Selective Janus Kinase 2 (JAK2) Pseudokinase Ligands with a Diaminotriazole Core

Article abstract of DOI:10.1021/acs.jmedchem.0c00192

Janus kinases (JAKs) are non-receptor tyrosine kinases that are essential components of the JAK-STAT signaling pathway. Associated aberrant signaling is responsible for many forms of cancer and disorders of the immune system. The present focus is on the d

Full text of DOI:10.1021/acs.jmedchem.0c00192

Subscriber access provided by BIU Pharmacie | Faculté de Pharmacie, Université Paris V
Article
Selective Janus Kinase 2 (JAK2) Pseudokinase
Ligands with a Diaminotriazole Core
Maria-Elena Liosi, Stefan G Krimmer, Ana S. Newton, Thomas Dawson, David E. Puleo,
Kara J Cutrona, Yoshihisa Suzuki, Joseph Schlessinger, and William L Jorgensen
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.0c00192 • Publication Date (Web): 24 Apr 2020
Downloaded from pubs.acs.org on April 24, 2020
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Selective Janus Kinase 2 (JAK2) Pseudokinase Ligands with a
Diaminotriazole Core
†,§
†,§
†
†
Maria-Elena Liosi, Stefan G. Krimmer, Ana S. Newton, Thomas K. Dawson, David E.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
‡
†
‡
‡
†,
Puleo, Kara J. Cutrona, Yoshihisa Suzuki, Joseph Schlessinger, and William L. Jorgensen *
†
‡
Department of Chemistry, Yale University, New Haven, Connecticut 06520-8107, Department
of Pharmacology, Yale University School of Medicine, New Haven, CT 06520-8066.
§
: Both authors contributed equally to this work.
S Supporting Information
ABSTRACT: Janus kinases (JAKs) are non-receptor tyrosine kinases that are essential
components of the JAK-STAT signaling pathway. Associated aberrant signaling is responsible for
many forms of cancer and disorders of the immune system. The present focus is on the discovery
of molecules that may regulate the activity of JAK2 by selective binding to the JAK2 pseudokinase
domain, JH2. Specifically, the Val617Phe mutation in JH2 stimulates the activity of the adjacent
kinase domain (JH1) resulting in myeloproliferative disorders. Starting from a non-selective
screening hit, we have achieved the goal of discovering molecules which preferentially bind to the
ATP binding site in JH2 instead of JH1. We report the design and synthesis of the compounds and
binding results for the JH1, JH2, and JH2 V617F domains, as well as five crystal structures for
JH2 complexes. Testing with a selective and non-selective JH2 binder on the autophosphorylation
of wild-type and V617F JAK2 is also contrasted.
KEYWORDS: Janus Kinase 2, JAK2, JAK2 JH2, Pseudokinase, V617F, Structure-Based Drug
Design, Fluorescence Polarization, Protein Crystallography.
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 51
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
INTRODUCTION
JAK1, JAK2, JAK3, and TYK2 are members of the Janus family of non-receptor tyrosine kinases,
which are activated by and mediate the cellular responses induced by binding of a variety of
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
cytokines to specific cytokine receptors. Cytokine-induced activation of the JAK-STAT signaling
pathway and other intracellular pathways play important roles in the control of cell proliferation,
hematopoiesis, and immune functions. In addition to a canonical tyrosine kinase domain (JH1)
located in the C-terminal region, JAK proteins contain a pseudokinase domain designated JH2.
Though JH2 domains have an ATP-binding site, they show little or no catalytic activity. However,
JH2 domains do have a regulatory function for the JH1 kinase activity such that mutations in JH2
2
can cause hyperactivation leading to numerous diseases and cancer. In particular, the single point-
mutation Val617Phe (V617F) in JAK2 JH2 is responsible for the majority of myeloproliferative
disorders including polycythemia vera, myelofibrosis, and essential thrombocythemia.3,4
Undesirable side effects such as anemia and thrombocytopenia are associated with inhibition of
2
the wild-type kinase activity of the JH1 domain of JAK2. Thus, selective inhibition of V617F
JAK2 could be therapeutically valuable. Interestingly, mutations of residues that eliminate ATP-
binding by JAK2 JH2 were found to enhance the autophosphorylation of wild-type JAK2 by 2-5
5
fold, while they generally reduced the activity of V617F JAK2 by 5-15 fold. An exception was
L579F JAK2 mutation, which does not remove ATP binding, and showed reduced effect on V617F
JAK2 activation. This work has raised the possibility of selective reversal of the activating effect
5
,6
of the V617F mutation by blocking ATP binding to JAK2 JH2. To test this hypothesis, our first
goal was to discover small molecules that can bind strongly to the ATP binding site of JAK2 JH2,
but that bind at most weakly to the ATP binding site of JAK2 JH1. The next step is testing the
effects on the autophosphorylation activity of full-length wild-type and V617F JAK2.
2
ACS Paragon Plus Environment

Page 3 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Towards these ends, in seeking molecules that bind to the JH2 ATP-site, we reported
results of a high-throughput screen that yielded the known pan-CDK and pan-JAK inhibitor
7
,8
9
JNJ7706621 (1, Figure 1) as the strongest JH2 binder in a fluorescence polarization (FP) assay.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Subsequent analysis by isothermal titration calorimetry (ITC) yielded binding constants K of 106
d
Figure 1. Compound 1 is the known pan-CDK and pan-JAK kinase inhibitor JNJ7706621.
Compounds 2 – 15 are JAK2 JH2-binding molecules reported in this work.
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 51
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 2. (Left) 1 bound to JAK2 JH2 (PDB ID 5USZ); (Right) 1 bound to JAK2 JH1 (PDB
ID 5USY). Hydrogen bonds with rON < 3.6 Å are indicated with dashed lines. Selective
binding to JAK2 JH2 is sought by growing the ligands towards the vacant region near T555
and R715.
nM and 31 nM with JAK2 JH2 and JH1, respectively, showing an undesirable greater than three-
9
fold preference for binding JH1. It may be noted that these values are lower than an earlier K
d
8
report of 220 nM for 1 with a JH1-JH2 construct of JAK2 in a competition binding assay. Crystal
9
structures for 1 with JAK2 JH2 and JH1 were also determined, as illustrated in Figure 2. The
binding sites are similar in the hinge regions, Glu627-Phe628-Val629 for JH2, and Glu930-
Tyr931-Leu932 for JH1; the diaminotriazole fragment of 1 engages in three hydrogen bonds with
the backbone in both cases. In addition, for JAK2 JH2 there are hydrogen bonds between the amino
group of 1 and the sidechain oxygen of the gatekeeper, Gln626, and between the carbonyl group
of 1 and Lys581. The strong binding of 1 to both JAK2 JH1 and JH2 is consistent with the similar
binding modes in the hinge regions and the limited contact of the difluorophenyl substituent. The
molecule is U-shaped with the aryl group directed out of the binding site towards the solvent.
In order to seek molecules that may selectively bind to the ATP site in JAK2 JH2 and avoid
inhibition of the kinase activity of wildtype (WT) JAK2, the immediate goal was to discover small
4
ACS Paragon Plus Environment

Page 5 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
molecules whose affinities are at least as strong as 1 for JAK2 JH2, while showing much
diminished affinity for JAK2 JH1. For the present series, 1 was chosen as the starting point with
retention of the 3,5-diamino-1,2,4-triazole substructure. In viewing Figure 2, it is apparent that the
desired selectivity needs to exploit differences in the vacant eastern part of the binding clefts
heading towards Thr555 and Arg715 in JAK2 JH2. However, this is the region where the
triphosphate portion of ATP is bound with the terminal phosphate group in a hydrogen-bond
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
0
cluster with Thr555, Asn678, and Arg715, as shown in Figure 3. Thus, it is a very polar region
with the concomitant targeting difficulties owing to the competition between an envisioned polar
substituent being well-solvated in water and in the polar binding site. Nevertheless, the challenge
was taken on with the notion that precise fit of a relatively rigid, polar substituent ending in an
anionic mimic of the terminal phosphate might be successful. Competition with ATP is not
expected to be problematic since its binding is relatively weak with binding constants of 1.3 μM
Figure 3. Rendering from the crystal structure for ATP bound to JAK2 JH2 (PDB ID 4FVQ).
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 51
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
5
for both WT JAK2 JH2 and the V617F mutant. It should also be noted that the principal structural
1
0
changes for JAK2 V617F occur in this vicinity. Val617 is located above Phe595 in Figure 3 and
upon mutation, Phe617-Phe595-Phe594 form an edge-face-edge stack with downward
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
10
displacement of Phe594 towards the terminal phosphate in the ATP-bound structures. Phe595
has been shown to be indispensable for the constitutive activity of JAK2 V617F; mutation to non-
11
aromatic residues significantly decreases the activity. Thus, exploration of alternative ligands
that occupy the ATP site could reasonably be expected to affect the activity of JAK2 V617F, e.g.,
by reversing the displacement of F594.
RESULTS AND DISCUSSION
Design and Assays. For thoroughness, molecular design emphasized inspection of
hundreds of trial structures for JAK2 JH2 complexes built with the BOMB (Biochemical and
1
2
13
Organic Model Builder) program and energy-minimized with MCPRO using the OPLS-AA/M
force field for proteins14 and OPLS/CM1A for ligands.15 Free energy perturbation (FEP)
calculations were also carried out in some cases to better predict potential differences in free
1
2
energies of binding for analogues; these calculations included ca. 2000 explicit water molecules
and extensive sampling of configurations for the JAK2 JH2 protein, ligands and water. The
principal assays measured binding constants with JAK2 JH2, V617F JH2, and JH1 via
1
6
17
fluorescence polarization (FP) using a fluorescein-conjugate of 1 as the tracer. The assay
followed the previous description17 with minor changes to the buffer composition, as fully
described in the Experimental section. For several compounds, microscale thermophoresis was
carried out as an additional approach for obtaining K values. The present measurements like the
d
6
ACS Paragon Plus Environment

Page 7 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
prior ITC and FP assays were carried out at pH 8.0, which provided more consistent results than
current tests at pH 7.0 and 8.5.
For the FP assays with JAK2 JH2, 1 was used as a control compound so multiple K results
d
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
were obtained yielding an average value of 0.456 ± 0.124 µM (Table 1). The result from the MST
measurements, 0.489 ± 0.084 µM, was notably consistent given the differences in the utilized
buffers, as detailed in the SI. These values are somewhat lower than the previous FP result of 0.80
1
7
9
±
0.05 µM, and they are significantly higher than the ITC result of 0.106 µM, and of a recent
18
report of 0.094 µM from intrinsic tryptophan fluorescence. We have also assayed a second
compound, BI-D1870, that was tested in reference 18; our K of 2.0 ± 0.2 µM is again several-
d
1
8
fold higher than the 0.62 ± 0.04 µM in the prior report. Clearly alternative binding assays are
yielding somewhat different results, which likely reflects differences in the assay conditions and
preparation of the protein. Thus, it is advisable to make comparisons relative to a common control,
e.g., compound 1.
Table 1. Binding Affinity for JAK2 JH2 (K , µM) from the FPAssay
d
Compd
K
(µM)a
d
1
2
4
5
6
7
8
9
0.456 ± 0.124
16.7 ± 5.4
4.7 ± 0.8
47.8 ± 9.2
30.5 ± 3.6
12.3 ± 0.6
1.9 ± 0.1
0.643 ± 0.019
0.571 ± 0.034
4.0 ± 0.3
0.346 ± 0.034
0.439 ± 0.064
7.0 ± 0.1
10
11
12
13
14
15
0.374 ± 0.013
a
K
d
data from quadruplicate measurements in three independent assays. Mean ± SEM.
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 51
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Synthesis of New Compounds. The new compounds reported here are numbered 2 – 15
in Figure 1; complete synthetic details and characterization data are reported in the Experimental
section. However, the general approach can be illustrated with the synthetic route for 12 in Scheme
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
7
1
. The key coupling step is the regioselective acylation of a 1H-[1,2,4]triazole-3,5-diamine with
a phenylcarbamate (step e in Scheme 1). Of course, obstacles arose. Indeed, for the coupling, the
original intention was to employ a common procedure for urea formation, the reaction of two
amines with triphosgene in the presence of triethylamine. This strategy afforded compound 3 in
~
10% yield, but failed for more complex ureas, even when attempted in several different solvents
(
DCM, THF, dioxane, DMSO) and at various temperatures (r.t., 80 °C, 130 °C). The problems
appeared to be associated with the poor solubility in organic solvents for the polar diaminotriazole
precursor and the presence of its tautomeric 2-H form, which leads to the undesirable and difficult-
to-separate, regioisomeric 2-H byproduct. In search of a more efficient route, phenyl carbamates
1
9
were investigated as alternative carbonyl sources. This approach proved to be effective in
Scheme 1. Synthesis of Compound 12.a
aReagents and conditions: (a) Boc
dioxane, 110 °C, 19 h; (c) TFA, DCM, 0 °C to r.t., 30 min; (d) NaHCO
C, 50 min; (f) DBN, LiBr, MeCN, 2 vol % H O, r.t., 68h.
O, Et
N, iPrOH, r.t., 14 h; (b) 10 mol % Pd(OAc)
, 20 mol % JohnPhos, Cs
N, dioxane, 110
2
3
2
2 3
CO ,
3
, H O/THF, 0 °C, 1h; (e) Et
2
3
°
2
8
ACS Paragon Plus Environment

Page 9 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
generating the desired products when dioxane was used as the solvent, but the yields were not
optimal since the problems with the diaminotriazole precursor remained. Better results were
obtained by increasing the reaction temperature from 80 to 110 °C, extending the reaction time,
and diluting the mixture from 1.0 to 0.5 M to improve dissolution of the diaminotriazole. Finally,
the use of DCM:MeCN as the solvent system for the chromatography allowed for the successful
purification of ester precursor of 12.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
In several cases preparation of the carbamates also proved challenging. For 12, the original
plan to prepare the 2-aryloxazole (step c in Scheme 1) was to utilize regioselective, palladium-
catalyzed direct arylation of the commercially available ethyl oxazole-4-carboxylate with
unprotected 4-iodo aniline, under the conditions described by Verrier et al.20 However, the
suggested conditions (5 mol % Pd(OAc)
desired product in single-digit yields. Consequently, it was decided to protect the aniline and repeat
the arylation with increased catalyst and ligand loading (10 mol % Pd(OAc) , 20 mol % JohnPhos).
2
, 10 mol% JohnPhos) proved inefficient, affording the
2
This approach allowed access to the 2-anilinyloxazole in a viable yield (30%). And, finally,
hydrolysis of the ester precursor of 12 in the last step required care. Standard conditions using
strong, nucleophilic bases led to decomposition of the urea. Fortunately, heteroatoms were present
in the α or β-position of most ester precursors, which allowed use of the mild ester hydrolysis
21
method introduced by Mattsson et al. In this transformation, lithium coordination to the carbonyl
group and the neighboring heteroatom increases the electrophilicity and thus selectivity of the ester
toward nucleophilic attack, allowing hydrolysis by water at room temperature. However, long
reaction times were required due to the low solubility of the compounds in the reaction media.
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 51
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Lead Optimization. An initial question that was addressed was the importance of the
carbonyl group in 1 for binding to JAK2 JH2. Though the carbonyl oxygen atom participates in a
hydrogen bond with the ammonium group of Lys581 (rNO = 3.16 Å), it is in repulsive contact, 3.06
Å, with the sidechain carbonyl oxygen atom of Gln626 (Figure 2). Thus, 2, the desoxy analog of
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
, was prepared; it was found to yield a much-reduced affinity for JAK2 JH2 with a K of 16.7 μM
d
in the FP assay (Table 1). Apparently, the hydrogen bond with the charged ammonium group more
than offsets the repulsion with Gln626 and/or the reduced torsional flexibility of 1 compared to 2
is beneficial.
Having established the desirability of the carbonyl group, the next step was to append a
substituent that should project eastwardly past Lys581 towards Thr555 rather than outward from
the binding site as for 1. Though several alternatives were considered, a secondary amide linkage
modeled well, especially with an N-phenyl substituent that might form a cation-π interaction with
Lys581. This notion led to the synthesis of 3 and 4, which only differ by the p-anilinyl substituent
being cyano or sulfamyl. In general, we have observed little difference in binding affinity for these
alternatives, but the lesser solubility of the cyano analogs sometimes led to assay difficulties as for
3
. The FP result for the N-phenylamide 4 with a 3-fold improvement over 2 to 4.7 µM was
encouraging in view of the fact that the phenyl ring is being placed in a polar environment. It was
also possible to obtain crystal structures for the complexes of both 3 and 4 with JAK2 JH2 at 1.94-
Å and 1.90-Å resolution, respectively. The structures clearly reveal the desired eastwardly
projection and the cation-π interaction with Lys581, as illustrated in Figure 4 for 4. The hydrogen-
bond length between the amide carbonyl oxygen of 4 and the ammonium nitrogen atom of Lys581
is 2.85 Å, and the distances between the ammonium nitrogen and the carbon atoms of the phenyl
1
0
ACS Paragon Plus Environment

Page 11 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 4. Rendering from the 1.90-Å crystal structure for 4 bound to JAK2 JH2 (PDB ID 6OBB).
ring are as short as 3.92 Å for the ipso carbon atom and 3.77 Å for the nearer ortho carbon atom,
while the hydrogen bonding in the hinge region is essentially the same for 1, 3 and 4.
To expand the structure-activity data, 5, the tertiary methylamide corresponding to 4 was
checked. Since there is a water molecule hydrogen-bonded to the amide NH in 4 (rNO = 3.11 Å,
Figure 4), it might seem likely that 4 and 5 would have similar K values. However, a
d
conformational issue should be noted for acetanilides. From small-molecule crystallography and
quantum chemical calculations, it is known that acetanilide prefers the Z-conformer, while N-
methylacetanilide prefers the E-conformer.22,23 We confirmed this with DFT B3LYP/6-31G(d)
geometry optimizations and subsequent single-point energy evaluations at the MP2/6-311+G(d,p)
2
4
level. The B3LYP and MP2 calculations favor the Z-conformer for acetanilide by 3.19 and 1.29
kcal/mol, respectively, while the E-conformer is preferred by 3.87 and 3.39 kcal/mol for N-
methylacetanilide (Scheme 2). Assuming this preference carries over to 4 and 5, there would be a
1
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 51
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
significant conformational penalty for 5 to achieve the Z-conformation that is needed for the
cation-π interaction with Lys581. As listed in Table 2, FEP calculations were also performed for
the conversion of Z-5 to Z-4; when adjusted for the conformational penalty, weaker binding is
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
expected for 5. This was borne out by the measured K of 47.8 μM for 5 in Table 1.
d
Scheme 2. Computed Conformational Preferences (kcal/mol) for Acetanilides
Table 2. Differences in Free Energies of Binding with JAK2 JH2 from FEP Calculations
A → B
→ 4
→ 4
Transformation
(Z)-O=CNMePh → O=CNHPh
2-pyridinyl → phenyl
ΔG
b
(kcal/mol)a
1.47 ± 0.54
-0.43 ± 0.59
-1.44 ± 0.45
-0.64 ± 0.16
5
6
7 → 4
3-pyridinyl → phenyl
4-pyridinyl → phenyl
b
4
-Pyr → 4
a
4
-pyridinyl analogue of 6 or 7.
Pyridinyl replacements 6 and 7 for the phenyl group were also considered. Initial structure
building with BOMB indicated that constructive interaction of the pyridinyl nitrogen atoms with
Lys581 might be possible with N-N contacts of 3.32 Å for 6 (Figure 5) and 3.72 Å for 7. Energy
minimizations with the force-field find a more favorable interaction with the protein by 3.0
1
2
ACS Paragon Plus Environment

Page 13 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 5. Modeled structure illustrating the potential hydrogen bonding between Lys581 and the
amide carbonyl oxygen atom and the pyridinyl nitrogen atom for 2-pyridinyl analogues of 3 and 4
such as 6.
kcal/mol for 6 than 7; however, this is offset by a 3.4 kcal/mol conformational penalty for
placement of the pyridine nitrogen atom syn rather than anti to the carbonyl oxygen in 6, which is
needed for the hydrogen bond with Lys581. B3LYP/6-31G(d) optimizations for the corresponding
syn and anti conformers of N-(pyridine-2-yl)acetamide indicate an even higher conformational
24
penalty to achieve the syn conformer of 6, 8.98 kcal/mol. For both 6 and 7, the azine ring is now
basically in-plane with the amide group, which replaces the cation-π interaction found for 4 with
Lys581 (Figure 4) with a hydrogen bond (Figure 5). Results of FEP calculations were also obtained
for conversion of 6 and 7 to 4 with the predictions that the pyridinyl analogs would be weaker
binders for JAK2 JH2 (Table 2). The FEP calculations for 6 considered multiple starting
conformers; both conformers with the pyridine nitrogen syn or anti to the carbonyl group were
stable and the results indicated that the syn conformer is favored by 0.7 ± 0.2 kcal/mol. The
compounds were synthesized and the K
d
results of 30.5 (6) and 12.3 (7) μM confirmed the weaker
binding than for 4 (Table 1). These relatively simple changes illustrate well the challenges in
1
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 51
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
making even qualitatively correct predictions for differences in protein-ligand binding owing to
the complexities of the inter- and intra-molecular energetics and solvation.
The next step in the design was to consider substituents in the para position of the phenyl
ring of 4 that would extend further towards Thr555 and Arg715 (Figure 4). In view of the
placement of the terminal phosphate of ATP in this region (Figure 3), substituents terminating in
an anionic group seemed desirable. Modeling was performed for various carboxyalkyl and
carboxyalkoxy alternatives with the conclusion that carboxymethoxy should be particularly
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
promising; the OCH fragment should be in-plane with the phenyl ring, as in anisole, and the
2
carboxylate group should be extended to form hydrogen bonds with Thr555. For comparison, both
carboxyethyl and carboxyethoxy were expected to incorporate a gauche dihedral angle and be less
extended. Thus, the carboxymethoxy analogues 9 and 10 were synthesized and assayed along with
8
, the methyl-ester precursor of 10. The results were gratifying with a ca. 8-fold boost in binding
strength with JAK2 JH2 for 9 (0.64 μM) and 10 (0.57 μM) compared to 4. The corresponding
result for the ester 8 (1.9 μM) confirmed the importance of an anionic terminus for the substituent,
though it is not obvious intrinsically that a carboxylate group would yield any improvement in
binding since it would be well hydrated unbound in water. 11, the meta-substituted isomer of 9
was also synthesized and showed reduced binding with a K of 4.0 μM (Table 1). The modeling in
d
this case predicted that the substituent would likely be directed into the solvent from the upper
meta position in Figure 4.
It was also possible to obtain crystal structures for the complexes of 9 and 10 with JAK2
JH2 at resolutions of 2.06 Å and 1.71 Å, respectively. As shown in Figure 6, the carboxylate group
of 10 is extended, as expected, with a COCC dihedral angle of 170º. It is positioned similarly to
the terminal phosphate group of ATP in Figure 3, and it forms hydrogen bonds with the backbone
1
4
ACS Paragon Plus Environment

Page 15 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
NH and side-chain OH of Thr555 with lengths of 3.06 and 2.83 Å. As expanded upon in Figure 7,
the carboxylate group also participates in an extensive network of hydrogen bonds with localized
water molecules in the crystal structure including three hydrogen bonds between the carboxylate
oxygen atoms and water molecules.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
In viewing the crystal structure for 10, it was felt that a 1-Å additional extension of the
carboxylate group towards Arg715 to form a salt bridge would likely be beneficial for both binding
and selectivity. Addition of a methylene group was not expected to be optimal based on the
modeling noted above. Instead, replacement of the methoxy linking group with a five-membered
heterocycle yielded computed structures that delivered the desired contact. The added rigidity of
the heterocycle was also viewed as a desirable feature, though there were concerns about
Figure 6. Rendering from the 1.71-Å crystal structure for 10 bound to JAK2 JH2 (PDB ID 6OBF).
1
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 51
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 7. Illustration of the hydrogen-bonding in the vicinity of the carboxylate group of 10 from
the 1.71-Å crystal structure with JAK2 JH2 (PDB ID 6OBF). Red spheres represent oxygen atoms
of localized water molecules.
unfavorable disruption of the hydrogen-bonding networks (Figure 7). More than 30 alternative
structures were built with BOMB and visually inspected; our preference was for heterocycles with
multiple hydrogen-bonding sites. The syntheses brought new challenges; however, several
illustrative compounds were prepared starting with the oxazole-4-carboxylic acid 12 (Scheme 1).
The FP assay result of 0.346 ± 0.034 μM showed the validity of the design and yielded the first
compound that is a clearly stronger binder for JAK2 JH2 than JNJ7706621 (1). Again, we were
fortunate to obtain a crystal structure for this complex at a resolution of 2.03 Å (Figure 8). The
shortened contact with Arg715 was achieved with an O…N distance of 3.56 Å, while maintaining
the hydrogen bonds with the NH (2.98 Å) and side-chain OH (2.70 and 3.58 Å) of Thr555. There
are again three water molecules hydrogen-bonded to the carboxylate group with O…O separations
of 2.28, 2.82, and 3.09 Å. The water molecule with the 2.28-Å contact is part of a striking triangle
of hydrogen-bonded water molecules that bridges between the carboxylate groups of the ligand
and Asp699. The central water molecule in the cluster is also hydrogen-bonded to the nitrogen
1
6
ACS Paragon Plus Environment

Page 17 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
atom of the oxazole at a distance of 2.79 Å. Furthermore, the cluster is in van der Waals contact
with a meta carbon atom of Phe594 with separations of 3.37, 3.91, and 4.27 Å. Thus, this
interesting water trimer is providing a carefully arranged molecular cushion between the ligand
and JAK2 JH2 residues Asp699 and Phe594.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Among additional five-membered heterocycles as linking groups, the furan (13) and 1,3,4-
oxadiazole (14) analogues of 12 were synthesized. These were expected to be less robust binders
for JAK2 JH2 than the oxazole 12. First, a furan-type oxygen atom is a weak hydrogen-bond
acceptor and its replacement of the oxazole nitrogen atom in Figure 8 would weaken the link to
the water cluster. Furthermore, furan and thiophene are the most lipophilic five-membered
heteroaromatic molecules with octanol/water log P values of 1.34 and 1.81 versus, for example,
Figure 8. Rendering from the 2.03-Å crystal structure for 12 bound to JAK2 JH2 (PDB ID 6OCC).
Red spheres represent oxygen atoms of localized water molecules.
1
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 51
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
25
0
.12 for oxazole, 0.08 for isoxazole, and 0.44 for thiazole. Finally, the oxazole oxygen atom in
the crystal structure for 12 (Figure 8) does not participate in a hydrogen bond and, in fact, there is
no room for a neighboring water molecule. Thus, if the oxadiazole ring in 14 is oriented as for 12
in Figure 8, there would be a significant penalty for desolvation of one of the nitrogen atoms. The
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
qualitative expectations were borne out by the measured K
4 (Table 1).
From viewing the crystal structure for 10 (Figures 6 and 7) and modeling with BOMB, the
E)-cinnamic acid analog 15 was also suggested with the vinyl group replacing the OCH linker.
This compound was synthesized and turned out to be a somewhat better binder than 10 with a K
d
values of 0.44 and 7.0 μM for 13 and
1
(
2
d
of 0.37 μM. The improvement can be attributed to the greater rigidity of 15 stemming from
removal of one torsional degree of freedom. It is clear from the modeling of 15 and the crystal
structures of ATP, 10, and 12 bound to the JH2 domain in Figures 3, 6, and 8 that the terminal
carboxylate group in the designed compounds overlaps closely with the terminal phosphate of
ATP; in all cases hydrogen bonds are formed between these groups and both the hydroxyl group
and backbone NH of Thr555.
JH2/JH1 Selectivity. The next step was to study the selectivity of the reference compound
1
and the potent new compounds 10, 12 and 15 towards wild-type JAK2 JH1 and JH2, and the
V617F JH2 variant. K results were obtained in all cases via fluorescence polarization using the
d
1
7
same tracer. As listed in Table 3, the pan-JAK binder JNJ7706621 (1) shows no significant
selectivity in the FP assays towards the three proteins, which is in accord with prior reports and
expectations from the crystallography (Figure 2).8,9,18 However, as designed, compound 10
exhibits strong (ca. 75-fold) selectivity for binding the wild-type or V617F JH2 pseudokinase
domain over the JH1 kinase domain. For the oxazole 12 and cinnamic acid 15 there are also ca.
1
8
ACS Paragon Plus Environment

Page 19 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
2
0-fold preferences for binding wild-type JH2 over the kinase domain, and there are again modest
differences in binding for wild-type JH2 and the V617F variant.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
, µM) from the FP and MST Assays for JAK2 Domainsa
Table 3. Binding Affinities (K
d
JH1
JH2
JH2
V617F JH2
FP
V617F JH2
MST
Cmpd
1
FP
FP
MST
0.67 ± 0.18
42.3 ± 2.3
6.6 ± 0.9
6.0 ± 1.0
0.46 ± 0.12
0.57 ± 0.03
0.35 ± 0.03
0.37 ± 0.01
0.49 ± 0.08
0.31 ± 0.03
0.29 ± 0.07
0.20 ± 0.05
0.60 ± 0.09
0.54 ± 0.13
0.48 ± 0.14
0.18 ± 0.08
0.82 ± 0.12
0.40 ± 0.05
0.43 ± 0.03
0.23 ± 0.03
1
1
0
2
15
In addition, eight K
Optimization of this assay was possible for the wild-type JH2 and JH2 V617F domains, but an
adequate signal-to-noise ratio could not be obtained for JH1. MST was found to give similar K
values as FP (Table 3), though some differences could be expected in view of the variations in
optimized buffer compositions. With MST, compounds 10 and 12 are found to have identical K
d
values were determined using microscale thermophoresis.26
d
d
values within the error limits for JAK2 JH2 near 0.30 μM, which improves on the 0.49 μM result
for 1. The binding constants for JAK2 JH2 and the V617F variant are again close. Finally,
compound 15 is found with MST to be the most tenacious binder with K values at the 200-nM
d
level for both the JH2 and V617F JH2 domains.
At this point the primary goals were achieved. Compounds were discovered that show good
selectivity for binding to the JAK2 JH2 domain in preference to the JH1 kinase domain and that
were at least as potent as the non-selective JNJ7706621 (1), which has previously been reported
values of 0.106 and 0.094 µM.9,18
from alternative binding measurements to give K
d
Results with Full-Length Wild-Type and V617F JAK2. The next step was to test
compounds 1 and 10 for their influence on the autophosphorylation of wild-type and V617F JAK2
1
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 51
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
mutant. These compounds were chosen to contrast the non-selective JH1/JH2 binder 1 with the
JH2-selective 10. The expectation was that 1 would be an inhibitor of the JH1-based kinase
activity for both the WT and variant proteins, while the outcome with 10 was unclear. Though the
mutational studies suggested that displacement of ATP from the JH2 domain by selective binders
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5
could deactivate V617F JAK2, the JH2 binders may also be viewed as potential ATP surrogates
leading to no effect. Since activation of kinase activity with some mutations was also observed,
this outcome is also possible with the present compounds.
As detailed in the Experimental Section, full-length cDNAs encoding human WT and
V617F JAK2 (residues 1-1132) were amplified by PCR, cloned into an expression plasmid, and
the JAK2 constructs were expressed in HEK293T cells grown at 37 °C. The transfected cells were
lysed and the JAK2 proteins were immunoprecipitated from the supernatant and used to measure
3
2
kinase activity by treatment with [γ- P]ATP in the presence or absence of increasing
concentrations of 1 and 10. The autophosphorylation reaction was stopped after 15 min at 30 °C
and the extent of phosphorylation was gauged by autoradiography. Details of the protocols were
worked out with compound 1 to establish good signal-to-noise and reproducibility in triplicate
experiments.
Though 1 has previously been established to bind to the JH1 and JH2 domains of
JAK2,8,9,17,18 its status as an inhibitor has not been determined until now. As shown in Figure 9A,
we do find that 1 is an inhibitor of the autophosphorylation of both WT and V617F JAK2. There
are steady, dose-dependent decreases in phosphorylation in the presence of 1. It was possible to
quantify the gel bands from the phosphor autoradiography to yield IC50 values of 2.96 ± 0.41 μM
for 1 with WT JAK2 and 11.05 ± 3.21 μM with V617F JAK2. These results seem consistent with
the K of 0.67 μM for 1 with the JH1 kinase domain of JAK2 (Table 3).
d
2
0
ACS Paragon Plus Environment

Page 21 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
32
Figure 9. In vitro [γ- P]ATP kinase activity assay of immunoprecipitated full-length JAK2.
Autoradiography of autophosphorylated JAK2 wild-type and V617F mutant exposed to different
concentrations of (A) 1, and (B) 10. Shown is representative data from experiments performed in
triplicate.
In contrast, with the JH2-selective 10 higher concentrations are needed to see clear effects
(Figure 9B). Inhibition of WT JAK2 is apparent above ca. 10 μM, and an IC50 of 46.53 ± 4.74 μM
could be assigned. The weak inhibition of the kinase activity is consistent with the K of 42 μM
d
that was found for 10 with JAK2 JH1 (Table 3). Thus, it is expected that 10 is binding to the JH2
domain at low concentrations, but it has negligible effects on the kinase kinetics until significant
binding to the ATP-site in the JH1 domain sets in at concentrations above 10 μM. For V617F
JAK2, the effects of 10 are further diminished with little inhibition apparent below 25 μM, with
an IC50 in the 100 – 200 μM range. Although 10 is expected to be binding to the ATP-site in the
pseudokinase JH2 domain at low concentrations (Table 3), it is not affecting the kinase kinetics.
5
This result contrasts the expectations from the mutational studies, though it is consistent with an
2
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 51
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
alternative view of JH2-selective 10 as an ATP surrogate. The present finding for the one
compound 10 does not imply that all JH2-selective binders will yield qualitatively similar results.
If sufficient structural variety in JH2-binders can be explored, there will be a range of structural
effects on the JH2/JH1 interface that can be expected to provide alternative outcomes.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
CONCLUSION
Janus kinases are intracellular, non-receptor tyrosine kinases that are activated by cytokine binding
to cytokine receptors resulting in stimulation of the JAK-STAT signaling pathway. Aberrant
activation of this signaling pathway is associated with numerous immune disorders and cancer.
The present focus is on selective reduction of the hyperactivation that occurs with the V617F
mutation in the JH2 domain of JAK2 kinase, which is the primary source of myeloproliferative
neoplasms. The initial goal has been to discover small molecules that selectively bind to the ATP
binding site in the JH2 pseudokinase domain of JAK2 in preference to the ATP binding site in the
JH1 kinase domain. Starting from the non-selective, multi-kinase inhibitor JNJ7706621 (1), new
compounds were systematically designed, synthesized, and assayed that ultimately yielded
compounds such as 10, 12, and 15, which bind more strongly to JAK2 JH2 than 1 and that show a
substantial preference for binding JAK2 JH2 over JH1. The design was particularly challenging
owing to the polarity of the ATP binding sites in the area which is amenable to differentiation of
binding to JAK2 JH1 and JH2, and which is proximal to the region affected structurally by the
V617F mutation. The ability to obtain multiple crystal structures for complexes with JAK2 JH2
was most valuable in monitoring the progress of the designs, and the ability to perform
fluorescence polarization assays for binding to all three component proteins, JAK2 JH1, JH2, and
V617F, was essential. Initial further exploration of 1 and 10 was pursued by examining the effects
2
2
ACS Paragon Plus Environment

Page 23 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
of the compounds on autophosphorylation of full-length JAK2 and V617F JAK2. 1 was
established to be a low-μM inhibitor of both enzymes, while 10 showed little effects on the kinase
activity below concentrations of ca. 25 μM. The interpretation is that, in contrast to expectations
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5
from mutational studies, 10 binds well to the ATP-site of the pseudokinase JH2 domain and
performs much the same function as ATP. However, it may also be expected that alternative JH2
ligands may perturb the proximal JH1/JH2 interface in different manners and lead to possible
deactivation or activation of the oncogenic V617F JAK2.
EXPERIMENTAL SECTION
Generation of Recombinant Baculoviruses. The isolated JH1 and JH2 domains of human
JAK2 were expressed in baculovirus-infected Sf9 insect cells and purified similarly to the
previously reported procedure.9,17 The two JH2 domain constructs contain residues 536–812 (with
either mutations W659A, W777A, F794H, or mutations W777A, F794H, V617F), followed by a
C-terminal thrombin cleavage site and 6×His-tag. The reported JH1 domain construct included an
N-terminal 6×His-tag, followed by a TEV cleavage site and residues 840-1132. Recombinant
bacmid and baculoviruses were generated using the Bac-to-Bac baculovirus expression system
(
Invitrogene). DH10Bac competent cells were transformed with recombinant pFastBac plasmid
containing the gene of interest to generate the recombinant expression bacmid. P1, P2 and P3
baculovirus stocks were produced according to the manufacturer’s instructions. Sf9 cells were
grown in HyClone SFX-Insect cell culture media (GE Healthcare) at 27 °C.
JAK2 Protein Expression and Purification. Sf9 cells were grown in HyClone SFX-
6
Insect cell culture media to a density of 2.5–4.0×10 cells/mL, followed by transfection with P3
baculovirus stock. After incubation for 48 h at 27°C, cells were harvested and separated from the
supernatant by centrifugation (4000 rpm, 30 mins). Purification of the JH1 and JH2 domains of
JAK2 was performed in an identical manner. Cell pellets were resuspended in lysis buffer
composed of 20 mM Tris pH 8.0, 500 mM NaCl, 20% glycerol, 0.25 mM TCEP, and cOmplete
EDTA free protease inhibitor (Roche). Cells were lysed by sonication, followed by pressure
homogenization using an Emulsiflux cell disruptor (Avestin). Lysate was separated from cell
2
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 51
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
debris by centrifugation (45 min, 16,500 rpm). Ni-NTA agarose beads (Qiagen) were added in
batch mode, and incubated for 2 h at 4°C. Beads were washed with lysis buffer containing 10 mM
imidazole, and JAK2 protein was eluted with lysis buffer containing 200 mM imidazole. The
eluate was dialyzed overnight at 4°C using a MWCO 3.5 kDa Slide-A-Lyzer dialysis cassette
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(Thermo Fisher Scientific) against a low salt dialysis buffer composed of 20 mM Tris pH 8.0, 25
mM NaCl, 20% glycerol, and 0.25 mM TCEP. The dialysis product was filtered through a 0.45
µm membrane and loaded onto a pre-equilibrated Mono Q HR 16/19 column (GE Healthcare)
linked to an ÄKTA pure protein purification system (GE Healthcare). Protein was eluted applying
a linear gradient starting with dialysis buffer and ending with dialysis buffer containing 500 mM
NaCl. JAK2 fractions were pooled and applied to a Superdex 75 10/300 (GE Healthcare) pre-
equilibrated with a buffer composed of 20 mM Tris pH 8.0, 100 mM NaCl, 10% glycerol, and 1.0
mM TCEP. Purified protein was aliquoted, flash-frozen in liquid nitrogen, and stored at -80°C.
1
7
Competitive FP assay. In a flat black bottom 96 well plate (Corning), 200 µL of FP
buffer (20 mM Tris-HCl pH 8.0, 150 mM NaCl, 20% Glycerol, 0.5 mM TCEP, 0.01% Tween 20)
were added to column 1 (blank), 150 µL to column 2, and 140 µL to columns 3-12. 10 µL of 2.96
µM of JAK2-JH2 WT (3.52 µM for JAK2-JH2-VF, and 6.93 µM for JAK2-JH1), were added to
columns 3-12, followed by the addition of 2 µL of DMSO to columns 1-3. 2 µL of inhibitor in
DMSO at different concentrations were added from column 4 to 12. 50 µL of 24 nM of tracer were
added to columns 2-12. Fluorescence polarization was measured at λexc = 485 ± 20 nm, λem = 535
±
25 nm for 1 hour. Experiments were carried out in quadruplicate with three independent
experiments. Data were analyzed by a least-squares non-linear fit, generated using Prism 7 in order
to determine the compound’s IC50. K
d
values for each inhibitor were calculated using the following
푡
푑
equation based on the IC50, K
d
of the tracer (퐾 ), total (L
t
) and bound (L ) tracer, as well as total
b
27
protein concentration (푃 ).
푡
푡
푑
퐿 ꢀ퐶 퐾
푏
50
퐼
퐾 =
푑
푡
푑
푃 퐿 + 퐿 (푃 − 퐿 + 퐿 − 퐾 )
푡
푡
푏
푡
푡
푏
Microscale Thermophoresis Assay. MST measurements were performed with a Monolith
NT.115Pico device (NanoTemper Technologies). JH2 domain protein (triple mutant W659A,
W777A, F794H, or triple mutant W777A, F794H, V617F) was fluorescently labeled with the His-
tag labeling kit RED-tris-NTA 2nd generation. All dilutions were prepared with a buffer composed
of 20 mM HEPES pH 8.0, 150 mM NaCl, 5% glycerol, and 0.05% Tween. Protein was labeled by
2
4
ACS Paragon Plus Environment

Page 25 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
incubating a mixture of 150 nM protein and 50 nM dye for 30 min at ambient temperature. MST
measurements were performed with protein and dye concentrations adjusted to 30 nM and 10 nM,
respectively. The serial dilution of 1 and 10 ranged from 160 µM to 0.00488 µM, for 12 from 20
µM to 0.00488 µM, and for 15 from 20 to 0.00061 µM. All measurement samples contained a
constant DMSO concentration of 3%. Measurements were performed with standard capillaries
(mutant W659A, W777A, F794H) or premium capillaries (mutant W777A, F794H, V617F),
medium MST power, and 5% excitation power at ambient temperature. All measurements were
performed in triplicate, and were analyzed with the MO.Affinity Analysis software (NanoTemper).
All dose-response curves are displayed in Figure S2.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Protein Crystallization. Crystals of JAK2 JH2 were grown by hanging drop vapor
diffusion at 4 °C. Crystals were prepared by adding 1 µL of protein (6 mg/mL) in a solution
composed of 20 mM Tris pH 8.0, 100 mM NaCl, 10% glycerol, and 1 mM TCEP to 1 µL of
reservoir solution composed of 0.1 M Tris pH 8.0, 0.2 M sodium acetate, 12–24% PEG4000, and
1
mM TCEP. Crystallization was induced by streak seeding, and crystals grew to full size within
a week. Complexes of JH2 with small-molecule ligands were prepared by ligand soaking.
Therefore, crystals were transferred into a solution of 0.1 M Tris pH 8.0, 0.2 M sodium acetate,
22% PEG4000, 1 mM TCEP, 8% DMSO, and 4 mM of the respective small-molecule ligand. After
2
4 h of incubation, crystals were briefly exposed to a cryobuffer (0.1 M Tris pH 8.0, 0.2 M sodium
acetate, 22% PEG4000, 1 mM TCEP, 8% DMSO, 4 mM of the respective small-molecule ligand,
and 20% glycerol), and flash-frozen in liquid nitrogen.
Collection of X-ray Diffraction Data and Data Processing. Crystal structures for
complexes of five of the new compounds (3, 4, 9, 10 ,12) with JAK2 JH2 were obtained at 1.7 –
2
.1 Å resolution. All datasets were collected at 100 K. Datasets of JH2 in complex with 4, 10, and
12 were collected in-house on a Rigaku MicroMax-007HF X-ray generator (Cu rotating anode;
λ=1.54 Å) with a Dectris Pilatus 200K detector. The dataset of JH2 in complex with 3 was
collected at the Advanced Photon Source (APS; Argonne, Illinois, USA) on beamline 24-ID-E
with a Dectris Eiger 16M detector (λ=0.97918 Å). The dataset of JH2 in complex with 9 was
collected at the APS beamline 24-ID-C with a Dectris Pilatus 6M detector (λ=0.97910 Å). Datasets
ͅ
ͅ
ͅ
2
8
of 4, 10, and 12 were indexed, integrated, and scaled with HKL2000. Datasets of 3 and 9 were
29
indexed, integrated, and scaled with XDS. Diffraction data and refinement statistics are listed in
Table S1. Refinement procedures are also provided along with omit maps in Figure S3.
2
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 51
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Cloning and Expression of Full-Length JAK2. Full-length cDNA encoding human
JAK2 (residues 1-1132) wild-type (NP_001309123) and V617F mutant with C-terminal FLAG-
Tag (DYKDDDDK) were amplified by PCR, and subcloned into a modified pOptiVec expression
vector (Invitrogen). JAK2 constructs were expressed in HEK293T cells grown at 37 °C at 5% CO
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
in DMEM (Gibco) supplemented with 10% (v/v) FBS (Gibco) and 1% (v/v) penicillin-
streptomycin (Gibco). HEK-293T cells were transiently transfected using Lipofectamin 2000
(
Invitrogen) according to manufacturer’s instructions. 36 h post-transfection, cells were washed
twice with ice-cold PBS, and lysed with ice-cold lysis buffer (50 mM HEPES pH 7.4, 150 mM
NaCl, 10% (v/v) glycerol, 1% (v/v) Triton-X 100, 1 mM EDTA, 1 mM EGTA, 25 mM NaF, 1.5
mM MgCl
2
, 1.0 mM Na
3
VO , Roche complete mini EDTA-free protease inhibitor cocktail
4
mixture). The lysate was centrifuged (20 min, 13000×g), and the lysate supernatant was flash-
frozen in liquid nitrogen and stored at -80 °C.
Immunoprecipitation and In Vitro Kinase Assay. JAK2 protein was
immunoprecipitated from the lysate supernatant by adding anti-FLAG M2 antibody (Sigma-
Aldrich, no. F1804) and protein G-PLUS agarose (Santa Cruz Biotechnology, no. sc-2002)
followed by incubation overnight while rocking at 4°C. Immunoprecipitates were washed four
times with wash buffer (50 mM HEPES pH 7.4, 150 mM NaCl, 10% (v/v) glycerol, 0.1% (v/v)
Triton-X 100, 1 mM EDTA, 1 mM EGTA, 25 mM NaF, 1.5 mM MgCl
2
, 1.0 mM Na
3
VO , Roche
4
complete mini EDTA-free protease inhibitor cocktail mixture), and once with kinase reaction
3
2
buffer (50 mM HEPES pH 7.4, 150 mM NaCl, 0.5 mM DTT, 5 mM MnCl ). The [γ- P]ATP in
2
vitro kinase activity assay was performed based on a previously published protocols.30,31 Washed
immunoprecipitates were divided into equal parts, centrifuged, their residual solvent was removed,
and the resulting pellets were resuspended in 25 µL kinase reaction buffer containing different
concentrations of 1 or 10, followed by incubation for 1 h at 4°C. The autophosphorylation reaction
of JAK2 was initiated by adding 25 µL of phosphorylation mixture consisting of kinase reaction
buffer supplemented with 10 µM cold ATP, and 2 µCi (for JAK2 V617F) or 5 µCi (for JAK2 wild-
3
2
type) of [γ- P]ATP (EasyTides, PerkinElmer) per reaction. The mixture was allowed to react for
15 min at 30°C (within the linear range of kinase activity), and stopped by putting on ice and
adding 18 µL of reducing Laemmli sample buffer (4×). Samples were heated at 95 °C for 5 min,
and run by 7.5% SDS-PAGE. Gels were rocked in a solution of 10% glycerol, 20% ethanol for 30
min, dried with a vacuum drier, and autoradiographed using a phosphor imager. To calculate IC50
2
6
ACS Paragon Plus Environment

Page 27 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
32
values, phosphor autoradiography was quantified using ImageJ, and curves were plotted with
Prism 8.0 (GraphPad Software Inc., La Jolla, CA).
General Synthetic Methods. All purchased compounds were used as received from
vendors without further purification. Reactions were conducted under a nitrogen atmosphere and
monitored by thin layer chromatography on Merck silica gel plates pre-coated with fluorescent
indicator F254. Visualization of plates was achieved with UV light or potassium permanganate
stain. Mass analysis of intermediates was done with an Agilent 6120 Quadrupole LC/MS
instrument via electrospray ionization. Chromatographic purification was performed with a
Teledyne ISCO CombiFlash automated system employing RediSep Normal Phase Silica (particle
size: 35-70 μm; pore size: 60 Å) or RediSep Gold Normal Phase Silica (particle size: 20-40 μm;
pore size: 60 Å) disposable cartridge columns. RediSep Gold C18 reusable columns (particle size:
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
0 – 40 μm spherical; pore size: 100 Å) were employed for reversed-phase chromatography.
Preparative reverse phase HPLC was performed using a Shimadzu Prominence system equipped
with LC-20AP pumps, CBM-20A Communications BUS module, SPD-20A UV/vis detector, SIL-
1
0AP autosampler, FRC-10A fractions collector and a Waters SymmetryPrepTM C8, 19 x 300
mm column (particle size: 7 μm; pore size: 100 Å) with a gradient of 0.1 % formic acid in water
1
13
and 0.1 % formic acid in acetonitrile as the mobile phase. H and C NMR spectra were recorded
on Agilent DD 400 MHz, DD 500 MHz, or DD 600 MHz instruments. The IR spectrum was
2
2
2
recorded on Thermo Scientific Nicolet 6700 FT-IR spectrometer using a thin film of solid sample.
HRMS analysis of final products was done on a Waters Xevo QTOF with a Z-spray electrospray
ionization source. Purity of all assayed compounds was determined on a Shimadzu Prominence
HPLC equipped with an Agilent Poroshell 120 SB-C18 2.7 μm column, using 0.1% TFA in water
and 0.1% TFA in acetonitrile as the mobile phase. The compounds (2-4, 6-10, 12, 14, 15) were ≥
95% pure by analytical HPLC, while the purities of three non-key compounds, 5, 11, and 13, were
9
4%, 91%, and 90%, respectively. The HPLC traces of all compounds showed no minor peaks
above 3%. It is possible that the HPLC results may be affected by formation of dimers of the
carboxylic acids or alternative protonation states. The systematic SAR data and multiple crystal
structures reported here demonstrate that the reported binding data for this series is reflective of
the compounds themselves. Spectra of all new compounds are provided in the SI.
Synthesis of Intermediates and Final Compounds.
2
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 51
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Scheme 3. Synthesis of 17a-ba
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
aReagents and conditions: (a) THF, reflux, 24 h; (b) NaH, THF, reflux, 24 h; (c) Hydrazine, reflux,
-19 h.
8
Phenyl-N'-cyano-N-(4-sulfamoylphenyl)carbamimidate (16a). Sulfanilamide (4.0 g,
23.2 mmol) was combined with diphenyl cyanocarbonimidate (5.5 g, 23.2 mmol) in anhydrous
tetrahydrofuran (26 mL, 0.9 M). The reaction mixture was stirred at reflux for 24 h and afterwards
was concentrated under reduced pressure. The residue was suspended in dichloromethane, filtered,
washed with dichloromethane, and dried overnight to provide the title product as a pale white solid
1
(6.2 g, 84% yield). H NMR (500 MHz, DMSO-d
6
) δ 11.12 (s, 1H), 7.83 (d, J = 8.5 Hz, 2H), 7.66
1
3
(d, J = 8.5 Hz, 2H), 7.46 (t, J = 7.8 Hz, 2H), 7.36 (s, 2H), 7.35 – 7.29 (m, 3H). C NMR (126
MHz, DMSO-d
6
) δ 151.43, 140.97, 139.12, 129.85, 127.39, 126.51, 126.39, 123.32, 120.79,
+
112.39. ESI-MS m/z: [M+H] : 317.1.
Phenyl-N'-cyano-N-(4-cyanophenyl)carbamimidate (16b). Sodium hydride (760 mg
o
(60% dispersion in mineral oil), 19.0 mmol) was added at 0 C to a mixture of 4-aminobenzonitrile
o
(2.2 g, 19.0 mmol) in anhydrous tetrahydrofuran (65 mL, 0.3 M). The mixture was stirred at 60 C
for 45 min, followed by portion-wise addition of diphenyl cyanocarbonimidate (4.0 g, 17.0 mmol).
The reaction was brought to reflux and allowed to run for 24 h. The crude mixture was quenched
with methanol and solvent was evaporated under reduced pressure. Hot acetone was added until
complete dissolution of the residue, at which point hexanes were added until the solution became
cloudy. The precipitate was filtered to provide the title product as a white, fluffy solid (3.11 g,
1
7
8
0% yield). H NMR (400 MHz, DMSO-d
6
) δ 11.21 (s, 1H), 7.86 (d, J = 8.1 Hz, 2H), 7.66 (d, J =
13
.1 Hz, 2H), 7.46 (t, J = 7.6 Hz, 2H), 7.37 – 7.28 (m, 3H). C NMR (151 MHz, DMSO-d
6
) δ
157.71, 155.53, 154.16, 132.12, 128.84, 124.11, 123.67, 121.98, 121.78, 120.32, 100.73. ESI-MS
+
m/z: [M+H] : 263.1.
2
8
ACS Paragon Plus Environment

Page 29 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
7
4
-((5-amino-1H-1,2,4-triazol-3-yl)amino)benzenesulfonamide (17a). 16a (5.0 g, 15.9
mmol) was suspended in anhydrous tetrahydrofuran (18 mL, 0.9 M) and hydrazine (17.5 mL (1 M
o
in THF), 17.5 mmol) was added at 0 C. The reaction mixture was warmed to room temperature
and stirred at reflux for 7.5 h. The mixture was then filtered, washed three times with
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
tetrahydrofuran, and dried overnight to afford the title product as white solid (3.4 g, 84.5% yield).
1H NMR (600 MHz, DMSO-d
) δ 11.30 (s, 1H), 9.19 (s, 1H), 7.60 (q, J = 9.0 Hz, 4H), 7.03 (s,
6
1
3
2
1
H), 5.95 (s, 2H). C NMR (151 MHz, DMSO-d ) δ 157.22, 155.50, 145.46, 133.07, 126.72,
6
+
14.55. ESI-MS m/z: [M+H] : 255.1.
4
-((5-amino-1H-1,2,4-triazol-3-yl)amino)benzonitrile (17b). 16b (1.0 g, 3.8 mmol) was
suspended in dry tetrahydrofuran (11 mL, 0.3 M) and combined with hydrazine (7.6 mL (1 M in
o
THF), 7.6 mmol) at 0 C. The reaction mixture was brought to reflux and stirred for 19 h, then was
concentrated and purified by column chromatography (DCM/MeOH) to afford a yellow solid (461
1
mg, 61% yield). H NMR (400 MHz, DMSO-d
6
) δ 11.38 (s, 1H), 9.40 (s, 1H), 7.60 (q, J = 8.9 Hz,
1
3
4
1
H), 5.97 (s, 2H). C NMR (151 MHz, DMSO-d ) δ 155.23, 147.97, 146.51, 133.04, 120.16,
6
+
15.49, 99.10. ESI-MS m/z: [M+H] : 201.1.
Scheme 4. Synthesis of 18a-ba
aReagents and conditions: (a) K
Pd/C, THF:MeOH, 20h.
CO
, MeCN, r.t., 16.5 h; (b) K
CO
, acetone, reflux, 5 h (c) H ,
2
3
2
3
2
3
3
Methyl 2-(4-nitrophenoxy)acetate (18a’). Para-nitrophenol (1.5 g, 10.8 mmol) was
suspended in dry acetonitrile (50 mL, 0.2 M) and followed by addition of potassium carbonate
(
2.98 g, 21.6 mmol). The mixture was allowed to stir until homogenized, methyl-2-bromoacetate
(1.0 mL, 10.8 mmol) was added, and stirring continued for 16.5 h at room temperature. The crude
mixture was concentrated and diluted with ethyl acetate (50 mL), washed with water (2 x 20 mL)
and brine (2 x 15 mL), and was dried over anhydrous sodium sulfate. Solvent was evaporated
1
under vacuum to afford the title compound as a white solid (2.0 g, 89% yield). H NMR (400 MHz,
2
9
ACS Paragon Plus Environment