Linker-Region Modified Derivatives of the Deoxyhypusine Synthase Inhibitor CNI-1493 Suppress HIV-1 Replication

Article abstract of DOI:10.1002/ardp.201500323

The inhibition of cellular factors that are involved in viral replication may be an important alternative to the commonly used strategy of targeting viral enzymes. The guanylhydrazone CNI-1493, a potent inhibitor of the deoxyhypusine synthase (DHS), preve

Full text of DOI:10.1002/ardp.201500323

RCHH HARM
A P
Arch. Pharm. Chem. Life Sci. 2016, 349, 91–103
Archiv der Pharmazie
Full Paper
Linker-Region Modified Derivatives of the Deoxyhypusine
Synthase Inhibitor CNI-1493 Suppress HIV-1 Replication
1
2
2
3
3
€
€
€
Marcus Schroder , Adrian Kolodzik , Bjorn Windshugel , Marcel Krepstakies , Poornima Priyadarshini ,
Philip Hartjen4, Jan van Lunzen4, Matthias Rarey2, Joachim Hauber3, and Chris Meier1
1
Institute of Organic Chemistry, Department of Chemistry, Faculty of Sciences, University of Hamburg,
Hamburg, Germany
ZBH Center for Bioinformatics, University of Hamburg, Hamburg, Germany
Heinrich Pette Institute – Leibniz Institute for Experimental Virology, Hamburg, Germany
Department of Medicine, Infectious Diseases Unit, University Medical Center Hamburg-Eppendorf,
2
3
4
Hamburg, Germany
The inhibition of cellular factors that are involved in viral replication may be an important alternative
to the commonly used strategy of targeting viral enzymes. The guanylhydrazone CNI-1493, a potent
inhibitor of the deoxyhypusine synthase (DHS), prevents the activation of the cellular factor eIF-5A
and thereby suppresses HIV replication and a number of other diseases. Here, we report on the
design, synthesis and biological evaluation of a series of CNI-1493 analogues. The sebacoyl linker in
CNI-1493 was replaced by different alkyl or aryl dicarboxylic acids. Most of the tested derivatives
suppress HIV-1 replication efficiently in a dose-dependent manner without showing toxic side effects.
The unexpected antiviral activity of the rigid derivatives point to a second binding mode as previously
assumed for CNI-1493. Moreover, the chemical stability of CNI-1493 was analysed, showing a
successive hydrolysis of the imino bonds. By molecular dynamics simulations, the behaviour of the
parent CNI-1493 in solution and its interactions with DHS were investigated.
Keywords: Anti-HIV activity / Deoxyhypusine synthase / DHS / eIF-5A / Inhibitor synthesis
Received: September 14, 2015; Revised: December 2, 2015; Accepted: December 4, 2015
DOI 10.1002/ardp.201500323
Additional supporting information may be found in the online version of this article at the publisher’s web-site.
:
eIF-5A is catalysed by the sequential action of deoxyhypusine
synthase (DHS) and deoxyhypusine hydroxylase (DOHH;
Introduction
The eukaryotic initiation factor 5A (eIF-5A) is unique, since it is
to date the only known cellular protein that contains the
unusual amino acid hypusine (Ne-[4-amino-2-hydroxybutyl]-
lysine). This post-translational spermidine-dependent modifi-
cation is required for eIF-5A activity and occurs at Lys50 in the
human 154-amino acid eIF-5A protein [1, 2]. Hypusination of
Fig. 1). DHS mediates the transfer of the aminobutyl moiety
of spermidine to the e-NH2 group of Lys50 to form an
intermediate reaction substrate. This still inactive intermedi-
ate is subsequently hydroxylated by DOHH, resulting in the
active hypusine-modified protein.
Although eIF-5A has been initially described as an ‘initia-
tion’ factor of the protein synthesis [3, 4], more recent studies
demonstrated that eIF-5A rather promotes the elongation
step of translation [5, 6]. Interestingly, this eIF-5A activity
occurs in mammalian cells particularly during cellular stress
[7, 8], which may suggest that hypusinated eIF-5A acts in the
translation of a small and distinct subset of cellular transcripts.
In addition to affecting translational elongation, in higher
eukaryotes eIF-5A also participates in the nucleocytoplasmic
Correspondence: Prof. Chris Meier, Institute of Organic Chemistry,
Department of Chemistry, Faculty of Sciences, University of
Hamburg, Martin-Luther-King-Platz 6, 20146 Hamburg, Germany.
E-mail: chris.meier@chemie.uni-hamburg.de
Fax: þ49 40 428385592
ß 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.archpharm.com
91

RCHH HARM
A P
Arch. Pharm. Chem. Life Sci. 2016, 349, 91–103
M. Schroder et al.
€
Archiv der Pharmazie
(iNOS) [11]. It is noted, however, that nuclear eIF-5A has
been also reported to interact with XPO4/exportin4, another
CRM1-unrelated transport receptor [12]. Nevertheless, in
HIV-1 infection eIF-5A serves as a cellular cofactor of the
essential viral regulatory protein Rev, thereby promoting the
CRM1-dependent nuclear export of unspliced and incom-
pletely spliced retroviral mRNAs [13, 14]. Moreover, it was
demonstrated that inhibition of DHS by small-molecular
weight inhibitors or RNA interference (RNAi) suppressed the
cytoplasmic accumulation of Rev-regulated viral transcripts
and, in consequence, the formation of virus progeny [15, 16].
Thus, pharmacological interference with DHS activity may
provide a novel strategy to efficiently inhibit HIV-1 replica-
tion, particularly viral strains with high-level resistance to
inhibitors of current combinational antiretroviral therapy
(cART) [17]. In particular, the guanylhydrazone CNI-1493
(compound 1 in Fig. 2, the INN is semapimod), an experimental
anti-inflammatory drug [18], has been shown in vitro as well as
in living cells to be an efficient DHS inhibitor with activity not
only against HIV [15, 19] but also against other diseases or
pathogens [20, 21]. Concerning the ability of CNI-1493 to cross
cell membranes, it was shown that fetuin can mediate the
cellular uptake in macrophages [22]. Due to its size, CNI-1493
can not fit in the active site of the DHS. A recently solved co-
crystal structure of CNI-1493 bound to DHS revealed that
CNI-1493 was bound at the end of a narrow channel at the
surfaceoftheprotein.Intheboundform,themoleculeadopted
a U-shape structure with the guanidine-residues interacting
Figure 1. In vivo activation of eIF-5A catalysed by DHS and
DOHH.
translocation of specific mRNAs. For example, independent
evidence has been provided demonstrating that eIF-5A
regulates the nuclear exit of cellular transcripts via the
CRM1/exportin1 pathway, particularly encoding the surface
protein CD83 [9, 10] or inducible nitric oxide synthase
Figure 2. Summary of the synthesised CNI-1493 analogues 2–15.
ß 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.archpharm.com
92

RCHH HARM
A P
Arch. Pharm. Chem. Life Sci. 2016, 349, 91–103
HIV Inhibition by Blocking a Cellular Factor
Archiv der Pharmazie
with the protein side chains [23]. The combined data provided
the starting point for a structure-activity-relationship study to
developnovelinhibitorsofDHSthatmay,asoutlinedbefore,be
particularly useful in the treatment of diseases inflicted by
otherwise cART-resistant HIV-1. We report the synthesis and in
vitro test of various compounds derived from CNI-1493,
including the analysis of their DHS inhibitory capacity as well
as their effects on cellular metabolism and HIV-1 replication.
Moreover, the corresponding energy-minimised structures of
these compounds were superposed with that of CNI-1493 and
molecular dynamics simulation was employed for modelling
the binding between CNI-1493 and the DHS.
In summary, a series of CNI-1493 derivatives 2–14 were
synthesised for a structure-activity-relationship study. For this
small library of CNI-1493 analogues, an earlier reported
synthetic approach was used starting from 5-nitro isophthalic
acid 16, whose nitro group serves as an amino group
precursor [24, 25]. First, 5-nitro isophthalic acid 16 was
converted into the corresponding dichloride 18 using thionyl
chloride with a catalytic amount of DMF. In contrast to the
procedure reported in the literature, the following reaction
of 18 with diethyl magnesium malonate failed. Thus, in a one-
pot reaction, sodium hydride was used for the deprotonation
of diethylmalonate 17 before adding 5-nitro isophthaloyl
dichloride 18 leading to an intermediate, which was not
isolated but was directly reacted in a hydrolytic decarboxyl-
ation to 5-nitro-1,3-diacetylbenzene 19 (Scheme 1). The
reduction of the nitro group to yield 3,5-diacetylaniline 20
was achieved by tin(II) chloride [25]. Compound 15 was
commercially available.
Results and discussion
Synthesis of the CNI-1493 analogues
In order to analyse the structural requirements for
inhibition of DHS by CNI-1493 (1), the CNI-1493 structure
was varied by replacing the alkyl chain with more rigid
linkers or simply either shortened or elongated the alkyl
chain (Fig. 2).
The tetraketones 21–33 (Scheme 1) were prepared by
reaction of 20 with the corresponding diacid dichlorides in dry
CH2Cl2/pyridine according to
et al. [26].
a method from Bianchi
Scheme 1. Synthesis of 3,5-diacetylaniline 20 and its conversion to CNI-1493 (1) and the CNI-1493 analogues 2–13. Reagents and
conditions: i) SOCl2, DMF, reflux, [N2], 90 min; ii) NaH, THF, reflux, [N2], 4 h; iii) THF, reflux, [N2], 16 h; iv) AcOH/water/H2SO4, reflux,
3.5 h; i)–iv) 29%; v) SnCl2, conc. HCl, 50°C, 15 min, 98%; vi) 0.5 eq. dicarboxylic acid chloride, 2.7 eq. pyridine, CH2Cl2, rt, [N2], 2–15 h,
38–96%; vii) 5 eq. aminoguanidine hydrochloride, ethanol/water 9:1, conc. HCl, reflux, 3 h, 24–95%.
ß 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.archpharm.com
93

RCHH HARM
A P
Arch. Pharm. Chem. Life Sci. 2016, 349, 91–103
M. Schroder et al.
€
Archiv der Pharmazie
The final step was the conversion to the guanylhydrazone
derivatives 1–14 using aminoguanidine hydrochloride in
ethanol under acid catalysis [26]. As CNI-1493, all CNI-1493
analogues precipitated as pure compounds at ꢀ26°C and
were isolated as their hydrochlorides by filtration and
washing steps in yields from 24 to 95%.
Inhibition of the human DHS enzyme
In order to investigate the ability of the new compounds 2–14
to inhibit DHS, all compounds were tested in an enzymatic in
vitro DHS assay. Therefore, the DHS reaction was simulated
employing recombinant expressed eIF-5A and DHS together
with the substrates NADþ and 3H-labelled spermidine. The
reaction was performed with the indicated concentrations of
the derivatives 2–14 as well as with compound 15, CNI-1493 (1)
as a positive control or only DMSO as the negative control,
respectively. As a measure for DHS activity, the relative
amount of the transferred tritium-labelled aminobutyl
residue from spermidine to eIF-5A was detected. As shown
before, CNI-1493 (1) inhibited the DHS reaction in
a
concentration-dependent manner [15]. The inhibitory effect
was greater than 90% at the highest tested concentration
(5 mM). Interestingly, several of the CNI-1493 analogues
showed a similar effect on the enzymatic DHS reaction
(Fig. 3). Most of these active compounds belong to the group
with an aliphatic linker (Fig. 3A). For 7 (linker contains 12
methylene groups), the alkyl linker may be too long for
effective binding to DHS, which possibly explains the
observed decrease in activity.
In compound 2, the two head groups were directly linked
whereas the linker lengths of the other compounds varied
between two and twelve methylene groups. In comparison,
CNI-1493 (1) contains a (CH2)8 chain as linker. The amino-
guanidine moiety 15 alone, which was found in the
spermidine pocket of the CNI-1493-DHS co-crystal structure,
had no effect on DHS activity. Additionally, the truncated
derivative 14 showed nearly no inhibitory effect on DHS.
Among the CNI-1493 derivatives with a linker bearing a rigid
aromatic residue compound 12, in which the two head groups
are linked in 4 positions of the phenyl ring, showed an activity
similar to 1 (Fig. 3B).
Figure 3. DHS inhibition assay with aliphatic CNI-1493 ana-
logues (A) and with the aromatic CNI-1493 derivatives (B).
CNI-1493 (1) was used as a reference.
The four most promising inhibitors from this series of the
aliphatic and the aromatic linker class (Fig. 4) were further
tested in a HIV infection assay.
Inhibition of HIV-1 replication in cell culture
DHS-inhibitor candidates 3, 6, 12 and 13 as well as CNI-1493
(1) as a positive control were tested in regard to their
potential to inhibit the replication of HIV-1 in cell culture. For
this purpose, PM1 cells were preincubated for 8–10 days in
presence of the compounds at concentrations of 0.5, 1 and
2 mM or DMSO (solvent control) and then infected with the
CCR5-tropic HIV-1 strain BaL. After further cultivation for
7 days, p24 antigen levels in the cell culture supernatants and
cellular viability (measured by AlamarBlue staining) were
determined.
Figure 4. DHS inhibition potency of the derivatives selected for
the HIV infection assay.
ß 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.archpharm.com
94

RCHH HARM
A P
Arch. Pharm. Chem. Life Sci. 2016, 349, 91–103
HIV Inhibition by Blocking a Cellular Factor
Archiv der Pharmazie
As expected, CNI-1493 (1) effectively inhibited HIV-1 replica-
tion in a dose-dependent manner beginning at a concentra-
tion of 0.5 mM. CNI-1493 analogue 3 exhibited a very similar
inhibitory potential in this assay, showing efficient inhibition
of the replication of HIV-1 BaL beginning at 0.5 mM as well
(Fig. 5).
It is not clear why 6, which proved to be active in the in vitro
DHS-activity assay, showed no anti-HIV-1 activity in cell
culture at similar concentrations as the other DHS inhibitors
(Fig. 5). The aromatic CNI-1493 analogues, 12 and 13, also
inhibited HIV-1 replication albeit less potently, beginning at a
concentration of 2 mM (Fig. 6). The result that none of the
compounds exerted cellular toxicity in the tested concentra-
tion range (0.5–2 mM) is important to note (Figs. 5 and 6, lower
panels). Since the cells are cultivated for at least 17 days in
presence of CNI-1493 (1) and its analogues 3, 6, 12 and 13, the
maintained cellular metabolic activity including the measured
redox-potential indicated a good cellular tolerance of the
tested CNI-1493 derivatives.
inactive in the DHS assay (Fig. 3). However, this fragment
may be formed by hydrolysis of 1 in the crystallisation buffer
at longer periods of time. For a correct interpretation of the
in vitro assays reported herein, it was required to prove that 1
and thus the new analogues 2–14 were stable under the
conditions used in the assays. The hydrolysis of CNI-1493 using
the conditions summarised in Table 1 was monitored by
means of HPLC. 30-Azidothymidine (AZT) was used as an
internal standard. The UV-absorption spectrum of 1 showed
an intense absorption at 264 nm.
Conditions 3–5 represent the media used in HIV-inhibition
assays (3), crystallisation buffer experiments (4) and the
DHS-inhibition assay (5), respectively. PBS buffer was used as
an additional buffer. In acidic and slightly acidic media (pH
1.9–6.8), no hydrolysis of CNI-1493 (1) was observed for at
least 5 days. However, under neutral as well as higher pH
values, successive hydrolysis of 1 was detected (conditions 1, 4,
5, 10–13; Table 1).
In general, all observed hydrolysis products were more
lipophilic resulting in a shift to higher retention times on the
RP-18 column. Conditions 1 and 2 resulted in the same
decomposition products independent of the temperature.
However, 1 seemed to be more stable in DMSO alone than in
DMSO/water as concluded from a longer half-life of about
Hydrolysis of CNI-1493
In the co-crystal structure of DHS and CNI-1493 (1), an
aminoguanidine fragment (15) was observed in the active
site. It was discussed above that this fragment 15 was
Figure 5. Anti-retroviral activity and cytotoxicity of the aliphatic CNI-1493 derivatives 3 and 6. Mean values of at least six
independent infections are shown. Upper panels show the percentage of virus replication as compared with replication in solvent
control (DMSO)-cultures, which is arbitrarily set to 100% and depicted as a dotted line. AlamarBlue assays from the same cultures are
shown in the lower panels.
ß 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.archpharm.com
95

RCHH HARM
A P
Arch. Pharm. Chem. Life Sci. 2016, 349, 91–103
M. Schroder et al.
€
Archiv der Pharmazie
Figure 6. Antiretroviral activity and toxicity of the aromatic CNI-1493 derivatives 12 and 13. Mean values of at least six independent
infections are shown. Upper panels show the percentage of virus replication as compared with replication in solvent control
(DMSO)-cultures, which is arbitrarily set to 100% and depicted as a dotted line. AlamarBlue assays from the same cultures are shown
in the lower panels.
140 h compared to 87 h, respectively. In more basic buffers (pH
>7.5), the hydrolysis products formed were found to be
identical with those detected in H2O (pH 7.0) and in DMSO.
However, the hydrolysis in basic PBS buffer was significantly
faster (half-live about 45 h) than in water (87 h). The hydrolysis
of 1 in glycine buffer was analysed by mass spectrometry as an
example for the identification of the hydrolysis products.
Two hydrolysis products could be identified via LC-MS as the
monoketone (tR ¼ 8.85 min) and the diketone (tR ¼ 10.35 min)
(Supporting Information Figs. S1 and S2). The third hydrolysis
peak (tR ¼ 12.65 min) was not assigned by LC-MS, but is
assumed to be the triketone. Due to the MS analysis, it could
not be determined whether the second imine bond is cleaved
at the same side as the first one or at the second head group.
Table 1. Summary of the conditions used in the hydrolysis studies.
Condition
Buffer
pH Value
Condition
Buffer
pH Value
1
2
3
4
5
6
7
DMSO
Water
–
8
9
10
11
12
13
PBSd
PBSd
PBSd
PBSd
PBSd
PBSd
6.0
6.8
7.5
8.5
9.0
7.0
7.2
8.5
9.0
1.9
4.0
RPMI-1640a
Trisb
Glycinec
PBSd
11.0
PBSd
a RPMI-1640/heat-inactivated FCS (10%), 37°C.
b 100 mM Tris, 63% 2-methyl-2,4-pentanediol, 3 mM NAD, rt.
c 300 mM glycine, 37°C.
d 50 mM phosphate buffer, 37°C.
ß 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.archpharm.com
96

RCHH HARM
A P
Arch. Pharm. Chem. Life Sci. 2016, 349, 91–103
HIV Inhibition by Blocking a Cellular Factor
Archiv der Pharmazie
The half-live of 1 in the glycine buffer was determined to
about 139 h, which was rather long regarding the basic buffer
(condition 5, Table 1). Surprisingly, the hydrolysis in the
crystallisation buffer (condition 4, Table 1) resulted in a fast
decomposition of 1 with a half-live of only 22 h, which
perfectly explains the appearance of the fragment in the DHS
active site in the X-ray crystal structure analysis. In RPMI-1640,
cell media CNI-1493 was found to be stable at 37°C for at least
4 days. All results indicated a successive hydrolysis of the imine
bonds to the corresponding ketones under neutral and basic
buffer conditions. A cleavage of the amide bond in the
backbone was not observed, because no formation of the
arylamine 14 was detected. Due to these results and the fact,
that the CNI-1493 derivatives are modified in the diamide
linker moiety, they should display comparable hydrolysis
properties as obtained for compound 1.
Figure 7. Superpositioning of minimised structures of CNI-1493
(1) (green) and 11 (carbon atoms in grey, nitrogen atoms in
blue, oxygen atoms in red and hydrogen atoms in white) with
̊
an RMSD of 0.18 A.
In silico docking studies
In a recent co-crystal structure [23], CNI-1493 (1) was found to
bind to DHS at the surface of the protein, next to the entry to
the binding site in a sandwich-like conformation. To investi-
gate this binding site in more detail, a number of CNI-1493
analogues 2–14 were studied by computational means in
order to find out if the CNI-1493 analogues reported herein
were able to adopt a similar conformation at a low calculated
potential energy. Therefore, the conformation adopted by
CNI-1493 in the co-crystal structure was energy minimised
using the MMFFx94 force field within MOE [27]. The aliphatic
residue of the linker in CNI-1493 (1) has a length of eight
methylene groups. The length of the linker was modified in
the range of two to twelve methylene groups. In addition, the
oxalyl group was studied as a linker (Fig. 2). Afterwards,
energetically optimised conformations of these modifications
of CNI-1493 were generated using the MMFF94x force field.
Root-mean-square deviations (RMSD) were calculated with
YASARA [28]. To calculate the RMSD, hydrogen atoms were
neglected and structures were aligned and superposed
allowing flipping of symmetrical groups. These minimised
structures were superposed to the sandwich-like conforma-
tion of CNI-1493 as found in the co-crystal structure of the
DHS/CNI-1493 complex (Fig. 7).
were performed (Supporting Information Fig. S3). X-ray
crystal data of DHS complexed with CNI-1493 revealed the
ligand to bind next to the entry to the spermidine pocket in a
sandwich-like conformation. The guanidinium group of the
inner CNI-1493 head group (directly bound to DHS) forms p–p
interactions with the phenylalanine side chain of Phe247 and
may interact with Glu180. The phenyl ring of the inner CNI-
1493 half forms p–p stacking interactions with the phenyl ring
of the outer half. This part of the molecule also forms two salt
bridges between one of its guanidinium moieties and Glu180.
At first, the stability of the reported DHS/CNI-1493 inter-
actions was evaluated with an MD simulation using the
GROMACS software. As ring orientations are difficult to
compute in GROMACS, interactions between 1 and Phe247
were monitored indirectly by calculating the number of
van-der-Waals (vdW) interactions between both moieties.
The guanidinium group of the inner CNI-1493 half
extensively interacts with Phe247 (Supporting Information
Fig. S3A). On average, 11.1 vdW contacts between atoms of
the guanidinium moiety and the aromatic side chain were
measured. Compared to the crystal structure, we observed
significantly more vdW interactions (on average 5.15)
between the inner half’s phenyl ring and Phe247. In contrast
to 1, 14 revealed much less stable contacts between the
guanidinium moiety and Phe247 (Supporting Information
Fig. S3B). Only in a single simulation the molecule stably
interacted with the phenyl ring throughout the course of the
simulations. Also, interactions between the aromatic group of
the ligand and Phe247 are marginal. Moreover, we have not
found evidences for cation–p interactions. On average, only a
single vdW contact was observed. In simulations of the DHS
tetramer with bound CNI-1493, the salt bridge between the
ligand and Glu180 is stable in most simulations (average bond
>0.5), although with large fluctuations (Supporting Informa-
tion Fig. S3C). In addition, a salt bridge between 1 and Asp243,
located in the tunnel that connects the spermidine pocket
Although the optimal conformations of the CNI-1493
modifications slowly change with increasing length of the
linker, the sandwich-like conformation becomes unlikely for
shorter or rigid linkers. Therefore, the observed inhibitory
activity of the CNI-1493 analogues 2–4, 8, 9, 12 and 13 point to
a binding mode different to the observed sandwich-like
conformation, whereas 5–7, 10 and 11 can adopt
a
conformation similar to the sandwich-like form.
MD simulations
In order to evaluate the stability of the CNI-1493 binding
mode observed in the DHS:CNI-1493 complex and to investi-
gate why the truncated version 14 is unable to block the
DHS/eIF-5A interaction, comparative molecular dynamics
(MD) simulations of DHS in complex with 1 as well as 14
ß 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.archpharm.com
97

RCHH HARM
A P
Arch. Pharm. Chem. Life Sci. 2016, 349, 91–103
M. Schroder et al.
€
Archiv der Pharmazie
with the protein surface, is observed (Supporting Information
Fig. S3E). In comparison with 1, simulations including the
truncated molecule 14 did not reveal any significant interac-
tion with Glu180 (Supporting Information Fig. S3D) and
contacts with Asp243 were found only in a single simulation
run (Supporting Information Fig. S3F). Compared to the
extensive interactions between the guanidinium group and
Phe247, the aromatic rings of 1 interact much less with each
other (Supporting Information Fig. S3G). The number of vdW
contacts is even lower than the interactions between the
aromatic ring of the inner CNI-1493 head group and Phe247
(on average 3.2 vdW contacts). The significantly less pro-
nounced interactions are due to reorientation of the outer
half of the ligand. The RMSD-plots of CNI-1493 heavy atoms
clearly indicate the conformational change in four of the six
molecules simulated (Supporting Information Fig. S3H). Only
two CNI-1493 molecules revealed the stacked-like conforma-
tion observed in the X-ray crystal structure. In those cases, an
intramolecular hydrogen bond either formed between a
guanidinium moiety of one CNI-1493 half and the carbonyl
group of the second half or both amide moieties was observed
that restricts the conformational flexibility (Supporting
Information Fig. S3H).
The results confirm several protein–ligand interactions
observed between 1 and DHS in the X-ray crystal structure.
Moreover, an additional contact with an amino acid of the
tunnel that connects the spermidine pocket with the solvent
could be detected. In contrast to X-ray crystal data, the
simulations suggest that the stacked conformation of 1 as
observed in the X-ray crystal structure is not very stable. Only
in few simulations the aromatic rings of both head groups
were found to interact with each other. The simulations
suggested that 14 will not stably bind to DHS as none of the
interactions formed between DHS and 1 were present in the
DHS/14 complex. Comparing both simulations, the head
group alone seems to be not sufficient for inhibiting the
interaction of eIF-5A and DHS. The linker as well as the second
head group needs to establish additional contacts with DHS
and thus stabilise the interactions of the inner head group
with Glu180, Asp243 and Phe247.
are unable to adopt the conformation of 1 as observed in the
co-crystal structure with DHS. This may point to an alternative
binding mode of these compounds to the enzyme. This is
supported by the fact that 11 turned out to only marginally
inhibit the DHS, even though it can perfectly adapt the
sandwich-like shape as calculated by minimised energy
simulations. MD simulations showed that the stacked
conformation of 1 in the crystal seems to be not very stable
in solution, supporting the experimentally measured activity
of the rigid derivatives. Moreover, the simulations provide
structural evidence for the inability of 14 to inhibit DHS. None
of the new derivatives tested in cell culture against HIV-1
showed cytotoxicity in concentrations up to 2 mM. 3, 12 and 13
proved to inhibit the HIV-1 replication efficiently in a dose-
dependent manner. In addition to the results of the enzymatic
assay, these results support the hypothesis of a different
binding mode of these rigid derivatives. The chemical stability
was analysed using 1 as an example and showed successive
cleavage of the imine bonds under basic conditions, whereas
acidic conditions did not affect the stability of 1. This is
consistent with the observation that the aminoguanidine
fragment 15 was found in the spermidine pocket of the DHS
under basic conditions. Experiments to co-crystallise the
CNI-1493 derivatives reported herein with DHS and tests
against cART resistant virus strains are currently ongoing.
Experimental
Chemistry
General
All air- or water-sensitive reactions were performed in flame-
dried glassware under a nitrogen atmosphere. Commercially
available solvents and reagents were used without further
purification with the following exceptions: dichloromethane
(CH2Cl2) was distilled from calcium hydride and stored over
activated molecular sieve. Tetrahydrofurane (THF) was dried
over potassium/benzophenone, distilled under nitrogen and
stored over activated molecular sieve. Petroleum ether 50–70,
EtOAc, CH2Cl2 and CH3OH employed in chromatography were
distilled before use. For column chromatography, silica gel 60,
230–400 mesh was used. Pre-coated aluminium plates 60 F254
with 0.2 mm layer of silica gel containing a fluorescence
indicator were used for thin layer chromatography (TLC). NMR
spectra were recorded on 400 or 500MHz spectrometers
(Bruker AMX 400, Bruker AV 400 or Bruker DRX 500). All 1H
and 13C NMR chemical shifts (d) are quoted in parts per million
(ppm) and were calibrated on solvent signals. High resolution
mass spectra were obtained with a VG Analytical VG/70-250F
spectrometer (FAB, matrix was m-nitrobenzyl alcohol). HR-ESI
mass spectra were obtained with an Agilent Technologies ESI-
TOF 6224spectrometer.IRspectra wereacquiredusinga Bruker
Alpha-P IR spectrometer within 400–4000 cmꢀ1 working in the
attenuated total reflection (ATR) mode. Analytical HPLC was
carried out on a VWR-Hitachi LaChromElite HPLC system
consisting of a VWR-Hitachi L-2130 pump, L-2200 autosampler
Conclusion
For closer insights into the structural factors influencing the
antiviral activity and the binding mode of CNI-1493, several
analogues of 1 were synthesised and analysed regarding their
inhibitory potency against DHS in vitro. It was shown that the
truncated derivative 14 and the aminoguanidine fragment 15
were found to be inactive. Most of the other analogues,
containing different linkers between the head groups,
showed dose-dependent activity against DHS. It can be
concluded that two head groups linked at least by an
oxaloyl-moiety are necessary for an inhibitory effect of these
CNI-1493 derivatives. Interestingly, the rigid derivatives 12
and 13 were also found to inhibit the DHS even though they
ß 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.archpharm.com
98

RCHH HARM
A P
Arch. Pharm. Chem. Life Sci. 2016, 349, 91–103
HIV Inhibition by Blocking a Cellular Factor
Archiv der Pharmazie
and a diode array detector L-2455. The column used was a
Nucleodur1 C18 Gravity (EC, 125/4.6, 5 mm, Macherey–Nagel).
Elution was performed using a 100mM ammonium formate
buffer(pH3.0)/acetonitrile(Sigma–Aldrich,HPLCgrade)eluent:
5–44% CH3CN (0–12 min), flow rate 1.3 mL/min and UV
detection at 264nm. The purity of tested CNI-1493 derivatives
1–14 was checked by means of HPLC and was in all cases ꢁ95%.
LC-MS analysis was carried out with the same solvents noted
above using an Agilent Technologies Extend-C18 (1.8mm,
2.1 mm ꢂ 50 mm) column on an Agilent Technologies 1200
series HPLC equipped with an Agilent Technologies ESI-TOF
6224 spectrometer.
N,N0-Bis[3,5-bis[1(aminoiminomethyl)-hydrazoethyl]-
phenyl]-butane diamide tetrahydrochloride 3
The reaction was performed according to the general
procedure using 22 (148 mg, 0.339 mmol), amino guanidine
hydrochloride (192 mg, 1.73 mmol) and 3 mL ethanol/water
(9:1) yielding a pale grey solid (259 mg, 0.322 mmol, 95%). 1H
NMR (400 MHz, DMSO-d6): d ¼ 11.28 (brs, 4H, guanidino-NH),
10.31 (s, 2H, amide-NH), 8.13 (s, 4H, H-2/-6), 8.04 (s, 2H, H-4),
7.82 (brs, 12H, guanidino-N2H3), 2.72 (s, 4H, butyl-H), 2.38 (s,
12H, CH3) ppm. 13C NMR (101 MHz, DMSO-d6): d ¼ 170.1
(amide-C), 156.1 (guanidino-C), 151.8 (imino-C), 139.5 (C-1),
137.7 (C-3/-5), 118.8 (C-4), 118.7 (C-2/-6), 56.0 (butyl-C), 15.2
ꢃ
(CH3) ppm. IR (neat) n ¼ 3236, 3075, 1667, 1590 and 1355 cmꢀ1
.
General procedure for the preparation of the
aminoguanidine compounds 1–14
HRMS (FAB) m/z ¼ calcd. 661.3659 [MþH]þ, found 661.3660
[MþH]þ.
The corresponding diamide 20–33 and 5 eq. aminoguanidine
hydrochloride were dissolved in 3–5 mL ethanol/water (9:1)
and 2 drops of conc. hydrochloric acid were added. After that,
the reaction mixture was refluxed for 3 h. The pure product
precipitated as its hydrochloride after storing the reaction
mixture at ꢀ26°C. The precipitated product was collected by
filtration, washed with 10 mL ꢀ26°C cold ethanol and 2 mL
cold water and dried in vacuo. (See also the Supporting
Information for the InChI codes.)
N,N0-Bis[3,5-bis[1(aminoiminomethyl)-hydrazoethyl]-
phenyl]-hexane diamide tetrahydrochloride 4
The reaction was performed according to the general
procedure using 23 (137 mg, 0.294 mmol), amino guanidine
hydrochloride (151 mg, 1.36 mmol) and 3 mL ethanol/water
(9:1) yielding a colourless solid (121 mg, 0.145 mmol, 49%). 1H
NMR (400 MHz, DMSO-d6): d ¼ 11.33 (brs, 4H, guanidino-NH),
10.30 (s, 2H, amide-NH), 8.17 (s, 4H, H-2/-6), 8.05 (s, 2H, H-4),
7.85 (brs, 12H, guanidino-N2H3), 2.41 (s, 4H, H-a), 2.39 (s, 12H,
CH3), 1.67 (s, 4H, H-b) ppm. 13C NMR (101 MHz, DMSO-d6):
d ¼ 171.3 (amide-C), 156.0 (guanidino-C), 151.6 (imino-C),
N,N0-Bis[3,5-bis[1(aminoiminomethyl)-hydrazoethyl]-
phenyl]-decane diamide tetrahydrochloride 1
The reaction was performed according to the general
procedure using 25 (533 mg, 1.02 mmol), amino guanidine
hydrochloride (579 mg, 5.24 mmol) and 5 mL ethanol/water
(9:1) yielding a colourless solid (696 mg, 0.781 mmol, 76%). 1H
NMR (400 MHz, DMSO-d6): d ¼ 11.23 (brs, 4H, guanidino-NH),
10.17 (s, 2H, amide-NH), 8.12 (d, J ¼ 1.6 Hz, 4H, H-2/-6), 8.04 (t,
J ¼ 1.6 Hz, 2H, H-4), 7.81 (brs, 12H, guanidino-N2H3), 2.37 (s,
12H, CH3), 2.33 (t, J ¼ 7.4 Hz, 4H, H-a), 1.62–1.59 (m, 4H, H-b),
1.33–1.30 (m, 8H, H-c/-d) ppm. 13C NMR (101 MHz, DMSO-d6):
d ¼ 171.6 (amide-C), 156.0 (guanidino-C), 151.8 (imino-C),
139.4 (C-1), 137.5 (C-3/-5), 119.4 (C-4), 118.8 (C-2/-6), 36.3 (C-a),
ꢃ
24.6 (C-b), 15.1 (CH3) ppm. IR (neat) n ¼ 3107, 1666, 1593 and
1353 cmꢀ1. HRMS (FAB) m/z ¼ calcd. 689.3973 [MþH]þ, found
689.3985 [MþH]þ.
N,N0-Bis[3,5-bis[1(aminoiminomethyl)-hydrazoethyl]-
phenyl]-octane diamide tetrahydrochloride 5
The reaction was performed according to the general
procedure using 24 (201 mg, 0.408 mmol), amino guanidine
hydrochloride (221 mg, 2.00 mmol) and 3 mL ethanol/water
(9:1) yielding a colourless solid (203 mg, 0.238 mmol, 58%). 1H
NMR (400 MHz, DMSO-d6): d ¼ 11.32 (brs, 4H, guanidino-NH),
10.22 (s, 2H, amide-NH), 8.15 (d, J ¼ 1.4 Hz, 4H, H-2/-6), 8.04 (s,
2H, H-4), 7.85 (brs, 12H, guanidino-N2H3), 2.38 (s, 12H, CH3),
2.36–2.28 (m, 4H, H-a), 1.70–1.58 (m, 4H, H-b), 1.35 (s, 4H, H-c)
ppm. 13C NMR (101 MHz, DMSO-d6): d ¼ 171.4 (amide-C), 156.0
(guanidino-C), 151.6 (imino-C), 139.4 (C-1), 137.4 (C-3/-5),
139.4 (C-1), 137.4 (C-3/-5), 118.9 (C-2/-4/-6), 36.4 (C-a), 28.7 (C-
ꢃ
c/-d), 25.0 (C-b), 15.2 (CH3) ppm. IR (neat) n ¼ 2987, 2901, 1451,
1075 and 892 cmꢀ1
.
HRMS (FAB) m/z ¼ calcd. 745.4598
[MþH]þ, found 745.4614 [MþH]þ.
N,N0-Bis[3,5-bis[1(aminoiminomethyl)-hydrazoethyl]-
phenyl]ethyl diamide tetrahydrochloride 2
The reaction was performed according to the general
procedure using 21 (91 mg, 0.22 mmol), amino guanidine
hydrochloride (125 mg, 1.13 mmol) and 5 mL ethanol/water
(9:1) yielding a colourless solid (94 mg, 0.12 mmol, 55%). 1H
NMR (500 MHz, DMSO-d6): d ¼ 11.31 (brs, 4H, guanidino-NH),
10.81 (s, 2H, amide-NH), 8.43 (d, J ¼ 1.5 Hz, 4H, H-2/-6), 8.17 (s,
2H, H-4), 7.82 (brs, 12H, guanidino-N2H3), 2.41 (s, 12H, CH3)
ppm. 13C NMR (125 MHz, DMSO-d6): d ¼ 158.3 (amide-C), 156.1
119.4 (C-4), 118.7 (C-2/-6), 36.2 (C-a), 28.3 (C-c), 24.7 (C-b), 15.1
ꢃ
(CH3) ppm. IR (neat) n ¼ 3338, 3167, 1678, 1587 and 1355 cmꢀ1
.
HRMS (FAB) m/z ¼ calcd. 717.4286 [MþH]þ, found 717.4304
[MþH]þ.
N,N0-Bis[3,5-bis[1(aminoiminomethyl)-hydrazoethyl]-
phenyl]dodecane diamide tetrahydrochloride 6
The reaction was performed according to the general
procedure using 26 (158 mg, 0.289 mmol), amino guanidine
hydrochloride (175 mg, 1.59 mmol) and 3 mL ethanol/water
(9:1) yielding a colourless solid (181 mg, 0.197 mmol, 68%). 1H
NMR (400 MHz, DMSO-d6): d ¼ 11.24 (brs, 4H, guanidino-NH),
(guanidino-C), 151.3 (imino-C), 137.8 (C-3/-5), 137.7 (C-1),
ꢃ
120.9 (C-4), 119.8 (C-2/-6), 14.7 (CH3) ppm. IR (neat) n ¼ 3305,
3151, 1669, 1588 and 1348 cmꢀ1. HRMS (ESIþ) m/z ¼ calcd.
633.3341 [MþH]þ, found 633.3347 [MþH]þ.
ß 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.archpharm.com
99

RCHH HARM
A P
Arch. Pharm. Chem. Life Sci. 2016, 349, 91–103
M. Schroder et al.
€
Archiv der Pharmazie
10.16 (s, 2H, amide-NH), 8.12 (d, J ¼ 1.5 Hz, 4H, H-2/-6), 8.04 (t,
J ¼ 1.5 Hz, 2H, H-4), 7.81 (brs, 12H, guanidino-N2H3), 2.37 (s,
12H, CH3), 2.33 (t, J ¼ 7.5 Hz, 4H, H-a), 1.62–1.57 (m, 4H, H-b),
1.28 (s, 12H, H-c/-d/-e) ppm. 13C NMR (101 MHz, DMSO-d6):
d ¼ 171.5 (amide-C), 156.0 (guanidino-C), 151.6 (imino-C),
139.4 (C-1), 137.5 (C-3/-5), 118.8 (C-4), 118.7 (C-2/-6), 36.2 (C-a),
(s, 12H, CH3) ppm. 13C NMR (101 MHz, DMSO-d6): d ¼ 162.6
(amide-C), 156.1 (guanidino-C), 151.7 (imino-C), 139.0 (C-1),
137.8 (C-3/-5), 134.1 (C-a), 120.2 (C-4), 119.2 (C-2/-6), 15.1 (CH3)
ꢃ
ppm. IR (neat) n ¼ 2987, 2971, 2900, 1394, 1250 and
1065 cmꢀ1. HRMS (FAB) m/z ¼ calcd. 659.3503 [MþH]þ, found
659.3505 [MþH]þ.
28.8, 28.7, 28.6 (C-c/-d/-e), 24.9 (C-b), 15.1 (CH3) ppm. IR (neat)
ꢃ
n ¼ 3130, 1659, 1587 and 1355 cmꢀ1. HRMS (FAB) m/z ¼ calcd.
N,N0-Bis[3,5-bis[1(aminoiminomethyl)-hydrazoethyl]-
phenyl]-1,4-phenylacetyl diamide tetrahydrochloride 10
The reaction was performed according to the general
procedure using 30 (120 mg, 0.230 mmol), amino guanidine
hydrochloride (129 mg, 1.20 mmol) and 5 mL ethanol/water
(9:1) yielding a colourless solid (171 mg, 0.194 mmol, 84%).
1H NMR (400 MHz, DMSO-d6): d ¼ 11.20 (brs, 4H, guanidino-
NH), 10.53 (s, 2H, amide-NH), 8.13 (s, 2H, H-4), 8.04 (s, 2H, H-
2/-6), 7.79 (brs, 12H, guanidino-N2H3), 7.32 (s, 4H, H-20), 3.66
(s, 4H, CH2), 2.36 (s, 12H, CH3) ppm. 13C NMR (101 MHz,
DMSO-d6): d ¼ 169.6 (amide-C), 156.0 (guanidino-C), 151.8
(imino-C), 139.4 (C-1), 137.6 (C-3/-5), 134.0 (C-10), 129.1 (C-20),
773.4912 [MþH]þ, found 773.4938 [MþH]þ.
N,N0-Bis[3,5-bis[1(aminoiminomethyl)-hydrazoethyl]-
phenyl]tetradecane diamide tetrahydrochloride 7
The reaction was performed according to the general
procedure using 27 (136 mg, 0.236 mmol), amino guanidine
hydrochloride (130 mg, 1.20 mmol) and 5 mL ethanol/water
(9:1) yielding a colourless solid (80 mg, 84 mmol, 36%). 1H NMR
(400 MHz, DMSO-d6): d ¼ 11.24 (brs, 4H, guanidino-NH), 10.15
(s, 2H, amide-NH), 8.12 (d, J ¼ 1.2 Hz, 4H, H-2/-6), 8.04 (t,
J ¼ 1.4 Hz, 2H, H-4), 7.81 (brs, 12H, guanidino-N2H3), 2.37 (s,
12H, CH3), 2.33 (t, J ¼ 7.6 Hz, 4H, H-a), 1.59 (q, J ¼ 6.7 Hz, 4H, H-
b), 1.32–1.23 (m, 16H, H-c/-d/-e/-f) ppm. 13C NMR (101 MHz,
DMSO-d6): d ¼ 171.6 (amide-C), 156.0 (guanidino-C), 151.8
(imino-C), 139.5 (C-1), 137.6 (C-3/-5), 118.9 (C-4), 118.4 (C-2/-6),
118.9 (C-2/-6), 118.6 (C-4), 42.6 (CH2), 15.1 (CH3) ppm. IR
ꢃ
(neat) n ¼ 2987, 2901, 1588, 1406, 1229, 1066 and 899 cmꢀ1
.
MS (ESIþ) m/z ¼ calcd. 737.3967 [MþH]þ, found 737.3956
[MþH]þ.
36.4 (C-a), 29.0, 28.9, 28.7 (C-c/-d/-e/-f), 25.0 (C-b), 15.2 (CH3)
ꢃ
ppm. IR (neat) n ¼ 3139, 1921, 1670, 1586, 1443, 1273 and
N,N0-Bis[3,5-bis[1(aminoiminomethyl)-hydrazoethyl]-
phenyl]-1,3-phenylacetyl diamide tetrahydrochloride 11
The reaction was performed according to the general
procedure using 31 (120 mg, 0.230 mmol), amino guanidine
hydrochloride (130 mg, 1.17 mmol) and 5 mL ethanol/water
(9:1) yielding a colourless solid (81 mg, 92 mmol, 40%). 1H NMR
(400 MHz, DMSO-d6): d ¼ 11.21 (brs, 4H, guanidino-NH), 10.55
(s, 2H, amide-NH), 8.15 (s, 2H, H-4), 8.05 (s, 2H, H-2/-6), 7.80
(brs, 12H, guanidino-N2H3), 7.33 (s, 1H, H-2’), 7.31–7.25 (m, 3H,
H-40/-50/-60), 3.69 (s, 4H, CH2), 2.36 (s, 12H, CH3) ppm. 13C NMR
(101 MHz, DMSO-d6): d ¼ 169.4 (amide-C), 156.0 (guanidino-
C), 151.8 (imino-C), 139.4 (C-1), 137.7 (C-3/-5), 135.7 (C-10/-30),
1123 cmꢀ1. MS (FAB) m/z ¼ calcd. 801.5 [MþH]þ, found 801.5
[MþH]þ.
N,N0-Bis[3,5-bis[1(aminoiminomethyl)-hydrazoethyl]-
phenyl] (cis/trans)-1,4-cyclohexyl diamide
tetrahydrochloride 8
The reaction was performed according to the general
procedure using 28 (200 mg, 0.410 mmol), amino guanidine
hydrochloride (225 mg, 2.04 mmol) and 5 mL ethanol/water
(9:1) yielding a pale brown solid (70 mg, 90 mmol, 24%). 1H
NMR (400 MHz, DMSO-d6): d ¼ 11.17 (brs, 4H, guanidino-NH),
10.06 (s, 2H, amide-NH), 8.16–8.14 (m, 4H, H-2/-6), 8.05 (brs,
2H, H-4), 7.78 (brs, 12H, guanidino-N2H3), 2.62–2.55 (m, 2H,
cyclohexyl-H-a), 2.37 (s, 12H, CH3), 2.01–1.95 (m, 4H, cyclo-
hexyl-H-b), 1.68–1.50 (m, 4H, cyclohexyl-H-b) ppm. 13C NMR
(101 MHz, DMSO-d6): d ¼ 174.2 (amide-C), 156.0 (guanidino-
C), 151.8 (imino-C), 139.7 (C-1), 137.6 (C-3/-5), 118.9 (C-4),
130.1 (C-20), 128.3, 127.6 (C-40/-50/-60), 119.3 (C-2/-6), 118.6 (C-
ꢃ
4), 43.2 (CH2), 15.2 (CH3) ppm. IR (neat) n ¼ 3120, 1667, 1584,
1440, 1354, 1127 and 874 cmꢀ1. HRMS (FAB) m/z ¼ calcd.
737.3973 [MþH]þ, found 737.4009 [MþH]þ.
N,N0-Bis[3,5-bis[1(aminoiminomethyl)-hydrazoethyl]-
phenyl]-1,4-benzene diamide tetrahydrochloride 12
The reaction was performed according to the general
procedure using 32 (226 mg, 0.447 mmol), amino guanidine
hydrochloride (263 mg, 2.38 mmol) and 5 mL ethanol/water
(9:1) yielding a colourless solid (198 mg, 0.279 mmol, 60%). 1H
NMR (400 MHz, DMSO-d6): d ¼ 11.35 (brs, 4H, guanidino-NH),
10.71 (s, 2H, amide-NH), 8.42 (s, 4H, H-2/-6), 8.24 (s, 4H, H-b),
8.13 (s, 2H, H-4), 7.87 (brs, 12H, guanidino-N2H3), 2.43 (s, 12H,
CH3) ppm. 13C NMR (101 MHz, DMSO-d6): d ¼ 164.7 (amide-C),
156.1 (guanidino-C), 151.6 (imino-C), 139.1 (C-1), 137.6 (C-3/-
118.7 (C-2/-6), 41.4 (cyclohexyl-C-a), 26.5 (cyclohexyl-C-b), 15.2
ꢃ
(CH3) ppm. IR (neat) n ¼ 3138, 1667, 1589, 1443, 1138 and
862 cmꢀ1. HRMS (FAB) m/z ¼ calcd. 715.4129 [MþH]þ, found
715.4143 [MþH]þ.
N,N0-Bis[3,5-bis[1(aminoiminomethyl)-hydrazoethyl]-
phenyl]-2-butene diamide tetrahydrochloride 9
The reaction was performed according to the general
procedure using 29 (150 mg, 0.350 mmol), amino guanidine
hydrochloride (191 mg, 1.70 mmol) and 5 mL ethanol/water
(9:1) yielding a pale grey solid (156 mg, 0.194 mmol, 55%).
1H NMR (400 MHz, DMSO-d6): d ¼ 11.26 (brs, 4H, guanidino-
NH), 10.93 (s, 2H, amide-NH), 8.23 (s, 4H, H-2/-6), 8.11 (s, 2H,
H-4), 7.82 (brs, 12H, guanidino-N2H3), 7.26 (s, 2H, H-a), 2.41
5), 137.1 (C-a), 127.9 (C-b), 120.7 (C-2/-6), 120.5 (C-4), 15.2
ꢃ
(CH3) ppm. IR (neat) n ¼ 3340, 3150, 1668 and 1357 cmꢀ1
.
HRMS (FAB) m/z ¼ calcd. 709.3659 [MþH]þ, found 709.3671
[MþH]þ.
ß 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.archpharm.com
100

RCHH HARM
A P
Arch. Pharm. Chem. Life Sci. 2016, 349, 91–103
HIV Inhibition by Blocking a Cellular Factor
Archiv der Pharmazie
N,N0-Bis[3,5-bis[1(aminoiminomethyl)-hydrazoethyl]-
phenyl]-1,3-benzene diamide tetrahydrochloride 13
The reaction was performed according to the general
procedure using 33 (210 mg, 0.412 mmol), amino guanidine
hydrochloride (228 mg, 2.06 mmol) and 5 mL ethanol/water
(9:1) yielding a colourless solid (231 mg, 0.270 mmol, 66%). 1H
NMR (400 MHz, DMSO-d6): d ¼ 11.31 (brs, 4H, guanidino-NH),
10.91 (s, 2H, amide-NH), 9.05 (s, 1H, H-20), 8.59 (s, 4H, H-2/-6),
8.25 (d, J ¼ 7.7 Hz, 2H, H-40/-60), 8.13 (s, 2H, H-4), 7.86 (brs, 12H,
guanidino-N2H3), 7.74 (t, J ¼ 7.7 Hz, 1H, H-50), 2.44 (s, 12H, CH3)
ppm. 13C NMR (101 MHz, DMSO-d6): d ¼ 164.7 (amide-C), 156.1
(guanidino-C), 151.7 (imino-C), 139.4 (C-1), 137.6 (C-3/-5),
soluble fraction was separated by centrifugation at 12000 rpm
for 15 min. About 1 mL bed volume of buffer-equilibrated Ni-
NTA beads was added to the supernatant and incubated with
gentle rocking in an end-to-end rotor at 4°C for 1 h. The beads
were then packed in a column and washed with wash buffer
(50 mM Tris pH 8, 300 mM NaCl and 30 mM imidazole),
followed by second wash buffer (50 mM Tris pH 8, 300 mM
NaCl and 50 mM imidazole). The protein was eluted in 50 mM
Tris pH 8, 300 mM NaCl and 300 mM imidazole. The purified
His-DHS protein was analysed on SDS-PAGE and was dialysed
against 300 mM glycine NaOH, pH 9 and 10 % glycerol and
stored at ꢀ20°C.
134.3 (C-10/-30), 130.9 (C-40/-60), 128.8 (C-50), 126.0 (C-20), 120.3
ꢃ
(C-4), 119.7 (C-2/-6), 15.1 (CH3) ppm. IR (neat) n ¼ 2987, 2900,
eIF-5A expression and purification
1406, 1250, 1066 and 892 cmꢀ1. HRMS (FAB) m/z ¼ calcd.
709.3660 [MþH]þ, found 709.3680 [MþH]þ.
pGEX-eIF-5A clone was used to prepare eIF-5A protein as
described in [19], with minor modifications. GST-eIF-5A
overexpression was performed as described for DHS. IPTG-
induced cell pellet was resuspended in 40 mL lysis buffer
(10 mM PBS pH 7.4, 1 mM DTT, 5 mM EDTA, 2.6 mM MnCl2,
26 mM MgCl2) with 0.5 mg/mL DNAse and protease inhibitors
(0.1 mM PMSF, 2 mg/mL leupeptin, 2 mg/mL aprotinin) and
lysed by sonication. To the total lysate, 1% Triton X-100 was
added (Triton X-100 to be diluted in 1ꢂ PBS and then used)
and incubated at 4°C for 1 h on an end-to-end rotor. The
soluble fraction was separated by centrifuging at 30000g for
15 min. To the supernatant, 1 mL bed volume of buffer-
equilibrated GSH beads were added and incubated with
gentle rocking in an end-to-end rotor at 4°C for 1–2 h. The
beads are then washed with 100 mL of wash buffer (50 mL PBS
pH 7.4, 1 mM DTT and 5 mM EDTA) and with 150 mL of second
buffer (50 mM Tris pH 8, 200 mM NaCl and 5 mM EDTA). The
protein was eluted in 50 mM Tris-HCl, 200 mM NaCl and
20 mM reduced glutathione (pH 8.0). The purified protein
fractions were dialysed against the factor Xa cleavage buffer,
50 mM Tris-HCl, 100 mM NaCl and 1 mM CaCl2. Factor Xa
cleavage and final purification of eIF-5A were performed
exactly as described in [19]. eIF-5A was dialysed in assay buffer
(300 mM glycine/NaOH, pH 9) and stored at ꢀ20°C.
DHS activity assay
The deoxyhypusine synthase reaction was conducted in
principle as described previously [29]. Briefly, the reaction
mixture contained 5 mg eIF-5A precursor protein, 3 mg DHS
and 2 mCi [3H]-spermidine (32.4 Ci/mmol) in a total volume of
200 mL reaction buffer (300 mM glycine-NaOH, pH 9.0, 1 mM
NADþ, 1 mM DTT, 50 mg/mL BSA). In order to test a possible
influence of the new inhibitors on DHS activity, the enzymatic
reaction was performed in the presence of the indicated
concentrations of the indicated CNI-1493 derivatives, or with
CNI-1493 as a positive control or without compounds as a
solvent or negative control, respectively. After 1 h at 37°C, the
reaction was stopped by adding 200 mL of 20 mM spermidine
in PBS and transferred onto
a Millipore GSWPO2500
nitrocellulose membrane, which was previously blocked with
20 mM spermidine/PBS for 1 h. The reaction mixture was
vacuum filtered followed by washing the membrane with
5 mL PBS. Finally, the membranes were air dried and the
tritium signal was measured in a liquid scintillation counter.
Overexpression and purification of DHS and
eIF-5A
HIV-1 infection experiments
The Escherichia coli strain BL21 (DE3) (Novagen) was
employed for protein expression.
PM1-cells were preincubated for 8–10 days and splitted
according to culture density in presence of the inhibitor
candidates or DMSO as solvent control. Subsequently, the cells
were infected with the macrophage-tropic (R5) HIV-1 strain
BaL [30]. For infection, roughly 3 ꢂ 106 cells were resuspended
in 500 mL culture medium without drugs and incubated at 37°
C for 3 h with 3 ng of HIV-1 viral stock. After infection, cells
were washed twice with PBS and further cultivated in
presence of the inhibitor candidates or DMSO. At day 3 post
infection, the culture medium was replaced and cells were
split. At day 7 post infection, p24-levels in the supernatant
were determined by enzyme-linked immunosorbent assay
(Innogenetics NV). Cellular viability was tested in parallel
utilizing AlamarBlue (AbD Serotec). At least six independent
infections were carried out for each tested concentration.
PM1 cells were maintained in RPMI (Invitrogen) containing
His-DHS purification
pTricHis-DHS clone was used for the preparation for
recombinant DHS [19]. A 1 L culture of LB-ampicillin medium
was inoculated with a 50 mL overnight culture of E. coli BL21
(DE3) transformed with pTricHis-DHS (preinoculum) and
grown at 37°C until the optical density (OD) value at
600 nm reached 0.6–0.8 (ꢃ3 h). The Tre recombinase gene
expression was induced by addition of 0.5 mM of IPTG, and
was further grown for a period of 4 h at 37°C. Post incubation
the bacteria were harvested by centrifugation (10 min, 5000g,
4°C) and stored at ꢀ80°C until further use. The cell pellet was
resuspended in 30 mL lysis buffer (50 mM Tris pH 8, 300 mM
NaCl and 30 mM imidazole) and lysed by sonication. The
ß 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.archpharm.com
101

RCHH HARM
A P
Arch. Pharm. Chem. Life Sci. 2016, 349, 91–103
M. Schroder et al.
€
Archiv der Pharmazie
10% FCS (Biochrom AG), 100 units/mL penicillin and 100 mg/
mL streptomycin (Invitrogen). PM1 cells and the HIV-1 isolate
BaL were obtained from the NIH AIDS Research and Reference
Reagent Program.
Infection’. We also thank Silke Kummer for excellent technical
assistance and the other research groups collaborating in this
€
€
project: Prof. Dr. T. Brummendorf (Universitatsklinikum
€
€
Aachen) and Dr. S. Balabanov (UniversitatsSpital Zurich).
MD simulations
The authors have declared no conflict of interest.
Three-dimensional coordinates of the DHS tetramer in
complex with two CNI-1493 molecules were taken from the
unpublished crystal structure [23]. The protein was placed in a
box with an edge length of 7.90 nm ꢂ 9.45 nm ꢂ 11.97 nm.
Protonation states of all amino acid residues were adjusted
according to the pKa values in a medium of pH 7.4. Water
molecules were represented using the SPC model [31]. In
order to ensure neutrality of the system and to mimic a
physiological NaCl solution, 77 Naþ and 49 Clꢀ ions were
added to the systems.
References
[1] M. H. Park, J. Biochem. 2006, 139, 161–169.
[2] E. C. Wolff, K. R. Kang, Y. S. Kim, M. H. Park, Amino Acids
2007, 33, 341–350.
[3] W. M. Kemper, K. W. Berry, W. C. Merrick, J. Biol. Chem.
1976, 251, 5551–5557.
[4] R. Benne, M. L. Brown-Luedi, J. W. Hershey, J. Biol. Chem.
1978, 253, 3070–3077.
[5] P. Saini, D. E. Eyler, R. Green, T. E. Dever, Nature 2009,
459, 118–121.
[6] A. P. Gregio, V. P. Cano, J. S. Avaca, S. R. Valentini, C. F.
Zanelli, Biochem. Biophys. Res. Commun. 2009, 380,
785–790.
[7] C. H. Li, T. Ohn, P. Ivanov, S. Tisdale, P. Anderson, PLoS
ONE 2010, 5, 9942.
[8] R. D. Robbins, S. A. Tersey, T. Ogihara, D. Gupta, T. B.
Farb, J. Ficorilli, K. Bokvist, B. Maier, R. G. Mirmira, J. Biol.
Chem. 2010, 285, 39943–39952.
[9] M. Kruse, O. Rosorius, F. Kratzer, D. Bevec, C. Kuhnt, A.
Steinkasserer, G. Schuler, J. Hauber, J. Exp. Med. 2000,
191, 1581–1590.
[10] J. Chemnitz, N. Turza, I. Hauber, A. Steinkasserer, J.
Hauber, Immunobiology 2010, 215, 370–379.
[11] B.Maier,T.Ogihara,A.P.Trace,S.A.Tersey,R.D.Robbins,S.
K.Chakrabarti,C.S.Nunemaker,N.D.Stull,C.A.Taylor,J.E.
Thompson, R. S. Dondero, E. C. Lewis, C. A. Dinarello, J. L.
Nadler, R. G. Mirmira, J. Clin. Invest. 2010, 120, 2156–2170.
[12] G. Lipowsky, F. R. Bischoff, P. Schwarzmaier, R. Kraft, S.
Kostka, E. Hartmann, U. Kutay, D. Gorlich, EMBO J. 2000,
19, 4362–4371.
[13] D. Bevec, H. Jaksche, M. Oft, T. Woehl, M. Himmelspach,
A. Pacher, M. Schebesta, K. Koettnitz, M. Dobrovnik,
Science 1996, 271, 1858–1860.
Full-length CNI-1493, its truncated version and NAD were
parameterised manually according to the GROMOS-96 topol-
ogy and tested for reliability separately in a small water box
(data not shown). Atomic point charges were calculated using
the MMFF94x force field [30] within MOE [27]. All simulations
were performed using GROMACS version 3.3.3 (parameter set
43a1)[32]. Theresultingsystemswereminimisedina three-step
procedure. Initially, positions of hydrogen atoms were opti-
mised using 100 steps conjugate gradients with fixed heavy
atoms. Subsequently, the whole system was minimised using
twominimizationprocedures with1000stepssteepestdescents
and conjugate gradients, respectively [33]. In order to minimise
edge effects periodic boundary conditions were employed. The
particle-mesh Ewald method was applied to accurately deter-
mine the long-range electrostatic interactions [34, 35]. VdW
̊
interactions were considered by applying a cutoff of 14A. The
timestep for the simulations was set to 2 fs. The systems were
gradually heated from 0 to 298 K using a simulated annealing
protocol followed by an equilibration period of 5 ns. To keep
the systems at constant temperature, a Berendsen thermostat
was applied using a coupling time of 0.1 ps [36]. Constant
pressure was maintained by coupling to an external bath with a
reference value of 105 Pa, a coupling time of 1.0 ps and an
isothermal compressibility of 4.5 ꢂ 10ꢀ10 Paꢀ1. Bonds between
heavy atoms and corresponding hydrogen atoms were con-
strained to their equilibrium bond lengths using the SHAKE
algorithm [37]. Subsequent production runs of 50ns length
were performed using a Nose Hoover heat bath (coupling time
€
€
[14] W. Hofmann, B. Reichart, A. Ewald, E. Muller, I. Schmitt, R.
H. Stauber, F. Lottspeich, B. M. Jockusch, U. Scheer, J.
Hauber,M.-C. Dabauvallea,J.CellBiol. 2001, 152, 895–910.
[15] I. Hauber, D. Bevec, J. Heukeshoven, F. Kraetzer, F. Horn,
A. Choidas, T. Harrer, J. Hauber, J. Clin. Invest. 2005, 115,
76–85.
[16] R. A. Hart, J. N. Billaud, S. J. Choi, T. R. Phillips, Virology
2002, 304, 97–104.
[17] A. Kaiser, Amino Acid 2010, 42, 679–684.
[18] S. V. Sitaraman, M. Hoteit, A. T. Gewirtz, Curr. Opin.
Investig. Drug 2003, 4, 1363–1368.
0.5 ps) and Parrinello–Rahman pressure coupling with
a
coupling time of 5.0 ps [38–41]. For each system three
independent simulations were performed. Representative
frames were obtained by clustering the trajectories over
backbone atoms [42]. VdW contacts are defined as heavy atom
̊
distances ꢄ4.0 A. Hydrogen bonds are defined by a donor–
̊
acceptor cutoff distance of ꢄ3.5 A and a cutoff-angle of ꢄ30°.
Figures were prepared using VMD and POV-ray [43, 44].
The authors are grateful to the Federal Ministry of Education
and Research (BMBF) for financial support under grant 01 GU
0715–0718 ‘Combating drug resistance in CML and HIV-1
[19] M. N. Sommer, D. Bevec, B. Klebl, B. Flicke, K. Hoelscher,
T. Freudenreich, I. Hauber, J. Hauber, H. Mett, J. Biomol.
Screen. 2004, 9, 434–438.
ß 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.archpharm.com
102

RCHH HARM
A P
Arch. Pharm. Chem. Life Sci. 2016, 349, 91–103
HIV Inhibition by Blocking a Cellular Factor
Archiv der Pharmazie
[20] P. Ziegler, T. Chahoud, T. Wilhelm, N. Paellman, M. Braig,
V. Wiehle, S. Ziegler, M. Schroeder, C. Meier, A. Kolodzik,
M. Rarey, J. Panse, J. Hauber, S. Balabanov, T. H.
Bruemmendorf, Invest. New Drug 2012, 30, 2274–2283.
[21] M. Woriedh, I. Hauber, A. L. Martinez-Rocha, C. Voigt, F.
J. Maier, M. Schroeder, C. Meier, J. Hauber, W. Schaefer,
Mol. Plant Microbe Interact. 2011, 24, 619–627.
[22] H. Wang, M. Zhang, M. Bianchi, B. Sherry, A. Sama, K. J.
Tracey, Proc. Natl. Acad. Sci. USA 1998, 95, 14429–14434.
[23] R. Hilgenfeld, 2014, unpublished.
Reidel Publishing Company, Dordrecht 1981, pp.
331–342.
[32] T. A. Halgren, J. Comput. Chem. 1996, 17, 490–519.
[33] D. van der Spoel, E. Lindahl, B. Hess, G. Groenhof, A. E.
Mark, H. J. Berendsen, J. Comput. Chem. 2005, 16,
1701–1718.
€
[34] W. R. P. Scott, Hunenberger, I. G. Tironi, A. E. Mark, S. R.
€
Billeter, J. Fennen, A. E. Torda, T. Huber, P. Kruger, W. F.
van Gunsteren, J. Phys. Chem. A 1999, 103, 3596–3607.
[35] T. A. Darden, D. York, L. Pedersen, J. Chem. Phys. 1993,
98, 10089–10092.
[36] U. Essmann, L. Perera, M. L. Berkowitz, T. Darden, H. Lee,
L. G. Pedersen, J. Comp. Phys. 1995, 103, 8577–8592.
[37] H. J. C. Berendsen, J. P. M. Postma, A. Dinola, J. R. Haak, J.
Chem. Phys. 1984, 81, 3684–3690.
[38] J. P. Ryckaert, G. Ciccotti, H. J. C. Berendsen, J. Comp.
Phys. 1977, 23, 327–341.
[39] W. G. Hoover, Phys. Rev. A. 1985, 31, 1695–1697.
[24] D. M. McKinnon, J. Y. Wong, Can. J. Chem. 1971, 49,
2018–2022.
[25] P. Ulrich, A. Crami, J. Med. Chem. 1984, 27, 35–40.
[26] M. Bianchi, P. Ulrich, O. Bloom, M. III Meistrell, G. A.
Zimmerman, H. Schmidtmayerova, M. Bukrinsky, T.
Donnelley, R. Bucala, B. Sherry, K. R. Manogue, A. J.
Tortolani, A. Cerami, K. J. Tracey, Mol. Med. 1995, 1,
254–266.
ꢀ
[27] MOE, v 2010-10, Chemical Computing Group, http://
www.chemcomp.com/software.htm.
[40] A. Nose, Mol. Phys. 1984, 52, 255–268.
[41] M. Parrinello, A. Rahman, J. Appl. Phys. 1981, 52,
7182–7190.
[42] X. Daura, K. Gademann, B. Jaun, D. Seebach, W. F. van
Gunsteren, A. E. Mark, Angew. Chem. Int. Ed. 1999, 38,
236–240.
[43] W. Humphrey, A. Dalke, K. Schulten, J. Mol. Graphics
1996, 14, 33–38.
[44] POVRAY, version 3.6, www.povray.org.
[28] YASARA, version 8-10-13, http://www.yasara.org.
[29] D. Bevec, B. Kappel, H. Jaksche, R. Csonga, J. Hauber, H.
Klier, A. Steinkasserer, FEBS Lett. 1996, 378, 195–198.
[30] S. Gartner, P. Markovits, D. M. Markovitz, M. H. Kaplan,
R. C. Gallo, M. Popovic, Science 1986, 233, 215–219.
[31] H. J. C. Berendsen, J. P. M. Postma, W. F. van Gunsteren, J.
Hermans, in Intermolecular Forces. (Ed.: B. Pullman), D.
ß 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.archpharm.com
103