye Xu; Na Man; Daniel Karl; Concepcion Martinez; Fan Liu; Jun Sun; Camilo Jose Martinez; Gloria Mas Martin; Felipe Beckedorff; Fan Lai; Jingyin Yue; Alejandro Roisman; Sarah Greenblatt; Stephanie Duffort; Lan Wang; Xiaojian Sun; Maria Figueroa; Ramin Shiekhattar; Stephen Nimer Journal: Nat Commun Date: 2019-10-29 Impact factor: 14.919
Authors: Lukas Leiendecker; Pauline S Jung; Izabela Krecioch; Tobias Neumann; Alexander Schleiffer; Karl Mechtler; Thomas Wiesner; Anna C Obenauf Journal: EMBO Mol Med Date: 2020-10-07 Impact factor: 12.137
Figure 1
Figure 2
Figure 3
Scheme 1
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8