Alternative reagents for methotrexate as immobilizing anchor moieties in the optimization of MASPIT: Synthesis and biological evaluation

Article abstract of DOI:10.1002/cbic.201402702

We report the evaluation of two alternative chemical dimerizer approaches aimed at increasing the sensitivity of MASPIT, a three-hybrid system that enables small-molecule target protein profiling in intact human cells. To circumvent the potential limitati

Full text of DOI:10.1002/cbic.201402702

DOI: 10.1002/cbic.201402702
Full Papers
Alternative Reagents for Methotrexate as Immobilizing
Anchor Moieties in the Optimization of MASPIT: Synthesis
and Biological Evaluation
[
a]
[a]
[a]
[b]
Dries J. H. De Clercq, Martijn D. P. Risseeuw, Izet Karalic, Anne-Sophie De Smet,
[b]
[b]
[b]
[a]
Dieter Defever, Jan Tavernier, Sam Lievens,* and Serge Van Calenbergh*
We report the evaluation of two alternative chemical dimerizer
approaches aimed at increasing the sensitivity of MASPIT,
a three-hybrid system that enables small-molecule target pro-
tein profiling in intact human cells. To circumvent the potential
limitations related to the binding of methotrexate (MTX) to en-
dogenous human dihydrofolate reductase (DHFR), we explored
trimethoprim (TMP) as an alternative prokaryote-specific DHFR
ligand. MASPIT evaluation of TMP fusion compounds with ta-
moxifen, reversine, and simvastatin as model baits, resulted in
dose–response curves shifted towards lower EC50 values than
those of their MTX congeners. Furthermore, a scalable azido-
TMP reagent was synthesized that displayed a similar improve-
ment in sensitivity, possibly owing to increased membrane per-
meability relative to the MTX anchor. Applying the SNAP-tag
approach to introduce a covalent bond into the system, on
the other hand, produced an inferior readout than in the MTX-
or TMP-tag based assay.
Introduction
Methods that allow high-throughput identification of the cellu-
lar targets of bioactive small molecules are invaluable assets in
pharmaceutical research. They are useful in mechanism-of-
action studies of hits identified by phenotypic screening,
which is increasingly being applied in both academic and in-
MASPIT (mammalian small molecule–protein interaction
trap) is the three-hybrid component of the MAPPIT technology
[7]
platform, which enables the identification of interactions be-
tween small organic compounds and their cytosolic target pro-
[8]
teins in living human cells (Scheme 1).
[
1]
dustrial programs. Additionally, they can uncover unexpected
targets of established drugs that could contribute to their ther-
apeutic efficacy or cause unwanted side effects (polypharma-
A prerequisite for successful MASPIT analysis is the availabili-
ty of appropriate synthetic probes. We previously presented
a scalable synthesis of a versatile methotrexate (MTX) reagent
that allows the rapid g-selective conjugation to alkyne-func-
tionalized bioactive small molecules to yield MTX fusion com-
[
2]
cology). Finally, such methods can also lead to the identifica-
tion of new therapeutic applications of existing drugs within
[
3]
[9]
the scope of drug-repositioning projects. Over the past
decade, numerous case studies within the target-profiling field
that apply, for example, activity-based protein profiling
pounds (MFCs) appropriate for MASPIT. Here, we take the
next step and discuss our efforts to optimize the MASPIT sys-
tem’s sensitivity based on chemical dimerizers with tamoxifen
(TAM) as the model “bait”. The latter is a selective estrogen re-
ceptor (ER) modulator that has been part of the standard ther-
[
4]
(
(
ABPP) or compound-centric chemical proteomics methods
[
5]
CCCP) have proved successful. However, despite these suc-
[10]
cess stories, target deconvolution often remains an important
bottleneck in drug-discovery research, as a generally applicable
apy for ER-positive breast cancer treatment since the 1970s.
However, TAM has been reported to induce apoptosis even in
ER-negative cancer cells, thus suggesting that it can also oper-
[
6]
methodology is still lacking.
[11]
ate by modulating alternative targets. As yet, the exact mo-
lecular mechanism of action underlying the apparent promis-
cuity of this blockbuster drug remains elusive. Taking this to-
gether and building on prior experience in constructing and
[
a] D. J. H. De Clercq, Dr. M. D. P. Risseeuw, I. Karalic,
Prof. Dr. S. Van Calenbergh
Laboratory for Medicinal Chemistry
Faculty of Pharmaceutical Sciences, Ghent University
Ottergemsesteenweg 460, 9000 Gent (Belgium)
E-mail: serge.vancalenbergh@ugent.be
[9]
evaluating various TAM-MFCs, we judged this bait would be
a particularly interesting test case for our optimization work.
First, we tried to circumvent any potential limitations related
to the tight binding of MTX to endogenous human dihydrofo-
late reductase (DHFR), which might titrate out a portion of the
fusion compound and induce cellular toxicity through pertur-
bation of the endogenous folate metabolism. As an Escherichia
coli enzyme is employed in MASPIT, we explored trimethoprim
[b] A.-S. De Smet, D. Defever, Prof. Dr. J. Tavernier, Dr. S. Lievens
Cytokine Receptor Laboratory
Department of Medical Protein Research, VIB, Gent and
Department of Biochemistry
Faculty of Medicine and Health Sciences, Ghent University
Albert Baertsoenkaai 3, 9000 Gent (Belgium)
E-mail: sam.lievens@vib-ugent.be
(
TMP; Scheme 2A) as an alternative prokaryote-specific DHFR
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cbic.201402702.
ligand. Whereas MTX binds both prokaryotic and mammalian
ChemBioChem 2015, 16, 834 – 843
834
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
tion azido-TMP reagent is depicted in Scheme 3. The
synthesis began with the generation of a tosyl/azido
[15]
bifunctionalized hexa(ethylene glycol) linker (4).
This spacer was used to alkylate phenol 5, which was
[16]
obtained by acidic hydrolysis of TMP, to afford the
desired ligation handle 6. However, purification of 6
required RP-HPLC to remove a side product formed
by alkylation at the benzhydrilic position (see the
Supporting Information). Presumably, this double al-
kylated TMP analogue was formed by oxidative con-
version of phenol 5 to a reactive pyrimidine imino-
[17]
quinone methide intermediate. The latter can be
converted through resonance stabilization to the cor-
responding para-quinone methide form, which has
a prochiral activated exocyclic methylene group. This
side reaction rendered this alkylation unsuitable for
scale-up. Therefore, we examined alternative alkyla-
tion conditions, varying the base (DBU, Na CO ), leav-
Scheme 1. Outline of the MASPIT system. E. coli dihydrofolate reductase (eDHFR) is fused
to a cytokine receptor (CR) that is rendered signaling-deficient by mutating STAT3 re-
cruitment sites in its cytoplasmic tail (grey dots). A prey protein is tethered to a gp130
CR fragment containing functional STAT3 transcription factor docking sites (black dots).
When a fusion compound consisting of a small molecule of interest (asterisk) coupled to
methotrexate (MTX) is added to the cells, MTX binds to eDHFR, resulting in the com-
pound of interest being displayed as bait. Upon administration of the appropriate cyto-
kine ligand (cyt), the CR–eDHFR chimeric receptor undergoes a conformational change,
activating the associated JAK2 kinases through crossphosphorylation (P). Interaction be-
tween the small-molecule bait and the prey–gp130 fusion protein brings the latter into
proximity of the activated JAK kinases, reconstituting a functional JAK-STAT signaling
pathway. Sequential phosphorylation of STAT3 docking sites on the gp130 chain (P),
STAT3 recruitment, and STAT3 phosphorylation (P) ultimately leads to activated STAT3
dimers that induce the expression of a luciferase reporter gene coupled to a STAT3-
dependent promoter.
2
3
ing group (I, OMs), solvent (DMF), temperature (RT)
and reaction time (16–48 h), but none yielded a more
favorable regioselectivity profile.
Subsequently, first-generation trimethoprim fusion
compounds (TFCs) 7a–c were prepared by ligating
DHFR with similar affinity, TMP displays a 12000-fold binding
preference for E. coli over human DHFR (K =80 pm vs. 960 nm),
i
[12]
thus reflecting its use as a selective antibiotic in the clinic.
Cornish and Sheetz have previously demonstrated the compat-
ibility of the TMP-tag with mammalian systems for intracellular
[
13]
live-cell imaging.
Furthermore, so as to stabilize the ternary complex (CR–
fusion compound–prey chimera) in order to improve the sys-
tem’s sensitivity, we introduced the concept of covalent bond-
ing into the MASPIT assay, which currently relies on reversible
interactions on both ends of the MFCs. As a starting point, we
selectively and covalently immobilized the fusion compound
[
14]
to the CR by using a SNAP-tag-based system. This strategy is
6
centered around the human DNA-repair protein O -alkylgua-
nine-DNA alkyltransferase (hAGT). Johnsson and co-workers
have exploited the low substrate specificity of this enzyme to
covalently label SNAP-tag-fused proteins in vivo with a ligand
of interest by conjugating the latter to the para position of
6
[14]
O -benzylguanine (BG). By fusing hAGT to the cytoplasmic
domain of the CR and synthesizing the appropriate BG fusion
compound (BGFC), we we planned to present a covalently cou-
pled bait small molecule (Scheme 2B).
Results and Discussion
TMP-tag approach
We sought to develop a scalable synthesis of a versatile TMP
reagent with an azide ligation handle as an alternative for the
earlier MTX congener. A synthetic route towards a first-genera-
Scheme 2. A) Folic acid (FA) and the DHFR inhibitors methotrexate and tri-
methoprim; B) Mechanism of hAGT-mediated covalent immobilization of the
6
O -benzylguanine fusion compound (BGFC) to the cytokine receptor (CR).
ChemBioChem 2015, 16, 834 – 843
www.chembiochem.org
835
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Scheme 3. Synthesis of the first-generation TMP-N
1%; d) 5, K CO , acetone, D, 39%.
3 3 2 2 3 3 2 2
reagent 6. a) TsCl, Et N, CH Cl , 08C; b) NaN , DMF, 608C, 64% over two steps; c) TsCl, Et N, CH Cl , 08C,
9
2
3
[
18]
azido reagent 6 by CuAAC
to the previously described
Varying the solvent (DMF, acetone), temperature (408C, D) and
reaction time (6–60 h) had negligible effects on the regioselec-
tivity and reaction progress/yield. However, encouraged by the
significant positive trend we observed earlier for 6 upon
switching to a softer counterion for the carbonate base (K CO
[
9]
[9]
alkyne-functionalized tamoxifen, reversine and simvasta-
[19]
tin, respectively (Figure 1). MASPIT evaluation of these TMP
conjugates against their established primary target preys
showed decreased EC50 values with respect to the correspond-
ing MFCs (Figure 1). These data imply a successful improve-
ment in the system’s sensitivity by introducing TMP as an alter-
native immobilizing anchor moiety, despite the fact that it ex-
2
3
vs. Na CO ) and further guided by a procedure by Chen
2
3
[20]
et al.,
we ultimately opted for Cs CO3 which successfully
2
gave access to acid 9. The latter was subsequently condensed
[
12]
[9]
hibits lower affinity for eDHFR than MTX (K =1.0 pm).
with PEG-4 azidoamine 10 in the presence of 2-(2-pyridon-1-
i
In a next step, we explored a different alkylation strategy
starting from phenol 5 (Scheme 4). Initial alkylations using
ethyl 5-bromovalerate in combination with K CO consistently
yl)-1,1,3,3-tetramethyluronium tetrafluoroborate (TPTU) to
obtain a second-generation azido-TMP reagent in satisfactory
yield on a multigram scale. This reagent has the same spacer
length, with respect to the number of atoms and chemical
2
3
provided an intractable mixture of overalkylated products.
Figure 1. Evaluation of small-molecule TFCs 7a–c (TMP I, *) and 12 (TMP II,
) against the corresponding MFCs (MTX, *) for binding to their primary target
&
protein in MASPIT. A) Tamoxifen–ER1; B) reversine–TTK; C) simvastatin–HMGCR. Luciferase signals are expressed as fold induction relative to a control sample
treated with cytokine without fusion compound. In (A), the signal decrease at the highest concentrations is due to the cellular toxicity of the tamoxifen
fusion compound (data not shown).
ChemBioChem 2015, 16, 834 – 843
www.chembiochem.org
836
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Scheme 4. Synthesis of the second-generation TMP-N
3
reagent 11 and TAM-TFC 12. a) Ethyl 5-bromovalerate, Cs
2
CO
3
, DMF, 708C, 66%; b) i: NaOH, MeOH; ii:
[9]
HCl, 92%; c) 10, TPTU, Et N, DMF, 78%; d) alkyne-functionalized tamoxifen, CuSO , Na ascorbate, TBTA, H O/tBuOH, 808C, 56%.
3
4
2
bonds, as the original 6 and was analogously click-coupled to
the tamoxifen model bait to generate the second-generation
TAM-TFC 12.
Both generations of TAM-TFCs in MASPIT yielded approxi-
mately coinciding curves, again shifted towards lower EC50
values compared to the original MFC (Figure 1A). Hereby, we
demonstrated the biological equivalence of the scalable 11 to
transcription activation as efficiently as the corresponding MTX
probe, whereas Bertozzi concluded exactly the opposite for
the SLF bait (a synthetic analogue of FK506). They postulated
that the disparity in activity between the two CID anchors
might be attributed, for example, to different cell-permeability
properties. To our knowledge, this aspect has not been experi-
mentally clarified in the context of compound profiling strat-
egies so far. Hence, in an effort to elucidate the origin of the
superior performance of TMP fusions in the assay, we studied
the uptake of MTX- versus TMP-linked fluorophores in
HEK293T cells, the cell line employed in MASPIT. The known
azido-MTX and optimized azido-TMP reagents were fused to
6
, and confirmed the increase in sensitivity for TFCs. Moreover,
next to its superior behavior in the MASPIT assay, the newly de-
veloped second-generation TMP reagent offers a number of
important advantages over the existing MTX anchor from
a chemical perspective. The TMP reagent lacks chirality and
shows increased solubility, thus making it in general more
practical. Furthermore, the reagent has enhanced stability in
CuAAC reactions, whereas, under typical conditions, the azido-
MTX reagent suffered from significant degradation to the cor-
responding amine. This is possibly due to residual trifluoroace-
tic acid in the material from cleavage of the a-tert-butylester
[23]
an alkyne-functionalized BODIPY analogue
by CuAAC to
yield 13 and 14, respectively (Figure 2). Subsequently, the per-
meability of both conjugates was tracked in a fluorescence-ac-
tivated cell-sorting (FACS) experiment. First, we measured the
fluorescence intensity of both fusion molecules in solution in
order to exclude the possibility that the inherent fluorescence
of the BODIPY fluorophore was affected by fusion to MTX or
TMP. As the concentration–fluorescence curves of both BODIPY
fusions closely overlap, this does not seem to be the case (Fig-
ure 2A). Next, HEK293T cells were incubated for a fixed time
(15 min) with increasing concentrations of either BODIPY
fusion molecule. The mean fluorescence of the viable cell
subset was measured by FACS, and showed a dose-dependent
increase that was significantly higher for the TMP-linked fluoro-
phore than for the MTX fusion molecule (Figure 2B). Addition-
[9]
precursor. In the case of TMP, which does not require these
protecting-group manipulations, CuAACs can be performed
without the formation of degradation or by-products, thereby
facilitating purification of the final conjugates.
[
21]
[22]
Previously, Cornish and Bertozzi obtained contradictory
outcomes when comparing the effectiveness of MTX- and
TMP-based chemical inducers of dimerization (CIDs) in the
same yeast three-hybrid system. For the dexamethasone
ligand, Cornish found that the TMP-based CID did not induce
ChemBioChem 2015, 16, 834 – 843
www.chembiochem.org
837
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Figure 2. Evaluation of the membrane permeability of BODIPY conjugates of MTX (13; *) and TMP (14; *). A) Dose–response curves showing the inherent
fluorescence of the conjugates. B) Dose–response curves of the cellular uptake of the BODIPY conjugates. The graph shows the mean fluorescence as mea-
sured by FACS analysis of cells stained with increasing concentrations of the conjugates. C) Kinetic analysis of cellular staining with the MTX– or TMP–BODIPY
fusion molecules. Mean fluorescence measured by FACS is plotted against time after the addition of 1 mm of either conjugate.
ally, the rate of dye fusion molecule uptake was followed as
a function of time at a set concentration of 1 mm. Cells turned
out to take up the BODIPY conjugates rapidly, with maximal
fluorescence being reached within 30 s (Figure 2C). Consistent
with the dose–response analysis, both the rate of conjugate
uptake and the plateau level of fluorescence were markedly
higher for the TMP–BODIPY fusion. These results indicate that
the increased sensitivity obtained when using TMP fusion com-
pounds in MASPIT can be attributed, at least partly, to a signifi-
cantly higher membrane permeability than for MTX-linked
compounds, although we cannot exclude that this trend might
be affected by the nature of the bait.
sic thermal degradation issues that result in the elimination of
the terminal azidoethyl group (see the Supporting Information)
and make it less attractive than the optimized TMP reagent. In
order to enable BGFC incorporation in the MASPIT system, in
the plasmid encoding the CR fusion protein, the eDHFR coding
sequence was replaced by a DNA fragment encoding a hAGT
mutant that exhibits increased activity towards BG and which
has been optimized with regard to mammalian codon usage
(see the Experimental Section).
The SNAP-tag MASPIT version was evaluated for the interac-
tion between tamoxifen and its primary target, ER1. Cells ex-
pressing both the CR-hAGT and ER1 prey fusion proteins and
treated with the cytokine ligand and increasing concentrations
of the tamoxifen–BGFC 20 exhibited a dose-dependent in-
crease in luciferase reporter activity (Figure 3). This indicated
that a ternary complex containing the hAGT and ER1 fusion
proteins and the tamoxifen–BGFC is indeed formed, likely in-
volving a covalent bond between hAGT and the BG conjugate.
However, the induction window turned out to be significantly
lower (sixfold vs. 177-fold) and the EC markedly higher (2.4”
SNAP-tag approach
6
To implement the SNAP-tag strategy, we needed a general O -
benzylguanine (BG) reagent, para-substituted with a PEG linker
and with a terminal azide ligation handle. Therefore, the azido
[
24]
spacer 3 was first benzylated and then purinated with 15
and 18,
[
25]
respectively (Scheme 5). Subsequent CuAAC be-
50
À7
À8
tween the resulting BG-based building block 19 and alkyne-
functionalized tamoxifen readily afforded the desired TAM–
BGFC 20. However, the BG-PEG -N reagent suffers from intrin-
10 m vs. 2.5”10 m) than in the case of the TMP-tag setup.
Note, we tested a variety of experimental conditions (varying
CR-hAGT and ER1 prey expression levels and time between
6
3
ChemBioChem 2015, 16, 834 – 843
www.chembiochem.org
838
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Scheme 5. Synthesis of BG-PEG
tionalized tamoxifen, CuSO , Na ascorbate, TBTA, H O/tBuOH, 808C, 48%.
4 2
6
-N
3
reagent 19 and TAM-BGFC 20: a) 15, NaH, DMF, 08C, 59%; b) HF·pyr, THF, 80%; c) 18, KOtBu, DMF, 25%; d) alkyne-func-
[9]
system’s sensitivity by introducing trimethoprim as an alterna-
tive immobilizing anchor moiety. This improvement is possibly
due to its significantly higher membrane permeability than
that of MTX-based fusion compounds. In addition, we present-
ed a scalable synthesis of a versatile TMP reagent that proved
superior to the original MTX probe with respect to solubility
and stability under various reaction conditions. In a next step,
our newly developed second-generation azido-TMP reagent
will be applied to uncovering new intracellular targets of small
[27]
molecules of interest by MASPIT cell-array screening.
Figure 3. Evaluation of the interaction of TAM-BGFC 20 with ER1 in the
SNAP-tag MASPIT setup. The graph shows the fold change in luciferase activ- Experimental Section
ity for a concentration gradient of TAM-BGFC relative to a control treated
with the cytokine ligand without TAM-BGFC. The signal decrease at the
Synthesis
highest concentration is due to the cellular toxicity of the tamoxifen fusion
compound (data not shown).
General: All reactions were performed under nitrogen and at am-
bient temperature, unless stated otherwise. Reagents and solvents
were purchased from Sigma–Aldrich, Acros Organics, or TCI
Europe, and used as received. Reactions were monitored by thin-
layer chromatography on TLC aluminum sheets (Macherey–Nagel,
Alugram Sil G/UV254) with detection by spraying with a solution of
BGFC addition and luciferase readout) for the interaction be-
tween the tamoxifen–BGFC and the ER1 prey, as well as several
other tamoxifen (off- or on-) target proteins recently identified
by our group (S.L., D.J.H.D.C., S.V.C., J.T., unpublished results),
and in none of these cases did the SNAP-tag approach perform
better than the MTX- or TMP-tag-based assay (data not
shown). There might be a number of reasons for this, including
À1
À1
(NH
)
Mo
O
24·4H
O (25 gL ) and (NH
)
Ce(SO
)
·2H
O (10 gL ) in
4
6
7
2
4
4
4
4
2
À1
À1
H SO (10%) or KMnO (20 gL ) and K CO (10 gL ) in water fol-
2
4
4
2
3
lowed by charring. Column chromatography was performed man-
ually with 60 Š silica gel (Grace, Davisil, 40–63 mm) and/or automat-
ically on a Grace Reveleris X2 flash system equipped with disposa-
ble silica gel cartridges (Grace, Reveleris). LC-MS analyses were car-
ried out on a Waters Alliance 2695 XE separation Module by using
a Phenomenex Luna reversed-phase C18 column (100”2.00 mm,
[
14]
titration of the supplied BGFC by endogenous AGT and CR-
[
26]
hAGT fusion protein degradation upon alkyl transfer.
3
mm) and a gradient system of HCOOH in H O (0.1%, v/v)/HCOOH
2
À1
in CH CN (0.1%, v/v) at a flow rate of 0.4 mLmin . High-resolution
3
Conclusions
spectra were recorded on a Waters LCT Premier XE Mass spectrom-
1
13
eter. H and C NMR spectra were measured on a Varian Mercury-
00BB (300/75 MHz) spectrometer. NMR solvents were purchased
from Euriso-top. Chemical shifts are given in ppm relative to tetra-
In this study, we evaluated two approaches, that is, TMP- and
SNAP-tag, to increase the sensitivity of MASPIT. Unexpectedly,
the implementation of the SNAP-tag approach, in which the
bait fusion molecule is coupled covalently to the CR chimera,
did not yield the hypothesized increase in sensitivity. On the
other hand, we have demonstrated a clear improvement in the
3
1
13
methylsilane ( H NMR) or CDCl , CD OD or SO(CD ) ( C NMR) as
3
3
3 2
internal standards. Preparative TLC purification was carried out on
glass-backed Uniplate TLC plates (Analtech, Silica gel GF, UV254,
20”20 cm, 2000 mm). Preparative HPLC purifications were carried
ChemBioChem 2015, 16, 834 – 843
www.chembiochem.org
839
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
out by using a Laprep preparative RP-HPLC system equipped with
a Phenomenex Luna C18 column (21.20”250 mm, 5 mm) with a
gradient system of HCOOH in H O (0.2%, v/v)/CH CN at a flow rate
MeCN, 15 min run); HRMS: calcd for C H N O : 481.2627
[M+2H] , found: 481.2622.
53
70
8
9
2
+
2
3
À1
First-generation reversine TFC (7b): Azide 6 (67.9 mg, 120 mmol,
of 17.5 mLmin .
1
.7 equiv) was taken up in H O/tBuOH (2.1 mL, 1:1, v/v), and
2
[9]
alkyne-functionalized reversine
(30.3 mg, 70 mmol, 1.0 equiv),
a-Azido,a-deoxyhexa(ethylene glycol) (3): Triethylamine (8.4 mL,
0.4 mmol) and TsCl (5.62 g, 29.5 mmol) were added to an ice-
CuSO (28.0 mL, 0.5m, 0.2 equiv), and Na ascorbate (140 mL, 0.5m,
6
4
1
.0 equiv) were added. The resulting mixture was charged with
cooled solution of hexa(ethylene glycol) (9.58 g, 33.9 mmol) in an-
hydrous CH Cl (150 mL). This solution was stirred overnight, and
[28]
a catalytic amount of both Et N and TBTA and heated to 808C
3
2
2
with vigorous stirring. Upon completion of the reaction (88 h), the
solution was cooled to RT and concentrated in vacuo. The residue
was purified by preparative RP-HPLC (10–100% MeCN) to yield the
title compound (36.6 mg, 37 mmol, 53%) as a white amorphous
the temperature was allowed to rise to RT. The reaction mixture
was concentrated in vacuo, and the crude tosylate was taken up in
DMF (250 mL). Sodium azide (6.71 g, 103.2 mmol) was added, and
the mixture was vigorously stirred overnight at 608C. The mixture
was concentrated in vacuo, and the residue was purified by silica
gel chromatography (MeOH/CH Cl 2:98, v/v) to yield the title com-
solid. LC-HRMS: t =5.30 min (10–100% MeCN, 15 min run); HRMS:
R
3
+
calcd for C H N O : 332.8557 [M+3H] , found: 332.8524.
49 72 15
8
2
2
pound (5.80 g, 18.9 mmol, 64%) as a pale yellow viscous liquid.
First-generation simvastatin TFC (7c): Azide
6
(84.8 mg,
1
H NMR (300 MHz, CDCl ): d=3.74–3.69 (m, 2H), 3.69–3.64 (m,
3
1
50 mmol, 1.5 equiv) was taken up in H O/DMF (3.3 mL, 1:1, v/v),
2
[19]
1
1
6
8H), 3.62–3.58 (m, 2H), 3.39 (t, J=5.0 Hz, 2H), 3.05 ppm (brs,
and alkyne-functionalized simvastatin
(42.9 mg, 100 mmol,
1
3
H); C NMR (75 MHz, CDCl ): d=72.5, 70.5–70.3 (wide peak), 70.2,
3
1
.0 equiv), CuSO4 (40.0 mL, 0.5m, 0.2 equiv), and Na ascorbate
9.9, 61.5, 50.5 ppm; HRMS: calcd for C H N O Na: 330.1636
12
25
3
6
(
200 mL, 0.5m, 1.0 equiv) were added. The resulting mixture was
+
[M+Na] , found: 330.1649.
[28]
charged with a catalytic amount of TBTA and heated to 758C for
6 h with vigorous stirring. After being cooled to RT, the solution
1
a-Azido,a-deoxy,w-p-toluenesulfonylhexa(ethylene glycol) (4):
was concentrated in vacuo. The residue was purified by prepara-
tive RP-HPLC (30–50% MeCN) to yield the title compound
Alcohol 3 (1.53 g, 5.0 mmol) was dissolved in anhydrous CH Cl2
2
(
1
25 mL), and the solution was cooled on ice. Triethylamine (2.1 mL,
5.1 mmol) and TsCl (2.38 g, 12.5 mmol) were added. The resulting
(16.7 mg, 17 mmol, 17%) as a white amorphous solid. LC-HRMS:
tR =7.39 min (82.03% area under the curve (AUC; see the Support-
mixture was stirred overnight, and the temperature was allowed to
rise to RT. The mixture was concentrated in vacuo, and the residue
was purified by silica gel chromatography (MeOH/CH Cl 1–3%, v/
ing Information), 10–100% MeCN, 15 min run); HRMS: calcd for
2
+
C H N O : 497.7784 [M+2H] , found: 497.7700.
51
77
7
13
2
2
v) to yield the title compound (2.10 g, 4.5 mmol, 91%) as a brown-
Ethyl 5-(p-trimethoprimoxy)valerate (8): Cs CO3 (11.80 g,
2
1
ish-yellow viscous liquid. H NMR (300 MHz, CDCl ): d=7.79 (d, J=
3
3
1
6.2 mmol, 2.0 equiv) and ethyl 5-bromovalerate (2.87 mL,
8.1 mmol, 1.0 equiv) were added to a solution of phenol 5
8
3
.4 Hz, 2H), 7.35 (d, J=8.4 Hz, 2H), 4.15 (t, J=5.0 Hz, 2H), 3.70–
.56 (m, 20H), 3.37 (t, J=5.0 Hz, 2H), 2.44 ppm (s, 3H); C NMR
[16]
13
(5.00 g, 18.1 mmol, 1.0 equiv) in DMF (250 mL). The resulting solu-
(
75 MHz, CDCl ): d=144.6, 132.8, 129.7, 127.7, 70.5–70.2 (wide
3
tion was heated to 708C for 7 h. After being cooled to RT, the mix-
ture was concentrated in vacuo, and the residue was purified by
silica gel chromatography (MeOH/CH Cl 0–13%, v/v) to yield the
peak), 69.8, 69.1, 68.4, 50.4, 21.4 ppm; HRMS: calcd for
+
C H N O S: 479.2170 [M+NH ] , found: 479.2176.
1
9
35
4
8
4
2
2
title compound (4.81 g, 11.9 mmol, 66%) as an off-white solid.
First-generation TMP-N reagent (6): K CO (138 mg, 1.00 mmol)
3
2
3
1
H NMR (300 MHz, CD OD): d=7.51 (s, 1H), 6.50 (s, 2H), 4.11 (q, J=
3
and PEG linker 4 (369 mg, 0.80 mmol) were added to a solution of
[16]
7.1 Hz, 2H), 3.89 (t, J=6.0 Hz, 2H), 3.76 (s, 6H), 3.62 (s, 2H), 2.38 (t,
J=7.2 Hz, 2H), 1.87–1.75 (m, 2H), 1.75–1.64 (m, 2H), 1.23 ppm (t,
phenol 5 (138 mg, 0.50 mmol) in acetone (5 mL). The resulting
mixture was refluxed at 758C for 60 h with vigorous stirring. After
being cooled to RT, the mixture was concentrated in vacuo. The
residue was repeatedly purified by silica gel chromatography
13
J=7.1 Hz, 3H); C NMR (75 MHz, CD OD): d=175.5, 164.4, 163.1,
3
1
3
,
55.8, 154.8, 136.7, 136.2, 108.1, 106.7, 73.6, 61.4, 56.5, 34.7, 34.5,
+
0.4, 22.7, 14.6 ppm; HRMS: calcd for C H N O : 405.2133 [M+H]
20
29
4
5
(
MeOH/CH Cl 0–15%, v/v) to give a semipure product, which was
2 2
found: 405.2113.
further purified to homogeneity by preparative RP-HPLC (10–100%
MeCN) to yield the title compound (111.2 mg, 197 mmol, 39%) as
5
-(p-Trimethoprimoxy)valeric acid (9): A solution of ester 8
1
a clear, colorless oil. H NMR (300 MHz, CD OD): d=7.35 (s, 1H),
3
(370 mg, 0.92 mmol, 1.0 equiv) in MeOH (9 mL) was treated with
NaOH (1.0 mL, 4.0m). The resulting mixture was stirred overnight
at RT, then neutralized by the addition of HCl (1.33 mL, 3.0m); this
gave fine precipitates. The latter were filtered, washed with a mini-
mal amount of cold MeOH and dried overnight under high
vacuum to afford the title compound (317 mg, 0.84 mmol, 92%) as
6
3
.54 (s, 2H), 4.05 (t, J=4.8 Hz, 2H), 3.79 (s, 6H), 3.77–3.71 (m, 2H),
.71–3.60 (m, 20H), 3.35 ppm (t, J=5.0 Hz, 2H); C NMR (75 MHz,
13
CD OD): d=164.9, 161.3, 154.8, 151.8, 136.4, 136.0, 108.8, 106.9,
3
7
3.4, 71.5–71.4 (wide peak), 71.34, 71.27, 71.0, 56.6, 51.8, 34.3 ppm;
LC-HRMS: t =5.90 min (10–100% MeCN, 15 min run); HRMS: calcd
R
+
1
for C H N O : 566.2933 [M+H] , found: 566.2919.
2
5
40
7
8
an off-white solid. H NMR (300 MHz, SO(CD ) ): d=7.51 (s, 1H),
3
2
6
.54 (s, 2H), 6.09 (s, 2H), 5.71 (s, 2H), 3.77 (t, J=6.0 Hz, 2H), 3.71 (s,
H), 3.52 (s, 2H), 2.26 (t, J=6.9 Hz, 2H), 1.71–1.54 ppm (m, 4H);
First-generation tamoxifen TFC (7a): Azide
6
(166.5 mg,
6
2
94 mmol, 2.3 equiv) was taken up in H O/tBuOH (2.5 mL, 1:1, v/v),
13
2
C NMR (75 MHz, SO(CD ) ): d=174.8, 162.23, 162.18, 155.6, 152.9,
[9]
3 2
and alkyne-functionalized tamoxifen
.0 equiv), CuSO4 (50.4 mL, 0.5m, 0.2 equiv), and Na ascorbate
252 mL, 0.5m, 1.0 equiv) were added. The resulting mixture was
charged with a catalytic amount of tris((1-benzyl-1H-1,2,3-triazol-4-
(50.0 mg, 126 mmol,
1
35.7, 134.8, 105.85, 105.76, 71.9, 55.8, 33.8, 33.0, 29.1, 21.3 ppm;
1
(
À
HRMS: calcd for C H N O : 375.1674 [MÀH] , found: 375.1669.
18
23
4
5
Second-generation TMP-N reagent (11): TPTU (2.72 g, 9.2 mmol,
1.2 equiv) and Et N (10.6 mL, 76.3 mmol) were added to a solution
of acid 9 (2.87 g, 7.6 mmol, 1.0 equiv) in DMF (33.5 mL). The result-
ing preactivation mixture was stirred for 5 min at RT. Subsequently,
3
[28]
yl)methyl)amine (TBTA) and heated to 808C for 16 h with vigo-
rous stirring. After being cooled to RT, the solution was concentrat-
ed in vacuo. The residue was purified by preparative RP-HPLC (10–
3
[9]
1
00% MeCN) to yield the title compound (90.9 mg, 95 mmol, 75%)
a solution of spacer 10 (1.58 g, 7.2 mmol, 0.95 equiv) and Et N
3
as a white amorphous solid. LC-HRMS: t =6.96 min (10–100%
(10.6 mL, 76.3 mmol) in DMF (4.5 mL) was added dropwise to this
R
ChemBioChem 2015, 16, 834 – 843
www.chembiochem.org
840
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
mixture. The resulting reaction mixture was stirred for 5 h at RT,
then concentrated in vacuo, and the residue was purified by silica
gel chromatography (MeOH/CH Cl 0–12%, v/v) to yield the title
centrated in vacuo, and the residue was purified by silica gel chro-
matography (MeOH/CH Cl 0–8% v/v) to yield the title compound
2
2
1
(3.58 g, 6.6 mmol, 59%) as a clear, colorless oil. H NMR (300 MHz,
2
2
1
compound (3.27 g, 5.7 mmol, 78%) as
a
beige oil. H NMR
CDCl ): d=7.30 (brs, 4H), 4.73 (s, 2H), 4.55 (s, 2H), 3.69–3.59 (m,
3
(
300 MHz, CD OD): d=7.29 (s, 1H), 6.55 (s, 2H), 3.92 (t, J=6.2 Hz,
22H), 3.38 (t, J=5.0 Hz, 2H), 0.94 (s, 9H), 0.09 ppm (s, 6H);
C NMR (75 MHz, CDCl ): d=140.8, 136.9, 127.7, 126.1, 73.1, 70.7–
3
3
13
2
3
H), 3.80 (s, 6H), 3.68–3.57 (m, 12H), 3.54 (t, J=5.6 Hz, 2H), 3.40–
.32 (m, 4H), 2.28 (t, J=7.2 Hz, 2H), 1.87–1.65 ppm (m, 4H);
70.5 (wide peak), 70.0, 69.3, 64.8, 50.7, 26.0, 18.4, À5.2 ppm; HRMS:
1
3
+
C NMR (75 MHz, CD OD): d=176.1, 164.9, 160.8, 154.9, 150.9,
calcd for C H N O Si: 559.3522 [M+NH ] , found: 559.3511.
3
26 51
4
7
4
1
5
36.8, 135.4, 109.0, 106.9, 73.8, 71.50, 71.49, 71.4, 71.2, 71.0, 70.5,
a-Azido,a-deoxy,w-4-(hydroxymethyl)benzylhexa(ethylene
glycol) (17): hydrogen fluoride–pyridine complex (HF·pyr;
.9 mL, 38.5m, 34.7 mmol, 16.6 equiv) was added dropwise to a so-
6.6, 51.7, 40.3, 36.6, 34.2, 30.5, 23.6 ppm; HRMS: calcd for
+
A
C H N O : 577.3093 [M+H] , found: 577.3090.
2
6
41
8
7
0
lution of compound 16 (1.13 g, 2.1 mmol, 1.0 equiv) in anhydrous
THF (30 mL) in a plastic flask. The resulting mixture was gently
stirred at RT for 6 h, then neutralized with NaOH (10 mL, 4.0m).
The basic solution was concentrated in vacuo, resuspended in
Second-generation tamoxifen TFC (12): Azide 11 (132.6 mg,
2
30 mmol, 2.3 equiv) was taken up in H O/tBuOH (1.6 mL, 1:1, v/v),
2
[9]
and alkyne-functionalized tamoxifen
(39.6 mg, 100 mmol,
1
.0 equiv), CuSO4 (40 mL, 0.5m, 0.2 equiv), and Na ascorbate
MeOH (40 mL), and filtered. The filtrate was dried over Na SO , fil-
tered, concentrated in vacuo, and finally coevaporated twice with
toluene to remove any traces of pyridine. The residue was purified
(
200 mL, 0.5m, 1.0 equiv) were added. The resulting mixture was
2
4
[28]
charged with a catalytic amount of TBTA and heated to 808C for
6 h with vigorous stirring. After being cooled to RT, the solution
1
by silica gel chromatography (MeOH/CH Cl 0–8% v/v) to yield the
was concentrated in vacuo. The residue was purified by prepara-
tive RP-HPLC (10–100% MeCN) to yield the title compound
2
2
title compound (716 mg, 1.68 mmol, 80%) as a clear, colorless oil.
1
H NMR (300 MHz, CDCl ): d=7.30 (brs, 4H), 4.60 (s, 2H), 4.52 (s,
(
54.6 mg, 56 mmol, 56%) as a white amorphous solid. LC-HRMS:
3
2
1
7
H), 3.67–3.57 (m, 22H), 3.34 (t, J=5.0 Hz, 2H), 3.20 ppm (brs,
t =6.98 min (10–100% MeCN, 15 min run); HRMS: calcd for
R
13
2
+
H); C NMR (75 MHz, CDCl ): d=140.6, 137.1, 127.6, 126.6, 72.7,
C H N O : 486.7707 [M+2H] , found: 486.7708.
3
5
4
71
9
8
0.4–70.2 (wide peak), 69.7, 69.1, 64.3, 50.4 ppm; HRMS: calcd for
+
BODIPY MFC (13): The previously described a-tert-butylester-pro-
C H N O : 445.2657 [M+NH ] , found: 445.2673.
20
37
4
7
4
[9]
tected MTX-PEG -N reagent (79.9 mg, 100 mmol) was taken up in
6
3
BG-PEG -N reagent (19): Benzyl alcohol 17 (711 mg, 1.66 mmol,
6
3
TFA/CH Cl (2.2 mL, 1:1, v/v), and the solution was stirred for
2
2
1
(
.0 equiv) was dissolved in DMF (4.2 mL) under nitrogen. KOtBu
746 mg, 6.65 mmol, 4.0 equiv) was added to this solution, and the
mixture was stirred for 30 min at RT, turning dark red. After this
40 min at RT. The mixture was then evaporated, coevaporated
twice with toluene, and concentrated under high vacuum for 1 h.
The resulting deprotected MTX-PEG -N reagent residue was taken
6
3
[25]
time, compound 18 (423 mg, 1.66 mmol, 1.0 equiv) was added
slowly, and the resulting mixture was stirred for 22 h at RT. The
mixture was concentrated in vacuo, and the residue was purified
by silica gel chromatography (MeOH/CH Cl 0–10%, v/v) to yield
up in H O/DMF (2.6 mL, 1:1, v/v), and CuSO4 (120 mL, 0.5m,
2
0
.6 equiv), Na ascorbate (600 mL, 0.5m, 3.0 equiv), and a catalytic
[28]
amount of TBTA were added. Finally, the pH of the resulting re-
2
2
action mixture was adjusted to 8 by the addition of Et N (168 mL,
3
[23]
the title compound (227 mg, 0.41 mmol, 25%) as a pale yellow oil.
1
2 equiv), and BODIPY–alkyne
(108.3 mg, 330 mmol, 3.3 equiv)
1
H NMR (300 MHz, SO(CD ) ): d=7.82 (brs, 1H), 7.48 (d, J=7.8 Hz,
3
2
was added. The reaction mixture was stirred for 24 h at RT and
concentrated in vacuo. The residue was purified by preparative RP-
HPLC (10–100% MeCN) to yield the title compound (63.7 mg,
2
2
1
1
H), 7.35 (d, J=7.8 Hz, 2H), 6.27 (brs, 2H), 5.48 (s, 2H), 4.50 (s,
[29] 13
H), 3.61–3.48 ppm (m, 22H);
C NMR (75 MHz, SO(CD ) ): d=
3 2
59.9 (weak), 159.7, 155.2 (weak), 138.4, 137.8 (weak), 135.9, 128.4,
27.5, 113.5 (weak), 71.7, 69.86–69.78 (wide peak), 69.70, 69.2, 69.1,
5
9 mmol, 59%) as an orange amorphous solid. LC-HRMS: t =
R
7
.08 min (10–100% MeCN, 15 min run); HRMS: calcd for
2
+
66.5, 50.0 ppm; LC-HRMS: t =6.47 min (91.03% AUC, 10–100%
MeCN, 15 min run); HRMS: calcd for C H N O : 561.2780 [M+H] ,
R
C H N O BF : 536.2789 [M+2H] , found: 536.2791.
5
1
71 14
9
2
+
25
37
8
7
BODIPY TFC (14): Azide 11 (57.7 mg, 100 mmol, 1.0 equiv) was
found: 561.2735.
[23]
taken up in H O/DMF (2.6 mL, 1:1, v/v), and BODIPY–alkyne
2
Tamoxifen BGFC (20): Azide 19 (227.3 mg, 405 mmol, 3.2 equiv)
was taken up in H O/tBuOH (2.5 mL, 1:1, v/v), and alkyne-function-
alized tamoxifen (50.0 mg, 126 mmol, 1.0 equiv), CuSO
(
75.5 mg, 230 mmol, 2.3 equiv), CuSO (80 mL, 0.5m, 0.4 equiv), and
4
2
Na ascorbate (400 mL, 0.5m, 2.0 equiv) were added. The resulting
[9]
[28]
(50.4 mL,
4
mixture was charged with a catalytic amount of TBTA, stirred for
9 h at RT, and concentrated in vacuo. The residue was purified by
preparative TLC (MeOH/CH Cl 1:4, v/v with 1.0% NH OH) to yield
0
.5m, 0.2 equiv), and Na ascorbate (252 mL, 0.5m, 1.0 equiv) were
2
added. The resulting mixture was charged with a catalytic amount
2
2
4
[28]
of TBTA and heated to 808C for 16 h with vigorous stirring. After
the title compound (37.0 mg, 41 mmol, 41%) as an orange amor-
being cooled to RT, the solution was concentrated in vacuo. The
residue was purified by preparative RP-HPLC (30–100% MeCN) to
yield the title compound (58.1 mg, 61 mmol, 48%) as an off-white
amorphous solid. LC-HRMS: t =7.89 min (10–100% MeCN, 15 min
run); HRMS: calcd for C H N O : 478.7551 [M+2H] , found:
phous solid. LC-HRMS: t =7.22 min (10–100% MeCN, 15 min run);
R
2
+
HRMS: calcd for C H N O BF : 453.2544 [M+2H]
, found:
4
5
65 10
7
2
4
53.2533.
R
2
+
a-Azido,a-deoxy,w-4-(((TBDMS)oxy)methyl)benzylhexa(ethylene
glycol) (16): Azido spacer 3 (3.63 g, 11.8 mmol, 1.05 equiv) was dis-
solved in DMF (100 mL), and the solution was cooled on ice under
nitrogen. Sodium hydride (60% dispersion in mineral oil, 450 mg,
53 67
9
8
4
78.7526.
Molecular biology
1
1.25 mmol, 1.0 equiv) was added to this solution, and the mixture
Plasmid constructs: The CR-eDHFR chimeric protein consisting of
the full-size mouse leptin receptor F3 mutant fused to E. coli DHFR
is encoded by the pCLL-eDHFR plasmid described previously. The
pCLG-hAGT plasmid encodes the CR-hAGT fusion protein, which is
made up of the extracellular, transmembranal, and membrane-
proximal portions of the mouse leptin receptor tethered to
[24]
was stirred for 10 min. A solution of compound 15
(4.28 g,
[9]
11.8 mmol, 1.05 equiv) in DMF (20 mL) was then added, and the
resulting mixture was stirred overnight; the temperature was al-
lowed to rise to RT. The mixture was then concentrated in vacuo,
resuspended in EtOAc (300 mL), and filtered. The filtrate was con-
ChemBioChem 2015, 16, 834 – 843
www.chembiochem.org
841
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
a mutant hAGT protein. This plasmid was generated by amplifying
the hAGT coding sequence of the pSNAPf vector (New England
Biolabs) by using forward primer 5’-CCCGA GCTCA ATGGA CAAAG
ACTGC GAAAT G-3’ and reverse primer 5’-GGGGC GGCCG CTTAA
CCCAG CCCAG GCTTG CCCAG-3’ to introduce a stop codon down-
stream of hAGT. The amplicon was cloned into the pCLG vector
ing mean BODIPY fluorescence of the viable cells (propidium
iodide negative).
Acknowledgements
[27]
backbone, which has been described previously, by using SacI
and NotI restriction enzymes. The ER1 prey, consisting of full-size
ER1 fused N-terminally to a gp130 receptor fragment containing
multiple STAT3 recruitment sites, was generated by Gateway
recombinatorial cloning (Invitrogen) into a Gateway-compatible
pMG1C vector. To generate this destination vector, first the pMet7-
Flag Gateway destination vector was made by amplifying the Gate-
J.T. was supported by a grant from IUAP P6/36, and is a recipient
of an ERC Advanced grant (CYRE,340941). S.V.C. and J.T. received
support from the Fund for Scientific Research–Flanders (FWO-V).
D.D.C. obtained a PhD grant from the Special Research Fund of
Ghent University (BOF).
[27]
way cassette from the Gateway-compatible pMG1 vector
by
Keywords: dimerization · immobilization · MASPIT · target
identification · trimethoprim conjugates
using forward primer 5’-CCCCA ATTGA CAAGT TTGTA CAAAA
AAGC-3’ and reverse primer 5’-GGGTC TAGAC TACTT ATCGT CGTCA
TCCTT GTAAT CTTTA ATTAA AACCA CTTTG TACAA GAAAG C-3’ (the
latter containing the Flag epitope coding sequence), and inserting
[1] D. C. Swinney, J. Anthony, Nat. Rev. Drug Discovery 2011, 10, 507–519.
[2] A. L. Hopkins, Nat. Chem. Biol. 2008, 4, 682–690.
[3] G. Jin, S. T. C. Wong, Drug Discovery Today 2014, 19, 637–644.
[7a]
this amplicon into the pMet7 expression vector by ligation of
the MfeI-XbaI restriction enzyme product into the EcoRI-XbaI cut
pMet7 vector. Next, the gp130 coding sequence of the pMG1
[
4] B. F. Cravatt, A. T. Wright, J. W. Kozarich, Annu. Rev. Biochem. 2008, 77,
383–414.
[27]
vector was amplified by using forward primer 5’-CCCTT AATTA
ACGGA GGGAG TATCT CGACC GTGGT ACACA GTG-3’ and reverse
primer 5’-GGGTT AATTA ACCCC TGAGG CATGT AGCCG C-3’, which
was inserted into the PacI site of the pMet7-Flag Gateway destina-
tion vector. Finally, the ER1 open reading frame was transferred
into the pMG1C destination vector from the ER1 entry clone of the
[
[
5] U. Rix, G. Superti-Furga, Nat. Chem. Biol. 2009, 5, 616–624.
6] S. Ziegler, V. Pries, C. Hedberg, H. Waldmann, Angew. Chem. Int. Ed.
2
013, 52, 2744–2792; Angew. Chem. 2013, 125, 2808–2859.
[
7] a) S. Eyckerman, A. Verhee, J. Van der Heyden, I. Lemmens, X. Van Os-
tade, J. Vandekerckhove, J. Tavernier, Nat. Cell Biol. 2001, 3, 1114–1119;
b) S. Lievens, F. Peelman, K. De Bosscher, I. Lemmens, J. Tavernier, Cyto-
kine Growth Factor Rev. 2011, 22, 321–329.
[30]
hORFeome collection. The pMG1-TTK prey plasmid encoding the
[9]
C-terminal fusion of full-size TTK to the gp130 fragment and the
pMG1-HMGCR plasmid coding for the C-terminal fusion of the
[8] M. Caligiuri, L. Molz, Q. Liu, F. Kaplan, J. P. Xu, J. Z. Majeti, R. Ramos-
Kelsey, K. Murthi, S. Lievens, J. Tavernier, N. Kley, Chem. Biol. 2006, 13,
711–722.
[19]
statin-binding cytoplasmic domain to the gp130 fragment have
been described elsewhere, as has the STAT3-dependent firefly luci-
[
9] M. D. P. Risseeuw, D. J. H. De Clercq, S. Lievens, U. Hillaert, D. Sinnaeve,
F. Van den Broeck, J. C. Martins, J. Tavernier, S. Van Calenbergh, Chem-
MedChem 2013, 8, 521–526.
[7a]
ferase reporter pXP2d2-rPAPI-luciferase.
[10] C. K. Osborne, N. Engl. J. Med. 1998, 339, 1609–1618.
[11] E. P. Gelmann, J. Natl. Cancer Inst. 1996, 88, 224–226.
[12] J. R. Appleman, N. Prendergast, T. J. Delcamp, J. H. Freisheim, R. L. Blak-
ley, J. Biol. Chem. 1988, 263, 10304–10313.
MASPIT assays: HEK293T cells were cultured in 96-well microtiter
plates in Dulbecco’s modified Eagle’s medium supplemented with
1
0% fetal calf serum, incubated at 378C, under 8% CO , and trans-
2
fected with pCLL-eDHFR or pCLG-hAGT CR fusion protein plasmid
10 ng per 96 wells), pMG1C-ER1, pMG1-TTK, or pMG1-HMGCR
prey plasmid (100 ng per 96 wells) and pXP2d2-rPAPI-luciferase
[
13] L. W. Miller, Y. Cai, M. P. Sheetz, V. W. Cornish, Nat. Methods 2005, 2,
(
255–257.
[14] A. Keppler, S. Gendreizig, T. Gronemeyer, H. Pick, H. Vogel, K. Johnsson,
(
5 ng per 96 wells) by applying a standard calcium phosphate
Nat. Biotechnol. 2002, 21, 86–89.
15] M. J. Hynes, J. A. Maurer, Angew. Chem. Int. Ed. 2012, 51, 2151–2154;
Angew. Chem. 2012, 124, 2193–2196.
[27]
[
[
[
[
transfection method, as described earlier.
Twenty-four hours
À1
after transfection, cells were stimulated with leptin (100 ngmL
final concentration) alone or in combination with the indicated
concentration of fusion compound. After another 24 h, luciferase
activity was measured by using the Luciferase Assay System kit
16] M. J. C. Long, Y. Pan, H.-C. Lin, L. Hedstrom, B. Xu, J. Am. Chem. Soc.
2
011, 133, 10006–10009.
17] W. G. Lai, N. Zahid, J. P. Uetrecht, J. Pharmacol. Exp. Ther. 1999, 291,
92–299.
2
(
Promega) on a Envision plate reader (PerkinElmer). Luciferase data
18] V. V. Rostovtsev, L. G. Green, V. V. Fokin, K. B. Sharpless, Angew. Chem.
Int. Ed. 2002, 41, 2596–2599; Angew. Chem. 2002, 114, 2708–2711.
[19] S. Lievens, S. Gerlo, I. Lemmens, D. J. H. De Clercq, M. D. P. Risseeuw, N.
Vanderroost, A.-S. De Smet, E. Ruyssinck, E. Chevet, S. Van Calenbergh, J.
Tavernier, Mol. Cell. Proteomics 2014, 13, 3332–3342.
[20] Z. Chen, C. Jing, S. S. Gallagher, M. P. Sheetz, V. W. Cornish, J. Am. Chem.
Soc. 2012, 134, 13692–13699.
[21] S. S. Gallagher, L. W. Miller, V. W. Cornish, Anal. Biochem. 2007, 363, 160–
are the average of three technical replicates, error bars indicate
standard deviation.
Fluorescence measurement: The fluorescence of the MTX– and
TMP–BODIPY fluorophores was measured by preparing a serial di-
lution of the molecules in PBS in 96-well microtiter plates, and ana-
lyzing the plates in an Envision plate reader equipped with a FITC
filter set. Fluorescence data are the average of three replicates,
error bars indicate standard deviation. The curve in Figure 2A was
fit by using four-parameter nonlinear regression in GraphPad
Prism.
162.
[
22] J. L. Czlapinski, M. W. Schelle, L. W. Miller, S. T. Laughlin, J. J. Kohler, V. W.
Cornish, C. R. Bertozzi, J. Am. Chem. Soc. 2008, 130, 13186–13187.
23] M. Verdoes, U. Hillaert, B. I. Florea, M. Sae-Heng, M. D. P. Risseeuw, D. V.
Filippov, G. A. van der Marel, H. S. Overkleeft, Bioorg. Med. Chem. Lett.
2007, 17, 6169–6171.
[
7
À1
FACS analysis: HEK293T cells at a density of 10 mL were incu-
bated with propidium iodide (3 mm), and the indicated concentra-
tion of MTX– or TMP–BODIPY (1 mm final concentration in the
kinetic experiment in Figure 2C) for the indicated time (15 min in
the dose–response experiment in Figure 2B). Samples were ana-
lyzed by using a FACSCalibur instrument (BD Biosciences), measur-
[
24] G. Gorins, L. Kuhnert, C. R. Johnson, L. J. Marnett, J. Med. Chem. 1996,
3
9, 4871–4878.
25] M. Kindermann, N. George, N. Johnsson, K. Johnsson, J. Am. Chem. Soc.
003, 125, 7810–7811.
[
2
[26] A. Keppler, H. Pick, C. Arrivoli, H. Vogel, K. Johnsson, Proc. Natl. Acad.
Sci. USA 2004, 101, 9955–9959.
ChemBioChem 2015, 16, 834 – 843
www.chembiochem.org
842
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
[
[
[
27] S. Lievens, N. Vanderroost, J. Van der Heyden, V. Gesellchen, M. Vidal, J.
Tavernier, J. Proteome Res. 2009, 8, 877–886.
28] T. R. Chan, R. Hilgraf, K. B. Sharpless, V. V. Fokin, Org. Lett. 2004, 6, 2853–
[30] P. Lamesch, N. Li, S. Milstein, C. Fan, T. Hao, G. Szabo, Z. Hu, K. Venkate-
san, G. Bethel, P. Martin, J. Rogers, S. Lawlor, S. McLaren, A. Dricot, H.
Borick, M. E. Cusick, J. Vandenhaute, I. Dunham, D. E. Hill, M. Vidal, Ge-
nomics 2007, 89, 307–315.
2
855.
29] Note that the small shoulder peaks upfield of the right aromatic
7.35 ppm) and benzylic (4.50 ppm) proton signals (i.e., the PEG attach-
(
ment side) originate from the azidoethyl-eliminated product and ac-
count for approximately 9% contamination, as estimated from the cor-
responding peak integrals.
Received: December 12, 2014
Published online on February 16, 2015
ChemBioChem 2015, 16, 834 – 843
www.chembiochem.org
843
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim