Liam Hudson1, James Mui1, Santiago Vázquez2, Diana M Carvalho1, Eleanor Williams3, Chris Jones1, Alex N Bullock3, Swen Hoelder1. 1. Institute of Cancer Research , 15 Cotswold Road , Sutton , Surrey SM2 5NG , United Kingdom. 2. Laboratori de Química Farmacèutica (Unitat Associada al CSIC), Facultat de Farmàcia i Ciències de l'Alimentació, and Institute of Biomedicine (IBUB) , Universitat de Barcelona , Av. Joan XXIII s/n , Barcelona E-08028 , Spain. 3. Structural Genomics Consortium , University of Oxford , Old Road Campus Research Building, Roosevelt Drive , Oxford OX3 7DQ , United Kingdom.
Abstract
Structure-activity relationship and crystallographic data revealed that quinazolinone-containing fragments flip between two distinct modes of binding to activin receptor-like kinase-2 (ALK2). We explored both binding modes to discover potent inhibitors and characterized the chemical modifications that triggered the flip in binding mode. We report kinase selectivity and demonstrate that compounds of this series modulate ALK2 in cancer cells. These inhibitors are attractive starting points for the discovery of more advanced ALK2 inhibitors.
Structure-activity relationship and crystallographic data revealed that quinazolinone-containing fragments flip between two distinct modes of binding to activin receptor-like kinase-2 (ALK2). We explored both binding modes to discover potent inhibitors and characterized the chemical modifications that triggered the flip in binding mode. We report kinase selectivity and demonstrate that compounds of this series modulate ALK2 in cancer cells. These inhibitors are attractive starting points for the discovery of more advanced ALK2 inhibitors.
ALK2 (gene: ACVR1)
is a serine/threonine kinase in the bone morphogenetic
protein (BMP) pathway and one of seven (ALK1–7) type-I receptors
in the BMP and transforming growth factor beta (TGFβ) signaling
pathways.[1]Signaling through ALK2
is deregulated in two disease contexts.
The first is fibrodysplasia ossificans progressiva (FOP), an extremely
rare condition of ectopic bone formation resulting in a progressive
loss of mobility.[2] Heterozygous missense
mutations in ACVR1 (most commonly R206H) lead to neofunction in response
to activin A as well as increased BMP signaling through SMAD1/5/8—the
driving force in FOP.[3] The second disease
is diffuse intrinsic pontine glioma (DIPG), a highly infiltrative
tumor originating in the pons of the brainstem.[4] DIPGs arise with a peak age of incidence of 6–7
years and have a fatality of 100%; median survival is 9–12
months.[4] As DIPGs grow diffusely throughout
the vital midline brain region, surgical resection is generally considered
impossible, and the current treatment of radiotherapy, while providing
short-term relief of symptoms, does not prevent rapid disease progression.[5]ACVR1 mutations are observed in 24% of
DIPG patients.[4] These mutations occur in
the cytoplasmic domains
of ALK2 and modulate kinase activity though (i) destabilizing the
inactive conformation of the kinase and (ii) disrupting the binding
of a negative regulator protein, FKBP12.[2] The high frequency of ALK2 mutations in DIPG strongly suggests a
contribution to disease phenotype.ALK2 inhibitors have been
reported and fall into two series (Figure ). The first contains
a pyrazolo[1,5-a]pyrimidine core and derives from
dorsomorphin. Dorsomorphin was initially reported as an adenosine
monophosphate (AMP)-activated protein kinase inhibitor, but was found
in a zebrafish embryo dorsalization assay to selectively inhibit BMP
signaling through SMAD1/5/8.[6] Further development
led to LDN-193189,[7]LDN-212854,[8] and ML347,[9] which had improved microsomal stability,
potency, and selectivity. In addition, these molecules demonstrated
efficacy in mouse models of FOP.[8] However,
these compounds have a number of kinase off-targets and display dose-limiting
toxicity with a 10% loss in body weight in animal models.[10] A second series of inhibitor based on a pyridine
core (e.g., K02288(11) and LDN-214117(12)), with equivalent
biochemical potency and improved kinome selectivity, has also been
reported (Figure ).
Figure 1
A selection
of ALK2 and ALK5 inhibitors (the atom marked in red
is the main anchor point to the kinase hinge residues).
A selection
of ALK2 and ALK5 inhibitors (the atom marked in red
is the main anchor point to the kinase hinge residues).
Results
We identified 6-pyrazole
quinazolinone, 1, as a ligand
efficient inhibitor of ALK2 (IC50 = 8.2 μM; LE =
0.45) through cross-screening of a focused kinase fragment library,
using Invitrogen’s LanthaScreen binding assay. 1 shares features with reported ALK5 inhibitors such as PF-03671148(13) and compound 19.[14] These are known to bind to the hinge of ALK5
through a single polar contact at the N-1 position
of the quinazolinone moiety, with the 2-methylpyridine directing toward
the ALK5 Ser280 gatekeeper residue, forming a key water mediated hydrogen
bond to Lys232 (Figure ).[14] Interestingly, PF-03671148 has been shown to be selective against ALK1 likely due a larger
gate keeper (Thr) causing a clash with the pyridine substituent.[13] Given that ALK2 also features a threonine gatekeeper
residue (Thr283), it is also unlikely to be potently inhibited by PF-03671148.
Figure 2
(A) Docking pose of 4 in ALK2 (structure used: 3Q4U).
(B) X-ray structure of analogous naphthyridine-based inhibitor (compound 19) of ALK5[14] (PDB 1VJY).
(A) Docking pose of 4 in ALK2 (structure used: 3Q4U).
(B) X-ray structure of analogous naphthyridine-based inhibitor (compound 19) of ALK5[14] (PDB 1VJY).Unfortunately, attempts to solve the structure
of 1 bound to ALK2 were not successful. However, since 1 did not feature the pyridine substituent of the published
ALK5 inhibitor
(compound 19), we initially hypothesized that it explored
a similar binding mode when binding to ALK5 (Figure ).We sought to investigate whether
quinazolinone 1 could
be optimized into an independent series of ALK2 inhibitor and characterize
its binding mode.We started our investigations at the quinazolinone
6-position,
by modifying the pyrazole moiety (Table ). N-Methylation of pyrazole
(2) had a minor negative effect on ALK2 binding affinity,
whereas substitution at the pyrazole 3-position was tolerated. Interestingly,
increasing the size from methyl to cyclopropyl (4) was
not only tolerated but led to a 21-fold potency increase. A 3,5-dimethylpyrazole
(6) maintained sub-μM potency and was the most
efficiently binding pyrazole derivative with an impressive LE of 0.51.
This initial set of compounds, particularly compounds 2 and 6, suggested that the pyrazole was not the hinge
binding motif and instead binds to Lys235 through a water bridge,
as observed for compound 19 in Figures and 2. This prompted
us to replace the pyrazole for a bicycle (7–9). Since DIPG requires brain penetrable drugs, this had the
added benefit of reducing the TPSA, a well-established predictor for
passive permeability across the blood–brain barrier.[15] Gratifyingly, this afforded a 40–80-fold
improvement in potency over the initial hit. As the 4- and 5-quinoline
examples (7 and 8) represented the most
promising compounds so far, with sub-μM potency, good ligand
efficiency and low TPSA, we decided to focus further modification
on these isomers.
Table 1
SAR at Quinazolinone 6-Positiona
IC50 data is an average
of 2–4 measurements by LanthaScreen Eu kinase binding assay.
IC50 data is an average
of 2–4 measurements by LanthaScreen Eu kinase binding assay.On the basis of the observed
structure–activity relationship
(SAR), we hypothesized that the compounds presented in Table adopted a binding mode in which
the variable heteroaromatic group occupied the ALK2 central pocket,
extending toward the catalytic lysine (Lys235), similarly to reported
ALK5 inhibitors (Figure ).[14] We also expected that this binding
mode would be preserved between the pyrazoles and quinolines, which
are present in several reported ALK2 inhibitors (Figure ). However, it was still not
clear how the quinazolinone core interacted with the hinge region.
In an effort to shed light on this question, we positioned a methyl
group at various positions around the quinazolinone ring (Table ).
Table 2
Quinazolinone Methyl Scan SARa
IC50 data is an average
of 2–4 measurements by LanthaScreen Eu kinase binding assay.
IC50 data is an average
of 2–4 measurements by LanthaScreen Eu kinase binding assay.When compared to the desmethyl
analogues (7 and 8), a 5-methyl group (10 and 11)
did not alter potency. This was consistent with the binding mode suggested
in Figure , where
the additional methyl group would direct toward solvent. 3-Methyl
isomers 12 and 13 were also of equivalent
potency, which again was consistent with a binding mode as in Figure where the 3-methyl
directs toward the solvent channel of ALK2. Surprisingly, analogous
2-methyl isomers (14 and 15) also displayed
equal potency at ALK2. This was an unexpected result given that a
2-position substituent should clash with protein under the proposed
binding model.While this surprising tolerance to substitution
at all three positions
(Table ) did not clearly
suggest a preferred hinge binding motif, the 3-methyl isomer 12 stood out due to the low TPSA and closest analogy to published
inhibitors.[7]We decided to explore
further substitution at the quinazolinone
3-position and incorporated solvent channel groups that had led to
activity gains in other series, e.g., the pyrazolo[1,5-a]pyrimidine series. Introduction of a 4-morpholinophenyl group (16) was found to increase the potency 5-fold (Table ) and yielded the most potent
compound up to this point. However, the additional 12 heavy atoms
from the 4-morpholinophenyl group did inflict a 0.11–0.19 LE
unit penalty, suggesting that the solvent channel group was suboptimal.
Table 3
SAR at Quinazolinone 3-Positiona
IC50 data is an average
of 2–4 measurements by LanthaScreen Eu kinase binding assay.
IC50 data is an average
of 2–4 measurements by LanthaScreen Eu kinase binding assay.To understand why only modest
potency gains were afforded a series
of truncated and modified analogues were prepared: phenyl (17), cyclohexyl (18), 3-morpholinopropyl (19), and 4-dimethyl aniline (20). Comparison of these
analogues showed that the phenyl ring alone leads to a loss of activity
compared to the methyl derivative, 12. Replacement by
a cyclohexyl group did not drastically reduce binding further, suggesting
that saturated or otherwise three-dimensional groups may be tolerated
in this region of the ALK2 pocket. The 30-fold gain in potency achieved
through the addition of morpholine (16) to phenyl-only
compound 17 suggests that the phenyl ring is acting as
a linker for this polar function. However, a simple propyl linker
(19) to the morpholine unit does not afford a similar
potency boost. In addition, maintaining the phenyl ring, but removing
the ether functionality [of 16], leaving a dimethyl aniline
(20) had a negligible effect on binding—strongly
suggesting that the amine, and its precise position, are key to achieving
potent inhibitors of ALK2.With more potent compounds in hand,
we reattempted cocrystallization
and indeed managed to obtain the cocrystal structure of 16 with ALK2 to 2.2 Å resolution (Figure ). The compound indeed displayed a binding
mode similar to the reported inhibitor LDN-193189—quinazolinone N-1 interacts as an HBA for the ALK2 hinge residue His286,
the 4-morpholino phenyl group directs through the solvent channel,
and the 4-quinoline occupies the central pocket, with the quinoline-N-atom forming a water-bridged interaction to Lys235. Other
similarities between the solved structure of 16 and LDN-193189 include the variable position of morpholine and
methylpiperidine in the solvent channel—multiple conformations
exist for both compounds in their respective asymmetric units suggesting
that while the phenyl-bonded N atoms may have a significant role in
binding the distal N-Me/O atom is unlikely to contribute heavily.
We can only speculate on the reason why the morpholine amine contributes
so strongly to binding. It is unlikely that this amine is charged
due to the aniline character and electron withdrawing ether function.
However, upon inspection of the electron density map for 16 a water bridged hydrogen bond to V214 is visible in one ALK2 chain
in the asymmetric unit and may contribute to the gain in potency (Figure A). A notable difference
versus available structures of pyrazolo[1,5-a]pyrimidine-based
ALK2 inhibitors (PDB 3Q4U) is that the central pocket quinoline penetrates deeper toward Lys235
(Figure B), by virtue
of the larger hinge-binding core.
Figure 3
(A) Cocrystal structure of 16 in ALK2 showing three
H-bond acceptors for H286, K235, and V214 (PDB 6GIN). (B) Overlay of 16 (brown) with LDN-193189 (green, PDB 3Q4U) showing deeper
penetrance of 4-quinoline toward K235 for 16, and slight
change in position of the solvent channel group.
(A) Cocrystal structure of 16 in ALK2 showing three
H-bond acceptors for H286, K235, and V214 (PDB 6GIN). (B) Overlay of 16 (brown) with LDN-193189 (green, PDB 3Q4U) showing deeper
penetrance of 4-quinoline toward K235 for 16, and slight
change in position of the solvent channel group.The binding mode revealed by the crystal structure was consistent
with our earlier hypothesis of preferred binding mode (cf. Figure with the key hinge
binding interaction between the quinazolinone N-1
and His286). An interesting observation was that some compounds (e.g., 14 and 15) maintained potency even though the
additional 2-methylation would likely cause a clash with the hinge
residues of ALK2 suggesting that they bind through an alternate binding
mode. We therefore sought to solve additional structures and were
able to obtain the structure of compound 11. Interestingly, 11 displayed a flipped binding mode (Figure A) in which the quinazolinone core bound
to His286 through the amide, as a donor and acceptor. This also altered
the vector from the quinazolinone 6-position such that the quinoline
also flipped in order to fill the same volume in the ALK2 central
pocket as for 16 (Figure B).
Figure 4
(A) Cocrystal structure of 11 with ALK2 (PDB 6GI6). (B) Superposition
of costructures of 11 (blue) and 16 (brown).
(C) Superposition of the costructures of 11 (blue) and 21 (orange, PDB 6GIP).
(A) Cocrystal structure of 11 with ALK2 (PDB 6GI6). (B) Superposition
of costructures of 11 (blue) and 16 (brown).
(C) Superposition of the costructures of 11 (blue) and 21 (orange, PDB 6GIP).Inspection of the compound
bound in the flipped binding mode (11) suggested that
introduction of small hydrophobic groups
in the 2-position should not only be tolerated but would lead to additional
hydrophobic interaction (Table ). Interestingly, while introduction of the 2- and 5-methyl
groups individually (Table ) was tolerated but did not lead to an increase in activity,
introduction of both (21) afforded a 6-fold increase
in potency (Table ). We hypothesized that this cooperative increase in potency can
be explained by the change in binding mode: 2-methylation forces adoption
of the flipped, and presumably lower preference binding mode [for
unsubstituted quinazolinones] (Figure C); however, there is no change in potency due to the
addition of a methyl group in a relatively nonpolar part of the binding
pocket. 5-Methylation fills the excluded volume created by flipping
binding mode, which in the absence of 2-methylation is apparently
enough to compensate for adopting the less energetically favored flipped
binding mode. Once the flipped mode is induced by one methyl substituent,
addition of the second does not incur a further penalty and reaps
the benefit of the newly formed interaction.
Table 4
SAR of
Flipped Binding Seriesa
IC50 data is an average
of 2–4 measurements by LanthaScreen Eu Kinase Binding Assay.
IC50 data is an average
of 2–4 measurements by LanthaScreen Eu Kinase Binding Assay.To assess the available space
at the quinazolinone 2-position of
the inhibitors binding in the flipped binding mode, we designed and
prepared a few additional compounds. Crystallography suggested that
the 2-methyl group could be replaced by larger groups. This was indeed
the case. Increasing the size of the 2-substitutent to ethyl (22) or benzyl (23) was tolerated but did lead
to a modest reduction in potency—mirroring the phenyl-only
compound for the precedented binding mode (13), suggesting
that there is scope to further investigate substituents at this position.Furthermore, as the flipped binders do not appear to make use of N-1 for interaction with ALK2, we hypothesized that replacement
with a C atom would result in additional hydrophobic interaction and
loss of the desolvation penalty upon binding due to the polar heteroatom.
The resulting isoquinolinone 24 indeed displayed 10 nM
potency with a superb LE of 0.49.At this stage, two divergent
series with quinazolinone cores were
developed to the stage of a having potency < 50 nM. We hypothesized
that inhibitors of the normal and flipped modes would display distinct
selectivity profiles. To test this hypothesis, the most potent normal
and flipped binding quinazolinones (16 and 21 respectively) were assessed in DiscoveRx’s scanEDGE panel
of 97 diverse kinases, plus all reported off-targets of the pyrazolo[1,5-a]pyrimidine and pyridine series (Table
S1).[8,9,12]Both
quinazolinone based inhibitors displayed excellent selectivity
in the kinome panel tested, though some notable differences were evident
(Figure ). At 1 μM
the normal binder, 16, showed similarly potent inhibition
of ALK6, platelet-derived growth factor receptor (PDGFR) A and B and
Proto-Oncogene receptor tyrosine kinase KIT. While the flipped binder, 21, showed selectivity confined to the tyrosine kinase-like
(TKL) family of kinases—a surprising result given its low molecular
weight. Aside from the highly homologous ALKs 1, 4, 5, and 6, RAF1
and BRAF were also inhibited by 21. These results indicate
that the two different compounds have distinct inhibition profiles
within the kinome, versus each other as well as reported inhibitors
and that both have a highly encouraging level of selectivity against
the panel of kinases screened.
Figure 5
Kinase selectivity profile of 16 and 21 at 1 μM. Blue circle – ALK2; Green
circles ≤
65% probe displacement; small red circle = 65–90% probe displacement;
intermediate circle = 90–95% probe displacement; larger red
circle = 95–99% probe displacement.
Kinase selectivity profile of 16 and 21 at 1 μM. Blue circle – ALK2; Green
circles ≤
65% probe displacement; small red circle = 65–90% probe displacement;
intermediate circle = 90–95% probe displacement; larger red
circle = 95–99% probe displacement.KD values were obtained for ALKs
1–6
and all off-targets identified
in Figure , for both 16 and 21 (Table ) using DiscoverX’s KdELECT assay. 16 had similar affinity for ALKs 1, 2,
3, and 6 (which all signal through SMADs 1, 5, and 8) with moderate
selectivity over ALKs 4 and 5 (which signal through SMADs 2 and 3).
This profile is similar to reported ALK2 inhibitors.[8,9,12] The smaller, flipped binder, 21, displayed an atypical profile—with lower selectivity
over ALKs 4 and 5 and considerable variance in affinity for ALKs 1,
2, 3, and 6. The fact that 21 binds ALKs 1 and 6 16-fold
more potently than ALK2 shows that 21 is not yet favorable
for FOP and DIPG which harbor various activating mutations in ALK2.
However, this is the first case, to our knowledge, of a small molecule
with ALK1 selectivity over ALK2, which shares 79% sequence identity
in their kinase domains.[11] This is a potentially
useful property if developed further to investigate the role of ALK1
in various disease contexts, particularly in regulating angiogenesis.[16]
Table 5
KD’s
for ALKs 1–6 and All off-Targets Identified for Compounds 16 and 21a
KD (nM)
kinase
16
21
ALK1
420
41
ALK2
330
640
ALK3
610
2700
ALK4
11000
690
ALK5
19000
1000
ALK6
96
39
BRAF
65
RAF1
330
KIT
54
PDGRFA
43
PDGRFB
250
KD’s
calculated from duplicate 11-point dose–response curves; KdELECT,
DiscoverX
KD’s
calculated from duplicate 11-point dose–response curves; KdELECT,
DiscoverXFinally, to achieve
proof of concept that this class of compound
shows activity in cells, we tested if 24 modulates the
BMP pathway downstream of ALK2 (Figure ) in HSJD-DIPG-007 patient derived cells (H3F3A K27M and ACVR1 R206H) at three different concentrations
(0.1 μM, 1 μM, and 10 μM). Encouragingly, 24 displayed a dose-dependent reduction of pSMAD1/5/8 and
ID1.
Figure 6
Dose-dependent reduction in markers of ALK2 inhibition in HSJD-DIPG7
cells.
Dose-dependent reduction in markers of ALK2 inhibition in HSJD-DIPG7
cells.Compounds were prepared principally
by a multicomponent reaction
between orthoesters, amines, and anthranillic acids, followed by a
Suzuki coupling (Scheme ). For noncommercially available pyrazole boronic acids/esters, the
parent pyrazole was brominated, then tosyl-protected (ix–x), which allowed for a one-pot borylation and Suzuki protocol.
Scheme 1
General Route to Quinazolinone Derivatives
(a) 110 °C, 16–100%;
(b) PdCl2(PPh3)2, NaOH(aq), 1,4-dioxane,
120–150 °C, 12–100%.
General Route to Quinazolinone Derivatives
(a) 110 °C, 16–100%;
(b) PdCl2(PPh3)2, NaOH(aq), 1,4-dioxane,
120–150 °C, 12–100%.2-Benzyl
quinazolinones were prepared by treatment of benzylcyanide
with hydroxylamine, forming an intermediate amidoxime (Scheme ).[17] In the same pot, reaction with anthranillic acid, 25, afforded 6-bromo intermediate 26, which was functionalized
by Suzuki coupling.
Scheme 2
Synthesis of 2-Benzyl Quinazolinones
(a) (i) 50% NH2OH(aq), 120 °C, (ii)
25, 150 °C, 24%; (b) PdCl2(PPh3)2, NaOH(aq), 1,4-dioxane,
100 °C, 51–100%.
Synthesis of 2-Benzyl Quinazolinones
(a) (i) 50% NH2OH(aq), 120 °C, (ii)
25, 150 °C, 24%; (b) PdCl2(PPh3)2, NaOH(aq), 1,4-dioxane,
100 °C, 51–100%.Isoquinolinones
were prepared by a Doebner-modified Knoevenagel
condensation with 4-bromo-3-methylbenzaldehyde, leading to 27 (Scheme ).[18] Preparation of a mixed anhydride with ethyl
chloroformate followed by treatment with sodium azide gave 28, and subsequent Curtius rearrangement allowed for intramolecular
trapping of the isocyanate to yield a 3:2 mixture of isoquinolinone
regioisomers (29 and 30). After chromatographic
separation, the major product (29) was functionalized
by a Suzuki coupling.
Scheme 3
Synthesis of Isoquinolinones
(a) Piperidine, 110 °C,
64%; (b) (i) NEt3, acetone, ethyl chloroformate, 0 °C
– RT; (ii) NaN3, water, RT, 97%; (c) I2, 1,2-dichlorobenzene, 140–180 °C, 50% [3:2] 29:30; (d)
PdCl2(PPh3)2, Na2CO3(aq), 1,4-dioxane, 150 °C, 61%.
Synthesis of Isoquinolinones
(a) Piperidine, 110 °C,
64%; (b) (i) NEt3, acetone, ethyl chloroformate, 0 °C
– RT; (ii) NaN3, water, RT, 97%; (c) I2, 1,2-dichlorobenzene, 140–180 °C, 50% [3:2] 29:30; (d)
PdCl2(PPh3)2, Na2CO3(aq), 1,4-dioxane, 150 °C, 61%.
Discussion
and Conclusions
In conclusion, we identified a ligand efficient
quinazolinone fragment
(1) for inhibition of the ALK2 kinase domain through
systematic cross screening. We found that the 6-pyrazole of 1 could be modified or replaced with bicyclic groups for gains
in potency and that the activity was surprisingly tolerant to addition
of methyl groups at the quinazolinone 2-, 3-, and 5-positions. Guided
by crystallography, we were able to rationalize this tolerance through
a flipped binding mode. We explored both binding modes to discover
potent inhibitors. Our work shows that there is scope for further
investigation of the SAR of both series and that the kinome selectivity
profiles for example compounds (16 and 21) are not only distinct from one another but also do not hit off-targets
for previously reported ALK2 inhibitors at the concentration tested.
Finally we demonstrated that compound 24 modulates ALK2
in cells in a dose-dependent manner.The compounds presented
here thus represent attractive starting
points to discover potent and selective inhibitors of activin receptor-like
kinases.
Experimental Section
Unless otherwise
stated, commercially available reagents and solvents
were used without further purification. Yields were not optimized.
NMR experiments were performed on a BrukerAvance 500 MHz spectrometer
using an internal deuterium lock. Chemical shifts were measured in
parts per million (ppm) relative to the residual signal of the deuterated
solvent. Data are presented in the following format: chemical shift
(multiplicity [s = singlet, d = doublet, t = triplet, q = quartet,
p = pentet, m = multiplet, and br = broad], coupling constants (J in Hz), integration). Where mixed solvent systems were
used residual solvent peaks to which the spectra are referenced are
underlined. Assignment of 13C NMR data was achieved through
analysis of 2D NMR spectra (HSQC, HMBC, COSY, and NOSEY). LCMS analyses
and high resolution mass spectrometry were performed on an Agilent
1200 series HPLC and diode array detector coupled to a 6210 time-of-flight
mass spectrometer with dual multimode APCI/ESI source (methods A and
B) or a Waters Acquity UPLC and diode array detector coupled to a
Waters G2 QToF mass spectrometer fitted with a multimode ESI/APCI
source (method C); see Supporting Information for method details. All compounds described herein exhibited spectral
data consistent with their proposed structures and, with the exception
of 6 (purity >90%), had purities >95% as determined
by
HPLC.
General Procedure 1: Multicomponent Reaction
A mixture
of an anthranillic acid or isatoic anhydride (1.0 equiv), amine (1.0
equiv), and orthoester (1.0 equiv) was stirred for 1–16 h at
110 °C, until LCMS indicated that reaction was complete. The
mixture was cooled to RT, and the product was purified either by crystallization
or trituration with hot EtOH unless otherwise indicated.
6-Bromo-3-(4-morpholinophenyl)quinazolin-4(3H)-one, i
6-Bromo-2H-benzo[d][1,3]oxazine-2,4(1H)-dione
(1.00 g; 4.15 mmol;
1.0 equiv), NH4OAc (385 mg; 5.00 mmol; 1.2 equiv), and
triethoxymethane (673 mg; 4.15 mmol; 1.0 equiv) were stirred at 110
°C for 18 h. The crude reaction product was loaded onto silica
using CH2Cl2 and purified by column chromatography
with a solvent system of 0–8% MeOH in CH2Cl2. 6-Bromo-2-methylquinazolin-4(3H)-one (212
mg; 0.891 mmol; 21%) was isolated as a cream solid. 1H
NMR (500 MHz, CDCl/MeOH) δ 8.28 (d, J = 2.3 Hz, 1H), 7.76
(dd, J = 8.7, 2.4 Hz, 1H), 7.45 (d, J = 8.7 Hz, 1H), 2.40 (s, 3H). 13C NMR (126 MHz, CDCl/MeOH) δ
161.87, 154.36, 147.56, 137.82, 128.72, 128.20, 121.88, 119.82, 21.36.
HRMS (ESI) m/z calc C9H8BrN2O [M + H]+ 238.9815; found
= 238.9822.
6-Bromo-2,5-dimethylquinazolin-4(3H)-one, vii
6-Amino-3-bromo-2-methylbenzoic acid (950 mg;
4.15 mmol; 1.0 equiv),
NH4OAc (385 mg; 5.00 mmol; 1.2 equiv) and triethoxymethane
(673 mg; 4.15 mmol; 1.0 equiv) were stirred at 110 °C for 48
h. The crude reaction product was loaded onto silica using CH2Cl2 and purified by column chromatography with
a solvent system of 0–8% MeOH in CH2Cl2. 6-Bromo-2,5-dimethylquinazolin-4(3H)-one (255
mg; 1.01 mmol; 24%) was isolated as a cream solid. 1H NMR
(500 MHz, CDCl/MeOH) δ 7.77 (dd, J = 8.8, 1.5 Hz,
1H), 7.25 (d, J = 8.3 Hz, 1H), 2.90 (s, 3H), 2.33
(s, 3H). 13C NMR (126 MHz, CDCl3/MeOH) δ
166.76, 157.95, 153.44, 143.71, 142.04, 129.43, 127.89, 124.22, 25.22,
24.73. HRMS (ESI) m/z calc C10H10BrN2O
[M + H]+ 252.9971; found = 252.9980.
6-Bromo-2-ethyl-5-methylquinazolin-4(3H)-one, viii
6-Amino-3-bromo-2-methylbenzoic
acid (950 mg;
4.15 mmol; 1.0 equiv), NH4OAc (385 mg; 5.00 mmol; 1.2 equiv),
and triethyl orthopropionate (731 mg; 4.15 mmol; 1.0 equiv) were stirred
at 110 °C for 48 h. The crude reaction product was loaded onto
silica using CH2Cl2 and purified by column chromatography
with a solvent system of 0–8% MeOH in CH2Cl2. 6-Bromo-2-ethyl-5-methylquinazolin-4(3H)-one (418 mg; 1.57 mmol; 38%) was isolated as an off-white solid. 1H NMR (500 MHz, DMSO) δ 12.17 (s, 1H), 7.91 (d, J = 8.7 Hz, 1H), 7.36 (d, J = 8.7 Hz, 1H),
2.92 (s, 3H), 2.58 (q, J = 7.5 Hz, 2H), 1.23 (t, J = 7.5 Hz, 3H). 13C NMR (126 MHz, DMSO) δ
162.45, 159.18, 150.45, 138.98, 137.79, 127.17, 122.94, 121.18, 27.87,
21.54, 11.57. HRMS (ESI) m/z calc
C11H12BrN2O [M + H]+ 267.0128;
found = 267.0135.
4-Bromo-3-cyclopropyl-1-tosyl-1H-pyrazole, ix
To a solution of 4-bromo-3-cyclopropyl-1H-pyrazole (100 mg; 0.538 mmol; 1.0 equiv) in CH2Cl2 (10 mL; 0.054 M) was added NaOH(aq) (118
μL; 5 M; 0.591 mmol; 1.1 equiv) and 4-toluenesulfonyl chloride
(113 mg; 0.591 mmol; 1.1 equiv). The mixture was heated to reflux
and stirred for 72 h before being diluted with water (400 mL) and
extracted with CH2Cl2 (2 × 200 mL). The
combined organic portions were washed with brine, dried (Na2SO4), and filtered, and solvent was removed under reduced
pressure. The crude product was washed through an SCX-II column with
MeOH to yield 4-bromo-3-cyclopropyl-1-tosyl-1H-pyrazole
(103 mg; 0.303 mmol; 56%) as a white solid after removal of solvent
under reduced pressure. 1H NMR (500 MHz, DMSO) δ
8.68 (s, 1H), 7.86–7.81 (m, 2H), 7.48 (d, J = 8.1 Hz, 2H), 2.40 (s, 3H), 1.84 (tt, J = 8.3,
4.9 Hz, 1H), 0.99–0.93 (m, 2H), 0.78–0.73 (m, 2H). 13C NMR (126 MHz, DMSO) δ 158.61, 146.80, 133.46, 133.17,
130.88, 128.12, 99.30, 21.64, 8.83, 7.76. HRMS (ESI) m/z calc C13H13BrN2O2SNa [M + Na]+ 362.9779; found = 362.9771.
4-Bromo-3-ethyl-1-tosyl-1H-pyrazole, x
p-Toluene sulfonyl chloride (194
mg; 1.02 mmol; 1.1 equiv) was added to a solution of 4-bromo-3-ethyl-1H-pyrazole (162 mg; 0.930 mmol; 1.0 equiv) and [5 M] NaOH
solution (0.2 mL; 1.02 mmol; 1.1 equiv) in CH2Cl2 (18.5 mL), and the mixture was stirred at 38 °C. After 16 h
the reaction mixture was washed with water (15 mL) and the aqueous
layer was extracted with CH2Cl2 (2 × 20
mL). The combined organic fractions were washed with brine, dried
over MgSO4, filtered, and concentrated under vacuum. The
crude mixture was purified by normal phase flash chromatography in
a solvent system of 0–5% EtOAc in cyclohexane to afford 4-bromo-3-ethyl-1-tosyl-1H-pyrazole (246 mg; 0.826 mmol; 81%) as a white solid. 1H NMR (500 MHz, CDCl3) δ 8.03 (s, 1H), 7.85–7.93
(m, 2H), 7.30–7.38 (m, 2H), 2.62 (q, J = 7.6,
2H), 2.44 (s, 3H), 1.22 (t, J = 7.6, 3H). 13C NMR (126 MHz, CDCl3) δ 158.69, 145.92, 133.86,
128.13, 98.52, 21.74, 20.38, 12.23. HRMS (ESI) m/z calc C12H14BrN2O2S [M + H]+ 328.9954; found = 328.9954.
General Procedure
2: Suzuki Coupling
To a Biotage microwave
vial was added the required bromo-aryl (1.0 equiv), boronic acid or
ester (1.0–1.2 equiv), PdCl2(PPh3)2 (0.05 equiv), Na2CO3(aq) (0.5 M; 1.0
equiv), and 1,4-dioxane or DME. The vial was sealed and purged of
air with 3 rounds of vacuum and N2 or Ar. The mixture was
heated to 120–150 °C for 1 h under microwave irradiation.
The reaction mixture was filtered through Celite, washed with water,
and extracted twice with CH2Cl2. Combined organic
portions were dried (Na2SO4), filtered, and
had solvent removed under reduced pressure. The crude product was
purified by column chromatography in a solvent system of MeOH in CH2Cl2.
6-(1H-Pyrazol-4-yl)quinazolin-4(3H)-one, 1
Following general procedure
2, 6-bromoquinazolin-4(3H)-one (44 mg; 0.20 mmol)
was reacted with [1-(tert-butoxycarbonyl)-1H-pyrazol-4-yl]boronic acid pinacol
ester (69 mg; 0.24 mmol) in DME to yield 6-(1H-pyrazol-4-yl)quinazolin-4(3H)-one (22 mg; 54%) as an off-white solid. Reaction temperature:
150 °C. Purification: 1–8% MeOH in CH2Cl2. 1H NMR (500 MHz, DMSO) 13.06 (br-s, 1H), 12.21
(br-s, 1H), 8.38 (br-s, 1H), 8.28 (d, J = 2.2 Hz,
1H), 8.08 (dd, J = 8.4, 2.2 Hz, 1H), 8.05 (s, 1H),
8.04 (br-s, 1H), 7.65 (d, J = 8.4 Hz, 1H). 13C NMR (126 MHz, DMSO) 161.2, 147.4, 145.0, 136.9, 132.1, 128.2, 126.6,
123.5, 121.3, 120.6. HRMS (ESI) m/z calc C11H9N4O [M + H]+ 213.0771; found = 213.0779.
4-Bromo-3-ethyl-1-tosyl-1H-pyrazole (66 mg; 0.194 mmol; 1.0 equiv), KOAc (57 mg;
0.583 mmol; 1.1 equiv), bis(pinacolato)diboron (55 mg; 0.214 mmol;
1.1 equiv), and PdCl2(PPh3)2 (6.8
mg; 9.70 μmol; 0.05 equiv) were charged to a microwave vial.
The vial was sealed and purged of air with avacuum and N2 five times. 1,4-Dioxane (4.4 mL; 0.044 M) was added, and the mixture
was heated at 120 °C for 1 h under microwave irradiation. At
RT, 6-bromoquinazolin-4(3H)-one (44 mg; 0.194 mmol;
1.0 equiv), [0.5 M] Na2CO3(aq) (389 μL;
0.194 mmol; 1.0 equiv), and PdCl2(PPh3)2 (6.8 mg; 9.70 μmol; 0.05 equiv) were added to the reaction
mixture before the vial was sealed and purged of air with vacuum and
N2 five times. The mixture was heated to 120 °C for
1 h under microwave irradiation. The mixture was then diluted with
water (100 mL) and extracted with CH2Cl2 (3
× 75 mL). Combined organic portions were washed with brine (100
mL), dried (Na2SO4), and filtered, and solvent
was removed under reduced pressure. The crude product was purified
by column chromatography in a solvent system of 2–15% MeOH
in CH2Cl2 to yield 6-(3-ethyl-1-tosyl-1H-pyrazol-4-yl)quinazolin-4(3H)-one (13
mg; 0.03 mmol; 15%) as a white solid. A solution of the tosyl-protected
intermediate was formed in EtOH (1.1 mL), to which was added [1 M]
NaOH(aq) (0.8 mL; 0.82 mmol). The mixture was stirred at
50 °C for 14 h and then neutralized using 1 M HCl and purified
by SCX-2 chromatography (2 g, 10 mL MeOH then 30 mL 1 M NH3 in MeOH). The crude mixture was further purified by Biotage column
chromatography (5–12% MeOH in DCM) and concentrated under a
vacuum. The resulting solid was triturated with n-hexane (3 × 1 mL) to afford 6-(3-ethyl-1H-pyrazol-4-yl)quinazolin-4(3H)-one (7 mg, 0.025 mmol, 82%) as an off-white solid. 1H NMR (500 MHz, DMSO) δ 12.77 (s, 1H), 12.25 (s, 1H),
8.09 (d, J = 2.1 Hz, 1H), 8.06 (d, J = 3.4 Hz, 1H), 7.90 (dd, J = 8.4, 2.2 Hz, 1H),
7.68 (d, J = 8.5 Hz, 1H), 2.82 (t, J = 7.6 Hz, 2H), 1.22 (t, J = 7.6 Hz, 4H). 13C NMR (126 MHz, DMSO) δ 161.18, 147.14, 145.16, 133.73, 132.92,
128.12, 123.32, 122.97, 18.25, 13.86. HRMS (ESI) m/z calc C13H13N4O [M + H]+ 241.1084; found = 241.1084.
4-Bromo-3-cyclopropyl-1-tosyl-1H-pyrazole (30 mg; 0.0882 mmol; 1.0 equiv), KOAc (26 mg;
0.265 mmol; 1.1 equiv), bis(pinacolato)diboron (25 mg; 0.0971 mmol;
1.1 equiv), and PdCl2(PPh3)2 (3.1
mg; 4.41 μmol; 0.05 equiv) were charged to a microwave vial.
The vial was sealed and purged of air with a vacuum and N2 five times. 1,4-Dioxane (2 mL; 0.044 M) was added, and the mixture
was heated at 120 °C for 1 h under microwave irradiation. 6-Bromoquinazolin-4(3H)-one (20 mg; 0.0882 mmol; 1.0 equiv), [0.5 M] Na2CO3(aq) (176 μL; 0.0882 mmol; 1.0 equiv), and PdCl2(PPh3)2 (3.1 mg; 4.41 μmol; 0.05
equiv) were added to the reaction mixture (<40 °C) before
the vial was sealed and purged of air with a vacuum and N2 five times. The mixture was heated to 120 °C for 1 h under
microwave irradiation. The mixture was diluted with water (100 mL)
and extracted with CH2Cl2 (3 × 75 mL).
The combined organic portions were washed with brine (100 mL), dried
(Na2SO4) and filtered and solvent was removed
under reduced pressure. The crude intermediate was purified by column
chromatography in a solvent system of 2–15% MeOH in CH2Cl2 to yield 6-(3-cyclopropyl-1-tosyl-1H-pyrazol-4-yl)quinazolin-4(3H)-one (31
mg; 0.0763 mmol; 87%) as a white solid. A mixture of 6-(3-cyclopropyl-1-tosyl-1H-pyrazol-4-yl)quinazolin-4(3H)-one (5
mg; 0.0123 mmol; 1.0 equiv), [1 M] NaOH(aq) (308 μL;
0.308 mmol; 25 equiv), and EtOH (1 mL; 0.012 M) was heated at 50 °C
for 2 h before being cooled to RT and neutralized with [1 M] HCl(aq). The mixture had solvent removed under reduced pressure,
and the residue was purified by column chromatography in a solvent
system of 2–10% MeOH in CH2Cl2. 6-(3-Cyclopropyl-1H-pyrazol-4-yl)quinazolin-4(3H)-one (2.1
mg; 0.00833 mmol; 68%) was isolated as a white solid. 1H NMR (500 MHz, MeOD) δ 8.49 (q, J = 2.1 Hz,
1H), 8.12 (d, J = 8.6 Hz, 1H), 8.10 (s, 1H), 7.87
(s, 1H), 7.75 (d, J = 8.4 Hz, 1H), 2.13–2.06
(m, 1H), 1.12–0.82 (m, 4H). 13C NMR (126 MHz, MeOD)
δ 161.90, 146.50, 144.37, 133.76, 133.02, 128.41, 126.83, 125.55,
123.26, 122.52, 118.86, 6.36, 6.35. HRMS (ESI) m/z calc C14H13N4O [M + H]+ 253.1084; found = 253.1089.
(4-Oxo-3,4-dihydroquinazolin-6-yl)boronic
acid (30 mg; 0.158 mmol; 1.0 equiv), 3-bromopyrazolo[1,5-a]pyridine (31 mg; 0.158 mmol; 1.0 equiv), K2CO3 (66 mg; 4.74 mmol; 3.0 equiv), bis(di-tert-butyl(4-dimethylaminophenyl)phosphine)dichloropalladium(II)
(5.6 mg; 0.00790 mmol; 5 mol %), DME (1 mL; 0.079 M), and water (1
mL; 0.079 M) were charged to a microwave vial. The vial was sealed
and purged of air with five rounds of vacuum and N2. The
mixture was stirred at 150 °C for 1 h under microwave irradiation.
The reaction mixture was cooled to <40 °C and diluted with
CH2Cl2 (50 mL) and passed through a syringe
filter, and solvent was removed under a vacuum. The residue was purified
initially by column chromatography using a solvent system of 0–8%
MeOH in CH2Cl2, followed by purification by
preparative HPLC (see Supporting Information for details). 6-(Pyrazolo[1,5-a]pyridin-3-yl)quinazolin-4(3H)-one (14 mg; 0.0537 mmol; 34%) was isolated as a white
solid. 1H NMR (500 MHz, MeOD) δ 8.53 (dt, J = 7.0, 1.1 Hz, 1H), 8.47 (d, J = 2.2
Hz, 1H), 8.25 (s, 1H), 8.07 (dd, J = 8.4, 2.2 Hz,
1H), 8.01 (s, 1H), 7.99 (dt, J = 9.1, 1.2 Hz, 1H),
7.79 (d, J = 8.5 Hz, 1H), 7.34 (ddd, J = 9.0, 6.7, 1.1 Hz, 1H), 6.94 (td, J = 6.9, 1.3
Hz, 1H). 13C NMR (126 MHz, MeOD) δ 161.94, 146.63,
144.09, 140.21, 137.18, 133.48, 132.45, 128.82, 127.64, 125.37, 123.09,
123.04, 117.31, 113.02, 111.42. HRMS (ESI) m/z calc C15H11N4O [M + H]+ 263.0927; found = 263.0929.
2-Phenylacetonitrile (500
mg; 4.27 mmol; 1.0 equiv)
and hydroxylamine [50% in water] (424 μL; 6.41 mmol; 1.5 equiv)
were stirred at 120 °C for 2 h. The mixture was cooled to <40
°C, and 6-amino-3-bromo-2-methylbenzoic acid (985 mg; 4.27 mmol;
1.0 equiv) was added. The mixture was heated to 150 °C, stirred
for 2 h, and cooled to <40 °C. EtOH (20 mL) was added to the
reaction mixture which induced precipitation. The solvent was decanted
three times from EtOH, and the solid residue was dried under high
vacuum. 2-Benzyl-6-bromo-5-methylquinazolin-4(3H)-one
(341 mg; 1.04 mmol; 24%) was isolated at a white solid. 1H NMR (500 MHz, DMSO) δ 12.15 (s, 1H), 7.91 (d, J = 8.7 Hz, 1H), 7.38–7.34 (m, 3H), 7.34–7.30 (m, 2H),
7.26–7.22 (m, 1H), 3.89 (s, 2H), 2.90 (s, 3H). 13C NMR (126 MHz, DMSO) δ 157.94, 156.89, 150.36, 138.88, 137.89,
136.83, 129.32, 128.96, 127.27, 127.14, 122.99, 121.14, 40.81, 21.54.
HRMS (ESI) m/z calc C16H14BrN2O [M + H]+ 329.0284; found
= 329.0284.
To a suspension of 3-(4-bromo-3-methylphenyl)-2-methylacrylic
acid (4.80 ;, 18.8 mmol; 1.0 equiv) in acetone (100 mL) at 0 °C
was added triethylamine (3.41 mL; 24.5 mmol; 1.3 equiv). A clear solution
was formed. Then, ethyl chloroformate (2.16 mL; 22.6 mmol; 1.2 equiv)
was added, and a thick suspension was formed. The suspension was allowed
to warm to room temperature. After 2 h, sodium azide (1.83 g; 28.2
mmol; 1.5 equiv) in water (15 mL) was added dropwise, and the suspension
was stirred for a further 2 h. The suspension was cooled down and
diluted with water (50 mL). Ethyl acetate (200 mL) was added, the
layers were separated, and the aqueous layer was extracted with more
ethyl acetate (2 × 50 mL). The organic were washed with sat.
aq. NaHCO3 (2 × 50 mL), dried (MgSO4),
filtered and concentrated (without heating) to yield the product as
a white solid (5.1 g, 97% yield). 1H NMR (500 MHz, CDCl3) δ 7.65 (d, J = 1.5 Hz, 1H), 7.58
(d, J = 8.3 Hz, 1H), 7.29–7.28 (m, 1H), 7.12
(dd, J = 8.3, 2.3 Hz, 1H), 2.44 (s, 3H), 2.12 (d, J = 1.4 Hz, 3H). 13C NMR (126 MHz, CDCl3) δ 174.36, 139.94, 138.23, 134.36, 132.54, 132.22, 129.99,
128.59, 125.92, 22.98, 13.87.
7-Bromo-3,8-dimethyl-2H-isoquinolin-1-one,
29 and 7-bromo-3,6-dimethyl-2H-isoquinolin-1-one, 30
A suspension of 3-(4-bromo-3-methylphenyl)-2-methylacryloyl
azide (4.76 g; 17.0 mmol; 1.0 equiv) in 1,2-dichlorobenzene (70 mL)
was heated to 140 °C for 1 h. Then, iodine (431.5 mg; 1.7 mmol;
0.1 equiv) was added, and the mixture was heated at reflux (ca. 180
°C) for 6 h and cooled to room temperature. The solution was
stirred overnight at room temperature, and the solid formed was filtered
and washed with 20 mL of cyclohexane. The solid was dried for 1 h
at 50 °C and 20 Torr to give a mixture of 29 and 30 (2.16 g; 50% yield) in a ratio 3:2 (1H NMR).
Normal phase flash chromatography of the mixture (0–25% EtOAc
in cyclohexane) gave 116 mg of pure 29, 660 mg of mixture
and 260 mg of pure 30. For 29: 1H NMR (500 MHz, DMSO) δ 11.24 (s, 1H), 7.77 (d, J = 8.5 Hz, 1H), 7.29 (d, J = 8.5 Hz, 1H), 6.27 (dd, J = 2.0, 1.1 Hz, 1H), 2.95 (s, 3H), 2.16 (d, J = 1.1 Hz, 3H). 13C NMR (126 MHz, DMSO) δ 163.21,
140.03, 139.80, 139.40, 136.12, 125.89, 124.19, 123.08, 103.48, 21.76,
18.77. HRMS (ESI) m/z calc for C11H1179BrNO [M + H]+ 252.0019,
found = 252.0014. For 30: 1H NMR (DMSO, 500
MHz) δ 11.30 (br-s, 1H), 8.20 (s, 1H), 7.51 (a, 1H), 6.26 (s,
1H), 2.43 (s, 3H), 2.18 (s, 3H). 13C NMR (DMSO, 126 MHz)
δ 161.2, 141.8, 139.5, 137.6, 129.6, 127.4, 123.8, 121.1, 102.0,
22.8, 18.8. HRMS (ESI) m/z calc
for C11H1179BrNO [M + H]+ 252.0019, found = 252.0015.
Authors: Mark L Boys; Feng Bian; James B Kramer; Christopher L Chio; Xiao-Dan Ren; Huifen Chen; Stephen D Barrett; Karen E Sexton; Donna M Iula; Gary F Filzen; Maria N Nguyen; Paul Angell; Victoria L Downs; Zhi Wang; Neil Raheja; Edmund L Ellsworth; Stephen Fakhoury; Larry D Bratton; Paul R Keller; Richard Gowan; Elena M Drummond; Samarendra N Maiti; Mostofa A Hena; Leroy Lu; Patrick McConnell; John D Knafels; Venkataraman Thanabal; Fang Sun; Diane Alessi; Ann McCarthy; Erli Zhang; Barry C Finzel; Sneha Patel; Susan M Ciotti; Rone Eisma; N A Payne; Richard B Gilbertsen; Catherine R Kostlan; David J Pocalyko; Deepak S Lala Journal: Bioorg Med Chem Lett Date: 2012-04-10 Impact factor: 2.823
Authors: Bruno Larrivée; Claudia Prahst; Emma Gordon; Raquel del Toro; Thomas Mathivet; Antonio Duarte; Michael Simons; Anne Eichmann Journal: Dev Cell Date: 2012-03-13 Impact factor: 12.270
Authors: Paul B Yu; Charles C Hong; Chetana Sachidanandan; Jodie L Babitt; Donna Y Deng; Stefan A Hoyng; Herbert Y Lin; Kenneth D Bloch; Randall T Peterson Journal: Nat Chem Biol Date: 2007-11-18 Impact factor: 15.040
Authors: Darren W Engers; Audrey Y Frist; Craig W Lindsley; Charles C Hong; Corey R Hopkins Journal: Bioorg Med Chem Lett Date: 2013-04-11 Impact factor: 2.823
Authors: Apirat Chaikuad; Ivan Alfano; Georgina Kerr; Caroline E Sanvitale; Jan H Boergermann; James T Triffitt; Frank von Delft; Stefan Knapp; Petra Knaus; Alex N Bullock Journal: J Biol Chem Date: 2012-09-12 Impact factor: 5.157
Authors: Sayantani Sinha; Christina Mundy; Till Bechtold; Federica Sgariglia; Mazen M Ibrahim; Paul C Billings; Kristen Carroll; Eiki Koyama; Kevin B Jones; Maurizio Pacifici Journal: PLoS Genet Date: 2017-04-26 Impact factor: 5.917
Authors: Caroline E Sanvitale; Georgina Kerr; Apirat Chaikuad; Marie-Christine Ramel; Agustin H Mohedas; Sabine Reichert; You Wang; James T Triffitt; Gregory D Cuny; Paul B Yu; Caroline S Hill; Alex N Bullock Journal: PLoS One Date: 2013-04-30 Impact factor: 3.240
Authors: Agustin H Mohedas; You Wang; Caroline E Sanvitale; Peter Canning; Sungwoon Choi; Xuechao Xing; Alex N Bullock; Gregory D Cuny; Paul B Yu Journal: J Med Chem Date: 2014-09-04 Impact factor: 7.446
Authors: Gregory R Gipson; Erich J Goebel; Kaitlin N Hart; Emily C Kappes; Chandramohan Kattamuri; Jason C McCoy; Thomas B Thompson Journal: Bone Date: 2020-07-27 Impact factor: 4.398