Small molecule receptor protein tyrosine phosphatase γ (RPTPγ) ligands that inhibit phosphatase activity via perturbation of the tryptophan-proline-aspartate (WPD) loop

ARTICLE

pubs.acs.org/jmc

Small Molecule Receptor Protein Tyrosine Phosphatase γ (RPTPγ)

Ligands That Inhibit Phosphatase Activity via Perturbation of the

TryptophanꢀProlineꢀAspartate (WPD) Loop^†

Steven Sheriﬀ,*^,‡Brett R. Beno,^§Weixu Zhai,^§Walter A. Kostich,^§Patricia A. McDonnell,^‡Kevin Kish,^‡

Valentina Goldfarb,^‡Mian Gao,^‡Susan E. Kiefer,^‡Joseph Yanchunas,^‡Yanling Huang,^§Shuhao Shi,^§

Shirong Zhu,^§Carolyn Dzierba,^§Joanne Bronson,^§John E. Macor,^§Kingsley K. Appiah,^§

Ryan S. Westphal,^§Jonathan O’Connell,^§and Samuel W. Gerritz^§

^‡Bristol-Myers Squibb Research and Development, P.O. Box 4000, Princeton, New Jersey 08543-4000, United States

^§Bristol-Myers Squibb Research and Development, 5 Research Parkway, Wallingford, Connecticut 06492, United States

S Supporting Information

b

ABSTRACT: Protein tyrosine phosphatases (PTPs) catalyze the dephosphorylation of

tyrosine residues, a process that involves a conserved tryptophanꢀprolineꢀaspartate

(WPD) loop in catalysis. In previously determined structures of PTPs, the WPD-loop has

been observed in either an “open” conformation or a “closed” conformation. In the

current work, X-ray structures of the catalytic domain of receptor-like protein tyrosine

phosphatase γ (RPTPγ) revealed a ligand-induced “superopen” conformation not

previously reported for PTPs. In the superopen conformation, the ligand acts as an

apparent competitive inhibitor and binds in a small hydrophobic pocket adjacent to, but

distinct from, the active site. In the open and closed WPD-loop conformations of RPTPγ,

the side chain of Trp1026 partially occupies this pocket. In the superopen conformation,

Trp1026 is displaced allowing a 3,4-dichlorobenzyl substituent to occupy this site. The bound ligand prevents closure of the WPD-

loop over the active site and disrupts the catalytic cycle of the enzyme.

’ INTRODUCTION

tail suspension model is consistent with an antidepressant-like

activity. These observations suggest that RPTPγ may be a worth-

while target for drug discovery eﬀorts.

Protein tyrosine phosphatases (PTPs), along with protein

tyrosine kinases (PTKs), play a central role in inter- and

intracellular processes via phosphorylation (PTKs) and depho-

sphorylation (PTPs) of tyrosine residues in proteins.^1ꢀ8PTPs

have been implicated in a number of human diseases including

various forms of cancer, diabetes, multiple sclerosis, autism, and

others.^1ꢀ5,7ꢀ10Therefore, considerable interest exists in PTPs as

targets for pharmaceutical agents. Eﬀorts to identify and/or

design molecules that modulate the functions of various PTPs

have been actively pursued.^{3,4,7,10ꢀ14}

RPTPγ is a member of the receptor-like (RPTP) class of

protein tyrosine phosphatases. It is a type 1 integral membrane

protein comprising an extracellular receptor-like domain, a

region containing a catalytic PTP domain proximal to the

membrane, and a catalytically inactive PTP-fold domain distal

to the membrane.¹⁵Although RPTPγ has not been studied

extensively, literature reports suggest a connection between

RPTPγ and certain types of cancer.^16ꢀ20In addition, disruption

of the RPTPγ gene with a β-galactosidase reporter gene revealed

that this tyrosine phosphatase is predominantly expressed in

pyramidal and sensory neurons and may serve as a marker for

these neurons.²¹In this report, subtle behavioral changes were

noted including enhanced prepulse inhibition and longer latency

to ﬁrst immobilization in tail suspension.²¹The phenotype in the

The WPD-loop (Trp-Pro-Asp), which is one of the character-

istic structural motifs found in PTPs, contains an aspartic acid

residue that plays a central role in the dephosphorylation

mechanism.²²In PTPs, the WPD-loop adopts distinctly dif-

ferent conformations in the presence or absence of bound

peptide or small molecule ligands [cf. PTP1B X-ray crystal

structures 2HNP²³and 1PTY²⁴in the Protein Data Bank

(PDB)].²⁵With no ligand bound in the active site, the WPD-

loop in PTPs typically adopts an “open” conformation as

exempliﬁed by 2HNP. In comparison, ligand-bound PTPs

usually assume a “closed” WPD-loop conformation as in struc-

ture 1PTY. Although the open and closed conformations are

perceived to be the relevant ones in the catalytic cycle, other

conformations of the WPD-loop have been reported, e.g., an

intermediate²⁶between the open and closed conformations and

“atypical” conformations,²⁷where the WPD-loop in the open

position is shifted laterally.

The “closed” or active conformation of PTPs may be an

unsuitable target for drug development because of diﬃculties

Received: March 31, 2011

Published: September 01, 2011

r

2011 American Chemical Society

6548

dx.doi.org/10.1021/jm2003766 J. Med. Chem. 2011, 54, 6548–6562

|

Journal of Medicinal Chemistry

ARTICLE

in both dimensions consistent with a folded protein (Figure 2a).

The single isolated domain 1 protein had activity in the biological

assay comparable to that of the HTS cytoplasmic protein (data

not shown). This suggested that the phosphatase activity of

RPTPγ resides in domain 1, in accord with the phosphatase

literature.^6,8

The optimized domain 1 protein was a valuable tool for

compound binding, structure determination, and activity mea-

surements. Initial direct binding experiments consisted of NMR

protein-detected experiments (transverse relaxation optimized

spectroscopy, TROSY), NMR ligand-detected experiments (T₂

relaxation), and a thermal shift assay.³⁰Subsequent direct bind-

ing experiments used only the NMR protein-detected experi-

ment. Out of ∼30 HTS hits, only compound 1 exhibited direct

binding using biophysical measurements. Compound 1 caused

chemical shift perturbations in the ¹H/¹⁵N TROSY experiment

of the domain 1 RPTPγ protein (Figure 2a) and showed a drastic

reduction in its T₂relaxation curve (data not shown). In addition,

the melting temperature (T_m) of the protein increased from 39.4

to 40.8 °C in the presence of 1 (100 μM) and the ΔT_mtitrated

with increasing concentrations of compound (Figure 2b). Since

both NMR and thermal shift experiments indicated that 1 bound

to RPTPγ, we employed kinetic studies to further elucidate the

mechanism of inhibition.

RPTPγ/Compound 1 Kinetic Studies. The mechanism of

RPTPγ inhibition by 1 was characterized using a FRET-based

peptide dephosphorylation assay (Z⁰-Lyte system). This assay

used the full RPTPγ cytoplasmic domain protein (used for the

HTS screen) rather than the shorter catalytic domain protein

that was used for NMR, thermal shift assay, and crystallography

efforts. The kinetic parameters (K_m, V_max) of both proteins were

very similar, suggesting that the absence of the C-terminal

protein sequence had minimal effect on the enzymatic activity

of the truncated catalytic domain protein.

Inhibition of RPTPγ by 1 exhibited modest time dependence.

A ∼5-fold potency increase was observed with a 20 min

compound/enzyme preincubation (IC₅₀= 2.6 ( 0.1 μM)

relative to controls with no preincubation (IC₅₀= 12.8 (

0.2 μM; see Figure 3a). Extending the preincubation to 40 min

did not further improve the potency (IC₅₀= 2.5 ( 0.1 μM).

The reversibility of RPTPγ inhibition by 1 was determined

through comparison of progress curves from RPTPγ pretreated

with inhibitor (followed by 12-fold dilution to reduce inhibitor

concentration) vs pretreatment with vehicle. The progress curves

under both conditions were essentially identical, suggesting that

RPTPγ inhibition by 1 is readily reversible (Figure 3b).

The mode by which 1 inhibits RPTPγ was determined by

measuring initial velocity as a function of substrate concentration

at several ﬁxed concentrations of inhibitor. The data were ﬁtted

to the MichealisꢀMenton equation and for graphical purposes

were depicted in a LineweaverꢀBurke plot (Figure 4a).³¹Com-

pound 1 increased the apparent K_mvalue without altering the

apparent V_max, consistent with a competitive mode of action. The

LineweaverꢀBurke plot shows the expected characteristic inter-

secting Y-intercepts. The K_ifrom a plot of apparent K_mvs [1] is

2.5 ( 0.6 μM (Figure 4b).

Figure 1. Chemical structure and measured IC₅₀values for HTS hit

compound 1.

arising from (1) the high degree of identity in the active site

among PTPs and (2) the limited ability of ligands bound to the

active site to access adjacent sites that do diﬀer.

We report the discovery, biophysical characterization, and

optimization of a series of small molecule RPTPγ inhibitors that

bind adjacent to the active site and perturb the WPD-loop

(Trp1026-Pro1027-Asp1028) into a novel “superopen” confor-

mation via displacement of Trp1026 from a small hydrophobic

pocket. This pocket is partially occupied by Trp1026 in the open

and closed enzyme conformations, while substituted aromatic

moieties of bound ligands ﬁll the pocket in the superopen

conformation. Furthermore, movement of the WPD-loop into

this superopen conformation modiﬁes the shape and electro-

static characteristics of the active site. This superopen WPD-loop

conformation and associated binding pocket are novel and have

not been reported for other PTPs. The unique movement of the

WPD-loop may be exploited to design and/or identify com-

pounds that selectively interact with RPTPγ relative to other

PTPs. Typically, PTP active site inhibitors require acidic moieties

to mimic the substrate phosphate and provide adequate

aﬃnity.^{3,7,10,11,13,14}It may be possible to design RPTPγ ligands

with more desirable physicochemical properties by targeting the

binding pocket that is exposed by the movement of the WPD-

loop.

’ RESULTS

We identiﬁed the initial hit, compound 1 (Figure 1), in a high

throughput screen (HTS) that is described elsewhere.²⁸Subse-

quently, we characterized 1 via enzyme kinetic studies, direct

binding studies, and X-ray crystallography. The insights gained

from the X-ray cocrystal structure and other biophysical studies

of RPTPγ with 1 facilitated the design of analogues with

signiﬁcantly improved in vitro activity. We also report the

cocrystal structures of several key analogues with RPTPγ.

Characterization of the HTS Hit, Compound 1. To support

lead optimization of compound 1,²⁸we characterized the struc-

ture of RPTPγ and its interactions with this HTS hit. The HTS

assay used a RPTPγ protein consisting of residues 790ꢀ1445,

which is the cytoplasmic portion of the protein. This construct

included domains 1 and 2 of RPTPγ, 37 amino acids of the full

length protein N-terminal to domain 1, and 16 non-native amino

acids at the N-terminus.²⁸Initial efforts to use this protein

construct for structural studies met with little success. Protein

recovery was poor during the buffer exchange necessary for direct

binding experiments of HTS compounds, and no crystals of this

protein were obtained in broad screening trials. To overcome

these difficulties, a single domain 1 construct consisting of

residues 825ꢀ1128 was selected for structural work.²⁹NMR

spectra of this domain 1 construct contain resonance dispersion

X-ray Structure and SAR Insights. Molecular Replacement

and Model Building of Apo Orthorhombic Form. While the kinetic

studies indicated that 1 inhibited RPTPγ in an apparent compe-

titive manner, crystal structures of RPTPγ, both in apo form and

in complex with 1 and related analogues, provided a detailed

picture of the manner in which the ligands compete with

6549

dx.doi.org/10.1021/jm2003766 |J. Med. Chem. 2011, 54, 6548–6562

Journal of Medicinal Chemistry

ARTICLE

Figure 2. (a) Chemical shift perturbations conﬁrm direct binding of 1 to RPTPγ. The ¹H/¹⁵N TROSY NMR spectra of 125 μM RPTPγ (black) and

upon the addition of 500 μM of 1 (red) are shown. (b) Eﬀect of 1 on the thermal stability of RPTPγ. Error bars represent standard deviation from

multiple replicates.

Figure 3. RPTPγ inhibition by 1 is time dependent but readily reversible. (a) Compound 1 dose response curves demonstrate a preincubation time

dependent increase in the potency of RPTPγ inhibition. Compound 1 achieves maximal inhibition by the 20 min time point. The results are from a

representative experiment from two trials. (b) RPTPγ pretreated with 1 (followed by 12-fold dilution at 0 min to reduce [1]) generates an overlapping

progress curve with vehicle pretreated RPTPγ.

substrate. The first crystals obtained with the RPTPγ domain 1

construct were orthorhombic with two molecules per asymmetric

unit. As described in the methods section, molecular replacement

was used with a model derived from the published RPTPμ crystal

structure (1RPM).³²By use of the AMoRe^33ꢀ35rotation

function, two peaks emerged well above the noise continuum,

but 46 other peaks had heights that were at least 50% of the

maximum. The two peaks did not show an obvious symmetry

relationship, and both yielded clean translation function solu-

tions. Rigid body fitting improved the statistical indicators as did

combining the two solutions. The statistical indicators for

molecular replacement were correlation coefficients on F and I,

respectively, of 0.523 and 0.569, and the R-value was 0.436.

Following a round of refinement, model completion proceeded

and the following sequence stretches between 825 and 1128

could be traced: A chain, 825ꢀ897, 901ꢀ1000, 1014ꢀ1122; B

chain, 826ꢀ898, 901ꢀ1002, 1013ꢀ1122. The chains were quite

similar to each other with an rmsd of 0.6 Å for 282 of 283 Cα

atom pairs. Both chains resembled a typical PTP domain with an

rmsd of 1.3 Å for 277 of 278 Cα atom pairs relative to RPTPμ

and an rmsd of 1.6 Å for 269 of 281 Cα atom pairs relative to

PTP1B (2HNP).²³

The orthorhombic apo crystals were grown in the presence

of ammonium sulfate. Sulfate anions were localized to the PTP

active site, which is deﬁned by Cys1060 in RPTPγ. Other

residues that make up the active site binding pocket include

Ser1061, Ala1062, Gly1063, Val1064, and Arg1066. While this

work was underway, an apo structure of domain 1 of RPTPγ

(2H4V) in a crystal form that is very similar to this structure and a

structure of domain 1 and domain 2 (2NLK) were deposited.²⁷

6550

dx.doi.org/10.1021/jm2003766 |J. Med. Chem. 2011, 54, 6548–6562

Journal of Medicinal Chemistry

ARTICLE

Figure 4. (a) LineweaverꢀBurke plot for RPTPγ inhibition by 1. (b) K_idetermination for inhibition of RPTPγ by 1. The results are from a

representative experiment from two trials.

In the latter structure, domain 1 is similar to our apo structure of

domain 1. Another RPTPγ domain 1 structure with sulfate

bound in the active site was also reported (2PBN)³⁶and is very

similar to our apo structure.

WPD-loop in the RPTPγ complex with 1 adopted a diﬀerent

conformation. Moreover, although it was expected that 1 would

bind at the active site, since it displayed apparent competitive

kinetics, it was obvious that the ligand did not bind there; i.e., it

did not overlap the position of phosphotyrosine superimposed

from 1PTY.²⁴Instead, compound 1 binds in a volume partially

overlapping that occupied by part of the WPD-loop (especially

the side chain of Trp1026 and the main chain of Val1031 and

Pro1032) in the open conformation of RPTPγ and forces the

WPD-loop to move away from the active site into a conformation

we have termed superopen. In the A chain, 1 only partially

occupied the site and the rest of the time the WPD-loop was in

the open conformation, whereas in the B chain, 1 appeared to

bind with full occupancy aﬀording a more well-deﬁned electron

density map (Figure 5). Residues that comprise the binding site

for 1 include Ile950, Tyr1023, Trp1026, Val1031, Pro1032,

Arg1066, Thr1069, Thr1070, Thr1105, Glu1107, Gln1108, and

Phe1111.

As shown in Figure 6, 1 binds proximal to, but not within, the

active site and does not interact with the catalytic Cys1060

residue either directly or indirectly. The dichlorophenyl moiety

of 1 is completely enclosed in a hydrophobic pocket formed by

Ile950, Tyr1023, Trp1026, Pro1032, Val1038, Phe1041, Val1042,

Thr1069, Tyr1070, Ile1073, and Phe1111. The dichlorophenyl

ring and Tyr1023 form an “edge to face” π-stacking interaction.

The thiophene moiety of 1 forms hydrophobic contacts with side

chain methylenes of Glu1107 and Gln1108 and with Val1031,

Pro1032, and Phe1111. The thiophene and Phe1111 form an

oﬀset “edge to face” π-stacking interaction. The thiomethylene

linker in 1 makes hydrophobic contacts with Trp1026, Pro1032,

Val1038, Phe1111, and the side chain methylene moieties of

Arg1066. Polar interactions formed between 1 and RPTPγ

domain 1 include a direct hydrogen bond to the Arg1066

guanidine moiety and a water-bridged hydrogen bond to the

hydroxyl moiety of Thr1105. One of the carboxylate oxygen atoms

of 1 is 3.4 Å from the side chain amide nitrogen atom of Gln1108,

suggesting a relatively weak hydrogen bonding interaction.

In order for 1 to bind to RPTPγ as described above, the WPD-

loop must adopt a conformation distinct from either the open or

Molecular Replacement and Model Building of Vanadate

Complex. Both orthorhombic crystals and crystals of a second,

trigonal crystal form of RPTPγ, which contained only one

molecule in the asymmetric unit,²⁹were placed in an artificial

crystal growth solution containing 10 mM Na₃VO₄. The struc-

ture of the orthorhombic crystal form was obtained using a dimer

consisting of chains A and B, and that of the trigonal crystal was

determined by molecular replacement using chain A of the

orthorhombic crystal form as initial models. In both crystal

forms, the WPD-loop was found in the closed conformation

with the vanadium atom liganded to Cys1060 Sγ and with the

four oxygen atoms liganded to (1) Asp1028 Oδ1, Gln1104 Nε2,

and a water molecule, which was liganded to Asp1028 Oδ2; (2)

Ser1061 N and Ala1062 N; (3) Gly1063 N and Val1064 N; (4)

Arg1066 N and a water molecule, which was liganded to

Met1029 N and Gln1108 Nε2.

Molecular Replacement and Model Building of RPTPγ with

Compound 1 Bound. The X-ray crystal structure of the apo

catalytically active phosphatase domain 1 of RPTPγ reported

here (and previously reported by other workers in PDB entries

2H4V²⁷and 2PBN³⁶) exhibits a WPD-loop conformation very

similar to the open conformation reported for other PTPs, while

the X-ray crystal structure of RPTPγ domain 1 in complex with

vanadate is in good agreement with the aforementioned closed

conformation. The X-ray crystal structure of RPTPγ domain 1

complexed with 1 revealed an unexpected and unprecedented

third conformation for the WPD-loop.

The model of the dimeric apo RPTPγ molecule discussed

above was used as an initial model for molecular replacement for

the orthorhombic crystals of RPTPγ either cocrystallized with 1

or with 1 soaked into the crystal. The electron density maps for

crystals where 1 was either cocrystallized or soaked into extant

crystals showed very similar electron density patterns that both

diﬀered from the apo structure. Comparison to the structure of

PTP1B bound to a ligand (1PTY²⁴) clearly showed that the

6551

dx.doi.org/10.1021/jm2003766 |J. Med. Chem. 2011, 54, 6548–6562

Journal of Medicinal Chemistry

ARTICLE

closed conformations of apo and vanadate-bound RPTPγ,

respectively. Comparison of the vanadate-bound, apo, and

compound 1-bound X-ray structures of RPTPγ as shown in

Figure 7 highlights the extent of this conformational change.

Figure 7a depicts the displacement of Asp1028 in the middle of

the WPD-loop from the closed to open conformation and the

additional large shift from the open to superopen conformation.

A modest change in the position of Trp1026 at the N-terminal

end of the WPD-loop is observed between the closed and open

conformations; the indole rings of Trp 1026 partially overlap

(Figure 7c). However, in the compound 1-bound superopen

conformation, the indole is shifted substantially and does not

overlap with the indole rings in the open and closed conforma-

tions. As shown in Figure 7b and Figure 7c, this is necessary to

allow 1 to ﬁt into the binding pocket.

These results indicate that the RPTPγ WPD-loop can stably

occupy at least three diﬀerent conformations: (1) closed with

traditional inhibitors, e.g., vanadate bound in the active site; (2)

open with nothing other than solvent molecules, including

sulfate, bound in the active site; (3) a new conformation termed

superopen (Figure 7). The superopen conformation is observed

when ligands such as 1 bind to a new site adjacent to the active

site, displacing Trp1026. Note that Asp1028, which has been

implicated in the catalytic mechanism, is distant from the active

site in the open conformation and even farther away in the

superopen conformation.

Benzylthioether Analogues of 1. Initial synthetic efforts

focused on modifications to the 3,4-dichlorobenzyl thioether

and carboxylic acid substituents of 1. As shown in Table 1, the

SAR of the 3,4-dichlorobenzyl thioether proved to be quite

narrow and steep, particularly with respect to the position and

extent of halogen substitution. For example, removal of both

chloro groups of 1 afforded an inactive compound (2, R = Ph),

while removal of the 3-chloro (3) or 4-chloro (4) substituent of 1

provided analogues with weak PTPγ activity. Similarly, the 2,3-

dichlorophenyl analogue (5) is 4-fold less active than 1. Isosteric

chloro replacements such as the 3,4-dimethylphenyl (6) and 2-

naphthyl (7) analogues were also inactive, but the 3,4-dibro-

mophenyl analogue 8 was equipotent with 1, suggesting a

preference for halogen substituents. Attempts to modify the

thioether linker of 1 (X = S) were similarly unsuccessful,

as exemplified by compounds 9 (X = O), 10 (X = NH), and 11

(X = SO₂).

Figure 5. Compound 1 bound to RPTPγ. Selected portions of RPTPγ

are shown either as a worm representing the backbone or with side chain

atoms represented by sticks. RPTPγ carbon atoms are shown in cyan.

Compound 1 carbon atoms are shown in green. Nitrogen atoms are

shown in blue, oxygen atoms in red, sulfur atoms in yellow, and chlorine

atoms in light green. Initial (i.e., prior to ﬁtting 1) 2F_oꢀ F_celectron

density is shown as magenta caged contours at 1σ. Hydrogen bonds are

shown in black as a series of small prolate ellipsoids. Image was created

with PyMOL.³⁷

Analogues of 1 That Extend into the Active Site. In contrast to

the narrow SAR observed for the 3,4-dichlorobenzyl thioether

side chain, the carboxylic acid of 1 could be replaced with an

acylmethylsulfonamide (12) or an acylphenylsulfonamide (13)

without affecting RPTPγ activity (Table 2). At this juncture,

Figure 6. (a) Compound 1 bound to RPTPγ. The location of the active site is indicated by Cys1060, and the Trp, Pro, and Asp residues of the WPD-

loop are shown (stick representation, orange carbon atoms). Other residues proximal to the bound ligand are shown in stick representation with cyan

carbon atoms, and the protein backbone is depicted with a gray cartoon representation. Water molecules are shown as red spheres, and the protein

surface is shown in gray. (b) Detailed view of the binding pocket for 1 with proximal RPTPγ residues depicted with cyan carbon atoms. Carbon atoms

from 1 are in green in parts a and b. Hydrogen bonds are shown in black as a series of small prolate ellipsoids. Images were created with PyMOL.³⁷

6552

dx.doi.org/10.1021/jm2003766 |J. Med. Chem. 2011, 54, 6548–6562

Journal of Medicinal Chemistry

ARTICLE

Figure 7. Superposition of the three states of the WPD-loop of RPTPγ. The closed form with vanadate bound is shown in magenta. The open form of the

apo protein is shown in orange, and the superopen form with 1 bound is shown in cyan. The carbon atoms for 1 are shown in green. Vanadium is shown in

gray, nitrogen in blue, oxygeninred, sulfurinyellow, and chlorine in lightgreen. (a, b) Orthogonalviews in the vicinity ofthe bindingsite for 1 and the active

site. (c) Close-up showing the positions of Trp1026 in the closed, open, and superopen conformations. Images were created with PyMOL.³⁷

Table 1. RPTPγ IC₅₀for Benzylthioether Analogues of 1

Table 2. RPTPγ IC₅₀for Acylsulfonamide Analogues

IC₅₀, μM

compd

R

X

RPTPγ

CD45

PTP1B

1

3,4-di-Cl-Ph

Ph

S

6.6

>100

>120

100

79

2

>120

80

50

3

4-Cl-Ph

>120

90

4

3-Cl-Ph

80

5

2,3-di-Cl-Ph

3,4-di-Me-Ph

2-naphthyl

3,4-di-Br-Ph

3,4-di-Cl-Ph

20

40

6

110

>120

3.4

>120

60

7

8

9

O

>120

80

10

11

NH

SO₂

>120

the X-ray structures of 1 and 12 bound to RPTPγ were

determined and provided additional insights into analogue de-

sign. Specifically, the proximity of the binding site for 1 and 12 to

the RPTPγ active site, and the bound conformation of the acyl

methylsulfonamide moiety in 12, which directed the methyl

substituent into the active site, prompted the synthesis of a

number of acylphenylsulfonamide analogues. These analogues

incorporated phenyl substituents capable of accepting additional

hydrogen bonds from the many backbone and side chain amide

NH and NH₂moieties present in this region of the protein (e.g.,

Gln1108, Arg1066, Gln1104, Gly1065, Val1064).

Figure 8a and Figure 8b show the structures of the complexes

of acylmethylsulfonamide 12 and acyl-m-aminophenylsulfona-

mide 14 bound to RPTPγ domain 1. The thiophene and

dichlorophenyl moieties of these compounds bind in a fashion

similar to that described for 1 above. Similar to 1, these

compounds also form water-bridged hydrogen bonds to the side

chain hydroxyl moiety of Thr1105. However, they diﬀer from

each other and from 1 in how they engage Arg1066.

Compound 12 forms two hydrogen bonds to the guanidine

moiety of Arg1066 through the acyl carbonyl and sulfonyl

moieties. The OꢀN distances for both of these are relatively

long, at 3.1 and 3.0 Å, respectively. The conformation of the

Arg1066 side chain diﬀers in the RPTPγꢀ14 complex, and only

one, shorter (2.5 Å) and presumably stronger, hydrogen bond

forms between the ligand sulfonyl group and the Arg1066

guanidine moiety. Both the methyl group in 12 and the

m-aminophenyl moiety in 14 occupy part of the active site.

However, neither the larger phenyl ring (13) nor the addition of

the potentially hydrogen bond-donating aminophenyl moiety

(14) provided a signiﬁcant improvement in potency relative to

6553

dx.doi.org/10.1021/jm2003766 |J. Med. Chem. 2011, 54, 6548–6562

Journal of Medicinal Chemistry

ARTICLE

Figure 8. (a) Structure of the complex of RPTPγ with acylmethylsulfonamide 12. (b) Structure of the complex of RPTPγ with acyl-m-

aminophenylsulfonamide 14. The ligands are depicted in stick representation with green carbon atoms. Residues proximal to the bound ligands are

shown in stick representation with cyan carbon atoms, and the protein backbone is depicted with a gray cartoon representation. Water molecules are

shown as red spheres, and the protein surface is shown in gray. Hydrogen bonds are shown in black as a series of small prolate ellipsoids. Images were

created with PyMOL.³⁷

two hydrogen bonds with the Arg1066 side chain guanidine

moiety, and a water-mediated hydrogen bond to Thr1105 in the

same manner as 12. The OꢀN distance between a side chain

carboxylate oxygen atom of Glu1107 and the bridging water

molecule is 3.5 Å, indicating an additional weak water-bridged

hydrogen bond to the N atom of the ligand acylsulfonamide

moiety. The carboxylate forms direct hydrogen bonds to the

backbone NH of Arg1066 and the side chain amides of Gln1104

and Gln1108. Water-mediated interactions with the backbone

NH moieties of Val1064 and Gly1065 and the SH of Cys1060

were observed as well. While the potency improvement for 15

relative to 12 was encouraging, it was achieved through the

introduction of an additional acidic moiety. Since the physical

properties of small molecules bearing two acidic moieties are not

generally conducive to membrane permeability, as required for

cellular uptake and general bioavailability, including brain uptake,

additional modiﬁcations were explored.

Analogues Reaching through “Tunnel” Acylsulfonamide ana-

logues such as 15 provided a significant improvement in RPTPγ

Figure 9. Structure of the complex of RPTPγ with 15. The ligand is

depicted in stick representation with green carbon atoms. Residues

activity relative to 1. However, substituents bearing a carboxylic acid

(or bioisostere) were required to achieve submicromolar potency.

Because of the need to achieve suitable druglike properties, we

turned our attention to removing acidic moieties and gaining

binding affinity through alternative interactions in a new pocket

that we had observed in previous X-ray structures: namely, a region

that is accessible via a narrow cavity or “tunnel” extending from the

5-position of the thiophene. This new region is replete with residues

and affords multiple potential hydrogen bonding opportunities.

forming direct or water-bridged hydrogen bonds to the ligand and the

residues of the “WPD” motif are depicted in large diameter stick

representation, and other proximal residues are depicted with thin sticks.

Protein carbon atoms are colored cyan. Water molecules are shown as

red spheres. Hydrogen bonds are shown in black as a series of small

prolate ellipsoids. Image was created with PyMOL.³⁷

the methyl group in 12. As shown in Figure 8b, the amino moiety

in 14 is directed away from the binding site and does not form

any direct or water-bridged hydrogen bonds with the enzyme.

As shown in Table 2, introduction of a m-benzoic acid

substituent (15) improved potency by an order of magnitude

with a concomitant enhancement in selectivity versus PTP1B

and a concomitant increase in the T_mfrom 39.4 to 44.5 °C

(Supporting Information Figure S1). The corresponding methyl

ester (16) was signiﬁcantly less active, suggesting that the

carboxylate was required for increased activity. This was subse-

quently conﬁrmed via the X-ray structure of 15 bound to RPTPγ.

As shown in Figure 9, the acylsulfonamide moiety in 15 forms

Our initial synthetic approach to 5-position analogues is

shown in eq 1, wherein a 2,3,5-trisubstituted thiophene was

6554

dx.doi.org/10.1021/jm2003766 |J. Med. Chem. 2011, 54, 6548–6562

Journal of Medicinal Chemistry

ARTICLE

Table 3. RPTPγ IC₅₀for Alkynylbenzoic Acid Analogues

and 23 was surprisingly inactive. Compounds 21 and 23 dis-

played greater potency against PTP1B, one of the two PTPs

employed to assess selectivity (along with CD45), than RPTPγ.

We have not determined cocrystal structures of PTP1B with any

compound 1 analogues, and it is unknown how 21 and 23 bind

to PTP1B.

The X-ray cocrystal structure of 20 bound to RPTPγ was

determined and is shown in Figure 10b. Although the alkynyl

analogues did not provide a signiﬁcant improvement in activity

relative to 1, the X-ray structure of 20 bound to RPTPγ

demonstrated that additional residues proximal to the enzyme

active site (e.g., Tyr841, Ile1110, Asp1114) could be targeted via

substituents on the 2-(3,4-dichlorobenzylthio)benzoic acid core

(17) that projected through “the tunnel” as predicted.

’ DISCUSSION

NMR and thermal shift studies conﬁrmed that 1, which was

identiﬁed through high-throughput screening, binds to RPTPγ

domain 1. This ligand inhibits RPTPγ-catalyzed dephosphoryla-

tion of substrate, and kinetic studies indicate that the mode of

inhibition is apparent-competitive. Determination of the cocrys-

tal structure of 1 with RPTPγ revealed a new and unique

hydrophobic ligand binding site on the surface of RPTPγ domain

1 which is exposed by movement of the WPD-loop into a novel

superopen conformation. This site is adjacent to the enzyme

active site and is occupied by the Trp1026 side chain in the

typical open and closed conformations of the enzyme. Com-

pound 1 binds in this pocket, and when bound, it prevents the

WPD-loop from closing over the enzyme active site. Residue

Asp1028 from the WPD-loop is involved in the catalytic mechan-

ism of PTPs, and thus, even if substrates could bind to the

RPTPγ active site at the same time as 1, catalytic dephosphor-

ylation of substrate should still be inhibited. Although 1 does not

directly block the active site, analysis of the RPTPγ/1 cocrystal

structure reported herein supports the results of the kinetic

studies that indicate apparent competitive inhibition of the

enzyme reaction. The discovery that compound 1 did not bind

to the active site was a surprise because we anticipated that the

carboxylate would act as a phosphate mimetic as has been seen

with PTP1B active site inhibitors (e.g., see Iversen et al.,³⁸

1C86,³⁸1C87,³⁸1C88³⁸). An attempt to locate the binding site

for compound 1 in PTP1B by soaking into extant crystals did not

reveal any binding, so the binding site for compound 1 and

analogues in other PTPs is unclear.

Cocrystal structures of complexes of RPTPγ with acylsulfo-

namide compounds 12 and 15 also show the ligands bound to

this site. The cocrystal structure of the RPTPγ/15 complex

demonstrates the possibility of designing ligands that occupy this

novel pocket and project into the enzyme active site. On the basis

of this ﬁnding, a lead optimization eﬀort resulted in the synthesis

of additional compounds with ∼10-fold potency improvements

over the original hit, 1. However, the results to date suggest that it

may prove diﬃcult to exploit interactions with binding site

residues without appending moieties that impart undesirable

physicochemical properties to the resultant compounds. At-

tempts to show signiﬁcant compound mediated inhibition of

PTPγ in a cellular context were unsuccessful (data not shown).

This may be due to the modest compound potency for inhibition

of PTPγ and/or the relatively poor compound membrane

permeability. This diﬃculty has plagued previous eﬀorts to

design PTP inhibitors with druglike properties.^{7,10,11,13,14}

envisioned to undergo a Sonogashira reaction with a terminal

alkyne to provide the corresponding alkynyl product. Unfortu-

nately, all attempts to synthesize useful quantities of the thio-

phene precursor were unsuccessful, and so attention was focused

on analogues of compound 17 (Table 3), in which the thiophene

core was replaced by a phenyl ring. Compound 17 was equipo-

tent with 1, but its corresponding methylacylsulfonamide (18)

was surprisingly inactive (cf. 12 and 18).

A subsequent X-ray cocrystal structure of 17 bound to RPTPγ

(Figure 10a) highlighted a major diﬀerence in the binding mode

of the carboxylate compared to 1 and suggested that the

corresponding acylsulfonamide would be unable to exploit the

key interactions observed in the cocrystal structure of RPTPγ

with 12. Speciﬁcally, the carboxylate of 1 is twisted out of the

plane of the thiophene ring by 6°, while the carboxylate in 17

deviates from coplanarity with the attached phenyl ring by 47°.

Along with other geometric diﬀerences resulting from the shapes

of the ﬁve- and six-membered rings in 1 and 17, respectively, the

larger deviation from coplanarity with the attached ring allows 17

to form shorter hydrogen bonds to Arg 1066 and Gln 1108

compared to 1. It also suggests that acylsulfonamides based on 17

will not orient the sulfonamide substituent in the desired

direction. As a consequence, the new vector was explored via the

p-bromobenzoic acid analogue 19, which can be synthesized in large

quantities and is an excellent substrate for the Sonogashira reaction

with a variety of terminal alkynes.

As shown in Table 3, 17 and the corresponding p-bromo

analogue 19 had similar potencies. Treatment of 19 with terminal

alkynes under Sonogashira conditions yielded alkynyl analogues

20ꢀ23 and a number of interesting SAR observations. Com-

pounds 20 and 21 provided modest improvements in potency.

Compound 22 was an order of magnitude more potent than 19,

6555

dx.doi.org/10.1021/jm2003766 |J. Med. Chem. 2011, 54, 6548–6562

Journal of Medicinal Chemistry

ARTICLE

Figure 10. (a) Structure of the complex of RPTPγ with 17. (b) Structure of the complex of RPTPγ with 20. The ligands are depicted in stick

representation with green carbon atoms. The location of the active site is indicated by Cys1060 which is shown in stick representation with orange carbon

atoms along with the Trp, Pro, and Asp residues of the WPD-loop. Other residues proximal to the bound ligands are shown in stick representation with

cyan carbon atoms, and the protein backbone is depicted with a gray cartoon representation. Water molecules are shown as red spheres, and the protein

surface is shown in gray. Hydrogen bonds are shown in black as a series of small prolate ellipsoids. Image was created with PyMOL.³⁷

While this work was in progress, the New York Structural

Genomics Consortium deposited coordinates of RPTPγ (2HY3)

with vanadate bound and the WPD-loop in an open confor-

mation.³⁶It is unclear what might be preventing the WPD-

loop from closing in that structure. The closed conformations of

the WPD-loop in RPTPγ complexed with vanadate in both crystal

forms reported here were very similar, and the position of the

WPD-loop resembled that for PTP1B with phosphotyrosine bound

(1PTY)²⁴rather than PTP1B with tungstate bound (2HNQ),²³

where it is also unclear what prevents the WPD-loop from closing.

Two structures of PTP1B (3I7Z, 3I80)²²with vanadate, represent-

ing two diﬀerent transition state analogues for two diﬀerent catalytic

steps, were reported while this manuscript was in preparation. In

both cases the WPD-loop is in the closed position. A search of the

PDB for a sequence similar to the phosphatase domain of RPTPγ

and either tungstate or vanadate bound provided eight examples.

In addition to 2HNQ²³(WO₄^3ꢀ) and 2HY3³⁶(VO₄^3ꢀ), which

3ꢀ

have the WPD-loop in the open conformation, and 3I7Z (VO₄

)

and 3I80 (VO₄^3ꢀ), which have the WPD-loop in the closed

position, four other examples were found with the WPD-loop in

the closed conformation: Yersinia PTPase (1YTW,³⁹WO₄^3ꢀ),

CDC14 (1OHD,⁴⁰WO₄^3ꢀ), PTPβ (2IE4,⁴¹VO₄^3ꢀ), and the

archaeal Sulfolobus solfataricus PTP (2I6M,⁴²VO₄^3ꢀ).

Figure 11. Diﬀerent WPD-loop conformations of PTPs. The closed

form of RPTPγ with vanadate bound is shown in magenta. The open

form with nothing bound is shown in orange, and the superopen form

with 1 bound is shown in cyan. The carbon atoms for 1 are shown in

green. The PTP active site is denoted by Cys1060. The intermediate

WPD-loop conformation is represented by the PCPTP1 structure

(2A8B,²⁶yellow), and atypical conformations are represented by

structures of GLEPP1 (2GJT,²⁷green), STEP (2BIJ,²⁶blue), and Lyp

(2P6X,²⁷red). Image was created with PyMOL.³⁷

In a recent comprehensive analysis of PTP crystal structures,

Barr and co-workers highlighted two other conformations that

the WPD-loops of PTPs can adopt.²⁷These include an “inter-

mediate” conformation of PCPTP1 (2A8B²⁶) in which the

WPD-loop occupies a position between the open and closed

states and “atypical” conformations for GLEPP1 (2GJT²⁷),

STEP (2BIJ²⁶), and Lyp (2P6X²⁷). In the atypical conforma-

tions, the WPD-loop moves away from the active site laterally,

while the closed, intermediate, open, and superopen conforma-

tions involve more vertical displacement of the WPD-loop over

the active site (Figure 11). Thus, the superopen RPTPγ con-

formation reported here appears to be more related to the open,

intermediate, and closed WPD-loop conformations.

(PTP1B) inhibitor reported by Wiesmann and co-workers to

inhibit PTP1B noncompetitively with respect to p-nitrophenyl

phosphate substrate (compound 24, Figure 12).⁴³Compound

24 binds to PTP1B at a site distant from the active site and forms

multiple contacts with the protein including π-stacking interac-

tions with Phe280. Based in part on a comparison of crystal

structures of PTP1B with compound 24 (or close analogues)

bound, PTP1B in the unliganded open conformation (2HNP²³),

and a Cys215Ser PTP1B mutant bound to phosphotyrosine in

the closed conformation (1PTY²⁴), Wiesmann et al. postulated

The binding mode of 1 may be contrasted to that of a

previously reported allosteric protein tyrosine phosphatase 1B

6556

dx.doi.org/10.1021/jm2003766 |J. Med. Chem. 2011, 54, 6548–6562

Journal of Medicinal Chemistry

ARTICLE

WPD-loop into the closed conformation is hindered by steric

interactions that prevent the Trp residue of the WPD-loop from

occupying the hydrophobic pocket where it resides in the closed

conformation. The diﬀerent origin of the steric hindrance, direct

in the case of the RPTPγ inhibitors vs indirect (through Phe191)

for the PTP1B inhibitors, may help explain the mechanistic

diﬀerence (apparent competitive vs noncompetitive) between

1 and 24.

The unique movement of the RPTPγ WPD-loop was

exploited to design compounds that were selective relative to

other PTPs. Currently, the ligand-induced superopen WPD-loop

conformation described herein is unique to RPTPγ; however,

the WPD-loop is a highly conserved and mechanistically im-

portant motif in the PTP superfamily. This structural homology

may portend the discovery of small molecules that inhibit other

PTPs via displacement of the WPD-loop tryptophan.

Figure 12. Superposition of the PTP1B/compound 24 (1T49,⁴³

magenta), the RPTPγ/1 (cyan protein representation, ligand depicted

with green carbon atoms), and the PTP1B(C215S)/pTyr (1PTY,²⁴

yellow) cocrystal structures. The Phe191 side chain in the PTP1B/

compound 24 complex and 1 in the RPTPγ/1 complex sterically hinder

movement of the WPD-loop Trp side chain (Trp179 in PTP1B and

Trp1026 in RPTPγ) to the position it occupies in the catalytically active

closed conformations of the respective enzymes. The PTP WPD-closed

conformation is represented by the PTP1B(C215S)/pTyr structure.

Image was created with PyMOL.³⁷

’ EXPERIMENTAL SECTION

RPTPγ Cloning and Expression. The RPTPγ catalytic domain

construct (Figure 13) used for direct binding studies (NMR and thermal

shift assay) and X-ray crystallography consisted of residues 825ꢀ1128

and is described elsewhere.²⁹Briefly, transformed E. coli Rossetta (DE3)

(Novagen, Madison, WI) at 37 °C were induced at OD₆₀₀= 1.0 with

0.25 mM IPTG and grown for 30 h at 20 °C. The protein was isolated

using ammonium sulfate precipitation, followed by Ni affinity, Q

Sepharaose, and SP Sepharose chromatography. The protein could be

concentrated to 12 mg/mL in 25 mM Tris, pH 7.5, 200 mM NaCl, 2 mM

EDTA, 5 mM DTT, and typical yields were 40 mg/L.

The RPTPγ catalytic domain construct used for enzyme assays

consisted of residues 790ꢀ1445 (full cytoplasmic domain) with an

N-terminal His-myc tag. Brieﬂy, transformed E. coli BL21 (DE3)

growing at 37 °C were induced at OD₅₀₀= 4.5 with 10% lactose and

grown for 18 h at 25 °C. The cells were lysed using a MicroFluidizer, and

the protein was isolated using Ni aﬃnity and gel ﬁltration (Superdex 75)

chromatography. Finally the puriﬁed protein was dialyzed into 20 mM

HEPES, pH 8.0, 200 mM NaCl, 5 mM DTT, and 40% glycerol.

Protein Crystallization. The purification and crystallization of the

catalytic domain of RPTPγ have been described in detail elsewhere.²⁹

Briefly, a construct was designed that encompassed residues 825ꢀ1128

of RPTPγ and an N-terminal His₆tag (Figure 13). The protein was

expressed in E. coli. RPTPγ was purified by sequential column chroma-

tography involving Ni affinity, ion exchange, and molecular sieves.

Two crystal forms were examined in detail, one with symmetry

consistent with space group P2₁2₁2₁and unit cell parameters of a =

75 (1Å, b=80 (2.5 Å, c=125 (2.5 Å, which had two catalytic domains

per asymmetric unit. In particular, the b and c axes of this crystal form varied

depending upon whether or not a ligand was bound and at which site

binding occurred. A second crystal form had symmetry consistent with

space group P3₂21 and unit cell parameters of a = b = 75 Å, c = 152 (

1.5 Å. The c-axis of this crystal form varied depending upon whether or not a

ligand was bound and at which site binding occurred.

that compound 24 and analogues inhibit PTP1B by locking the

enzyme in an open conformation in which the WPD-loop cannot

close over the active site.⁴³Support for this hypothesis is

provided by a targeted molecular dynamics study of the complex

of a compound 24 analogue with PTP1B that showed reduced

WPD-loop ﬂexibility compared to apo-PTP1B and the PTP1B-

pTyr complex.⁴⁴In their analysis, Wiesmann and co-workers

highlighted the importance of Phe191 and pointed out that when

compound 24 is bound to PTP1B, the side chain of Phe191

partially ﬁlls the cavity that Trp179 from the WPD-loop occupies

in the closed conformation and prevents the WPD-loop from

closing.⁴³

Although the RPTPγ inhibitors reported herein do not bind

to RPTPγ at a site analogous to the allosteric site where

compound 24 binds to PTP1B, their putative modes of action

exhibit intriguing similarities. Figure 12 shows a superposition of

the RPTPγꢀ1 cocrystal structure, the cocrystal structure of

PTP1B with compound 24 (1T49⁴³), and the cocrystal structure

of a Cys215Ser mutant of PTP1B with phosphotyrosine bound

(1PTY²⁴). This latter structure represents a closed conformation

of PTP1B. Comparison of the PTP1B/24 and PTP1B/pTyr

cocrystal structures shows that the Phe191 side chain in the

PTP1B/24 structure projects into a cavity occupied by Trp179 in

the PTP1B/pTyr closed structure as originally pointed out by

Wiesmann, et al.⁴³In the RPTPγꢀ1 complex, the ligand

occupies part of the volume that the WPD-loop Trp1026 residue

ﬁlls in the closed conformation. Thus, both the substrate non-

competitive allosteric PTP1B inhibitors and the apparent-sub-

strate-competitive RPTPγ inhibitors reported here appear to act

by perturbing the dynamics of the WPD-loop and preventing its

closure over the active site. In the case of the allosteric PTP1B

inhibitors, the WPD-loop is locked into a typical open confor-

mation, while the RPTPγ inhibitors force that enzyme into a

novel superopen conformation. In both cases movement of the

Data Collection and Processing. Data were collected either in

the laboratory or at APS ID17 (IMCA-CAT) and are reported in detail

elsewhere.²⁹

Structure Determination by Molecular Replacement. The

CCP4³³version of AMoRe^34,35was used to determine the structure of

RPTPγ. The probe model used to determine the structure of the “apo”

form by molecular replacement was derived from RPTPμ (1RPM).³²

Stretches of residues, i.e., principally loops that were thought to differ

between RPTPμ and RPTPγ, were removed. The remaining noniden-

tical residues from 1RPM were shorn to their minimal isostructural

component, e.g., typically alanine, but for example, if RPTPμ had Phe

6557

dx.doi.org/10.1021/jm2003766 |J. Med. Chem. 2011, 54, 6548–6562

Journal of Medicinal Chemistry

ARTICLE

Figure 13. Amino acid sequence of the His-Tb-RPTPγ construct used to produce protein for NMR and crystallization studies. The hexa-His tag and the

thrombin cleavage site are underlined, and the amino acids from residues 825 and 1128 of RPTPγ are not underlined.

and RPTPγ had Tyr at a particular position, then Phe would have been

used in the model. A FORTRAN program, COALESCE, was written to

bring the individual molecular replacement solutions for the orthor-

hombic crystal, which had two molecules in the asymmetric unit, close

together in space. Following molecular replacement and interspersed

with refinement and model-building, the CCP4³³programs NCSMASK

and DM were used for NCS map averaging and other density

modifications.

to determine T_mvalues (T_m, the temperature at which the reaction was

half-complete).

RPTPγ Enzyme Assay. Dephosphorylation of phosphotyrosine

peptide substrate was monitored using the FRET based Z⁰-Lyte system

(Invitrogen). The assay was carried out in a total volume of 12 μL in a

buffer containing 25 mM MOPS, 50 mM NaCl, 0.5 mM EDTA, 1 mM

DTT, 0.1% Pluronic acid F-68, and 30 pM RPTPγ full length cytoplas-

mic domain. Inhibitors were dissolved and serially diluted in DMSO.

They were carried through a buffer intermediate dilution, added to the

assay, and allowed to preincubate with enzyme. For time dependence

experiments preincubation times of 0, 10, 20, and 40 min were used. All

other studies used a 20 min preincubation time. The final DMSO

concentration in the assay was 1.25%. The reaction was started with the

addition of phosphotyrosine substrate, Z⁰-Lyte Phospho-Tyr 1 peptide

(Invitrogen), and allowed to incubate at room temperature until the

desired time-point within the assay linear range. For kinetic analysis

these time points were 0, 3, 6, 9, and 12 min. The phosphatase reaction

was terminated with the addition of 6 μL of stop/development buffer

consisting of development buffer B (Invitrogen) with 0.45 mg/mL

development reagent B (Invitrogen) and 6 mM sodium vanadate. The

terminated reaction mixture was incubated for 60 min at room tem-

perature to allow development reagent B to cleave nonphosphorylated

peptide. The plate was then imaged using a ViewLux plate reader

(Perkin-Elmer) at an excitation wavelength of 400 nm and emission

wavelengths of 460 and 530 nm. Percent phosphorylation was derived

from the 460/530 emission ratio applied to a substrate phosphopeptide

standard curve. Kinetic analysis was carried out using Prizm 4.0 non-

linear curve fitting (Graphpad).

Fitting and Refinement of the Model. CNX⁴⁵(release 2005;

Accelrys, Inc., San Diego, CA, U.S.) was used for refinement, and the

QUANTA Modeling Environment (release 2005; Accelrys, Inc., San

Diego, CA, U.S.) was employed for model fitting. In the later stages

(compounds 20 and S2) BUSTER/TNT⁴⁶was used for refinement and

COOT for model fitting.^47,48Figures were generated with PyMOL.³⁷

Direct Binding Assays. Uniformly ²H/¹⁵N labeled RPTPγ

(825ꢀ1128) for 2D TROSY experiments was expressed in minimal

2

medium supplemented with H glucose, ¹⁵N ammonium sulfate, and

98% D₂O. All samples were approximately 500 μL in a 5 mm NMR tube.

1

The H/¹⁵N 2D TROSY experiments were carried out at 20 °C on a

1

Bruker NMR spectrometer (operating at a 700 MHz H frequency)

equipped with a Bruker 5 mm ¹H/¹³C/¹⁵N triple resonance, triple-axis

PFG probe. The samples were typically 125 μM ²H/¹⁵N RPTPγ,

500 μM compound in 50 mM sodium phosphate, pH 7.2, 100 mM

NaCl, 2 mM dithiothreitol-d₁₀, 1 mM EDTA-d₁₆, 7% D₂O.

For the NMR T₂relaxation experiments, two samples were prepared.

One sample contained ligand alone at 50 μM, and the other sample

contained 50 μM ligand + 20 μM RPTPγ in 20 mM Hepes-d₁₈, pD 6.8,

50 mM NaCl, 5 mM DTT-d₁₀, 0.5 mM EDTA-d₁₆, 1.0% DMSO-d₆, 99%

D₂O. A 50 ms spin-lock time with 512 scans was collected.

Reversibility experiments were carried out as above with a few

modiﬁcations. Assay preincubation conditions were 1 at 3 ꢁ K_i(7.5 μM)

and RPTPγ at 714 pM. After a 30 min preincubation the reaction was

diluted 12-fold with buﬀer and phosphopeptide substrate was added at

3 ꢁ K_m(6 μM). A parallel vehicle-only control reaction was carried out as a

measure of uninhibited RPTPγ activity.

The eﬀect of inhibitors on the thermal stability of RPTPγ was

measured with a ThermoFluor instrument (3-Dimensional Pharmaceu-

ticals, Inc. Yardley, PA)^49,50by monitoring the ﬂuorescence enhance-

ment of an external probe (1-anilino-8-naphthalene sulfonate [ANS]),

as it binds preferentially to the unfolded protein. Fluorescence was

monitored with a digital camera equipped with a 500 ( 30 nm ﬁlter for

detection. The unfolding reactions were carried out in a 384-well plate in

5 μL volume with 0.2 mg/mL RPTPγ (5 μM), 100 μM ANS, and 2.5%

(v/v) DMSO in buﬀer containing 25 mM PIPES, pH 7.0, 200 mM NaCl,

5 mM MgCl₂, and 1 mM β-mercaptoethanol. The contents were

overlaid with 3 μL of polydimethylsiloxane DC200 oil (Sigma) to

prevent evaporation. Unfolding reactions were monitored by ramping

the temperature in 1 °C increments from 25 to 80 °C with 60 s

equilibration time followed by four 10 s exposures (plus one dark ﬁeld

image) per temperature point. An average intensity per well was

calculated by integrating pixel intensities per well and averaging the

four exposures. The resulting denaturation curves were analyzed with

ThermoFluor analysis software (3-Dimensional Pharmaceuticals, Inc.)

Selectivity Assays. Inhibitor activity was evaluated at two addi-

tional phosphatases, PTP-1B and CD45 (both from Upstate, Lake

Placid, NY). The assay conditions were similar to those described above

for RPTPγ with the exception of the phosphatase enzyme concentra-

tions. The phosphatases PTP-1B and CD45 were used at 50 and 500

pM, respectively. As with RPTPγ, the assays were carried out using 2 μM

p-Tyr1 substrate peptide. The apparent K_mwe determined for PTP1B

and CD45 was 0.5 and 2.1 μM, respectively (data not shown).

Analogue Synthesis. All reactions were conducted in distilled

solvents under an inert atmosphere. Unless otherwise noted, all reagents

were purchased from commercial vendors and used without further

purification. Samples were purified via reverse-phase HPLC chroma-

1

tography, and final products were analyzed via HPLCꢀMS and H

6558

dx.doi.org/10.1021/jm2003766 |J. Med. Chem. 2011, 54, 6548–6562

Journal of Medicinal Chemistry

ARTICLE

NMR. All NMR spectra were recorded on a Bruker DPX-500A or Bruker

DPX-300B NMR spectrometer. Chemical shifts are expressed in parts

per million (δ) using a residual solvent proton as a reference. LC/MS

analysis was performed on Shimadzu HPLC equipped with a Micromass

spectrometer: column, Luna 3.0 mm ꢁ 50 mm; gradient time, 2 min,

from 0% B to 100% B (A, 95% H₂O/5% ACN with 10 mM NH₄OAc; B,

5% H₂O/95% ACN with 10 mM NH₄OAc). The purity of all com-

pounds was g95% as determined by HPLC analysis (UV detection at

220 nm).

Compound 1 [251096-84-1] was purchased from Key Organics

(Cornwall, U.K., catalog no. 10G-330S). ¹H NMR (DMSO-d₆) δ 7.90

(d, J = 5.2 Hz, 1H), 7.84 (s, 1H), 7.70 (d, J = 8 Hz, 1H), 7.61 (d, J = 8.2

Hz, 1H), 7.34 (d, J = 5.2 Hz, 1H), 4.53 (s, 2H). LC/MS (ESI) m/z 318

(M + H)⁺, t_R= 0.92 min.

Compounds 2ꢀ8 were prepared as follows: To a solution of 3-

mercaptothiophene-2-carboxylic acid (1.0 equiv) in DMSO were added

sodium methoxide (25 wt % in methanol, 2.0 equiv) and the appropriate

benzyl bromide (1.0 equiv). The resulting mixture was stirred at 25 °C

for 3 h, then concentrated in vacuo and puriﬁed by reverse-phase

preparative HPLC to aﬀord the desired product.

3-(Benzylthio)thiophene-2-carboxylic Acid (2). From 16 mg of 3-

mercaptothiophene-2-carboxylic acid, 11.4 mg (46%) of 2 was obtained.

¹H NMR **(MeOH-d₄) δ 7.66(d, J = 5.4 Hz, 1H), 7.45 (d, J = 7.4 Hz,

2H), 7.38 (t, J = 7.4 Hz, 2H), 7.28ꢀ7.23 (m, 1H), 7.16 (d, J = 5.4 Hz,

1H), 4.31 (s, 2H). LC/MS(ESI) m/z 251.1 (M + H)⁺, t_R= 1.45 min.

3-(4-Chlorobenzylthio)thiophene-2-carboxylic Acid (3). From 16

mg of 3-mercaptothiophene-2-carboxylic acid, 9.1 mg (32%) of 3 was

obtained. ¹H NMR (MeOH-d₄) δ 7.66 (d, J = 5.4 Hz, 1H), 7.44 (d, J =

8.4 Hz, 2H), 7.38 (d, J = 8.4 Hz, 2H), 7.14 (d, J = 5.4 Hz, 1H), 4.29 (s,

2H). LC/MS(ESI) m/z 284.1 (M + H)⁺, t_R= 1.61 min.

1H), 7.54 (d, J = 8.2 Hz, 1H), 7.43 (d, J = 8.2 Hz, 1H), 7.04 (d, J = 5.5 Hz,

1H), 5.26 (s, 2H). LC/MS(ESI) m/z 303 (M + H)⁺, t_R= 2.01 min.

3-(3,4-Dichlorobenzylamino)thiophene-2-carboxylic Acid (10).

Step 1. To a solution of methyl 3-aminothiophene-2-carboxylate(314 mg,

1.0 equiv) and 3,4-dichlorobenzaldehyde (350 mg, 2.0 equiv) in THF

(6 mL) were added dibutyldichlorostannane (0.4 mL, 0.02 equiv)

and phenylsilane (217 μL, 2.2 equiv). The resulting mixture was stirred

at 25 °C for 3 h. Then the solvent was removeed in vacuo and

the crude product was purified by preparative HPLC to afford 210 mg

(33%) of the intermediate methyl 3-(3,4-dichlorobenzylamino)thiophene-

2-carboxylate.

Step 2. To a solution of methyl 3-(3,4-dichlorobenzylamino)-

thiophene-2-carboxylate (78 mg) in 1.3 mL of EtOH was added 50%

NaOH_(aq)(0.3 mL). The mixture was heated to reﬂux for 1 h, then

cooled to 25 °C and concentrated in vacuo. The crude product was

diluted with ethyl acetate and water and acidiﬁed with 6 N HCl to

pH ∼3. The layers were separated, and the aqueous layer was back-

extracted with EtOAc (3ꢁ). The combined organic extracts were dried

(Na₂SO₄) and concentrated in vacuo to aﬀord 59 mg (76%) of 10. ¹H

NMR (MeOH-d₄) δ 7.51ꢀ7.49 (m, 1H), 7.46 (d, J = 5.4 Hz, 1H), 7.43

(s, 1H), 7.27ꢀ7.30 (m, 1H), 6.62 (d, J = 5.5 Hz, 1H), 4.52 (s, 2H).

LC/MS(ESI) m/z 302 (M + H)⁺, t_R= 1.03 min.

3-(3,4-Dichlorobenzylsulfonyl)thiophene-2-carboxylic Acid (11).

To a solution of 1 (7.8 mg, 1.0 equiv) in DCM (3 mL) was added

m-chloroperbenzoic acid (57ꢀ86%, purchased from Aldrich, 15.5 mg,

2.2 equiv) at 25 °C. The resulting mixture was stirred at 25 °C for 16 h,

then concentrated in vacuo and resuspended in MeOH and purified

by preparative HPLC to afford 6.6 mg of 11 (77%). ¹H NMR (MeOH-

d₄) δ 7.74 (s, 1H), 7.61 (s, 1H), 7.49ꢀ7.48 (m, 2H), 7.32 (s, 1H),

7.24ꢀ7.23 (m, 1H), 5.01 (s, 2H). LC/MS(ESI) m/z 351(M + H)⁺,

t_R= 0.82 min.

3-(3-Chlorobenzylthio)thiophene-2-carboxylic Acid (4). From 16 mg

of 3-mercaptothiophene-2-carboxylic acid, 4.8 mg (17%) of 4 was

Compounds 12ꢀ14 and 16 were prepared as follows: To a solution

of 1 (1.0 equiv) in DMF at 25 °C were added EDC (1.0 equiv), DMAP

(1.0 equiv), and RSO₂NH₂(1.0 equiv). The resulting mixture was

stirred at 25 °C for 6 h, then concentrated in vacuo and puriﬁed by

reverse-phase preparative HPLC to aﬀord the desired product.

3-(3,4-Dichlorobenzylthio)-N-(methylsulfonyl)thiophene-2-carbox-

1

obtained. H NMR (MeOH-d₄) δ 7.66 (d, J = 5.4 Hz, 1H), 7.48 (s,

1H), 7.38ꢀ7.26 (m, 3H), 7.14(d, J = 5.4 Hz, 1H), 4.29 (s, 2H). LC/

MS(ESI) m/z 285.1 (M + H)⁺, t_R= 1.58 min.

3-(2,3-Dichlorobenzylthio)thiophene-2-carboxylic Acid (5). From

16 mg of 3-mercaptothiophene-2-carboxylic acid, 9.2 mg (29%) of 5 was

obtained. ¹H NMR (DMSO-d₆) δ 7.69 (d, J = 5.4 Hz, 1H), 7.48(d, J =

7.9 Hz, 2H), 7.26 (t, J = 7.9 Hz, 1H), 7.13 (d, J = 5.4 Hz, 1H), 4.48 (s,

2H). LC/MS(ESI) m/z 319.1 (M + H)⁺, t_R= 1.68 min.

3-(3,4-Dichlorobenzylthio)thiophene-2-carboxylic Acid (6). From

16 mg of 3-mercaptothiophene-2-carboxylic acid, 18.1 mg (65%) of 6

was obtained. ¹H NMR (MeOH-d₄) δ 7.64 (d, J = 5.1 Hz, 1H),

7.21ꢀ7.15 (m, 3H), 7.09 (s, 1H), 4.22 (s, 2H), 2.26 (s, 3H), 2.25 (s,

3H). LC/MS(ESI) m/z 279 (M + H)⁺, t_R= 0.9 min.

3-(Naphthalen-2-ylmethylthio)thiophene-2-carboxylic Acid (7).

From 16 mg of 3-mercaptothiophene-2-carboxylic acid, 7.5 mg (25%) of

7 was obtained. ¹H NMR (MeOH-d₄) δ 7.91 (s, 1H), 7.86ꢀ7.84 (m,

3H), 7.65 (d, J = 5.4 Hz, 1H), 7.59 (d, J = 8.4 Hz, 1H), 7.50ꢀ7.48 (m,

2H), 7.21 (d, J = 5.4 Hz, 1H), 4.48 (s, 2H). LC/MS(ESI) m/z 301 (M +

H)⁺, t_R= 2.08 min.

1

amide (12). From 12 mg of 1, 4.5 mg (28%) of 12 was obtained. H

NMR (DMSO-d₆) δ 7.70 (d, J = 6.5 Hz, 1H), 7.61 (s, 1H), 7.47 (d, J =

8.4 Hz, 1H), 7.37 (d, J = 6.5 Hz, 1H), 7.12 (d, J = 8.4 Hz, 1H), 4.29 (s,

2H), 3.21 (s, 3H). LC/MS(ESI) m/z 396 (M + H)⁺, t_R= 0.90 min.

3-(3,4-Dichlorobenzylthio)-N-(phenylsulfonyl)thiophene-2-carbox-

1

amide (13). From 14 mg of 1, 2.6 mg (13%) of 13 was obtained. H

NMR (DMSO-d₆) δ: 8.07 (d, J = 7.7 Hz, 2H), 7.75ꢀ7.69 (m, 4H), 7.62

(d, J = 7.6 Hz, 2H), 7.38ꢀ7.36 (m, 2H), 7.17ꢀ7.14 (m, 2H), 4.17 (s,

2H). LC/MS(ESI) m/z 458 (M + H)⁺, t_R= 1.10 min.

N-(3-Aminophenylsulfonyl)-3-(3,4-dichlorobenzylthio)thiophene-

2-carboxamide (14). From 11 mg of 1, 5 mg (31%) of 14 was obtained.

¹H NMR (MeOH-d₄) δ: 7.79ꢀ7.77 (m, 1H), 7.54 (s, 1H), 7.45 (d, J =

8.4 Hz, 1H), 7.40ꢀ7.36 (m, 3H), 7.19ꢀ7.17 (m, 2H), 7.14ꢀ7.11 (m,

1H). LC/MS(ESI) m/z 473 (M + H)⁺, t_R= 0.99 min.

3-(3,4-Dibromobenzylthio)thiophene-2-carboxylic Acid (8). From

16 mg of 3-mercaptothiophene-2-carboxylic acid, 16.1 mg (40%) of 6

was obtained. ¹H NMR (MeOH-d₄) δ 7.78 (d, J = 2.2 Hz, 1H),

7.66ꢀ7.61 (m, 2H), 7.35 (dd, J = 2.2, 8.2 Hz, 1H), 7.11 (d, J = 5.1

Hz, 1H), 4.27 (s, 2H). LC/MS(ESI) m/z 407 (M + H)⁺, t_R= 0.94 min.

3-(3,4-Dichlorobenzyloxy)thiophene-2-carboxylic Acid (9). Com-

pound 9 was prepared by analogy to compounds 2ꢀ8, substituting

methyl 3-hydroxythiophene-2-carboxylate for 3-mercaptothiophene-2-

carboxylic acid as the starting material and utilizing 3,4-dichlorobenzyl

bromide as the alkylating agent. The methyl ester thus obtained was

hydrolyzed with 50% NaOH using the same procedure as described in

step 2 of the preparation of compound 10 to afford 7.8 mg of 9 (26%,

two steps). ¹H NMR (MeOH-d₄) δ 7.72 (s, 1H), 7.61 (d, J = 5.5 Hz,

5-(N-(3-(3,4-Dichlorobenzylthio)thiophene-2-carbonyl)sulfamoyl)-

2-methoxybenzoic Acid (15). Treatment of 16 (12 mg) with 3 M aq

LiOH (15 equiv) in MeOH provided 10.5 mg (94%) of 15. ¹H

NMR (DMSO-d₆) δ: 8.22 (s, 1H), 8.08 (dd, J = 2.2 Hz, 7.6 Hz, 1H),

7.92 (d, J = 5 Hz, 1H), 7.64 (br s, 1H), 7.53 (m, 1H), 7.36 (m,

2H), 7.23 (d, J = 5 Hz, 1H). LC/MS(ESI) m/z 532 (M + H)⁺, t_R=

0.82 min.

Methyl 5-(N-(3-(3,4-Dichlorobenzylthio)thiophene-2-carbonyl)-

sulfamoyl)-2-methoxybenzoate (16). From 57 mg of 1, 21.5 mg (23%)

of 16 was obtained. ¹H NMR (DMSO-d₆) δ: 8.45 (s, 1H), 8.22 (dd, J =

2.4 Hz, 8.8 Hz, 1H), 7.75 (d, J = 8.0 Hz, 1H), 7.38ꢀ7.34 (m, 3H), 7.16

(m, 2H), 4.15 (s, 2H), 3.99 (s, 3H), 3.92 (s, 3H). LC/MS(ESI) m/z 546

(M + H)⁺, t_R= 1.21 min.

6559

dx.doi.org/10.1021/jm2003766 |J. Med. Chem. 2011, 54, 6548–6562

Journal of Medicinal Chemistry

ARTICLE

2-(3,4-Dichlorobenzylthio)benzoic Acid (17). To a solution of

2-mercaptobenzoic acid (34.7 mg, 0.225 mmol) and 3,4-dichlorobenzyl

bromide (54 mg, 0.225 mmol) in 1.2 mL of DMSO was added 0.1 mL of

sodium methoxide (25 wt %, 0.45 mmol) at 25 °C. The resulting mixture

was stirred at 25 °C for 6 h, then concentrated in vacuo and purified by

preparative HPLC to afford 47 mg (67%) of 17. ¹H NMR (DMSO-d₆) δ

7.95 (dd, 1H, J = 1.5 Hz, J = 8.0 Hz), 7.61 (d, 1H, J = 1.5 Hz), 7.51ꢀ7.45

(m, 3H), 7.36ꢀ7.21 (m, 1H), 4.21 (s, 2H). LC/MS (ESI) m/z 311

(M ꢀ H)^ꢀ, t_R= 0.90 min.

dichlorobenzylthio)benzoic acid 19 (0.0375 mmol), 3 mg (18%) of 23

was obtained. H NMR (DMSO-d₆) δ 7.92 (d, J = 8.1 Hz, 1H), 7.77

(d, J = 2.2 Hz, 1H), 7.69 (t, J = 2.2 Hz, 1H), 7.63ꢀ7.56 (m, 2H), 7.55

(t, J = 2.2 Hz, 1H), 7.53 (s, 1H), 7.51ꢀ7.47 (m, 2H), 7.40 (dd, J =

1.5 Hz, 8 Hz, 1H), 4.32 (s, 2H). LC/MS(ESI) m/z 447 (M + H)⁺,

t_R= 1.44 min.

1

’ ASSOCIATED CONTENT

2-(3,4-Dichlorobenzylthio)-N-(methylsulfonyl)benzamide (18). Com-

pound 18 was prepared from compound 17 by analogy to the preparation

of compound 12. From 32 mg of 17, 33 mg (81%) of 18 was obtained.

¹H NMR (DMSO-d₆) δ: 7.62 (s, 1H), 7.57ꢀ7.55 (m, 2H), 7.47ꢀ7.45

(m, 2H), 7.35 (d, J = 8.3 Hz, 1H), 7.31ꢀ7.28 (m, 1H), 4.24 (s, 2H).

LC/MS(ESI) m/z 390 (M + H)⁺, t_R= 2.12 min.

S

Supporting Information. Procedures for the synthesis of

b

3-mercaptothiophene-2-carboxylic acid and compound 1; ﬁgure

showing impact of compounds 1 and 15 on the thermal stabil-

ity of RPTPγ; reﬁnement statistics for X-ray crystal struc-

tures (RPTPγ orthorhombic apo structure (PDB code 3QCB);

RPTPγ trigonal apo structure (3QCN); RPTPγ orthorhombic

vanadate complex (3QCC); RPTPγ trigonal vanadate complex

(3QCD); RPTPγ/1 complex (3QCE, soak, 3QCF, cocrystal);

RPTPγ/S1 complex (3QCG); RPTPγ/12 complex (3QCH);

RPTPγ/14 complex (3QCI); RPTPγ/15 complex (3QCJ);

RPTPγ/17 complex (3QCK); RPTPγ/20 complex (3QCL);

RPTPγ/S2 complex (3QCM)); ﬁgures showing X-ray cocrystal

structures with electron density for complexes of RPTPγ with

compounds 1, 12, 14, 15, 17, 20, and two additional analogues.

This material is available free of charge via the Internet at http://

pubs.acs.org.

4-Bromo-2-(3,4-dichlorobenzylthio)benzoic Acid (19). Compound

19 was prepared by analogy to the preparation of compound 17. From

4-bromo-2-mercaptobenzoic acid (387 mg, 1.66 mmol) and 3,4-dichlor-

obenzyl bromide (438 mg, 1.83 mmol), 480 mg (74%) of 20 was

obtained. ¹H NMR (MeOH-d₄) δ 7.86 (d, J = 8.4 Hz, 1H), 7.62 (d, J =

2.1 Hz, 1H), 7.53 (d, J = 2.2 Hz, 1H), 7.46 (d, J = 8 Hz, 1H), 7.38 (d, J =

2 Hz, 1H), 7.36 (d, J = 2 Hz, 1H), 4.21 (s, 2H). LC/MS(ESI) m/z 391

(M + H)⁺, t_R= 1.08 min.

2-(3,4-Dichlorobenzylthio)-4-(4-hydroxybut-1-ynyl)benzoic Acid

(20). A mixture of 4-bromo-2-(3,4-dichlorobenzylthio)benzoic acid 19

(14.7 mg, 0.0375 mmol), but-3-yn-1-ol (5.2 mg, 0.075 mmol), tetrakis-

(triphenylphosphine)palladium (1 mg, 0.00087 mmol), copper(II)

iodide (1 mg, 0.0053 mmol), and pyrrolidine (0.006 mL, 0.073 mmol)

in H₂O (1.5 mL) was heated to 70 °C for 90 min, then cooled to 25 °C,

extracted with EtOAc (3ꢁ), dried (Na₂SO₄), and concentrated in vacuo

to afford an oil which was purified by preparative HPLC to afford 8.8 mg

(62%) of 20 as a white solid. ¹H NMR (MeOH-d₄) δ 7.91 (d, J = 8.1 Hz,

1H), 7.61 (d, J = 2.2 Hz, 1H), 7.49ꢀ7.43 (m, 2H), 7.38 (d, J = 2.2 Hz,

1H), 7.36 (d, J = 2.2 Hz, 1H), 7.22 (d, J = 7.5 Hz, 1H), 4.20 (s, 2H), 3.75

(t, J = 6.6 Hz, 2H), 2.65 (t, J = 6.6 Hz, 2H). LC/MS (ESI) m/z 381(M +

H)⁺, t_R= 1.09 min.

Accession Codes

^†PDB codes of RPTPγ complexes are the following: 3QCB,

3QCC, 3QCD, 3QCE, 3QCF, 3QCG, 3QCH, 3QCI, 3QCJ,

3QCK, 3QCL, 3QCM, 3QCN.

’ AUTHOR INFORMATION

Corresponding Author

*Telephone: 609-252-5934. Fax: 609-252-6012. E-mail: steven.

sheriﬀ@bms.com.

4-((3-Aminophenyl)ethynyl)-2-(3,4-dichlorobenzylthio)benzoic

Acid (21). Compound 21 was prepared by analogy to the preparation of

compound 20 by using 1-amino-3-ethynylbenzene as the alkyne com-

ponent and triethylamine as the base. From 25 mg of 4-bromo-2-(3,4-

dichlorobenzylthio)benzoic acid 19 (0.064 mmol), 7.1 mg (26%) of 21

was obtained. ¹H NMR (DMSO-d₆) δ 7.90 (d, J = 8.0 Hz, 1H), 7.74 (d,

J = 2.0 Hz, 1H), 7.62 (d, J = 8.3 Hz, 1H), 7.54 (d, J = 1.3 Hz, 1H), 7.45

(dd, J = 8.4, 2.1 Hz, 1H), 7.35 (dd, J = 8.0, 1.5 Hz, 1H), 7.09 (d, J =

7.8 Hz, 1H), 6.80ꢀ6.71 (m, 2H), 6.70ꢀ6.62 (m, 1H), 4.33 (s, 2H).

4-((3-(2-Aminoacetamido)phenyl)ethynyl)-2-(3,4-dichloroben-

zylthio)benzoic Acid (22). A solution of 4-((3-aminophenyl)ethynyl)-

2-(3,4-dichlorobenzylthio)benzoic acid (21, 4 mg, 9.34 μmol), Boc-Gly-

OH (6.54 mg, 0.037 mmol), 2-(3H-[1,2,3]triazolo[4,5-b]pyridin-3-yl)-1,

1,3,3-tetramethylisouronium hexafluorophosphate(V) (HATU) (5.33

mg, 0.014 mmol), and N-ethyl-N-isopropylpropan-2-amine (3.02 mg,

0.023 mmol) in 1.2 mL of NMP was stirred at 25 °C for 16 h. The

product was isolated by preparative HPLC (MeOH/water with 0.1%

TFA) and treated with TFA (0.1 mL, 1.3 mmol) in DCM (0.1 mL) at

25 °C for 1 h. The solvent was evaporated in vacuo to give the product

4-((3-(2-aminoacetamido)phenyl)ethynyl)-2-(3,4-dichlorobenzylthio)-

benzoic acid (22) as a TFA salt (1.5 mg, 2.5 μmol, 30% yield over two

steps). ¹H NMR (MeOH-d₄) δ 7.97 (d, J = 8.0 Hz, 1H), 7.88 (s, 1H),

7.62 (d, J = 1.8 Hz, 1H), 7.58ꢀ7.51 (m, 2H), 7.47 (d, J = 8.3 Hz, 1H),

7.42ꢀ7.29 (m, 4H), 4.23 (s, 2H), 3.87 (s, 2H).

’ ACKNOWLEDGMENT

We thank Brian Carpenter for additional cloning work, Li Tao

for mass spectrometry, David Langley for generating an initial

model of RPTPγ prior to cloning and expression, and William

Metzler and James Bryson for initial construct design. Use of the

Advanced Photon Source (for the two vanadate structures and

the compound 1 soak) was supported by the U.S. Department of

Energy, Oﬃce of Science, Oﬃce of Basic Energy Sciences, under

Contract No. W-31-109-Eng-38. Use of the IMCA-CAT beam-

line 17-ID at the Advanced Photon Source was supported by the

companies of the Industrial Macromolecular Crystallography

Association through a contract with the Center for Advanced

Radiation Sources at the University of Chicago, IL.

’ ABBREVIATIONS USED

HTS, high throughput screen; PTP, protein tyrosine phospha-

tase; RPTPγ, receptor protein tyrosine phosphatase γ; T_m, melt-

ing temperature; TROSY, transverse relaxation optimized

spectroscopy; WPD, tryptophan, proline, aspartate; DCM, di-

chloromethane; DMF, dimethylformamide; DMSO, dimethyl-

sulfoxide; EtOAc, ethyl acetate; THF, tetrahydrofuran; EDC,

N-(3-dimethylaminopropyl)-N-ethylcarbodiimide; DMAP, 4-di-

methylaminopyridine

4-((3-Chlorophenyl)ethynyl)-2-(3,4-dichlorobenzylthio)benzoic

Acid (23). Compound 23 was prepared by analogy to the preparation of

compound 20 by using 1-chloro-3-ethynylbenzene as the alkyne com-

ponent and triethylamine as the base. From 14.7 mg of 4-bromo-2-(3,4-

6560

dx.doi.org/10.1021/jm2003766 |J. Med. Chem. 2011, 54, 6548–6562

Journal of Medicinal Chemistry

ARTICLE

’ REFERENCES

(21) Lamprianou, S.; Vacaresse, N.; Suzuki, Y.; Meziane, H.; Buxbaum,

J. D.; Schlessinger, J.; Harroch, S. Receptor Protein Tyrosine Phosphatase γ

Is a Marker for Pyramidal Cells and Sensory Neurons in the Nervous

System and Is Not Necessary for Normal Development. Mol. Cell. Biol.

2006, 26, 5106–5119.

(22) Brand~ao, 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.

(23) Barford, D.; Flint, A. J.; Tonks, N. K. Crystal Structure of Human

Protein Tyrosine Phosphatase 1B. Science 1994, 263, 1397–1404.

(24) Puius, Y. A.; Zhao, Y.; Sullivan, M.; Lawrence, D. S.; Almo,

S. C.; Zhang, Z.-Y. Identiﬁcation of a Second Aryl Phosphate-Binding

Site in Protein-Tyrosine Phosphatase 1B: A Paradigm for Inhibitor

Design. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 13420–13425.

(25) Berman, H. M.; Westbrook, J.; Feng, Z.; Gilliland, G.; Bhat,

T. N.; Weissig, H.; Shindyalov, I. N.; Bourne, P. E. The Protein Data

Bank. Nucleic Acids Res. 2000, 28, 235–242.

(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) 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.

(28) Appiah, K. K.; Kostich, W. A.; Gerritz, S. W.; Huang, Y.;

Hamman, B. D.; Allen, J.; Zhang, W.; Lanthorn, T. H.; Albright, C. F.;

Westphal, R.; Banks, M.; O’Connell, J. C. A High-Throughput Screen

for Receptor Protein Tyrosine Phosphatase-γ Selective Inhibitors. J. Biomol.

Screening 2011, 16, 476–485.

(29) Kish, K.; McDonnell, P. A.; Goldfarb, V.; Gao, M.; Metzler,

W. J.; Langley, D. R.; Bryson, J. W.; Kiefer, S. E.; Carpenter, B.; Kostich,

W. A.; Westphal, R. S.; Sheriﬀ, S. Cloning, Puriﬁcation, Crystallization

and Preliminary X-ray Analysis of the Catalytic Domain of Human

Receptor-like Protein Tyrosine Phosphatase γ in Three Diﬀerent

Crystal Forms. Acta Crystallogr., Sect. F: Struct. Biol. Cryst. Commun.

2011, 67, 768–774.

(1) Alonso, A.; Sasin, J.; Bottini, N.; Friedberg, I.; Friedberg, I.;

Osterman, A.; Godzik, A.; Hunter, T.; Dixon, J.; Mustelin, T. Protein

Tyrosine Phosphatases in the Human Genome. Cell 2004,

117, 699–711.

(2) Andersen, J. N.; Jansen, P. G.; Echwald, S. M.; Mortensen, O. H.;

Fukada, T.; Del Vecchio, R.; Tonks, N. K.; Moller, N. P. H. A Genomic

Perspective on Protein Tyrosine Phosphatases: Gene Structure,

Pseudogenes, and Genetic Disease Linkage. FASEB J. 2004, 18, 8–30.

(3) Bialy, L.; Waldmann, H. Inhibitors of Protein Tyrosine Phos-

phatases: Next-Generation Drugs? Angew. Chem., Int. Ed. 2005,

44, 3814–3839.

(4) Dewang, P. M.; Hsu, N.-M.; Peng, S.-Z.; Li, W.-R. Protein

Tyrosine Phosphatases and Their Inhibitors. Curr. Med. Chem. 2005,

12, 1–22.

(5) Hubbard, S. R.; Till, J. H. Protein Tyrosine Kinase Structure and

Function. Annu. Rev. Biochem. 2000, 69, 373–398.

(6) Tabernero, L.; Aricescu, A. R.; Jones, E. Y.; Szedlacsek, S. E.

Protein Tyrosine Phosphatases: StructureꢀFunction Relationships.

FEBS J. 2008, 275, 867–882.

(7) Tautz, L.; Pellecchia, M.; Mustelin, T. Targeting the PTPome in

Human Disease. Expert Opin. Ther. Targets 2006, 10, 157–177.

(8) Tonks, N. K. Protein Tyrosine Phosphatases: From Genes, to

Function, to Disease. Nat. Rev. Mol. Cell Biol. 2006, 7, 833–846.

€

(9) Ostman, A.; Hellberg, C.; B€ohmer, F. D. Protein-Tyrosine

Phosphatases and Cancer. Nat. Rev. Cancer 2006, 6, 307–320.

(10) Vintonyak, V. V.; Antonchick, A. P.; Rauh, D.; Waldmann, H.

The Therapeutic Potential of Phosphatase Inhibitors. Curr. Opin. Chem.

Biol. 2009, 13, 272–283.

(11) Combs, A. P. Recent Advances in the Discovery of Competitive

Protein Tyrosine Phosphatase 1B Inhibitors for the Treatment of

Diabetes, Obesity, and Cancer. J. Med. Chem. 2010, 53, 2333–2344.

(12) Jiang, Z.-X.; Zhang, Z.-Y. Targeting PTPs with Small Molecule

Inhibitors in Cancer Treatment. Cancer Metastasis Rev. 2008,

27, 263–272.

(13) Tam, S.; Saiah, E. Recent Advances in the Discovery and

Development of PTP-1B Inhibitors. Drugs Future 2008, 33, 175–185.

(14) Zhang, S.; Zhang, Z.-Y. PTP1B as a Drug Target: Recent

Developments in PTP1B Inhibitor Discovery. Drug Discovery Today

2007, 12, 373–381.

(15) Barnea, G.; Silvennoinen, O.; Shaanan, B.; Honegger, A. M.;

Canoll, P. D.; D’Eustachio, P.; Morse, B.; Levy, J. B.; Laforgia, S.;

Huebner, K.; Musacchio, J. M.; Sap, J.; Schlessinger, J. Identiﬁcation of a

Carbonic Anhydrase-like Domain in the Extracellular Region of RPTP

Gamma Deﬁnes a New Subfamily of Receptor Tyrosine Phosphatases.

Mol. Cell Biol. 1993, 13, 1497–1506.

(30) McDonnell, P. A.; Yanchunas, J.; Newitt, J. A.; Tao, L.; Kiefer,

S. E.; Ortega, M.; Kut, S.; Burford, N.; Goldfarb, V.; Duke, G. J.; Shen,

H.; Metzler, W.; Doyle, M.; Chen, Z.; Tarby, C.; Borzilleri, R.; Vaccaro,

W.; Gottardis, M.; Lu, S.; Crews, D.; Kim, K.; Lombardo, L.; Roussell,

D. L. Assessing Compound Binding to the Eg5 Motor Domain Using a

Thermal Shift Assay. Anal. Biochem. 2009, 392, 59–69.

(31) Copeland, R. A. Enzymes: A Practical Introduction to Structure,

Mechanism, and Data Analysis; Wiley: New York, 2000.

(16) LaForgia, S.; Morse, B.; Levy, J.; Barnea, G.; Cannizzaro, L. A.;

Li, F.; Nowell, P. C.; Boghosian-Sell, L.; Glick, J.; Weston, A.; Harris,

C. C.; Drabkin, H.; Patterson, D.; Croce, C. M.; Schlessinger, J.;

Huebner, K. Receptor Protein-Tyrosine Phosphatase Gamma Is a

Candidate Tumor Suppressor Gene at Human Chromosome Region

3p21. Proc. Natl. Acad. Sci. U.S.A. 1991, 88, 5036–5040.

(17) Liu, S.; Sugimoto, Y.; Kulp, S. K.; Jiang, J.; Chang, H.-L.; Park,

K.-Y.; Kashida, Y.; Lin, Y. C. Estrogenic Down-Regulation of Protein

Tyrosine Phosphatase γ (PTPγ) in Human Breast Is Associated with

Estrogen Receptor α. Anticancer Res. 2002, 22, 3917–3923.

(18) Liu, S.; Sugimoto, Y.; Sorio, C.; Tecchio, C.; Lin, Y. C. Function

Analysis of Estrogenically Regulated Protein Tyrosine Phosphatase γ

(PTP γ) in Human Breast Cancer Cell Line MCF-7. Oncogene 2004,

23, 1256–1262.

(19) Tsukamoto, T.; Takahashi, T.; Ueda, R.; Hibi, K.; Saito, H.;

Takahashi, T. Molecular Analysis of the Protein Tyrosine Phosphatase

γ Gene in Human Lung Cancer Cell Lines. Cancer Res. 1992, 52,

3506–3509.

(20) van Niekerk, C. C.; Poels, L. G. Reduced Expression of Protein

Tyrosine Phosphatase Gamma in Lung and Ovarian Tumors. Cancer Lett.

1999, 137, 61–73.

(32) Hoﬀmann, K. M. V.; Tonks, N. K.; Barford, D. The Crystal

Structure of Domain 1 of Receptor Protein-Tyrosine Phosphatase μ.

J. Biol. Chem. 1997, 272, 27505–27508.

(33) Collaborative Computational Project, Number 4. The CCP4

Suite: Programs for Protein Crystallography. Acta Crystallogr., Sect. D:

Biol. Crystallogr. 1994, 50, 760ꢀ763.

(34) Navaza, J. AMoRe: An Automated Package for Molecular

Replacement. Acta Crystallogr., Sect. A: Found. Crystallogr. 1994, 50,

157–163.

(35) Navaza, J.; Vernoslova, E. On the Fast Translation Functions

for Molecular Replacement. Acta Crystallogr., Sect. A: Found. Crystallogr.

1995, 51, 445–449.

(36) Almo, S. C.; Bonanno, J. B.; Sauder, J. M.; Emtage, S.;

Dilorenzo, T. P.; Malashkevich, V.; Wasserman, S. R.; Swaminathan,

S.; Eswaramoorthy, S.; Agarwal, R.; Kumaran, D.; Madegowda, M.;

Ragumani, S.; Patskovsky, Y.; Alvarado, J.; Ramagopal, U. A.; Faber-Barata,

J.; Chance, M. R.; Sali, A.; Fiser, A.; Zhang, Z.-y.; Lawrence, D. S.; Burley,

S. K. Structural Genomics of Protein Phosphatases. J. Struct. Funct. Genomics

2007, 8, 121–140.

(37) The PyMOL Molecular Graphics System, version 1.2; Schr€odinger,

LLC: New York, NY.

6561

dx.doi.org/10.1021/jm2003766 |J. Med. Chem. 2011, 54, 6548–6562

Journal of Medicinal Chemistry

ARTICLE

(38) Iversen, L. F.; Andersen, H. S.; Branner, S.; Mortensen, S. B;

Peters, G. H.; Norris, K.; Olsen, O. H.; Jeppersen, C. B.; Lundt, B. F.;

Ripka, W.; Møller, K. B.; Møller, N. P. H. Structure-Based Design of a

Low Molecular Weight, Nonphosphorus, Nonpeptide, and Highly

Selective Inhibitor of Protein-Tyrosine Phosphatase 1B. J. Biol. Chem.

2000, 275, 10300–10307.

(39) Fauman, E. B.; Yuvaniyama, C.; Schubert, H. L.; Stuckey, J. A.;

Saper, M. A. The X-ray Crystal Structures of Yersinia Tyrosine Phos-

phatase with Bound Tungstate and Nitrate: Mechanistic Implications.

J. Biol. Chem. 1996, 271, 18780–18788.

(40) Gray, C. H.; Good, V. M.; Tonks, N. K.; Barford, D. The

Structure of the Cell Cycle Protein Cdc14 Reveals a Proline-Directed

Protein Phosphatase. EMBO J. 2003, 22, 3524–3535.

(41) Xing, Y.; Xu, Y.; Chen, Y.; Jeﬀrey, P. D.; Chao, Y.; Lin, Z.; Li, Z.;

Strack, S.; Stock, J. B.; Shi, Y. Structure of Protein Phosphatase 2A Core

Enzyme Bound to Tumor-Inducing Toxins. Cell 2006, 127, 341–353.

(42) Chu, H.-M.; Wang, A. H.-J. EnzymeꢀSubstrate Interactions

Revealed by the Crystal Structures of the Archaeal Sulfolobus PTP-Fold

Phosphatase and Its Phosphopeptide Complexes. Proteins: Struct.,

Funct., Bioinf. 2007, 66, 996–1003.

(43) 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 Phospha-

tase 1B. Nat. Struct. Mol. Biol. 2004, 11, 730–737.

(44) Kamerlin, S. C. L.; Rucker, R.; Boresch, S. A Molecular

Dynamics Study of WPD-Loop Flexibility in PTP1B. Biochem. Biophys.

Res. Commun. 2007, 356, 1011–1016.

(45) Br€unger, A. T.; Adams, P. D.; Clore, G. M.; DeLano, W. L.;

Gros, P.; Grosse-Kunstleve, R. W.; Jiang, J.-S.; Kuszewski, J.; Nilges, M.;

Pannu, N. S.; Read, R. J.; Rice, L. M.; Simonson, T.; Warren, G. L.

Crystallography & NMR System: A New Software Suite for Macro-

molecular Structure Determination. Acta Crystallogr., Sect. D: Biol.

Crystallogr. 1998, 54, 905–921.

(46) Blanc, E.; Roversi, P.; Vonrhein, C.; Flensburg, C.; Lea, S. M.;

Bricogne, G. Reﬁnement of Severely Incomplete Structures with Max-

imum Likelihood in BUSTER-TNT. Acta Crystallogr., Sect. D: Biol.

Crystallogr. 2004, 60, 2210–2221.

(47) Emsley, P.; Cowtan, K. Coot: Model-Building Tools for

Molecular Graphics. Acta Crystallogr., Sect. D: Biol. Crystallogr. 2004,

60, 2126–2132.

(48) Emsley, P.; Lokhamp, B.; Scott, W. G.; Cowtan, K. Features and

development of Coot. Acta Crystallogr., Sect. D: Biol. Crystallogr. 2010,

66, 486–501.

(49) Carver, T. E.; Bordeau, B.; Cummings, M. D.; Petrella, E. C.;

Pucci, M. J.; Zawadzke, L. E.; Dougherty, B. A.; Tredup, J. A.; Bryson,

J. W.; Yanchunas, J., Jr.; Doyle, M. L.; Witmer, M. R.; Nelen, M. I.;

DesJarlais, R. L.; Jaeger, E. P.; Devine, H.; Asel, E. D.; Springer, B. A.;

Bone, R.; Salemme, F. R.; Todd, M. J. Decrypting the Biochemical

Function of an Essential Gene from Streptococcus pneumoniae Using

ThermoFluor Technology. J. Biol. Chem. 2005, 280, 11704–11712.

(50) Pantoliano, M. W.; Petrella, E. C.; Kwasnoski, J. D.; Lobanov,

V. S.; Myslik, J.; Graf, E.; Carver, T.; Asel, E.; Springer, B. A.; Lane, P.;

Salemme, F. R. High-Density Miniaturized Thermal Shift Assays as a

General Strategy for Drug Discovery. J. Biomol. Screening 2001,

6, 429–440.

6562

dx.doi.org/10.1021/jm2003766 |J. Med. Chem. 2011, 54, 6548–6562

Article Doi

Small molecule receptor protein tyrosine phosphatase γ (RPTPγ) ligands that inhibit phosphatase activity via perturbation of the tryptophan-proline-aspartate (WPD) loop

DOI: 10.1021/jm2003766

Source and publish data:

Authors:

Article abstract of DOI:10.1021/jm2003766

Full text of DOI:10.1021/jm2003766

Products guided by the article

R&D Labs maybe for 1336982-74-1

Relevant to this article

Hot Product