Allosteric Activation of Striatal-Enriched Protein Tyrosine Phosphatase (STEP, PTPN5) by a Fragment-like Molecule

Article abstract of DOI:10.1021/acs.jmedchem.8b00857

Protein tyrosine phosphatase non-receptor type 5 (PTPN5, STEP) is a brain specific phosphatase that regulates synaptic function and plasticity by modulation of N-methyl-d-aspartate receptor (NMDAR) and α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor (AMPAR) trafficking. Dysregulation of STEP has been linked to neurodegenerative and neuropsychiatric diseases, highlighting this enzyme as an attractive therapeutic target for drug discovery. Selective targeting of STEP with small molecules has been hampered by high conservation of the active site among protein tyrosine phosphatases. We report the discovery of the first small molecule allosteric activator for STEP that binds to the phosphatase domain. Allosteric binding is confirmed by both X-ray and ¹⁵N NMR experiments, and specificity has been demonstrated by an enzymatic test cascade. Molecular dynamics simulations indicate stimulation of enzymatic activity by a long-range allosteric mechanism. To allow the scientific community to make use of this tool, we offer to provide the compound in the course of an open innovation initiative.

Full text of DOI:10.1021/acs.jmedchem.8b00857

Article
pubs.acs.org/jmc
Cite This: J. Med. Chem. XXXX, XXX, XXX−XXX
Allosteric Activation of Striatal-Enriched Protein Tyrosine
Phosphatase (STEP, PTPN5) by a Fragment-like Molecule
†
‡
†
†
Christofer S. Tautermann,*,† Florian Binder, Frank H. Buttner, Christian Eickmeier, Dennis Fiegen,
Ulrike Gross,† Marc A. Grundl,† Ralf Heilker,‡ Scott Hobson,§ Stefan Hoerer,† Andreas Luippold,‡
Volker Mack,§ Florian Montel,† Stefan Peters,† Supriyo Bhattacharya,∥ Nagarajan Vaidehi,∥
Gisela Schnapp,† Sven Thamm,‡ and Markus Zeeb†
̈
‡
§
†Medicinal Chemistry, Drug Discovery Sciences, and CNS Diseases Research, Boehringer Ingelheim Pharma GmbH & Co KG,
Birkendorfer Straße 65, D-88397 Biberach an der Riss, Germany
∥Department of Molecular Immunology, Beckman Research Institute of the City of Hope, 1500, E. Duarte Road, Duarte, California
91010, United States
S
* Supporting Information
ABSTRACT: Protein tyrosine phosphatase non-receptor type 5 (PTPN5, STEP)
is a brain specific phosphatase that regulates synaptic function and plasticity by
modulation of N-methyl-D-aspartate receptor (NMDAR) and α-amino-3-hydroxy-
5-methyl-4-isoxazolepropionic acid receptor (AMPAR) trafficking. Dysregulation
of STEP has been linked to neurodegenerative and neuropsychiatric diseases,
highlighting this enzyme as an attractive therapeutic target for drug discovery.
Selective targeting of STEP with small molecules has been hampered by high
conservation of the active site among protein tyrosine phosphatases. We report the
discovery of the first small molecule allosteric activator for STEP that binds to the phosphatase domain. Allosteric binding is
confirmed by both X-ray and 15N NMR experiments, and specificity has been demonstrated by an enzymatic test cascade.
Molecular dynamics simulations indicate stimulation of enzymatic activity by a long-range allosteric mechanism. To allow the
scientific community to make use of this tool, we offer to provide the compound in the course of an open innovation initiative.
changes in synaptic plasticity such as long-term potentiation.11
Currently, the only available pharmacological tool to modulate
INTRODUCTION
■
STEP (striatal-enriched protein tyrosine phosphatase) is a
critical regulator of glutamatergic synapse function and
neuronal survival. STEP acts on targets directly involved in
STEP activity is TC-2153 (1 in Figure 1), which has been used
successfully in several pharmacological studies.8,12,13 TC-2153
is an inhibitor covalently binding to the catalytic cysteine
regulation of synaptic plasticity such as the NMDA and AMPA
C496, with moderate in vitro selectivity over other tyrosine
receptors, as well as several targets regulating these receptors
phosphatases such as HePTP and PTP1B.8 In addition a quite
(e.g., ERK1/2, p38, FYN, and PYK2).1,2 Dysregulation of
potent substrate derived orthosteric inhibitor has recently been
STEP has been attributed to the pathophysiology of various
reported.14 Numerous lines of evidence suggesting that STEP
devastating neuropsychiatric disorders.1 Whereas an upregula-
tion of STEP has been reported in indications such as
Alzheimer’s disease (AD),3 Parkinson’s disease,4 schizophre-
nia,5 or fragile X syndrome,6 a downregulation has been linked
to Huntington’s disease,7,8 alcohol abuse,9 cerebral ischemia,10
and stress related disorders.2 Therefore, modulation of STEP
activity holds promise for the treatment of multiple CNS
diseases with unmet medical need.
The efficacy of STEP modulation to alter synaptic function
and affect processes of learning and memory has been
indicated by genetic depletion of the enzyme5,7 as well as
Figure 1. (1) Orthosterically binding STEP inhibitor TC-2153 and
pharmacologically with a moderately selective orthosteric
inhibitor.8 STEP depletion led to enhanced synaptic plasticity
and rescued a behavioral phenotype in triple transgenic AD
mice.3 STEP knockout mice have been used to validate various
downstream effects such as the modification of the
phosphorylation state and surface expression of target
NMDA9 and AMPA10 receptors, which have been linked to
(2) the fragment screening hit binding to an allosteric pocket of the
catalytic domain of STEP.
Special Issue: Allosteric Modulators
Received: May 30, 2018
© XXXX American Chemical Society
A
DOI: 10.1021/acs.jmedchem.8b00857
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 2. Sequence identities throughout the class of human PTPs. Lower triangle: sequence identity within 6 Å of the active site cysteine. Upper
triangle: sequence identity within 6 Å of compound 2 bound to the allosteric site. Species conservation is displayed in the upper left 3-by-3
submatrix.
is a valuable target with the lack of selective inhibitors triggered
discovery efforts toward a better starting point for STEP
modulation. However, given the high polarity of the catalytic
pocket and its strong structural conservation among the family
of PTPs (Figure 2, lower triangle), the development of
selective and potent inhibitors of PTPs is a daunting challenge.
This is even more true for STEP inhibitors, for which CNS
penetration is mandatory.15 Therefore, selective tool ligands
have been reported for only very few PTP family members;16
notably to be mentioned is the recent SHP2 allosteric inhibitor
by Chen et al.17 Due to the lower conservation of remote
pockets, the detection of allosterically acting ligands has been
brought into focus, and for other phosphatases, such as
PTP1B, allosteric inhibitors have been reported.18,19 In the
present study we describe the identification and optimization
of allosteric STEP ligands that were validated as specific
binders by biophysical methods. Beyond that, the most potent
compound in this study (BI-0314) has been shown to act as a
selective allosteric STEP activator in several biochemical
assays.
domain bears various conserved structural motifs, such as the
WPD loop, which is crucial for the catalytic step, as its
aspartate (D461) mediates proton transfer to the phosphate
leaving group.22 However, similarity analysis of the active sites
of the PTP family (Figure 2, lower triangle) shows strong
sequence conservation, indicating that the identification of
selective active site binders will be very demanding, if not
impossible. Therefore, focus was put on the identification of
ligands that do not bind to the active site of STEP.
Fragment based screening has proven to be a very powerful
method to detect ligands of low molecular weight binding to a
target of interest.23 A biophysical readout allows the detection
of fragments binding to any site of the construct used,
including allosteric pockets.24,25 To probe for allosteric pockets
on STEP, a fragment based screen was conducted on
hSTEP46. A library of 3083 fragments was tested using
complementary technologies, i.e., saturation transfer difference
NMR (STD NMR), differential scanning fluorometry (DSF),
and microscale thermophoresis (MST). The primary screen hit
rate for STD NMR was very high (>15%), whereas DSF and
MST yielded low hit rates (0.5% and 3%, respectively), which
is indicative of hits with rather low affinity. All primary hits
were subjected to 2D 1H,15N TROSY NMR, and finally, seven
hits could be confirmed to bind to the PTP domain.
Comparison of the chemical shift perturbation (CSP) pattern
caused by the fragment hits with the ones caused by the
orthosteric pan-phosphatase inhibitor Na vanadate demon-
strated that an additional binding site for 2 (Figure 1) had
been identified on hSTEP (Figure S1).
RESULTS AND DISCUSSION
■
Detection of Allosterically Binding Fragments. STEP
is a multidomain protein that exists as two splice variants, the
membrane anchored longer isoform STEP61 and the cytosolic
STEP46.20 Both isoforms share the identical kinase interaction
motif (KIM) and the protein tyrosine phosphatase (PTP)
domain with the phosphatase consensus motif C(X)5R. The
KIM domain mediates binding to various target kinases with
high affinity, while the PTP domain catalyzes their subsequent
dephosphorylation.21 To avoid developing ligands that
potentially suffer from substrate specificity, we preferred
targeting the PTP domain over the KIM domain. The PTP
Following the fragment based screen, the X-ray structure of
the PTP domain of hSTEP in complex with 2 was solved. The
overall structure superimposes very well with published STEP
structures.22,26 Residues around the active site P- and WPD-
B
DOI: 10.1021/acs.jmedchem.8b00857
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 3. (A) Crystal structure of human STEP with compound 2 (orange sticks) and the distances to the active site and the WPD loop. (B)
Binding mode of 2 in human STEP. (C) Binding mode of BI-0314 in murine STEP.
loops are well resolved in the electron density, while the
density for the loop between β3 and β4, namely residues 400−
408, is weakly defined. The active site C496 likely shows an
alternative conformation and positive difference density that
could be interpreted as an acetyl moiety (X-ray statistics in
Table S1).
by 2D 15N TROSY NMR (based on a published assignment of
mSTEP,27 Figure S1).
At this point, we had identified a new binding pocket in
STEP that was also shown to be the specific ligand binding site
of 2 in solution, and we had optimized the crystal system to be
suitable for fragment growing by switching to a different
species with a conserved allosteric binding site.
Compound 2 was found to bind to a remote pocket about
20 Å distant from the active site, on the back side of the
protein (Figure 3), which will be referred to as “allosteric
pocket”. The sequence conservation of the allosteric pocket
residues across the PTP family is very low (Figure 2, upper
triangle), making this pocket ideal for the design of STEP
selective ligands. Fortunately, the allosteric pocket is quite
conserved within STEP orthologs (mouse, rat), reducing the
risk of potential species selectivity issues. Within the allosteric
pocket, compound 2 is surrounded by helices α3, α4 and the
loop/310-helix between β-sheets β1 and β2 (nomenclature
based on Eswaran et al.26). Its trifluoromethyl moiety points
into a hydrophobic cleft formed by the side chains of Y369,
I492, F506, and S510. The positively charged amine moiety of
2 forms a salt bridge with E365 and in addition makes water
bridged contacts to C489 and Q514. The hydroxy moiety of 2
interacts with the side chain of E479 via a H-bond. In direct
proximity to the binding site and the compound are residues of
a symmetry related molecule (helix α2′) that closes the pocket
to the solvent (Figure S2).
As fragment growing of 2 would lead to a clash with this
symmetry related molecule, a new crystal system for fragment
optimization had to be established. For this purpose, murine
STEP (mSTEP) was shown to be ideally suited, as a different
crystal packing allowed for a fragment binding site without any
close crystal contacts (Figure S2). The X-ray structure of 2 in
mSTEP revealed a binding pose identical to the one in hSTEP
(not shown). Importantly, specific binding of 2 to the allosteric
pocket of mSTEP could also be confirmed in aqueous solution
Fragment Optimization. As determined by 15N NMR
dose−response testing, the affinity of fragment 2 was >5 mM
(data not shown). However, the fragment still showed
significant shifts at 500 μM when tested by 15N NMR on
hSTEP (Figure S1). These results qualified 15N NMR to
become the driving assay for fragment optimization. Three 15N
NMR peaks were selected, and the cumulated H and N shifts
were taken as surrogates for the changes in affinity for every
compound (see Figure S1 for peak definition). The
optimization of the fragments focused on the three substituents
of 2 as depicted in Figure 4. The CF3 group was found to be
superior to other substituents at this position. The OH group
of 2 can be exchanged by a variety of groups that occupy the
crevice on the protein surface, still leading to measurable NMR
shifts. Substitution by NH2 (compound 3) or omission
(compound 4) led to slightly weaker NMR signals compared
to 2 (Table S2). The basic side chain of 2 proved to be
essential for binding, which can be rationalized by the key
interaction with E365. A scan of basic side chains eventually
led to the discovery of 5, revealing stronger NMR shifts
compared to 4 due to additional hydrogen bonds beyond the
salt bridge to E365 (Figure 3). Indeed, various other guanidine
bearing analogs also displayed strong NMR shifts (data not
shown). Further optimization efforts resulted in BI-0314,
which was characterized by various biophysical techniques.
Crystallization trials with BI-0314 and hSTEP were not
successful due to crystal contacts in the vicinity of the allosteric
binding site. Nevertheless, the X-ray co-structure of BI-0314
could be determined using mSTEP (Figure 3). Confirmation
C
DOI: 10.1021/acs.jmedchem.8b00857
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
(ITC) and 15N NMR on hSTEP (Figures S3 and S4).
Depending on the conditions and the experimental technique,
the resulting affinity was found to be 38 μM by ITC
measurements (n = 3) and around 800 μM by 15N NMR
titration experiments. ITC experiments also revealed that
binding of BI-0314 is mostly entropy driven (Figure S4).28 To
have a first estimate for selectivity on a biophysical level,
binding of BI-0314 was tested on the prototypic protein
phosphotyrosine phosphatase 1B (hPTP1B) by 15N NMR,
where no binding could be observed (data not shown).
In summary, the optimization efforts yielded the fragment
like (i.e., MW < 300 Da) compound BI-0314, which is
specifically binding to STEP. It has a significantly improved
affinity over 2 and is selective over hPTP1B and therefore
ready for characterization in functional assays.
Detection of Functional Allostery. To be close to the
physiological situation, the primary enzymatic assay was
performed with a phosphorylated peptide substrate derived
from the endogenous substrate FYN, called pFYN-peptide. In
this assay, concentration dependent amplification of hSTEP
catalytic activity by BI-0314 could be observed, using both
AlphaLISA and mass spectrometry (MS) as readout. The
turnover of pFYN-peptide is increased by ≈30% at 100 μM
and ≈60% at 500 μM. A whole set of complementary
experiments was performed to demonstrate specificity of
hSTEP activation (see Table 1).
Figure 4. Snapshots of the fragment optimization of compound 2,
eventually resulting in BI-0314.
of the specificity of the binding site in solution was achieved by
15N NMR on mSTEP (Figure 5, top left).27 In analogy, very
strong chemical shift perturbations are detected for human
STEP in complex with BI-0314, which clearly differ from 15N
NMR shifts induced by the orthosteric inhibitor vanadate
(Figure 5, lower panel). Furthermore, the binding affinity of
BI-0314 was determined by isothermal titration calorimetry
Substrate dependency was tested using DiFMUP, which is a
generic tool-substrate for phosphatases, changing its fluo-
rescence when dephosphorylated.29 Similar to the results
obtained for the pFYN substrate, activation of hSTEP by BI-
Figure 5. Superposition of 2D 15N TROSY NMR spectra of 100 μM 2H,15N labeled human and mouse STEP PTP domain in the absence (black)
or presence of BI-0314 or Na vanadate (red). Left panel: mouse STEP PTP domain in complex with 500 μM BI-0314 (top) or 2 mM Na vanadate
(bottom). Cross peaks were assigned according to ref 27. Cross peaks reflecting the allosteric or orthosteric site are highlighted in green or black,
respectively. Right panel: human STEP PTP domain in complex with 500 μM BI-0314 (top) or 2 mM Na vanadate (bottom). Cross peaks
indicative for orthosteric (black) or allosteric (green) site binding are circled.
D
DOI: 10.1021/acs.jmedchem.8b00857
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Table 1. Enzymatic Assays To Test for Specificity of STEP Activation by BI-0314
assay
technology
enzyme
hSTEP46
substrate
effect of BI-0314 in assay
learning
pFYN-peptide
AlphaLISA
Activating, 56% 5% at 500 μM (n = 8),
33% 12% at 100 μM (n = 12)
BI-0314 has an activating effect on hSTEP
dephosphorylation capability on pFYN-peptide.
hSTEP46
pFYN-peptide
pFYN-peptide
DiFMUP
RapidFire
(MS)
Activating, 28% 5% at 100 μM (n = 4)
Activating effect is assay technology independent.
PTP domain of
hSTEP
AlphaLISA
Activating, 61% 6% at 500 μM (n = 10)
Activation also observed on the PTP domain itself.
hSTEP46
hPTP1B
hTCPTP
hPTP1B
Fluorescence
AlphaScreen
AlphaScreen
AlphaScreen
Activating, 48% 8% at 1000 μM (n = 3),
27% 5% at 300 μM (n = 3)
Activating effect is also observed for another substrate and
different assay readout.
Insulin receptor
peptide
No effect at 500 μM (n = 2)
No effect at 500 μM (n = 2)
No effect at 500 μM (n = 2)
Selectivity over hPTP1B
Insulin receptor
peptide
Selectivity over hTCPTP
pFYN-peptide
Activation effect is not pFYN peptide mediated.
Table 2. Michaelis−Menten Parameter Values for the Dephosphorylation of pFYN-Peptide by hSTEP PTP Domain with and
without BI-0314
n
KM [μM]
kcat [s−1
]
k
cat/KM [M−1 s−1
]
hSTEP + pFYN-peptide
BI-0314 + hSTEP + pFYN-peptide
4
8
60.8 7.9
65.4 13.9
0.024 0.002
0.042 0.005
400 20
650 50
0314 could also be demonstrated using DiFMUP. AlphaScreen
assays were performed on the closely related phosphatases
hPTP1B and hTCPTP with an insulin receptor derived
peptide to demonstrate selectivity. BI-0314 was inactive on
both targets. Additionally, BI-0314 did not enhance the
dephosphorylation of the pFYN-peptide by hPTP1B, which
again confirms a hSTEP dependent mechanism rather than a
substrate specific effect. To narrow down the structural basis of
action of BI-0314, we also demonstrated activation in an assay
with the PTP domain of hSTEP. A trend toward activation can
also be demonstrated on mouse STEP PTP domain (Figure
S5), but the effect is smaller than on human STEP.
ceptors.30 Two of the strongest pipelines emerging from the
active site connect to the allosteric binding site of BI-0314, as
shown in Figure 6. On the basis of the mutual information
In order to gain insight into the mechanism of action of BI-
0314 on hSTEP, we derived Michaelis−Menten parameters for
the hSTEP/pFYN-peptide system in the absence or presence
of 500 μM BI-0314 employing the PTP domain of hSTEP and
RapidFire readout (Table 2 and Figure S6). Determined values
for KM are well in line with values reported by Li et al.;21
however kcat is somewhat lower than reported.21 These
differences may arise from different assay conditions employed.
Addition of BI-0314 caused an acceleration of the enzymatic
reaction catalyzed by hSTEP, which is reflected in an increase
of kcat while keeping the KM for the pFYN-peptide unchanged.
Mechanistically this implies that binding of BI-0314 to the
allosteric site causes a higher turnover of the pFYN-peptide
without changing the affinity of this substrate for hSTEP.
Therefore, the kcat/KM ratio, which is a measure for the
catalytic efficiency, is significantly increased by BI-0314. In
analogy to allosteric noncompetitive inhibition, where the
catalytic efficiency is decreased by a reduced kcat with a
constant KM, we identify BI-0314 as an allosteric activator of
hSTEP.
Figure 6. Allosteric pipelines in hSTEP communicating between the
active site and the BI-0314 binding site displaying the allosterically
active residues in both pockets (green residues). Important allosteric
hub residues along the communication pipelines are shown in
magenta.
analysis, residues K407, S497, and R502 in the active site show
highly correlated motion with the allosteric site. Specifically,
S497, which is neighboring the catalytic C496, is known to be
involved in the pTyr dephosphorylation mechanism.31 On the
other hand, looking at the allosteric site, the residues R517 and
E479 show strong allosteric communication with the active
site. These two residues form a salt bridge and also make close
contacts with BI-0314. We speculate that the interaction of BI-
0314 with this salt bridge mediates the allosteric communica-
tion with the active site and thereby stabilizes the active
conformation of STEP. Comparison of apo and ligand bound
Mechanism of Allostery. To investigate the molecular
mechanism of communication between the allosteric binding
site and the active site, we calculated the allosteric
communication pipelines in STEP using the Allosteer method
(Figure S7).30 This method relies on a graph theory algorithm
combined with the mutual information in torsion angles
between residue pairs and has been successfully validated to
study long-range communication in G-protein-coupled re-
E
DOI: 10.1021/acs.jmedchem.8b00857
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
a
Scheme 1. Synthesis of BI-0314
a
Reagents and conditions: (a) Cs2CO3, Pd(dppf)Cl2, dioxane, 100 °C, 82%; (b) RaNi, H2, EtOAc, EtOH, 50 °C, 78%; (c) NH2C(NH)SO2H,
MeOH, 100 °C, 13%; (d) TFA, DCM, rt, 22%.
structures of mSTEP confirms a change in conformation of the
salt bridge (Figure S8) and therefore strengthens our
hypothesis. Further support is coming from a comparison of
residues significantly contributing to allosteric communication
(allosteric hubs in Figure 6) to the corresponding residues in
PTP1B (Figure S11). Upon ligand binding the flexibility of
these residues in PTP1B is altered,32,33 experimentally
demonstrating the allosteric communication along these
pathways.
When corroborating these results with extensive plain MD
simulations, we observe a strong impact on the WPD loop.
Upon binding of BI-0314, the flexibility of the WPD loop
decreases (Figure S9), again proving the long-range allosteric
action extending more than 20 Å across the PTP domain. This
decrease in flexibility and increase in stabilizing interactions
could be confirmed experimentally using differential scanning
fluorometry. Heat induced unfolding of hSTEP revealed an
elevated melting temperature in the presence of 500 μM BI-
0314 (Tm shift of 1.4 °C (n = 7)), whereas for 2 no
stabilization of the native conformation could be detected
(Figure S10).
change in the flexibility of the catalytically important WPD
loop.
The discovery of a functionally active allosteric binding site
on STEP is another example demonstrating that fragment-
based drug discovery is a very powerful method to identify
chemical tools after limited cycles of optimization. As shown
by many groups, the discovery of STEP selective ligands is a
challenging task when employing classical high throughput
screening approaches.8 Until now, no selective small molecule
activator of STEP has been reported. Therefore, the discovery
of an allosteric site on STEP that is amenable to designing
selective modulators of its catalytic activity represents a
breakthrough to successfully targeting STEP. However, to
transform BI-0314 into a ligand that is suitable for in vivo
testing, further optimization or specifically screening for
ligands binding to the reported allosteric site will be required.
Upon having demonstrated hSTEP activation by BI-0314, no
further optimization on ligands targeting this allosteric site was
performed, since we originally had set out to identify STEP
inhibitors. The discovery of BI-0314 and its allosteric site
demonstrates that phosphatases are indeed druggable but will
require new and smart approaches to identify selective ligands
that have the potential to become drugs.
Taking these results together, we are confident that the two
pockets are connected by allosteric communication and
hypothesize that binding of BI-0314 employs this mechanism
to impact the catalytic efficiency of STEP.
The disclosure of BI-0314 should spark research in the
direction of developing allosteric enhancers of phosphatases.
Therefore, we offer to share BI-0314 in the course of an open
innovation initiative (https://opnme.com/) where the com-
pound can be obtained by research groups free of charge.
CONCLUSIONS
■
BI-0314 is the first allosteric ligand reported for hSTEP, which
is a highly interesting target to potentially treat various
psychiatric conditions as well as neurodegenerative diseases.
The allosteric binding pocket on STEP was identified by a
fragment-based screen, where one very weakly binding primary
hit was optimized, yielding a compound with sufficient binding
affinity to allow testing in biochemical assays. To rule out that
the observed effects were assay or substrate specific artifacts, a
set of experiments assessing different substrates, other
phosphatases, and different assay technologies were performed.
BI-0314 is a hSTEP specific positive allosteric modulator and
has been shown to be inactive on two other related tyrosine
phosphatases due to the low conservation of the residues in the
allosteric site. Derivation of the Michaelis−Menten parameters
of hSTEP mediated dephosphorylation reaction of pFYN-
peptide in the presence or absence of BI-0314 showed that the
ligand causes an increase in kcat leading to an enhanced
turnover rate. On the basis of MD simulations, the mechanism
can be explained by a long-range interaction via allosteric
communication between the allosteric binding site and the
active site, where BI-0314 binding results in a significant
EXPERIMENTAL SECTION
■
Compound Purity. Compounds 2, 3, 4, 5, and BI-0314 have been
determined to be >95% pure by HPLC/MS.
Compound Synthesis. Compounds 2−5 are commercially
available.
Synthesis of BI-0314. The synthesis of BI-0314 is shown in
Scheme 1 and is described below.
Tert-butyl 4-[3-nitro-5-(trifluoromethyl)phenyl]- 1,2,3,6-tet-
rahydropyridine-1-carboxylate (S3). 3-Bromo-5-nitrobenzotri-
fluoride (S2) (405 mg, 1.50 mmol) and boronic ester S1 (557 mg,
1.80 mmol, 1.2 equiv) were dissolved in dioxane (18 mL). The
solution was degassed with argon. An aqueous solution of Cs2CO3 (2
M, 3 mL, 6 mmol, 4 equiv) was added, and the reaction mixture was
degassed with argon. The reaction mixture was heated to 100 °C
overnight. It was poured into water (60 mL) and was extracted with
CHCl3 (2 × 30 mL). The combined extracts were washed with
aqueous solution of HCl (0.1 M, 2 × 30 mL) and brine (1 × 30 mL)
and were filtered through a plug of Celite and then through a plug of
Alox. The filtrate was evaporated in vacuo to yield the desired product
S3 (542 mg, 85% purity by HPLC, 1.24 mmol, 83%). HPLC−MS
(method A): tR = 1.28 min. MS: (M + H − CH2C(CH3)2)+ 317.
F
DOI: 10.1021/acs.jmedchem.8b00857
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Tert-butyl 4-[3-amino-5- (trifluoromethyl)phenyl]-
piperidine-1-carboxylate (S4). S3 (542 mg, 85% purity, 1.24
mmol) was dissolved in EtOAc (15 mL) and EtOH (15 mL). RaNi
(500 mg, 8.52 mmol, 7 equiv) and H2 (50 bar) were added. The
reaction was stirred at 50 °C for 1 d. The catalyst was removed by
filtration. The filtrate was evaporated in vacuo. The crude product was
purified by preperative HPLC to yield the desired product S4 (332
mg, 0.964 mmol, 78%). HPLC−MS (method A): tR = 1.14 min. MS:
(M + H − CH2C(CH3)2)+ 289.
Tert-butyl 4-[3-carbamimidamido-5- (trifluoromethyl)-
phenyl]piperidine-1-carboxylate (S5). S4 (200 mg, 0.581
mmol) and formamidine sulfinic acid (144 mg, 1.16 mmol, 2.0
equiv) were dissolved in degassed MeOH (4 mL). The reaction
mixture was stirred at 100 °C for 3 d. ACN and water were added and
the mixture was purified by preparative HPLC to yield the desired
product (29 mg, 0.075 mmol, 13%). HPLC−MS (method A): tR =
0,94 min. MS: (M + H)+ 387.
N-[3-(piperidin-4-yl)-5- (trifluoromethyl)phenyl]guanidine
(BI-0314). S5 (29 mg, 0.075 mmol) was dissolved in DCM (1
mL). TFA (95%, 1 mL) was added at rt. After 2 h the mixture was
evaporated to dryness. The residue was dissolved in an ACN/water
mixture and was purified by preparative HPLC to yield the desired
product BI-0314 as double TFA salt (8.0 mg, 0.016 mmol, 22%).
HPLC−MS (method A): tR = 0.63 min. MS: (M + H)+ 287. 1H NMR
(400 MHz, DMSO-d6) δ 10.15 (s, 1H), 8.70 (br s, 1H), 8.42 (br s,
1H), 7.74 (s, 4H), 7.44−7.47 (m, 2H), 7.38 (dd, J = 1.0 Hz, 1H),
3.40 (d, J = 1.0 Hz, 2H), 2.94−3.06 (m, 3H), 1.99 (br d, J = 13.2 Hz,
2H), 1.81 (qd, J = 13.1, 4.0 Hz, 2H).
Analytical Methods. NMR spectra were recorded on a Bruker
400 MHz instrument. Chemical shifts are given in parts per million
(ppm) downfield from internal reference tetramethylsilane in δ units.
Selected data are reported in the following manner: chemical shift,
multiplicity (s = singlet, d = doublet, dd = double doublet, t = triplet,
q = quartet, m = multiplet), coupling constants (J), integration.
Mass spectra were obtained using a liquid chromatography mass
spectrometer (HPLC−MS) that consisted of Agilent 1200 LC with
DA- and MS-detector.
handler35 to achieve NMR sample tube filling. Transfer of the samples
(140 μL in 2.5 mm NMR tubes) to the magnet was accomplished
with the Bruker Sample Rail system. Fragments were either purchased
from various external vendors or selected from internal proprietary
sources.
Confirmation of primary FBS hits obtained from STD-NMR, DSF,
and MST were performed using two-dimensional 1H,15N TROSY
NMR spectra36 collected on a Bruker Avance III 600 MHz
spectrometer equipped with a 5 mm z-gradient TCI cryoprobe and
a Bruker Sample Rail. Each sample contained 40 μM 15N labeled
STEP PTP in 20 mM Tris-d11, 100 mM NaCl, 1 mM TCEP, pH 7.4,
and 8% (v/v) D2O. The protein was incubated with 500 μM fragment
in a 2.5 mm NMR tube at 298 K and a DMSO-d6 concentration of
1%. Spectra were recorded with 192 transients and 64 data points in
the indirect dimension. Processing and analysis were done with the
Topspin 3.0 software (Bruker BioSpin). Binders were identified by
manual comparison of spectra in the presence of a fragment and the
STEP PTP reference spectrum.
Differential Scanning Fluorimetry (DSF). DSF experiments
were carried out in a Bio-Rad CFX384 real-time system
(C1000Touch Thermal Cycler) in sealed Hard-Shell PCR 384-well
plates (no. HSP3805; PCR Sealers; no. MSB1001; Bio-Rad) and a
total volume of 10 μL. The assay was optimized regarding protein
consumption and SYPRO Orange dye excess (5000× concentration
in DMSO, Invitrogen) to obtain a reliable fluorescence signal.
Compound dilutions in assay buffer (20 mM HEPES, pH 7.5; 50 mM
NaCl; 1 mM TCEP) were prepared as duplicates by a Hamilton
Microlab Starlet pipetting robot, whereas the protein/dye stock
solution was added manually with a multichannel pipet. The reaction
mixture contained 8 μL compound dilution and 2 μL of the STEP46
SYPRO Orange stock mix to yield final concentrations of 5 μM
STEP46, 25× SYPRO Orange and 500 μM fragment (1% DMSO).
Samples were heated at 1 °C/min from 25 to 95 °C with fluorescence
readings every 0.5 °C. Melting curves were analyzed with the Bio-Rad
CFX Manager data analysis software by plotting the first derivative of
the recorded fluorescence intensity versus the temperature where the
minimum of such a curve corresponds to the Tm value. Thermal shifts
of ΔTm ≥ 0.5 °C (ΔTm = Tm,frag − Tm,DMSO) were assumed significant
and defined as primary FBS hits.
Retention times (tR) were obtained using Agilent 1200 liquid
chromatography−mass spectrometer (HPLC−MS), using the follow-
ing method shown in Table 3.
Xray. Protein crystallization was done using the hanging drop
method by mixing 1 μL of hSTEP complexed with 5 mM compound
2 (from a 100 mM DMSO stock solution) with 1 μL of reservoir
solution (24% PEG 3350, 0.2 M ammonium sulfate, and 0.1 M Bis-
Tris, pH 5.5) at 4 °C. Crystals were flash frozen in liquid nitrogen.
Crystallization of mSTEP was also done via the hanging drop method
at 4 °C in a 24-well plate. The reservoir solution contained 26% PEG
3350, 0.1 M Tris, pH 7.8, 0.2 M ammonium citrate. For the drops 1
μL of apo protein was mixed with 1 μL of the reservoir. First crystals
appeared after 8 days. Apo crystals were transferred to a soaking and
cryo buffer containing 25% PEG3350, 0.2 M Bis-Tris, pH 7.5, 0.2 M
disodium malonate and soaked for 4 h in the presence of 100 mM BI-
0314. Crystals were flash frozen in liquid nitrogen, and data were
collected at the SLS beamline X06DA (Swiss Light Source, Paul
Scherrer Institute) at a wavelength of 1 Å using the PILATUS 2M
detector. The crystals belonged to space group P212121 (hSTEP) or
P41212 (mSTEP) and contained one monomer per asymmetric unit.
Images were processed with autoPROC.37 The resolution limits were
set using default autoPROC settings. The structures were solved by
molecular replacement using the STEP structure 2BIJ26 as a search
model. Subsequent model building and refinement was done using
standard protocols using CCP4,38 COOT39 and autoBUSTER.40
The unit cell parameters of the hSTEP−compound 2 complex were
a = 53.1 Å, b = 64.4 Å, c = 100.8 Å and α, β, γ = 90°, and the structure
was refined to Rwork and Rfree values of 20.0% and 23.9%, respectively,
with 95% of the residues in Ramachandran favored regions as
validated with Molprobity.41 The unit cell parameters of the mSTEP−
compound BI-0314 were a = 59.3 Å, b = 59.3 Å, c = 181.6 Å, α, β, γ =
90° and was refined to R/Rfree = 18.1/21.2% with 95% of the residues
in Ramachandran favored regions.
a
Table 3. HPLC−MS Method A
gradient/solvent
time
[min]
% sol
[H2O, 0.1 % TFA]
% sol
[ACN]
flow
temp
[mL/min]
[°C]
0
97
97
0
0
0
3
3
2.2
2.2
2.2
3
60
60
60
60
60
0.2
1.2
1.25
1.4
100
100
100
3
a
Column: Sunfire C18, 3.0 mm × 30 mm, 2.5 μm. Column supplier:
Waters.
NMR Spectroscopy. STD-NMR experiments experiments34 were
carried out on a Bruker Avance II 600 MHz instrument equipped with
a QCI cryogenic probe and z-gradients. Selective irridation for 3 s was
achieved with a Gaussian pulse train at −0.2 ppm for the on-
resonance and 60 ppm for the off-resonance spectrum. A 30 ms spin-
lock pulse was used to suppress protein signals, and both spectra were
subtracted after acquisition to yield the difference spectrum.
Fragments were tested in mixtures of four compounds (each at 250
μM) in 20 mM Tris-d11, 100 mM NaCl, 1 mM TCEP, pH 7.4, in
D2O in the presence of 4 μM STEP46. Signals in the difference
spectrum are indicative for binding, and hits were identified by
comparing to individual prerecorded reference spectra. The entire
fragment library was characterized with respect to purity, identity,
buffer solubility, and self-STD effects. Samples were freshly prepared
just in time with an in-house customized Tecan Freedom Evo liquid
G
DOI: 10.1021/acs.jmedchem.8b00857
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
ITC (Isothermal Titration Calorimetry) Analysis. Experiments
using isothermal titration calorimetry were conducted on a MicroCal
auto iTC200 instrument with hSTEP PTP domain that had been
passed through a PD-10 desalting column (GE Healthcare)
equilibrated with 10 mM HEPES, 200 mM NaCl, and 0.5 mM
TCEP, pH 7.5. Complete saturation of 200 μM of hSTEP was
typically achieved by injecting 18 × 2 μL aliquots of 2 mM BI-0314 at
25 °C. The thermodynamic binding parameters were extracted by
nonlinear regression analysis of the binding isotherms (MicroCal
Origin software, version 7.0). A single-site binding model was applied
yielding the binding enthalpy (ΔH), stoichiometry (n), entropy (ΔS),
and association constant (Ka). Data are mean values from three
independent titrations with calculated standard deviation.
Protein Expression and Purification. The gene corresponding
to amino acid residues 197−565 (STEP46) or 281−563 (STEP PTP
domain) of human STEP (PTN5 human, Uniprot code P54829) was
produced by gene synthesis (optimized for E. coli expression) with an
N-terminal 6 His-tag followed by a TEV protease site and cloned into
pET24a (Novagen) with NdeI and EcoRI. The gene corresponding to
amino acid residues 244−539 of mouse STEP (PTN5 mouse,
Uniprot code P54830) was synthesized and cloned identical to human
STEP.
(AlphaLISA (“high” control) − AlphaLISA (“low” control)).
Optimized assay conditions for the AlphaLISA assay on hSTEP
PTP domain were as above but with 20 nM hSTEP PTP and 3 nM
pFYN peptide.
STEP-RapidFire MS/MS Assay. Nonbiotinylated Fyn-peptide
(BioTrend no. 435305) is diluted in assay buffer (1 M HEPES,
Invitrogen no. 15630-056 (50 mM f.c.); Brij35, Roth no. CN21.1
(0.02% f.c), DTT, Sigma no. D5545 (1 mM f.c.). In a 384-well
Optiplate (PerkinElmer no. 6007290), 5 μL/well of the diluted
substrate, 5 μL/well of STEP46 (50 nM f.c.; 46 kDa fragment of
STEP (STEP46) produced in the laboratory of Dr. Gisela Schnapp,
Boehringer Ingelheim Pharma GmbH & Co. KG) and 5 μL/well of
prediluted compound (500 μM f.c.) are mixed and incubated for 1 h
at room temperature. The reaction was stopped by adding 5 μL/well
of a sodium orthovanadate solution (12 μM f.c.). An amount of 15 μL
of the stopped enzymatic reaction is transferred to a 384-well
polypropylene microplate (Masterblock 22 mm; Greiner no. 781280)
and mixed with 135 μL of Fyn labeled internal standard solution
(GLA[R(13C6; 15N4)]LIEDNEYTARQGAKFPI), Thermo Fisher).
The analytical sample handling was performed by a rapid-injecting
RapidFire autosampler system (Agilent, Waldbronn, Germany)
coupled to a triple quadrupole mass spectrometer (TSQ Vantage,
ThermoFisher, San Jose, CA, USA). Liquid sample is aspirated by a
vacuum pump into a 10 μL sample loop for 250 ms and subsequently
flushed for 3000 ms onto a C18 cartridge (Agilent, Waldbronn,
Germany) with the aqueous mobile phase (99.5% water, 0.5% acetic
acid, flow rate 1.5 mL/min). The solid phase extraction step retains
the analyte while removing interfering matrix (e.g., buffer
components). The analyte is desorbed and back-eluted from the
cartridge for 3000 ms with an organic mobile phase (99.5% methanol,
0.5% acetic acid, flow rate 1.25 mL/min) and flushed into the mass
spectrometer for detection in MRM mode. The MRM transition for
the FYN peptide is 788.48 < 994.27 Da (S-lense 156 V, collision
energy 28 V) and for the internal standard 791.66 < 1073.18 Da (S-
lense 200 V, collision energy 22 V). Dwell time for each MRM
transition is 25 ms, and pause time between MRMs is 5 ms. The mass
spectrometer is equipped with the HESI II ion source and operated in
positive ionization mode (ion spray voltage 2500 V, capillary temp
325 °C, sheath gas pressure 30 Au, vaporizer temp 500 °C, aux gas
pressure 45 Au). While performing the back flush into the mass
spectrometer, the sample loop and relevant tubing were flushed with
the organic mobile phase to prevent carryover of analyte or matrix
components into the next sample. Equilibration time for the system is
500 ms. In order to further minimize carryover effects, the wash
station of the RapidFire system was used to perform needle washes
with pure water (100%) and pure methanol (100%) between samples.
The solvent delivery setup of the RapidFire system consists of three
continuously running and isocratically operating HPLC pumps
(G1310A, Agilent, Waldbronn, Germany). MS data processing was
performed in GMSU (Alpharetta, GA, USA), and peak area ratio FYN
peptide/internal standard was reported for IC50 calculation.
STEP-DiFMUP Assay. Buffer for the DiFMUP assay was the same
as listed previously. DiFMUP (Life Tech no. D6567) stock solutions
were prepared at 10 mM in DMSO and stored at −20 °C. Test
compounds were received as 50 mM stock solutions in DMSO and
stored at −20 °C. Purified human STEP46 was the same as previously
listed.
Compounds were diluted to 20× final test concentration in DMSO.
An amount of 2.5 μL was then pipetted into a 384-well black, flat-
bottom assay plate (Corning no. 3573). Purified human STEP46 was
prediluted in ice cold assay buffer to 210 nM, after which an amount
of 23.5 μL was added to the assay plate and incubated for 10 min at
room temperature. DiFMUP substrate was diluted with buffer to 2
μM, after which the assay was initiated by adding 24 μL to the assay
plate, giving final reaction conditions of 50 mM HEPES, 0.02% Brij
35, 1 mM DTT, 1 μM DiFMUP, 100 nM STEP46, and 5% DMSO.
Fluorescence intensity (excitation 340 nm; emission 460 nm) was
monitored in a Fluoroskan Ascent (PerkinElmer) at 1 min intervals
for 30 min at 37 °C. Background signal, determined by not adding
STEP46 enzyme, was subtracted from assay signals. The linear signal
The expression plasmids for STEP proteins were transformed in
BL21 (DE3) cells (Novagen), and cells were grown in LB containing
30 μM kanamycin at 37 °C to mid log phase, and protein expression
was induced with 1 mM IPTG. The temperature was then reduced,
and cells were grown for further 16 h at 18 °C. Cells were harvested,
resuspended in lysis buffer (50 mM HEPES, 500 mM NaCl, 0.5 mM
TCEP, and protease inhibitors (complete EDTA free, Roche
Diagnostics), pH 7.5, and lysed by the Constant Cell Disruption
System. The fusion protein was purified using Ni-NTA (Qiagen)
affinity and eluted with buffer containing 250 mM imidazole. STEP
protein was further purified by gel filtration using Superdex 200 HR
column (GE Healthcare) equilibrated in 10 mM HEPES, pH 7.5, 200
mM NaCl, and 0.5 mM TCEP. The protein was concentrated to 10
mg/mL for crystallization and to 5 mg/mL for NMR studies. For 15N
or 2H,15N labeling STEP was expressed in Silantes E. coli OD2 N, 15
N
medium or in Silantes E. coli OD2 DN, 2H, 15N medium, respectively.
STEP-pFYN-peptide AlphaLISA Assay. Compounds are serial
diluted in 100% DMSO in a 384-well PP microplate V-shape (Greiner
no. 781280) using an Agilent Bravo pipetting robot. With an
intermediate dilution step in AlphaLISA assay buffer (PerkinElmer no.
AL000F) in a 384-well clear PS microplate (Greiner no. 781101), the
DMSO concentration is reduced to achieve 1% DMSO in the final
assay reaction. The assay in a 384-well Optiplate (PerkinElmer no.
6007290) starts by addition of 5 μL of diluted compound, 5 μL of
STEP46 (50 nM f.c.; 46 kDa fragment of STEP (STEP46) produced
in the laboratory of Dr. Gisela Schnapp, Boehringer Ingelheim
Pharma GmbH & Co. KG) and 5 μL pFyn peptide (20 nM f.c.;
(biotin-GLARLIEDNE-Y(H2PO3)-TARQGAKFPI-OH), purity
>95%, custom synthesized, and purchased via Biotrend), followed
by short centrifugation. The assay mixture is incubated at room
temperature for 60 min in the dark. After incubation, 5 μL/well of anti
non-phospho Src/Tyr416 antibody (Cell Signaling Technology no.
2102), diluted in enzyme stop buffer (AlphaLISA ImmunoAssay
buffer, supplemented with 10 μM sodium orthovanadate) is added to
each well (final antibody dilution 1:26 000 in 15 μL of detection
volume). Next, 5 μL/well of protein A-coated ALPHA acceptor beads
(PerkinElmer no. 6760617M) in enzyme stop buffer is added to each
well. The assay mixture is incubated for 15 min at room temperature
covered with a plate lid. After incubation, 5 μL/well of streptavidin-
coated ALPHA donor beads in enzyme stop buffer is added to each
well. The assay mixture is incubated for 15 min at room temperature
with a plate lid. Finally, the signal is measured using an Envision
reader (PerkinElmer). Each assay plate contains wells with “high” and
“low” controls, where “high” controls (100% CTL) are defined as 50
nM STEP46 incubated with 20 nM pFyn peptide and “low” controls
(0% CTL) are defined as 20 nM pFyn peptide without STEP46. The
“percent control” (% CTL) values are calculated as follows: % CTL =
100 × (AlphaLISA (sample) − AlphaLISA (“low” control))/
H
DOI: 10.1021/acs.jmedchem.8b00857
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
increase (Δsignal increase/min) for STEP46 in the absence of
compound or at different compound concentrations was calculated,
and STEP46 in the absence of compound was set to 100%.
further analysis because the ligand rmsd was <3.5 Å throughout the
MD, corresponding to proper binding. The simulations were repeated
with deprotonated Cys496, resulting in basically identical results for
the B-factor analysis. Twenty trajectories each, from the apo
simulation and from the liganded simulations, were subjected to the
Allosteer30 analysis to identify the allosteric pathways.
PTP1B, TC-PTP-insulin Receptor Derived Substrate and
PTP1B pFYN Substrate Assays. For all phosphatase assays the
AlphaScreen technology was employed. Assays were set up as
competition assays generating a high assay signal using biotin-Ahx-
EDT-pY-Y-RKG-amide and anti-phosphotyrosine antibody 4G10.
Unbiotinylated substrate is competing with the antibody binding and
therefore reducing the AlphaScreen signal. For every phosphatase
assay white 384-well microplates Lumitrac 200 (Greiner, Friggenhau-
sen, Germany, catalog no. 781075) were employed. PTP1B (1−405
AA) and TC-PTP (2−345AA) were produced in house (Boehringer-
Ingelheim Pharma GmbH & Co. KG). The substrates for PTP1B and
TC-PTP EDT-pY-Y-RKG-amide and biotin-Ahx-EDT-pY-Y-RKG-
amide were purchased from Biosyntan (Berlin, Germany), whereas
pFYN for the PTP1B pFYN assay was purchased from Biotrend
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.jmed-
chem.8b00857.
■
S
15N TROSY NMR spectra of 2 with human and mouse
STEP PTP; crystal contacts in mouse and human STEP
close to the allosteric binding site; 15N TROSY NMR
spectrum of BI-0314 on human STEP and determi-
nation of the KD; ITC measurements of the binding
affinity of BI-0314 toward hSTEP; activation of mouse
STEP PTP domain by BI-0314; enzyme progress curves
for human STEP PTP domain; allosteric binding sites
and allosteric pipelines; comparison of mSTEP, apo vs
bound to BI-0314; calculated B-factors upon binding of
BI-0314; DSF spectrum of BI-0314 binding to human
STEP; allosteric pipelines and comparison to PTP1B; X-
ray statistics; NMR chemical shift perturbation; other
comments on hit finding; compound characterization;
1H NMR of BI-0314; 13C NMR of BI-0314 (PDF)
Molecular formula strings (CSV)
̈
(Koln, Germany, catalog no. 435305). Anti-phosphotyrosine antibody
clone 4G10 was purchased at Merck Millipore (Darmstadt, Germany,
catalog no. 05-321). Orthovanadate was bought at Biolabs (Ipswhich,
USA, catalog no. P0758). The AlphaScreen IgG detection kit was
obtained from PerkinElmer (Rodgau, Germany, catalog no.
6760617R). All other materials were of highest grade commercially
available.
An amount of 100 nL of a test compound in various concentrations
up to 500 μM was added into a microtiter plate, followed by the
addition of 5 μL of a phosphatase (final concentration 500 pM). After
an incubation time of 10 min at 24 °C 5 μL of the EDT-pY-Y-RKG-
amide substrate (final concentration of 1.4 μM) or pFYN substrate
(final concentration of 20 μM) were added. Plates were then
incubated in a humidified incubator at 24 °C for 1.5 h. To stop the
enzymatic reaction, 5 μL of orthovanadate (final concentration 200
μM) was added. Each microtiter plate contained low controls (buffer
instead of enzyme) and high controls (enzyme). Furthermore, each
microtiter plate contained a standard curve using EDT-pY-Y-RKG-
amide in the concentration range 30 μM/3 μM/1 μM/300 nM/100
nM/30 nM/10 nM/1 nM or pFYN in the range of 300 μM/100 μM/
30 μM/3 μM/1 μM/300 nM/100 nM/30 nM.
Accession Codes
Atomic coordinates of human STEP in complex with 2 (PDB
code 6H8R) and of mouse STEP in complex with BI-0314
(PDB code 6H8S) have been deposited with the Protein Data
Bank.
5 μL of biotin-Ahx-EDT-pY-Y-RKG-amide and streptavidin donor
bead mix (final concentration of 2.5 nM substrate and 5 μg/mL donor
beads) were added, followed by the addition of 5 μL acceptor bead
and antibody mix (final concentration of 5 μg/mL acceptor bead and
500 pM antibody). After an incubation time of 2 h at 24 °C the
AlphaScreen signal was measured with an Envision reader
(PerkinElmer). Data were analyzed by using the standard curve on
each plate.
Computational Methods. Protein Preparation. hSTEP PTP
domain X-ray structure bound to 2 was superposed with PDB code
2BV5 to model the missing loop (manual transfer) and the acetylation
of the active site Cys496 was removed. 2 was replaced by BI-0314 by
superposing mouse and human STEP structures. The resulting
structure was treated by the protein preparation procedure including
protonation by protonate3D as implemented in MOE,42 and a
stepwise energy minimization was done employing the MMFF94x
force field (position restraints 0.0, 1.0 on heavy atoms, followed by 1.0
on the backbone). The protonation state of the active site cysteine
Cys496 was manually adjusted.
MD Simulations. Molecular dynamics simulations were carried out
using the AMBER 14 package43 employing the ff14SB force field and
the TIP3P water model. The AM1-BCC method was used to obtain
partial charges for the ligands with the antechamber tool. A truncated
octahedral solvation bounding box of 90 Å and periodic boundary
conditions with the Ewald approach were employed. Equilibration of
the system was done by a stepwise schedule, starting with heating
from 100 to 300 K in the NVT ensemble, transitioning to the NPT
ensemble at 300 K with stepwise lowering of the constraints on the
protein heavy atoms. For the production runs 40 MD simulations of
apo STEP and 60 runs of BI-0314 bound STEP (each 100 ns),
employing a time step of 2 fs with the SHAKE algorithm, were
performed. 44 trajectories of BI-0314 bound STEP qualified for
AUTHOR INFORMATION
Corresponding Author
*E-mail: christofer.tautermann@boehringer-ingelheim.com.
ORCID
Christofer S. Tautermann: 0000-0002-6935-6940
Nagarajan Vaidehi: 0000-0001-8100-8132
Notes
■
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank Anita Bloching, Heidi Bosch, Jessica Bretzel, Julian
Friedl, and Franziska Moser for excellent technical assistance;
Stephanie Merkle, Katja Philipp, and Achim Bixenmann for
compound syntheses; Kathrin Feuerstein for compound
analytics; Dominik Frey and Achim Lietz for compound
profiling (biochemical assays); and Markus Holstein for setting
up and performing RapidFire MS/MS analytics.
ABBREVIATIONS USED
■
PTPN5, protein tyrosine phosphatase non-receptor type 5;
STEP, striatal-enriched protein tyrosine phosphatase;
NMDAR, N-methyl-D-aspartate receptor; AMPAR, α-amino-
3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor; AD,
Alzheimer’s disease; KIM, kinase interaction motif; PTP,
protein tyrosine phosphatase; STD NMR, saturation transfer
difference NMR; DSF, differential scanning fluorometry; MST,
microscale thermophoresis; CSP, chemical shift perturbation;
ITC, isothermal titration calorimetry
I
DOI: 10.1021/acs.jmedchem.8b00857
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
(15) Rankovic, Z. Cns Drug Design: Balancing Physicochemical
Properties for Optimal Brain Exposure. J. Med. Chem. 2015, 58,
2584−2608.
REFERENCES
■
(1) Goebel-Goody, S. M.; Baum, M.; Paspalas, C. D.; Fernandez, S.
M.; Carty, N. C.; Kurup, P.; Lombroso, P. J. Therapeutic Implications
for Striatal-Enriched Protein Tyrosine Phosphatase (STEP) in
Neuropsychiatric Disorders. Pharmacol. Rev. 2012, 64, 65−87.
(2) Kurup, P.; Zhang, Y.; Xu, J.; Venkitaramani, D. V.; Haroutunian,
V.; Greengard, P.; Nairn, A. C.; Lombroso, P. J. Aβ-Mediated NMDA
Receptor Endocytosis in Alzheimer’s Disease Involves Ubiquitination
of the Tyrosine Phosphatase STEP61. J. Neurosci. 2010, 30, 5948−
5957.
(3) Zhang, Y.; Kurup, P.; Xu, J.; Carty, N.; Fernandez, S. M.;
Nygaard, H. B.; Pittenger, C.; Greengard, P.; Strittmatter, S. M.;
Nairn, A. C.; Lombroso, P. J. Genetic Reduction of Striatal-Enriched
Tyrosine Phosphatase (STEP) Reverses Cognitive and Cellular
Deficits in an Alzheimer’s Disease Mouse Model. Proc. Natl. Acad.
Sci. U. S. A. 2010, 107, 19014−19019.
(4) Kurup, P. K.; Xu, J.; Videira, R. A.; Ononenyi, C.; Baltazar, G.;
Lombroso, P. J.; Nairn, A. C. STEP61 is a Substrate of the E3 Ligase
Parkin and is Upregulated in Parkinson’s Disease. Proc. Natl. Acad. Sci.
U. S. A. 2015, 112, 1202−1207.
(5) Carty, N. C.; Xu, J.; Kurup, P.; Brouillette, J.; Goebel-Goody, S.
M.; Austin, D. R.; Yuan, P.; Chen, G.; Correa, P. R.; Haroutunian, V.;
Pittenger, C.; Lombroso, P. J. The Tyrosine Phosphatase STEP:
Implications in Schizophrenia and the Molecular Mechanism
Underlying Antipsychotic Medications. Transl. Psychiatry 2012, 2,
e137.
(6) Goebel-Goody, S. M.; Wilson-Wallis, E. D.; Royston, S.;
Tagliatela, S. M.; Naegele, J. R.; Lombroso, P. J. Genetic
Manipulation of STEP Reverses Behavioral Abnormalities in a Fragile
X Syndrome Mouse Model. Genes, Brain and Behavior 2012, 11, 586−
600.
(7) Reinhart, V. L.; Nguyen, T.; Gerwien, R.; Kuhn, M.; Yates, P. D.;
Lanz, T. A. Downstream Effects of Striatal-Enriched Protein Tyrosine
Phosphatase Reduction on RNA Expression in Vivo and in Vitro.
Neuroscience 2014, 278, 62−69.
(8) Xu, J.; Chatterjee, M.; Baguley, T. D.; Brouillette, J.; Kurup, P.;
Ghosh, D.; Kanyo, J.; Zhang, Y.; Seyb, K.; Ononenyi, C.; Foscue, E.;
Anderson, G. M.; Gresack, J.; Cuny, G. D.; Glicksman, M. A.;
Greengard, P.; Lam, T. T.; Tautz, L.; Nairn, A. C.; Ellman, J. A.;
Lombroso, P. J. Inhibitor of the Tyrosine Phosphatase STEP Reverses
Cognitive Deficits in a Mouse Model of Alzheimer’s Disease. PLoS
Biol. 2014, 12, e1001923.
(9) Venkitaramani, D. V.; Moura, P. J.; Picciotto, M. R.; Lombroso,
P. J. Striatal-Enriched Protein Tyrosine Phosphatase (STEP)
Knockout Mice Have Enhanced Hippocampal Memory. European
Journal of Neuroscience 2011, 33, 2288−2298.
(16) He, R.; Zeng, L.-F.; He, Y.; Zhang, S.; Zhang, Z.-Y. Small
Molecule Tools for Functional Interrogation of Protein Tyrosine
Phosphatases. FEBS J. 2013, 280, 731−750.
(17) Chen, Y.-N. P.; LaMarche, M. J.; Chan, H. M.; Fekkes, P.;
Garcia-Fortanet, J.; Acker, M. G.; Antonakos, B.; Chen, C. H.-T.;
Chen, Z.; Cooke, V. G.; Dobson, J. R.; Deng, Z.; Fei, F.; Firestone, B.;
Fodor, M.; Fridrich, C.; Gao, H.; Grunenfelder, D.; Hao, H.-X.; Jacob,
J.; Ho, S.; Hsiao, K.; Kang, Z. B.; Karki, R.; Kato, M.; Larrow, J.; La
Bonte, L. R.; Lenoir, F.; Liu, G.; Liu, S.; Majumdar, D.; Meyer, M. J.;
Palermo, M.; Perez, L.; Pu, M.; Price, E.; Quinn, C.; Shakya, S.;
Shultz, M. D.; Slisz, J.; Venkatesan, K.; Wang, P.; Warmuth, M.;
Williams, S.; Yang, G.; Yuan, J.; Zhang, J.-H.; Zhu, P.; Ramsey, T.;
Keen, N. J.; Sellers, W. R.; Stams, T.; Fortin, P. D. Allosteric
Inhibition of SHP2 Phosphatase Inhibits Cancers Driven by Receptor
Tyrosine Kinases. Nature 2016, 535, 148.
(18) Krishnan, N.; Koveal, D.; Miller, D. H.; Xue, B.; Akshinthala, S.
D.; Kragelj, J.; Jensen, M. R.; Gauss, C.-M.; Page, R.; Blackledge, M.;
Muthuswamy, S. K.; Peti, W.; Tonks, N. K. Targeting the Disordered
C Terminus of PTP1B with an Allosteric Inhibitor. Nat. Chem. Biol.
2014, 10, 558−566.
(19) Wiesmann, C.; Barr, K. J.; Kung, J.; Zhu, J.; Erlanson, D. A.;
Shen, W.; Fahr, B. J.; Zhong, M.; Taylor, L.; Randal, M.; McDowell,
R. S.; Hansen, S. K. Allosteric Inhibition of Protein Tyrosine
Phosphatase 1B. Nat. Struct. Mol. Biol. 2004, 11, 730−737.
(20) Sharma, E.; Zhao, F.; Bult, A.; Lombroso, P. J. Identification of
Two Alternatively Spliced Transcripts of Step: A Subfamily of Brain-
Enriched Protein Tyrosine Phosphatases. Mol. Brain Res. 1995, 32,
87−93.
(21) Li, R.; Xie, D.-D.; Dong, J.-h.; Li, H.; Li, K.-s.; Su, J.; Chen, L.-
Z.; Xu, Y.-F.; Wang, H.-M.; Gong, Z.; Cui, G.-Y.; Yu, X.; Wang, K.;
Yao, W.; Xin, T.; Li, M.-Y.; Xiao, K.-H.; An, X.-f.; Huo, Y.; Xu, Z.-g.;
Sun, J.-P.; Pang, Q. Molecular Mechanism of ERK Dephosphorylation
by Striatal-Enriched Protein Tyrosine Phosphatase. J. Neurochem.
2014, 128, 315−329.
(22) Barr, A. J.; Ugochukwu, E.; Lee, W. H.; King, O. N. F.;
Filippakopoulos, P.; Alfano, I.; Savitsky, P.; Burgess-Brown, N. A.;
Muller, S.; Knapp, S. Large-Scale Structural Analysis of the Classical
̈
Human Protein Tyrosine Phosphatome. Cell 2009, 136, 352−363.
(23) Erlanson, D. A.; Fesik, S. W.; Hubbard, R. E.; Jahnke, W.; Jhoti,
H. Twenty Years On: The Impact of Fragments on Drug Discovery.
Nat. Rev. Drug Discovery 2016, 15, 605−619.
́
(24) Jahnke, W.; Rondeau, J.-M.; Cotesta, S.; Marzinzik, A.; Pelle,
̈
X.; Geiser, M.; Strauss, A.; Gotte, M.; Bitsch, F.; Hemmig, R.; Henry,
C.; Lehmann, S.; Glickman, J. F.; Roddy, T. P.; Stout, S. J.; Green, J.
R. Allosteric Non-Bisphosphonate FPPS Inhibitors Identified by
Fragment-Based Discovery. Nat. Chem. Biol. 2010, 6, 660−666.
(25) Cox, A. D.; Fesik, S. W.; Kimmelman, A. C.; Luo, J.; Der, C. J.
Drugging the Undruggable RAS: Mission Possible? Nat. Rev. Drug
Discovery 2014, 13, 828−851.
(10) Zhang, Y.; Kurup, P.; Xu, J.; Anderson, G. M.; Greengard, P.;
Nairn, A. C.; Lombroso, P. J. Reduced Levels of the Tyrosine
Phosphatase STEP Block Beta Amyloid-Mediated GluA1/GluA2
Receptor Internalization. J. Neurochem. 2011, 119, 664−672.
(11) Olausson, P.; Venkitaramani, D. V.; Moran, T. D.; Salter, M.
W.; Taylor, J. R.; Lombroso, P. J. The Tyrosine Phosphatase STEP
Constrains Amygdala-Dependent Memory Formation and Neuro-
plasticity. Neuroscience 2012, 225, 1−8.
(26) Eswaran, J.; von Kries, J. P.; Marsden, B.; Longman, E.;
́
Debreczeni, J. E.; Ugochukwu, E.; Turnbull, A.; Lee, W. H.; Knapp,
S.; Barr, A. J. Crystal Structures and Inhibitor Identification for
PTPN5, PTPRR and PTPN7: A Family of Human MAPK-Specific
Protein Tyrosine Phosphatases. Biochem. J. 2006, 395, 483−491.
(27) Francis, D. M.; Page, R.; Peti, W. Sequence-Specific Backbone
1H, 13C and 15N Assignments of the 34 kDa Catalytic Domain of
PTPN5 (STEP). Biomol. NMR Assignments 2014, 8, 185−188.
(28) Klebe, G. Applying Thermodynamic Profiling in Lead Finding
and Optimization. Nat. Rev. Drug Discovery 2015, 14, 95−110.
(12) Xu, J.; Kurup, P.; Baguley, T. D.; Foscue, E.; Ellman, J. A.;
Nairn, A. C.; Lombroso, P. J. Inhibition of the Tyrosine Phosphatase
STEP61 Restores BDNF Expression and Reverses Motor and
Cognitive Deficits in Phencyclidine-Treated Mice. Cell. Mol. Life Sci.
2016, 73, 1503−1514.
(13) Kulikova, E. A.; Volcho, K. P.; Salakhutdinov, N. F.; Kulikov, A.
V. Benzopentathiepine Derivative, 8-(Trifluoromethyl)-1,2,3,4,5-
Benzopentathiepin-6-Amine Hydrochloride (TC-2153), as a Promis-
ing Antidepressant of New Generation. Lett. Drug Des. Discovery 2017,
14, 974−984.
(29) Welte, S.; Baringhaus, K.-H.; Schmider, W.; Muller, G.; Petry,
̈
S.; Tennagels, N. 6,8-Difluoro-4-Methylumbiliferyl Phosphate: A
Fluorogenic Substrate for Protein Tyrosine Phosphatases. Anal.
Biochem. 2005, 338, 32−38.
(30) Bhattacharya, S.; Vaidehi, N. Differences in Allosteric
Communication Pipelines in the Inactive and Active States of a
GPCR. Biophys. J. 2014, 107, 422−434.
(14) Witten, M. R.; Wissler, L.; Snow, M.; Geschwindner, S.; Read,
̈
J. A.; Brandon, N. J.; Nairn, A. C.; Lombroso, P. J.; Kack, H.; Ellman,
J. A. X-Ray Characterization and Structure-Based Optimization of
Striatal-Enriched Protein Tyrosine Phosphatase Inhibitors. J. Med.
Chem. 2017, 60, 9299−9319.
J
DOI: 10.1021/acs.jmedchem.8b00857
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
(31) Brandao, T. A. S.; Hengge, A. C.; Johnson, S. J. Insights into the
̃
Reaction of Protein-Tyrosine Phosphatase 1B: Crystal Structures for
Transition State Analogs of Both Catalytic Steps. J. Biol. Chem. 2010,
285, 15874−15883.
(32) Choy, M. S.; Li, Y.; Machado, L. E. S. F.; Kunze, M. B. A.;
Connors, C. R.; Wei, X.; Lindorff-Larsen, K.; Page, R.; Peti, W.
Conformational Rigidity and Protein Dynamics at Distinct Timescales
Regulate PTP1B Activity and Allostery. Mol. Cell 2017, 65, 644−658.
(33) Lipchock, J. M.; Hendrickson, H. P.; Douglas, B. B.; Bird, K. E.;
Ginther, P. S.; Rivalta, I.; Ten, N. S.; Batista, V. S.; Loria, J. P.
Characterization of Protein Tyrosine Phosphatase 1B Inhibition by
Chlorogenic Acid and Cichoric Acid. Biochemistry 2017, 56, 96−106.
(34) Mayer, M.; Meyer, B. Characterization of Ligand Binding by
Saturation Transfer Difference NMR Spectroscopy. Angew. Chem., Int.
Ed. 1999, 38, 1784−1788.
(35) Ross, A.; Senn, H. Automation of Biomolecular NMR
Screening. Curr. Top. Med. Chem. 2003, 3, 55−67.
(36) Pervushin, K. V.; Wider, G.; Wuthrich, K. Single Transition-to-
̈
Single Transition Polarization Transfer (ST2-PT) in [15N,1H]-
TROSY. J. Biomol. NMR 1998, 12, 345−348.
(37) Vonrhein, C.; Flensburg, C.; Keller, P.; Sharff, A.; Smart, O.;
Paciorek, W.; Womack, T.; Bricogne, G. Data Processing and Analysis
with the Autoproc Toolbox. Acta Crystallogr., Sect. D: Biol. Crystallogr.
2011, 67, 293−302.
(38) Collaborative Computational Project, Number 4.. The CCP4
Suite: Programs for Protein Crystallography. Acta Crystallogr., Sect. D:
Biol. Crystallogr. 1994, 50, 760−763.
(39) Emsley, P.; Lohkamp, B.; Scott, W. G.; Cowtan, K. Features
and Development of Coot. Acta Crystallogr., Sect. D: Biol. Crystallogr.
2010, 66, 486−501.
(40) Bricogne, G.; Blanc, E.; Brandl, M.; Flensburg, C.; Keller, P.;
Paciorek, W.; Roversi, P.; Sharff, A.; Smart, O. S.; Vonrhein, C.;
Womack, T. O. Buster, version 2.11.2; Global Phasing Ltd., 2017.
(41) Chen, V. B.; Arendall, W. B., III; Headd, J. J.; Keedy, D. A.;
Immormino, R. M.; Kapral, G. J.; Murray, L. W.; Richardson, J. S.;
Richardson, D. C. Molprobity: All-Atom Structure Validation for
Macromolecular Crystallography. Acta Crystallogr., Sect. D: Biol.
Crystallogr. 2010, 66, 12−21.
(42) Molecular Operating Environment (MOE), version 2016.08;
Chemical Computing Group, 2017.
(43) Case, D. A.; Babin, V.; Berryman, J. T.; Betz, R. M.; Cai, Q.;
Cerutti, D. S.; Cheatham, T. E.; Darden, T. A.; Duke, R. E.; Gohlke,
H.; Goetz, A. W.; Gusarov, S.; Homeyer, N.; Janowski, P.; Kaus, J.;
́
Kolossvary, I.; Kovalenko, A.; Lee, T. S.; LeGrand, S.; Luchko, T.;
Luo, R.; Madej, B.; Merz, K. M.; Paesani, F.; Roe, D. R.; Roitberg, A.;
Sagui, C.; Salomon-Ferrer, R.; Seabra, G.; Simmerling, C. L.; Smith,
W.; Swails, J.; Walker, R. C.; Wang, J.; Wolf, R. M.; Wu, X.; Kollman,
P. A. AMBER 14; University of California, San Francisco, 2014.
K
DOI: 10.1021/acs.jmedchem.8b00857
J. Med. Chem. XXXX, XXX, XXX−XXX