Selective inhibition of extracellular thioredoxin by asymmetric disulfides

Article abstract of DOI:10.1021/jm301775s

Whereas the role of mammalian thioredoxin (Trx) as an intracellular protein cofactor is widely appreciated, its function in the extracellular environment is not well-understood. Only few extracellular targets of Trx-mediated thiol-disulfide exchange are known. For example, Trx activates extracellular transglutaminase 2 (TG2) via reduction of an intramolecular disulfide bond. Because hyperactive TG2 is thought to play a role in various diseases, understanding the biological role of extracellular Trx may provide critical insight into the pathogenesis of these disorders. Starting from a clinical-stage asymmetric disulfide lead, we have identified analogs with >100-fold specificity for Trx. Structure-activity relationship and computational docking model analyses have provided insights into the features important for enhancing potency and specificity. The most active compound identified had an IC ₅₀ below 0.1 μM in cell culture and may be appropriate for in vivo use to interrogate the role of extracellular Trx in health and disease.

Full text of DOI:10.1021/jm301775s

Article
pubs.acs.org/jmc
Selective Inhibition of Extracellular Thioredoxin by Asymmetric
Disulfides
Thomas R. DiRaimondo,† Nicholas M. Plugis,‡ Xi Jin,‡ and Chaitan Khosla*,†,‡
†Department of Chemical Engineering and ‡Department of Chemistry, Stanford University, Stanford, California 94305, United States
S
* Supporting Information
ABSTRACT: Whereas the role of mammalian thioredoxin
(Trx) as an intracellular protein cofactor is widely appreciated,
its function in the extracellular environment is not well-
understood. Only few extracellular targets of Trx-mediated
thiol−disulfide exchange are known. For example, Trx activates
extracellular transglutaminase 2 (TG2) via reduction of an
intramolecular disulfide bond. Because hyperactive TG2 is
thought to play a role in various diseases, understanding the biological role of extracellular Trx may provide critical insight into
the pathogenesis of these disorders. Starting from a clinical-stage asymmetric disulfide lead, we have identified analogs with >100-
fold specificity for Trx. Structure−activity relationship and computational docking model analyses have provided insights into the
features important for enhancing potency and specificity. The most active compound identified had an IC50 below 0.1 μM in cell
culture and may be appropriate for in vivo use to interrogate the role of extracellular Trx in health and disease.
intramolecular disulfide bond of Trx is reduced by TrxR with
the concomitant consumption of one NADPH molecule.1
Thus, intracellular Trx acts as a protein cofactor for disulfide
bond reduction. In contrast, because the extracellular environ-
ment lacks TrxR or NADPH, secreted Trx is presumably a
stoichiometric reagent outside the cell. This singular difference
could be exploited in the design of small molecule inhibitors
that oxidize extracellular Trx via selective disulfide bond
exchange.
One such agent is 1 (PX-12, 1-methyl-1-propyl-2-imidazolyl
disulfide),15,16 which has undergone clinical evaluation in
cancer patients.17,18 1 oxidizes the active site cysteines of Trx
into a disulfide bond (Figure 1B).19 Using a combination of
novel assay development and structure−activity relationship
analysis, we have identified analogs with markedly improved
potency, selectivity, and activity in cell culture systems.
Additionally, a docking model was developed to rationalize
the observed improvements against human Trx.
INTRODUCTION
■
Mammalian thioredoxin-1 (Trx) is an archetypal protein
cofactor. In partnership with thioredoxin reductase (TrxR), it
plays a crucial role in many intracellular redox reactions and
regulatory processes.1 Additionally, mammalian cells are also
known to secrete Trx, although the mechanism and biological
implications of this phenomenon are poorly understood.1−5
For example, the TRPC ion channel6 and the HIV-1 envelope
protein gp1207 are activated by extracellular Trx. Selective small
molecule inhibitors could therefore be useful tools for decoding
the biology of extracellular Trx.
Our own interest in extracellular thioredoxin was motivated
by the related observations that, not only does this cofactor
activate transglutaminase 2 (TG2) via disulfide bond reduction
(Figure 1A) but also that interferon-γ (IFN-γ) induces Trx
secretion from monocytic cells in amounts that are sufficient to
activate TG2 in the extracellular matrix.8,9 These observations
are especially relevant to celiac disease pathogenesis, because
TG2 activation is required for post-translational modification of
antigenic gluten peptides in the small intestine10−12 and IFN-γ
is the principal inflammatory cytokine secreted by disease-
specific T cells.13 Together, these findings have led to the
hypothesis that extracellular Trx is the missing link in a
maladaptive, gluten-induced amplificatory loop between
inflammatory T cells and TG2 (Figure 2). Because Trx-null
mice show embryonic lethality,14 testing of this hypothesis
requires the engineering of small molecule inhibitors that
inactivate extracellular Trx to the exclusion of its intracellular
counterpart.
RESULTS
■
Design, Synthesis, and Biochemical Analysis of
Human Trx Inhibitors. The general synthesis of Trx
inhibitors is outlined in Scheme 1. Our starting point for
developing a pharmacologically useful inhibitor of extracellular
Trx was 1 (Table 1), a small molecule Trx inhibitor that has
entered human clinical trials in cancer patients.17,18
1
oxidatively inactivates extracellular Trx without appreciably
altering the levels of reduced intracellular Trx, because the latter
pool of Trx has access to TrxR, whereas the former does not.
Unfortunately, the compound has an extremely short half-life in
All Trx proteins have a highly conserved WCGPC catalytic
motif. The Cys residues of this motif are responsible for
reducing disulfide bonds in oxidized proteins. Inside the cell,
the overall catalytic cycle is completed when the resulting
Received: December 3, 2012
Published: January 17, 2013
© 2013 American Chemical Society
1301
dx.doi.org/10.1021/jm301775s | J. Med. Chem. 2013, 56, 1301−1310

Journal of Medicinal Chemistry
Article
Figure 1. (A) Thioredoxin (Trx) mediated activation of transglutaminase 2 (TG2). (B) Reaction of a small molecule asymmetric disulfide with
Trx.19
Figure 2. Proposed pathogenic role of Trx in celiac disease. In celiac disease patients, extracellular transglutaminase 2 (TG2) facilitates a robust T
cell (Th1) response to dietary gluten. Proteolytically resistant, Pro- and Gln-rich peptides derived from gluten accumulate in the gut lumen. By
themselves, these peptides are weak antigens; their affinity for HLA-DQ2 or DQ8 is markedly enhanced by TG2-catalyzed deamidation of selected
Gln residues. However, extracellular TG2 in the small intestinal mucosa is maintained in an inactive state via an intramolecular disulfide bond
between vicinal Cys residues. TG2 is activated by Th1-derived IFN-γ, which promotes Trx release. In turn, secreted Trx reduces the disulfide bond in
TG2 with high specificity. Thus, Trx is proposed as the missing link that mutually amplifies Th1 activity and TG2 activation in celiac disease, leading
to an inflammatory response.
a
aromatic functional groups could be readily accommodated
without penalty. We therefore synthesized and tested a series of
substituted benzothiazole derivatives 5−11. In general,
electron-withdrawing substituents showed a modest increase
in activity, whereas electron-donating groups were slightly
deleterious. The 6-Cl analog 7 had the highest specificity, with a
Trx/DTT reactivity ratio of 120.
Scheme 1
a
Reagents and conditions: (a) thiourea, HCl, H2O2, H2O/EtOH, 0
°C, 11−91%; (b) NaHCO3, MeOH, 58−93%.
Previous studies by Kirkpatrick et al. together with our LC−
MS data suggested that the thiol−disulfide exchange reaction
between 1 and Trx involves an initial attack that results in the
release of 2-mercaptoimidazole.19 To assess the influence of
alternative alkyl substituents on molecular recognition by
human Trx, we synthesized analogues of compound 14 bearing
alternative primary, secondary, and tertiary thiol substituents
(12−17, Table 1). The primary thiol 12 was most reactive
against Trx, while the tertiary thiol 13 was least reactive. The
reaction rates of secondary thiols did not vary significantly as a
function of size or conformation. Cyclic alkanes showed
moderately improved specificity, a feature that was verified
through the synthesis and evaluation of two other heterocycles,
18 and 19. Additionally, cycloalkanethiols would also have
reduced odor-related toxicity due to higher boiling points. For
these reasons, the cyclohexanethiol moiety was selected as the
preferred alkyl substituent for future analogs.
humans,18 which precludes derivation of a reliable correlation
between drug dose and pharmacological response. We have
confirmed this short half-life in our own preliminary studies in
rats (data not shown). Additionally, the release of sec-butyl thiol
from the Trx reaction (Figure 1B) induces a noxious response,
thereby limiting the maximum tolerable dose of 1.
To quantify the specificity of a small molecule disulfide for
human Trx, we established a kinetic assay to measure the
bimolecular rate constant (kinh/Ki) of this reaction. Reduced
dithiothreitol (DTT) was used as a representative control
capable of undergoing intramolecular dithiol → disulfide
oxidation. Representative data for 1 and DTT is shown in
Figure S1 of the Supporting Information. From these results, it
can be estimated that 1 has 7-fold specificity for human Trx
over DTT (Table 1).
An initial set of 1 analogs was synthesized in which both
halves of the asymmetric disulfide were systematically varied
(Table 1). Comparison of imidazolyl 1 versus thiazolyl 2
compounds suggested that the latter functional group was
moderately preferred. Similarly, comparison of the imidazolyl 1
and benzimidazolyl 14 compounds, or alternatively thiazolyl 2
and benzothiazolyl 5 compounds, suggested that polycyclic
To seek further improvement in specificity, an additional
series of compounds 20−30 (Table 1) was synthesized, and
promising new analogs were identified. Of greatest interest are
compounds 21 and 23, both of which show more than 20-fold
enhanced specificity than 1. Moreover, these efforts also yielded
1302
dx.doi.org/10.1021/jm301775s | J. Med. Chem. 2013, 56, 1301−1310

Journal of Medicinal Chemistry
Article
Table 1. Structures and Trx Inhibitory Activities of Various Asymmetric Disulfides
a series of compounds (e.g., 21, 23, 25, 26, 29, and 30) with
systematic differences in potency and specificity. It was
therefore anticipated that, not only would this compound set
provide insight regarding the most important physicochemical
determinants of biological activity, but it could also facilitate
development of a computational model for Trx−inhibitor
interactions.
Biological Evaluation of Trx Inhibitors. Our strategy for
designing a biological assay for Trx inhibitors that would be
informative in the context of celiac disease is summarized in
Figure 3. Our previous studies had shown that monolayers of
mature T84 enterocytic cells naturally express TG2 in their
extracellular matrix but that the enzyme is maintained in a
predominantly inactive state.20 We therefore sought to induce
extracellular TG2 activity in this cultured cell system by adding
recombinant human Trx (Figure 3A). TG2 activity was
quantified by the concomitant addition of 5-biotinamidopentyl-
amine (5BP), an amine substrate of TG2 that undergoes
covalent attachment to TG2 substrates in the extracellular
matrix (Figure 3B). As seen in Figure 3C, extracellular TG2 was
1303
dx.doi.org/10.1021/jm301775s | J. Med. Chem. 2013, 56, 1301−1310

Journal of Medicinal Chemistry
Article
Figure 3. An assay for Trx-mediated TG2 activation. (A) T84 enterocytic cells are grown in flat-bottom, 48-well plates for 7 days and then treated
with 0−10 μM Trx for a variable duration along with 200 μM 5BP for 3 h. Trx activates inactive extracellular TG2 (Figure 1). (B) Once activated,
TG2 cross-links a biotinylated amine, 5BP, to selected Gln residues in proteins such as fibronectin in the extracellular matrix. Using streptavidin
conjugated to horseradish peroxidase, the amount of 5BP incorporated is quantified as a measure of TG2 activity. (C) Trx-mediated TG2 activity
over 0−24 h, where the last 3 h were in the presence of 5BP. Maximal TG2 activity occurred within 2−4 h of Trx exposure. The maturity of the T84
monolayer did not significantly influence the result (Figure S2, Supporting Information). (D) At the 3 h time point, the EC50 of Trx-mediated TG2
activation was 25 3 nM. Addition of 25 μM TG2 inhibitor, ERW1041E, completely blocked Trx-mediated 5BP incorporation. Experiments were
performed in triplicate. Measurements are normalized to TG2 activity in the absence of Trx and are reported as average standard error.
Figure 4. Fluorescence microscopy of T84 monolayers treated with Trx. T84 enterocytic cells were treated with 0−10 μM Trx in the presence of
200 μM 5BP for 3 h with or without 25 μM of the selective TG2 inhibitor ERW1041E. Cultures were stained with antibodies against E-cadherin
(red) and TG2 (green) proteins, as well as with a streptavidin conjugate (blue). Images were obtained using a Zeiss LSM 510 Meta confocal
microscope with an oil immersion compatible 40× objective capable of digital imaging correlation. White bars in each panel correspond to 50 μm.
Each experiment was performed in triplicate. All images were processed in the same manner.
rapidly activated within an hour and reached peak activity 3 h
following Trx addition. Dose−response analysis revealed an
EC50 of 25 nM for Trx (Figure 3D), further reinforcing the
exquisite specificity of this protein−protein interaction. The
selectivity of the 5BP assay was verified by addition of
ERW1041E, a selective TG2 inhibitor,21 which fully suppressed
attachment of this amine to extracellular matrix protein (Figure
3D). Fluorescence microscopy was undertaken to verify the
1304
dx.doi.org/10.1021/jm301775s | J. Med. Chem. 2013, 56, 1301−1310

Journal of Medicinal Chemistry
Article
logic of this assay. As can be seen in Figure 4, TG2 is abundant
in the extracellular matrix of T84 cells. (Compare the
distribution of E-cadherin and TG2.) Whereas the amount of
TG2 protein (green) remains unchanged as a result of exposure
to Trx, its activity (blue) is induced in a dose-dependent
manner.
Using the above assay, the biological activity of selected Trx
inhibitors was quantified on the basis of their ability to block
5BP incorporation. Their IC50 values are summarized in Table
1 (see Figure S3 of the Supporting Information for titration
data). Of particular interest is compound 23, which has the
lowest IC50 value of 0.09 0.01 μM among all inhibitors tested.
This inhibitor is 24-fold more selective for Trx than 1, a value
that agrees well with biochemical data (Table 1).
Our results from the T84 enterocyte assay system also reveal
that the potency of an inhibitor in cell culture correlates both
with the absolute magnitude of its bimolecular rate constant
(kinh/Ki) as well as its specificity for Trx over DTT. For
example, compound 23 is more potent than the less reactive
but comparably specific compounds 21 or 25. Similarly,
compounds 1, 29, 26, and 25 are comparably reactive toward
DTT; they show progressive improvement in potency in a
manner that correlates well with Trx specificity. To verify that
DTT is indeed a representative nonspecific thiol, the reactivity
of selected inhibitors with DTT, cysteine, and glutathione was
compared (Table S1, Supporting Information). The specificity
of this class of asymmetric disulfides for Trx is evident from the
data summarized in Table S1 of the Supporting Information.
Computational and Mutagenic Analysis of Inhibitor−
Trx Interactions. Thioredoxin reduces disulfide bonds in
proteins via a bimolecular thiol−disulfide exchange mechanism.
The thiol−disulfide exchange reaction is a fundamental process
in biology and has therefore been extensively investigated via
experiment and theory.22−25 It is generally accepted that a
thiolate anion is the attacking species and that the reaction
proceeds via an SN2 transition state involving back-attack of the
disulfide bond by the thiolate (Figure 5A). The S−S bond
length in the reactant and product is 2.1 Å, while the S−S bond
length in the transition state is estimated to be slightly
elongated to 2.6−2.9 Å.25 Water molecules stabilize the
transition state by forming hydrogen bonds with partially
charged sulfur atoms.
To identify a putative binding site on Trx for compound 14,
we used the SwissDock server to derive a docking model. (See
Experimental Section for details.) Given that each asymmetric
disulfide consists of two rotamers that likely have differential
affinity and reactivity for Trx, both rotamers were docked using
the SwissDock server. Docking solutions were assessed on the
basis of three criteria. First, the distance between the Cys32
thiol in Trx and the sec-butyl thiol in compound 14 should be
less than 4.5 Å, which would allow thiol−disulfide exchange to
occur. Second, the angle of S−S···S(Cys32) atoms should be
greater than 135°, which allows back-attack of the incoming
thiolate. Third, Trx interacts predominantly with the
benzimidazolyl group and not significantly with the alkyl
group, as suggested by our SAR data.
Figure 5. A model for the binding of asymmetric disulfide 14 to
human Trx. (A) Thiol−disulfide SN2 exchange mechanism. Attack by
the thiolate leads to a linear S···S···S transition state. (B) Top view of
the inhibitor docking onto the active site of Trx. The benzene ring of
the benzimidazolyl group interacts with a tryptophan indole near the
Trx active site.
distance between benzimidazolyl moiety and the CH group of
Cys73 is only ∼3.4 Åclose enough for thiol−π inter-
action26,27and is thus consistent with recognition of the
aromatic rings of our best inhibitors by Trx. An alternate
docking model also satisfies the screening criteria (figure not
shown), in which the aromatic ring is oriented toward the
Trp31 residue and the sec-butyl group points in the opposite
direction. This model is less favored with a 145° S−S···S angle.
Furthermore, the former model is consistent with our
observation that compounds 5, 12, and 14, would react more
readily with Trx than compounds 2, 13, and 1, respectively, due
to interactions with residues in the vicinity of the Trx active
site.
To test the proposed docking model, a mutant W31A was
engineered in human Trx. This mutation was predicted to
reduce steric hindrance for the sec-butyl group of 14. The
insulin reduction assay revealed that the mutant had
comparable activity (78%) to wild-type Trx, a value consistent
with an analogous mutation in Escherichia coli Trx.28 However,
its reactivity toward compound 14 showed an 11-fold increase
(kinh/Ki = 1.42 μM−1 min−1 as compared to 0.13 μM−1 min−1).
Notwithstanding the results of the in silico model, it is possible
that both rotamers are responsible for activity in vitro, such that
the calculated kinh/Ki would be a composite value.
DISCUSSION AND CONCLUSIONS
■
Although the role of thioredoxin (Trx) in intracellular redox
homeostasis has been extensively investigated, the function of
this protein cofactor in the extracellular environment is poorly
understood.29 Recent observations have suggested that
activation of extracellular transglutaminase 2 (TG2) is a
major consequence of Trx secretion.9 TG2 is a ubiquitous
but tightly regulated multifunctional enzyme.30 Its aberrant
activity has been implicated in a number of pathological
conditions,31 the most well studied example being celiac
disease.32 Thus, a small molecule approach to interrogate the
role of extracellular Trx in celiac disease has the potential to
By screening hundreds of docking solutions, the most
favored one was identified and is shown in Figure 5B. Docking
the other rotamer of 14 in the Trx active site was unfavorable.
In this model, the distance between Cys32 thiol and sec-butyl
thiol is 3.7 Å, only 0.8 Å longer than the transition state. An
SN2 back-attack by Cys32 is also supported by a 160° S−S···S
angle. This model also accounts for our SAR data, namely, the
1305
dx.doi.org/10.1021/jm301775s | J. Med. Chem. 2013, 56, 1301−1310

Journal of Medicinal Chemistry
Article
as described by Kirkpatrick et al. with minor modification (Scheme
1).33 DMF was used as a cosolvent for aryl thiols with low solubility in
methanol. LC−MS and 1H NMR methods were employed to
determine the purity of the synthesized compounds. All new
compounds tested had purity at or above 95%.
provide fundamentally new insights into inflammatory biology.
In turn, small molecule Trx inhibitors could also become drug
leads due to their ability to indirectly block TG2 activity.
Two major challenges can be anticipated in the development
of a small molecule inhibition strategy for extracellular Trx
biology. First, an inhibitor must inactivate secreted Trx while
leaving intracellular Trx unperturbed. Second, unlike enzymes
that have well-defined pockets, Trx has a relatively shallow
active site pocket, making it more difficult to engineer highly
specific small molecule inhibitors. In this study, we addressed
both challenges, starting from a previously reported asymmetric
disulfide inhibitor, 1. SAR analysis of 1 suggested that half of
the molecule should ideally be an extended aromatic group with
electron-withdrawing substituents. These insights led to
substantial improvements over 1, both with respect to activity
and possibly also mitigation of malodorous dose-limiting
toxicity.
Cyclohexyl Carbamo(dithioperoxo)imidate Hydrochloride (A).
Cyclohexanethiol (7.9519 g, 65 mmol) and thiourea (3.806 g, 50
mmol) were dissolved in 22.5 mL of H2O and 65 mL of ethanol. The
flask was immersed in an ice−water bath, and concentrated
hydrochloric acid (5.5 mL) was carefully added. Hydrogen peroxide
(30%, 6.5 mL, ice-cold) was added dropwise over 30 min with
vigorous stirring, and the solution was continually stirred for another 3
h. The solvent was removed in vacuo and the residue was dissolved in
10 mL of ethanol, diluted with ether, and cooled overnight. Filtration
followed by recrystallization from ethanol−water yielded cyclohexyl
carbamo(dithioperoxo)imidate hydrochloride (10.3373 g, 91%) as
1
white crystals. H NMR (400 MHz, D2O): δ 0.96−1.28 (m, 5H),
1.41−1.44 (m, 1H), 1.58−1.61 (m, 2H), 1.80−1.83 (m, 2H), 2.83−
2.90 (m, 2H).
A novel biological assay that mimicked celiac disease
pathogenesis was developed to test candidate inhibitors. A
model was also developed to rationalize SAR data utilizing
computational simulation and site-directed mutagenesis; this
model suggested an important protein−small molecule
interaction at Trp31. Together, these accomplishments have
yielded a new class of chemical tools to study an emerging facet
of the biology of an important human protein.
Although the role of extracellular Trx in human biology and
pathophysiology is poorly understood, it appears to play a role
in allosterically regulating several extracellular proteins
including the TRPC ion channel, gp120 from HIV-1, and
transglutaminase 2 (TG2). Using the last of these targets to
devise a cell-based assay, we have engineered a class of
asymmetric disulfides to identify potent and selective inhibitors
of extracellular human Trx. Our chemical tools could be used to
understand the role of Trx in celiac disease, HIV, and diseases
of the synovial tissue.
sec-Butyl Carbamo(dithioperoxo)imidate Hydrochloride (B).
Compound B (121 mg, 0.60 mmol, 12%) was prepared by the
method used for A to yield a white solid. 1H NMR (400 MHz, D2O):
δ 0.77−0.81 (t, 3H, J = 7.2), 1.13−1.14 (d, 2H, J = 6.8), 1.35−1.55 (m,
2H), 2.86−2.93 (m, 1H).
Ethyl Carbamo(dithioperoxo)imidate Hydrochloride (C). Com-
pound C (330 mg, 1.9 mmol, 38%) was prepared by the method used
1
for A to yield a white solid. H NMR (400 MHz, D2O): δ 1.12−1.16
(t, 3H, J = 7.6), 2.70−2.75 (d, 2H, J = 7.2).
tert-Butyl Carbamo(dithioperoxo)imidate Hydrochloride (D).
Compound D (111 mg, 0.55 mmol, 11%) was prepared by the
method used for A to yield a white solid. 1H NMR (400 MHz, D2O):
δ 1.20 (s, 9H).
Cyclopentyl Carbamo(dithioperoxo)imidate Hydrochloride (E).
Compound E (160 mg, 0.76 mmol, 15%) was prepared by the method
used for A to yield a white solid. 1H NMR (400 MHz, D2O): δ 1.40−
1.59 (m, 4H), 1.81−1.90 (m, 2H), 3.29−3.36 (m, 1H).
Isopropyl Carbamo(dithioperoxo)imidate Hydrochloride (F).
Compound F (383 mg, 2.1 mmol, 41%) was prepared by the method
used for A to yield a white solid. 1H NMR (400 MHz, D2O): δ 1.14−
1.16 (d, 6H, J = 6.8), 3.06−3.16 (sep, 1H, J = 6.8).
EXPERIMENTAL SECTION
■
2-(sec-Butyldisulfanyl)-1H-imidazole (1). Commercially available
2-mercaptoimidazole (220 mg, 2.2 mmol) and 1-methyl-1-propylth-
ioisothiourea hydrochloride (366 mg, 2.5 mmol) were dissolved in 6
mL of methanol. Sodium bicarbonate (286 mg, 3.4 mmol) in 10 mL of
water was added dropwise with vigorous stirring at room temperature.
The resulting mixture was stirred for an additional 1 h and cooled
overnight at 4 °C. The crude product was collected by filtration and
recrystallized from an ethanol−water mixture to yield a white powder
Materials. All chemicals used in the syntheses described below
were from Sigma-Aldrich or Chemi-Block. DTT was from Invitrogen,
SDS−polyacrylamide gradient gels (5−20%) were from Bio-Rad, Ni-
NTA resin was from Qiagen, the HiTrap-Q anion exchange column
was from GE Healthcare, and the 7K MWCO spin columns were from
Pierce. T84 cells were from American Type Culture Collection
(ATCC). Tissue culture plates were from Corning Life Sciences. The
TG2 inhibitor ERW1041E [(S)-quinolin-3-ylmethyl 2-((((S)-3-
bromo-4,5-dihydroisoxazol-5-yl)methyl)carbamoyl)pyrrolidine-1-car-
boxylate] was synthesized as previously described.21 Tetramethylben-
zidine (TMB) ready mix substrate was from Sigma Aldrich.
Vectashield Mounting Media was purchased from Vector Laboratories.
Cell culture medium, fetal bovine serum, antibiotics, trypsin−EDTA,
sterile PBS, goat anti-rabbit (H-L) and goat anti-mouse (H-L)
secondary antibodies, streptavidin Alexa Fluor 647, and streptavidin−
HRP were from Invitrogen. Primary antibodies mouse anti-TG2
(TG100) and rabbit anti-E-cadherin were from Thermo Scientific and
Cell Signaling Technology, respectively.
1
(316 mg, 1.7 mmol, 76%). H NMR (400 MHz, CDCl3): δ 0.908−
0.945 (t, 3H, J = 7.4), 1.31 (d, 3H, J = 6.8), 1.499−1.607 (m, 1H),
1.682−1.792 (m, 1H), 2.910−2.994 (m, 1H), 7.08−7.11 (m, 1H),
9.45 (s, 1H). HRMS (Q-TOF MS ES+): m/z calcd for C7H12N2S2
188.0452, found 188.0451 (M + H)+.
2-(sec-Butyldisulfanyl)thiazole (2). Compound 2 (172 mg, 0.84
mmol, 84%) was prepared by the method used for 1 to yield a yellow
liquid. 1H NMR (400 MHz, DMSO): δ 0.911−0.948 (t, 3H, J = 7.4),
1.31 (m, 3H), 1.512−1.718 (m, 2H), 3.072−3.156 (m, 1H), 7.740 (s,
1H). HRMS (Q-TOF MS ES+): m/z calcd for C7H11NS3 205.0054,
found 205.0053 (M + H)+.
Preparation of Recombinant Human Thioredoxin. Trx was
expressed and purified in E. coli Rosetta 2/pCK11 as previously
described.9 Briefly, expression of Trx was induced by 0.2 mM IPTG at
18 °C for 12 h. The recombinant protein was then purified by Ni-NTA
and anion exchange chromatography at 4 °C with the presence of 1
mM DTT at all times. Purified Trx was then flash frozen and stored at
−80 °C. Trx mutants W31A and C73A were generated from pCK11
by site-directed mutagenesis, introduced to E. coli Rosetta 2, and
purified by the same procedure as the wild-type protein.
2-(sec-Butyldisulfanyl)pyridine (3). Compound 3 (172 mg, 0.86
mmol, 86%) was prepared by the method used for 1 to yield a yellow
1
oil. H NMR (300 MHz, DMSO): δ 0.95 (t, 3H), 1.24 (d, 3H), 1.57
(m, 2H), 2.98 (m, 1H), 7.25 (m, 1H), 7.76 (m, 2H), 8.46 (m, 1H).
HRMS (Q-TOF MS ES+): m/z calcd for C9H13NS2 199.0489, found
199.0491 (M + H)+.
2-(sec-Butyldisulfanyl)-3H-imidazo[4,5-c]pyridine (4). Compound
4 (304 mg, 1.27 mmol, 58%) was prepared by the method used for 1
1
to yield a gray solid. H NMR (300 MHz, CD3OD): δ 0.97 (t, 3H),
Synthesis of Disulfide Inhibitors. The alkyl asymmetrical
disulfides were synthesized by reacting substituted thioisothioureas
1.32 (d, 3H), 1.62 (m, 2H), 3.04 (m, 1H), 7.29 (t, 1H), 7.92 (d, 1H),
1306
dx.doi.org/10.1021/jm301775s | J. Med. Chem. 2013, 56, 1301−1310

Journal of Medicinal Chemistry
Article
8.32 (s, 1H). HRMS (Q-TOF MS ES+): m/z calcd for C10H13N3S2
239.0551, found 239.0552 (M + H)+.
6H), 3.25 (m, 1H), 7.24 (d, 2H), 7.51 (s, 2H). HRMS (Q-TOF MS
ES+): m/z calcd for C10H12N2S2 224.0442, found 224.0443 (M + H)+.
2-(Cyclopentyldisulfanyl)-1H-benzo[d]imidazole (16). Compound
16 (419 mg, 1.67 mmol, 76%) was prepared by the method used for 1
2-(sec-Butyldisulfanyl)benzo[d]thiazole (5). Compound 5 (494
mg, 1.94 mmol, 88%) was prepared by the method used for 1 to yield
a yellow liquid. 1H NMR (300 MHz, DMSO): δ 1.07 (t, 3H), 1.38 (d,
3H), 1.71 (m, 2H), 3.10 (m, 1H), 7.39 (m, 2H), 7.85 (m, 2H). HRMS
(Q-TOF MS ES+): m/z calcd for C11H13NS3 255.0210, found
255.02101 (M + H)+.
1
to yield a white solid. H NMR (300 MHz, CDCl3): δ 1.74 (m, 4H),
2.01 (m, 2H), 3.45 (m, 1H), 7.42 (s, 1H), 7.68 (s, 1H), 9.68 (s, 1H).
HRMS (Q-TOF MS ES+): m/z calcd for C12H14N2S2 250.0598, found
250.0600 (M + H)+.
2-(Cyclohexyldisulfanyl)-1H-benzo[d]imidazole (17). Compound
17 (360 mg, 1.36 mmol, 62%) was prepared by the method used for 1
to yield a white solid. 1H NMR (300 MHz, CD3OD): δ 1.32 (m, 6H),
1.78 (m, 2H), 2.03 (m, 2H), 3.00 (m, 1H), 7.24 (d, 2H), 7.51 (s, 2H).
HRMS (Q-TOF MS ES+): m/z calcd for C13H16N2S2 264.0755, found
264.0756 (M + H)+.
2-(sec-Butyldisulfanyl)-6-fluorobenzo[d]thiazole (6). Compound
6 (191 mg, 0.7 mmol, 87%) was prepared by the method used for 1 to
yield a yellow oil. 1H NMR (400 MHz, DMSO): δ 0.95−0.99 (t, 3H, J
= 7.6), 1.32−1.33 (d, 3H, J = 6.8), 1.52−1.75 (m, 2H), 3.14−3.23 (m,
1H), 7.32−7.37 (td, 1H, J = 2.4, 6.4), 7.83−7.87 (dd, 1H, J = 4.8, 4),
7.97−8.00 (dd, 1H, J = 2.8, 6).
2-(Cyclohexyldisulfanyl)benzo[d]thiazole (18). Compound 18
2-(sec-Butyldisulfanyl)-6-chlorobenzo[d]thiazole (7). Compound
7 (168 mg, 0.58 mmol, 92%) was prepared by the method used for 1
(2.047 g, 7.3 mmol, 90%) was prepared by the method used for 1
1
1
to yield a yellow oil. H NMR (400 MHz, DMSO): δ 1.11−1.47 (m,
to yield a yellow oil. H NMR (400 MHz, CDCl3): δ 0.95−0.98 (t,
5H), 1.50−1.57 (m, 1H), 1.67−1.74 (m, 2H), 1.98−2.04 (m, 2H),
3.12−3.19 (m, 1H), 7.36−7.41 (m, 1H), 7.44−7.49 (m, 1H), 7.81−
7.83 (m, 1H), 8.03−8.05 (m, 1H).
3H, J = 7.6), 1.31−1.33 (d, 3H, J = 6.8), 1.54−1.75 (m, 2H), 3.15−
3.24 (m, 1H), 7.49−7.52 (dd, 1H, J = 2.4, 6.4), 7.81−7.83 (d, 1H, J =
8.0), 8.21 (d, 1H, J = 2.4).
2-(Cyclohexyldisulfanyl)benzo[d]oxazole (19). Compound 19
(116 mg, 0.44 mmol, 87%) was prepared by the method used for 1
to yield a red oil. 1H NMR (400 MHz, DMSO): δ 1.12−1.43 (m, 5H),
1.50−1.57 (m, 1H), 1.67−1.73 (m, 2H), 1.96−2.02 (m, 2H), 3.13−
3.20 (m, 1H), 7.34−7.41 (m, 2H), 7.67−7.75 (m, 2H). HRMS (Q-
TOF MS ES+): m/z calcd for C13H16N2S2 265.0595, found 265.0597
(M + H)+.
2-(sec-Butyldisulfanyl)-6-iodobenzo[d]thiazole (8). Compound 8
(153 mg, 0.40 mmol, 80%) was prepared by the method used for 1 to
yield a yellow oil. 1H NMR (400 MHz, DMSO): δ 0.95−0.98 (t, 3H, J
= 7.2), 1.31−1.33 (d, 3H, J = 6.8), 1.52−1.75 (m, 2H), 3.15−3.23 (m,
1H), 7.60−7.62 (d, 1H, J = 8.4), 7.75−7.78 (dd, 1H, J = 1.6, 6.8), 8.48
(d, 1H, J = 1.6).
4-Bromo-2-(sec-butyldisulfanyl)benzo[d]thiazole (9). Compound
9 (193 mg, 0.58 mmol, 72%) was prepared by the method used for 1
2-(sec-Butyldisulfanyl)-6-chloro-5-fluoro-1H-benzo[d]imidazole
(20). Compound 20 (87 mg, 0.30 mmol, 60%) was prepared by the
1
to yield an orange oil. H NMR (400 MHz, DMSO): δ 0.96−0.99 (t,
1
method used for 1 to yield an off-white solid. H NMR (400 MHz,
3H, J = 7.2), 1.33−1.34 (d, 3H, J = 6.8), 1.55−1.76 (m, 2H), 3.17−
3.23 (m, 1H), 7.29−7.33 (t, 1H, J = 8), 7.70−7.72 (dd, 1H, J = 0.8,
6.8), 8.06−8.08 (dd, 1H, J = 0.8, 7.2). HRMS (Q-TOF MS ES+): m/z
calcd for C11H12BrNS3 332.9315, found 332.9313 (M + H)+.
CDCl3): δ 0.96−1.00 (t, 3H, J = 7.6), 1.32−1.34 (d, 3H, J = 6.8),
1.52−1.63 (m, 1H), 1.70−1.81 (m, 1H), 2.95−3.04 (m, 1H), 7.33−
7.35 (d, 1H, J = 9.2), 7.57−7.58 (d, 1H, J = 6.4). HRMS (Q-TOF MS
ES+): m/z calcd for C13H14ClFN2S2 316.0271, found 316.0271 (M +
H)+.
6-Chloro-2-(cyclohexyldisulfanyl)-5-fluoro-1H-benzo[d]imidazole
(21). Compound 21 (105 mg, 0.33 mmol, 66%) was prepared by the
method used for 1 to yield a beige solid. 1H NMR (400 MHz,
DMSO): δ 1.10−1.38 (m, 5H), 1.51−1.56 (m, 1H), 1.66−1.71 (m,
2H), 1.93−1.98 (m, 2H), 3.03−3.10 (m, 1H), 7.50−7.53 (d, 1H, J =
9.6), 7.66−7.68 (d, 1H, 6.8). HRMS (Q-TOF MS ES+): m/z calcd for
C11H12ClFN2S2 290.0115, found 290.0118 (M + H)+.
5-Bromo-2-(sec-butyldisulfanyl)benzo[d]thiazole (10). Com-
pound 10 (156 mg, 0.47 mmol, 59%) was prepared by the method
1
used for 1 to yield an orange oil. H NMR (400 MHz, DMSO): δ
0.95−0.99 (t, 3H, J = 7.2), 1.32−1.33 (d, 3H, J = 6.8), 1.54−1.75 (m,
2H), 3.18−3.23 (m, 1H), 7.54−7.57 (dd, 1H, J = 1.6, 6.8), 8.02−8.05
(m, 2H). HRMS (Q-TOF MS ES+): m/z calcd for C11H12BrNS3
332.9315, found 332.9317 (M + H)+.
2-(sec-Butyldisulfanyl)-6-nitrobenzo[d]thiazole (11). Compound
11 (132 mg, 0.44 mmol, 88%) was prepared by the method used for 1
to yield a yellow solid. 1H NMR (400 MHz, DMSO): δ 0.97−1.00 (t,
3H, J = 7.6), 1.34−1.36 (d, 3H, J = 6.8), 1.57−1.77 (m, 2H), 3.22−
3.30 (m, 1H), 7.99−8.01 (d, 1H, J = 8.8), 8.31 (dd, 1H, J = 2.8, 6.4),
9.12 (d, 1H, J = 2.4). HRMS (Q-TOF MS ES+): m/z calcd for
C11H12N2O2S3 300.0061, found 300.0062 (M + H)+.
2-(sec-Butyldisulfanyl)-5-nitro-1H-benzo[d]imidazole (22). Com-
pound 22 (507 mg, 1.79 mmol, 81%) was prepared by the method
1
used for 1 to yield a yellow solid. H NMR (400 MHz, CDCl3): δ
0.98−1.02 (t, 3H, J = 7.2), 1.35−1.36 (d, 3H, J = 6.8), 1.55−1.67 (m,
1H), 1.73−1.83 (m, 1H), 2.99−3.07 (m, 1H), 7.48−7.70 (dd, 1H, J =
8.4, 73), 8.18−8.21 (dd, 1H, J = 2, 6.8), 8.37−8.55 (d, 1H, J = 70),
10.14 (s, 1H). HRMS (Q-TOF MS ES+): m/z calcd for
C11H13N3O2S2 283.0449, found 283.0453 (M + H)+.
2-(Cyclohexyldisulfanyl)-5-nitro-1H-benzo[d]imidazole (23).
Compound 23 (2.310 g, 7.47 mmol, 93%) was prepared by the
method used for 1 to yield a brown solid. 1H NMR (400 MHz,
DMSO): δ 1.09−1.39 (m, 5H), 1.49−1.56 (m, 1H), 1.65−1.71 (m,
2H), 1.94−2.00 (m, 2H), 3.06−3.14 (m, 1H), 7.64−7.66 (d, 1H, J =
8.8), 8.07−8.10 (dd, 1H, J = 2.4, 6.8), 8.35 (d, 1H, J = 2). HRMS (Q-
TOF MS ES+): m/z calcd for C13H15N3O2S2 309.0606, found
309.0606 (M + H)+.
2-(Cyclohexyldisulfanyl)-5-ethoxy-1H-benzo[d]imidazole (24).
Compound 24 (130 mg, 0.42 mmol, 85%) was prepared by the
method used for 1 to yield a black oil. 1H NMR (400 MHz, DMSO): δ
1.10−1.38 (m, 8H), 1.48−1.57 (m, 1H), 1.63−1.73 (m, 2H), 1.93−
2.00 (m, 2H), 3.01−3.08 (m, 1H), 3.98−4.03 (m, 2H), 6.76−6.79 (dd,
1H, J = 2.4, 6.4), 6.96 (s, 1H), 7.35−7.37 (d, 1H, J = 8). HRMS (Q-
TOF MS ES+): m/z calcd for C15H20N2OS2 308.1017, found
308.1023 (M + H)+.
Phenyl(2-thioxo-2,3-dihydro-1H-benzo[d]imidazol-5-yl)-
methanone. In a synthesis adapted from Okolotowicz et al.,34 (3,4-
diaminophenyl)(phenyl)methanone (743 mg, 3.5 mmol) and
potassium ethyl xanthate (1074 mg, 6.7 mmol) were dissolved in
2-(Ethyldisulfanyl)-1H-benzo[d]imidazole (12). Compound 12
(326 mg, 1.55 mmol, 70%) was prepared by the method used for 1
1
to yield a white solid. H NMR (300 MHz, CDCl3): δ 1.37 (t, 3H),
2.88 (m, 2H), 7.42 (m, 1H), 7.68 (m, 1H), 9.78 (s, 1H). HRMS (Q-
TOF MS ES+): m/z calcd for C9H10N2S2 210.0285, found 210.0286
(M + H)+.
2-(tert-Butyldisulfanyl)-1H-benzo[d]imidazole (13). Compound
13 (299 mg, 1.25 mmol, 60%) was prepared by the method used
1
for 1 to yield a white solid. H NMR (300 MHz, CD3OD): δ 1.39 (s,
9H), 7.50 (m, 2H). HRMS (Q-TOF MS ES+): m/z calcd for
C11H14N2S2 238.0598, found 238.0600 (M + H)+.
2-(sec-Butyldisulfanyl)-1H-benzo[d]imidazole (14). Compound 14
(446 mg, 1.9 mmol, 85%) was prepared by the method used for 1 to
1
yield a white solid. H NMR (400 MHz, CDCl3): δ 0.963−0.999 (t,
3H, J = 7.2), 1.328−1.345 (d, 3H, J = 6.8), 1.525−1.634 (m, 1H),
1.720−1.826 (m, 1H), 2.937−3.021 (m, 1H), 7.216−7.254 (m, 2H),
7.394−7.439 (m, 1H), 7.647−7.692 (m, 1H), 9.57 (s, 1H). HRMS
(Q-TOF MS ES+): m/z calcd for C11H14N2S2 238.0598, found
238.0600 (M + H)+.
2-(Isopropyldisulfanyl)-1H-benzo[d]imidazole (15). Compound
15 (314 mg, 1.40 mmol, 64%) was prepared by the method used
for 1 to yield a white solid. 1H NMR (300 MHz, CD3OD): δ 1.34 (d,
1307
dx.doi.org/10.1021/jm301775s | J. Med. Chem. 2013, 56, 1301−1310

Journal of Medicinal Chemistry
Article
1.5 mL of H2O and 8 mL of ethanol and refluxed for 3 h. Activated
carbon was added and the solution was refluxed for 20 min and then
filtered. Ten milliliters of H2O was added to the filtrate, followed by
1.5 mL of 50% acetic acid. The suspension was cooled overnight at 4
°C and dried in vacuo, yielding the title compound (641 mg, 2.5
mmol, 72%) as an orange solid. H NMR (400 MHz, DMSO): δ
7.24−7.26 (d, 1H, J = 8.4), 7.45−7.46 (d, 1H, J = 1.2), 7.53−7.57 (m,
3H), 7.63−7.70 (m, 3H), 12.82 (s, 2H).
(2-(Cyclohexyldisulfanyl)-1H-benzo[d]imidazol-6-yl)(phenyl)-
methanone (25). Compound 25 (88 mg, 0.24 mmol, 72%) was
prepared by the method used for 1 to yield an orange solid. 1H NMR
(400 MHz, DMSO): δ 1.11−1.41 (m, 5H), 1.52−1.57 (m, 1H), 1.66−
1.72 (m, 2H), 1.95−2.01 (m, 2H), 3.06−3.13 (m, 1H), 7.53−7.74 (m,
8H), 7.84 (s, 1H). HRMS (Q-TOF MS ES+): m/z calcd for
C20H20N2OS2 368.1017, found 368.1021 (M + H)+.
2-Amino-8-(cyclohexyldisulfanyl)-7H-purin-6-ol (26). Compound
26 (74 mg, 0.25 mmol, 71%) was prepared by the method used for 1
to yield an off-white solid. 1H NMR (400 MHz, DMSO): δ 1.13−1.42
(m, 5H), 1.52−1.56 (m, 1H), 1.61−1.73 (m, 2H), 1.85−1.98 (m, 2H),
2.96−3.05 (m, 1H), 6.41−6.50 (m, 2H), 10.61 (s, 1H), 12.85 (s, 1H).
HRMS (Q-TOF MS ES+): m/z calcd for C11H15N5OS2 297.0718,
found 297.0720 (M + H)+.
initial inhibitor concentration, [E] is the thioredoxin concentration,
and k is kinh/Ki for Trx or ki for DTT, glutathione, or cysteine.
Insulin Turbidity Assay. The activities of wild-type and the W31A
mutant of human Trx were measured by insulin turbidity assay as
previously described with slight modification.35 Briefly, 4 μM Trx was
added into a mixture of 160 μM bovine insulin (Km = 30 μM for WT
Trx) and 1 mM DTT in 0.1 M phosphate−citrate buffer, pH 6.0.
Formation of the reduced insulin was monitored by following the
appearance of turbidity at 650 nm due to precipitation. The
regeneration of Trx by DTT is much faster than direct reduction of
insulin by DTT. Previous experiments have shown that ΔA650nm/time
is linear over the Trx concentration 0−8 μM (data not shown).
Therefore, the activity of the W31A mutant was estimated by
normalizing its ΔA650nm/time to that of wild-type Trx.
Docking Models. Docking between the structurally characterized
reduced human Trx (PDB 1ERT) and compound 14 was performed
on the Swiss Dock server.36 The coordinate file 1ERT was edited to
remove all water molecules. Compound 14 was optimized by MM2 in
Chem3D. The ligand-binding site was constrained in a 10 × 10 × 10
Å3 cube with (x_31, y_12.5, z_4.5) as the center. Flexible side chains
were allowed. All rotatable single bonds were allowed to rotate in the
ligand. Docking results were screened by Chimera.37
1
Cell Culture. T84 epithelial cells were grown in Dulbecco’s
modified Eagle’s medium/Ham’s F-12 (1:1) supplemented with
antibiotics (penicillin/streptomycin) and 5% (v/v) fetal bovine
serum. Cells are grown at 37 °C and 5% CO2. Medium was changed
every alternate day. When the cells reached greater than 90%
confluency, they were passaged using trypsin−EDTA. For the TG2
activity assays, T84 enterocytes were grown in standard flat-bottom
48-well plates. For fluorescent microscopy experiments, T84 enter-
ocytes were grown on collagen-coated permeable supports (5 μm pore
size, 6.5 mm in diameter) as described previously.20 T84 cells were
seeded at 3 × 104 cells/well in all experiments.
8-(Cyclohexyldisulfanyl)-7H-purin-6-amine (27). Compound 27
(69 mg, 0.24 mmol, 70%) was prepared by the method used for 1 to
yield an orange solid. 1H NMR (400 MHz, DMSO): δ 1.11−1.42 (m,
5H), 1.50−1.57 (m, 1H), 1.63−1.73 (m, 2H), 1.93−2.02 (m, 2H),
3.00−3.09 (m, 1H), 7.20 (s, 2H), 8.07 (s, 1H), 13.25 (s, 1H). HRMS
(Q-TOF MS ES+): m/z calcd for C11H15N5S2 281.0769, found
281.0772 (M + H)+.
2-(Cyclohexyldisulfanyl)-4H-benzo[d][1,3]thiazine (28). Com-
pound 28 (59 mg, 0.20 mmol, 79%) was prepared by the method
used for 1 to yield a brown oil. 1H NMR (400 MHz, DMSO): δ 1.13−
1.43 (m, 5H), 1.50−1.57 (m, 1H), 1.63−1.73 (m, 2H), 1.93−2.02 (m,
2H), 2.98−3.06 (m, 1H), 4.11 (s, 2H), 7.15−7.36 (m, 4H).
2-(Cyclohexyldisulfanyl)-5-phenyl-1H-imidazole (29). Compound
29 (115 mg, 0.40 mmol, 79%) was prepared by the method used for 1
to yield an off-white solid. 1H NMR (400 MHz, DMSO): δ 1.14−1.39
(m, 5H), 1.52−1.57 (m, 1H), 1.65−1.72 (m, 2H), 1.94−2.02 (m, 2H),
3.00−3.07 (m, 1H), 7.18−7.22 (t, 1H, J = 7.2), 7.33−7.37 (t, 2H, J =
7.6), 7.72−7.74 (d, 2H, J = 7.6). HRMS (Q-TOF MS ES+): m/z calcd
for C15H18N2S2 290.0911, found 290.0914 (M + H)+.
3-(Cyclohexyldisulfanyl)-5-phenyl-4H-1,2,4-triazol-4-amine (30).
Compound 30 (119 mg, 0.39 mmol, 78%) was prepared by the
method used for 1 to yield a white solid. 1H NMR (400 MHz,
DMSO): δ 1.16−1.42 (m, 5H), 1.50−1.55 (m, 1H), 1.68−1.74 (m,
2H), 1.95−2.01 (m, 2H), 3.10−3.17 (m, 1H), 6.22 (s, 2H), 7.50−7.55
(m, 3H), 8.00−8.04 (m, 2H). HRMS (Q-TOF MS ES+): m/z calcd
for C14H18N4S2 306.0973, found 306.0975 (M + H)+.
In Vitro Thioredoxin Activity Assay. Before use, recombinant
human Trx was freshly reduced with a 10-fold molar excess of DTT on
ice. The excess DTT was removed by passing the solution through a
7K MWCO spin column. Trx concentration was determined by A280nm
(ε = 7570 M−1 cm−1), and the protein was used freshly within 2 h.
Prereduced Trx was reacted with disulfide inhibitors in 0.1 M
phosphate−citrate buffer, pH 6.0, at a concentration of 10 μM each at
room temperature.
Thioredoxin-Mediated TG2 Activity Assay. For measurement
of TG2 activity, T84 monolayers were grown in standard 48-well flat-
bottom plates for 1−14 days (Supporting Information, Figure S2) as
described above. Recombinant human thioredoxin was reduced using a
10 molar excess of DTT, buffer exchanged into T84 cell culture
medium, and used fresh as described in the in vitro thioredoxin activity
assay procedure above. Thioredoxin was diluted into fresh medium
and added to the T84 cells for 1−24 h (Figure 3). For the last 3 h of
thioredoxin exposure, 200 μM 5BP was spiked into the medium such
that each well was exposed to 5BP for a total of 3 h. For thioredoxin
treatments less than 3 h, 5BP was spiked into wells prior to
thioredxoin exposure. As a control, TG2 activity was inhibited with 25
μM ERW1041E for the duration of 5BP exposure. After 5BP exposure,
T84 monolayers were washed with warm PBS three times and fixed
with 4% (w/v) paraformaldehyde in PBS for 15 min at room
temperature. Following three washes with PBS, monolayers were
blocked overnight at 4 °C using 5% (w/v) BSA in PBS supplemented
with 0.1% (v/v) Tween-20. Then, monolayers were treated with
horseradish peroxidase conjugated streptavidin, diluted according to
manufacturer’s recommendations, in blocking buffer overnight at 4 °C.
Cells were washed four times with PBS + 0.1% Tween-20. A ready mix
solution of tetramethylbenzidine (TMB) substrate was added to the
transwell chambers for 5−10 min, after which 100 μL was removed
and quenched in equal volume of 1 M hydrochloric acid in a 96-well
plate format. Absorbance at 450 nm was measured using a 96-well
plate reader. The fold-change in 5BP incorporation was normalized to
the 0 μM thioredoxin exposure time. Each set of conditions was
assayed in at least triplicate wells.
Reduction of the inhibitor disulfide was monitored spectrophoto-
metrically on the basis of an increase in absorbance at 252 nm due to
the release of 2-mercaptoimidazole (Δε = 9400 M−1 cm−1) or
corresponding substituents. The time-dependent absorption curve was
recorded until no significant change was observed. From this data, the
rate of 2-mercaptoimidazole and oxTrx formation was calculated on
the basis of the above Δε value. In a separate measurement, 250 μM
DTT was reacted with 10 μM of each inhibitor to calculate
bimolecular rate constants for the reaction between the inhibitor
and a reference thiol reducing agent. The resulting time-dependent
curves were fitted to the following equation to yield kinh/Ki or ki
values: [P] = [I]0(1 − e−k[E]t), where [P] is the accumulated
concentration of the aromatic reaction product at time t, [I]0 is the
Fluorescent Microscopy. T84 cells were grown to maturity on
permeable supports, as described previously.20 For fluorescence
staining experiments, recombinant human thioredoxin was reduced
using a 10 molar excess of DTT, buffer exchanged into T84 cell culture
medium, and used fresh as described in the in vitro thioredoxin activity
assay procedure above. Thioredoxin was diluted into fresh medium
containing 200 μM 5-biotinamido pentylamine (5BP) and added to
both top and bottom compartments of the T84 transwells for 3 h. As a
control, TG2 activity was inhibited with 25 μM ERW1041E for the
1308
dx.doi.org/10.1021/jm301775s | J. Med. Chem. 2013, 56, 1301−1310

Journal of Medicinal Chemistry
Article
duration of thioredoxin and 5BP exposure. After the 3 h thioredoxin
and 5BP incubation, monolayers were washed with warm PBS three
times and fixed with 4% (w/v) paraformaldehyde in PBS for 15 min.
Cells were then washed three times with PBS and blocked overnight at
4 °C using 5% (w/v) BSA in PBS supplemented with 0.1% (v/v)
Tween-20. Primary antibodies, rabbit anti-E-cadherin IgG mAb (1:200
dilution in blocking buffer) and mouse anti-TG2 IgG mAb (1:200
dilution in blocking buffer), were used to label E-cadherin and TG2,
respectively, in an overnight incubation at 4 °C. E-cadherin was used
as a marker of cell−cell contacts in the enterocyte monolayer. Cells
were washed three times with PBS + 0.1% (v/v) Tween-20 and
incubated once again overnight at 4 °C with secondary antibodies,
goat anti-rabbit (H-L) IgG Alexa Fluor 555 conjugate (2 μg/mL in
blocking buffer) and goat anti-mouse (H-L) IgG Alexa Fluor 488
conjugate (2 μg/mL in blocking buffer). To visualize TG2 activity,
streptavidin Alexa Fluor 647 conjugate (2 μg/mL in blocking buffer)
was also added. Subsequently, cells were washed four times with PBS +
0.1% (v/v) Tween-20. Permeable supports were cut from their frames
with a sterile razor blade and mounted on slides using Vectashield
mounting media. Coverslips were sealed onto the mounting slides with
multiple coats of clear nail polish. Sealed slides were stored in the dark
at 4 °C and imaged using a Zeiss LSM 510 Meta Confocal
Microscope, Zeiss Immersol 518F oil, and a 40× Plan-Neo/1.3 NA
oil objective lens with digital image correlation (DIC) capability.
(5) Nakamura, H. Thioredoxin as a key molecule in redox signaling.
Antioxid. Redox Signal. 2004, 6, 15−17.
(6) Xu, S.-Z.; Sukumar, P.; Zeng, F.; Li, J.; Jairaman, A.; English, A.;
Naylor, J.; Ciurtin, C.; Majeed, Y.; Milligan, C. J.; Bahnasi, Y. M.; Al-
Shawaf, E.; Porter, K. E.; Jiang, L.-H.; Emery, P.; Sivaprasadarao, A.;
Beech, D. J. TRPC channel activation by extracellular thioredoxin.
Nature 2008, 451, 69−72.
(7) Azimi, I.; Matthias, L. J.; Center, R. J.; Wong, J. W. H.; Hogg, P. J.
Disulfide bond that constrains the HIV-1 gp120 V3 domain is cleaved
by thioredoxin. J. Biol. Chem. 2010, 285, 40072−40080.
(8) Kim, S.; Oh, J.; Choi, J.; Jang, J.; Kang, M. Identification of
human thioredoxin as a novel IFN-gamma induced factor: Mechanism
of induction and its role in cytokine production. BMC Immunol. 2008,
9, 1−15.
(9) Jin, X.; Stamnaes, J.; Klock, C.; Diraimondo, T. R.; Sollid, L. M.;
̈
Khosla, C. Activation of extracellular transglutaminase 2 by
thioredoxin. J. Biol. Chem. 2011, 286, 37866−37873.
(10) Molberg, O.; Mcadam, S. N.; Korner, R.; Quarsten, H.;
̈
Kristiansen, C.; Madsen, L.; Fugger, L.; Scott, H.; Noren
́
, O.;
Roepstorff, P.; Lundin, K. E.; Sjostrom, H.; Sollid, L. M. Tissue
̈
̈
transglutaminase selectively modifies gliadin peptides that are
recognized by gut-derived T cells in celiac disease. Nat. Med. 1998,
4, 713−717.
(11) van de Wal, Y.; Kooy, Y.; van Veelen, P.; Pena, S.; Mearin, L.;
̃
Papadopoulos, G.; Koning, F. Selective deamidation by tissue
transglutaminase strongly enhances gliadin-specific T cell reactivity.
J. Immunol. 1998, 161, 1585−1588.
(12) Siegel, M.; Strnad, P.; Watts, R. E.; Choi, K.; Jabri, B.; Omary,
M. B.; Khosla, C. Extracellular transglutaminase 2 is catalytically
inactive, but is transiently activated upon tissue injury. PLoS ONE
2008, 3, e1861.
(13) Nilsen, E. M.; Lundin, K. E.; Krajci, P.; Scott, H.; Sollid, L. M.;
Brandtzaeg, P. Gluten specific, HLA-DQ restricted T cells from coeliac
mucosa produce cytokines with Th1 or Th0 profile dominated by
interferon gamma. Gut 1995, 37, 766−776.
(14) Matsui, M.; Oshima, M.; Oshima, H.; Takaku, K.; Maruyama,
T.; Yodoi, J.; Taketo, M. M. Early embryonic lethality caused by
targeted disruption of the mouse thioredoxin gene. Dev. Biol. 1996,
178, 7−7.
(15) Kirkpatrick, D. L.; Ehrmantraut, G.; Stettner, S.; Kunkel, M.;
Powis, G. Redox active disulfides: The thioredoxin system as a drug
target. Oncol. Res. 1997, 9, 351−356.
(16) Kunkel, M. W.; Kirkpatrick, D. L.; Johnson, J. I.; Powis, G. Cell
line-directed screening assay for inhibitors of thioredoxin reductase
signaling as potential anti-cancer drugs. Anticancer Drugs 1997, 12,
659−670.
(17) Ramanathan, R. K.; Kirkpatrick, D. L.; Belani, C. P.; Friedland,
D.; Green, S. B.; Chow, H.-H. S.; Cordova, C. A.; Stratton, S. P.;
Sharlow, E. R.; Baker, A.; Dragovich, T. A phase I pharmacokinetic and
pharmacodynamic study of PX-12, a novel inhibitor of thioredoxin-1,
in patients with advanced solid tumors. Clin. Cancer Res. 2007, 13,
2109−2114.
(18) Ramanathan, R. K.; Abbruzzese, J.; Dragovich, T.; Kirkpatrick,
L.; Guillen, J. M.; Baker, A. F.; Pestano, L. A.; Green, S.; Hoff, D. D. A
randomized phase II study of PX-12, an inhibitor of thioredoxin in
patients with advanced cancer of the pancreas following progression
after a gemcitabine-containing combination. Cancer Chemother.
Pharmacol. 2011, 67, 503−509.
ASSOCIATED CONTENT
■
S
* Supporting Information
Data demonstrating the specificity of Trx for TG2 compared to
DTT, the growth time of T84 cells and Trx ability to activate
extracellular TG2, biological activity curves for Trx inhibitors,
and comparison of Trx inhibitor reactivity for DTT,
glutathione, and cysteine. This material is available free of
charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
*Phone: (650) 723-6538. E-mail: khosla@stanford.edu.
Author Contributions
Authors T.R.D., N.M.P., and X.J. contributed equally.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This research was supported by a grant from the NIH (R01
DK063158) to C.K.
ABBREVIATIONS USED
■
Trx, thioredoxin, thioredoxin-1; TG2, tissue transglutaminase
2; TrxR, thioredoxin reductase; IFN- γ, interferon- γ; DTT,
dithiothreitol; 5BP, 5-biotinamidopentylamine; SAR, struc-
ture−activity relationship.
REFERENCES
■
(1) Holmgren, A.; Lu, J. Thioredoxin and thioredoxin reductase:
Current research with special reference to human disease. Biochem.
Biophys. Res. Commun. 2010, 396, 120−124.
(2) Gromer, S.; Urig, S.; Becker, K. The thioredoxin system−From
science to clinic. Med. Res. Rev. 2004, 24, 40−89.
(3) Lillig, C. H.; Holmgren, A. Thioredoxin and related molecules−
From biology to health and disease. Antioxid. Redox Signal. 2007, 9,
25−47.
(4) Rubartelli, A.; Bajetto, A.; Allavena, G.; Wollman, E.; Sitia, R.
Secretion of thioredoxin by normal and neoplastic cells through a
leaderless secretory pathway. J. Biol. Chem. 1992, 267, 24161−24164.
(19) Kirkpatrick, D. L.; Kuperus, M.; Dowdeswell, M.; Potier, N.;
Donald, L. J.; Kunkel, M.; Berggren, M.; Angulo, M.; Powis, G.
Mechanisms of inhibition of the thioredoxin growth factor system by
antitumor 2-imidazolyl disulfides. Biochem. Pharmacol. 1998, 55, 987−
994.
(20) Diraimondo, T. R.; Klock, C.; Khosla, C. Interferon-γ activates
̈
transglutaminase 2 via a phosphatidylinositol-3-kinase-dependent
pathway: Implications for celiac sprue therapy. J. Pharmacol. Exp.
Ther. 2012, 341, 104−114.
1309
dx.doi.org/10.1021/jm301775s | J. Med. Chem. 2013, 56, 1301−1310

Journal of Medicinal Chemistry
Article
(21) Watts, R. E.; Siegel, M.; Khosla, C. Structure−activity
relationship analysis of the selective inhibition of transglutaminase 2
by dihydroisoxazoles. J. Med. Chem. 2006, 49, 7493−7501.
(22) Wilson, J.; Bayer, R.; Hupe, D. Structure−reactivity correlations
for the thiol−disulfide interchange reaction. J. Am. Chem. Soc. 1977,
99, 7922−7926.
(23) Szajewski, R.; Whitesides, G. Rate constants and equilibrium
constants for thiol−disulfide interchange reactions involving oxidized
glutathione. J. Am. Chem. Soc. 1980, 102, 2011−2026.
(24) Houk, J.; Whitesides, G. Structure−reactivity relations for thiol−
disulfide interchange. J. Am. Chem. Soc. 1987, 109, 6825−6836.
(25) Fernandes, P. A.; Ramos, M. J. Theoretical insights into the
mechanism for thiol/disulfide exchange. Chemistry 2004, 10, 257−266.
(26) Tauer, T. P.; Derrick, M. E.; Sherrill, C. D. Estimates of the ab
initio limit for sulfur−π interactions: The H2S−benzene dimer. J. Phys.
Chem. A 2005, 109, 191−196.
(27) Ringer, A. L.; Senenko, A.; Sherrill, C. D. Models of S/π
interactions in protein structures: Comparison of the H2S benzene
complex with PDB data. Protein Sci. 2007, 16, 2216−2223.
(28) Krause, G.; Holmgren, A. Substitution of the conserved
tryptophan 31 in Escherichia coli thioredoxin by site-directed
mutagenesis and structure−function analysis. J. Biol. Chem. 1991,
266, 4056−4066.
(29) Burke-Gaffney, A.; Callister, M. E. J.; Nakamura, H.
Thioredoxin: Friend or foe in human disease? Trends Pharmacol. Sci.
2005, 26, 398−404.
(30) Klock, C.; Khosla, C. Regulation of the activities of the
̈
mammalian transglutaminase family of enzymes. Protein Sci. 2012, 21,
1781−1791.
(31) Gundemir, S.; Colak, G.; Tucholski, J.; Johnson, G. V. W.
Transglutaminase 2: A molecular Swiss army knife. Biochim. Biophys.
Acta 2012, 1823, 406−419.
(32) Sollid, L. M.; Jabri, B. Celiac disease and transglutaminase 2: A
model for posttranslational modification of antigens and HLA
association in the pathogenesis of autoimmune disorders. Curr. Opin.
Immunol. 2011, 23, 732−738.
(33) Kirkpatrick, D.; Jimale, M.; King, K.; Chen, T. Synthesis and
evaluation of imidazolyl disulfides for selective cytotoxicity to hypoxic
EMT6 tumor cells in vitro. Eur. J. Med. Chem. 1992, 27, 33−37.
(34) Okolotowicz, K. J.; Shi, R.; Zheng, X.; MacDonald, M.; Reed, J.
C.; Cashman, J. R. Selective benzimidazole inhibitors of the antigen
receptor-mediated NF-κB activation pathway. Bioorg. Med. Chem.
2010, 18, 1918−1924.
́
(35) Arner, E. S.; Holmgren, A. Measurement of thioredoxin and
thioredoxin reductase. Curr. Protoc. Toxicol. 2001, 24, 7.4.1−7.4.14.
(36) Grosdidier, A.; Zoete, V.; Michielin, O. SwissDock, a protein−
small molecule docking web service based on EADock DSS. Nucleic
Acids Res. 2011, 39, W270−W277.
(37) Pettersen, E. F.; Goddard, T. D.; Huang, C. C.; Couch, G. S.;
Greenblatt, D. M.; Meng, E. C.; Ferrin, T. E. UCSF Chimera−A
visualization system for exploratory research and analysis. J. Comput.
Chem. 2004, 25, 1605−1612.
1310
dx.doi.org/10.1021/jm301775s | J. Med. Chem. 2013, 56, 1301−1310