Phosphono Bisbenzguanidines as Irreversible Dipeptidomimetic Inhibitors and Activity-Based Probes of Matriptase-2

Article abstract of DOI:10.1002/chem.201600206

Matriptase-2, a type II transmembrane serine protease, plays a key role in human iron homeostasis. Inhibition of matriptase-2 is considered as an attractive strategy for the treatment of iron-overload diseases, such as hemochromatosis and β-thalassemia. In the present study, synthetic routes to nine dipeptidomimetic inactivators were developed. Five active compounds (41–45) were identified and characterized kinetically as irreversible inhibitors of matriptase-2. In addition to a phosphonate warhead, these dipeptides possess two benzguanidine moieties as arginine mimetics to provide affinity for matriptase-2 by binding to the S1 and S3/S4 subpockets, respectively. This binding mode was strongly supported by covalent docking analysis. Compounds 41–45 were obtained as mixtures of two diastereomers and were therefore separated into the single epimers. Compound 45 A, with S configuration at the N-terminal amino acid and R configuration at the phosphonate carbon atom, was the most potent matriptase-2 inactivator with a rate constant of inactivation of 2790 m^?1s^?1and abolished the activity of membrane-bound matriptase-2 on the surface of intact cells. Based on the chemotyp of phosphono bisbenzguanidines, the design and synthesis of a fluorescent probe (51 A) by insertion of a coumarin label is described. The in-gel fluorescence detection of matriptase-2 was demonstrated by applying 51 A as the first activity-based probe for this enzyme.

Full text of DOI:10.1002/chem.201600206

DOI: 10.1002/chem.201600206
Full Paper
&
Enzyme Inhibition |Hot Paper|
Phosphono Bisbenzguanidines as Irreversible Dipeptidomimetic
Inhibitors and Activity-Based Probes of Matriptase-2
Daniela Häußler,[a] Martin Mangold,[a] Norbert Furtmann,[a, b] Annett Braune,[c]
Michael Blaut,[c] Jürgen Bajorath,[b] Marit Stirnberg,[a] and Michael Gütschow*[a]
Abstract: Matriptase-2, a type II transmembrane serine pro-
tease, plays a key role in human iron homeostasis. Inhibition
of matriptase-2 is considered as an attractive strategy for the
treatment of iron-overload diseases, such as hemochromato-
sis and b-thalassemia. In the present study, synthetic routes
to nine dipeptidomimetic inactivators were developed. Five
active compounds (41–45) were identified and characterized
kinetically as irreversible inhibitors of matriptase-2. In addi-
tion to a phosphonate warhead, these dipeptides possess
two benzguanidine moieties as arginine mimetics to provide
affinity for matriptase-2 by binding to the S1 and S3/S4 sub-
pockets, respectively. This binding mode was strongly sup-
ported by covalent docking analysis. Compounds 41–45
were obtained as mixtures of two diastereomers and were
therefore separated into the single epimers. Compound
45A, with S configuration at the N-terminal amino acid and
R configuration at the phosphonate carbon atom, was the
most potent matriptase-2 inactivator with a rate constant of
inactivation of 2790mÀ1 sÀ1 and abolished the activity of
membrane-bound matriptase-2 on the surface of intact cells.
Based on the chemotyp of phosphono bisbenzguanidines,
the design and synthesis of a fluorescent probe (51A) by in-
sertion of a coumarin label is described. The in-gel fluores-
cence detection of matriptase-2 was demonstrated by apply-
ing 51A as the first activity-based probe for this enzyme.
and iron-deficient phenotype in murine models.[3] Matriptase-2
is expressed mainly in hepatocytes and it negatively regulates
the production of hepcidin, the systemic iron-regulatory hor-
mone. Hepcidin regulates cellular iron export by promoting
the degradation of the iron transporter ferroportin.[4] The ex-
pression of HAMP, the gene of hepcidin, is controlled through
the bone morphogenetic protein/son of mothers against de-
capentaplegic (BMP/SMAD) pathway as follows. An activated
complex, composed of the BMP receptor, its endogenous
ligand BMP6, and the coreceptor hemojuvelin phosphorylates
the SMAD1/5/8–SMAD4 complex, which then causes increased
expression of HAMP. The presence of intact hemojuvelin is re-
quired for this signaling cascade[5] and, accordingly, causes lim-
ited iron absorption and iron release.[4] Matriptase-2 exerts its
influence as a regulator of this pathway by hydrolytically cleav-
ing hemojuvelin from the plasma membrane, which leads to
a down-regulation of HAMP expression. Putative cleavage sites
in hemojuvelin have been suggested after arginine residues.[5,6]
These cleavage sites correspond to the primary substrate spe-
cificity of matriptase-2 to hydrolyze a peptide bond after basic
amino acids, with a preference for arginine over lysine, in the
P1 position.[2] Matriptase-2 undergoes complex autoprocessing
that results in activation and release from the cell surface, as
shown in a transfected cell system.[7] These autoprocessing
sites also feature a P1 arginine residue. Attempts to develop
peptidic substrates for matriptase-2 revealed that, in addition
to the P1 arginine, further basic amino acids at positions P4–
P2 are favored,[8] which was confirmed by a combinatorial ap-
proach that identified the preferred P4–P1 substrate sequence
Introduction
The cell-surface protease matriptase-2, encoded by the
TMPRSS6 gene, is a member of the type II transmembrane
serine proteases. The family members exhibit a short intracellu-
lar N-terminal tail, a transmembrane domain, and a large extra-
cellular portion containing a variable stem region and a C-ter-
minal serine protease catalytic domain with a highly conserved
chymotrypsin fold. Matriptase-2 belongs to the trypsin-like pro-
teases, which preferentially cleave substrates with arginine or
lysine in the P1 position.[1,2]
Matriptase-2 has attracted strong attention because of its
contribution to the regulation of iron homeostasis. It has been
found that mutations in the TMPRSS6 gene cause iron-refracto-
ry iron deficiency anemia (IRIDA) in humans and an anemic
[a] D. Häußler, M. Mangold, N. Furtmann, Dr. M. Stirnberg,
Prof. Dr. M. Gütschow
Pharmaceutical Institute, Pharmaceutical Chemistry I
University of Bonn, An der Immenburg 4, 53121 Bonn (Germany)
E-mail: guetschow@uni-bonn.de
[b] N. Furtmann, Prof. Dr. J. Bajorath
Department of Life Science Informatics, B-IT
LIMES Program Unit Chemical Biology and Medicinal Chemistry
University of Bonn, Dahlmannstrasse 2, 53113 Bonn (Germany)
[c] Dr. A. Braune, Prof. Dr. M. Blaut
Department of Gastrointestinal Microbiology
German Institute of Human Nutrition Potsdam-Rehbruecke
Arthur-Scheunert-Allee 114–116, 14558 Nuthetal (Germany)
Supporting information for this article can be found under http://
dx.doi.org/10.1002/chem.201600206.
Chem. Eur. J. 2016, 22, 8525 – 8535
8525
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
as Ile–Arg–Ala–Arg.[9] So far, the characterization of the (patho)-
physiological role of matriptase-2 has been hampered by the
lack of specific antibodies and activity-based probes.
deep S1 pocket of the target enzyme.[11,29,30] Beside the ionic
interaction in the S1 binding pocket, an additional positively
charged moiety of an inhibitor should be able to interact with
the upper part of the S3/S4 binding pocket of matriptase-2,
where the backbone carbonyl oxygen atoms of Glu662,
Asp663, and Ser664 provide a negatively charged environ-
ment.[11,29,30] Among the few peptidomimetic inhibitors report-
ed for matriptase-2 so far, the presence of two basic substruc-
tures turned out to be favorable.[11,29,31] In particular, this was
the case for a series of peptidyl benzothiazole ketones, for
which the following P4–P1 sequences afforded the most
potent inhibitors of matriptase-2: Arg–Gln–Tyr–Arg, Trp–Cys–
Tyr–Arg, Arg–Gln–Phe–Arg, and Lys–Trp–Trp–Arg.[32] The pres-
ence of a second basic amino acid in three of these sequences
supported the design of dibasic inhibitors for matriptase-2 and
was in agreement with the preferred aforementioned P4–P1
substrate sequences.[7,8]
Matriptase-2 is involved in disease states related to human
iron homeostasis. Mutations in the TMPRSS6 gene lead to
IRIDA, a rare autosomal recessive disorder.[10] Matriptase-2
emerges as a potentially important pharmaceutical target for
the treatment of iron-overload diseases, based on the consid-
eration that inhibition of matriptase-2 activity increases HAMP
expression and, accordingly, attenuates the plasma iron
level.[11,12] Hereditary hemochromatosis and b-thalassemia rep-
resent prominent examples of iron-overload diseases. It has
been shown in mouse models of hereditary hemochromatosis
or b-thalassemia that iron overload can be prevented by tar-
geting TMPRSS6. The b-thalassemia mouse model mimics non-
transfusion-dependent thalassemia conditions that are associ-
ated with a suppression of HAMP. In both models, the treat-
ment of such mice with TMPRSS6 small interfering RNA dimin-
ishes iron overload by up-regulation of hepcidin expression.[13]
Similarly, an antisense oligonucleotide strategy was pursued,
which demonstrated decreased serum iron, transferrin satura-
tion, and iron accumulation in the liver of mice affected by he-
reditary hemochromatosis and an improved anemia in b-tha-
lassemia mice.[14] Therefore, selective, low-molecular-weight
matriptase-2 inhibitors might be beneficial as pharmacological
tools to investigate further the exact role of the protease in
regulating iron homeostasis and, moreover, as leads for the
therapeutic treatment of iron-overload diseases, in particular,
the iron overload in thalassemic patients receiving multiple
blood transfusions.[4,11,12]
Protein profiling with activity-based probes has emerged as
a powerful strategy to study enzyme functions and activities in
complex proteomes.[20,21,33–35] The visualization of the target is
restricted to the active form of the enzyme, which is of particu-
lar importance, because enzyme expression levels do not
always correlate with the activity or the activity might be regu-
lated post-translationally.[34,36] These probes consist of three es-
sential components: a reactive group, a recognition element,
and a reporter tag. The reactive group, a so-called warhead,
leads to a covalent attachment to an active-site nucleophile.
Electrophilic moieties applied for serine protease probing com-
prise fluorophosphonates, diphenylphosphonates, sulfonyl flu-
orides, and isocoumarins.[35]
For this study, we considered peptidomimetic phosphonates
as potential inhibitors of matriptase-2. The irreversible mode of
action is based on nucleophilic attack of the active-site serine
and the simultaneous release of one phenolic residue. The re-
sulting serine phosphono diester might undergo a slow aging
upon loss of the second phenoxy group to form a phosphono
monoester. An advantage of phosphonates compared with
other classes of inhibitors is a complete lack of activity against
cysteine or threonine proteases.[15] Covalent inhibitors might
have several potential benefits, such as 1) improved biochemi-
cal efficiency of target disruption, 2) prevention of drug resist-
ance, and 3) more stable pharmacokinetics.[16] However, irrever-
sible inhibitors could also modify off-targets and may lead to
a risk of organ toxicity because such electrophiles are able to
react with circulating nucleophiles, for example, glutathione.[17]
To attain sufficient potency and the desired selectivity of
phosphonate inhibitors for a trypsin-like serine protease, their
structures might advantageously be equipped with residues of
basic amino acids at the P1 position or with other positively
charged groups, such as amidines or guanidines.[18–26] Beside
their utilization as arginine mimetics, benzamidine and benz-
guanidine moieties have been successfully introduced into an-
tiprotozoal and antifungal compounds.[27] In certain cases, their
fluorescent properties have been shown to depend on the in-
sertion into the minor groove of double-strand DNA.[28] For our
design of matriptase-2 inhibitors, the introduction of an argi-
nine mimetic was intended to enable an interaction with the
Results and Discussion
We aimed at synthesizing dipeptidomimetic inhibitors for ma-
triptase-2 with the following structural features. An amino-
phosphonate moiety in place of the C-terminal amino acid was
supposed to act as a warhead and enable irreversible inhibi-
tion of the target protease. In addition, two arginine-mimetic
substructures would be introduced for an attractive interaction
with the S1 and S3/S4 subsites: one guanidinophenyl group as
an a-substituent at the aminophosphonate and an N-terminal
guanidinobenzoyl capping group (Figure 1). Our synthetic
strategy involved the construction of the entire phosphono di-
peptide, followed by the generation of the guanidine moieties
from the nitro groups of the precursors. Five points of diversity
were planned to be introduced, that is, the position of the
guanidine group at both benzguanidines, the nature (and ste-
Figure 1. Schematic composition of the phosphono bisbenzguanidines de-
scribed herein.
Chem. Eur. J. 2016, 22, 8525 – 8535
www.chemeurj.org
8526
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
reochemistry) of the N-terminal amino acid, and the O sub-
stituents at the phosphonate moiety.
bamate (4), and an aldehyde. Either 3-nitrobenzaldehyde (5) or
4-nitrobenzaldehyde (6) was used to introduce structure diver-
sity. Whereas triphenylphosphite (1) was commercially avail-
able, tris(4-(methylthio)phenyl)phosphite (3) was synthesized
by heating PCl3 with 4-(methylthio)phenol (2) in acetonitrile at
reflux.[19,38,39] The Cbz-protected aminophosphonates 8–10
were obtained as racemic mixtures in good yields. Deprotec-
Scheme 1 outlines the eight-step synthetic approach toward
the bisguanidine phosphonates 39–45. The linear synthesis
started with the Oleksyszyn reaction, a three-component reac-
tion to obtain the a-aminophosphonate building block.[18,37]
This transformation involved a phosphite (1 or 3), benzyl car-
Scheme 1. Synthesis of compounds 39–45. Reaction conditions: a) PCl3, MeCN, reflux; b) AcOH, 908C; c) AcOH/HBr, RT; d) HATU, DIPEA, DMF, RT; e) TFA,
CH2Cl2, RT; f) HATU, DIPEA, DMF, RT; g) SnCl2·2H2O, H2O, EtOAc, reflux; h) HgCl2, TEA, CH2Cl2, RT; i) TFA, CH2Cl2, RT, prep. HPLC, HCl, lyophilization. Cbz: benzyl-
oxycarbonyl; HATU: N-[(dimethylamino)-1H-1,2,3-triazole[4,5-b]-pyridin-1-ylmethylene]-N-methylmethanaminium hexafluorophosphate; DIPEA: diisopropyl-
ethylamine; TFA: trifluoroacetic acid; TEA: triethylamine.
Chem. Eur. J. 2016, 22, 8525 – 8535
www.chemeurj.org
8527
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
tion of the amino group was achieved after treatment with
a solution of HBr in acetic acid.[20,21,33,37,38,40] The bromide salts
13 were directly subjected to a coupling reaction with the
amino acids Boc-l-Phe-OH (14; Boc: tert-butoxycarbonyl), Boc-
d-Phe-OH (15), and Boc-l-Lys(Cbz)-OH (16), respectively, by ap-
plying HATU as coupling reagent and DIPEA as base.[21,33,41]
Other guanidinium/uronium coupling reagents or carbodii-
mides have also been employed for the preparation of several
phosphono dipeptides.[18,20,22,24,33,38,42] The resulting Boc-pro-
tected peptides 17–21 occurred as a mixture of two diastereo-
mers after the introduction of a second chiral center, which ori-
ginated from the introduced amino acid.
47 (see Table 1 for its structure), a bisiodine phosphonate ana-
logue of 42, was synthesized (Scheme S2 in the Supporting In-
formation). The general synthetic approach to the protected a-
aminophosphonates was performed with 4-iodobenzaldehyde
to place the iodine at the P1 residue. After Cbz deprotection,
coupling to Boc-l-Phe-OH (14), and Boc deprotection, the
second iodine functionality was introduced by using 4-iodo-
benzoic acid.
The phosphono dipeptides 39–47 were evaluated at matrip-
tase-2 and three related serine proteases, that is, human
thrombin, bovine factor Xa, and bovine trypsin; selected com-
pounds were also evaluated at human trypsin. These enzymes
were assayed with fluorogenic peptide substrates and the re-
actions were followed over 30 min. The active compounds
were measured at five different inhibitor concentrations and
the progress curves were analyzed. In most cases, the active
compounds exhibited a time-dependent inhibition; non-linear
regression of the progress curves provided pseudo-first-order
rate constants of irreversible inhibition, kobs. These kobs values
were then plotted versus the inhibitor concentrations and the
parameters for the inactivation (kinac/Ki) were calculated from
the second-order rate constants, kobs/[I], obtained by linear re-
gression,. The results of the kinetic experiments are summar-
ized in Table 1.
In order to remove the Boc protecting group, compounds
17–21 were treated with trifluoroacetic acid to obtain the am-
monium salts 24. To insert a further point of diversity, the free
amine was coupled with either 3-nitrobenzoic acid (25) or 4-ni-
trobenzoic acid (26). For this purpose, the protocol with HATU
and DIPEA was reapplied to yield the bisnitro compounds 28–
34. For the subsequent reduction of the nitro groups, SnCl2
acted as the reducing agent.[19,26] In accordance with a previ-
ously described method,[19,20,26] the so-obtained aniline deriva-
tives 36 were converted into protected guanidines in a HgCl2-
promoted reaction with N,N’-di-Boc-S-methylisothiourea (37)
and Et3N. Somewhat different approaches have been reported
for phosphonates with Boc-protected benzguanidine moieties
performing the aniline guanylation with N,N’-di-Boc-1-guanyl-
pyrazole,[23,24] or N,N’-di-Boc-2-(trifluoromethylsulfonyl)guani-
dine.[21] The desired guanidine phosphonates 39–45 were ob-
tained after removing the Boc protecting groups with trifluoro-
acetic acid in dichloromethane.[19] Afterwards, the products
were purified by means of preparative HPLC and lyophilization
to give the final hydrochloride salts 39–45.
In addition to the phosphonate diphenyl esters 39–44 and
the bismethylthiophenyl ester 45, a compound was designed
in which the aryloxy moieties were exchanged for ethoxy
groups. Its structure, 46, is shown in Table 1. The preparation
of 46 is depicted in Scheme S1 in the Supporting Information.
The transesterification of diphenyl phosphonates constitutes
one efficient route to dialkyl phosphonates.[43] Thus, with the
diphenyl phosphonate 9 (structure in Scheme 1) as the starting
material, a Cbz-protected a-amino diethyl phosphonate was
prepared in ethanol in the presence of KF/18-crown-6
ether.[33,43] By following a route analogous to that described
above, via the corresponding mononitro and bisnitro inter-
mediates, the guanidine phosphonate diethyl ester 46 was
achieved. For this exemplary synthesis, the substitution pattern
of 42 was chosen by taking the bioactivity data of the phen-
oxy derivatives into account (see below). With consideration of
previous results on the inhibition of human thrombin and
bovine trypsin by peptide phosphonate derivatives,[18] a diethyl
substitution at the phosphorus was not expected to cause an
improved potency. Nevertheless, it was intended to verify this
at matriptase-2.
Initially, the influence of the positions of the guanidine moi-
eties on the inhibitory activity toward matriptase-2 was investi-
gated. We compared compounds 39, 40, 41, and 42, which
possess a different guanidino substitution pattern but share
the remaining structural features. Whereas a meta substitution
at the “Eastern” part (position 3) led to inactive compounds
(39, 40), a guanidine group in the para position (position 4)
was advantageous (41, 42) and, in combination with a para
substitution at the “Western” part (position 4’), gave a kinac/Ki
value of 120mÀ1 sÀ1 for the inactivation of matriptase-2 by
compound 42. The guanidino substitution pattern clearly af-
fected the selectivity of the phosphono bisbenzguanindines
39–42. A meta substitution at the “Eastern” part (position 3)
caused inhibition of thrombin, whereas matriptase-2, factor Xa,
and bovine trypsin were not affected by 39 and 40. Unexpect-
edly, the inhibition of thrombin was not time dependent. Rates
in the presence of different inhibitor concentrations were de-
termined from linear regression of the progress curves to
obtain IC50 values, which are unsuitable for a comparison with
the second-order rate constants of inactivation. The influence
of the substrate concentration on thrombin inhibition by 39
was analyzed and revealed a noncompetitive mode of inhibi-
tion (see Figure S1 in the Supporting Information). Further
studies might clarify the inhibition of thrombin by 39 and 40
and address the question of to what extent the phosphonate
moiety might be involved in the enzyme–inhibitor interaction.
As para-guanidine groups at the “Eastern” (position 4) and
“Western” part (position 4’), as present in 42, were beneficial
for matriptase-2 inhibition, this substitution pattern was kept
in the other phosphono dipeptides of this study (43–47).
These compounds did not inhibit either thrombin or factor Xa,
but the inactivation of matriptase-2 (by 42–45) was generally
accompanied by inactivation of bovine trypsin. We also deter-
It has been reported that positively charged groups (e.g.,
guanidines, amidines) of protein ligands can be exchanged for
halogen atoms that are able to perform protein–ligand interac-
tions in so-called halogen bonding.[44] Accordingly, compound
Chem. Eur. J. 2016, 22, 8525 – 8535
www.chemeurj.org
8528
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Table 1. Inhibition of serine proteases by phosphono dipeptides 39–47.
[b,c]
Compound[a]
Positions of the guanidine residues R1
Enzyme inactivation, kinac/Ki [mÀ1 sÀ1
Human matriptase-2 Human thrombin
]
Comp.
R2
Bovine factor Xa Bovine trypsin
39 (S, rac) 3’
40 (S, rac) 4’
3
3
H
H
Ph
Ph
n.i.
n.i.
IC50 =(1.66Æ0.12) mm[d] n.i.
n.i.
n.i.
IC50 =(2.72Æ0.20) mm[d] n.i.
41 (S, rac)
41A (S, R) 3’
41B (S, S)
59.1Æ4.7
197Æ4
n.i.
n.i.
n.i.
n.i.
n.i.
n.i.
n.i.
1360Æ90
1740Æ200
1290Æ50
4
4
4
4
4
H
H
H
H
Ph
Ph
Ph
42 (S, rac)
42A (S, R) 4’
42B (S, S)
120Æ10
283Æ14
n.i.
n.i.
n.i.
n.i.
n.i.
n.i.
n.i.
568Æ36
1570Æ150[e]
143Æ18[e]
43 (R, rac)
43A (R, S) 4’
43B (R, R)
58.7Æ4.6
n.i.
154Æ20
n.i.
n.i.
n.i.
n.i.
n.i.
n.i.
359Æ26
166Æ7
1550Æ70
44 (S, rac)
44A (S, R) 4’
44B (S, S)
65.0Æ2.3
n.i.
n.i.
n.i.
n.i.
n.i.
n.i.
872Æ91
1450Æ80
511Æ65
(CH2)3NHÀCbz 73.6Æ5.3
n.i.
45 (S, rac)
45A (S, R) 4’
45B (S, S)
1750Æ40
n.i.
n.i.
n.i.
n.i.
n.i.
n.i.
5900Æ80
SMe Ph
2790Æ60
1280Æ50
5370Æ360[e]
5430Æ190[e]
46 (S, rac)
47 (S, rac)
n.i.
n.i.
n.i.
n.i.
n.i.
n.i.
n.i.
n.i.
[a] The stereochemistry is given in parentheses. The first entry refers to the configuration at the amino acid, the second to the configuration at the amino-
phosphonate. [b] IC50 and kinac/Ki values were determined from duplicate measurements with five different inhibitor concentrations. [c] n.i.: no inhibition re-
lates to more than 30% product formation after 30 min at an inhibitor concentration of 30 mm, which corresponds to an IC50 value of >13 mm. Determined
from duplicate measurements. [d] No time-dependent inhibition was observed. Rates in the presence of five different inhibitor concentrations, obtained by
linear regression of the progress curves, were plotted against inhibitor concentrations to obtain IC50 values. [e] The following kinac/Ki values were obtained
for inactivation of human trypsin: (217Æ36)mÀ1 sÀ1 for (S,R)-42A, (65Æ13)mÀ1 sÀ1 for (S,S)-42B, (1480Æ214)mÀ1 sÀ1 for (S,R)-45A, and (1780Æ193)mÀ1 sÀ1
for (S,S)-45B.
mined the inactivation of human trypsin by selected com-
pounds (see below). Lower kinac/Ki values were obtained for the
human enzyme. Nevertheless, inhibitors 42–45, and 41, were
not selective for matriptase-2 over trypsin.
At this stage of the chemical optimization, methylthio sub-
stituents were introduced into the phenoxy groups of 42,
which resulted in derivative 45. In accordance with previous
reports on the inactivation of other serine proteases by a-ami-
nophosphonates,[19,26,38] this modification was also favorable
for matriptase-2 inhibition and led to a kinac/Ki value of
1750mÀ1 sÀ1. In contrast, the replacement of the unsubstituted
phenoxy groups in 42 by two ethoxy residues completely ab-
rogated the inhibitory effect of the corresponding compound
46 toward matriptase-2 (Table 1). It is likely that the leaving
group abilities of the corresponding P substituents (4-(methyl-
thio)phenoxy>phenoxy>ethoxy) contributed to these differ-
ences.
Next, we examined the effect of the configuration of the
N-terminal amino acid on the inactivator potency toward
matriptase-2. Compounds 42 and 43 share the guanidino sub-
stitution pattern (para, para) and the unsubstituted phenoxy
groups (R1: H), but they differ in the configuration of the
phenylalanine moiety. It turned out that the natural S con-
figuration caused stronger inhibition of matriptase-2. In the
phosphono dipeptide 44, l-phenylalanine was exchanged
for Cbz-protected l-lysine, which reduced the kinac/Ki value
(44 relative to 42). Thus, l-phenylalanine was maintained as
the N-terminal amino acid in the phosphono dipeptides of this
study.
Recently, halogen-containing compounds have been recog-
nized to form noncovalent interactions with Lewis bases. Halo-
gen bonding in protein–ligand interactions is driven by the
Chem. Eur. J. 2016, 22, 8525 – 8535
www.chemeurj.org
8529
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
anisotropy of electron density on the halogen, which leads to
a positively charged region and increases in the order Cl<Br<
I. Accordingly, by targeting a negatively charged aspartate resi-
due, strong halogen bonding might occur. For non-phospho-
nate inhibitors of factor Xa, an amidine group could be suc-
cessfully substituted by iodine.[44] The bisiodine derivative 47
was included in this study in order to examine whether the
same strategy would be applicable to our compounds. Howev-
er, 47 was inactive (Table 1), which indicated a limited ability
for competing with the substrate to interact with the S1 and
S3/S4 binding pockets.
between the active site and the covalently bound inhibitor.
Therefore, it is very likely that this epimer of 42 represented
the inactive diastereomer and we assigned the (S,S) configura-
tion to 42B. Accordingly, 42A possessed the (S,R) configura-
tion. The modeled complex that resulted from the interaction
of 42A with matriptase-2 and the subsequent hydrolytic
“aging” is depicted in Figure 2.
In order to elucidate the effect of the configuration at the
aminophosphonate carbon atom on the biological activity, the
single diastereomers were separated by preparative HPLC and
are listed in Table 1. The corresponding code (A or B) refers to
the pattern of NMR resonances as follows. Generally, the A dia-
stereomer showed a 1H NMR signal for the methine POCH
proton at higher field and had a larger heteronuclear H–P cou-
pling constant than the corresponding B diastereomer. For ex-
ample, the following signals were obtained for 41A and 41B:
3
2
41A: d=5.93 ppm (dd, J (H, H)=9.8 Hz, J (H, P)=22.4 Hz);
41B: d=6.00 ppm (dd, 3J (H, H)=9.8 Hz, 2J (H, P)=21.8 Hz).
The 13C NMR signal for the aminophosphonate POCH carbon
atom was shifted upfield in the A diastereomer relative to the
corresponding
B diastereomer; for example, 41A: d=
Figure 2. Covalent docking of compound 42A (or 45A) bound to Ser762 in
the binding site of a matriptase-2 homology model. Note that one phenoxy
group was removed because of the nucleophilic attack of the active-site
serine and formation of the SerÀOÀP bond, whereas the other phenoxy
group was removed through hydrolysis.
1
1
49.65 ppm (d, J (C, P)=156.4 Hz); 41B: d=50.03 ppm (d, J
(C, P)=156.4 Hz).
The five pairs of separated epimers clearly differed in their
matriptase-2 inhibitory activity. In the cases of 41, 42, 44, and
45, the A diastereomer was the more active one, whereas the
corresponding B diastereomer showed a weaker activity; for
example, kinac/Ki =2790mÀ1 sÀ1 for 45A compared with
1280mÀ1 sÀ1 for 45B. The assignment of the configuration at
the phosphonate carbon atom of the active epimer was tenta-
tively achieved by molecular docking. The crystal structures of
complexes of serine proteases with phosphonate inactivators
revealed that the initially formed serine phosphono diester
was converted by hydrolytic “aging” into a phosphono mono-
ester.[45] The phosphono monoester complexes of matriptase-2
with the two diastereomers of 42 were generated by covalent
docking as follows. The single bond between the active-site
serine oxygen atom and the phosphorus atom was manually
formed and both phenoxy groups were removed from the in-
hibitor structure. The covalently bound inhibitor was energy
minimized within the complex and subjected to covalent dock-
ing by using the AutoDock 4.2 program. This procedure yield-
ed a reasonable orientation of the ligand in the case of 42
with an R-configured phosphonate carbon atom. Docking indi-
cated that the “Eastern” benzguanidine moiety likely occupied
the S1 pocket and formed potential hydrogen bonds to resi-
dues Asp756 and Gly787, whereas the “Western” benzguani-
dine addresses the S3/S4 subsite and is well positioned for hy-
drogen-bonding interactions to residues Ser664 and Gln741.
Moreover, the phenylalanine residue of 42 was likely to form
p–p interactions with His617. However, covalent docking with
42 if the phosphonate carbon atom was S configured resulted
in a less plausible binding mode because of steric constraints
On the basis of this molecular docking approach, the (S,R)
configuration of the active epimer was concluded to be also
valid for 41A, 44A, and 45A (Table 1), whereby 42A and 45A,
which only differ in the aryloxy leaving groups, essentially pro-
duce the same phosphono monoester complex with a serine
protease. Our findings are in accordance with previous reports
on diastereomerically pure phosphonic peptides as serine pro-
tease inactivators. First, Boduszek et al. observed a single phos-
phonate diastereomer to be tenfold more potent at dipeptidyl
peptidase IV than the epimeric mixture.[42] Subsequently,
Walker et al. prepared two diphenyl aminophosphonates by
asymmetric synthesis and found that the corresponding R
enantiomer was the stronger inactivator of cathepsin G, chy-
motrypsin, and human neutrophil elastase.[46] Winiarski et al.
were able to isolate single diastereomers and compared the in-
hibitory activity towards human neutrophil elastase, porcine
pancreatic elastase, chymotrypsin, and trypsin. In all cases, the
epimer with an R-configured phosphonate warhead was the
more potent inactivator.[38]
In the case of the two epimers of 43 (Table 1), the (R,R) con-
figuration was assigned to the active matriptase-2 inhibitor
43B. Hence, the R configuration at the phosphonate carbon
atom permitted a more favorable accommodation of the
ligand in the active site, irrespective of the inverted configura-
tion at the amino acid, that is, R instead of S. The modeled
complex of 43B and matriptase-2 obtained by covalent dock-
Chem. Eur. J. 2016, 22, 8525 – 8535
www.chemeurj.org
8530
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
ing is presented in the Supporting Information (Figure S2).
Compounds 42A, 42B, 43A, and 43B represent the four ste-
reoisomers of this structure with 42A and 43A forming one
and 42B and 43B the other enantiomeric pair. As a matter of
course, each pair possessed the same NMR signals and, thus,
in the case of 43, the (R,R)-configured isomer has to be as-
signed as 43B. For the one enantiomeric pair, 42A was shown
to be the more active matriptase-2 inhibitor, that is, the euto-
mer. For the other pair, 43B represents the eutomer. The two
eutomers displayed a similar binding mode within the active
site of matriptase-2, as illustrated by an overlay shown in the
Supporting Information (Figure S3). Finally, a comparison of
the two eutomers revealed that, if the R configuration applies
to the phosphonate, the S configuration is preferred at the
amino acid. Accordingly, 42A exhibited the strongest inhibi-
tion among the four stereoisomers. Introduction of two para-
methylthio groups into 42A further improved the activity, and
the resultant 45A turned out to the most potent matriptase-2
inactivator of the whole series. The kinetic analysis of 45A is
reported in Figure 3.
activation of matriptase-2, as shown in Figure S4 in the Sup-
porting Information.
The efficient inactivation of matriptase-2 by phosphono bis-
benzguanidines prompted us to design a first activity-based
probe for matriptase-2 by utilizing this chemotype. Two of the
three required elements of an activity-based probe have al-
ready been assembled in the structures of inactivators 41–45,
that is, the phosphonate warhead and the benzguanidine rec-
ognition elements. It was intended to introduce the fluores-
cent reporter tag as part of the central amino acid. Thus, a fluo-
rescent amino acid of rather small size and with physicochemi-
cal properties similar to those of phenylalanine was chosen.
Coumarins with donor groups at position 7 represent a widely
used class of fluorescent dyes. Their small molecular size, high
fluorescence quantum yields, and large Stokes shifts, as well as
chemical and enzymatic stability, are favored properties.[41,47–49]
Thus, for the generation of the activity-based probe 51
(Scheme 2), we employed the coumarin-labeled amino acid
48.[47] Attempts to incorporate 48 into a bis(4-(methylthio)phe-
nylphosphono) bisbenzguanidine failed. However, we succeed-
ed in preparing the final compound 51, a structural analogue
of inactivator 42. The final steps in the preparation of 51 were
essentially the same as those described for the nonlabeled
phosphono bisbenzguanidines. Compound 51 contains two
benzguanidine moieties to interact with the S1 and S3/S4 spe-
cificity pockets, the phosphonate warhead for an irreversible
attachment of the probe to the target, an appropriate coumar-
in as a fluorescent label, and the central amino acid in the pre-
ferred S configuration. As executed for the other diastereo-
mers, the racemic compound 51 was separated into the two
epimers, designated as 51A and 51B. In Table 2, the kinetic
parameters of 51, 51A, and 51B are outlined, which revealed
again inhibitory activity for the R-configured phosphonate
(51A), whereas the epimer possessing the S-configured phos-
phonate (51B) was inactive toward matriptase-2.
The applicability of 51 A as an activity-based probe for ma-
triptase-2 was demonstrated by using the conditioned
medium of stably transfected HEK cells as a source of matrip-
tase-2, which was tagged with the c-Myc peptide and polyhis-
tidine (His6). The conditioned medium was collected and con-
centrated, then different amounts (20–100 mg of total protein)
were incubated with 100 mm compound 51 A for 30 min and
subjected to sodium dodecylsulfate polyacrylamide-gel electro-
phoresis (SDS-PAGE). A fluorescent band at 30 kDa was clearly
visible in samples containing matriptase-2 (Figure 4A, lanes 1–
5). This demonstrated the sensitivity of the probe to detect the
amount of matriptase-2 that is present in 20 mg of total pro-
tein. A detection limit for matriptase-2 could not be deter-
mined because of the lack of sufficiently purified enzyme. The
specificity of the probe for matriptase-2 was verified with HEK
mock cells as a negative control; these cells were carrying the
empty vector and expressing no endogenous matriptase-2.[2]
The conditioned medium of HEK mock cells (100 mg of total
protein) was treated likewise with 51 A and no fluorescent
band was visible at 30 kDa (Figure 4A, lane 6), which provided
evidence of labeled matriptase-2 in lanes 1–5. The protein pat-
terns of the conditioned media were visualized by Coomassie
Figure 3. Inhibition of human matriptase-2 by phosphono dipeptide 45A.
From top to bottom: uninhibited reaction, 0.4, 0.8, 1.2, 1.6, and 2 mm. Values
of the observed rate constant (kobs) were obtained from nonlinear regression
of the progress curves by using the equation [P]=vi(1ÀeÀk Ât)/kobs+d, in
obs
which [P] is the product concentration, vi is the initial rate, and d is the
offset. The inset shows a plot of kobs values (mean of duplicate measure-
ments) versus the inhibitor concentration [I]. Linear regression gave a kobs/[I]
value of (1240Æ30)mÀ1 sÀ1
.
Whereas the cell-free enzyme assays for matriptase-2 were
performed with the conditioned medium of transfected
human embryonic kidney (HEK) cells as the enzyme
source,[7,11,29] inactivator 45A was chosen for a further study
with intact cells. Matriptase-2 is a membrane-bound protease
with an extracellular catalytic domain, so inhibition of the cell-
surface-associated enzymatic activity was investigated. For this
purpose, adherent, transfected HEK cells in a 96-well plate
were washed and treated with 45A and the consumption of
the peptidic substrate was monitored. Actually, treatment with
750 nm 45 A led to an instantaneous and nearly complete in-
Chem. Eur. J. 2016, 22, 8525 – 8535
www.chemeurj.org
8531
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Scheme 2. Synthesis of the activity-based probe 51. Reaction conditions: a) AcOH/HBr, RT; b) HATU, DIPEA, DMF, RT; c) TFA, CH2Cl2, RT; d) HATU, DIPEA, DMF,
RT; e) SnCl2·2H2O, H2O, EtOAc, reflux; f) HgCl2, TEA, CH2Cl2, RT; g) TFA, CH2Cl2, RT, prep. HPLC, HCl, lyophilization.
Table 2. Inhibition of matriptase-2 by the activity-based probes of type
51.
[b,c]
Comp.[a]
Enzyme inactivation, kinac/Ki [MÀ1 sÀ1
]
Human
Human
Bovine
Bovine
matriptase-2 thrombin
factor Xa trypsin
51 (S, rac) 34.5Æ0.7
51A (S, R) 50.4Æ0.6
51B (S, S) n.i.
IC50 =(21.5Æ2.2) mm[d] n.i.
108Æ8
IC50 =(45.8Æ5.6) mm[d] n.i.
148Æ10[e]
IC50 =(15.5Æ1.7) mm[d] n.i.
92.2Æ12.1[e]
[a] The stereochemistry is given in parentheses. The first entry refers to
the configuration at the amino acid, the second to the configuration at
the aminophosphonate. [b] IC50 and kinac/Ki values were determined from
duplicate measurements with five different inhibitor concentrations.
[c] n.i.: no inhibition relates to less than 30% product formation after
30 min at an inhibitor concentration of 30 mm, which corresponds to an
IC50 value of >13 mm. Determined from duplicate measurements. [d] No
time-dependent inhibition was observed. Rates in the presence of five
different inhibitor concentrations, obtained by linear regression of the
progress curves, were plotted against inhibitor concentrations to obtain
IC50 values. [e] The following kinac/Ki values were obtained for inactivation
of human trypsin: (81.5Æ14.6)mÀ1 sÀ1 for (S,R)-51A and (58.7Æ
11.6)mÀ1 sÀ1 for (S,S)-51B.
Figure 4. Imaging of human matriptase-2 with the fluorescent probe 51A.
Different amounts of conditioned medium of transfected HEK cells express-
ing matriptase-2 (HEK-MT2; total protein amounts of 20 mg/18 mL to 100 mg/
18 mL) were incubated with 100 mm 51A. In a control experiment, a condi-
tioned medium of HEK mock cells (100 mg/18 mL) was treated with 100 mm
51A. Labeling reactions were performed for 30 min at 378C. The proteins
were separated by SDS-PAGE and visualized by A) fluorescence detection.
B) The untreated conditioned medium was analyzed by western blotting
with a monoclonal mouse anti-c-Myc antibody. M: molecular mass marker.
Chem. Eur. J. 2016, 22, 8525 – 8535
www.chemeurj.org
8532
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
brilliant blue staining (Figure S5C in the Supporting Informa-
tion), which indicated sufficiently selective labeling of the
target matriptase-2 in the presence of a large excess of other
proteins. The indistinct signal visualized by fluorescence detec-
tion in the range of 40–170 kDa (Figure 4A) resulted from un-
specific interaction of the probe with the proteins present in
high concentrations that originate from the medium used for
cell culturing and from the HEK cells (see Figure S5C and S5D
in the Supporting Information).
Experimental Section
Preparation of phosphono bisbenzguanidines 45A and 45B
Benzyl (bis(4-(methylthio)phenoxy)phosphoryl)(4-nitrophenyl)-
methylcarbamate (10): 4-(Methylthio)phenol (2; 4.2 g, 30.0 mmol)
was added to a flask containing phosphorus trichloride (1.37 g,
0.87 mL, 10.0 mmol) dissolved in acetonitrile (40 mL). The mixture
was heated under reflux for 3 h. The solvent was removed in
vacuo and the crude product, the triarylphosphite 3, was used di-
rectly for the amidoalkylation reaction. 4-Nitrobenzaldehyde (6;
1.51 g, 10.0 mmol), benzyl carbamate (4; 1.51 g, 10.0 mmol), and
phosphite 3 were combined with glacial acetic acid (50 mL). The
solution was stirred at 908C for 3 h. The acetic acid was removed
in vacuo and the crude product was dissolved in MeOH (100 mL).
The solution was kept at 48C overnight until the precipitation of
the protected a-aminophosphonate was complete. Compound 10
was obtained by filtration and drying as a white solid (3.66 g,
60%); m.p. 147–1508C (lit. [19]: m.p. 1358C); 1H NMR (500 MHz,
As a positive control to prove the identity of matriptase-2,
the conditioned media of HEK cells expressing matriptase-2
and HEK mock cells were analyzed by SDS-PAGE, without treat-
ment with 51 A (Figure 4B; corresponding Ponceau S staining
in Figure S5D in the Supporting Information) and matriptase-2
was visualized by western blot analysis by using a monoclonal
mouse anti-c-Myc antibody (Figure 4B, lanes 1–5). Notably, the
sensitivity of the activity-based probe 51A to label matriptase-
2 was comparable to the reliably efficient immunodetection of
the Myc tag. Although western blotting addresses the total
amount of the target, activity-based probes such as 51A lead
to selective labeling of only the catalytically active form.
2
[D6]DMSO, 308C): d=2.42 (s, 3H), 2.43 (s, 3H), 5.06 (d, J=12.3 Hz,
1H), 5.14 (d, 2J=12.3 Hz, 1H), 5.83 (dd, 3J=10.1 Hz, 2J=23.6 Hz,
3
3
1H), 6.97 (d, J=8.2 Hz, 2H), 7.02 (d, J=7.9 Hz, 2H), 7.21–7.23 (m,
4H), 7.29–7.39 (m, 5H), 7.91 (dd, J=1.9 Hz, 3J=8.8 Hz, 2H), 8.25
4
(d, 3J=8.5 Hz, 2H), 9.04 ppm (d, 3J=9.8 Hz, 1H); 13C NMR
(125 MHz, [D6]DMSO, 308C): d=15.38, 15.41, 52.63 (d, J=154.2 Hz),
66.53, 120.93 (d, J=4.0 Hz), 121.02 (d, J=3.7 Hz), 123.65, 127.71,
127.75, 128.14, 128.50, 129.79, 129.83, 135.04, 135.18, 136.62,
142.12, 147.28 (d, J=9.7 Hz), 147.48, 147.51, 147.61 (d, J=9.7 Hz),
156.08 ppm (d, J=8.4 Hz); LC-MS (ESI): purity 98%, m/z 611.16
[M+H]+, 608.90 [MÀH]À; elemental analysis: calcd for
C29H27N2O7PS2: C 57.04, H 4.46, N 4.59; found: C 57.02, H 4.45, N
4.78.
Conclusion
We discovered phosphono bisbenzguanidines as dipeptidomi-
metic inactivators for matriptase-2. The design was based on
the twofold introduction of benzguanidine moieties to occupy
the S1 and S3/S4 binding sites, respectively. Furthermore,
a phosphonate warhead was introduced to facilitate the cova-
lent linkage to the active-site serine. Nine compounds (39–47)
comprising this series were synthesized and evaluated toward
the serine proteases matriptase-2, human thrombin, bovine
factor Xa, and trypsin. Structural optimization and structure–ac-
tivity relationship analyses led to the development of com-
pound 45 with a kinac/Ki value of 1750mÀ1 sÀ1 towards matrip-
tase-2. Each product was obtained as a mixture of two diaste-
reomers and the biologically active products were separated
into their single epimers. The inhibitors with R configuration at
the phosphonate carbon atom were found to be more potent
than the corresponding S epimers. The R-configured 45A pos-
sessed a kinac/Ki value of 2790mÀ1 sÀ1 and represents the most
potent irreversible inhibitor of matriptase-2 known so far. This
phosphono bisbenzguanidine might serve as a starting point
for further structure modifications of matriptase-2 inactivators.
To verify the suitability of 45A, we demonstrated for the first
time that the activity of membrane-bound matriptase-2 of ad-
herent, intact cells was abolished upon treatment with a low-
molecular-weight inhibitor. By applying the chemotype of
phosphono bisbenzguanidines, we successfully employed the
activity-based probe 51A for direct in-gel fluorescence detec-
tion of matriptase-2 in a complex protein mixture. The phos-
phono bisbenzguanidine inactivators are expected to serve as
valuable tool compounds for future investigations of the key
role of matriptase-2 in iron homeostasis. In particular, 45A
might be useful to block matriptase-2 activity in control experi-
ments in the development of activity-based probes of other
chemotypes targeting matriptase-2.
(S)-tert-Butyl 1-((bis(4-(methylthio)phenoxy)phosphoryl)(4-nitro-
phenyl)methylamino)-1-oxo-3-phenylpropan-2-ylcarbamate (21):
The protected a-aminophosphonate 10 (0.696 g, 1.14 mmol) was
dissolved in a solution of HBr (33%) in acetic acid (20 mL) and
stirred for 30 min at room temperature. The solvent was removed
in vacuo and the crude product 13 was obtained after precipita-
tion in ethyl acetate (50 mL). Boc-l-Phe-OH (14; 0.302 g,
1.14 mmol) was dissolved in anhydrous DMF (20 mL) and activated
for 15 min at room temperature with HATU (0.433 g, 1.14 mmol)
and DIPEA (0.439 g, 3.42 mmol). Afterwards, amine 13 (0.635 g,
1.14 mmol) was added and the mixture was stirred overnight at
room temperature.[35] Product 21 was purified by column chroma-
tography on silica gel with petroleum ether:ethyl acetate (3:1) as
the eluent to obtain a white solid (0.512 g, 62%); m.p. 89–928C;
1H NMR (500 MHz, [D6]DMSO, 308C, mixture of diastereomers, ratio
1
of approximately 1:1 according to H NMR spectroscopy): d=1.25
(s, 9H), 1.31 (s, 9H), 2.40 (s, 3H), 2.41–2.42 (m, 9H), 2.68–2.76 (m,
2H), 2.81–2.91 (m, 2H), 4.40–4.47 (m, 1H), 4.49–4.56 (m, 1H), 6.02
3
2
2
(dd, J=9.5 Hz, J=23.0 Hz, 1H), 6.11 (dd, 3J=9.6 Hz, J=23.2 Hz,
3
1H), 6.95 (d, J=8.5 Hz, 2H), 6.98–7.07 (m, 8H), 7.16–7.30 (m, 18H),
3
3
3
7.77 (d, J=7.3 Hz, 2H), 7.92 (d, J=7.3 Hz, 2H), 8.22 (d, J=8.9 Hz,
2H), 8.29 (d, 3J=8.9 Hz, 2H), 9.38 (d, 3J=10.1 Hz, 1H), 9.63 ppm
(dd, 4J=2.7 Hz, 3J=9.6 Hz, 1H); 13C NMR (125 MHz, [D6]DMSO,
308C, mixture of diastereomers, ratio of approximately 1:1 accord-
1
ing to H NMR spectroscopy): d=15.34, 15.38, 28.21, 28.27, 37.26,
37.47, 49.68 (d, J=155.6 Hz), 50.05 (d, J=153.9 Hz), 55.57, 55.62,
78.26, 78.28, 121.00, 121.16 (d, J=3.5 Hz), 121.43 (d, J=3.2 Hz),
123.69, 126.34, 126.41, 127.71, 127.75, 128.01, 128.16, 129.26,
129.40, 129.63, 129.67, 129.80, 129.84, 135.22, 135.28, 137.73,
137.93, 141.92, 142.15, 147.19–147.51 (m), 155.43, 155.54, 172.30–
172.45 ppm (m); LC-MS (ESI): purity 95%, m/z 741.37 [M+NH4]+; el-
Chem. Eur. J. 2016, 22, 8525 – 8535
www.chemeurj.org
8533
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
emental analysis: calcd for C35H38N3O8PS2: C 58.08, H 5.29, N 5.81;
found: C 58.16, H 5.66, N 5.67.
1n HCl were added to the product to form the hydrochloride salt.
The final products 45A and 45B were obtained as white solids
after lyophilization (92 mg A and 7 mg B, 8% over three steps).
(S)-Bis(4-(methylthio)phenyl) (2-(4-nitrobenzamido)-3-phenylpro-
panamido)(4-nitrophenyl)methylphosphonate (34): The N-pro-
tected phosphonate 21 (2.512 g, 3.47 mmol) was dissolved in a so-
lution of dichloromethane and trifluoroacetic acid (1:1, 20 mL) and
stirred for 1 h at room temperature. Product 24 was obtained after
removal of the solvent in vacuo. 4-Nitrobenzoic acid (26; 0.580 g,
3.47 mmol) was dissolved in anhydrous DMF (20 mL) and activated
for 15 min at room temperature with HATU (1.319 g, 3.47 mmol)
and DIPEA (1.345 g, 10.41 mmol). Afterwards, amine 24 (2.560 g,
3.47 mmol) was added and the mixture was stirred overnight at
room temperature.[35] A volume of 1 mL of n-heptane was added
to the solution and a white precipitate was formed. The pure prod-
uct 34 was obtain by suction filtration (1.180 g, 44%); m.p. 191–
1938C; 1H NMR (500 MHz, [D6]DMSO, 308C, mixture of diastereo-
Diastereomer 45A with (COCH-S, POCH-R) configuration: M.p.
169–1728C; 1H NMR (500 MHz, [D6]DMSO, 308C): d=2.40 (s, 3H),
3
2.41 (s, 3H), 2.97–2.99 (m, 2H), 5.10 (q, J=8.7 Hz, 1H), 5.90 (dd,
3J=9.8 Hz, 2J=22.4 Hz, 1H), 6.99 (d, 3J=8.2 Hz, 2H), 7.00 (d, 3J=
3
3
8.2 Hz, 2H), 7.26–7.24 (m, 7H), 7.27 (t, J=8.2 Hz, 4H), 7.38 (d, J=
3
7.3 Hz, 2H), 7.56 (s, 4H), 7.64–7.66 (m, 6H), 7.92 (d, J=8.6 Hz, 2H),
8.78 (d, 3J=8.5 Hz, 1H), 9.78 (d, 3J=7.9 Hz, 1H), 10.08 ppm (brs,
2H); 13C NMR (125 MHz, [D6]DMSO, 308C): d=15.42, 15.43, 37.32,
49.55 (d, J=156.1 Hz), 54.82, 121.20 (d, J=3.7 Hz), 121.41 (d, J=
3.5 Hz), 123.12, 123.94, 126.45, 127.71, 128.13, 129.13, 129.41,
129.77, 129.81, 131.12, 132.09, 135.04, 135.57, 135.59, 138.08,
138.59, 147.38 (d, J=10.2 Hz), 147.48 (d, J=10.4 Hz), 155.98,
156.01, 165.52, 171.92 (d, J=7.2 Hz); LC-MS (ESI): purity 97%, m/z
797.4 [M+H]+; HRMS (ESI+): m/z calcd for C39H41N8O5PS2 [M+H]+:
797.2452; found: 797.2452.
1
mers, ratio of approximately 7:1 (A:B) according to H NMR spec-
troscopy): d=2.38 (s, 3H), 2.39 (s, 3H), 2.40 (s, 3H), 2.42 (s, 3H),
Diastereomer 45B with (COCH-S, POCH-S) configuration: M.p.
167–1708C; 1H NMR (500 MHz, [D6]DMSO, 308C): d=2.40 (s, 3H),
2.42 (s, 3H), 2.95 (d, J=6.6 Hz, 2H), 5.13 (q, J=6.6 Hz, 1H), 5.90
2.91–3.05 (m, 4H), 4.97–5.02 (m, 1H), 5.11–5.17 (m, 1H), 6.06 (dd,
2
3
2
3J=9.6 Hz, J=23.2 Hz, 1H), 6.16 (dd, J=9.8 Hz, J=22.7 Hz, 1H),
3
3
3
3
6.99 (t, J=8.0 Hz, 4H), 7.07 (d, J=7.9 Hz, 2H), 7.10–7.28 (m, 7H),
7.34 (d, 3J=7.0 Hz, 2H), 7.37 (d, 3J=7.0 Hz, 2H), 7.83 (dd, 4J=
3
2
3
(dd, J=8.1 Hz, J=18.5 Hz, 1H), 6.98 (d, J=7.1 Hz, 2H,), 6.99 (d,
3J=6.9 Hz, 2H), 7.14 (t, 3J=6.2 Hz, 1H), 7.18 (d, 3J=7.4 Hz, 2H),
3
3
1.9 Hz, J=8.8 Hz, 2H), 7.93–7.96 (m, 4H), 8.03 (d, J=9.1 Hz, 2H),
3
3
3
3
3
7.19 (d, J=7.2 Hz, 2H), 7.22 (t, J=6.3 Hz, 2H), 7.27 (d, J=6.9 Hz,
8.24 (d, J=8.8 Hz, 2H), 8.27–8.32 (m, 2H), 9.02 (d, J=8.6 Hz, 1H),
9.11 (d, 3J=8.6 Hz, 1H), 9.69 (dd, 4J=2.4 Hz, 3J=9.6 Hz, 1H),
9.89 ppm (dd, 4J=2.9 Hz, 3J=9.5 Hz, 1H); 13C NMR (125 MHz,
[D6]DMSO, 308C, mixture of diastereomers, ratio of approximately
7:1 (A:B) according to 1H NMR spectroscopy): d=15.33, 15.38,
37.06, 37.28, 49.87 (d, J=154.4 Hz), 54.77, 55.01, 121.01 (d, J=
4.0 Hz), 121.28 (d, J=3.5 Hz), 123.62, 123.73, 126.51, 126.58, 127.67,
127.70, 127.78, 128.14, 128.28, 129.00, 129.08, 129.22, 129.37,
129.70, 129.74, 129.95, 129.89, 135.24, 137.64, 137.94, 139.50,
139.57, 141.80, 141.92, 147.16–147.44 (m), 149.28, 164.98, 164.90,
171.66 ppm (d, J=7.7 Hz); the following C atoms of the B dia-
steromer were not detectable: POCH, CONHCP, and the C1 carbon
atoms of both phenoxy groups; LC-MS (ESI): purity 97%, m/z 773.3
[M+H]+; elemental analysis: calcd for C37H33N4O9PS2: C 57.51, H
4.30, N 7.25; found: C 57.37, H 4.52, N 7.25.
3
4H), 7.38 (d, J=6.1 Hz, 2H), 7.41 (s, 4H), 7.43–7.50 (m, 4H), 7.65
(dd, 4J=1.2 Hz, 3J=7.1 Hz, 2H), 7.90 (d, 3J=7.1 Hz, 2H), 8.71 (d,
3J=7.1 Hz, 1H), 9.78 (d, J=8.6 Hz, 1H), 9.63–9.92 ppm (brs, 2H);
3
13C NMR (125 MHz, [D6]DMSO, 308C): d=15.38, 15.39, 121.14 (d, J=
4.5 Hz), 121.40 (d, J=3.5 Hz), 123.37, 124.16, 127.68, 128.10, 129.10,
129.36, 129.73, 129.77, 132.19, 135.05, 155.65, 155.69, 164.79 ppm;
the following C atoms were not detectable: CONHCH2, COCH,
POCH, CONHCP, and the C1 carbon atoms of both phenoxy
groups; LC-MS (ESI): purity 94%, m/z 797.2 [M+H]+; HRMS (ESI+):
m/z calcd for C39H41N8O5PS2 [M+H]+: 797.2452; found: 797.2455.
Acknowledgements
D.H. holds a fellowship from the Bonn International Graduate
School of Drug Sciences (BIGS DrugS). D.H. was also supported
by a fellowship from the NRW International Graduate Research
School Biotech-Pharma, and N.F. by a fellowship from the
Jürgen Manchot Foundation, Düsseldorf, Germany. M.S. and
M.M. are supported by the Maria von Linden program of the
University of Bonn. M.S. is also supported by the German Re-
search Foundation (STI 660/1-1). The authors thank Tamara
Scheidt for assistance.
(S)-Bis(4-(methylthio)phenyl) (2-(4-guanidinobenzamido)-3-phe-
nylpropanamido)(4-guanidinophenyl)methylphosphonate dihy-
drochloride (45): SnCl2·2H2O (3.09 g, 13.7 mmol) and H2O (0.99 g,
54.8 mmol) were added to a stirred solution of 34 (1.061 g,
1.37 mmol) in EtOAc (40 mL). The solution was stirred under reflux
for 4 h. The solvent was removed in vacuo and the residue was
suspended in 2n NaOH (100 mL). The mixture was extracted with
dichloromethane (3”100 mL). The combined organic extracts were
washed with H2O (100 mL) and brine (100 mL). The material was
dried over Na2SO4 and evaporated to obtain the corresponding
amino phosphonate intermediate 36, which was purified by
column chromatography with ethyl acetate. The amino phospho-
nate 36 (0.600 g, 0.84 mmol) was treated with N,N’-di-Boc-S-meth-
ylisothiourea (37; 0.439 g, 1.51 mmol), HgCl2 (0.433 g, 1.60 mmol),
and Et3N (0.255 g, 2.52 mmol) in dichloromethane (20 mL) at room
temperature for 24 h. The solution was passed through a pad of
Celite and washed with dichloromethane (50 mL) and methanol
(50 mL). The solvents were removed in vacuo. The resulting Boc-
protected guanidine 38 was purified by column chromatography,
with ethyl acetate:petroleum ether:TEA (1:1:0.002). Compound 38
was dissolved in dichloromethane (5 mL) and a 50% solution of tri-
fluoroacetic acid in dichloromethane (10 mL) was added. After
being stirred at room temperature for 2 h, the solvent was evapo-
rated. The resulting oil was purified by preparative HPLC. The HPLC
method was as follows: MeOH:H2O (40:60), 17 min. A few drops of
Keywords: enzymes
· fluorescent probes · inhibitors ·
peptidomimetics · serine proteases
[1] T. H. Bugge, T. M. Antalis, Q. Wu, J. Biol. Chem. 2009, 284, 23177–23181.
[2] G. Velasco, S. Cal, V. Quesada, L. M. Sµnchez, C. López-Otín, J. Biol.
Chem. 2002, 277, 37637–37646.
[3] a) K. E. Finberg, M. M. Heeney, D. R. Campagna, Y. Aydinok, H. A. Pear-
son, K. R. Hartman, M. M. Mayo, S. M. Samuel, J. J. Strouse, K. Markianos,
N. C. Andrews, M. D. Fleming, Nat. Genet. 2008, 40, 569–571; b) X. Du,
E. She, T. Gelbart, J. Truksa, P. Lee, Y. Xia, K. Khovananth, S. Mudd, N.
Mann, E. M. Moresco, E. Beutler, B. Beutler, Science 2008, 320, 1088–
1092; c) A. R. Folgueras, F. M. de Lara, A. M. Pendµs, C. Garabaya, F. Ro-
dríguez, A. Astudillo, T. Bernal, R. Cabanillas, C. López-Otín, G. Velasco,
Blood 2008, 112, 2539–2645.
Chem. Eur. J. 2016, 22, 8525 – 8535
www.chemeurj.org
8534
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[4] a) K. Finberg, J. Clin. Invest. 2013, 123, 1424–1427; b) M. Stirnberg, M.
Gütschow, Curr. Pharm. Des. 2013, 19, 1052–1061; c) C. Y. Wang, D.
Meynard, H. Y. Lin, Front. Pharmacol. 2014, 5, 114; d) C. J. McDonald, L.
Ostini, N. Bennett, N. Subramaniam, J. Hooper, G. Velasco, D. F. Wallace,
V. N. Subramaniam, Am. J. Physiol. Cell Physiol. 2015, 308, C539–C547.
[5] L. Silvestri, A. Pagani, A. Nai, I. De Domenico, J. Kaplan, C. Camaschella,
Cell Metab. 2008, 8, 502–511.
[32] D. DuchÞne, E. Colombo, A. DØsilets, P. L. Boudreault, R. Leduc, E. Mar-
sault, R. Najmanovich, J. Med. Chem. 2014, 57, 10198–10204.
[33] a) E. Sabidó, T. Tarragó, S. Niessen, B. F. Cravatt, E. Giralt, ChemBioChem
2009, 10, 2361–2366; b) P. M. Mumford, G. J. Tarver, M. Shipman, J. Org.
Chem. 2009, 74, 3573–3575; c) E. Sabidó, T. Tarragó, E. Giralt, Bioorg.
Med. Chem. 2010, 18, 8350–8355.
[34] a) A. B. Berger, P. M. Vitorino, M. Bogyo, Am. J. Pharmacogenomics 2004,
´
4, 371–381; b) M. Fonovic, M. Bogyo, Expert Rev. Proteomics 2008, 5,
[6] a) J. E. Maxson, J. Chen, C. A. Enns, A. S. Zhang, J. Biol. Chem. 2010, 285,
39021–39028; b) M. Rausa, M. Ghitti, A. Pagani, A. Nai, A. Campanella,
G. Musco, C. Camaschella, L. Silvestri, J. Cell. Mol. Med. 2015, 19, 879–
888.
[7] M. Stirnberg, E. Maurer, A. Horstmeyer, S. Kolp, S. Frank, T. Bald, K.
Arenz, A. Janzer, K. Prager, P. Wunderlich, J. Walter, M. Gütschow, Bio-
chem. J. 2010, 430, 87–95.
[8] F. BØliveau, A. DØsilets, R. Leduc, FEBS J. 2009, 276, 2213–2226.
[9] M. Wysocka, N. Gruba, A. Miecznikowska, J. Popow-Stellmaszyk, M. Güt-
schow, M. Stirnberg, N. Furtmann, J. Bajorath, A. Lesner, K. Rolka, Bio-
chimie 2014, 97, 121–127.
721–730; c) J. Clerc, B. I. Florea, M. Kraus, M. Groll, R. Huber, A. S. Bach-
mann, R. Dudler, C. Driessen, H. S. Overkleeft, M. Kaiser, ChemBioChem
2009, 10, 2638–2643; d) D. A. Shannon, C. Gu, C. J. McLaughlin, M.
Kaiser, R. A. van der Hoorn, E. Weerapana, ChemBioChem 2012, 13,
2327–2330; e) F. Zou, M. Schmon, M. Sienczyk, R. Grzywa, D. Palesch,
B. O. Boehm, Z. L. Sun, C. Watts, R. Schirmbeck, T. Burster, Anal. Biochem.
2012, 421, 667–672; f) H. Y. Hu, D. Vats, M. Vizovisek, L. Kramer, C. Ger-
manier, K. U. Wendt, M. Rudin, B. Turk, O. Plettenburg, C. Schultz,
Angew. Chem. Int. Ed. 2014, 53, 7669–7673; Angew. Chem. 2014, 126,
7802–7806.
[35] L. E. Sanman, M. Bogyo, Annu. Rev. Biochem. 2014, 83, 249–273.
[36] a) L. I. Willems, H. S. Overkleeft, S. I. van Kasteren, Bioconjugate Chem.
2014, 25, 1181–1191; b) P. Yang, K. Liu, ChemBioChem 2015, 16, 712–
724.
[10] L. De Falco, M. Sanchez, L. Silvestri, C. Kannengiesser, M. U. Muckenthal-
er, A. Iolascon, L. Gouya, C. Camaschella, C. Beaumont, Haematologica
2013, 98, 845–853.
[11] M. T. Sisay, T. Steinmetzer, M. Stirnberg, E. Maurer, M. Hammami, J. Bajor-
ath, M. Gütschow, J. Med. Chem. 2010, 53, 5523–5535.
[12] a) Y. Ginzburg, S. Rivella, Blood 2011, 118, 4321–4330; b) P. J. Schmidt,
M. D. Fleming, Hematol. Oncol. Clin. North Am. 2014, 28, 387–401.
[13] P. J. Schmidt, I. Toudjarska, A. K. Sendamarai, T. Racie, S. Milstein, B. R.
Bettencourt, J. Hettinger, D. Bumcrot, M. D. Fleming, Blood 2013, 121,
1200–1208.
[37] J. Oleksyszyn, L. Subotkowska, P. Mastalerz, Synthesis 1979, 985–986.
´
[38] Ł. Winiarski, J. Oleksyszyn, M. Sienczyk, J. Med. Chem. 2012, 55, 6541–
6553.
[39] A. Belyaev, X. Zhang, K. Augustyns, A. M. Lambeir, I. De Meester, I. Ve-
dernikova, S. ScharpØ, A. Haemers, J. Med. Chem. 1999, 42, 1041–1052.
´
´
[40] M. Skorenski, J. Oleksyszyn, M. Sienczyk, Tetrahedron Lett. 2013, 54,
4975–4977.
[14] S. Guo, C. Casu, S. Gardenghi, S. Booten, M. Aghajan, R. Peralta, A. Watt,
S. Freier, B. P. Monia, S. Rivella, J. Clin. Invest. 2013, 123, 1531–1541.
ˇ
[41] M. D. Mertens, J. Schmitz, M. Horn, N. Furtmann, J. Bajorath, M. Mares,
M. Gütschow, ChemBioChem 2014, 15, 955–959.
´
[15] M. Sienczyk, J. Oleksyszyn, Curr. Med. Chem. 2009, 16, 1673–1687.
[42] B. Boduszek, J. Oleksyszyn, C. M. Kam, J. Selzler, R. E. Smith, J. C. Powers,
J. Med. Chem. 1994, 37, 3969–3976.
[16] a) M. H. Potashman, M. E. Duggan, J. Med. Chem. 2009, 52, 1231–1246;
b) D. S. Johnson, E. Weerapana, B. F. Cravatt, Future Med. Chem. 2010, 2,
949–964; c) J. Singh, R. C. Petter, T. A. Baillie, A. Whitty, Nat. Rev. Drug
Discovery 2011, 10, 307–317.
[17] R. A. Bauer, Drug Discovery Today 2015, 20, 1061–1073.
[18] J. Oleksyszyn, B. Boduszek, C. M. Kam, J. C. Powers, J. Med. Chem. 1994,
37, 226–231.
[43] B. Lejczak, P. Kafarski, J. Szewczyk, Synthesis 1982, 412–414.
[44] a) E. Parisini, P. Metrangolo, T. Pilati, G. Resnati, G. Terraneo, Chem. Soc.
Rev. 2011, 40, 2267–2278; b) L. A. Hardegger, B. Kuhn, B. Spinnler, L.
Anselm, R. Ecabert, M. Stihle, B. Gsell, R. Thoma, J. Diez, J. Benz, J. M.
Plancher, G. Hartmann, D. W. Banner, W. Haap, F. Diederich, Angew.
Chem. Int. Ed. 2011, 50, 314–318; Angew. Chem. 2011, 123, 329–334;
c) R. Wilcken, M. O. Zimmermann, A. Lange, A. C. Joerger, F. M. Boeckler,
J. Med. Chem. 2013, 56, 1363–1388; d) J. Ren, Y. He, W. Chen, T. Chen,
G. Wang, Z. Wang, Z. Xu, X. Luo, W. Zhu, H. Jiang, J. Shen, Y. Xu, J. Med.
Chem. 2014, 57, 3588–3593.
[45] a) J. A. Bertrand, J. Oleksyszyn, C. M. Kam, B. Boduszek, S. Presnell, R. R.
Plaskon, F. L. Suddath, J. C. Powers, L. D. Williams, Biochemistry 1996, 35,
3147–3155; b) E. Burchacka, M. Zdzalik, J. S. Niemczyk, K. Pustelny, G.
Popowicz, B. Wladyka, A. Dubin, J. Potempa, M. Sienczyk, G. Dubin, J.
Oleksyszyn, Protein Sci. 2014, 23, 179–189.
[46] B. Walker, S. Wharry, R. J. Hamilton, S. L. Martin, A. Healy, B. J. Walker,
Biochem. Biophys. Res. Commun. 2000, 276, 1235–1239.
[47] M. Frizler, M. D. Mertens, M. Gütschow, Bioorg. Med. Chem. Lett. 2012,
22, 7715–7718.
[48] a) T. Murase, T. Yoshihara, K. Yamada, S. Tobita, Bull. Chem. Soc. Jpn.
2013, 86, 510–519; b) D. Häußler, M. Gütschow, Synthesis 2016, 245–
255.
´
[19] M. Sienczyk, J. Oleksyszyn, Bioorg. Med. Chem. Lett. 2006, 16, 2886–
2890.
[20] S. Serim, S. V. Mayer, S. H. Verhelst, Org. Biomol. Chem. 2013, 11, 5714–
5721.
[21] S. Serim, P. Baer, S. H. G. Verhelst, Org. Biomol. Chem. 2015, 13, 2293–
2299.
[22] a) D. S. Jackson, S. A. Fraser, L. M. Ni, C. M. Kam, U. Winkler, D. A. John-
son, C. J. Froelich, D. Hudig, J. C. Powers, J. Med. Chem. 1998, 41, 2289–
2301; b) C. M. Brown, M. Ray, A. A. Eroy-Reveles, P. Egea, C. Tajon, C. S.
Craik, Chem. Biol. 2011, 18, 48–57.
[23] J. Joossens, O. M. Ali, I. El-Sayed, G. Surpateanu, P. Van der Veken, A. M.
Lambeir, B. Setyono-Han, J. A. Foekens, A. Schneider, W. Schmalix, A.
Haemers, K. Augustyns, J. Med. Chem. 2007, 50, 6638–6646.
[24] a) J. Joossens, P. Van der Veken, A. M. Lambeir, K. Augustyns, A. Haem-
ers, J. Med. Chem. 2004, 47, 2411–2413; b) J. Joossens, P. Van der Veken,
G. Surpateanu, A. M. Lambeir, I. El-Sayed, O. M. Ali, K. Augustyns, A.
Haemers, J. Med. Chem. 2006, 49, 5785–5793.
[49] a) T. Terai, T. Nagano, Curr. Opin. Chem. Biol. 2008, 12, 515–521; b) P. W.
Elsinghorst, W. Härtig, S. Goldhammer, J. Grosche, M. Gütschow, Org.
Biomol. Chem. 2009, 7, 3940–3946; c) R. S. Agnes, F. Jernigan, J. R. Shell,
V. Sharma, D. S. Lawrence, J. Am. Chem. Soc. 2010, 132, 6075–6080;
d) O. Vµzquez, M. I. Sµnchez, J. L. MascareÇas, M. E. Vµzquez, Chem.
Commun. 2010, 46, 5518–5520; e) S. Nizamov, K. I. Willig, M. V. Sednev,
V. N. Belov, S. W. Hell, Chem. Eur. J. 2012, 18, 16339–16348; f) L. Meime-
tis, J. C. Carlson, R. J. Giedt, R. H. Kohler, R. Weissleder, Angew. Chem. Int.
Ed. 2014, 53, 7531–7534; Angew. Chem. 2014, 126, 7661–7664; g) F.
Kohl, J. Schmitz, N. Furtmann, A. C. Schulz-Fincke, M. D. Mertens, J. Küp-
pers, M. Benkhoff, E. Tobiasch, U. Bartz, J. Bajorath, M. Stirnberg, M. Güt-
schow, Org. Biomol. Chem. 2015, 13, 10310–10323.
[25] L. M. Ni, J. C. Powers, Bioorg. Med. Chem. 1998, 6, 1767–1773.
´
[26] M. Sienczyk, A. Lesner, M. Wysocka, A. Legowska, E. Pietrusewicz, K.
Rolka, J. Oleksyszyn, Bioorg. Med. Chem. 2008, 16, 8863–8867.
[27] a) W. D. Wilson, F. A. Tanious, A. Mathis, D. Tevis, J. E. Hall, D. W. Boykin,
Biochimie 2008, 90, 999–1014; b) T. L. Huang, A. Mayence, J. J.
Van den Eynde, Bioorg. Med. Chem. 2014, 22, 1983–1992.
[28] a) O. Vµzquez, M. I. Sµnchez, J. Martínez-Costas, M. E. Vµzquez, J. L. Mas-
careÇas, Org. Lett. 2010, 12, 216–219; b) M. I. Sµnchez, O. Vµzquez, M. E.
Vµzquez, J. L. MascareÇas, Chem. Eur. J. 2013, 19, 9923–9929.
[29] S. Dosa, M. Stirnberg, V. Lülsdorff, D. Häußler, E. Maurer, M. Gütschow,
Bioorg. Med. Chem. 2012, 20, 6489–6505.
[30] E. Maurer, M. T. Sisay, M. Stirnberg, T. Steinmetzer, J. Bajorath, M. Güt-
schow, ChemMedChem 2012, 7, 68–72.
[31] M. Hammami, E. Rühmann, E. Maurer, A. Heine, M. Gütschow, G. Klebe,
T. Steinmetzer, Med. Chem. Commun. 2012, 3, 807–813.
Received: January 15, 2016
Published online on May 23, 2016
Chem. Eur. J. 2016, 22, 8525 – 8535
www.chemeurj.org
8535
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim