Small Cause, Great Impact: Modification of the Guanidine Group in the RGD Motif Controls Integrin Subtype Selectivity

Article abstract of DOI:10.1002/anie.201508713

Due to its unique role as a hydrogen-bond donor and its positive charge, the guanidine group is an important pharmacophoric group and often used in synthetic ligands. The chemical modification of the guanidine group is often considered to destroy its function. Herein, we show that the N-methylation, N-alkylation, or N-acylation of the guanidine group can be used to modify the receptor subtype specificity of the integrin ligand cilengitide. Using the αvβ6/α5β1-biselective ligand c(isoDGRkphg) and the αvβ6-specific ligand c(FRGDLAFp(NMe)K(Ac) as examples, we show that the binding affinities of the ligands can be fine-tuned by this method to enhance the selectivity for αvβ6. Furthermore, we describe a new strategy for the functionalization of integrin ligands. By introducing longer N-alkylguanidine and N-acylguanidine groups, we are able to simultaneously identify a hitherto unknown anchoring point and enhance the subtype selectivity of the ligand.

Full text of DOI:10.1002/anie.201508713

Angewandte
Communications
Chemie
International Edition: DOI: 10.1002/anie.201508713
German Edition: DOI: 10.1002/ange.201508713
Integrin Ligands (2)
Small Cause, Great Impact: Modification of the Guanidine Group in
the RGD Motif Controls Integrin Subtype Selectivity
Tobias G. Kapp, Maximilian Fottner, Oleg V. Maltsev, and Horst Kessler*
Abstract: Due to its unique role as a hydrogen-bond donor
and its positive charge, the guanidine group is an important
pharmacophoric group and often used in synthetic ligands. The
chemical modification of the guanidine group is often consid-
ered to destroy its function. Herein, we show that the N-
methylation, N-alkylation, or N-acylation of the guanidine
group can be used to modify the receptor subtype specificity of
the integrin ligand cilengitide. Using the avb6/a5b1-biselective
ligand c(isoDGRkphg) and the avb6-specific ligand c-
The initial goal of early ligand development was primarily
to obtain ligands possessing high avb3 affinity. At that time,
selectivity for other integrin subtypes (except integrin aIIbb3)
was only a minor issue. The spatial screening of cyclic RGD-
containing peptides yielded the pentapeptide c(RGDf-
[
8,9]
(NMe)V) (cilengitide)
as the most potent avb3 ligand
(0.54 nm); cilengitide also has a relatively high a5b1 binding
[8,10]
affinity (15.4 nm).
When the importance of different
integrins in biology became evident, the need for integrin
subtype selective peptides became clear. In a parallel
approach, peptidomimetics were investigated by our group
(
FRGDLAFp(NMe)K(Ac) as examples, we show that the
binding affinities of the ligands can be fine-tuned by this
method to enhance the selectivity for avb6. Furthermore, we
describe a new strategy for the functionalization of integrin
ligands. By introducing longer N-alkylguanidine and N-
acylguanidine groups, we are able to simultaneously identify
a hitherto unknown anchoring point and enhance the subtype
selectivity of the ligand.
[11]
and others, leading to completely selective avb3 and a5b1
ligands. These molecules could be functionalized for medical
(e.g. molecular imaging) and biophysical applications. Deter-
mining the distinct differences in the biological functions of
these two subtypes opens the door for their future applica-
[12]
tions (e.g. their use in personalized medicine).
The binding mode of RGD integrin ligands was elucidated
I
ntegrins sense and regulate cell attachment both between
from the crystal structure of cilengitide bound to the avb3
[13]
cells and to the extracellular matrix (inside-out signaling and
extracellular domain.
The key elements of this binding
[
1]
outside-in signaling, respectively). One common feature of
the integrin family is a heterodimeric structure that consists of
interaction in all RGD-binding integrins include the metal-
ion-dependent adhesion site (MIDAS), located in the b
subunit, and the strong side-on bidendate H-bonding inter-
action (guanidine carboxylate) in the a subunit. Recently, the
[
2]
a and b subunits. These structures form 24 different
subtypes in mammals, which can be classified according to
their binding partners (e.g. laminin, collagen). Eight of these
subtypes form the Arg-Gly-Asp (RGD)-binding class. This
tripeptide sequence is present in a variety of extracellular
crystal structure of the a5b1 headpiece in complex with
[
3]
[14]
a peptidic ligand
unveiled a remarkable and important
difference in the binding modes of the guanidine group in the
av- and a5-binding pockets. In the a5 subunit the guanidine
group participates in two interactions: the bidendate side-on
binding of HN and HN to Asp227, and the end-on binding
[4]
matrix proteins such as fibronectin and vitronectin. Differ-
ent integrins are involved in many pathological processes such
[
5]
as metastasis and tumor vascularization. The structural
differences among the subtypes are often small because most
of them share different natural RGD-containing proteins;
however, they exhibit drastically different binding affinities.
For instance, the subtype avb3 can bind to the ECM ligands
vitronectin, von-Willebrand-factor, fibrinogen, osteopontin,
and fibronectin. This makes it challenging to develop
selective compounds, especially for subtypes that are often
co-expressed and present in different cellular and patholog-
ical situations. In this context, the fibronectin binding
subtypes avb3 and a5b1 are particularly important; these
subtypes both play important roles in angiogenesis and tumor
d
w1
interaction of HNw1 and HNw2 with Gln221. In the av pocket,
the only interaction of guanidine is the bidendate H-bonding
to Asp218 (Figure 1).
Provided with this structural insight, we were interested in
whether this small difference could be used to rationally
design subtype-selective peptidic integrin ligands. Theoret-
ically, two molecules with differently modified guanidine
group in the arginine side chain and opposing selectivity
should be obtainable in a model system of a biselective ligand.
In one case, blocking the end-on interaction of the molecule
[6]
with the receptor (modification of N ) would result in higher
w1
[7]
development.
avb3 selectivity since the significant binding energy contri-
buting to the a5b1 interaction would be lost. In the opposite
case, disturbing the side-on interaction by blocking HNd
would allow only an end-on interaction, resulting in a higher
selectivity for the subtype a5b1. As a model system, we chose
the avb3/a5b1 biselective cyclic peptide cilengitide. To block
the interaction, distinct hydrogen atoms in the guanidine
group were replaced by methyl groups. This alteration
prevents hydrogen bonding while introducing only minimal
[
*] T. G. Kapp, M. Fottner, Dr. O. V. Maltsev, Prof. Dr. H. Kessler
Institute for Advanced Study and Center for Integrated Protein
Science (CIPSM), Technische Universität München
Lichtenbergstrasse 4, 85747 Garching b. München (Germany)
E-mail: kessler@tum.de
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201508713.
1
540
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 1540 –1543

Angewandte
Communications
Chemie
Figure 2. Cilengitide (1) and its substituted guanidine derivatives 2–7.
methylation of the ornithine side chain on a solid support,
cyclization, o-Ns deprotection in solution, and guanidinyla-
tion with N,N’-di-Boc-1H-pyrazole-1-carboxamidine (SI).
The results of the in vitro evaluation of the ligands
correspond surprisingly well with the theoretical model
Figure 1. a) Linear RGD-ligand binding to the subtypes a5b1 (top) and
avb3 (bottom). Differences are found in the interaction of arginine
guanidine group with the a5 and av subunit, respectively, as shown
enlarged in the box. Two interactions are found for a5b1 (side-on and
end-on); for avb3 only side-on is observed. b) Concept of subtype
selectivity: a biselective guanidine compound (middle) is methylated
described above. For 2, which is methylated at N , the
w1
a5b1 binding affinity vanished completely (> 1 mm), whereas
the avb3 binding affinity decreased from 0.54 to 8.4 nm
(Table 1). The terminal methyl group prevents the end-on
on N (right) ! improvement of avb3 selectivity by blocking end-on
w
interaction to Q221; N -methylation (left) blocks side-on interaction to
d
binding interaction. In contrast, methylation at N blocks the
d
D218, end-on interaction is still possible ! improvement of a5b1
side-on binding interaction. In av, where the latter is the only
selectivity.
steric and electronic changes into the system. To
probe this hypothesis, four methylated analogues
of cilengitide were synthesized, including one di-
and one tetramethylated compound (Figure 2).
The syntheses of 2, 4, 5, 6 and 7 began with
solid-phase synthesis (SPPS) of the linear, orthog-
onally Dde-protected peptide H-Asp(OtBu)-d-
Phe-(NMe)Val-Orn(Dde)-Gly-OH, followed by
cyclization and subsequent orthogonal Dde depro-
tection (Figure 3). The final compounds were
obtained after guanidinylation followed by depro-
tection of the acid-labile side-chain protecting
groups. For the guanidinylation step, two different
Figure 3. Synthesis of compounds 2–7.
protocols were used. The terminally monoalky-
lated compounds 2 and 6 were obtained after
guanidinylation with alkylated N,N’-di-Boc-1H-
pyrazole-1-carboxamidine precursors. These are obtained by
Mitsunobu reaction of N,N’-di-Boc-1H-pyrazole-1-carboxa-
midine with the corresponding alcohol, for example, ethanol
interaction, this modification leads to a total loss of binding
affinity. In the a5 pocket, this side-on interaction is also
blocked, whereas the end-on interaction is still possible,
resulting in a a5b1-selective peptide (3) with reduced overall
binding affinity (51 nm). Furthermore, a compound with
a higher degree of methylation and, consequently, blocking of
side-on and end-on binding, should possess no binding affinity
at all. The biological evaluation of 4 and 5 demonstrated no
affinity (> 1000 nm) for both subtypes avb3 and a5b1.
[15]
for the generation of an ethyl group. The precursor for the
acetylated guanidine 7 was synthesized starting from S-
[16]
methylpseudothiourea (Supporting Information, SI). For 4
and 5, the ornithine side-chain amine groups were guanidiny-
lated using the commercially available N,N’-dimethylthiourea
and N,N,N’,N’-tetramethylthiourea in the presence of NEt3
and HgCl as base and activator; 3 was obtained after the N-
2
Angew. Chem. Int. Ed. 2016, 55, 1540 –1543
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1541

Angewandte
Communications
Chemie
Table 1: Binding affinities of cilengitide (1) and its derivatives 2–7 for
avb3 and a5b1.
Table 2: Binding affinities of compounds 10–14 for avb6 and a5b1.[a] as
[
a]
determined in a solid phase binding assay using a previously established
[20]
protocol. The selectivity profile for each compound against other
integrin subtypes can be found in the SI.
avb3 [nm]
a5b1 [nm]
[
b]
1
2
3
4
5
6
7
0.54Æ0.04
8.4Æ0.8
i.a.
15.4Æ1.7
i.a.
aVb6 [nm]
a5b1 [nm]
51Æ4.2
10
11
12
13
14
c(phgisoDGRk)
c(phgisoDGR(Me)k)
c(FRGDLAFp(NMe)K(Ac))
c(FR(Me)GDLAFp(NMe)K(Ac))
18Æ2.5
13Æ2.1
8.3Æ0.7
120Æ15
72Æ9
i.a.
[b]
i.a.
i.a.
i.a.
i.a.
i.a.
0.26Æ0.04
1.3Æ0.23
0.28Æ0.03
[
b]
0.93Æ0.1
[
b]
6.5Æ0.5
15.3Æ0.9
c(FR(Et)GDLAFp(NMe)K(Ac))
i.a.
[
a] The binding affinities were determined in a solid-phase binding assay
[a] The binding affinities were determined in a solid-phase binding assay
using a previously established protocol. The selectivity profile for each
[
20]
[20]
using a previously established protocol. The selectivity profile for each
compound can be found in the SI. [b] Cilengitide. i.a. >1000 nm.
compound can be found in the SI. [b] Modification of arginine on the
terminal Nw1 is indicated with the substituent brackets following R
(R(Alk)). i.a. >1000 nm; lowercase letters (p,h,g,k) denote d-amino
acids.
Having closely examined the crystal structure model of
the av binding pocket and taken into consideration the
tolerance of the 6-methyl-2-aminopyridine scaffold in pre-
120 nm, whereas avb6 affinity slightly increased from 18 to
13 nm (Table 2). In another example, we investigated the
recently developed helix-mimicking 9-mer cyclic peptide
[
17]
viously described avb3 peptidomimetic ligands,
we sup-
posed that alkyl groups, which are larger than the methyl
group in 2, might also be tolerated. Indeed, we found that the
presence of ethyl guanidine in this position resulted in
a ligand with much better binding affinity than the methylated
derivative. With a binding affinity of 0.93 nm, 6 retained
almost full avb3 affinity of cilengitide while no binding could
be observed for a5b1 and all other investigated subtypes.
Hence, 6 shows one of the best avb3 binding affinities and
subtype selectivity profiles ever described (see SI). It is well
[22]
c(FRGDLAFp(NMe)K(Ac)) (12) developed in our lab.
This peptide shows very high binding affinity (0.26 nm) for
avb6 and a residual affinity of 72 nm for a5b1. As in the
previously discussed examples, the method could be success-
fully applied, as demonstrated by the reduction in binding
affinity for a5b1 to > 1000 nm by methylation of the Nw1 of the
arginine guanidine group (13). Nevertheless, this methylation
also significantly reduced the avb6 affinity to 1.3 nm, which
could be overcome by introducing an ethyl substituent (14), as
in the above example of cilengitide.
As demonstrated above, the substitution of guanidine
allows us to influence and tune the subtype selectivity for av-
and a5-containing integrins. Considering the structural fea-
tures of the binding pockets of both a subunits, it appears that
space in elongation of HNw2 is not restricted. A long alkyl or
acyl chain on Nw2 would allow a new way of functionalizing
integrin ligands by simultaneously interacting with the
guanidine and linking out of the binding pocket.
[18]
known that in acylguanidines, the basicity is reduced by 4–5
orders of magnitude (pK ꢀ 8) compared to basic guanidine
a
compounds (pK ꢀ 13); for alkylated compounds, the pK is
a
a
[19]
reported to be 11–12. This implies that at physiological pH,
alkylguanidines are generally positively charged, whereas
acylguanidines are at least partly uncharged. This reduced
basicity is often used to tune the pharmacokinetic properties
of bioactive molecules to, for example, improve their
bioavailability. Furthermore, in both the charged and
uncharged states, acylguanidines form planar systems. In
addition to the N-alkylated molecules described above, one
compound (7) with a terminally acetylated guanidine Nw1 was
also investigated. Interestingly, the shift in selectivity seen for
the alkylated molecules was not observed for 7. Whereas the
binding affinity for a5b1 remained unchanged (approx.
For a proof-of-concept study, two functional compounds
were synthesized: one acylated on Nw2 with 6-acetyl-amino-
hexanoic acid (8) and the other alkylated on Nw2 with 6-
acetylaminohexane (9) (Figure 4). To synthesize 8 and 9, the
orthogonally Dde-deprotected peptide P1 was guanidinylated
with the corresponding tailor-made N,N’-di-Boc-1H-pyra-
1
5 nm), the avb3 affinity was reduced from 0.54 to 6.5 nm,
indicating a strong shift in selectivity towards a5b1. To
explain this effect, we hypothesized that the carbonyl of the
acetyl group can interact with one amide proton of Gln221 in
the a5 subunit, compensating for the lost interaction between
HNw1 and the carbonyl group of Gln221.
Surprised by the drastic effect on selectivity caused by
these rather simple modifications, we decided to attempt to
apply this technique in a wider sense to tune and adjust the
binding affinity profiles of integrin ligands. In particular, we
wanted to investigate whether terminal alkylation could be
used as a general approach to remove or significantly reduce
a5b1 binding affinity by blocking the a5b1-specific interac-
tion with Gln221. Our first system was the avb6/a5b1-
[
21]
biselective peptide c(isoDGRkphg) (10).
After methyla-
tion (11) of N , binding affinity for a5b1 decreased from 8 to
Figure 4. Functionalization of cilengitide via guanidine.
w1
1
542
www.angewandte.org
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 1540 –1543

Angewandte
Communications
Chemie
zole-1-carboxamidine (alkylation) and N-(tert-butoxycar-
bonyl)-S-methylisothiourea precursor (acylation; see SI).
The results obtained in the in vitro binding assay clearly
verified our hypothesis. With both modifications, the first
guanidine-functionalized integrin ligands were obtained and
showed affinities in the range of approximately 20 nm for
avb3 with complete selectivity against a5b1. The alternative
interaction seen in 7, which resulted in a good a5b1 affinity, is
not possible here due to steric hindrance. Instead, the chain is
oriented out of the binding pocket. The same is observed for
the hexylguanidine linker. The binding affinities of this new
class of functionalized molecules are comparable to that of
long-established c(RGDfK(AhxAc)) (24 nm), the “gold stan-
dard” of functionalized integrin ligands.
[10] M. Aumailley, M. Gurrath, G. Müller, J. Calvete, R. Timpl, H.
Kessler, FEBS Lett. 1991, 291, 50 – 54.
[
11] a) R. Dayam, F. Aiello, J. Deng, Y. Wu, A. Garofalo, X. Chen, N.
Neamati, J. Med. Chem. 2006, 49, 4526 – 4534; b) D. Heckmann,
A. Meyer, L. Marinelli, G. Zahn, R. Stragies, H. Kessler, Angew.
Chem. Int. Ed. 2007, 46, 3571 – 3574; Angew. Chem. 2007, 119,
3
641 – 3644; c) F. Zanardi, P. Burreddu, G. Rassu, L. Auzzas, L.
Battistini, C. Curti, A. Sartori, G. Nicastro, G. Menchi, N. Cini,
A. Bottonocetti, S. Raspanti, G. Casiraghi, J. Med. Chem. 2008,
5
1, 1771 – 1782; d) L. Gentilucci, G. Cardillo, S. Spampinato, A.
Tolomelli, F. Squassabia, R. De Marco, A. Bedini, M. Baiula, L.
Belvisi, M. Civera, J. Med. Chem. 2010, 53, 106 – 118; e) L.
Marinelli, A. Meyer, D. Heckmann, A. Lavecchia, E. Novellino,
H. Kessler, J. Med. Chem. 2005, 48, 4204 – 4207; f) D. Heckmann,
A. Meyer, B. Laufer, G. Zahn, R. Stragies, H. Kessler,
ChemBioChem 2008, 9, 1397 – 1407.
[
[
12] a) F. Rechenmacher, S. Neubauer, J. Polleux, C. Mas-Moruno,
M. De Simone, E. A. Calvalcanti-Adam, J. P. Spatz, R. Fässler,
H. Kessler, Angew. Chem. Int. Ed. 2013, 52, 1572 – 1575; Angew.
Chem. 2013, 125, 1612 – 1616; b) S. Neubauer, F. Rechenmacher,
A. J. Beer, F. Curnis, K. Pohle, C. D’Alessandria, H.-J. Wester, U.
Reuning, A. Corti, M. Schwaiger, H. Kessler, Angew. Chem. Int.
Ed. 2013, 52, 11656 – 11659; Angew. Chem. 2013, 125, 11870 –
11873.
13] a) J. P. Xiong, T. Stehle, B. Diefenbach, R. Zhang, R. Dunker,
D. L. Scott, A. Joachimiak, S. L. Goodman, M. A. Arnaout,
Science 2001, 294, 339 – 345; b) J. P. Xiong, T. Stehle, R. Zhang,
A. Joachimiak, M. Frech, S. L. Goodman, M. A. Arnaout,
Science 2002, 296, 151 – 155.
In conclusion, we were able to show that simple modifi-
cation of the guanidine group in the integrin ligand cilengitide
can be used to tune integrin subtype selectivity in two
directions, leading to two isomeric compounds varying only in
the position of a methyl group but having completely opposite
selectivity properties for av and a5 integrins. Furthermore,
the applicability of this procedure was shown for two more
systems, significantly improving selectivity against a5b1.
Finally, we were able show that by substituting the guanidine
group, functionalization of molecules is also possible.
[
[
14] M. Nagae, S. Re, E. Mihara, T. Nogi, Y. Sugita, J. Takagi, J. Cell
Biol. 2012, 197, 131 – 140.
15] a) G. Maccari, S. Sanfilippo, F. De Luca, D. Deodato, A. Casian,
M. C. D. Lang, C. Zamperini, E. Dreassi, G. M. Rossolini, J.-D.
Docquier, M. Botta, Bioorg. Med. Chem. Lett. 2014, 24, 5525 –
Acknowledgements
T.G.K acknowledges the International Graduate School for
Science and Engineering (IGSSE) of the Technische Uni-
versität München for financial support. H.K. thanks also the
CIPSM and the Deutsche Forschungsgemeinschaft for sup-
port.
5
529; b) S. Rossiter, C. L. Smith, M. Malaki, M. Nandi, H. Gill,
J. M. Leiper, P. Vallande, D. Selwood, J. Med. Chem. 2005, 48,
670 – 4678.
4
[16] S. Weiss, M. Keller, G. Bernhardt, A. Buschauer, B. Kçnig,
Bioorg. Med. Chem. 2010, 18, 6292 – 6304.
[
17] S. Neubauer, F. Rechenmacher, R. Brimioulle, F. S. Di Leva, A.
Bochen, T. R. Sobahi, M. Schottelius, E. Novellino, C. Mas-
Moruno, L. Marinelli, H. Kessler, J. Med. Chem. 2014, 57, 3410 –
Keywords: cilengitide · integrin ligands · peptide modifications ·
receptor subtype selectivity · RGD motif
3
417.
How to cite: Angew. Chem. Int. Ed. 2016, 55, 1540–1543
Angew. Chem. 2016, 128, 1564–1568
[
[
[
18] R. Kleinmaier, M. Keller, P. Igel, A. Buschauer, R. M.
Gschwind, J. Am. Chem. Soc. 2010, 132, 11223 – 11233.
19] S.-X. Xie, P. Ghorai, Q.-Z. Ye, A. Buschauer, R. Seifert, J.
Pharmacol. Exp. Ther. 2006, 317, 1139 – 1146.
20] A. O. Frank, E. Otto, C. Mas-Moruno, H. B. Schiller, L. Marine-
lli, S. Cosconati, A. Bochen, D. Vossmeyer, G. Zahn, R. Stragies,
E. Novellino, H. Kessler, Angew. Chem. Int. Ed. 2010, 49, 9278 –
[
1] S. J. Shattil, C. Kim, M. H. Ginsberg, Nat. Rev. Mol. Cell Biol.
010, 11, 288 – 300.
2
[
[
[
[
2] M. J. Humphries, Biochem. Soc. Trans. 2000, 28, 311 – 339.
3] R. O. Hynes, Cell 2002, 110, 673 – 687.
4] M. D. Pierschbacher, E. Rouslahti, Nature 1984, 309, 30 – 33.
5] a) S. L. Goodman, M. Picard, Trends Pharmacol. Sci. 2012, 33,
9
281; Angew. Chem. 2010, 122, 9465 – 9468.
[
[
21] A. Bochen, U. K. Marelli, E. Otto, D. Pallarola, C. Mas-Moruno,
F. S. Di Leva, H. Boehm, J. P. Spatz, E. Novellino, H. Kessler, L.
Marinelli, J. Med. Chem. 2013, 56, 1509 – 1519.
22] O. V. Maltsev, U. K. Marelli, T. G. Kapp, F. S. Di Leva, S.
Di Marco, M. Nieberler, U. Reuning, M. Schwaiger, E. Novel-
lino, L. Marinelli, H. Kessler, Angew. Chem. Int. Ed. 2015, 54,
DOI: 10.1002/anie.201508709; Angew. Chem. 2015, 127, DOI:
4
05 – 412; b) T. G. Kapp, F. Rechenmacher, T. R. Sobahi, H.
Kessler, Expert Opin. Ther. Pat. 2013, 23, 1273 – 1295; c) J. S.
Desgrosellier, D. A. Cheresh, Nat. Rev. Cancer 2010, 10, 9 – 22;
d) U. K. Marelli, F. Rechenmacher, T. R. A. Sobahi, C. Mas-
Moruno, H. Kessler, Front. Oncol. 2013, 3, 222.
[
6] E. F. Plow, T. A. Haas, L. Zhang, J. Loftus, J. W. Smith, J. Biol.
Chem. 2000, 275, 21785 – 21788.
1
0.1002/ange.201508709.
[
[
7] R. O. Hynes, Nat. Med. 2002, 8, 918 – 921.
8] M. A. Dechantsreiter, E. Planker, B. Mathä, E. Lohof, G.
Hçlzemann, A. Jonczyk, S. L. Goodman, H. Kessler, J. Med.
Chem. 1999, 42, 3033 – 3040.
Received: September 17, 2015
Revised: October 26, 2015
Published online: December 9, 2015
[
9] C. Mas-Moruno, F. Rechenmacher, H. Kessler, Anti-Cancer
Agents Med. Chem. 2010, 10, 753 – 768.
Angew. Chem. Int. Ed. 2016, 55, 1540 –1543
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1543