In Situ Proteome Profiling and Bioimaging Applications of Small-Molecule Affinity-Based Probes Derived from DOT1L Inhibitors

Article abstract of DOI:10.1002/chem.201600259

DOT1L is the sole protein methyltransferase that methylates histone H3 on lysine 79 (H3K79), and is a promising drug target against cancers. Small-molecule inhibitors of DOT1L such as FED1 are potential anti-cancer agents and useful tools to investigate the biological roles of DOT1L in human diseases. FED1 showed excellent in vitro inhibitory activity against DOT1L, but its cellular effect was relatively poor. In this study, we designed and synthesized photo-reactive and "clickable" affinity-based probes (AfBPs), P1 and P2, which were cell-permeable and structural mimics of FED1. The binding and inhibitory effects of these two probes against DOT1L protein were extensively investigated in vitro and in live mammalian cells (in situ). The cellular uptake and sub-cellular localization properties of the probes were subsequently studied in live-cell imaging experiments, and our results revealed that, whereas both P1 and P2 readily entered mammalian cells, most of them were not able to reach the cell nucleus where functional DOT1L resides. This offers a plausible explanation for the poor cellular activity of FED1. Finally with P1/P2, large-scale cell-based proteome profiling, followed by quantitative LC-MS/MS, was carried out to identify potential cellular off-targets of FED1. Amongst the more than 100 candidate off-targets identified, NOP2 (a putative ribosomal RNA methyltransferase) was further confirmed to be likely a genuine off-target of FED1 by preliminary validation experiments including pull-down/Western blotting (PD/WB) and cellular thermal shift assay (CETSA). Minimalist "clickable" probes: Small-molecule probes P1 and P2 based on FED1 (a known DOT1L inhibitor) were developed and successfully used in experiments including live-cell imaging, in situ proteome profiling, and off-target identification (see scheme).

Full text of DOI:10.1002/chem.201600259

DOI: 10.1002/chem.201600259
Full Paper
&
Inhibitors
In Situ Proteome Profiling and Bioimaging Applications of Small-
Molecule Affinity-Based Probes Derived From DOT1L Inhibitors
Biwei Zhu,[a] Hailong Zhang,[a] Sijun Pan,[a] Chenyu Wang,[a] Jingyan Ge,[b] Jun-Seok Lee,[c]
and Shao Q. Yao*[a]
Abstract: DOT1L is the sole protein methyltransferase that
methylates histone H3 on lysine 79 (H3K79), and is a promis-
ing drug target against cancers. Small-molecule inhibitors of
DOT1L such as FED1 are potential anti-cancer agents and
useful tools to investigate the biological roles of DOT1L in
human diseases. FED1 showed excellent in vitro inhibitory
activity against DOT1L, but its cellular effect was relatively
poor. In this study, we designed and synthesized photo-reac-
tive and “clickable” affinity-based probes (AfBPs), P1 and P2,
which were cell-permeable and structural mimics of FED1.
The binding and inhibitory effects of these two probes
against DOT1L protein were extensively investigated in vitro
and in live mammalian cells (in situ). The cellular uptake and
sub-cellular localization properties of the probes were subse-
quently studied in live-cell imaging experiments, and our re-
sults revealed that, whereas both P1 and P2 readily entered
mammalian cells, most of them were not able to reach the
cell nucleus where functional DOT1L resides. This offers
a plausible explanation for the poor cellular activity of FED1.
Finally with P1/P2, large-scale cell-based proteome profiling,
followed by quantitative LC-MS/MS, was carried out to iden-
tify potential cellular off-targets of FED1. Amongst the more
than 100 candidate off-targets identified, NOP2 (a putative
ribosomal RNA methyltransferase) was further confirmed to
be likely a genuine off-target of FED1 by preliminary valida-
tion experiments including pull-down/Western blotting (PD/
WB) and cellular thermal shift assay (CETSA).
and transcriptional elongation.[4] DOT1L is critically involved in
tumorigenesis and thus an attractive target for potential
cancer treatments. For example, in mixed lineage leukemia
(MLL), the translocation of the MLL gene generates fusion pro-
teins with AF10, AF4, and ENL proteins in which DOT1L is re-
cruited.[5–7] This causes mis-regulation of DOT1L’s activity culmi-
nating in increased methylation levels of H3K79 at hox genes,
which are well-recognized markers for MLL-rearranged leuke-
mia.[2,8] Due to DOT1L’s unique structural fold and therapeutic
potential, small-molecule inhibitors have been developed.[9]
Among them, FED1, a S-adenosylmethionine (SAM) analogue,
is a potent DOT1L inhibitor (IC50 =(2.9Æ0.2) nm;[1] Figure 1) ca-
pable of selectively inhibiting DOT1L methylation of H3K79 by
occupying the SAM-binding site. Analogues of FED1 (i.e., EPZ-
5676) have recently entered clinical trials.[10] Relatively poor cel-
lular activities of FED1 and its analogues (e.g., 4 day incubation
of cells with FED1 was needed in cell-based experiments[11]),
have thus far restricted these new chemical agents to be used
more widely in the study of DOT1L and its potential therapeu-
tics. Little is known about why these compounds were ineffec-
tive in situ (e.g., in live cells). Similarly, their potential cellular
off-targets have not been comprehensively studies under
native cellular environments.
Introduction
Protein methyltransferases (PMTs) play essential roles in epige-
netics by the regulation of gene expression and chromatin-de-
pendent signaling through the methylation of chromatin-asso-
ciated substrates (e.g., histones). There are 60 PMTs encoded in
the human genome, including 51 protein lysine methyltrans-
ferases (PKMTs) and 9 protein arginine methyltransferases
(PRMTs).[1] Abnormal activities of PMTs lead to many diseases,
especially cancers. Selective inhibition of protein methyltrans-
ferases is a promising new approach in drug discovery.
Most PKMTs possess a conserved SET domain, except
DOT1L, which has a protein fold closely resembling PRMTs[2]
and is the only PKMT that methylates histone H3 on Lysine 79
(H3K79),[1,3] a chromatin mark associated with active chromatin
[a] B. Zhu, Dr. H. Zhang, S. Pan, C. Wang, Prof. Dr. S. Q. Yao
Department of Chemistry, National University of Singapore
3 Science drive 3, Singapore, 117543 (Singapore)
E-mail: chmyaosq@nus.edu.sg
[b] Dr. J. Ge
Institute of Bioengineering, Zhejiang University of Technology
Hangzhou, 310014 (P.R. China)
[c] Prof. Dr. J.-S. Lee
Molecular Recognition Research Center
Korea Institute of Science and Technology (KIST)
and Department of Biological Chemistry
University of Science & Technology (Republic of Korea)
Activity-based protein profiling (ABPP) is an important
method that enables proteome-wide profiling of target–probe
interactions under native cellular settings.[12–16] By using the so-
called activity-based probes (ABPs) or affinity-based probes
(AfBPs) that were derived from bioactive compounds, this
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201600259.
Chem. Eur. J. 2016, 22, 7824 – 7836
7824
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 1. a) Overall workflow of our probes used in proteome profiling, bioimaging, and pull-down analysis for target identification in this study. Reaction of
the alkyne handle on the probe with rhodamine-azide through click chemistry allowed visualization of target proteins by in-gel fluorescence scanning and
fluorescence microscopy. Target identification was done by pull-down experiments followed by quantitative LC-MS/MS (SILAC) experiments with biotin-azide.
b) Synthetic scheme of probe P1, P2 and control inhibitors FED1, C1, and C2.
strategy has recently become a powerful tool for large-scale
identification of potential “on” and “off” cellular targets of drug
candidates.[13] Up to now, we and others have successfully ap-
plied such “in situ drug profiling” strategies to a number of
biologically important drugs (and candidates), including the
FDA-approved anti-obesity drug Orlistat,[17] several anti-cancer
small-molecule kinase inhibitors,[18–22] and the newly developed
epigenetic protein–protein interaction (PPI) inhibitor JQ-1.[23]
Furthermore, a “clickable” affinity-based probe derived from 3-
Deazaneplanocin A (DzNep), which is a global histone methyl-
ation inhibitor that also targets many other SAM-binding pro-
teins, was recently reported.[24] This probe, named DZ-1, is the
only known in situ-enabled AfBP against PMTs. Due to its rela-
tively weak binding and low selectivity toward most PMTs, this
probe is however not suitable for direct labeling of endoge-
nous PMTs including DOT1L whose expression levels are
known to be quite low in most cell lines. In the present study,
we have successfully developed two new AfBPs, P1 and P2,
which are “minimalist” probes of FED1 (Figure 1). Both in vitro
and in situ experiments indicated both probes were indeed ex-
cellent structural mimics of FED1 and possessed good cell per-
meability and cellular activities in targeting DOT1L. The probes
were subsequently used, in live-cell imaging experiments, to
study their cellular uptake and sub-cellular localization proper-
ties. These results revealed that, whereas both P1 and P2 read-
ily entered mammalian cells, most of them were not able to
reach the cell nucleus where functional DOT1L resides. This
offers a plausible explanation for the poor cellular activity of
FED1. Finally, potential cellular off-targets of FED1 were identi-
fied by carrying out large-scale cell-based proteome profiling
experiments with P1/P2, followed by quantitative LC-MS/MS
(Figure 1a). Amongst the more than 100 candidate off-targets
identified, NOP2 (a putative ribosomal RNA methyltransferase)
was further confirmed to be likely a genuine off-target of FED1
by preliminary validation experiments.
Chem. Eur. J. 2016, 22, 7824 – 7836
www.chemeurj.org
7825
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
As shown in Figure 1b, the synthesis of P1 commenced
from 1, which was installed with a trifluoromethyl ketone
through a Grignard reaction. Subsequent nucleophilic aromatic
substitution of the product by NaN3 yielded 3, which was then
condensed with hydroxylamine to form an oxime intermediate,
whose OH group was directly tosylated to afford 4. Then liquid
ammonia was added to produce diaziridine 5. The diazirine 6b
was generated first by the reduction of the azide followed by
oxidation of the diaziridine. After that, by utilizing the coupling
agent triphosgene, 6(a,b) was first coupled with 11 to form
Results and Discussion
Design and Synthesis of P1 and P2
As a SAM analogue, FED1 may bind to not only DOT1L but
also many other potential proteins that require SAM. Thus, pro-
teome-wide target profiling experiments of FED1 may help
answer its poor cellular activities and identify potential off-tar-
gets. Compartmentalization of cellular proteins is a highly
regulated process, ensuring precise localization of active pro-
teins in designated sub-cellular organelles. Consequently, accu-
rate localization of a drug inside cells is intimately linked to its
potential cellular efficacy. Clearly, a mis-localized FED1 in cells
will significantly diminish the therapeutic effect of this drug
against DOT1L since functional DOT1L is known to localize ex-
clusively in the cell nucleus. To reveal the sub-cellular localiza-
tion properties of FED1, it is imperative that both P1 and P2
were able to faithfully reproduce the drug’s structural and
target-binding properties to deceive the cells. Recently, two
FED1-based probes were reported;[11] with the direct introduc-
tion of a biotin or a fluorophore to FED1, the probes were suc-
cessfully used for in vitro biochemical screening of DOT1L in-
hibitors, but were too bulky to enter live mammalian cells.
They were therefore not suitable for in situ target identification
and cellular localization studies.[13]
7
(a,b). Subsequently, acetal deprotection of 7(a,b) was initiated
by THF/HCl (2:1) to produce the corresponding aldehyde,
which instantaneously cyclized to generate 8(a,b). The reaction
between 8(a,b) and 15(c,d) proceeded through reductive amina-
tion, and upon subsequent removal of the acetal protecting
group, P1 was successfully obtained. FED1/C1/C2 were similar-
ly prepared (see the Supporting Information, Scheme S1 for
details). P2 was prepared through another synthetic route
using some of the intermediates described above. The mini-
malist linker L7 was synthesized as previous reported,[19] and
was used to alkylate the amino group in compound 16 under
basic conditions. After alkylation, the azide group was reduced
with PPh3 to afford compound 18. Finally, compounds 8a and
18 reacted through reductive amination followed by subse-
quent removal of the acetal protecting group, generating
probe P2.
To achieve excellent cell permeability and structural mimics
of FED1, we first strategically retained the core structure of
FED1 in the two AfBPs, P1 and P2 (Figure 1a); specifically P1
was designed by modifying the tert-butyl group located on
the aromatic group in FED1 with trifluoromethyl diazirine, and
the N-isopropyl group with a terminal alkyne. With a high
structural similarity between P1 and FED1, we expected P1 to
maintain most of FED1s wildtype biological activities, and
hence it is qualified as a “minimalist” tagging approach. P2, on
the other hand, was designed by conveniently attaching our
previously developed “minimalist” linker to FED1. We previous-
ly showed such a photo-cross-linker having both an alkyl di-
azirine and a terminal alkyne could effectively minimize poten-
tial interference of the probe upon binding to the target pro-
teins.[19] This strategy has thus far been applied to tagging a va-
riety of bioactive compounds with reasonable success.[19,23,24]
Since both P1 and P2 were carefully designed to resemble
FED1’s core structure with minimal modifications, they should
be suitable for accurate interrogation of FED1 in live cellular
environments. With both probe designs, upon photo-irradia-
tion, the diazirine moiety in each case would generate a highly
reactive carbene, resulting in an effective covalent linkage with
nearby residues from the probe-bound target protein. Subse-
quently, the terminal alkyne in each probe would serve as an
analytical handle that reacts with azide-containing reporters
through copper(I)-catalyzed azide–alkyne cycloaddition
(CuAAC) for visualization or enrichment of the captured pro-
tein–probe complexes.[25–28] To better evaluate how the two
modifications of FED1 to P1 might influence the FED1-protein-
binding properties, two additional control probes, C1 and C2,
were designed, with each possessing either a terminal alkyne
or a diazirine, respectively (Figure 1a).
Molecular modeling
FED1 is a nanomolar inhibitor of recombinant DOT1L protein
in vitro, though its cellular activities were significantly poorer
as earlier discussed. Since the crystal structure of FED1 com-
plexed to DOT1L is available, the molecular docking experi-
ments were carried out with our probes/controls to gain
better insights into their potential bindings with DOT1L. As
shown in Figure 2a, although there were some differences in
the protein-binding modes beween P1/P2 and the wildtype
FED1, the critical binding sites were practically identical,
strongly suggesting that the probes were able to faithfuly re-
capitulate FED1’s binding to DOT1L and the “minimalist”
chemical modifications in our probes designs were appropri-
ate. As expected, the photoreactive group in P1 matched the
tert-butyl group in FED1 by fitting snuggly into the same hy-
drophobic pocket where the tert-butyl group occupied in
DOT1L. The terminal alkynes in P1 and P2 were shown to have
projected towards the solvent-accessible surface of DOT1L,
thus ensuring minimized disruption of the protein–probe inter-
action and facilitating subsequent CuAAC with an azide-con-
taining reporter. Similar binding modes were also observed for
the control inhibitors C1 and C2 (the Supporting Information,
Figure S1).
Labeling efficiency of recombinant DOT1L by P1 and P2
The affinity-based labeling efficiency of the two probes was
first examined with recombinant DOT1L under UV-irradiated
conditions as previously described.[19] Unpurified recombinant
Chem. Eur. J. 2016, 22, 7824 – 7836
www.chemeurj.org
7826
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 2. a) Molecular docking results of probes against DOT1L. FED1, P1and P2 are shown in yellow, green and red, respectively. b) Dose-dependent and
competitive labeling of recombinant DOT1L (1–420, molecular weight: ꢀ48 kDa) overexpressed in BL21 (DE3) pRARE2 E. coli strain by using P1 and P2. C1
(10” concentrated) was used as a competitor. c) ITC affinity determination of P1 and P2 against the purified recombinant DOT1L, giving the corresponding
Kd. d) In situ cellular thermal shift assay (CETSA) of probes/inhibitors in A431 cells. Probe concentration: 5 mm. Cells were incubated for 4 days. Error bar: n=2.
e) Immunoblot assay of H3K79me2 to analyze DOT1L’s methylation activity in A431 cells; the cells were treated with P1, P2 or FED1 at 1, 5, and 10 mm and in-
cubated for 4 days. f) Immunofluorescence (IF) of H3K79me2 to analyze DOT1L’s endogenous methylation activity. Cells were incubated with probes/inhibitors
(5 mm) for 4 days.
DOT1L protein overexpressed from bacteria transformed with
a suitable DOT1L construct was used. For dose-dependent in
vitro bacterial lysate labeling, 50 mg of the freshly prepared
bacterial lysates overexpressing DOT1L (1–420) were incubated
with probes/inhibitors at various concentrations for 2 h at
room temperature, followed by UV irradiation for 20 min
(ꢀ350 nm) on ice and subsequent reaction with RhÀN3
through CuAAC click chemistry for 2 h at room temperature.[18]
The labeling events were finally separated by SDS-PAGE and
analyzed by in-gel fluorescence scanning. For inhibitor-com-
petitive reactions, lysates were pretreated with C1 (10”). The
results showed that the probe-labeled DOT1L protein band
(ꢀ48 kDa) was well labeled by both P1 and P2 in an activity-
dependent manner. The fluorescence of the labeled bands in-
creased with increasing probe concentrations, but decreased
significantly with pretreatment of the competitors FED1, C1,
and C2 (Figure 2b). All three competitors behaved similarly, in-
dicating both C1 and C2 were excellent mimics of the wildtype
inhibitor FED-1. Overall, P1 demonstrated a better DOT1L la-
beling efficiency than P2 under the same labeling conditions.
to measure the Kd between the purified recombinant DOT1L
and each of the probe/inhibitor (Figure 2c). The titration
curves showed that P1 was a slightly better binder than P2
against DOT1L, having the binding constants of Kd,P1 =100 nm
and Kd,P2 =133 nm, respectively. These values are similar to
those previously reported with FED1 derivatives.[11]
To further investigate the physical interaction of our probes
with endogenous DOT1L in live cells, the cellular thermal shift
assay (CETSA) was applied according to published proce-
dures.[29] Considering the relative poor cellular activity of FED1
and its related probes,[11] A431 cells were incubated with the
probes/inhibitors (5 mm) for 4 days. Such conditions were pre-
viously optimized for FED-1 and shown not to cause cell apop-
tosis.[11] Then the cells were irradiated (ꢀ350 nm) on ice for
20 min and heated at different temperature (40–738C, with
38C difference at each point). Subsequently, the cells were
lysed with liquid nitrogen and spun down; the supernatants
were analyzed by Western blotting (WB) to quantify the DOT1L
amount. Here, DMSO and NP were negative controls, whereas
the wildtype inhibitor FED1 was a positive control. As shown
in Figure 2d, both P1 and P2 were engaged in physical bind-
ing to endogenous DOT1L in live cells, in a manner similar to
FED1, resulting in stabilization of DOT1L. This was clearly evi-
dent from the corresponding right shifts as observed in the
fitted curves.
Probe–DOT1L binding affinity determined by isothermal ti-
tration calorimetry (ITC) and cellular thermal shift assay
(CETSA)
To quantitatively study the direct binding affinity of the probes
to DOT1L, isothermal titration calorimetry (ITC) was first used
Chem. Eur. J. 2016, 22, 7824 – 7836
www.chemeurj.org
7827
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Cellular H3K79me2 suppression assays
consistent with the earlier results of CETSA and WB-based
H3K79me2 suppression assay. Taken together, all these lines of
evidence prove that our probes can effectively bind and inhibit
the intended cellular targets of FED1.
As earlier discussed, DOT1L is the only PKMT that methylates
histone H3K79 to produce H3K79me2. Thus, the inhibition of
endogenous DOT1L activity will lead to the suppression of
H3K79me2, which can be used as a marker for the assessment
of DOT1L inhibitory efficiency in live mammalian cells. A431
cells were incubated with FED1/P1/P2 (1, 5, 10 mm) for 4 days,
then harvested followed by histone extraction. The H3K79me2
expression levels were then determined by WB, with the total
histone amount as internal loading control (Figure 2e). The re-
sults showed that the H3K79me2 expression levels decreased
with increasing concentrations of P1/P2 in a manner similar to
FED1. This confirmed that our newly designed probes P1 and
P2 were indeed good mimics of FED1 that possessed similar
cellular activities and caused endogenous DOT1L inhibition
(see the Supporting Information, Figure S6 for further details).
Similar endogenous DOT1L inhibitory efficiency was also
measured in an imaging-based experiments by immunofluor-
escence (IF) measurement of the level of endogenous
H3K79me2 with the corresponding antibody (Figure 2 f). A431
cells were first treated with DMSO or FED1/P1/P2 (5 mm) for 4
days. The cells were then fixed with 3.7% formaldehyde, per-
meabilized by 0.1% Triton X-100, and further treated with anti-
H3K79me2 antibody and a fluorophore-labeled secondary anti-
body. Finally, cells were stained with Hoechst before being
imaged with a confocal fluorescence microscope. Strong im-
munofluorescence signals were detected in DMSO-treated
cells, and the signals were significantly reduced in the pres-
ence of a high concentration of FED1/P1/P2 (Figure 2 f and
the Supporting Information, Figure S7). These results clearly in-
dicated that P1 and P2 were able to inhibit endogenous
DOT1L activities and prevent the methylation of H3K79 to
form H3K79me2 in live mammalian cells. These results are also
Cellular imaging of P1 and P2
Small-molecule-based cellular imaging offers an attractive way
to study the cellular uptake and sub-cellular localization of
a drug, as well as report potential endogenous enzyme–drug
interactions.[30–37] In recent years, small-molecule-based probes
especially dual-purpose probes have been used to successfully
identify and interrogate endogenous cellular targets.[22,23,35–38]
Since FED1, P1, and P2 appeared to possess similar cellular ac-
tivities against DOT1L, we inferred that P1 and P2 might serve
as potential imaging probes to study endogenous DOT1L–
FED1 interactions. To confirm this hypothesis, A431 cells were
treated with DMSO or an optimized amount of P1/P2 (2 mm)
for 4 h, 1 day or 4 days, respectively. Subsequently, cells were
irradiated (ꢀ350 nm, 20 min) on ice and fixed with 3.7% form-
aldehyde, permeabilized by 0.1% Triton X-100, and clicked
with Rh-PEG-N3 by following the published protocols.[17,22] The
same cells were further treated with anti-DOT1L antibodies for
IF and stained with Hoechst, before the measurements were
carried out with confocal fluorescence microscopy under four
different imaging channels (Figure 3a and the Supporting In-
formation, Figure S8). Under the same conditions, P2 consis-
tently produced higher fluorescence signals than P1 (Fig-
ure 3a, panels 5 and 9). In addition, P2 demonstrated better
cellular permeability (the Supporting Information, Figure S8).
As a methyltransferase, functionally active DOT1L is situated in
the nucleus. However, in our imaging results, most of the P1/
P2 signals were shown to have localized outside of the nu-
cleus. This could be due to several reasons: 1) the presence of
Figure 3. a) Confocal fluorescence images of A431 cells labeled with P1 and P2. A total of four imaging channels were employed, including bright-field (DIC),
probe-treated channel (upon click chemistry with Rh-PEG-N3), immunofluorescence (IF) and Hoechst-stained channel (NU). Cells were incubated with probes/
inhibitors (2 mm) for 4 h, 1 day and 4 days. Full images are shown in the Supporting Information. Scale Bar: 10 mm. b) In vitro dose-dependent and competi-
tive proteome labeling profiles of P1 and P2 in A431 cells. c) In situ proteome profiles of P1 and P2 in A431 cells g. d) In vitro PD/WB results of P1/P2 against
endogenous DOT1L. The size of DOT1L is around 185 kDa. A431 cells were treated with P1/P2 (1 and 5 mm) and competitor C1 (10”); cells treated with C1
were negative controls. e) In situ PD/WB results of P1/P2 against endogenous DOT1L. A431 cells were treated with P1/P2 (5 and 10 mm) and competitor C1
(10”).
Chem. Eur. J. 2016, 22, 7824 – 7836
www.chemeurj.org
7828
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
the nuclear envelope might have prevented the probes from
effectively entering the nucleus, and 2) the potential cellular
off-targets in the cytosol may have caused the non-target
binding, effectively trapping the probes from entering the nu-
cleus in sufficient amount. This assumption may be verified by
further proteome profiling and large-scale target identification
experiments (see below).
ly profile their intended cellular target (e.g., DOT1L) may also
be due to the mis-localization of the drugs, as inferred from
our earlier live-cell imaging results.
Proteome-wide target identification and validation
Our earlier results indicated that the probes were not able to
directly visualize labeled DOT1L under in-gel fluorescence
scanning settings, either due to low abundance of endoge-
nous DOT1L or prominent off-target labeling, or both. We
therefore introduced an additional affinity enrichment step to
boost the amount of probe-labeled endogenous DOT1L in our
labeling reactions. Upon in vitro or in situ labeling with the
probes, followed by click chemistry with biotin-N3, the samples
were pulled-down by avidin agarose beads, followed by WB
with anti-DOT1L antibody (Figure 3d and e and the Supporting
Information, Figure S10). Successful detection of the endoge-
nous DOT1L (ꢀ185 KDa) in the P1/P2-labeled proteomes,
under both in vitro and in situ experiments, showed that our
probes indeed were efficient AfBPs of DOT1L capable of cova-
lent labeling/capturing of endogenous DOT1L from complex
mammalian protein lysates and live mammalian cells. The pres-
ence of labeled DOT1L band was highly dependent on the
presence/absence of competitor C1 (10”), indicating the label-
ing was protein activity-dependent, and that C1 was able to
effectively compete with the probe for binding to DOT1L in
complex cellular environments. These results thus indirectly
confirm the successful labeling of endogenous DOT1L by P1/
P2 with good efficiency. Unlike the in vitro experiments, the in
situ results produced significantly weaker labeled DOT1L sig-
nals even after PD/WB experiments (compare Figure 3d and
3e), providing another strong evidence that mis-localization of
the drug in live cells might be the main cause; a conclusion
that was also suggested from our earlier imaging results (Fig-
ure 3a). With these issues in mind, we next carried out large-
scale proteome profiling followed by PD/LC-MS/MS experi-
ments to identify potential off-targets of the drug.
Cell-based proteome profiling of P1 and P2
We next carried out cell-based proteome profiling experiments
under both in vitro (with mammalian cell lysates) and in situ
(with live mammalian cells; Figure 3b, c and the Supporting In-
formation, Figure S9). For in vitro labeling, 100 mg of freshly
prepared A431 cell lysates were treated with P1/P2 at different
concentrations (with or without pretreatment with 10”C1). C1
was chosen for such experiments because it possessed similar
competitive property as FED-1, as shown in Figure 2b, and
also more closely resembled P1 than FED-1) by following the
optimized procedures earlier described for bacterial lysate la-
beling. For in situ labeling, live A431 cells were incubated with
the probe (1 and 5 mm, ÆC1 (10”)) for 4 days. Then the cells
were irradiated and lysed. The rest of the procedures were the
same as the in vitro labeling experiment. In-gel fluorescence
scanning results of the in vitro labeling experiment showed
some distinct fluorescent bands at around 40–70 kDa, but not
at ꢀ185 kDa where the endogenous DOT1L was expected to
be located. This indicates that P1 and P2 were not able to di-
rectly profile the presence of endogenous DOT1L expression
under our current experimental conditions, presumably due to
either the extremely low endogenous expression level of the
enzyme in mammalian cells, or insufficient detection sensitivity
of the probes as well competitive labeling from potential cellu-
lar off-targets. The results of the in situ labeling also showed
that the endogenous DOT1L could not be directly profiled by
P1/P2. Further, most of the fluorescent bands on the SDS-
PAGE gel appeared at regions below 70 kDa, indicating poten-
tial off-targets of the probes/FED1, and were later identified
large-scale PD followed by quantitative LC-MS/MS. In addition,
based on the proteome profiling results, the labeling patterns
of the two probes were similar, especially from the in vitro ex-
periments (Figure 3b). For in situ labeling (Figure 3c), P2 was
surprisingly more active in producing stronger fluorescent
bands. We believe this was due to the difference in the photo-
reactive properties between the trifluoromethyl diazirine in P1
and the alkyl diazirine in P2. Nevertheless, there were clearly
differences in the fluorescently labeled bands from the two
probes under both in vitro and in situ conditions, indicating
the presence of potentially different sets of off-targets. It was
also observed that whereas P1 appeared to possess better la-
beling efficiency in the in vitro experiments, the opposite was
true for P2 in the in situ experiments. This may be attributed
to our “minimalist” linker used in P2, which may have
equipped the probe with a better cellular permeability. On the
other hand, the amino group on the adenine of FED1 may be
important for target binding, which enables P1 to label more
efficiently in vitro. Finally, the inability of both probes to direct-
We carried out stable isotope labeling by amino acids in cell
culture (SILAC) experiments with mammalian cells charged
with either light (L) or heavy (H) amino acids, by following
published procedures (see the Supporting Information for de-
tails).[39,40] Subsequently, in situ labeling of the cells with our
probes were conducted, followed by click chemistry and PD
experiments with protocols optimized earlier. Similar SILAC ex-
periments were carried out to further quantitatively compare
the difference in the labeling performance between P1 and
P2. Briefly, in a typical competitive SILAC experiment, heavy
cells were first treated with the probe (2 mm) followed by UV-
initiated photo-cross-linking. Light cells were similarly treated
(probe/C1, 2/20 mm). Upon cell lysis, equal amounts of the two
resulting lysates were combined into a single SILAC sample,
which was subsequently clicked with biotin-N3, enriched by
avidin agarose beads and separated by SDS-PAGE. The separat-
ed protein extracts were then subjected to in-gel trypsin diges-
tion, and analyzed by quantitative LC-MS/MS (under SILAC set-
tings). The complete LC-MS/MS data are shown in the Support-
ing Information (Tables S2–S4) with select results summarized
Chem. Eur. J. 2016, 22, 7824 – 7836
www.chemeurj.org
7829
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Table 1. Summary of proteins identified in the SILAC PD samples.
Description
P1
P2
Localization
Score
Heavy/Light Score
Heavy/Light
Methyltransferases
isoform 2 of putative ribosomal RNA
methyltransferase NOP2
4.67 15.297
–
–
nucleus; preribosome, large subunit precursor
putative rRNA methyltransferase 3
serine hydroxymethyltransferase
13
1.083
0.581
13.48
47.8
0.777
–
nucleus; preribosome, small subunit precursor
plastid
24.8
Proteins
peptidyl-prolyl cis-trans isomerase A
160.15
49.19
36.32
32.53
2.507
2.893
2.497
2.836
157.76
31.41
27.27
39.17
2.113
2.084
2.209
2.076
cytoplasm; secreted
nucleus
splicing factor 3A subunit 1 isoform 2
thymosin beta-4-like protein 3
cytoplasm
isoform 4 of plasminogen activator inhibitor 1
RNA-binding protein
cytoplasm; nucleus; perinuclear region
28 kDa heat- and acid-stable phosphoprotein
thyroid hormone receptor-associated protein 3
serpin H1
26.42
36.32
20.57
17.68
14.56
2.861
2.497
2.906
2.726
2.627
26.46
27.27
15.51
9.8
2.899
2.209
2.192
3.467
2.362
cytoplasm; membrane
nucleus
endoplasmic reticulum lumen
cell membrane; cell projection
isoform 1 of brain acid soluble protein 1
Na(+)/H(+) exchange regulatory cofactor
NHE-RF1
10.29
cytoplasm; apical cell membrane; endomembrane system;
peripheral membrane protein; cell projection
GUCA1B protein (Fragment)
UPF0556 protein C19orf10
14.21
10.57
3.386
3.306
23.03
10.5
2.169
2.347
cytoplasm
endoplasmic reticulum-golgi intermediate compartment;
extracellular region; endoplasmic reticulum lumen
aspartate aminotransferase, mitochondrial
8.67
2.487
5.45 20.27
19.42 57.631
mitochondria; cytoplasm
isoform 1 of eukaryotic translation
initiation factor 5A-1
4.73 39.559
cytoplasm; nucleus; endoplasmic reticulum membrane;
peripheral membrane protein; cytoplasmic side;
nuclear pore complex
in Table 1 (Figure 4a). Like most large-scale LC-MS/MS experi-
ments, a large number of proteins hits were identified in our
results (around 1000 different proteins). Among them, many
were inevitably highly abundant, nonspecific proteins that
were labeled indiscriminately in live cells due to their “sticky”
nature. As a result, they were labeled by all the compounds
(including negative control probes). These proteins were auto-
matically eliminated as false positives. Of the remaining pro-
teins, we focused on those that met the following criteria:
1) proteins labeled by both probes and at the same time
showing a high SILAC ratio (>2), which may represent the
most potent/genuine cellular targets; 2) proteins having inter-
Figure 4. a) SILAC ratio plots for A431 samples labeled by P1 and P2, respectively. b) In vitro (top) and in situ (bottom) PD/WB validation results of putative
identified off-target ribosomal RNA methyltransferase NOP2. c) In situ cellular thermal shift assay (CETSA) of probes against endogenous NOP2 in live A431
cells. Probe concentration: 5 mm. Cells were incubated with the compound for 4 days before CETSA protocols. Error bar: n=2.
Chem. Eur. J. 2016, 22, 7824 – 7836
www.chemeurj.org
7830
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
esting cellular functions that warrant further considerations; or
3) putative/previously known methyltransferases. It should be
noted that we were unable to detect DOT1L MS signals in our
SILAC experiments, despite successful detection of this target
from our proteome labeling/PD/WB experiments (Figure 3d
and 3e), presumably again due to the low abundance of this
protein in mammalian cells. It should be noted that WB is
more specific and sensitive than LCMS, and it is common that
a protein was successfully detected in WB but failed in LCMS
experiments.[13] Of the proteins positively labeled by both
probes, we further identified 17 proteins that showed high
SILAC ratios (>2) (Figure 4a). The molecular weight of the in-
teresting proteins were also checked, most of which were
around 40–70 kDa; this agreed well with our earlier proteome
profiling results in which most of the fluorescently labeled
bands were also located around the regions of 40–70 kDa.
Among them, the protein that showed the highest SILAC ratio
was identified to be the isoform 1 of eukaryotic translation ini-
tiation factor 5A-1 (highlighted with an arrow in Figure 4a).
Further analysis of the SILAC results indicated that most of
these positively identified “hits” were located in the cytosol,
with diverse biological functions ranging from RNA splicing/
processing to other metabolic functions. These results again
corroborated well with our earlier imaging/proteome labeling
experiments that already indicated most of the internalized
P1/P2 (and therefore FED1) likely ended up in the cytosolic
space of the cells, rather than in the nucleus in which function-
ally active DOT1L was located. The same reason probably also
accounted for the observed poor cellular activities of FED1 (as
well as our probes) in spite of its nanomolar inhibitory activity
against purified DOT1L in vitro.
ther chosen as a representative example for preliminary target
validation by PD/WB and CETSA experiments (Figure 4c and
d). As expected, NOP2 was significantly enriched and detected
only in P1-treated A431 cells under both in vitro and in situ
settings (Figure 4c). In the presence of the competitor C1,
however, most of the labeled NOP2 band was effectively com-
peted away. These result thus confirmed that NOP2 was likely
a genuine off-target of FED1. This conclusion was further sup-
ported by the CETSA experiment, which showed a direct physi-
cal interaction between P1 and endogenous NOP2 in A431
cells (Figure 4d). These stabilizing effects of endogenous NOP2
were also positively observed for FED1, which produced
a larger thermal shift than either that of C1 or C2. However, P2
showed no cellular thermal shift as the negative control
DMSO, which was consistent with the previous PD/WB experi-
ments of NOP2 (Figure 4c and d). These results indicated that
the amino group on the adenine of FED1 may be critical for
the NOP2 binding. At the moment, we do not know if endoge-
nous off-targeting of NOP-2 by FED-1 has any biological conse-
quences. Much more extensive molecular and cellular studies
will need to be performed to answer this question.
Conclusions
As it is the only PKMT known to methylate H3K79, DOT1L and
its potential therapeutics have been extensively studied in
recent years. Several small-molecule DOT1L inhibitors are al-
ready in clinical trials. FED1 is one of the best-known small-
molecule inhibitors against DOT1L, yet its relatively poor cellu-
lar activities have not been properly investigated. In this study,
based on the core structure of FED1, we have successfully de-
signed and synthesized new small-molecule AfBPs, P1 and P2.
During our investigation, we found P2 consistently produced
better signals in live-cell imaging and in situ labeling experi-
ments, presumably due to better cell permeability. Despite
these unexpected results, P1 was superior in most in vitro ex-
periments, suggesting that the amino group on the adenine of
FED1 may be important for target binding. Judging from the
imaging results, both P1 and P2 were mostly trapped in the
cytosolic space of live mammalian cells (A431), and were not
properly localized to the nucleus in which functionally active,
endogenous DOT1L resides. These results thus explain, for the
first time, the most likely reason why FED1 and its derivatives
showed potent inhibition against recombinant DOT1L under in
vitro conditions, but significantly weaker cellular activities
against endogenous DOT1L in cell-based experiments. Such in-
formation is invaluable for the future design of improved
DOT1L-targeting drugs. With our newly designed cell-perme-
able probes, we were able to carry out subsequent cell-based
proteome profiling followed by quantitative LC-MS/MS experi-
ments, to identify potential cellular off-targets of FED1 for the
first time. In addition to unambiguously confirm the mis-locali-
zation of P1/P2 (and thus possibly FED1), our MS results pro-
vided a list of putative off-targets of FED1, including several
methyltransferases. One such hit, ribosomal RNA methyltrans-
ferase (NOP2), was further validated by independent experi-
ments. Further study will focus on improvement of the probe’s
By further comparing the identified protein targets from our
SILAC data, it was found that there were some differences be-
tween the proteomes labeled by P1 and P2, which may be
caused by the subtle structural differences in these two
probes. From the molecular docking results, it was clear that
the binding modes of P1 and P2, though mostly similar, were
also distinctly different in some ways. Therefore, this gave us
a clue that different tagging approaches on the same bioactive
compound may inevitably result in different potential cellular
hits. Hence, to improve the accuracy of target identification,
a better choice might be to label the same drug with as many
different minimalist tags as possible. Such a conclusion is
based on the underlined hypothesis that, in a large-scale drug
profiling experiment, those common proteins positively la-
beled/identified by differently tagged probes (of the same
drug) may stand a better chance to be the genuine cellular tar-
gets of the original drug.[23]
Finally, mindful that FED-1 is a known PKMT-targeting drug,
we decided to closely examine other methyltransferases that
were positively identified from our SILAC experiments (Fig-
ure 4a; see the Supporting Information for complete data). All
the identified methyltransferases showed reasonable protein
scores and SILAC ratios. Among them, isoform 2 of putative ri-
bosomal RNA methyltransferase (NOP2; Table 1, top row) was
successfully detected in P1-labeled SILAC samples, and
showed the highest SILAC ratio. This protein was therefore fur-
Chem. Eur. J. 2016, 22, 7824 – 7836
www.chemeurj.org
7831
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
concentrated in vacuo. The residue was purified by silica gel
column chromatography (hexane, Rf =0.6) to afford 3 (8.1 g, 79%)
as a yellow liquid. 1H NMR (300 MHz, CDCl3): d=8.13 (dd, J=0.9,
8.7 Hz, 2H), 7.19–7.32 ppm (m, 2H); 13C NMR (75 MHz, CDCl3): d=
179.08 (2 JCÀF =34.86 Hz) (s), 147.77 (s), 132.29 (s), 126.48 (s), 116.83
(s), 116.84 (1JCÀF =289.26 Hz) (s).
detection sensitivity and labeling efficiency, as well as making
such profiling and target validation approaches a routine en-
deavor for future chemical and cellular studies of other bioac-
tive compounds.
1-(4-Azidophenyl)-2,2,2-trifluoroethanone oxime (4): Trifluoro-
methyl ketone 3 (4.3 g, 20.0 mmol) was dissolved in 60 mL of EtOH
containing 10 mL of pyridine. Hydroxylamine hydrochloride (1.68 g,
1.2 equiv) was then added, and the mixture was heated at reflux
for 12 h before the solvent was removed under vacuum and the
residue was purified by chromatography (hexane/EtOAc=20:1).
The oxime products were thus obtained as a mixture of isomers
Experimental Section
General information
All chemicals were purchased from commercial vendors and used
without further purification, unless otherwise noted. N,N-Dimethyl-
formamide (DMF) and dichloromethane (CH2Cl2, DCM) were dis-
tilled over CaH2. All non-aqueous reactions were carried out under
nitrogen atmosphere in oven-dried glassware. Reaction progress
1
(3.5 g, 75%). H NMR (300 MHz, CDCl3): d=9.52 (s, 1H), 7.60 (d, J=
8.5 Hz, 2H), 7.05–7.17 ppm (m, 2H); 13C NMR (75 MHz, CDCl3): d=
146.5 (d, J=31.5 Hz), 142.6, 130.5, 122.0, 120.5 (q, J=819.7 Hz),
119.1 ppm.
was monitored by TLC on pre-coated silica plates (Merck 60 F254
,
250 mm thicknesses) and spots were visualized by ceric ammonium
molybdate, ninhydrin, basic KMnO4, UV light, or iodine. Flash
column chromatography was carried out using Merck silica gel
(0.040–0.063). 1H NMR and 13C NMR spectra were recorded on
a Bruker model Avance 300 MHz or DPX-500 MHz NMR spectrome-
ter. Chemical shifts are reported in parts per million relative to in-
ternal standard tetramethylsilane (Si(CH3)4 =0.00 ppm) or residual
The oximes prepared above (3.45 g, 15.0 mmol) were treated with
4-toluenesulfonyl chloride (1.1 equiv) in acetone (50 mL) and tri-
ethylamine (10 mL). After 1 h, the solution was filtered, evaporated
to dryness, and purified by flash chromatography (hexane/EtOAc=
40:1). The product 4 was obtained as a mixture of isomers (5.2 g,
90%). 1H NMR (500 MHz, CDCl3): d=7.82–7.98 (m, 2H), 7.50–7.32
(m, 4H), 7.01–7.18 (m, 2H), 2.47 ppm (d, J=7.4 Hz, 3H); 13C NMR
(126 MHz, CDCl3): d=153.0 (q, J=107.5 Hz), 146.2, 143.8, 131.2,
130.4, 129.8, 129.1, 124.0, 120.6, 119.3, 21.7 ppm.
1
solvent peaks (CDCl3 =7.26 ppm, MeOD=3.31 ppm). H NMR cou-
pling constants (J) are reported in Hertz (Hz) and multiplicity is in-
dicated as follows: s (singlet), d (doublet), t (triplet), m (multiplet),
dd (doublet of doublet), dt (doublet of triplet). MS spectra were re-
corded on a Finnigan LCQ mass spectrometer, a Shimadzu LC-IT-
TOF spectrometer or a Shimadzu LC-ESI spectrometer. Analytical
HPLC was carried out on Shimadzu LC-IT-TOF or LC-ESI systems
equipped with an autosampler, using reverse-phase Phenomenex
Luna 5 mm C18 100 Š 50”3.0 mm columns. Preparative HPLC was
carried out on Gilson preparative HPLC system using Trilution soft-
ware and reverse-phase Phenomenex Luna 5 mm C18 (2) 100 Š 50”
30.00 mm column. 0.1% TFA/H2O and 0.1% TFA/acetonitrile were
used as eluents for all HPLC experiments. The flow rate was
0.6 mLminÀ1 for analytical HPLC and 8 mLminÀ1 for preparative
HPLC. UV/Vis absorption and fluorescence spectra were measured
by using a Shimadzu UV/Vis spectrophotometer and a PerkinElmer
LS50 spectrofluorometer, respectively. All the measurements were
performed at room temperature.
3-(4-Azidophenyl)-3-(trifluoromethyl)diaziridine (5): The tosylox-
ime 4 (3.2 g, 8.3 mmol) was dissolved in diethyl ether (20 mL) and
added to 20–25 mL of liquid NH3 at À788C. The mixture was
stirred for 4 h before being brought to room temperature, over-
night. The mixture was then evaporated to dryness and purified by
flash chromatography (hexane/EtOAc=70:30), to obtain the diaziri-
dine 5 (1.5 g, 77%). 1H NMR (300 MHz, CDCl3): d=7.60 (d, J=
8.5 Hz, 2H), 6.95–7.16 (m, 2H), 2.27 ppm (br, 2H); 13C NMR
(75 MHz, CDCl3): d=142.0, 129.7, 128.3 (d, J=189.7 Hz), 128.1,
123.4 (d, J=276.7 Hz), 119.3 ppm.
4-(3-(Trifluoromethyl)-3H-diazirin-3-yl)aniline (6b): The azide 5
(1.5 g, 6.5 mmol) was dissolved in dry THF (100 mL), before being
cooled to 08C. Subsequently, LiAlH4 (240 mg, 1.0 equiv) was added
portion-wise over 10 min. After the starting material was complete-
ly consumed as monitored by TLC, the reaction was quenched
with saturated NaHCO3 (10 mL), and then extracted with EtOAc
(3”30 mL). The combined organic extracts were dried (Na2SO4)
and concentrated. Purification by flash column chromatography
(hexane/ethyl acetate=2:1) afforded 0.93 g (70%) of the diazirinyl
hydrazine. 1H NMR (300 MHz, CDCl3): d=7.37 (d, J=8.3 Hz, 2H),
6.66 (d, J=8.7 Hz, 2H), 3.79 (s, 2H), 2.70 (s, 1H), 2.15 ppm (d, J=
7.6 Hz, 1H); 13C NMR (75 MHz, CDCl3): d=147.9, 129.2, 125.4, 121.8,
121.0, 114.6 ppm; MS (ESI): m/z calcd for C8H8F3N3: 204.07 [M+H]+;
found: 204.07.
Synthesis of probes and controls
2,2,2-Trifluoro-1-(4-fluorophenyl)ethan-1-one (2): To a solution of
dry THF (100 mL) at À788C was added 4-fluorophenyl magnesium
bromide (20 g, 100.3 mmol) under a closed system flushed with N2.
The reaction was stirred for 15 min at À788C before slow addition
of CF3COOEt (28.4 g, 200.62 mmol). The solution was stirred at
À788C for 2 h before adjusting the pH to 2 with 1m HCl. The or-
ganic layers were extracted with EtOAc twice, after which they
were combined, washed with brine and dried over Na2SO4. The sol-
vents were evaporated in vacuo. The residue was purified by silica
gel column chromatography (hexane, Rf =0.45) to afford 2 (9.5 g,
The diazirinyl hydrazine (0.41 g, 2 mmol) was dissolved in DCM at
08C, followed by addition of triethylamine (TEA) and I2 until the
purple color of I2 persisted. The solution was then stirred at 08C
for 1 h, quenched with Na2S2O3, extracted with EtOAc and purified
by column chromatography (hexane/EtOAc=85:15) to afford com-
1
50%) as a colorless liquid. H NMR (300 MHz, CDCl3): d=8.13–8.17
(m, 2H), 7.22–7.30 ppm (m, 2H); 13C NMR (75 MHz, CDCl3): d=
178.86 (2JCÀF =35.25 Hz) (s), 167.01 (s), 132.96 (s), 118.14 (1JCÀF
=
1
pound 6b (0.30 g, 75%). H NMR (300 MHz, CDCl3): d=7.00 (d, J=
292.5 Hz)(s), 116.49 (s), 116.41 ppm (s).
8.6 Hz, 2H), 6.56–6.73 (m, 2H), 3.79 ppm (s, 2H); 13C NMR (75 MHz,
1-(4-Azidophenyl)-2,2,2-trifluoroethanone (3): To a solution of 2
(9.2 g, 47.91 mmol) in dry DMF (80 mL) was added NaN3 (15.5 g,
239.5 mmol), and the reaction was stirred at room temperature for
12 h. The reaction mixture was quenched with water and the re-
sulting mixture was extracted with EtOAc twice. The combined or-
ganic layers was washed with brine, dried over Na2SO4, filtered and
CDCl3): d=147.7, 127.9, 124.1, 120.5, 118.2, 114.8 ppm.
1-(2-(1,3-Dioxolan-2-yl)ethyl)-3-(4-(3-trifluoromethyl)-3H-diazir-
in-3-yl)phenyl)urea (7b): To a solution of triphosgene (100 mg,
0.34 mmol) in 5 mL of dry DCM was slowly added a mixture of
amine 6b (1 mmol, 201 mg) and TEA (1.1 eq) in 2 mL DCM over
Chem. Eur. J. 2016, 22, 7824 – 7836
www.chemeurj.org
7832
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
a period of 10 min at room temperature under nitrogen atmos-
phere. After a further 30 min of stirring, a solution of amine 11
(1 mmol) and TEA (2.2 mmol) in 2 mL dry DCM was added. The re-
action mixture was stirred for 1 h at room temperature, then
evaporated to dryness, before being purified by column chroma-
tography (hexane/EtOAc=2:1) to afford compound 7b (173 mg,
((3aR,4R,6R,6aR)-6-(6-Amino-9H-purin-9-yl)-2,2-dimethyltetra-
hydrofuro[3,4-d][1,3]dioxol-4-yl)methyl methanesulfonate (14):
To a solution of 13 (200 mg, 0.651 mmol) in dry DCM (4 mL) and
TEA (0.23 mL, 1.63 mmol) at 08C was added MsCl (0.06 mL,
0.716 mmol) slowly. The ice bath was removed and the reaction
was left to stir at room temperature for 1 h, before being
quenched with H2O. The resulting mixture was extracted with DCM
twice and the combined organic layers were washed with brine,
dried over Na2SO4, filtered and concentrated in vacuo. The residue
was purified by silica gel column chromatography (EtOAc) to
afford 14 (181 mg, 72%) as a white powder. 1H NMR (300 MHz,
DMSO): d=9.44 (d, J=30.6 Hz, 2H), 8.67 (s, 1H), 8.51 (s, 1H), 6.70
(s, 1H), 5.01 (d, J=10.0 Hz, 3H), 4.40–4.64 (m, 2H), 2.31 (s, 3H),
1.48 (s, 3H), 1.23 ppm (s, 3H); 13C NMR (75 MHz, DMSO): d=157.10
(s), 149.64 (s), 140.48 (s), 139.63 (s), 120.00 (s), 112.83 (s), 91.22 (s),
85.64 (s), 84.55 (s), 80.26 (s), 57.80 (s), 26.18 (s), 24.74 ppm (s).
1
50%). H NMR (500 MHz, CDCl3): d=7.33 (d, J=8.7 Hz, 2H), 7.08 (d,
J=8.5 Hz, 2H), 4.92 (t, J=4.3 Hz, 1H), 3.89–4.00 (m, 2H), 3.74–3.89
(m, 2H), 3.30–3.49 (m, 2H), 1.81–1.97 ppm (m, 2H); 13C NMR
(125 MHz, CDCl3): d=155.5, 140.3, 127.3, 123.1, 119.6, 103.3, 64.8,
35.4, 33.1 ppm; MS (ESI): m/z calcd for C14H15F3N4O3: 345.12 [M+
H]+, found: 345.12.
(Z)-2-((4-(3-(Trifluoromethyl)-3H-diazirin-3-yl)phenyl)imino)-1,3-
oxazinan-6-ol (8b): To a solution of 7b (100 mg, 0.29 mmol) in THF
(2 mL) was added 1m HCl (1 mL). The reaction mixture was left to
stir at room temperature for 12 h. The reaction was quenched by
saturated NaHCO3 (aq.), until the solution was neutralized. Subse-
quently, the solution was extracted with EtOAc twice. The com-
bined organic extracts were dried (Na2SO4) and concentrated. Fur-
ther purification by flash column chromatography (hexane/
EtOAc=1:2) afforded 52 mg (60%) of the desired product 8b.
1H NMR (300 MHz, MeOD): d=7.32–7.42 (m, 2H), 7.19 (d, J=8.3 Hz,
2H), 5.19 (t, J=2.7 Hz, 1H), 3.56 (td, J=12.5, 4.2 Hz, 1H), 3.10–3.32
(m, 2H), 1.81–2.12 ppm (m, 2H); 13C NMR (75 MHz, MeOD): d=
157.1, 145.5, 129.6, 128.4, 127.9, 125.7, 81.4, 36.3, 30.6 ppm; MS
(ESI): m/z calcd for C12H11F3N4O2: 301.09 [M+H]+; found: 301.09.
9-((3aR,4R,6R,6aR)-6-((Isopropylamino)methyl)-2,2-dimethyltet-
rahydrofuro[3,4-d][1,3]dioxol-4-yl)-9H-purin-6-amine (15c): Iso-
propylamine (5 mL) and 4-dimethylaminopyridine (DMAP) (2 mg,
0.016 mmol) was added to 14 (50 mg, 0.13 mmol) and left to stir at
608C for 22 h. The reaction mixture was quenched with H2O and
the resulting mixture was extracted with DCM twice. The com-
bined organic layers were washed with brine, dried over Na2SO4,
filtered and concentrated in vacuo. The residue was purified by
silica gel column chromatography (CHCl3/MeOH=97:3) to obtain
1
15c (19 mg, 42%) as a white powder. H NMR (300 MHz, CDCl3): d=
8.25 (d, J=2.8 Hz, 1H), 7.89 (d, J=1.0 Hz, 1H), 6.56 (s, 2H), 5.97 (d,
J=2.8 Hz, 1H), 5.42 (dd, J=4.0, 2.4 Hz, 1H), 4.97 (dd, J=6.2,
3.4 Hz, 1H), 4.24–4.38 (m, 1H), 2.82 (s, 2H), 2.68 (d, J=2.1 Hz, 1H),
2.09 (s, 1H), 1.54 (s, 3H), 1.31 (s, 3H), 0.95 ppm (s, 6H); 13C NMR
(75 MHz, CDCl3): d=155.87 (s), 152.92 (s), 149.14 (s), 139.70 (s),
120.03 (s), 114.39 (s), 90.61 (s), 85.64 (s), 83.38 (s), 82.21 (s), 48.71
(d, J=13.5 Hz), 27.12 (s), 25.32 (s), 22.55 ppm (d, J=13.7 Hz).
2-(1,3-Dioxolan-2-yl)ethanamine (11): Compound 11 was synthe-
sized by following a previously published procedure.[41]
1-(2-(1,3-Dioxolan-2-yl)ethyl-3-(4-(tert-butyl)phenyl) urea (7a):
Synthesis of 7a was performed according to a previously published
procedure.[1] 1H NMR (300 MHz, CDCl3): d=7.82 (s, 1H), 7.20–7.32
(m, 4H), 6.15 (s, 1H), 4.92 (t, J=4.5 Hz, 1H), 3.90 (dt, J=13.0,
8.4 Hz, 2H), 3.81 (dt, J=9.0, 8.4 Hz, 2H), 3.40 (dd, J=12.2, 6.3 Hz,
2H), 1.90 (dd, J=11.1, 6.5 Hz, 2H), 1.31 ppm (s, 9H); 13C NMR
(75 MHz, CDCl3): d=156.80 (s), 145.63 (s), 136.46 (s), 125.65 (s),
120.05 (s), 103.14 (s), 64.71 (s), 35.29 (s), 34.11 (s), 33.66 (s),
31.33 ppm (s).
9-((3aR,4R,6R,6aR)-6-Azidomethyl-2,2-dimethyltetrahydro-
furo[3,4-d]-1,3-dioxol-4-yl)-9H-purin-6-ylamine (16): To 2’,3’-O-
isopropylideneadenosine (1.00 g, 3.25 mmol) suspended in 1,4-di-
oxane (15 mL) was added diphenylphosphoryl azide (DPPA, 1.7 g,
6.51 mmol) and 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU, 1.40 g,
9.77 mmol) at room temperature under N2. After the addition of
NaN3 (846 mg, 13.02 mmol), the reaction mixture was heated and
the solution was stirred for 16 h. After another 4 h, the formed
solid was removed by filtration. The solvent was evaporated and
the crude product was purified by column chromatography
(EtOAc/Et2O=1:1!MeOH/EtOAc=1:5, Rf =0.4). Upon drying in
vacuo, a colorless solid 16 (610 mg, 57%) was obtained. 1H NMR
(300 MHz, CDCl3): d=8.31 (s, 1H), 7.96 (s, 1H), 7.10 (s, 2H), 6.14 (d,
J=1.8 Hz, 1H), 5.48 (dd, J=1.8, 6.3 Hz, 1H), 5.06 (dd, J=3.3,
6.3 Hz, 1H), 4.36 (dd, J=5.4, 8.7 Hz, 1H), 3.53–4.05 (m, 2H), 1.58 (s,
3H), 1.36 ppm (s, 3H); 13C NMR (75 MHz, CDCl3): d=155.97 (s),
152.70 (s), 148.61 (s), 140.19 (s), 129.46 (s), 125.33 (s), 119.56 (s),
114.18 (s), 90.07 (s), 85.11 (s), 83.75 (s), 81.77 (s), 51.91 (s), 26.13 (s),
24.28 ppm (s); MS (IT-TOF): m/z calcd for C13H16N8O3: 333.14 [M+
H+]; found: 333.12.
1-(4-(tert-Butyl)phenyl)-3-(3-oxopropyl)urea (8a): Synthesis of 8a
was performed according to a previously published procedure.[1]
1H NMR (300 MHz, CDCl3): d=7.42–7.44 (m, 2H), 7.23–7.26 (m, 2H),
5.23 (d, J=2.7, Hz), 3.61–3.70 (m, 1H), 1.97–2.18 (m, 2H), 1.45 ppm
(s, 9H); 13C NMR (75 MHz, CDCl3): d=155.91 (s), 149.32 (s), 139.19
(s), 127.20 (s), 125.29 (s), 79.87 (s), 34.54 (s), 33.84 (s), 30.25 (s),
28.85 ppm (s); MS (IT-TOF): m/z calcd for C14H20N2O2: 249.16 [M+
H+]; found: 249.14.
((3aR,4R,6R,6aR)-6-(6-Amino-9H-purin-9-yl)-2,2-dimethyltetra-
hydrofuro[3,4-d][1,3]dioxol-4-yl)methanol (13): To a solution of
12 (500 mg, 1.87 mmol) in dry acetone at 08C was slowly added
HClO4, until 12 was completely dissolved. The reaction was stirred
at 08C for 1 h, then the ice bath was removed and the reaction
was left to stir at room temperature for 20 h. Next, the reaction
mixture was placed in an ice bath and the pH of the reaction mix-
ture was adjusted to ꢀ7 with NaHCO3. The resulting mixture was
then filtered and concentrated in vacuo. The residue was purified
by silica gel column chromatography (DCM/MeOH=96:4) to afford
13 (448 mg, 78%) as a white powder. 1H NMR (300 MHz, DMSO):
d=8.34 (s, 1H), 8.15 (s, 1H), 7.35 (s, 2H), 6.12 (d, J=3.1 Hz, 1H),
5.34 (dd, J=6.1, 3.1 Hz, 1H), 5.24 (t, J=5.5 Hz, 1H), 4.96 (dd, J=
6.2, 2.4 Hz, 1H), 4.21 (dd, J=7.2, 4.7 Hz, 1H), 3.54 (dd, J=11.1,
5.8 Hz, 2H), 2.08 (s, 1H), 1.54 (s, 3H), 1.32 ppm (s, 3H); 13C NMR
(75 MHz, DMSO): d=156.52 (s), 153.01 (s), 149.18 (s), 140.08 (s),
119.48 (s), 113.41 (s), 90.01 (s), 86.72 (s), 83.59 (s), 81.73 (s), 61.96
(s), 27.43 (s), 25.54 ppm (s).
7-((3aR,4R,6R,6aR)-6-(Aminomethyl)-2,2-dimethyltetrahydro-
furo[3,4-d][1,3]dioxol-4-yl)-7H-pyrrolo[2,3-d]pyrimidin-4-amine
(17): Compound 16 (610 mg, 1.8 mmol), palladium (300 mg, 10%
w/w on activated carbon), and EtOH (15 mL) were added to
a
100 mL round-bottomed flask. The flask was sealed with
a septum and while stirring, H2 gas was bubbled through the reac-
tion mixture from a balloon, while venting with a needle. After
5 min, the venting needle was removed and the reaction was con-
tinuously stirred under a positive pressure (1 atm) of H2 gas for
12 h. The mixture was filtered through a pad (2”2 cm) of Celite.
Chem. Eur. J. 2016, 22, 7824 – 7836
www.chemeurj.org
7833
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
The filter cake was washed with MeOH (2”25 mL) and the filtrate
further concentrated. Purification of the resulting mixture by silica
column chromatography (CHCl3/MeOH/conc. NH4OH (aq)=91:8:1,
Rf =0.3) yielded 17 (380 mg, 68%) as a white solid. 1H NMR
(300 MHz, [D6]DMSO): d=8.35 (s, 1 H), 8.14 (s, 1H), 8.08 (brs, 2H),
7.37 (s, 2H), 6.23 (d, J=2.4 Hz, 1H), 5.39 (dd, J=2.7, 6.3 Hz, 1H),
5.09–5.12 (m, 1H), 4.32–4.38 (m, 1H), 3.15–3.18 (m, 2H), 1.53 (s,
3H), 1.30 ppm (s, 3H); 13C NMR (75 MHz, CDCl3): d=156.01 (s),
152.47 (s), 148.35 (s), 140.83 (s), 129.37 (s), 119.39 (s), 114.54 (s),
90.80 (s), 83.83 (s), 83.02 (s), 81.87 (s), 41.13 (s), 25.92 (s), 23.99 ppm
(s); MS (IT-TOF): m/z calcd for C13H18N6O3: 307.15 [M+H+]; found:
307.11.
orless crude product was further purified by silica column chroma-
tography (CHCl3/MeOH/conc. NH4OH=89:10:1) to yield C1 (10 mg,
1
33%) as a white solid. H NMR (300 MHz, MeOD): d=8.28 (s, 1H),
8.21 (s, 1H), 7.18–7.36 (m, 4H), 6.00 (d, J=4.4 Hz, 1H), 4.69–4.79
(m, 1H), 4.29 (t, J=5.4 Hz, 1H), 4.21 (dt, J=9.8, 5.0 Hz, 1H), 3.51 (d,
J=2.1 Hz, 2H), 3.24 (t, J=6.6 Hz, 2H), 2.86–3.04 (m, 2H), 2.70 (t,
J=6.9 Hz, 2H), 2.59 (t, J=2.2 Hz, 1H), 1.64–1.79 (m, 2H), 1.29 ppm
(s, 9H); 13C NMR (126 MHz, MeOD): d=158.8, 157.6, 154.2, 150.8,
146.8, 141.6, 138.4, 126.8, 120.7, 90.8, 84.3, 79.4, 75.0, 75.1, 73.8,
57.2, 53.3, 43.8, 39.6, 35.3, 32.1, 28.8 ppm; HRMS: m/z calcd for
C27H36N8O4: 559.2932 [M+Na]+; found: 559.2938.
1-(3-((((2R,3S,4R,5R)-5-(6-amino-9H-purin-9-yl)-3,4-dihydroxy-
tetrahydrofuran-2-yl)methyl)(isopropyl)amino)propyl)-3-(4-(3-
(trifluoromethyl)-3H-diazirin-3-yl)phenyl)urea (C2): An oven-
dried 10 mL round-bottomed flask was charged with amine 15c
(20 mg, 57.5 mmol), oxazinanol 8b (18 mg, 60.0 mmol), 4 Š molecular
sieves (ca. 150 mg; finely powdered with a mortar and pestle) and
EtOH (1.0 mL). After stirring the mixture at room temperature for
5 min, NaBH3CN (0.24 mL, 0.50m solution in EtOH, 0.12 mmol) and
HOAc (0.13 mL, 1.0m solution in EtOH, 0.13 mmol) were added se-
quentially through a syringe. The reaction mixture was heated at
458C for 12 h. On cooling to room temperature, the mixture was
filtered through a pad of Celite and the filter cake washed with
CHCl3. The filtrate was diluted with half-saturated NaHCO3 and ex-
tracted with CHCl3. The organic extracts were pooled, dried
(Na2SO4) and concentrated. This crude product was added to a solu-
tion of 9:1 TFA/H2O (1 mL), stirred for 2 h, then concentrated to
dryness from toluene (3”2 mL), giving a colorless residue. Further
purification by silica column chromatography (CHCl3/MeOH/conc.
NH4OH=89:10:1) yielded C2 (16 mg, 47%) as a yellow solid.
1H NMR (500 MHz, MeOD): d=8.27 (s, 1H), 8.19 (s, 1H), 7.43 (d, J=
8.7 Hz, 2H), 7.09 (d, J=8.5 Hz, 2H), 5.99 (d, J=4.5 Hz, 1H), 4.75 (t,
J=4.9 Hz, 1H), 4.30 (t, J=5.3 Hz, 1H), 4.17 (dd, J=11.8, 5.1 Hz,
1H), 3.17–3.32 (m, 2H), 3.06 (dt, J=12.7, 6.3 Hz, 1H), 2.92 (dd, J=
13.9, 4.5 Hz, 1H), 2.75 (dd, J=13.9, 7.2 Hz, 1H), 2.60 (t, J=6.6 Hz,
2H), 1.64–1.78 (m, 2H), 1.06 (d, J=6.5 Hz, 3H), 1.01 ppm (d, J=
6.5 Hz, 3H); 13C NMR (75 MHz, MeOD): d=157.9, 157.6, 154.1,
150.8, 143.3, 141.6, 128.4, 125.9, 122.9, 122.2, 120.9, 119.9, 90.7,
85.0, 75.0, 73.8, 53.9, 52.2, 49.6, 39.8, 29.4, 18.7, 18.0 ppm; HRMS:
m/z calcd for C25H31F3N10O4: 615.2555 [M+Na]+; found: 615.2551.
9-((3aR,4R,6R,6aR)-2,2-Dimethyl-6-((prop-2-yn-1-ylamino)meth-
yl)tetrahydrofuro[3,4-d][1,3]dioxol-4-yl)-9H-purin-6-amine (15d):
Compound 15d was synthesized according to the published proce-
dure.[42]
1-(3-((((2R,3S,4R,5R)-5-(6-Amino-9H-purin-9-yl)-3,4-dihydroxy-
tetrahydrofuran-2-yl)methyl)(prop-2-yn-1-yl)amino)propyl)-3-(4-
(3-(trifluoromethyl)-3H-diazirin-3-yl)phenyl)urea (P1): An oven-
dried 10 mL round-bottomed flask was charged with amine 15d
(12 mg, 34.7 mmol), oxazinanol 8b (11 mg, 36.6 mmol), 4 Š molecular
sieves (ca., 80 mg; finely powdered with a mortar and pestle) and
EtOH (1.0 mL). After the mixture was stirred at room temperature
for 5 min, NaBH3CN (0.15 mL, 0.50m solution in EtOH, 0.075 mmol)
and HOAc (0.08 mL, 1.0m solution in EtOH, 0.08 mmol) were
added sequentially through a syringe. The reaction mixture was
heated at 458C for 12 h. Upon cooling to room temperature, the
mixture was filtered through a pad of Celite and the filter cake was
washed with CHCl3. The filtrate was diluted with half-saturated
NaHCO3 and extracted with CHCl3. The organic extracts were
pooled, dried with Na2SO4 and concentrated. This crude product
was added to a solution of 9:1 TFA/H2O (1 mL) and stirred further
for 2 h, then concentrated to dryness from toluene (3”2 mL),
giving a colorless residue. Further purification by silica column
chromatography (CHCl3/MeOH/conc. NH4OH=89:10:1) yielded P1
1
(8.1 mg, 40%) as a yellow solid. H NMR (300 MHz, MeOD): d=8.26
(s, 1H), 8.19 (s, 1H), 7.37–7.48 (m, 2H), 7.09 (d, J=8.5 Hz, 2H), 5.99
(d, J=4.4 Hz, 1H), 4.68–4.75 (m, 1H), 4.28 (t, J=5.3 Hz, 1H), 4.17–
4.24 (m, 1H), 3.52 (d, J=2.3 Hz, 2H), 3.24 (t, J=6.6 Hz, 2H), 2.94
(dd, J=5.6, 3.2 Hz, 2H), 2.70 (t, J=6.9 Hz, 2H), 2.59 (t, J=2.3 Hz,
1H), 1.70 ppm (p, J=6.7 Hz, 2H); 13C NMR (75 MHz, MeOD): d=
157.99 (s), 157.59 (s), 154.16 (s), 150.79 (s), 143.35 (s), 141.58 (s),
130.16 (s), 128.50 (s), 125.84 (s), 122.97 (s), 120.91 (s), 119.99 (s),
90.84 (s), 84.25 (s), 79.37 (s), 75.13 (d, J=3.0 Hz), 73.84 (s), 57.19 (s),
53.39 (s), 43.80 (s), 39.65 (s), 28.61 ppm (s); HRMS: m/z calcd for
C25H27F3N10O4: 611.2242 [M+Na]+; found: 611.2240.
1-(3-((((2R,3S,4R,5R)-5-(6-amino-9H-purin-9-yl)-3,4-dihydroxy-
tetrahydrofuran-2-yl)methyl)(isopropyl)amino)propyl)-3-(4-(tert-
butyl) phenyl)urea (FED1): An oven-dried 10 mL round-bottomed
flask was charged with amine 15c (25 mg, 73.5 mmol), oxazinanol 8a
(20 mg, 80.0 mmol), 4 Š molecular sieves (ca. 150 mg; finely pow-
dered with a mortar and pestle) and EtOH (1.0 mL). After stirring
the mixture at room temperature for 5 min, NaBH3CN (0.30 mL,
0.50m solution in EtOH, 0.15 mmol) and HOAc (0.16 mL, 1.0m solu-
tion in EtOH, 0.16 mmol) were added sequentially through a sy-
ringe. The reaction mixture was heated at 458C for 24 h. On cool-
ing to room temperature the mixture was filtered through a pad
of Celite and the filter cake washed with CHCl3. The filtrate was di-
luted with half saturated NaHCO3 and extracted with CHCl3. The or-
ganic extracts were pooled, dried (Na2SO4) and concentrated. This
crude product was added to a solution of 9:1 TFA/H2O (1 mL),
stirred for 2 h, then concentrated to dryness from toluene (3”
2 mL) to leave a colorless residue. Purification by silica column
chromatography (CHCl3/MeOH/conc. NH4OH=89:10:1) yielded
1-(3-((((2R,3S,4R,5R)-5-(6-amino-9H-purin-9-yl)-3,4-dihydroxy-
tetrahydrofuran-2-yl)methyl)(prop-2-yn-1-yl)amino)propyl)-3-(4-
(tert-butyl)phenyl)urea (C1): An oven-dried 10 mL round-bot-
tomed flask was charged with amine 15d (20 mg, 57.5 mmol), oxazi-
nanol 8a (15 mg, 60.0 mmol), 4 Š molecular sieves (ca., 120 mg;
finely powdered with a mortar and pestle) and EtOH (1.0 mL). After
stirring the mixture at room temperature for 5 min, NaBH3CN
(0.24 mL, 0.50m solution in EtOH, 0.12 mmol) and HOAc (0.13 mL,
1.0m solution in EtOH, 0.13 mmol) were added sequentially
through a syringe. The reaction mixture was heated at 458C for
12 h. On cooling to room temperature, the mixture was filtered
through a pad of Celite and the filter cake was washed with CHCl3.
The filtrate was diluted with half-saturated NaHCO3 and extracted
with CHCl3. The organic extracts were pooled, dried over Na2SO4
and concentrated. This crude product was added to a TFA/H2O co-
solvent (9:1, 1 mL) and the mixture was stirred for 2 h. Following
concentration to dryness from toluene (3”2 mL), the resulting col-
1
FED1 (16 mg, 39%) as a white solid. H NMR (500 MHz, MeOD): d=
8.25 (s, 1H), 8.18 (s, 1H), 7.22–7.30 (m, 2H), 7.16–7.22 (m, 2H), 5.98
(d, J=4.6 Hz, 1H), 4.75 (t, J=5.0 Hz, 1H), 4.54–4.72 (m, 1H), 4.31 (t,
J=5.3 Hz, 1H), 4.16 (dd, J=11.9, 5.0 Hz, 1H), 3.16–3.26 (m, 2H),
3.01–3.10 (m, 1H), 2.92 (dd, J=13.8, 4.1 Hz, 1H), 2.77 (s, 1H), 2.59
Chem. Eur. J. 2016, 22, 7824 – 7836
www.chemeurj.org
7834
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
(d, J=6.2 Hz, 2H), 1.63–1.72 (m, 2H), 1.27 (s, 9H), 1.05 (d, J=
6.6 Hz, 3H), 0.99 ppm (d, J=6.5 Hz, 3H); 13C NMR (75 MHz, MeOD):
d=158.8, 157.6, 154.1, 150.8, 146.7, 141.8, 138.4, 126.8, 120.9,
120.6, 90.8, 85.0, 74.9, 73.8, 53.9, 52.5, 49.7, 39.7, 35.3, 32.1, 29.5,
18.8, 17.9 ppm; HRMS: m/z calcd for C27H40N8O4: 540.3173 [M+
Na]+; found: 50.3178.
73.5, 70.3, 53.6, 52.1, 49.4, 39.4, 35.0, 33.7, 33.4, 31.8, 29.2, 27.8,
18.4, 17.6, 13.8 ppm; HRMS: m/z calcd for C34H48N10O4: 683.3936
[M+Na]+; found: 683.3933.
Acknowledgements
9-((3aR,4R,6R,6aR)-6-(Aminomethyl)-2,2-dimethyltetrahydro-
furo[3,4-d][1,3]dioxol-4-yl)-N-(2-(3-(but-3-yn-1-yl)-3H-diazirin-3-
yl)ethyl)-9H-purin-6-amine (18): Synthesis of the minimalist linker
(L7) was based on previously published protocols.[19] NaH (95%)
(50 mg, 2.0 mmol) was slowly added to compound 16 (332 mg,
1.0 mmol) in anhydrous DMF (5 mL) at room temperature under
Funding was provided by the National Medical Research Coun-
cil (CBRG/0038/2013) and Ministry of Education (MOE2013-T2-
1-048) of Singapore.
a
nitrogen atmosphere. After stirring for 2 h, L7 (248 mg,
Keywords: antitumor agents · cancer · proteins · proteomics ·
1.0 mmol) in DMF (2 mL) was slowly added to the mixture. The
mixture was stirred for 1 h at room temperature, followed by an
additional 2 h at 608C. The mixture was poured into cold water
(20 mL) and extracted with EtOAc (2”20 mL). The extracts were
washed with cold water (3”10 mL) and the solvent was removed.
PPh3 (262 mg, 1.0 mg) was next added to the residue in 5 mL THF
and 0.5 mL H2O. After the reaction was stirred for 12 h, the sol-
vents were removed and the resulting mixture was purified by
column chromatography (DCM/MeOH=1:1), giving 18 (65 mg,
15.3%). 1H NMR (300 MHz, MeOD): d=8.27 (s, 1H), 8.23 (s, 1H),
6.15 (d, J=3.1 Hz, 1H), 5.48 (dd, J=6.4, 3.1 Hz, 1H), 5.02 (dd, J=
6.4, 3.4 Hz, 1H), 4.24 (td, J=5.7, 3.5 Hz, 1H), 3.38–3.64 (m, 2H),
2.91 (d, J=5.9 Hz, 2H), 2.27 (t, J=2.7 Hz, 1H), 2.05 (td, J=7.3,
2.5 Hz, 2H), 1.83 (t, J=7.2 Hz, 2H), 1.68 (t, J=7.4 Hz, 2H), 1.61 (s,
3H), 1.39 ppm (s, 3H); 13C NMR (75 MHz, MeOD): d=156.0, 154.0,
141.4, 121.1, 115.6, 91.5, 88.4, 84.9, 83.6, 83.2, 70.3, 44.5, 36.5, 33.6,
33.4, 27.9, 27.5, 25.5, 13.8 ppm.
imaging agents
[1] W. Yu, E. J. Chory, A. K. Wernimont, W. Tempel, A. Scopton, A. Federa-
tion, J. J. Marineau, J. Qi, D. Barsyte-Lovejoy, J. Yi, R. Marcellus, R. E.
Iacob, J. R. Engen, C. Griffin, A. Aman, E. Wienholds, F. Li, J. Pineda, G.
Estiu, T. Shatseva, T. Hajian, R. Al-awar, J. E. Dick, M. Vedadi, P. J. Brown,
C. H. Arrowsmith, J. E. Bradner, M. Schapira, Nat. Commun. 2012, 3,
1288–1298.
[2] J. Min, Q. Feng, Z. Li, Y. Zhang, R.-M. Xu, Cell 2003, 112, 711 –723.
[3] Y. Okada, Q. Feng, Y. Lin, Q. Jiang, Y. Li, V. M. Coffield, L. Su, G. Xu, Y.
Zhang, Cell 2005, 121, 167–178.
[4] F. van Leeuwen, P. R. Gafken, D. E. Gottschling, Cell 2002, 109, 745–756.
[5] Q. Feng, H. Wang, H. H. Ng, H. Erdjument-Bromage, P. Tempst, K. Struhl,
Y. Zhang, Current biology 2002, 12, 1052–1058.
[6] T. Burmeister, C. Meyer, S. Schwartz, J. Hofmann, M. Molkentin, E.
Kowarz, B. Schneider, T. Raff, R. Reinhardt, N. Gçkbuget, D. Hoelzer, E.
Thiel, R. Marschalek, Blood 2009, 113, 4011–4015.
[7] C. Meyer, E. Kowarz, J. Hofmann, A. Renneville, J. Zuna, J. Trka, R.
Ben Abdelali, E. Macintyre, E. De Braekeleer, M. De Braekeleer, E. Dela-
besse, MP de Oliveira, H. Cave’, E. Clappier, J. J. M. van Dongen, B. V. Bal-
gobind, M. M. van den Heuvel-Eibrink, H. B. Beverloo, R. Panzer-Grümay-
er, A. Teigler-Schlegel, J. Harbott, E. Kjeldsen, S. Schnittger, U. Koehl, B.
Gruhn, O. Heidenreich, L. C. Chan, S. F. Yip, M. Krzywinski, C. Eckert, A.
Mçricke, M Schrappe, C. N. Alonso, B. W. Schäfer, J. Krauter, D. A. Lee, U.
zur Stadt, G. Te Kronnie, R. Sutton, S. Izraeli, L Trakhtenbrot, L. Lo Nigro,
G. Tsaur, L. Fechina, T. Szczepanski, S. Strehl, D. Ilencikova, M. Molkentin,
T. Burmeister, T. Dingermann, T. Klingebiel, R. Marschalek, Leukemia
2009, 23, 1490–1499.
[8] P. M. Ayton, M. L. Cleary, Genes Dev. 2003, 17, 2298–2307.
[9] a) W. Yu, D. Smil, F. Li, W. Tempel, O. Fedorov, K. T. Nguyen, Y. Bolshan, R.
Al-Awar, S. Knapp, C. H. Arrowsmith, M. Vedadi, P. J. Brown, M. Schapira,
Bioorg. Med. Chem. 2013, 21, 1787–1794; b) Y. Yao, P. Chen, J. Diao, G.
Cheng, L. Deng, J. L. Anglin, B. V. V. Prasad, Y. Song, J. Am. Chem. Soc.
2011, 133, 16746–16749; c) S. R. Daigle, E. J. Olhava, C. A. Therkelsen,
C. R. Majer, C. J. Sneeringer, J. Song, L. D. Johnston, M. P. Scott, J. J.
Smith, Y. Xiao, L. Jin, K. W. Kuntz, R. Chesworth, M. P. Moyer, K. M. Bernt,
J.-C. Tseng, A. L. Kung, S. A. Armstrong, R. A. Copeland, V. M. Richon,
R. M. Pollock, Cancer Cell 2011, 20, 53–65.
[10] C. R. Klaus, D. Iwanowicz, D. Johnston, C. A. Campbell, J. J. Smith, M. P.
Moyer, R. A. Copeland, E. J. P. Olhava, R. M. Pollock, S. R. Daigle, A. Rai-
mondi, J. Pharmacol. Exp. Ther. 2014, 350, 646–656.
[11] J. S. Yi, A. J. Federation, J. Qi, S. Dhe-Paganon, M. Hadler, X. Xu, R. St.
Pierre, A. C. Varca, L. Wu, J. J. Marineau, W. B. Smith, A. Souza, E. J.
Chory, S. A. Armstrong, J. E. Bradner, ACS Chem. Biol. 2015, 10, 667–674.
[12] M. J. Evans, B. F. Cravatt, Chem. Rev. 2006, 106, 3279–3301.
[13] Y. Su, J. Ge, B. Zhu, Y.-G. Zheng, Q. Zhu, S. Q. Yao, Curr. Opin. Chem. Biol.
2013, 17, 768–775.
1-(3-((((2R,3S,4R,5R)-5-(6-((2-(3-(But-3-yn-1-yl)-3H-diazirin-3-yl)-
ethyl)amino)-9H-purin-9-yl)-3,4-dihydroxytetrahydrofuran-2-yl)-
methyl)(isopropyl)amino)propyl)-3-(4-(tert-butyl)phenyl)urea
(P2): A flask was charged with amine 18 (21.3 mg, 0.05 mmol),
MeOH (0.5 mL) and acetone (75 mL, 10% v/v solution in MeOH,
0.10 mmol). After the solution had been stirred for 5 min, NaBH3CN
(0.19 mL, 0.50m solution in MeOH, 0.10 mmol) and HOAc (0.10 mL,
1.0m solution in MeOH, 0.10 mmol) were added sequentially. The
reaction was stirred for 2 h, then oxazinanol 8a (15 mg, 0.6 mmol)
was added in one portion, followed by 4 Š molecular sieves (ca.
70 mg; finely powdered with a mortar and pestle). After 5 min,
NaBH3CN (0.19 mL, 0.50m solution in MeOH, 0.10 mmol) and HOAc
(0.10 mL, 1.0m solution in MeOH, 0.10 mmol) were added sequen-
tially and the reaction mixture stirred rapidly at 558C for 24 h. The
mixture was allowed to cool to room temperature and filtered
through a pad (2”4 cm) of Celite. The filter cake was washed with
CHCl3 (2”10 mL). The filtrate was diluted with half-saturated
NaHCO3 solution (25 mL) and extracted with CHCl3 (3”25 mL). The
organic extracts were pooled, dried (Na2SO4), filtered and concen-
trated. This crude product was added to a solution of TFA/CHCl3 =
1:9 (1 mL) and stirred for 12 h, then concentrated to dryness from
toluene (3”2 mL), giving a colorless residue. Further purification
by silica column chromatography (CHCl3/MeOH/conc. NH4OH=
89:10:1) yielded P2 (12 mg, 36.3%) as a yellow solid. 1H NMR
(300 MHz, MeOD): d=8.25 (s, 1H), 8.22 (s, 1H), 7.23 (td, J=9.0,
2.4 Hz, 4H), 5.99 (d, J=4.5 Hz, 1H), 4.73–4.79 (m, 1H), 4.31 (t, J=
5.3 Hz, 1H), 4.17 (dd, J=11.9, 5.0 Hz, 1H), 3.50 (s, 2H), 3.22 (td, J=
6.5, 3.0 Hz, 2H), 3.01–3.15 (m, 1H), 2.93 (dd, J=14.0, 4.6 Hz, 1H),
2.72–2.84 (m, 1H), 2.63 (dd, J=14.9, 8.4 Hz, 2H), 2.26 (t, J=2.6 Hz,
1H), 2.05 (td, J=7.4, 2.6 Hz, 2H), 1.82 (t, J=7.2 Hz, 2H), 1.60–1.74
(m, 4H), 1.29 (s, 10H), 1.07 (d, J=6.6 Hz, 3H), 1.01 ppm (d, J=
6.5 Hz, 3H); 13C NMR (75 MHz, MeOD): d=158.5, 155.9, 153.8,
146.4, 140.9, 138.0, 132.3, 126.5, 121.1, 120.3, 90.4, 84.7, 83.6, 74.6,
[14] J. Lee, M. Bogyo, Curr. Opin. Chem. Biol. 2013, 17, 118–126.
[15] J. Park, M. Koh, S. B. Park, Mol. BioSyst. 2013, 9, 544–550.
[16] M. Gersch, J. Kreuzer, S. A. Sieber, Nat. Prod. Rep. 2012, 29, 659–682.
[17] P.-Y. Yang, K. Liu, M. H. Ngai, M. J. Lear, M. R. Wenk, S. Q. Yao, J. Am.
Chem. Soc. 2010, 132, 656–666.
[18] H. Shi, C.-J. Zhang, G. Y. J. Chen, S. Q. Yao, J. Am. Chem. Soc. 2012, 134,
3001–3014.
Chem. Eur. J. 2016, 22, 7824 – 7836
www.chemeurj.org
7835
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[19] Z. Li, P. Hao, L. Li, C. Y. J. Tan, X. Cheng, G. Y. J. Chen, S. K. Sze, H.-M.
Shen, S. Q. Yao, Angew. Chem. Int. Ed. 2013, 52, 8551–8556; Angew.
Chem. 2013, 125, 8713–8718.
[20] B. R. Lanning, L. R. Whitby, M. M. Dix, J. Douhan, A. M. Gilbert, E. C. Hett,
T. Johnson, C. Joslynl, J. C. Kath, S. Niessen, L. R. Roberts, M. E. Schnute,
C. Wang, J. J. Hulce, B. X. Wei, L. O. Whiteley, M. M. Hayward, B. F. Cra-
vatt, Nat. Chem. Biol. 2014, 10, 760–767.
[21] R. Krishnamurty, J. L. Brigham, S. E. Leonard, P. Ranjitkar, E. T. Larson, E. J.
Dale, E. A. Merritt, D. J. Maly, Nat. Chem. Biol. 2012, 9, 43–50.
[22] Y. Su, S. Pan, Z. Li, L. Li, X. Wu, P. Hao, S. K. Sze, S. Q. Yao, Sci. Rep. 2015,
5, 7724.
[31] M. Fernµndez-Suµrez, A. Y. Ting, Nat. Rev. Mol. Cell Biol. 2008, 9, 929–
943.
[32] H. Wang, E. Nakata, I. Hamachi, ChemBioChem 2009, 10, 2560–2577.
[33] D. W. Domaille, E. L. Que, C. J. Chang, Nat. Chem. Biol. 2008, 4, 168–175.
[34] A. Baruch, D. A. Jeffery, M. Bogyo, Trends Cell Biol. 2004, 14, 29–35.
[35] K. S. Yang, G. Budin, T. Reiner, C. Vinegoni, R. Weissleder, Angew. Chem.
Int. Ed. 2012, 51, 6598–6603; Angew. Chem. 2012, 124, 6702–6707.
[36] L. Li, X. Shen, Q.-H. Xu, S. Q. Yao, Angew. Chem. Int. Ed. 2013, 52, 424–
428; Angew. Chem. 2013, 125, 442–446.
[37] L. Li, C. Zhang, G. Y. J. Chen, B. Zhu, C. Chai, Q.-H. Xu, E.-K. Tan, Q. Zhu,
K.-L. Lim, S. Q. Yao, Nat. Commun. 2014, 5, 3276.
[38] L. Li, C.-W. Zhang, J. Ge, L. Qian, B.-H. Chai, Q. Zhu, J.-S. Lee, K.-L. Lim,
S. Q. Yao, Angew. Chem. Int. Ed. 2015, 54, 10821–10825; Angew. Chem.
2015, 127, 10971–10975.
[39] B. R. Lanning, L. R. Whitby, M. M. Dix, J. Douhan, A. M. Gilbert, E. C. Hett,
T. O. Johnson, C. Joslyn, J. C. Kath, S. Niessen, L. R. Roberts, M. E.
Schnute, C. Wang, J. J. Hulce, B. Wei, L. O. Whiteley, M. M. Hayward, B. F.
Cravatt„ Nat. Chem. Biol. 2008, 4, 168–175.
[40] M. J. Niphakis, K. M. Lum, A. B. Cognetta III, F. Piscitelli, H. Rosen, B. F.
Cravatt, Cell 2015, 161, 1668–1680.
[41] G. W. Gribble, F. L. Switzer, Synth. Commun. 1987, 17, 377–383.
[42] S. Pochet, G. Labesse, M. Gelin; L. Assairi, O. Dussurget, G. Poncet-Mon-
tange, International Application Published Under the Patent Cooperation
Treaty (PCT), EP2658863A1 2012, (in English).
[23] Z. Li, D. Wang, L. Li, S. Pan, Z. Na, C. Y. J. Tan, S. Q. Yao, J. Am. Chem. Soc.
2014, 136, 9990–9998.
[24] E. K. W. Tam, Z. Li, Y. L. Goh, X. Cheng, S. Y. Wong, S. Santhanakrishnan,
C. L. L. Chai, S. Q. Yao, Chem. Asian J. 2013, 8, 1818–1828.
[25] H. C. Kolb, K. B. Sharpless, Drug Discovery Today 2003, 8, 1128–1137.
[26] M. Meldal, C. W. Tornøe, Chem. Rev. 2008, 108, 2952–3015.
[27] E. M. Sletten, C. R. Bertozzi, Angew. Chem. Int. Ed. 2009, 48, 6974–6998;
Angew. Chem. 2009, 121, 7108–7133.
[28] K. A. Kalesh, H. Shi, J. Ge, S. Q. Yao, Org. Biomol. Chem. 2010, 8, 1749–
1762.
[29] a) Z. Na, B. Peng, S. Ng, S. Pan, J.-S. Lee, H. M. Shen, S. Q. Yao, Angew.
Chem. Int. Ed. 2015, 54, 2515–2519; Angew. Chem. 2015, 127, 2545–
2549; b) D. M. Molina, R. Jafari, M. Ignatushchenko, T. Seki, E. A. Larsson,
C. Dan, L. Sreekumar, Y. Cao, P. Nordlund, Science 2013, 341, 84–87.
[30] R. Y. Tsien, Angew. Chem. Int. Ed. 2009, 48, 5612–5626; Angew. Chem.
2009, 121, 5721–5736.
Received: January 19, 2016
Published online on April 26, 2016
Chem. Eur. J. 2016, 22, 7824 – 7836
www.chemeurj.org
7836
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim