Dihydroisoxazole analogs for labeling and visualization of catalytically active transglutaminase 2

Article abstract of DOI:10.1016/j.chembiol.2010.11.004

We report the synthesis and preliminary characterization of "clickable" inhibitors of human transglutaminase 2 (TG2). These inhibitors possess the 3-halo-4,5-dihydroisoxazole warhead along with bioorthogonal groups such as azide or alkyne moieties that enable subsequent covalent modification with fluorophores. Their mechanism for inhibition of TG2 is based on halide displacement, resulting in the formation of a stable imino thioether. Inhibition assays against recombinant human TG2 revealed that some of the clickable inhibitors prepared in this study have comparable specificity as benchmark dihydroisoxazole inhibitors reported earlier. At low micromolar concentrations they completely inhibited transiently activated TG2 in a WI-38 fibroblast scratch assay and could subsequently be used to visualize the active enzyme in situ. The potential use of these inhibitors to probe the role of TG2 in celiac sprue as well as other diseases is discussed.

Full text of DOI:10.1016/j.chembiol.2010.11.004

Chemistry & Biology
Article
Dihydroisoxazole Analogs for Labeling
and Visualization of Catalytically Active
Transglutaminase 2
Laila Dafik1 and Chaitan Khosla1,2,
*
1Departments of Chemistry and Chemical Engineering
2Department of Biochemistry
Stanford University, Stanford, CA 94305, USA
*Correspondence: khosla@stanford.edu
DOI 10.1016/j.chembiol.2010.11.004
SUMMARY
et al., 1994; Singh et al., 1995). TG2 activity is also subject to
regulation by local redox conditions (Stamnaes et al., 2010).
We report the synthesis and preliminary character- Under oxidizing conditions, the formation of an intramolecular
disulfide bond inactivates the enzyme; activity is reversibly
restored under reducing conditions. Given the ubiquitous nature
ization of ‘‘clickable’’ inhibitors of human transgluta-
minase 2 (TG2). These inhibitors possess the
of TG2 and its ability to form covalent ‘‘scars,’’ its catalytic
activity must be tightly regulated in mammals. However, the
physiological signals that interconvert these active and inactive
3-halo-4,5-dihydroisoxazole warhead along with
bioorthogonal groups such as azide or alkyne moie-
ties that enable subsequent covalent modification
with fluorophores. Their mechanism for inhibition of
TG2 is based on halide displacement, resulting in
the formation of a stable imino thioether. Inhibition
forms of the enzyme remain largely unknown. For example,
whereas the majority of extracellular TG2 in the mouse small
intestine is inactive, it can be rapidly activated at the villus tips
by administration of poly(I:C), a ligand of the TLR-3 innate
assays against recombinant human TG2 revealed
that some of the clickable inhibitors prepared in
immune receptor (Siegel et al., 2008). Therefore, chemical
probes of in vivo TG2 biology should be of considerable use in
this study have comparable specificity as benchmark resolving some of these fundamental mysteries associated
with allosteric TG2 regulation.
From a translational standpoint, TG2 is implicated in the
pathogenesis of a number of unrelated disorders, including
dihydroisoxazole inhibitors reported earlier. At low
micromolar concentrations they completely inhibited
transiently activated TG2 in a WI-38 fibroblast
scratch assay and could subsequently be used to
visualize the active enzyme in situ. The potential
use of these inhibitors to probe the role of TG2 in
celiac sprue as well as other diseases is discussed.
neurological diseases such as Huntington’s, Alzheimer’s, and
Parkinson’s disease, certain types of cancers and renal
diseases, and celiac sprue (Ruan and Johnson, 2007; Shweke
et al., 2008; Verma and Mehta, 2007). For example, in celiac
sprue, peptides derived from dietary gluten are deamidated by
TG2 to enhance their affinity toward the disease-associated
class II major histocompatibility complexes, which in turn
INTRODUCTION
activates inflammatory Th1 cells (Sollid and Lundin, 2009).
Therefore, TG2 may be a suitable target for celiac sprue therapy.
Mammalian transglutaminases play important roles in many However, absent knowledge of the location, timing, and mecha-
biological processes such as blood clotting, wound healing, nism of TG2 activation in the celiac small intestine, this hypoth-
extracellular matrix formation, and cell adhesion and motility esis cannot be rationally tested. Again, chemical probes could
(Akimov and Belkin, 2001; Fesus et al., 1987; Iismaa et al., provide answers to these questions. Last, but not least, small
2009; Lorand, 2007; Lorand and Graham, 2003; Piacentini molecule TG2 inhibitors could also serve as drug leads and even-
et al., 1991; Zemskov et al., 2006). Transglutaminase 2 (TG2), tually as drug candidates for one or more of the above diseases
a ubiquitous member of this enzyme family, is found in intracel- (Siegel and Khosla, 2007).
lular as well as extracellular environments of many organs. In the
Many classes of irreversible, active site inhibitors of human
presence of calcium and the absence of GTP or GDP, TG2 TG2 have been reported thus far (Case et al., 2005;
crosslinks selected g-glutaminyl and 3-lysine residues on Choi et al., 2005; Freund et al., 1994; Halim et al., 2007; Hausch
proteins, leading to the formation of proteolytically resistant et al., 2003; Keillor, 2005; Pardin et al., 2006, 2008; Schaertl
covalent bonds. This catalytically active form of TG2 exists in et al., 2010). One example is the 3-bromo-4,5-dihydroisoxazoles
an extended (‘‘open’’) conformation (Pinkas et al., 2007). In (DHIs) family of compounds (Castelhano et al., 1988; Choi et al.,
contrast, in the absence of calcium and the presence of GTP 2005; Watts et al., 2006). Guided by the excellent tolerance of
or GDP, TG2 assumes an inactive (‘‘closed’’) form (Liu et al., these molecules in acute and chronic administration studies in
2002) and functions as a G protein in the phospholipase C signal rodents (Choi et al., 2005; Yuan et al., 2007), here, we have
transduction cascade of the a-adrenergic receptor (Nakaoka used the DHI scaffold to develop a class of covalent TG2 probes
58 Chemistry & Biology 18, 58–66, January 28, 2011 ª2011 Elsevier Ltd All rights reserved

Chemistry & Biology
Tools for Visualizing Transglutaminase Activity
Figure 1. Labeling and Visualization of Catalytically Active (but Not Catalytically Inactive) TG2 with Alkynyl-DHI Inhibitors
Similar labeling strategy can be employed using azido-DHI inhibitors.
that bind to active but not inactive TG2 and can, therefore, be DHI inhibitor. Both azide and alkyne groups were introduced as
used to visualize the enzyme via a selective (bioorthogonal) substitutions on the carbobenzyloxy (Cbz) moiety because
chemical reaction (Figure 1). Analogous strategies have been previous studies had highlighted the advantage of an aromatic
used to modify cell surface glycans and proteins (Chin et al., group at this position (Watts et al., 2006). To identify a clickable
2002; Deiters and Schultz, 2005; Wang and Schultz, 2004), to inhibitor with optimal activity, we also varied the amino acid
profile protein glycosylation (Agard and Bertozzi, 2009), and for moiety of our DHI inhibitors using side chains that had previously
the detection of lipid-modified proteins (Hang et al., 2007). been found to be beneficial. The compounds synthesized and
Recently, bioorthogonal chemical reporters have also been evaluated in this study are summarized in Table 1; all of them
used in living animals as noninvasive imaging tools (Laughlin were made in good yield and purity (see Supplemental Experi-
et al., 2008; Laughlin and Bertozzi, 2009; Prescher and Bertozzi, mental Procedures and Figure S1 available online). Four
2005).
previously synthesized DHI inhibitors, 13–16, were chosen as
reference compounds in this study; their TG2 inhibitory activities
vary over an 8-fold range (Table 1).
RESULTS
Design and Synthesis of Azido- and Alkynyl-DHI
Inhibitors
Inhibition of Recombinant Human TG2
by Azido- and Alkynyl-DHI Inhibitors
In this study we describe the synthesis and properties of DHI The potencies of alkynyl- and azido-DHI inhibitors 1–12 were
inhibitors of human TG2 that are conjugated to alkyne or azide evaluated in a kinetic assay using recombinant human TG2. In
functional groups. We incorporated alkyne and azide groups this assay, TG2-catalyzed deamidation of the dipeptide
because of their bioorthogonality (Sletten and Bertozzi, 2009). substrate Cbz-Gln-Gly is monitored under steady-state condi-
Figure 2 describes the structure and synthesis of a representative tions. Analogous to the parameters of the Michaelis-Menten
Figure 2. General Synthesis of Azido- and Alkynyl-DHI Inhibitors
See Figure S1.
Chemistry & Biology 18, 58–66, January 28, 2011 ª2011 Elsevier Ltd All rights reserved 59

Chemistry & Biology
Tools for Visualizing Transglutaminase Activity
Table 1. TG2 Inhibitory Activity of Enantiopure Azido- and Alkynyl-DHI Compounds
Compounds 17 and 18 have a stereochemically altered DHI moiety and are, therefore, considerably less potent than their diastereomeric analogs, 11
and 12, respectively.
equation, the first order rate constant kinh is a measure of the for this inverse correlation between affinity and specificity
reactivity of the DHI warhead with the active site sulfhydryl of remains to be understood, in subsequent studies we placed
TG2, whereas Ki estimates the reversible binding affinity of the greater emphasis on biological evaluations of the most prom-
inhibitor to the active site. The bimolecular rate constant kinh/Ki ising alkynyl-DHI inhibitors.
can be thought of as the specificity constant of TG2 for the
inhibitor.
Inhibition of TG2 in a Tissue Culture Wounding Model
As summarized in Table 1, inhibitors harboring tryptophan To evaluate the activity of the clickable TG2 inhibitors identified
analogs are more potent than those with tyrosine as the amino above in a representative biological assay, we used a tissue
acid residue. Among tryptophan analogs, the 5-fluorotryptophan culture wounding model. TG2 plays a role in wound healing in
analogs have the highest specificity constants. The relative mammals by crosslinking extracellular matrix proteins after
trends are similar between the azido-DHI and alkynyl-DHI inhib- a wound has been inflicted. Recently, we demonstrated that,
itor series. The most active alkynyl- and azido-DHI inhibitors (12 when
a confluent monolayer of WI-38 fibroblast cells is
and 6, respectively) have specificity constants that are within scratched, TG2 is rapidly and transiently activated in the zone
1.3- and 2.3-fold, respectively, of the most potent DHI inhibitor surrounding the ‘‘wound’’ (Siegel et al., 2008). This activated
16 identified to date. Together, these results encouraged us to TG2 enzyme can be visualized using 5-biotinyl pentylamine
evaluate some of these compounds in a cell culture assay for (5BP), which presumably crosslinks onto glutamine side chains
activated TG2.
of fibronectin and other proteins along the scratch border that
It is noteworthy that the azido-DHI inhibitors prepared in this are recognized as TG2 substrates (Figure 3). Here, we have
study have moderately higher TG2 specificity (kinh/Ki) than their exploited this assay to compare the relative activities of TG2
alkynyl-DHI counterparts (Table 1). However, this benefit is inhibitors in a biological system.
primarily derived from higher reactivity toward the active site To quantify inhibitor potency, selected inhibitors were added
cysteine residue of TG2; the reversible binding affinity (Ki) of to scratched WI-38 cultures at incremental 2-fold dilutions. The
the alkynyl-DHI inhibitors is actually higher. Although the reason lowest inhibitor concentration at which complete inhibition of
60 Chemistry & Biology 18, 58–66, January 28, 2011 ª2011 Elsevier Ltd All rights reserved

Chemistry & Biology
Tools for Visualizing Transglutaminase Activity
Figure 3. Titration of 11 against Active TG2 in a Fibroblast Scratch Assay
TG2 activity was visualized in situ after scratching a confluent WI-38 monolayer with a small pipette tip. Significant TG2 activity was detected around the wound in
the presence of 300 mM 5BP and vehicle (DMSO) (A–C). Clear inhibition of active TG2 was observed in the presence of 6.25 mM 11 (G–I), and complete inhibition
was observed at or above 12.5 mM 11 (J–O). In this assay, active TG2 was detected by exposing fixed cultures to streptavidin Alexa Fluor 555 (A, D, G, J, and M).
The scratch geometry is conveniently visualized by co-staining with polyclonal anti-fibronectin antibody, followed by a secondary antibody conjugated Alexa
Fluor 488 (B, E, H, K, and N). Overlays of the left and middle images are in the right column (C, F, I, L, and O). Scale bar represents 200 mm and applies to all panels.
transiently activated TG2 was observed was recorded as the blocked TG2 activity around the wound at 3.1–6.25 mM, the
minimum inhibitory concentration (MIC) of that inhibitor. For less potent inhibitor 15 (kinh/KI = 6.1 mMÀ1 minÀ1) only showed
example the effect of increasing concentrations of 11 on the partial inhibition at 25 mM.
extent of 5BP crosslinking around WI-38 scratches is shown in
The above cell culture measurements verified that the biolog-
Figure 3; from these data, its MIC was estimated to lie between ical potencies of the most active azido-DHI inhibitor (6) and the
6.25 and 12.5 mM. The MIC values of different inhibitors are most active alkynyl-DHI inhibitors (11 and 12) were comparable
compared in Table 2 (see also Figures S2 and S3). Whereas to the benchmark DHI inhibitor of human TG2 (16) identified to
the most potent inhibitors (e.g., 6, 11, 14, and 16) completely date (Watts et al., 2006). This in turn encouraged us to develop
Chemistry & Biology 18, 58–66, January 28, 2011 ª2011 Elsevier Ltd All rights reserved 61

Chemistry & Biology
Tools for Visualizing Transglutaminase Activity
cule via a Cu1-catalyzed alkyne-azide [3+2] cycloaddition reaction
(Figure 1). Initial experiments were performed with recombinant
TG2 protein activated with Ca+2 or inhibited with GTP. The
proteins were resolved by gel electrophoresis and visualized by
western blotting using neutravidin-linked horseradish peroxidase.
Catalytically active TG2 treated with 12 can be efficiently conju-
gated with a suitable azide partner molecule and subsequently
visualized in a dose-dependent manner (Figure 4, lanes 1–4).
ThelimitofdetectionofactiveTG2bythiswesternblotting method
was below 1 mg. GTP-bound, catalytically inactive TG2 was not
labeled and, therefore, not visualized (Figure 4, lanes 5–8).
To investigate the utility of this method for visualizing catalyt-
ically active TG2 in biological systems, we performed fluores-
cence microscopy on scratched WI-38 fibroblast cultures
that had been incubated with two representative alkynyl-DHI
inhibitors, 11 and 12. In both cases, cells were extensively
washed to remove excess inhibitor, fixed with paraformalde-
hyde (PFA), permeabilized with 0.1% Triton X-100, washed
again, and subjected to the [3+2] cycloaddition reaction with
biotin azide TG2 (active + inactive). The biotin conjugate was
visualized by exposing fixed cultures to streptavidin conju-
gated with Alexa Fluor 555. The scratch was visualized with
rabbit anti-fibronectin antibody, followed by exposure to an
Alexa Fluor 488-conjugated goat anti-rabbit IgG secondary
antibody.
Table 2. Titration of DHI Inhibitors against TG2 in a Fibroblast
Wound Model
Each value is a representative of three or more scratches. See Figures S2
and S3.
Analogous to experiments with 5BP (see, for example, Fig-
ure 3), a strong red fluorescence signal was observed around
the scratch in samples treated with inhibitors and reacted with
biotin azide (Figure 5; Figure S4), corresponding to catalytically
active TG2. As negative controls, DMSO vehicle and two
substantially weaker inhibitors, 17 and 18 (Table 1), were evalu-
ated. These compounds, which are diastereomeric analogs of 11
and 12, respectively, were used because earlier work had shown
that TG2 specificity was strongly dependent upon the stereo-
chemistry of the chiral center in the dihydroisoxazole ring.
Scratches treated with 17 and 18 resulted in no red fluorescence
around the scratch (Figure 5; Figure S4).
procedures for visualizing inhibitor-bound TG2 in biological
samples, as described below. Future improvements to these
compounds will focus on improving their water solubility and
rendering them resistant to chymotrypsin-catalyzed hydrolysis.
Both properties would be necessary for chronic oral administra-
tion in animal studies directed at elucidating TG2 regulatory
mechanisms in the small intestine and can be achieved by incor-
porating alternative amino acid residues or isosteres in place of
5-fluorotryptophan in compounds 6 and 11.
Labeling and Visualization of Catalytically Active TG2
Two other observations were made during the course of these
In principle the alkyne and azide groups incorporated into DHI studies with WI-38 fibroblasts. First, the signal-to-noise was
inhibitors can be used as handles to selectively visualize the loca- significantly superior in scratches reacted with biotin azide as
tion of TG2 protein that is transiently activated in a biological compared to rhodamine azide (Figure S5). The reason for this
sample. Totestthishypothesis, alkynyl-inhibitedTG2waschemo- difference was not investigated. Second, 5BP yields a distinctly
selectively ligated to an azide-tagged biotin or fluorophore mole- different labeling pattern around the fibroblast scratch than the
Figure 4. Visualization of Recombinant TG2 with
Alkynyl-DHI Inhibitor 12
Recombinant TG2 was inhibited with 12 in the presence of
CaCl2 (5 mM) or GTP (500 mM) and MgCl2 (1 mM). The re-
sulting mixture was reacted with biotin-azide, resolved by
SDS-PAGE, and detected by western blotting with neutra-
vidin-HRP. Lane 1 shows TG2 (0.2 mg) + Ca2+ + 12; lane 2,
TG2 (0.4 mg) + Ca2+ + 12; lane 3, TG2 (1 mg) + Ca2+ + 12;
lane 4, TG2 (4 mg) + Ca2+ + 12; lane 5, TG2 (0.4 mg) +
GTP/MgCl2 + 12; lane 6, TG2 (1 mg) + GTP/MgCl2 + 12;
lane 7, TG2 (1 mg) + GTP/MgCl2 + 12; and lane 8, TG2
(4 mg) + GTP/MgCl2 + 12.
62 Chemistry & Biology 18, 58–66, January 28, 2011 ª2011 Elsevier Ltd All rights reserved

Chemistry & Biology
Tools for Visualizing Transglutaminase Activity
Figure 5. Visualization of Catalytically
Active TG2 in a Scratched WI-38 Fibroblast
Culture
TG2 activity was visualized after scratching
confluent WI-38 monolayers in the presence of
50 mM 11 (A–F) or 17 (G–I) for 1 hr. Cells were
then fixed, permeabilized, reacted with biotin
azide followed by streptavidin-Alexa Fluor 555
conjugates (red), and imaged by epifluorescence
microscopy through a 103 objective.
(A, D, and G) Rabbit anti-fibronectin antibody, fol-
lowed by a secondary antibody labeled with Alexa
Fluor 488, was used to visualize fibronectin on the
WI-38 monolayers.
(B, E, and H) Catalytically active TG2 was visual-
ized by coupling the alkynyl-DHI inhibitor to biotin
azide.
(C, F, and I) Overlays of the left and middle image
pairs are shown in the right column. Scale bar
represents 200 mm and applies to all panels.
See Figures S4 and S5.
for such a task. As shown, our clickable
TG2 inhibitors can visualize the active
enzyme in a manner that is only limited
by the resolution of optical microscopy.
Specifically, we have shown that com-
pounds 11 and 12 can be used to visu-
alize the burst of TG2 catalytic activity
that rapidly appears at the scratch
boundary of
culture system. Although the fibroblast
scratch assay mimics pathological
rather than a physiological state, its utility
a WI-38 fibroblast cell
a
clickable inhibitor. For example, whereas the overlap between for interrogating the allosteric transition of TG2 from an inactive
red and green signals is excellent in Figure 3A, the two colors to an active state has been established (Nicholas et al., 2003;
are proximal but clearly discernable in Figure 5E. Presumably, Siegel et al., 2008). Our method has two principal advantages
this is due to subtle differences in the distribution of catalytically over previous studies that visualized transiently activated TG2
active TG2 (i.e., the target of the clickable inhibitor) and its around scratched cells using amine substrates such as 5BP.
substrates (principally fibronectin, which is the target of 5BP) in First, whereas the use of amines only enables detection of
the vicinity of the scratch.
TG2 substrates (principally fibronectin) that have undergone
modification by catalytically active enzyme, our new method
enables visualization of the activated enzyme itself. As illus-
trated by the differences between the labeling patterns around
DISCUSSION
Human TG2 catalyzes transamidation and deamidation reac- scratches in Figures 3 and 5, these differences may prove
tions that are believed to contribute to a variety of physiological significant in further investigations of TG2 biology, especially
processes as well as to the pathogenesis of several diseases. in situations where subcellular resolution of the active enzyme
Therefore, understanding the mechanisms responsible for is desired. Second, because the specificity of transglutami-
spatiotemporal regulation of TG2 activity could have important nases for their electrophilic (glutamine-bearing) substrates is
implications for human biology and disease. However, because considerably higher than for nucleophilic (amine-bearing)
the protein is abundantly expressed in most organs and tissues, substrates, the development of isozyme-specific clickable DHI
most mechanisms of interest are allosteric in nature, and inhibitors should be feasible.
conventional molecular biology approaches are not applicable.
Pending verification of their in vivo stability, several important
Consequently, new tools are needed to visualize rapid, tissue- biological and medical problems could be addressed with click-
specific transitions of TG2 from a catalytically inactive to an able DHI inhibitors such as 11 and 12. In metastatic cancers,
active state (or vice versa). Because TG2 is localized in multiple TG2 is believed to play a crucial role in cell migration and has
intracellular compartments as well as the extracellular environ- been shown to localize primarily at the leading edge of
ment, the ideal tools should also be able to resolve the subcel- migrating cells (Antonyak et al., 2009). However, the relative
lular location of the target protein. Here, we have described importance of its catalytic versus signaling activity is unclear.
a new class of small molecule reagents that are well suited For example the catalytically inactive C277S mutant of TG2 is
Chemistry & Biology 18, 58–66, January 28, 2011 ª2011 Elsevier Ltd All rights reserved 63

Chemistry & Biology
Tools for Visualizing Transglutaminase Activity
(Atlanta Biologicals), and 100 U/ml penicillin + 100 mg/ml streptomycin
(GIBCO; 15140). Cells were typically grown to 80%–90% confluency before
treating with trypsin-EDTA (GIBCO; 15400) and splitting. All cells were grown
in a humidified incubator at 37ꢀC, 5% CO2. The antibodies used were anti-
fibronectin antibody produced in Rabbit (Sigma-Aldrich; F3648), Alexa Fluor
488 goat anti-rabbit IgG (H-L) antibody (Invitrogen; A-11008), and Alexa Fluor
555-labeled streptavidin (Invitrogen; S-21381).
indistinguishable from the wild-type protein with respect to cell
motility (Balklava et al., 2002). Visualizing the precise location(s)
of catalytically active TG2 over the course of a nascent and
metastasizing tumor could shed important light on the role of
protein crosslinking in cancer. In celiac sprue the pathogenic
role of catalytic TG2 is well established; however, the precise
location at which catalytically active TG2 encounters immuno-
toxic gluten peptides remains a topic of controversy. Again,
small molecule probes that can highlight the distribution of
active enzyme in normal versus inflamed intestinal mucosa
could be invaluable. More fundamentally, biochemical and
structural biological investigations have demonstrated that
TG2 undergoes allosteric changes between at least three
distinct conformational states, as summarized in the scheme
below:
Synthesis of Alkynyl- and Azido-DHI Inhibitors
Compounds 3–18 were synthesized as previously reported (Watts et al., 2006).
Briefly, 4-ethynylbenzyl alcohol was reacted with p-nitrophenyl chloroformate
in methylene chloride and N-methylmorpholine to give
a carbonate of
4-ethynyl benzyl and p-nitrophenol after purification by flash chromatography.
The carbonate was reacted with the methyl ester of the amino acid in DMF and
N-methylmorpholine to give 4-ethynyl benzyl carbamate protected amino acid
methyl ester. Methyl ester hydrolysis, followed by EDCI coupling of resulting
acid to the (S)-isomer of the halo-dihydroisoxazole, gave the indicated
compound as a final product after flash chromatography purification. The
azido-DHI compounds were synthesized in similar fashion.
TG2 Inhibition Assays
Recombinant human TG2 was expressed in E. coli and purified to >90%
homogeneity as described previously(Choi et al., 2005). The potency of indi-
vidual DHI inhibitors was assayed in a reaction mixture containing 200 mM
MOPS (pH 7.2), 5 mM CaCl2, 1 mM EDTA, 10 mM a-ketoglutarate, 18 U/ml
glutamate dehydrogenase, 0.4 mM NADH, 3.3% (v/v) DMSO, 0.5 mM TG2,
and 10, 20, and 30 mM Cbz-Gln-Gly (KM = 6.28 mM, kcat. = 37 minÀ1). The
concentration of each inhibitor was varied from 0.001 to 0.3 mM. The reaction
was initiated by adding TG2, and the consumption of NADH was monitored by
Tools that can distinguish between the Ca+2-bound active
state versus the GTP-bound or disulfide-bonded inactive states
could greatly facilitate a deeper understanding of the dynamics
of this multifunctional protein in biological systems.
In conclusion, we also note that our lead clickable inhibitors
reported here could be further improved with respect to several
characteristics. First, they have relatively low solubility in
aqueous buffers. The efficiency of labeling TG2-bound inhibitors
via [3+2] cycloaddition could also be improved by varying the
spacer length between the alkyne and the benzyl groups. Last,
but not least, the reactivity of these inhibitors with other mamma-
lian transglutaminases could be further engineered in order to
develop probes with orthogonal specificity for individual
isozymes of this protein family.
UV spectroscopy (340 nm, 3 = 6220 cmÀ1
M
À1). Kinetic parameters were
obtained by plotting the reaction progress curves against theoretical equa-
tions for irreversible enzyme inhibition.
Labeling and Visualization of TG2 with Clickable Inhibitors
TG2 protein was subjected to the probe labeling in 50 ml volume reaction con-
taining a final concentration of 2 mM of 12, and 5 mM CaCl2 or 500 mM GTP/
1 mM MgCl2 in 100 mM Tris HCl (pH 7.2) for 1 hr at RT. Then, 2 mM of biotin
azide was added, 1 mM tris(2-carboxyethyl)phosphine hydrochloride (TCEP;
Sigma-Aldrich) dissolved in water, 0.2 mM tris[(1-benzyl-1H-1,2,3-triazole-
4-yl)methyl amine (TBTA; Sigma-Aldrich) dissolved in DMSO/tert-butanol
(20%/80%), and 1 mM CuSO4 in PBS. The reaction was quenched with 23
loading buffer. Samples were resolved by SDS-PAGE using 4%–20% Tris-
glycine gels and transferred onto a nitrocellulose membrane. The membrane
was blocked with PBS, 0.1% Tween 20 (PBST), and 5% nonfat dried milk
for 1 hr at RT. The membrane was washed three times with PBST (5 min
each) and incubated with neutravidin-horseradish peroxidase (Invitrogen;
1:200 in 5% nonfat dried milk) for 1 hr at RT. The membrane was washed again
with PBST three times (10 min each) and developed according to manufac-
turer’s recommendation (Amersham Biosciences).
SIGNIFICANCE
We have developed nonradioactive chemical probes for de-
tecting and imaging catalytically active transglutaminase
2 (TG2) in mammalian cells and tissues. A series of electro-
philic dihydroisoxazole compounds have been engineered
with alkyne and azide groups that can be further modified
by ‘‘clickchemistry’’ (Cu1-catalyzedalkynyl-azide [3+2]cyclo-
addition). These compounds have high specificity for the TG2
active site and undergo an irreversible reaction with the cata-
lytically active cysteine residue. The biological utility of these
mechanism-based irreversible inhibitors is demonstrated by
theirabilitytovisualizetransientlyactivatedTG2inafibroblast
scratch assay. These probes should be broadly useful in
elucidating mechanisms responsible for allosteric regulation
of this multifunctional protein in cell biology, mammalian
development, as well as a variety of human diseases.
Detection of TG2 Activity in Fibroblast Scratch Assays Using 5BP
WI-38 fibroblasts were plated in an 8-well chamber glass slide at 2 3 105 cells/
well and grown for 3 days with medium change every 2 days. Small scratches
were made in the monolayer with 0.1–10 ml pipette tip, and the cells were
incubated with 5BP for 1 hr at 37ꢀC, 5% CO2. To capture the highest TG2
activity around the wound, we performed titration assays using 50, 100, 150,
and 300 mM 5BP; the best results were obtained at 300 mM 5BP. To quantify
TG2 inhibition in cultured cells, inhibitors 3–18 were added at the indicated
concentrations to the culture media for 10–15 min. An equal amount of vehicle
(DMSO) was added to the media of control cells at 37ꢀC. Then, 5BP was added
at a final concentration 300 mM and incubated for an additional 1 hr at 37ꢀC,
5% CO2. The cells were washed three times with PBS and fixed with À20ꢀC
methanol for 10 min. Fixed cells were washed again with PBS twice for
5 min, blocked with 1% BSA in PBS for 15 min at room temperature, and
washed twice more with PBS. Anti-fibronectin antibody produced in Rabbit
(F3648; Sigma-Aldrich) was diluted 1:500 in blocking buffer and incubated
for 1 hr at RT. Cells were washed three times with PBS and incubated with
Alexa Fluor 488 goat anti-rabbit IgG (H+L) (A-11034; Invitrogen) and Alexa Fluor
EXPERIMENTAL PROCEDURES
Cell Culture and Antibodies
WI-38 fibroblasts were grown in minimum essential media (GIBCO; 11095)
containing L-glutamine, Earle’s salts, nonessential amino acids (GIBCO;
11140), sodium pyruvate (GIBCO; 11360), 10% fetal bovine serum (FBS)
64 Chemistry & Biology 18, 58–66, January 28, 2011 ª2011 Elsevier Ltd All rights reserved

Chemistry & Biology
Tools for Visualizing Transglutaminase Activity
555-streptavidin conjugate in blocking buffer for 1 hr at RT. After three further
washes with PBS, 300 ml PBS was added onto each sample before visualiza-
tion via fluorescence microscopy. Note that Factor XIII, another ubiquitous
member of the transglutaminase family, is not expressed in WI-38 fibroblasts
(data not shown).
Chin, J.W., Santoro, S.W., Martin, A.B., King, D.S., Wang, L., and Schultz, P.G.
(2002). Addition of p-azido-L-phenylalanine to the genetic code of Escherichia
coli. J. Am. Chem. Soc. 124, 9026–9027.
Choi, K., Siegel, M., Piper, J.L., Yuan, L., Cho, E., Strnad, P., Omary, B., Rich,
K.M., and Khosla, C. (2005). Chemistry and biology of dihydroisoxazole
derivatives: selective inhibitors of human transglutaminase 2. Chem. Biol.
12, 469–475.
Detection of TG2 Activity in Fibroblast Scratch Assays Using
Clickable Inhibitors
Deiters, A., and Schultz, P.G. (2005). In vivo incorporation of an alkyne into
proteins in Escherichia coli. Bioorg. Med. Chem. Lett. 15, 1521–1524.
Following WI-38 fibroblast cell growth and scratching as above, compounds
11 and 12 were added to living cells for 10–15 min at 50 mM final concentration
at 37ꢀC, 5% CO2. Control cells were incubated with 300 mM 5BP and vehicle
(DMSO). Cells were washed three times with PBS to remove excess
compounds and fixed with 4% PFA for 10 min at RT. Cells were then permea-
bilized with PBS/0.1% Triton X-100 for 1–2 min at RT, washed extensively with
PBS, and subjected to the [3+2] cycloaddition reaction at RT for 1 hr in 175 ml
solution containing 1 mM biotin azide, 1 mM TCEP dissolved in water, and
1 mM CuSO4 in PBS. The labeled samples were rinsed extensively with PBS
and blocked in PBS/5% BSA for 45 min at RT. To visualize the fibronectin
protein, samples were treated with rabbit anti-fibronectin for 1 hr at RT in
PBS/5%BSA, then washed three times with PBS, and further stained with
the goat anti-Rabbit Alexa Fluor 488-conjugated secondary antibody for
45 min in PBS/5% BSA at RT. After three further washes, 300 ml of PBS was
added onto the cells for visualization via microscopy. Fluorescent images
were captured in an upright Zeiss microscope.
Fesus, L., Thomazy, V., and Falus, A. (1987). Induction and activation of tissue
transglutaminase during programmed cell death. FEBS Lett. 224, 104–108.
Freund, K.F., Doshi, K.P., Gaul, S.L., Claremon, D.A., Remy, D.C., Baldwin,
J.J., Pitzenberger, S.M., and Stern, A.M. (1994). Transglutaminase inhibition
by 2-[(2-oxopropyl)thio]imidazolium derivatives: mechanism of factor XIIIa
inactivation. Biochemistry 33, 10109–10119.
Halim, D., Caron, K., and Keillor, J.W. (2007). Synthesis and evaluation of
peptidic maleimides as transglutaminase inhibitors. Bioorg. Med. Chem.
Lett. 17, 305–308.
Hang, H.C., Geutjes, E.J., Grotenbreg, G., Pollington, A.M., Bijlmakers, M.J.,
and Ploegh, H.L. (2007). Chemical probes for the rapid detection of fatty-acyl-
ated proteins in mammalian cells. J. Am. Chem. Soc. 129, 2744–2745.
Hausch, F., Halttunen, T., Maki, M., and Khosla, C. (2003). Design, synthesis,
and evaluation of gluten peptide analogs as selective inhibitors of human
tissue transglutaminase. Chem. Biol. 10, 225–231.
Iismaa, S.E., Mearns, B.M., Lorand, L., and Graham, R.M. (2009).
Transglutaminases and disease: lessons from genetically engineered mouse
models and inherited disorders. Physiol. Rev. 89, 991–1023.
SUPPLEMENTAL INFORMATION
Supplemental Information includes Supplemental Experimental Procedures
and five figures and can be found with this article online at doi:10.1016/j.
chembiol.2010.11.004.
Keillor, J.W. (2005). Tissue transglutaminase inhibition. Chem. Biol. 12,
410–412.
Laughlin, S.T., and Bertozzi, C.R. (2009). In vivo imaging of Caenorhabditis
elegans glycans. ACS Chem. Biol. 4, 1068–1072.
ACKNOWLEDGMENTS
Laughlin, S.T., Baskin, J.M., Amacher, S.L., and Bertozzi, C.R. (2008). In vivo
imaging of membrane-associated glycans in developing zebrafish. Science
320, 664–667.
This work was supported by a grant from the NIH (R01 DK063158) to C.K. L.D.
is a recipient of a postdoctoral fellowship from ‘‘Fonds de Recherche sur la
Nature et les Technologies’’ (Canada).
Liu, S., Cerione, R.A., and Clardy, J. (2002). Structural basis for the guanine
nucleotide-binding activity of tissue transglutaminase and its regulation of
transamidation activity. Proc. Natl. Acad. Sci. USA 99, 2743–2747.
Received: July 27, 2010
Revised: November 2, 2010
Accepted: November 2, 2010
Published: January 27, 2011
Lorand, L. (2007). Crosslinks in blood: transglutaminase and beyond. FASEB
J. 21, 1627–1632.
Lorand, L., and Graham, R.M. (2003). Transglutaminases: crosslinking
enzymes with pleiotropic functions. Nat. Rev. Mol. Cell Biol. 4, 140–156.
REFERENCES
Nakaoka, H., Perez, D.M., Baek, K.J., Das, T., Husain, A., Misono, K., Im, M.J.,
and Graham, R.M. (1994). Gh: a GTP-binding protein with transglutaminase
activity and receptor signaling function. Science 264, 1593–1596.
Agard, N.J., and Bertozzi, C.R. (2009). Chemical approaches to perturb,
profile, and perceive glycans. Acc. Chem. Res. 42, 788–797.
Nicholas, B., Smethurst, P., Verderio, E., Jones, R., and Griffin, M. (2003).
Cross-linking of cellular proteins by tissue transglutaminase during necrotic
cell death: a mechanism for maintaining tissue integrity. Biochem. J. 371,
413–422.
Akimov, S.S., and Belkin, A.M. (2001). Cell surface tissue transglutaminase is
involved in adhesion and migration of monocytic cells on fibronectin. Blood 98,
1567–1576.
Antonyak, M.A., Li, B., Regan, A.D., Feng, Q.Y., Dusaban, S.S., and Cerione,
R.A. (2009). Tissue transglutaminase is an essential participant in the
epidermal growth factor-stimulated signaling pathway leading to cancer cell
migration and invasion. J. Biol. Chem. 284, 17914–17925.
Pardin, C., Gillet, S., and Keillor, J.W. (2006). Synthesis and evaluation of
peptidic irreversible inhibitors of tissue transglutaminase. Bioorg. Med.
Chem. 14, 8379–8385.
Pardin, C., Pelletier, J.N., Lubell, W.D., and Keillor, J.W. (2008). Cinnamoyl
Balklava, Z., Verderio, E., Collighan, R., Gross, S., Adams, J., and Griffin, M.
(2002). Analysis of tissue transglutaminase function in the migration of Swiss
3T3 fibroblasts: the active-state conformation of the enzyme does not affect
cell motility but is important for its secretion. J. Biol. Chem. 277, 16567–16575.
inhibitors of tissue transglutaminase. J. Org. Chem. 73, 5766–5775.
Piacentini, M., Autuori, F., Dini, L., Farrace, M.G., Ghibelli, L., Piredda, L., and
Fesus, L. (1991). Tissue transglutaminase is specifically expressed in neonatal
rat liver cells undergoing apoptosis upon epidermal growth factor-stimulation.
Cell Tissue Res. 263, 227–235.
Case, A., Ni, J., Yeh, L.A., and Stein, R.L. (2005). Development of a mecha-
nism-based assay for tissue transglutaminase—results of a high-throughput
screen and discovery of inhibitors. Anal. Biochem. 338, 237–244.
Pinkas, D.M., Strop, P., Brunger, A.T., and Khosla, C. (2007).
Transglutaminase 2 undergoes a large conformational change upon activa-
tion. PLoS Biol. 5, e327.
Castelhano, A.L., Billedeau, R., Pliura, D.H., Bonaventura, B.J., and Krantz, A.
(1988). Synthesis, chemistry, and absolute-configuration of novel transgluta-
minase inhibitors containing a 3-halo-4,5-dihydroisoxazole. Bioorg. Chem.
16, 335–340.
Prescher, J.A., and Bertozzi, C.R. (2005). Chemistry in living systems. Nat.
Chem. Biol. 1, 13–21.
Chemistry & Biology 18, 58–66, January 28, 2011 ª2011 Elsevier Ltd All rights reserved 65

Chemistry & Biology
Tools for Visualizing Transglutaminase Activity
Ruan, Q., and Johnson, G.V. (2007). Transglutaminase 2 in neurodegenerative
Sollid, L.M., and Lundin, K.E.A. (2009). Diagnosis and treatment of celiac
disorders. Front. Biosci. 12, 891–904.
disease. Mucosal Immunol. 2, 3–7.
Schaertl, S., Prime, M., Wityak, J., Dominguez, C., Munoz-Sanjuan, I., Pacifici,
R.E., Courtney, S., Scheel, A., and Macdonald, D. (2010). A profiling platform
for the characterization of transglutaminase 2 (TG2) inhibitors. J. Biomol.
Screen. 15, 478–487.
Stamnaes, J., Pinkas, D.M., Fleckenstein, B., Khosla, C., and Sollid, L.M.
(2010). Redox regulation of transglutaminase 2 activity. J. Biol. Chem. 285,
25402–25409.
Verma, A., and Mehta, K. (2007). Tissue transglutaminase-mediated chemore-
Shweke, N., Boulos, N., Jouanneau, C., Vandermeersch, S., Melino, G.,
Dussaule, J.C., Chatziantoniou, C., Ronco, P., and Boffa, J.J. (2008). Tissue
transglutaminase contributes to interstitial renal fibrosis by favoring accumula-
tion of fibrillar collagen through TGF-beta activation and cell infiltration. Am. J.
Pathol. 173, 631–642.
sistance in cancer cells. Drug Resist. Updat. 10, 144–151.
Wang, L., and Schultz, P.G. (2004). Expanding the genetic code. Angew.
Chem. Int. Ed. Engl. 44, 34–66.
Watts, R.E., Siegel, M., and Khosla, C. (2006). Structure-activity relationship
analysis of the selective inhibition of transglutaminase 2 by dihydroisoxazoles.
J. Med. Chem. 49, 7493–7501.
Siegel, M., and Khosla, C. (2007). Transglutaminase 2 inhibitors and their ther-
apeutic role in disease states. Pharmacol. Ther. 115, 232–245.
Siegel, M., Strnad, P., Watts, R.E., Choi, K.H., Jabri, B., Omary, M.B., and
Khosla, C. (2008). Extracellular transglutaminase 2 is catalytically inactive,
but is transiently activated upon tissue injury. PLoS ONE 3, e1861.
Yuan, L., Siegel, M., Choi, K., Khosla, C., Miller, C.R., Jackson, E.N., Piwnica-
Worms, D., and Rich, K.M. (2007). Transglutaminase 2 inhibitor, KCC009,
disrupts fibronectin assembly in the extracellular matrix and sensitizes ortho-
topic glioblastomas to chemotherapy. Oncogene 26, 2563–2573.
Singh, U.S., Erickson, J.W., and Cerione, R.A. (1995). Identification and
biochemical characterization of an 80 kilodalton GTP-binding transglutami-
nase from rabbit liver nuclei. Biochemistry 34, 15863–15871.
Zemskov, E.A., Janiak, A., Hang, J., Waghray, A., and Belkin, A.M. (2006). The
role of tissue transglutaminase in cell-matrix interactions. Front. Biosci. 11,
1057–1076.
Sletten, E.M., and Bertozzi, C.R. (2009). Bioorthogonal chemistry: fishing for
selectivity in a sea of functionality. Angew. Chem. Int. Ed. Engl. 48, 6974–6998.
66 Chemistry & Biology 18, 58–66, January 28, 2011 ª2011 Elsevier Ltd All rights reserved