Synthesis, Chiral Separation, Absolute Configuration Assignment, and Biological Activity of Enantiomers of Retro-1 as Potent Inhibitors of Shiga Toxin

Article abstract of DOI:10.1002/cmdc.201500139

The Shiga toxin (Stx) family is composed of related protein toxins produced by the bacteria Shigella dysenteriae and certain pathogenic strains of E.coli. No effective therapies for Stx intoxication have been developed yet. However, inhibitors that act on the intracellular trafficking of these toxins may provide new options for the development of therapeutic strategies. This study reports the synthesis, chromatographic separation, and pharmacological evaluation of the two enantiomers of Retro-1, a compound active against Stx and other such protein toxins. Retro-1 works by inhibiting retrograde transport of these toxins inside cells. In vitro experiments proved that the configuration of the stereocenter at position 5 is not crucial for the activity of this compound. X-ray diffraction data revealed (S)-Retro-1 to be slightly more active than (R)-Retro-1.

Full text of DOI:10.1002/cmdc.201500139

DOI: 10.1002/cmdc.201500139
Communications
Synthesis, Chiral Separation, Absolute Configuration
Assignment, and Biological Activity of Enantiomers of
Retro-1 as Potent Inhibitors of Shiga Toxin
[
a]
[b]
[b]
[a]
Hajer Abdelkafi, AurØlien Michau, Alexandra Clerget, David-Alexandre Buisson,
[c, d, e]
[b]
[b]
[a]
Ludger Johannes,
Daniel Gillet,* Julien Barbier, and Jean-Christophe Cintrat*
The Shiga toxin (Stx) family is composed of related protein
toxins produced by the bacteria Shigella dysenteriae and cer-
tain pathogenic strains of E. coli. No effective therapies for Stx
intoxication have been developed yet. However, inhibitors that
act on the intracellular trafficking of these toxins may provide
new options for the development of therapeutic strategies.
This study reports the synthesis, chromatographic separation,
and pharmacological evaluation of the two enantiomers of
Retro-1, a compound active against Stx and other such protein
toxins. Retro-1 works by inhibiting retrograde transport of
these toxins inside cells. In vitro experiments proved that the
configuration of the stereocenter at position 5 is not crucial for
the activity of this compound. X-ray diffraction data revealed
ing the severity of the disease symptoms. Hence, the use of
antibiotics for the management of STEC infection is controver-
[7]
[8]
sial or not recommended. Even the promising monoclonal
anti-C5 antibody eculizumab, used successfully to treat three
[
9]
3-year-old patients with neurological HUS complications,
[10]
gave mixed results during the German outbreak.
Shiga
toxins have one moiety (B-subunit) that binds to their respec-
tive cellular receptors, the glycosphingolipid Gb3. Shiga toxins
are then transported in a retrograde manner from the plasma
membrane via endosomes and the trans-Golgi network (TGN)
[11]
to the endoplasmic reticulum, before translocation of the en-
zymatic moiety (A-subunit) into the cytosol. Finally, the Shiga
toxin A subunit inactivates the 28S RNA of the 60S ribosomal
subunit (reviewed in references [11–14]). This is an irreversible
process that results in the inhibition of protein biosynthesis. In-
hibitors that act on the intracellular trafficking of these toxins
likely offer new options for the development of therapeutic
(S)-Retro-1 to be slightly more active than (R)-Retro-1.
The Shiga toxin (Stx) family is composed of related protein
[15]
toxins that are produced by the bacteria Shigella dysenteriae
strategies.
[
1,2]
and certain pathogenic strains of Escherichia coli.
Infections
In the course of a high-throughput screening campaign, we
have identified two compounds, Retro-1 and Retro-2, which
with Shiga-toxin-producing E. coli (STEC) are responsible for
[16]
acute and severe hemorrhagic diarrhea and its life-threatening
protect human cells against Stx.
These compounds were
[
3]
complication, hemolytic uremic syndrome (HUS). In 2011,
shown to act as inhibitors of the retrograde route used by
toxins to enter into cells. Retro-1 also exhibits an enhancement
of pharmacological action of antisense and splice-switching oli-
gonucleotides in vivo, although the mode of action might be
a major outbreak caused by E. coli O104:H4 spread throughout
[
4]
Germany and infected about 4000 people in Europe, causing
[
5]
more than 900 cases of HUS, resulting in 54 deaths. To date,
[
6]
[17]
there is no effective therapy for Stx intoxication. Many antibi-
otics used to treat bacterial infections, including quinolones,
stimulate the induction of Stx-converting prophages, enhanc-
different and remains elusive. Herein we report the synthesis
and evaluation of the two enantiomers of Retro-1 against
Shiga toxin and establish that the biological activity is almost
equally dispatched between the two enantiomers with a slight
preference for the S isomer, a less distinct behavior relative to
[
a] Dr. H. Abdelkafi, D.-A. Buisson, Dr. J.-C. Cintrat
CEA, iBiTec-S/SCBM, CEA-Saclay, LabEx LERMIT, 91191 Gif-sur-Yvette (France)
E-mail: jean-christophe.cintrat@cea.fr
[18]
Retro-2 analogues.
The synthesis of Retro-1 (6) was carried out by starting from
[b] Dr. A. Michau, A. Clerget, Prof. D. Gillet, Dr. J. Barbier
the
commercially
available
2-aminobenzophenone
CEA, iBiTec-S/SIMOPRO, CEA-Saclay, LabEx LERMIT, 91191 Gif-sur-Yvette
(
France)
1 (Scheme 1) in 39% yield over five steps. First, regioselective
bromination was performed with NBS with complete conver-
sion. Acetylation with bromoacetyl bromide was immediately
followed by cyclization with ammonia to yield benzodiazepine
E-mail: daniel.gillet@cea.fr
[
c] Dr. L. Johannes
Institut Curie, PSL Research University
Chemical Biology of Membranes and Therapeutic Delivery unit
4
in 66% yield over two steps. Reduction of the imino moiety
2
6 rue d’Ulm, 75248 Paris Cedex 05 (France)
with sodium cyanoborohydride offered a racemic mixture of
benzodiazepine 5, which was treated with propionyl chloride
to obtain Retro-1 as a 50:50 mixture of two enantiomers. In ad-
dition, rac-Retro-1 was obtained as a 1:1 mixture of conformers
[
d] Dr. L. Johannes
CNRS UMR3666, 75005 Paris (France)
[
e] Dr. L. Johannes
INSERM U1143, 75005 Paris (France)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.201500139: general procedures, characteri-
zation of new compounds, crystal structure determination of compound 6,
chemicals for in vitro experiments, intoxication assays, determination of
EC50 values.
1
13
as detected by H and C NMR spectroscopy. High-tempera-
ture NMR in various solvents allowed us to obtain coalescence
of the two conformers’ signals (Supporting Information).
ChemMedChem 2015, 10, 1153 – 1156
1153
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communications
for separate enantiomers. Amongst the various crys-
tallization conditions that we screened, a mixture of
dichloromethane and methanol (1:1) allowed us to
obtain monocrystals suitable for X-ray diffraction
(
slow evaporation method). These studies revealed
that Retro-1-E1 has the S configuration at the C5 po-
sition of the benzodiazepine core (Figure 1), whereas
Retro-1-E2 is the R isomer.
The biological activity of Retro-1 and each enantio-
mer was then evaluated against Stx (Figure 2). rac-
Retro-1 inhibited Stx action on HeLa cells with an
EC50 value of 6.2Æ0.7 mm (Figure 2A top and 2B)
with an R-value of 62.4Æ9.6 (Figure 2C; see Support-
ing Information for R-value calculation). We were also
interested to determine whether compounds 4 and 5
are active against Stx; both were completely inactive
at 30 mm (data not shown). These results confirm the
importance of the propionyl group for bioactivity,
but do not allow us to clearly assign the impact of
the conformation of the seven-membered ring (or
Scheme 1. Synthesis of Retro-1. Reagents and conditions: a) NBS, CH
RT, 2 h, 99%; b) NaHCO (aq. 2m), CH Cl , 08C, 2 h; c) NH
CN, AcOH, MeOH, RT, 6 h, 74%; e) Et
2
Cl
(7m in MeOH), 08C!RT, over-
N, CH Cl , RT, over-
2
, 08C, 1 h, then
3
2
2
3
night, 66% (two steps); d) NaBH
night, 79%.
3
3
2
2
To assign biological activity to one enantiomer and/or one
conformer, a chiral-phase HPLC separation was performed with
a ChiralPak IA column (see Figure 1 and Supporting Informa-
tion page 4). Among the four possible isomers, only two peaks
were detected that correspond to the two enantiomers (no
separation of the conformers). No contamination of Retro-1-E1
by Retro-1-E2 or vice versa could be identified on the chroma-
the rotamers around the amide bond) on biological activity.
Both (S)- and (R)-Retro-1 were evaluated for inhibition of Stx
cytotoxicity. (S)-Retro-1 inhibited Stx cellular action with an
EC50 value of 3.6Æ0.3 mm (Figure 2A middle and 2B), whereas
(R)-Retro-1 exhibited an EC50 value of 9.4Æ1.4 mm against Stx
(Figure 2A bottom and 2B). In addition, (S)-Retro-1 offered an
R-value of 144Æ24.5 (Figure 2C), and (R)-Retro-1 an R-value of
51.6Æ14.8 (Figure 2C). A eudismic ratio of 2.6 was calculated
from the EC50 values of cytotoxicity assays reported herein, re-
flecting low enantioselectivity. The same ratio was obtained in
comparing the R-values of each enantiomer.
1
tograms (Figure 1A). H NMR analysis was undertaken on each
enantiomer, confirming the presence of the two conformers.
To unequivocally assign the absolute configuration of each
enantiomer, X-ray crystal structure determination was obtained
Numerous proteins have been shown to regulate retrograde
transport at the early endosome–TGN interface. However, it
was previously demonstrated that only the SNARE protein syn-
taxin-5 is strongly relocalized in Retro-1-treated cells, whereas
[16]
the subcellular distribution of the Golgi was not affected.
We further investigated whether Retro-1 enantiomers can also
affect the subcellular distribution of syntaxin-5. Syntaxin-5
(
Figure 3, green) was strongly relocalized in Retro-1-treated
cells after 1 h (Figure 3B), in comparison with control (Fig-
ure 3A). The same phenotype was observed with either (S)- or
(
R)-Retro-1 used at their EC50 values (Figure 3C and D, respec-
tively). These data suggest that both enantiomers might have
the same cellular target.
The small difference in biological activities between the two
enantiomers cannot be readily assigned to a three-point con-
[19]
tact model. Nevertheless, a complex behavior based on the
influence of stereochemistry at C5 combined with the M and P
[20]
conformations of the seven-membered ring (and also possi-
bly rotamers around the amide bond) could justify the results
obtained in evaluating the enantiomers of Retro-1 against Stx,
showing similar activities between the two. Similar results
were previously demonstrated for the binding of benzodiaze-
[21]
pines to the GABA receptor.
In summary, we describe the synthesis and chromatographic
separation of Retro-1 enantiomers. In vitro experiments proved
that the configuration of the stereocenter at position 5 is not
Figure 1. A) Chiral HPLC separation of enantiomers of Retro-1. B) Structural
determination of the Retro-1 enantiomers by X-ray crystallography.
ChemMedChem 2015, 10, 1153 – 1156
www.chemmedchem.org
1154
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communications
crucial for the activity of this
compound, as both enantio-
mers are active at protecting
cells against Stx and relocaliz-
ing syntaxin-5. Nevertheless,
the absolute stereochemistry
of the eutomer has been as-
signed by X-ray diffraction ex-
periments.
The
S isomer
proved to be slightly more
active than the R isomer. Be-
cause no 1,3,4,5-tetrahydro-
2
H-benzo[e][1,4]diazepin-2-
ones possess bioactive proper-
ties, only little information is
available regarding the influ-
ence of chirality at C5, and we
are currently working on a few
hypotheses to decipher the
impact of M or P helicity on
the bioactive properties of
Retro-1 against Stx. We have
also started a structure–activi-
ty relationship study to opti-
mize the activity of the hit
compound in order to assess
the impact of the enantiose-
lectivity of analogues both in
vitro and in vivo; results from
these studies will be reported
in due course.
Figure 2. Biological activity of Retro-1 and its enantiomers toward Stx. A) HeLa cells were incubated for 4 h with
rac-Retro-1 (top), (S)-Retro-1 (middle), (R)-Retro-1 (bottom), or carrier only (DMSO, open circles) before the addition
of Stx at the indicated concentrations for 20 h. Media was removed and replaced with DMEM containing
14
À1
[
C]leucine at 0.5 mCimL for 7 h before quantifying radioactivity. Each data point represents the mean ÆSEM of
duplicates of a representative experiment. B) EC50 values ÆSEM of rac-Retro-1, (S)-Retro-1, and (R)-Retro-1 deter-
mined from five independent experiments. C) R-values ÆSEM of rac-Retro-1, (S)-Retro-1, and (R)-Retro-1 deter-
mined from five independent experiments.
Abbreviations
Stx: Shiga toxin, HUS: hemolytic uremic syndrome, Gb3: globo-
triaosylceramide, EC : 50% effective concentration, HPLC: high-
50
performance liquid chromatography, TGN: trans-Golgi network,
DMSO: dimethyl sulfoxide, SEM: standard error of the mean, NBS:
N-bromosuccinimide.
Acknowledgements
This work (Project RetroScreen) was funded by the French Na-
tional Agency for Research (ANR) (contract ANR-11-BSV2-0018),
the Joint Ministerial Program of R&D against CNRBE Risks, and
the CEA. This study benefited from the facilities and expertise
at the Imagif Cell Biology Unit of the Gif-sur-Yvette campus of
CNRS (www.imagif.cnrs.fr), which is supported by the Conseil
GØnØral de l’Essonne.
Keywords: benzodiazepines · Retro-1 · retrograde transport ·
Shiga toxin
Figure 3. Retro-1 and its enantiomers relocalize syntaxin-5. Cells were treat-
ed for 1 h with A) carrier (0.05% DMSO), B) rac-Retro-1 (6 mm), C) (S)-Retro-
[1] D. W. K. Acheson, G. T. Keusch in The Comprehensive Sourcebook of Bac-
terial Protein Toxins, 2nd ed. (Eds.: J. E. Alouf, J. H. Freer), Elsevier, Paris,
1999, pp. 229–242.
1
(3.5 mm), or D) (R)-Retro-1 (9.4 mm), then fixed and labeled for syntaxin-5
protein (green) and nuclei (blue). Relocalization of syntaxin-5 is observed
after treatment with rac-Retro-1, (S)-Retro-1, and (R)-Retro-1.
[2] A. D. O’Brien, R. K. Holmes, Microbiol. Rev. 1987, 51, 206–220.
ChemMedChem 2015, 10, 1153 – 1156
www.chemmedchem.org
1155
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communications
[
[
3] P. I. Tarr, C. A. Gordon, W. L. Chandler, Lancet 2005, 365, 1073–1086.
4] C. Frank, D. Werber, J. P. Cramer, M. Askar, M. Faber, M. an der Heiden,
H. Bernard, A. Fruth, R. Prager, A. Spode, M. Wadl, A. Zoufaly, S. Jordan,
M. J. Kemper, P. Follin, L. Muller, L. A. King, B. Rosner, U. Buchholz, K.
Stark, G. Krause, N. Engl. J. Med. 2011, 365, 1771–1780.
Schreiber, C. U. von Seydewitz, J. Steinhoff, S. Stracke, S. Suerbaum, A.
van de Loo, M. Vischedyk, K. Weissenborn, P. Wellhoner, M. Wiesner, S.
Zeissig, J. Buning, M. Schiffer, T. Kuehbacher, BMJ 2012, 345, e4565.
[11] L. Johannes, V. Popoff, Cell 2008, 135, 1175–1187.
[12] P. Watson, R. A. Spooner, Adv. Drug Delivery Rev. 2006, 58, 1581–1596.
[13] K. Sandvig, B. Spilsberg, S. U. Lauvrak, M. L. Torgersen, T. G. Iversen, B.
van Deurs, Int. J. Med. Microbiol. 2004, 293, 483–490.
[
5] H. Karch, E. Denamur, U. Dobrindt, B. B. Finlay, R. Hengge, L. Johannes,
E. Z. Ron, T. Tønjum, P. J. Sansonetti, M. Vicente, EMBO Mol. Med. 2012,
4
, 841–848.
6] P. I. Tarr, J. E. Sadler, W. L. Chandler, J. N. George, H. M. Tsai, Lancet 2012,
79, 516–517.
7] F. Proulx, J. P. Turgeon, G. Delage, L. Lafleur, L. Chicoine, J. Pediatr. 1992,
21, 299–303.
[14] J. M. Lord, L. M. Roberts, W. I. Lencer, Curr. Top. Microbiol. Immunol.
2005, 300, 149–168.
[15] J. Barbier, C. Bouclier, L. Johannes, D. Gillet, Toxins 2012, 4, 15–27.
[16] B. Stechmann, S. K. Bai, E. Gobbo, R. Lopez, G. Merer, S. Pinchard, L. Pan-
igai, D. Tenza, G. Raposo, B. Beaumelle, D. Sauvaire, D. Gillet, L. Jo-
hannes, J. Barbier, Cell 2010, 141, 231–242.
[
[
[
3
1
8] H. Trachtman, A. Cnaan, E. Christen, K. Gibbs, S. Zhao, D. W. Acheson, R.
Weiss, F. J. Kaskel, A. Spitzer, G. H. Hirschman, JAMA J. Am. Med. Assoc.
[17] X. Ming, K. Carver, M. Fisher, R. Noel, J. C. Cintrat, D. Gillet, J. Barbier, C.
Cao, J. Bauman, R. L. Juliano, Nucleic Acids Res. 2013, 41, 3673–3687.
[18] N. Gupta, V. Pons, R. Noel, D. A. Buisson, A. Michau, L. Johannes, D.
Gillet, J. Barbier, J. C. Cintrat, ACS Med. Chem. Lett. 2014, 5, 94–97.
[19] L. H. Easson, E. Stedman, Biochem. J. 1933, 27, 1257–1266.
[20] M. Czugler, A. Kµlmµn, J. Rçhricht, M. Low, L. Ürçgdi, L. Kisfaludy, Tetra-
hedron Lett. 1977, 18, 917–920.
2003, 290, 1337–1344.
[
9] A. L. Lapeyraque, M. Malina, V. Fremeaux-Bacchi, T. Boppel, M. Kirsch-
fink, M. Oualha, F. Proulx, M. J. Clermont, F. Le Deist, P. Niaudet, F. Schae-
fer, N. Engl. J. Med. 2011, 364, 2561–2563.
[
10] J. Menne, M. Nitschke, R. Stingele, M. Abu-Tair, J. Beneke, J. Bramstedt,
J. P. Bremer, R. Brunkhorst, V. Busch, R. Dengler, G. Deuschl, K. Feller-
mann, H. Fickenscher, C. Gerigk, A. Goettsche, J. Greeve, C. Hafer, F. Ha-
genmuller, H. Haller, S. Herget-Rosenthal, B. Hertenstein, C. Hofmann,
M. Lang, J. T. Kielstein, U. C. Klostermeier, J. Knobloch, M. Kuehbacher,
U. Kunzendorf, H. Lehnert, M. P. Manns, T. F. Menne, T. N. Meyer, C. Mi-
chael, T. Munte, C. Neumann-Grutzeck, J. Nuernberger, H. Pavenstaedt,
L. Ramazan, L. Renders, J. Repenthin, W. Ries, A. Rohr, L. C. Rump, O. Sa-
muelsson, F. Sayk, B. M. Schmidt, S. Schnatter, H. Schocklmann, S.
[21] C. Bertucci, A. Demuro, G. Giannaccini, J. Pharm. Biomed. Anal. 1997, 15,
803–809.
Received: March 31, 2015
Published online on June 1, 2015
ChemMedChem 2015, 10, 1153 – 1156
www.chemmedchem.org
1156
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim