Hannes Falke1, Apirat Chaikuad2, Anja Becker1, Nadège Loaëc3,4, Olivier Lozach4, Samira Abu Jhaisha5, Walter Becker5, Peter G Jones6, Lutz Preu1, Knut Baumann1, Stefan Knapp2, Laurent Meijer3, Conrad Kunick1. 1. †Institut für Medizinische und Pharmazeutische Chemie, Technische Universität Braunschweig, Beethovenstraße 55, 38106 Braunschweig, Germany. 2. ‡Nuffield Department of Clinical Medicine, Structural Genomics Consortium, University of Oxford, Old Road Campus Research Building, Roosevelt Drive, Headington, Oxford OX3 7DQ, U.K. 3. §ManRos Therapeutics, Perharidy Research Center, 29680 Roscoff, Bretagne, France. 4. ∥"Protein Phosphorylation and Human Disease" Group, Station Biologique de Roscoff, CNRS, 29680 Roscoff, France. 5. ⊥Institute of Pharmacology and Toxicology, RWTH Aachen University, Wendlingweg 2, 52074 Aachen, Germany. 6. #Institut für Anorganische und Analytische Chemie, Technische Universität Braunschweig, Hagenring 30, 38106 Braunschweig, Germany.
Abstract
The protein kinase DYRK1A has been suggested to act as one of the intracellular regulators contributing to neurological alterations found in individuals with Down syndrome. For an assessment of the role of DYRK1A, selective synthetic inhibitors are valuable pharmacological tools. However, the DYRK1A inhibitors described in the literature so far either are not sufficiently selective or have not been tested against closely related kinases from the DYRK and the CLK protein kinase families. The aim of this study was the identification of DYRK1A inhibitors exhibiting selectivity versus the structurally and functionally closely related DYRK and CLK isoforms. Structure modification of the screening hit 11H-indolo[3,2-c]quinoline-6-carboxylic acid revealed structure-activity relationships for kinase inhibition and enabled the design of 10-iodo-substituted derivatives as very potent DYRK1A inhibitors with considerable selectivity against CLKs. X-ray structure determination of three 11H-indolo[3,2-c]quinoline-6-carboxylic acids cocrystallized with DYRK1A confirmed the predicted binding mode within the ATP binding site.
The protein kinase DYRK1A has been suggested to act as one of the intracellular regulators contributing to neurological alterations found in individuals with Down syndrome. For an assessment of the role of DYRK1A, selective synthetic inhibitors are valuable pharmacological tools. However, the DYRK1A inhibitors described in the literature so far either are not sufficiently selective or have not been tested against closely related kinases from the DYRK and the CLK protein kinase families. The aim of this study was the identification of DYRK1A inhibitors exhibiting selectivity versus the structurally and functionally closely related DYRK and CLK isoforms. Structure modification of the screening hit 11H-indolo[3,2-c]quinoline-6-carboxylic acid revealed structure-activity relationships for kinase inhibition and enabled the design of 10-iodo-substituted derivatives as very potent DYRK1A inhibitors with considerable selectivity against CLKs. X-ray structure determination of three 11H-indolo[3,2-c]quinoline-6-carboxylic acids cocrystallized with DYRK1A confirmed the predicted binding mode within the ATP binding site.
Down syndrome (DS)
is one of the most frequent congenital disorders
in humans. Individuals with DS exhibit a complex phenotype of structural
alterations and intellectual disability. Many of them develop symptoms
of Alzheimer’s disease (AD) at a relatively young age, and
a high proportion of DS individuals develop dementia at a later stage.
DS is caused by a trisomy of chromosome 21 as a consequence of nondisjunction
during meiotic cell division.[1,2] Although the mechanisms
leading from trisomy 21 to the clinical disease pattern are not yet
understood, the 30 genes located in the so-called Down syndrome critical
region (DSCR; 21q22.1–22.3) of chromosome 21 are highly likely
to be associated with the disease.[3−5] There is mounting evidence
that overexpression of DYRK1A (dual-specificity tyrosine phosphorylation-regulated
kinase 1A), a protein kinase encoded by a gene located within DSCR,
contributes to mental retardation in DS. Since DYRK1A is also connected
to neurological disorders such as AD,[6,7] the early onset
of AD in DS individuals has been related to DYRK1A overexpression.[8] Two histopathological features are found in the
brains of ADpatients: extracellular β-amyloid plaques and intracellular
tangles consisting of hyperphosphorylated tau protein.[9] DYRK1A appears to connect these biochemical aberrations.[10] Hyperphosphorylation of tau by DYRK1A diminishes
its microtubule-stabilizing effects and increases aggregation.[11] The neurological impairment of mice with modified
DYRK1A expression has also been attributed to deregulated splicing,
leading to an imbalance of 3R-tau and 4R-tau isoforms.[12,13] The increased 3R-tau concentration (by a factor of up to 4) observed
in DS has been associated with changes of the neuronal cytoskeleton
and neurofibrillary degeneration.[12] DYRK1A-overexpressing
mice also exhibit increased production of Aβ peptide, which
has been attributed to phosphorylation of the amyloid precursor protein
(APP) and presenilin 1 by DYRK1A.[14,15]DYRK1A
is a member of the DYRK protein kinases (DYRK1A, DYRK1B,
DYRK2, DYRK3, DYRK4), which are part of the CMGC superfamily and share
structural similarity of the catalytic domain and a small sequence
nearby, the so-called DYRK homology box (DH-box).[16] Although DYRKs are serine/threonine kinases, autophosphorylation
occurs at a conserved tyrosine residue of the activation loop. This
autophosphorylation is constitutive and seems to be not related to
regulation.[17,18] In the light of its involvement
in DS and other neurodegenerative disorders, the inhibition of DYRK1A
with small chemical inhibitors has been suggested as a therapeutic
strategy.[19,20] Selective DYRK1A inhibitors would also be
valuable tools for the investigation of the role of DYRK1A in physiological
and pathobiochemical processes. To date, most protein kinase inhibitors
exert their activity by competing with ATP in the binding pocket of
the kinase. Since ATP binding sites of protein kinases share a common
structure with subtle differences, selectivity with regard to closely
related kinases is not easily achieved.[21,22] When selective
DYRK1A inhibitors are to be developed, special attention should therefore
be devoted to other members of the CMGC superfamily, consisting of
CDKs, MAPkinases (such as ERKs), GSK-3, and CDC-like kinases.[23]The structures and properties
of DYRK1A inhibitors have been summarized
in recent reviews.[19,24] A well established DYRK1A inhibitor
is the β-carboline alkaloidharmine (1, Chart 1), which however also shows high affinity for serotonine
and tryptamine receptor binding sites,[25] acts as monoamine oxidase A (MAO A) inhibitor,[26] and also inhibits CLKs[27] and
therefore is inappropriate for use as a cellular chemical probe. Epigallocatechin
3-gallate (EGCG) is a polyphenolic constituent of green tea reported
to inhibit DYRK1A in a non-ATP-competitive manner,[28,29] but it is chemically reactive and also interacts with numerous intracellular
signaling pathways by other mechanisms.[30] The benzothiazole derivative INDY showed potent inhibition of DYRK1A
and related kinases (DYRK1B, DYRK2, DYRK3, CLK1, CLK4, CK1, and PIM1).
A prodrug of this compound, proINDY (2), has been shown
to protect Xenopus tadpoles that overexpress DYRK1A
against head malformation during development.[27] Leucettine L41 (3), derived from the marine natural
product leucettamine B, is an ATP-competitive inhibitor of DYRKs and
CLKs that also interacts with GSK-3 and CK2 to a lower extent. It
modulates pre-mRNA splicing, protects HT22 hippocampal cells from
glutamate-induced cell death, induces autophagy, and inhibits phosphorylation
of tau on Thr212.[31−33] Similar to the agents mentioned before, the recently
reported chemical DYRK1A inhibitors comprising meridianines,[34] indirubin 5′-carboxylates,[35] thiazolo[5,4-f]quinazolines,[36] pyrido[2,3-d]pyrimidines,[37] 3,5-diaryl-7-azaindoles (DANDYs),[38] KH-CB19,[39] 2,4-bisubstituted
thiophenes,[40] and hydroxybenzothiophenes[41] either show limited selectivity against structurally
closely related DYRK and CLK kinase isoforms or were not tested on
these enzymes. In this regard, a DYRK1A inhibitor with high selectivity
over other CMGC kinases would be useful for biochemical and cellular
studies and as lead motif for the development of new pharmaceuticals
targeting neurodegenerative diseases. To identify novel hit matter,
we tested a small diverse in-house compound library against DYRK1A
and the following CMGC protein kinases: CDK1/cyclin B, CDK2/cylin
A, CDK5/p25, CK1, GSK-3, and ERK2. The only compound showing a moderate
selective DYRK1A inhibition (IC50 = 2.6 μM) was the
11H-indolo[3,2-c]quinoline-6-carboxylic
acid 5a (Table 1).
Chart 1
Three Reported DYRK1A Inhibitors
Table 1
Kinase Inhibitory Activity (IC50, μM) of 11H-Indolo[3,2-c]quinoline-6-carboxylic Acids 5a–h on a Panel of Selected CMGC Protein
Kinasesa
IC50, μM
compd
R1
R3
R4
CDK1
CDK2
CDK5
CK1
DYRK1A
GSK-3
ERK2
5a
H
H
H
>10
>10
>10
>10
2.6
>10
>10
5b
H
C(CH3)3
H
>10
>10
>10
>10
>10
>10
>10
5c
H
CF3
H
>10
>10
>10
>10
>10
>10
>10
5d
H
NO2
H
>10
>10
>10
>10
>10
0.82
>10
5e
H
CH3
H
>10
>10
>10
>10
>10
>10
>10
5f
H
COOH
H
>10
>10
>10
>10
>10
0.35
>10
5g
I
Cl
H
>10
>10
>10
>10
>10
>10
>10
5h
H
H
Cl
>10
>10
>10
>10
0.031
>10
>10
All data points for construction
of dose–response curves were recorded in triplicate. Typically,
the standard deviation of single data points was below 10%. For test
conditions refer to the section Experimental Procedures.
Synthesis and
Evaluation of Analogues
In order to generate congeners of 5a with increased
potency, we synthesized a number of analogues 5b–h, starting from the readily available 7,12-dihydroindolo[3,2-d][1]benzazepin-6(5H)-ones 4 (paullones), a class of dual GSK-3/CDK inhibitors[42−53] that undergo a radical-induced rearrangement when stirred with Co(II)
acetate and N-hydroxyphthalimide (NHPI) in the presence
of oxygen[54] (Scheme 1).
Scheme 1
Synthesis of 11H-Indolo[3,2-c]quinoline-6-carboxylic
Acids 5 by Rearrangement of Paullones 4
Reagents and conditions:[54] Co(II) acetate, NHPI, air or O2,
DMF, room temperature to 70 °C (13–88%).
Synthesis of 11H-Indolo[3,2-c]quinoline-6-carboxylic
Acids 5 by Rearrangement of Paullones 4
Reagents and conditions:[54] Co(II) acetate, NHPI, air or O2,
DMF, room temperature to 70 °C (13–88%).Evaluation of these derivatives in the kinase panel revealed
that
substituents located in 8-position of the heterocyclic scaffold eliminated
the DYRK1A inhibitory activity, suggesting steric exclusion from the
ATP-binding pocket, the most likely binding site of the inhibitors.
Compounds with polar H-bond acceptor substituents in 8-position were
moderate and selective GSK-3 inhibitors but also were inactive on
DYRK1A. Strikingly, the 10-chloro derivative 5h showed
that DYRK1A inhibitory potency increased by 2 orders of magnitude
(IC50 = 31 nM) compared to 5a without inhibiting
other kinases of the panel displayed in Table 1.All data points for construction
of dose–response curves were recorded in triplicate. Typically,
the standard deviation of single data points was below 10%. For test
conditions refer to the section Experimental Procedures.In light of the results
shown in Table 1, we hypothesized that substituents
in 10-position of the parent
scaffold induce a strong and selective inhibitory activity of DYRK1A.
Chemistry
To expand the panel of potential selective DYRK1A
inhibitors and
to gain insight into SAR, a series of 10-substituted 11H-indolo[3,2-c]quinoline-6-carboxylic acids5h–w and the heterocyclic congeners 6 and 7 were synthesized from paullones 4 (or appropriate heterocyclic paullone analogues), applying
a method analogous to Scheme 1. All paullones 4 used in this paper as starting materials were prepared from
suitable 3,4-dihydro-1H-1-benzazepine-2,5-diones 8 and phenylhydrazines 9 by standard Fischer
indole synthesis procedures described earlier (e.g., Scheme 2).[43−45,51,54−56]
Scheme 2
Synthesis of Paullones 4
Reagents and conditions: (i)
(1) AcOH, NaOAc, 70 °C, (2) AcOH, H2SO4, 70 °C. For residues R1–R4 refer
to Tables 1 and 2.
Synthesis of Paullones 4
Reagents and conditions: (i)
(1) AcOH, NaOAc, 70 °C, (2) AcOH, H2SO4, 70 °C. For residues R1–R4 refer
to Tables 1 and 2.
Table 2
Inhibition of CMGC Protein Kinases
(IC50, μM) by 11H-Indolo[3,2-c]quinoline-6-carboxylic Acids 5h–w and by Heterocyclic Analogues 6 and 7a
IC50, μM
compd
R1
R2
R4
CDK1
CDK2
CDK5
CDK9
CK1
GSK-3
CLK1
CLK2
CLK3
CLK4
DYRK1A
DYRK1B
DYRK2
DYRK3
5h
H
H
Cl
>10
>10
>100
0.150
>100
>100
0.130
0.061
>100
0.045
0.031
0.210
0.040
>100
5i
H
H
Br
>100
>100
>100
0.160
>100
>100
0.032
0.055
>100
0.035
0.020
0.080
0.016
>100
5j
H
H
I
>10
>10
>10
>10
>10
>10
0.50
5.5
>10
0.21
0.006
0.600
>10
>10
5k
H
H
OMe
>10
>10
>10
0.120
>10
>10
1.1
0.032
5l
H
OMe
F
>10
>10
>10
>10
>10
>10
0.33
0.18
>10
0.21
0.32
3.0
8.00
2.3
5m
H
OMe
Cl
>10
>10
1.000
>10
>100
0.025
0.390
>10
0.068
0.023
1.000
0.0330
>10
5n
H
OMe
Br
>10
>10
1.500
>10
3.300
0.0210
0.1100
>10
0.042
0.018
0.230
0.023
1.4
5o
H
OMe
I
>10
>10
>10
>10
>10
>10
2.0
>10
>10
2.3
0.022
>10
>10
>10
5p
H
Cl
Cl
>10
>10
>10
0.94
2.3
5q
H
OEt
Cl
>10
>10
2.2
0.24
>10
>10
0.35
2.0
>10
>10
>10
5r
I
H
F
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
5s
I
H
Cl
>10
>10
2.2
>10
>10
3.4
1.3
>10
3.6
0.18
4.9
0.22
>10
5t
I
H
Br
>10
>10
1.5
>10
8.8
1.2
0.61
>10
1.1
0.12
1.4
0.13
>10
5u
OCH3
H
Cl
>10
>10
>10
0.21
2.3
3.0
0.10
0.039
5v
3-MeOPh
H
Cl
>10
>10
>10
3.1
>10
>10
0.22
0.51
>10
>10
0.041
1.9
0.079
0.3
5w
4-MeOPh
H
Cl
>10
>10
>10
>10
>10
>10
0.30
0.056
6
na
na
na
>10
>10
>10
0.350
>10
4.300
0.053
0.17
7
na
na
na
>10
>10
>10
2.0
>10
>10
0.28
1.1
>10
0.37
0.13
1.300
>10
>10
All data points for construction
of dose–response curves were recorded in triplicate. Typically,
the standard deviation of single data points was below 10%. A space
indicates that the compound was not tested on the indicated kinase.
For test conditions refer to the section Experimental
Procedures. na = not applicable.
The esters 10n,o,s,t were synthesized by treatment of the appropriate
carboxylic
acids 5n,o,s,t with ethanolic hydrogen chloride at room temperature (Scheme 3). An X-ray structure analysis of the ester 10o unambiguously confirmed the structure which was present
as its monohydrate in the isolated crystals (Figure 1).
Scheme 3
Synthesis of Ethyl 11H-Indolo[3,2-c]quinoline-6-carboxylates 10
For substitution pattern,
refer to Table 2. Reagents and conditions:
(i) EtOH, HCl (g), room temperature, 30 min (27–75%).
Figure 1
Crystal structure of the ethyl ester 10o as its monohydrate.
Synthesis of Ethyl 11H-Indolo[3,2-c]quinoline-6-carboxylates 10
For substitution pattern,
refer to Table 2. Reagents and conditions:
(i) EtOH, HCl (g), room temperature, 30 min (27–75%).Crystal structure of the ethyl ester 10o as its monohydrate.
Biological Evaluation and
Discussion
The new derivatives 5h–w and heterocyclic
analogues 6 and 7 were tested on an extended
panel of CMGC kinases comprising several DYRK and CLK isoforms (Table 2). Evaluation of the
results yielded a consistent activity/selectivity pattern depending
on the substituent in position 10. All compounds with a 10-chloro
substitution (5h,m,p,q,s,u–w, 6) exhibited DYRK1A inhibitory activity with IC50 values in double digit nanomolar to single digit micromolar concentrations.
However, similar to other DYRK1A inhibitors, the selectivity of these
compounds was low and other DYRKs (DYRK1B, DYRK2), CLKs (CLK1, -2,
-4), and CDK9 were also inhibited at comparable concentrations. The
lack of DYRK1A/CLK1 selectivity also applied to the heterocyclic analogues 6 and 7. All tested derivatives showed selectivity
versus CDKs 1, 2, 5 and CLK3. Within series 5, substitution
of the 10-chloro substituent by fluorine led to weaker activity (cf. 5l) or complete loss (cf. 5r) of kinase inhibitory
activity. Derivatives in which the 10-chloro substituents were replaced
by bromine (e.g., 5i, 5n, 5t) showed a slightly improved DYRK1A inhibition but with diminished
selectivity versus DYRK2. It is noteworthy that the 10-bromo substituted
1-aza analogue 7 did not inhibit DYRK2 but did inhibit
CLKs 1, 2, and 4, which are frequently affected by DYRK1A inhibitors.
The introduction of a 10-iodo substituent (compounds 5j and 5o) eventually led to the desired properties. These
compounds showed strong potency (DYRK1A IC50 of 6 and 22
nM, respectively) and good selectivity, as all tested competing kinases
were inhibited by 5j and 5o in concentrations
higher by at least 2 orders of magnitude. The selectivity for DYRK1A
over DYRK1B is particularly striking, since the catalytic domains
of these kinases are very closely related on the level of the amino
acid sequence (85% sequence identity). Upon replacement of the 10-halogen
by a methoxy group the DYRK1A inhibitor 5k was produced,
which also inhibited CDK9, and therefore 10-methoxy-substituted derivatives
were not further pursued. The finding that the carboxylic acid esters 10n,o,s,t failed to
inhibit any of the tested kinases underlines the importance of the
free carboxylic acid moiety at the 6-position of the molecule.All data points for construction
of dose–response curves were recorded in triplicate. Typically,
the standard deviation of single data points was below 10%. A space
indicates that the compound was not tested on the indicated kinase.
For test conditions refer to the section Experimental
Procedures. na = not applicable.To evaluate the most potent and rather selective DYRK1A
inhibitor 5j in a cellular assay, we analyzed its effect
on the phosphorylation
of Thr434 in splicing factor 3B1 (SF3B1) in HeLa cells. Phosphorylation
of this site depends on DYRK1A and serves as a useful indicator of
the activity of the cellular endogenous DYRK1A.[57,58] In this assay 5j dose-dependently reduced the phosphorylation
of Thr434 in SF3B1, showing its DYRK1A-inhibitory effect in living
cells (Figure 2A). Next we analyzed the effect
of 5j on tau protein as a potentially relevant target
in AD. Overexpression of DYRK1A in HEK293 cells caused an increased
phosphorylation of tau at Thr212, which was inhibited by 5j with an IC50 of 2.1 μM (Figure 2B). The high concentration margin between in vitro and cellular
DYRK1A inhibition by 5j is probably the result of poor
cellular uptake. Nevertheless, the potency of 5j as an
inhibitor of tau phosphorylation is in the same range as or better
than those of other DYRK1A inhibitors that were tested in this assay.[27,59]
Figure 2
Efficacy
of compound 5j in cell-based assays. (A)
HeLa cells expressing GFP-SF3B1 were treated with compound 5j or the reference inhibitor leucettine 41 (L41) as indicated. Phosphorylation
of SF3B1 was determined by immunoblotting with a pThr434 antibody,
and expression levels of GFP-SF3B1 were assessed with GFP antibody.
The left panel shows Western blots of a representative experiment,
and the graph presents the quantitative evaluation of three experiments
as the mean values of normalized pThr434 immunoreactivity relative
to that in untreated cells ± SEM. (B) Stably transfected HEK293-tau-DYRK1A
cells[59] were treated with doxycyclin (except
lane 1, Ctrl) to induce the expression of GFP-DYRK1A and compound 5j or the reference inhibitor leucettine 41 (L41). Phosphorylation
of tau on Thr212 (pT212) was detected by immunoblotting with a phosphospecific
antibody, and expression levels of the recombinant proteins were determined
with a GFP antibody. To correct for basal tau phosphorylation not
caused by DYRK1A, the pT212 immunoreactivity in L41 treated cells
was subtracted as background. The graph presents the results of three
experiments as the mean values of pThr212 immunoreactivity relative
to that in untreated cells ± SEM. Curve fitting yielded IC50 = 0.5 μM (95% confidence interval 0.3–0.8 μM)
for the inhibition of SF3B1 phosphorylation and IC50 =
2.1 μM (95% confidence interval 1.2–3.6 μM) for
the inhibition of tau phosphorylation.
Efficacy
of compound 5j in cell-based assays. (A)
HeLa cells expressing GFP-SF3B1 were treated with compound 5j or the reference inhibitor leucettine 41 (L41) as indicated. Phosphorylation
of SF3B1 was determined by immunoblotting with a pThr434 antibody,
and expression levels of GFP-SF3B1 were assessed with GFP antibody.
The left panel shows Western blots of a representative experiment,
and the graph presents the quantitative evaluation of three experiments
as the mean values of normalized pThr434 immunoreactivity relative
to that in untreated cells ± SEM. (B) Stably transfected HEK293-tau-DYRK1A
cells[59] were treated with doxycyclin (except
lane 1, Ctrl) to induce the expression of GFP-DYRK1A and compound 5j or the reference inhibitor leucettine 41 (L41). Phosphorylation
of tau on Thr212 (pT212) was detected by immunoblotting with a phosphospecific
antibody, and expression levels of the recombinant proteins were determined
with a GFP antibody. To correct for basal tau phosphorylation not
caused by DYRK1A, the pT212 immunoreactivity in L41 treated cells
was subtracted as background. The graph presents the results of three
experiments as the mean values of pThr212 immunoreactivity relative
to that in untreated cells ± SEM. Curve fitting yielded IC50 = 0.5 μM (95% confidence interval 0.3–0.8 μM)
for the inhibition of SF3B1 phosphorylation and IC50 =
2.1 μM (95% confidence interval 1.2–3.6 μM) for
the inhibition of tau phosphorylation.The relationship between molecular structure and DYRK1A inhibition
indicated a high significance of the halogen substitution in position
10 of the title compounds. The fact that on the one hand in the series 5h–j the iodo substituted derivative 5j exhibited the highest potency and on the other hand iodine
shows the highest tendency to form halogen bonds with carbonyl oxygen
atoms hints to halogen bond involvement in this area. However, the
data collected with the derivatives described in this report are not
sufficient to substantiate such a hypothesis. For example, in the
3-methoxy substituted series 5m–o the DYRK1A inhibition is independent of halogen nature. Further
evidence against the necessity of a halogen bond in this position
is the 10-methoxy derivative 5k, which despite lacking
a halogen retains DYRK1A inhibitory activity. For a better understanding
of the ligand–target interaction mode, docking studies and
X-ray structure analyses were performed with representative congeners.
Docking Studies
Because the majority of protein kinase inhibitors
act by competition
with ATP at its binding site, we assumed this binding mode also for
inhibitors of series 5. For the construction of a detailed
model of the inhibitor–protein interaction, representatives
of series 5 were docked into a published DYRK1A structure
(PDB code 2WO6), the ATP pocket of which is occupied by the ligand DJM2005.[60] As a control, redocking DJM2005 into chain B
of the protein reproduced the original pose of the inhibitor, albeit
not as the highest ranked pose. Docking of 5i into the
ATP binding pocket produced four different poses as the highest ranked
results (Figure 3). Although lacking the classical
acceptor–donor–acceptor hydrogen bond pattern between
the hinge region of the kinase and the inhibitor, the pose in Figure 3b was assessed as the most plausible, since it not
only comprises a salt bridge between the carboxylate and the protonated
ε-amino group of the conserved lysine but, compared to the orientation
in Figure 3a, accommodates a larger part of
the molecule in the binding pocket. Another convincing argument in
favor of orientation in Figure 3b was its high
similarity to the orientation of the inhibitor IRB in DYRK2 (PDB code 3KVW, not depicted here).
Although the program used for docking (GOLD, versions 4.0 and 5.1,
respectively) was not taking halogen bonding into account, the contact
between the bromo substituent in 10-position and the carbonyl oxygen
of Leu241 as shown in Figure 3b resembled an
atom alignment found with halogen bonds. It had been reported before
that halogen bonds may replace the canonical hinge–inhibitor
hydrogen-bond interaction.[39,61−64] However, given the uncertainty encountered with docking experiments,
in this case the resulting pose was not adequate to prove the de facto
existence of a halogen bond.
Figure 3
Representative orientations of the 10 highest-ranked
docking poses
of 5i in the ATP binding site of DYRK1A (template for
docking, 2WO6): (a) ranks in the scoring list, 1–3; (b) ranks in the scoring
list, 4 and 7; (c) ranks in the scoring list, 5, 6, 10; (d) ranks
in the scoring list, 8, 9. P.1 refers to the pose ranked no. 1.
Representative orientations of the 10 highest-ranked
docking poses
of 5i in the ATP binding site of DYRK1A (template for
docking, 2WO6): (a) ranks in the scoring list, 1–3; (b) ranks in the scoring
list, 4 and 7; (c) ranks in the scoring list, 5, 6, 10; (d) ranks
in the scoring list, 8, 9. P.1 refers to the pose ranked no. 1.
Crystal Structure Analysis
In order to validate our docking hypothesis experimentally, we
cocrystallized DYRK1A with the compounds 5s, 5t, and 5j as examples of chloro, bromo, and iodo substituted
congeners (Table 3, Figure 4). In all cases the orientation in Figure 3B generated by the docking experiment was confirmed. Figure 4A–C illustrates the binding modes of these
three inhibitors in the ATP binding site. In 5s and 5t-DYRK1A complexes, the iodide groups at the position 2 (R1) were observed to partially or completely fall off from the
main inhibitor scaffold with the free radicals still located adjacent
to its original position. Overall, the crystal structure revealed
that the inhibitor was kept in position by a salt bridge between the
carboxylate and the conserved lysine 188. As expected, the binding
mode was highly conserved in all three inhibitor complexes. A water-mediated
network of hydrogen bonds involving Ser242 and Asp307 probably contributed
to the binding affinity of the ligands. The P-loop residue Phe170
flipped into the ATP binding site potentially forming aromatic end-on
stacking interactions with the inhibitor.
Table 3
Data Collection and Refinement Statistics
of Inhibitor–DYRK1A complexes
complex (pdb
id)
DYRK1A-5j (4YLJ)
DYRK1A-5s (4YLK)
DYRK1A-5t (4YLL)
Data Collection
beamline
Diamond, beamline I04-1
Diamond, beamline I04-1
Diamond, beamline I04-1
wavelength (Å)
0.9200
0.9200
0.9200
resolutiona (Å)
48.41–2.58 (2.72–2.58)
30.28–1.40 (1.48–1.40)
28.21–1.40 (1.48–1.40)
space group
C 2
C 2
C 2
cell dimensions
a (Å)
265.5
100.0
100.0
b (Å)
65.5
69.9
69.9
c (Å)
139.4
67.9
67.9
α (deg)
90.0
90.0
90.0
β (deg)
114.6
117.7
117.7
γ (deg)
90.0
90.0
90.0
no. unique reflectionsa
68 848 (10 018)
78 842 (11 510)
77 083 (11 265)
completenessa (%)
99.6 (99.9)
97.1 (97.3)
95.0 (95.4)
I/σIa
16.6 (2.2)
10.7 (3.5)
12.4 (4.4)
Rmergea
0.066 (0.884)
0.077 (0.386)
0.059 (0.268)
redundancya
6.5 (6.6)
4.4 (4.3)
3.3 (3.3)
Refinement
no. atoms in
refinement (P/L/O)b
11 439/164/385
2880/45/647
2849/22/642
Rfact (%)
19.2
15.5
15.1
Rfree (%)
22.2
17.7
17.2
Bf (P/L/O)b (Å2)
79/73/73
13/11/24
13/9/ 23
rmsd
bondc (Å)
0.011
0.016
0.015
rms deviation
anglec (deg)
1.4
1.8
1.7
Molprobity
Ramachandran
favored
96.8
96.3
96.3
Ramachandran allowed
100.0
100.0
100.0
Values in brackets show the statistics
for the highest resolution shells.
P/L/O indicate protein, ligand molecule,
and other (water and solvent molecules), respectively.
rmsd indicates root-mean-square
deviation.
Figure 4
Crystal structure analyses of the binding modes
of halogen-substituted
DYRK1A inhibitors. Accommodations of three halogen derivatives comprising
iodo (5j, A), chloro (5s, B), and bromo
(5t, C) substitutions at the position 10 within the ATP
pocket share common features of hydrogen bond network interactions
(yellow dashed lines) with the kinase. Insets show electron density
maps for the bound inhibitors, and water molecules are shown in cyan
spheres. (D) Detailed distances and C–X···O
angles (putative σ-hole angles) between the halogen groups of
the inhibitors and the carbonyl atom of gk + 3 Leu241.
In all three inhibitor
complexes the halogen in position 10 was
located at a distance between 3.3 and 3.7 Å from the backbone
carbonyl oxygen of the gatekeeper + 3 (gk + 3) Leu241 (Figure 4d). While these distances would be compatible with
the formation of halogen bonds, the C–X···O
angles (σ-hole angles) of ∼135° are far from the
optimum σ-hole angle which is 180°. For the experimental
model of halogen substituted benzenes and N-methylacetamide,
Wilcken and Boeckler et al. have demonstrated that a deviation of
more than 40° from the optimal σ-hole angle leads to virtually
complete loss of the complex formation energy.[61] Furthermore, in 5s, 5t, and 5j the distance between the halogen atoms and the carbonyl
oxygen of Leu241 is increasing in the order Cl < Br < I. This
is contrary to an expected decrease of the distance, which should
follow the expected strength of putative halogen bonds. In conclusion,
we believe that halogen bond contribution to inhibitor binding is
in these cases either nonexistent or very weak. However, the negative
belt of a halogen atom, polarized by even very weak halogen bonding,
may play a role as hydrogen acceptor motif toward the gk + 3 Leu241
backbone nitrogen. This kind of dual halogen-bond/hydrogen-bond interaction,
emanating from a single halogen atom and forming two contacts with
the hinge region, was suggested by Wilcken and Boeckler et al.[61] for the inhibitor KH-CB19 with CLK3.[39] Along these lines, the inhibitory activity of
the 10-methoxy substituted derivative 5k, which is equipotent
to its chloro analogue5h, can also be explained. The
methoxy oxygen of 5k can mimic the hydrogen bond acceptor
role of the halogen of 5h, forming an analogous hydrogen
bond to the Leu241nitrogen.Crystal structure analyses of the binding modes
of halogen-substituted
DYRK1A inhibitors. Accommodations of three halogen derivatives comprising
iodo (5j, A), chloro (5s, B), and bromo
(5t, C) substitutions at the position 10 within the ATP
pocket share common features of hydrogen bond network interactions
(yellow dashed lines) with the kinase. Insets show electron density
maps for the bound inhibitors, and water molecules are shown in cyan
spheres. (D) Detailed distances and C–X···O
angles (putative σ-hole angles) between the halogen groups of
the inhibitors and the carbonyl atom of gk + 3 Leu241.Values in brackets show the statistics
for the highest resolution shells.P/L/O indicate protein, ligand molecule,
and other (water and solvent molecules), respectively.rmsd indicates root-mean-square
deviation.With double digit
nanomolar IC50 values measured for
DYRK1A, the inhibitors reported here showed comparable potency to
other DYRK inhibitors reported in the literature. However, both 10-iodo
derivatives 5j and 5o showed superior selectivity
toward closely related kinases. Because of the potential antagonistic
functions of DYRK1 family members, the compounds constitute important
chemical tools for further biochemical studies. Comparison of our
screening data identified the 10-iodo substituent as the relevant
group for the observed selectivity. Although a plausible explanation
at a molecular level is not yet available, it is tempting to speculate
that only DYRK1A, but not the related protein kinases, is able to
accommodate the large 10-iodo substituent of the selective derivatives
in the ATP binding pocket.
Conclusion
Experimental DYRK1A inhibitors
described so far have suffered from
poor selectivity toward other closely related protein kinases, namely,
other DYRKs and members of the CLK family. We have developed two 10-iodo-11H-indolo[3,2-c]quinoline-6-carboxylic acids5j and 5o as the first selective inhibitors of
DYRK1A. Compound 5j proved also to inhibit the enzyme
in cellular assays. Docking studies and X-ray structure analyses revealed
a nonclassical binding mode in which inhibitors were oriented via
the 10-halogen substituent toward the hinge region of the kinase.
Although the 10-substituent obviously was of central importance for
the hinge–inhibitor attraction, the nature of this interaction
is not yet clear. The title compounds demonstrate that within the
CMGC family of protein kinases selective inhibition of DYRK1A is possible.
With the aim of developing compounds that are better suited for cellular
assays or animal disease models, further studies in this compound
class will be directed to increase solubility and cell permeability.
Experimental Procedures
Synthetic Chemistry
Starting materials were purchased
from Acros Organics (Geel, Belgium) or Sigma-Aldrich (Steinheim, Germany).
Solvents were used as commercially available grades for synthesis,
with the exceptions of toluene, THF, CH2Cl2,
and diethyl ether, which were dried and purified by published methods.[65]5b, 5c, and 5f were prepared as reported previously.[54] Synthetic procedures and structure characterization data
for the following compounds are available in the Supporting Information: 5a, 5d–e, 5g–i, 5k–n, 5p–w, 6, 7, 10n,s,t.Melting
points (mp) were determined on an electric variable heater (Electrothermal
IA 9100, Barnstedt International, Southend-on-Sea, U.K.) in open glass
capillaries. IR spectra were recorded as KBr discs on a Thermo Nicolet
FT-IR 200. 1HNMR spectra and 13CNMR spectra
were recorded on the following instruments: Bruker Avance DRX-400,
Bruker Avance III-400, and Bruker Avance II-600 (Bruker, Billerica,
MA, USA); internal standard tetramethylsilane; signals in ppm (δ
scale). Signals in 13C spectra were assigned based on the
results of 13C DEPT135 experiments (NMR Laboratories of
the Chemical Institutes of the Technische Universität Braunschweig).
Elemental analyses were determined on a CE Instruments FlashEA 1112
elemental analyzer (Thermo Quest). Mass spectra were recorded on a
Finnigan-MAT 95 (Thermo Finnigan MAT, Bremen, Germany). Accurate measurements
were conducted according to the peak match method using perfluorokerosene
(PFK) as an internal mass reference. (EI) MS: ionization energy 70
eV (Department of Mass Spectrometry of the Chemical Institutes of
the Technische Universität Braunschweig). TLC: Polygram Sil
G/UV254, Macherey-Nagel, 40 mm × 80 mm, visualization
by UV illumination (254 and 366 nm). Purity was determined by HPLC
using the following devices and settings. Isocratic elution: Elite
LaChrom (Merck/Hitachi), pump L-2130, autosampler L-2200, diode array
detector L-2450, organizer box L-2000; column, Merck LiChroCART 1254,
LiChrosphere 100, RP 18, 5 μm; flow rate 1.000 mL/min; volume
of injection, 10 μL; detection (DAD) at 254 and 280 nm; AUC,
% method; time of detection 15 min, net retention time (tN), dead time (tm) related
to DMSO. Gradient elution: Elite LaChrom (Merck/Hitachi), pump L-2130,
autosampler L-2200, UV detector L-2400, organizer box L-2000; column,
Merck LiChroCART 125-4, LiChrosphere 100, RP 18, 5 μm; flow
rate 1.000 mL/min; volume of injection, 10 μL; detection at
254 nm; AUC, % method; net retention time (tN), dead time (tm) related to DMSO.
For all gradient runs, mixtures of ACN and water or aqueous formic
or trifluoroacetic acid were used as specified for particular compounds.
Preparation of H2O + (Et3NH)2SO4 buffer (pH 2.6) for isocratic HPLC: triethylamine (20.0 mL)
and sodium hydroxide (242 mg) are dissolved in water to 1 L. The solution
is adjusted to pH 2.6 by addition of sulfuric acid. All compounds
that were biologically tested were of >95% purity with the exception
of 5h (HPLC, 93.6% at 254 nm and 94.8% at 280 nm detection
wavelength) and 5w (HPLC, 95.1% at 254 nm and 94.2% at
280 nm detection wavelength). Absorption maxima (λmax) were extracted from the spectra recorded by the DAD in the HPLC
peak maxima in isocratic runs (software, EZ Chrom Elite Client/server,
version 3.1.3.).
General Procedure A for the Fischer Indole
Synthesis of Paullones 4 Used as Starting Materials for
Subsequent Syntheses
An appropriate substituted 3,4-dihydro-1H-1-benzazepine-2,5-dione
(1 equiv) and an appropriately substituted phenylhydrazine hydrochloride
(1.2–1.5 equiv) and sodium acetate (1.2–1.5 equiv) are
stirred in acetic acid (10 mL) at 70 °C for the indicated time.
Sulfuric acid (0.1 mL) is added, and stirring at 70 °C is continued
for the indicated time. After cooling to room temperature, the mixture
is poured into 5% aqueous sodium acetate solution (20 mL) and kept
at 8 °C for 2 h. A precipitate is formed, which is filtered off
with suction, washed with water and petrol ether. Purification is
accomplished by column chromatography or crystallization.
General Procedure
B1 for Oxidative Ring Contraction of Paullones 4 To Yield
11H-Indolo[3,2-c]quinoline-6-carboxylic
Acids 5j and 5o
The appropriate
paullone 4 (1 equiv), NHPI
(2 equiv), and cobalt(II) acetate tetrahydrate (0.25 equiv) are dissolved
in the given volume of DMF. Oxygen is bubbled through the stirred
mixture for the given time at the indicated temperature. The reaction
is monitored by TLC. The precipitate is filtered off and washed with
water. If no precipitate appears, water (5 mL) is added and the resulting
precipitate is separated and purified by centrifugation, decanting,
and washing with water. The resulting solid is washed with a small
amount of acetone. The predried product is finally dried at 130 °C
in vacuo.
General Procedure C for the Syntheses of
Ethyl 11H-Indolo[3,2-c]quinoline-6-carboxylates 10n,o,s,t by Esterification
Ethanol (100 mL) is added to a solution of an appropriate 11H-indolo[3,2-c]quinoline-6-carboxylic acid 5 in DMSO (5 mL). Hydrogen chloride gas is bubbled through
the solution for 30 min at room temperature. After removal of the
ethanol by evaporation, water (100 mL) is added and the mixture is
kept at 8 °C for 1 h. The resulting precipitate is filtered off,
washed with water, and dissolved in acetone (5 mL). After addition
of silica gel (2 g), the mixture is evaporated. The product is subsequently
eluted from the silica gel by flash chromatography using the indicated
eluent. Finally the product is dried at 100 °C in vacuo.
Synthesis
was according to general procedure C from 10-iodo-3-methoxy-11H-indolo[3,2-c]quinoline-6-carboxylic acid
(5o, 120 mg, 0.300 mmol). Purification by column chromatography
(toluene–ethyl acetate–diethylamine 2:2:1) yielded pale
yellow crystals (100 mg, 75%); mp 214–217 °C. IR (KBr):
3384 (NH), 1726 cm–1 (C=O). 1HNMR (DMSO-d6, 400.4 MHz): δ (ppm)
= 1.45 (t, 3H, J = 7.1 Hz), 3.99 (s, 3H), 4.59 (q,
2H, J = 7.1 Hz), 7.14 (t, 1H, J =
7.8 Hz), 7.46 (dd, 1H, J = 9.1/2.6 Hz), 7.63 (d,
1H, J = 2.6 Hz), 7.93 (dd, 1H, J = 7.6/1.0 Hz), 8.33 (dd, 1H, J = 8.1/0.9 Hz), 9.05
(d, 1H, J = 9.1 Hz), 12.34 (s, 1H, NH). 13CNMR (DMSO-d6, 100.7 MHz): δ (ppm)
= 14.2 (CH2–CH); 55.5 (OCH3); 61.7 (CH2); 108.7, 119.1, 122.0, 122.5, 124.7, 135.1 (CH); 76.8, 111.2, 111.8,
121.4, 141.1, 142.0, 145.1 146.0, 160.1 (C); 166.7 (C=O). C19H15IN2O3 (446.24); calcd
C 51.14, H 3.39, N 6.28, found C 51.21, H 3.41, N 6.01. MS (EI): m/z (%) = 446 [M]+• (22),
374 [M – C3H4O2]+ (100). HPLC (isocr): 99.7% at 254 nm and 99.8% at 280 nm, tN = 7.6 min, tM =
1.0 min (ACN/H2O 80:20). λmax: 243 and
286 nm. HPLC (grad): 99.7% at 254 nm, tN = 14.1 min, tM = 1.1 min (gradient,
ACN/H2O; 0 min, 10/90 → 13 min, 90/10 (linear);
20 min, 90/10). Crystals for X-ray structure analysis were prepared
by crystallization from ethanol (96%).
X-ray Crystal Structure
Analysis of 10o
Crystal data for 10o·H2O at 100 K: orthorhombic, P212121, a = 6.36362(12), b = 15.0401(3), c = 18.3209(4) Å, V = 1753.48 Å3, Z = 4.
A yellow irregular crystal 0.3 mm ×
0.2 mm × 0.15 mm was used to record 76097 intensities, 5096 independent
(Rint = 0.036) on an Oxford Diffraction
Xcalibur E diffractometer using Mo Kα radiation (λ = 0.710 73
Å, 2θmax = 60°). The structure was refined
anisotropically on F2 using the program
SHELXL-97 to wR2 = 0.049, R1 = 0.020 for 249 parameters; S = 1.12, max. Δρ = 1.4 e Å–3.
Crystallographic data have been deposited with the Cambridge Crystallographic
Data Centre as supplementary publication no. CCDC-1036693. Copies
of the data can be obtained free of charge from www.ccdc.cam.ac.uk/data_request/cif.
Kinase Expression and Activity Assays. Protein Kinase Assays
Buffer A consisted of 10 mM MgCl2, 1 mM EGTA, 1 mM DTT,
25 mM Tris-HCl, pH 7.5, 50 μg heparin/mL. Buffer C consisted
of 60 mM β-glycerophosphate, 15 mM p-nitrophenylphosphate,
25 mM MOPS (pH 7.2), 15 mM EGTA, 15 mM MgCl2, 1 mM DTT,
1 mM sodium vanadate, 1 mM phenylphosphate.Kinase activities
were assayed in buffer A or C at 30 °C at a final ATP concentration
of 15 μmol/L. Blank values were subtracted, and activities were
expressed in percent of the maximal activity, i.e., in the absence
of inhibitors. Controls were performed with appropriate dilutions
of DMSO. The GS-1, CKS, CDK7/9 tide, and RS peptide substrates were
obtained from Proteogenix (Oberhausbergen, France).CDK1/cyclin
B (M phase starfish oocytes, native), CDK2/cyclin A,
and CDK5/p25 (human, recombinant) were prepared as previously described.[68] Their kinase activity was assayed in buffer
A, with 1 mg of histone H1/mL, in the presence of 15 μmol/L
[γ-33P] ATP (3000 Ci/mmol; 10 mCi/mL) in a final volume of 30
μL. After a 30 min incubation at 30 °C, the reaction was
stopped by harvesting onto P81phosphocellulose supernatant (Whatman)
using a FilterMate harvester (Packard) and washing in 1% phosphoric
acid. Scintillation fluid was added and the radioactivity measured
in a Packard counter.CDK9/cyclin T (human, recombinant, expressed
in insect cells) was
assayed as described for CDK1/cyclin B but using CDK7/9 tide (YSPTSPSYSPTSPSYSPTSPSKKKK)
(8.1 μg/assay) as a substrate.GSK-3 (porcine brain, native,
affinity purified on axin1-Sepharose
beads) was assayed, as described for CDK1, with 0.5 mg of BSA/mL +
1 mM DTT and using a GSK-3 specific substrate (GS-1, YRRAAVPPSPSLSRHSSPHQSpEDEEE)
(pS stands for phosphorylated serine).[69]CK1δ/ε (porcine brain, native, affinity purified
on
axin2-Sepharose beads) was assayed as described for CDK1 but in buffer
C and using 25 μM CKS peptide (RRKHAAIGpSAYSITA), a CK1-specific
substrate.[70]CLK1, -2, -3, and -4
(mouse, recombinant, expressed in E. coli as GST
fusion proteins) were assayed as described
for CDK1/cyclin B with 0.5 mg BSA/mL + 1 mM DTT and RS peptide (GRSRSRSRSRSR)
(1 μg/assay) as a substrate.DYRK1A, -1B, -2, -3 (human,
recombinant, expressed in E.
coli as GST fusion proteins) and CLK1, -2, -3, and -4 (mouse,
recombinant, expressed in E. coli as GST fusion proteins)
were assayed in buffer A (supplemented extemporaneously with 0.15
mg of BSA/Ml + 1 mM DTT) with 1 μg of RS peptide (GRSRSRSRSRSR)
as a substrate.All data points for construction of dose–response
curves
were recorded in triplicate. Typically, the standard deviation of
single data points was below 10%.
Inhibition of Cellular
DYRK1A Activity
The assay for
inhibition of SF3B1 phosphorylation was performed as described previously.[58] For the assay of tau phosphorylation, we used
a HEK293 subclone with regulatable expression of GFP-DYRK1A and constitutive
expression of GFP-tau (HEK293-tau-DYRK1A) that was kindly provided
by Dr. Matthias Engel (Department of Pharmaceutical and Medicinal
Chemistry, Saarland University, Saarbrücken, Germany).[59] Cells were grown overnight in six-well plates
before expression of GFP-DYRK1A was induced with 2 μg/mL doxycyclin.
The inhibitors were then added from stock solutions in DMSO to the
desired final concentration and cells were further incubated for 20
h. Cells were lysed in SDS lysis buffer (20 mM Tris-HCl, pH 7.4, 1%
SDS). Samples were sonicated and cleared by centrifugation before
SDS–PAGE and immunoblotting with a goat antibody for GFP (no.
600-101-215, Rockland Immunochemicals, Gilbertsville, PA, USA) and
a phosphorylation state specific antibody directed against pThr212
in the tau protein (no. 44740G, Invitrogen, Camarillo, CA, USA). Immunoreactivities
were detected by enhanced chemiluminescence using HRP-coupled secondary
antibodies and quantified using the AIDA Image Analyzer 5.0 program
(Raytest, Straubenhardt, Germany). Relative tau phosphorylation was
calculated by normalization to total tau expression, as determined
from GFP immunoreactivity. To calculate relative DYRK1A activity,
the basal pT212 signal in control cells not treated with doxycyclin
was subtracted from all values, and the phosphorylation in DYRK1A
expressing cells not treated with inhibitors was set to 100%. Curve
fitting for IC50 determination was done with the help of
the GraphPad Prism 5.0 program (GraphPad Software, La Jolla, CA, USA).
Crystallography and Molecular Modeling. Protein Production,
Crystallization of DYRK1A–Inhibitor Complexes, Data Collection,
and Structure Determination
Recombinant DYRK1A was purified
as previously described[60] and was treated
with TEV protease to remove the N-terminal His6 tag. The
kinase at ∼13–15 mg/mL in 50 mM HEPES, pH 7.5, 500 mM
NaCl and 5 mM DTT was incubated with the inhibitors at 1 mM prior
to crystallization. Crystals were obtained using the sitting drop
vapor diffusion method at 4 °C using either 2 M ammonium sulfate,
0.2 M Na/K tartrate, 0.1 M citrate, pH 5.6 (for 5s and 5t), or 37% PEG 400, 0.2 M lithium sulfate, 0.1 M Tris, pH
8.8 (for 5j), as the reservoir solutions. Diffraction
data collected at Diamond Light Source, beamline I04-1, were processed
with XDS[71] or mosflm[72] and subsequently scaled with Scala from CCP4 suite.[73] Structures were determined by molecular replacement
method using Phaser[74] and the coordinates
of DYRK1A structure[60] as a search model.
Iterative cycles of manual model building in COOT[75] alternated with refinement using Refmac,[76] and a TLS model calculated from TLSMD server[77] was performed. The geometric correctness of
the final model was verified with MolProbity.[78] Data collection and refinement statistics are summarized in Table 3. DYRK1A-ligand complexes (pdb id): DYRK1A-5j (4YLJ); DYRK1A-5s (4YLK); DYRK1A-5t (4YLL).
Docking
Molecular
docking was performed using GOLD,[79] version
4.0, running under Linux Ubuntu Dapper
Drake. The poses obtained with GOLD 4.0 were later reproduced using
GOLD, version 5.1. The PDB file 2WO6 was downloaded from the Protein Data
Bank. Only the B chain present in the structure was used for docking.
The wizard integrated in GOLD was used for loading the structure into
the user interface HERMES. Hydrogen atoms were added, and the protonation/tautomerization
status of amino acid side chains within the ATP binding site was checked
and adjusted manually, if necessary. Ligand and water molecules were
deleted from the structure. A zone of 10 Å around the original
ligand was defined as relevant for binding. The options “flip
side chains of Asn/Gln” and “alter tautomers of His”
were switched off. Ligands were constructed with MOE,[80] energy minimized, and saved as mol2 files. The option “save
lone pairs” was switched off and “search efficiency”
was set to 200%. Ten GA runs were performed without constraints and
“allow early termination” option switched off. Scoring
of resulting poses was performed applying the integrated chemscore.kinase
template. Redocking of the originally bound ligand DJM2005 into chain
B of the protein reproduced the original pose of the inhibitor, albeit
not as highest ranked pose.
Authors: Z Rahmani; J L Blouin; N Creau-Goldberg; P C Watkins; J F Mattei; M Poissonnier; M Prieur; Z Chettouh; A Nicole; A Aurias Journal: Proc Natl Acad Sci U S A Date: 1989-08 Impact factor: 11.205
Authors: T Geoff G Battye; Luke Kontogiannis; Owen Johnson; Harold R Powell; Andrew G W Leslie Journal: Acta Crystallogr D Biol Crystallogr Date: 2011-03-18
Authors: M K McCormick; A Schinzel; M B Petersen; G Stetten; D J Driscoll; E S Cantu; L Tranebjaerg; M Mikkelsen; P C Watkins; S E Antonarakis Journal: Genomics Date: 1989-08 Impact factor: 5.736
Authors: Oleg Fedorov; Kilian Huber; Andreas Eisenreich; Panagis Filippakopoulos; Oliver King; Alex N Bullock; Damian Szklarczyk; Lars J Jensen; Doriano Fabbro; Jörg Trappe; Ursula Rauch; Franz Bracher; Stefan Knapp Journal: Chem Biol Date: 2011-01-28
Authors: Katrin de Graaf; Hanna Czajkowska; Sabine Rottmann; Len C Packman; Richard Lilischkis; Bernhard Lüscher; Walter Becker Journal: BMC Biochem Date: 2006-03-02 Impact factor: 4.059
Authors: Airlie J McCoy; Ralf W Grosse-Kunstleve; Paul D Adams; Martyn D Winn; Laurent C Storoni; Randy J Read Journal: J Appl Crystallogr Date: 2007-07-13 Impact factor: 3.304
Authors: Kunal Kumar; Peng Wang; Roberto Sanchez; Ethan A Swartz; Andrew F Stewart; Robert J DeVita Journal: J Med Chem Date: 2018-08-21 Impact factor: 7.446
Authors: Kunal Kumar; Peter Man-Un Ung; Peng Wang; Hui Wang; Hailing Li; Mary K Andrews; Andrew F Stewart; Avner Schlessinger; Robert J DeVita Journal: Eur J Med Chem Date: 2018-08-22 Impact factor: 6.514
Authors: Jan S Kramer; Stefano Woltersdorf; Thomas Duflot; Kerstin Hiesinger; Felix F Lillich; Felix Knöll; Sandra K Wittmann; Franca-M Klingler; Steffen Brunst; Apirat Chaikuad; Christophe Morisseau; Bruce D Hammock; Carola Buccellati; Angelo Sala; G Enrico Rovati; Matthieu Leuillier; Sylvain Fraineau; Julie Rondeaux; Victor Hernandez-Olmos; Jan Heering; Daniel Merk; Denys Pogoryelov; Dieter Steinhilber; Stefan Knapp; Jeremy Bellien; Ewgenij Proschak Journal: J Med Chem Date: 2019-09-17 Impact factor: 7.446
Authors: Mir Mohammad Shahroz; Hemant Kumar Sharma; Abdulmalik S A Altamimi; Mubarak A Alamri; Abuzer Ali; Amena Ali; Safar Alqahtani; Ali Altharawi; Alhumaidi B Alabbas; Manal A Alossaimi; Yassine Riadi; Ahmad Firoz; Obaid Afzal Journal: Molecules Date: 2022-02-09 Impact factor: 4.411
Authors: Alfredo Picado; Apirat Chaikuad; Carrow I Wells; Safal Shrestha; William J Zuercher; Julie E Pickett; Frank E Kwarcinski; Parvathi Sinha; Chandi S de Silva; Reena Zutshi; Shubin Liu; Natarajan Kannan; Stefan Knapp; David H Drewry; Timothy M Willson Journal: J Med Chem Date: 2020-11-20 Impact factor: 7.446
Authors: Katharina Rüben; Anne Wurzlbauer; Agnes Walte; Wolfgang Sippl; Franz Bracher; Walter Becker Journal: PLoS One Date: 2015-07-20 Impact factor: 3.240
Authors: Jochem M G Evers; Roman A Laskowski; Marta Bertolli; Jill Clayton-Smith; Charu Deshpande; Jacqueline Eason; Frances Elmslie; Frances Flinter; Carol Gardiner; Jane A Hurst; Helen Kingston; Usha Kini; Anne K Lampe; Derek Lim; Alison Male; Swati Naik; Michael J Parker; Sue Price; Leema Robert; Ajoy Sarkar; Volker Straub; Geoff Woods; Janet M Thornton; Caroline F Wright Journal: Hum Mol Genet Date: 2017-02-01 Impact factor: 6.150
Chart 1
Scheme 1
Scheme 2
Scheme 3
Figure 1
Figure 2
Figure 3
Figure 4