Molecular insights to the bioactive form of BV02, a reference inhibitor of 14-3-3σ protein-protein interactions

Article abstract of DOI:10.1016/j.bmcl.2015.12.066

BV02 is a reference inhibitor of 14-3-3 protein-protein interactions, which is currently used as chemical biology tool to understand the role of 14-3-3 proteins in pathological contexts. Due to chemical instability in certain conditions, its bioactive for

Full text of DOI:10.1016/j.bmcl.2015.12.066

Bioorganic & Medicinal Chemistry Letters 26 (2016) 894–898
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Molecular insights to the bioactive form of BV02, a reference
inhibitor of 14-3-3r protein–protein interactions
Daniela Valensin a,y, Ylenia Cau a,y, Pierpaolo Calandro a,à, Giulia Vignaroli a, Lucia Dello Iacono a,à
,
Mario Chiariello b, Mattia Mori a,c, Maurizio Botta a,d,
⇑
a Dipartimento di Biotecnologie, Chimica e Farmacia, Università degli studi di Siena, via A. Moro snc, Via Aldo Moro 2, 53100 Siena, Italy
b Istituto Toscano Tumori, Core Research Laboratory and Consiglio Nazionale delle Ricerche, Istituto di Fisiologia Clinica, sede di Siena, Via Fiorentina 1, 53100 Siena, Italy
c Center for Life Nano Science@Sapienza, Istituto Italiano di Tecnologia, viale Regina Elena 291, 00161 Roma, Italy
d Sbarro Institute for Cancer Research and Molecular Medicine, Center for Biotechnology, College of Science and Technology, Temple University, BioLife Science Building, Suite 333,
1900 N 12th Street, Philadelphia, PA 19122, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
BV02 is a reference inhibitor of 14-3-3 protein–protein interactions, which is currently used as chemical
biology tool to understand the role of 14-3-3 proteins in pathological contexts. Due to chemical instabil-
ity in certain conditions, its bioactive form has remained unclear. Here, we use NMR spectroscopy to
Received 20 November 2015
Revised 16 December 2015
Accepted 19 December 2015
Available online 21 December 2015
prove for the first time the direct interaction between the molecule and 14-3-3r, and to depict its bioac-
tive form, namely the phthalimide derivative 9. Our work provides molecular insights to the bioactive
form of the 14-3-3 PPI inhibitor and facilitates further development as candidate therapeutic agent.
Ó 2015 Elsevier Ltd. All rights reserved.
Keywords:
14-3-3
Drug discovery
Bioactive form
NMR
Lead compound
Chemical transformation
Chemical instability
The molecular characterization of the bioactive form of a lead
compound is one of the major challenges in drug discovery.
Although efficient and sophisticated drug design strategies are
available,1,2 the bioactivity of a small molecule in vitro or in vivo
may be due to a different chemical species with respect to the orig-
inal design. Starting from a lead compound, exposure to cell envi-
ronment, reactive-oxygen species (ROS), culture medium, specific
pH values or the transformation operated by metabolic enzymes
can facilitate the generation of new chemical entities. The newly
formed molecule may be definitely inactive in the biological con-
text, or may be the main responsible for the observed biological
effects.3 Especially in this second scenario, the chemical character-
ization of the bioactive form of a lead compound becomes of
extreme importance to drive further optimization or to depict its
mechanism of action. However, this is still a challenging task in
drug discovery, which benefits of high resolution techniques such
as Nuclear Magnetic Resonance (NMR).4
In last years, we devoted many efforts to the design, synthesis
and biochemical/biological characterization of small molecule
inhibitors of 14-3-3 protein–protein interactions (PPI).5–9 14-3-3
is a family of ubiquitous and highly conserved proteins expressed
in eukaryotes, which perform their regulatory activity in cell by
establishing PPI with a large number of mainly phosphorylated
partners. Based on their involvement in human diseases such as
cancer, neurodegenerative disorders, and infection by Giardia
intestinalis, 14-3-3 proteins are considered as attractive targets in
drug discovery.9–11 A number of small molecule modulators of
14-3-3 have been discovered, acting as stabilizers or inhibitors of
14-3-3-mediated PPI.9,12 Among them, the 14-3-3 PPI inhibitor
BV02 (Fig. 1) was first discovered by our own research group in
2010 by screening in silico a commercial compounds library
against the crystallographic structure of 14-3-3r 5
. In subsequent
works, BV02 has been used as starting point in lead optimization,
thus leading to a number of chemical derivatives endowed with
anticancer effects in cells, which were also able to synergize with
doxorubicin in multidrug resistant cancer cells.8 BV02 is still con-
sidered a reference inhibitor of 14-3-3 PPI and has been used in a
⇑
Corresponding author. Tel.: +39 0577 234306; fax: +39 0577 234333.
E-mail address: botta.maurizio@gmail.com (M. Botta).
Authors equally contributed to this work.
P.C. and L.D.I. contributed only in part to this work, with a partial support in
y
à
expression and purification of recombinant 14-3-3, respectively.
http://dx.doi.org/10.1016/j.bmcl.2015.12.066
0960-894X/Ó 2015 Elsevier Ltd. All rights reserved.

D. Valensin et al. / Bioorg. Med. Chem. Lett. 26 (2016) 894–898
895
Figure 1. The reversible dehydration path that converts 9 into BV02 and vice versa. The chemical structure of the molecules is provided, alongside with atom numbering
scheme adopted in NMR experiments.
number of works to monitor the role of 14-3-3 proteins in different
pathological contexts.11,13–16
Recently, we have highlighted the possible chemical instability
of BV02 and have identified the reversible dehydration path
that may transform BV02 into the corresponding phthalimide
derivative (namely 9) and vice versa (Fig. 1).8 Molecule 9 showed
a similar binding conformation to 14-3-3r as BV02 by molecular
docking, and was equipotent to BV02 in cell-based high-
throughput screening conditions.8 However, it is still unclear if
the inhibition of 14-3-3 PPI is due to BV02 or the corresponding
phthalimide 9.
Driven by the research interest in clarifying this issue and by
the multiple requests received by scientists interested in using
BV02 as reference 14-3-3 PPI inhibitor, here we address the chal-
lenging characterization of BV02 bioactive form by means of
NMR spectroscopy. Results of our NMR investigation unequivocally
show that the compound is able to interact with recombinant 14-
Figure 2. Aromatic region of 1H NMR spectra of 9 0.1 mM in TRIS buffer at pH 8.0, T
298 K at different times.
3-3r, and clarify that the bioactive form is represented by the
phthalimide derivative 9.
studies. Lowering pH has also a strong impact on the kinetic of
9-BV02 conversion as shown by the NMR spectra recorded at dif-
ferent pH values (Supplementary information, Fig. S2). Table 1
The behavior of 9 in solution was monitored by NMR. In partic-
ular, the chemical shift variation of aromatic protons proved to be a
suitable tool to monitor the interconversion of 9 into BV02. NMR
spectra of 9, recorded in TRIS buffer at pH 8.0, indicate that the
molecule is not stable in solution but it rather converts in BV02
after 50 h (Fig. 2). In detail, NMR spectra recorded after 24 h show
the appearance of new resonances, whose intensity is gradually
increasing. On the other hand, the signals belonging to 9 are about
50% less intense after 24 h, to be completely vanished out after
50 h, thus indicating that the completed conversion of 9 into
BV02 does occur within this timeframe. The chemical identity of
BV02 generated by the path described in Figure 1 was further con-
firmed by MS spectrometry (m/z = 396.0).
00
00
reports the intensity percentage of H3 and H7 signals of 9 at dif-
ferent pH at both 0 and 24 h. These data clearly indicate that 9 is
more stable in solution at lower pH values, as very small percent-
age of BV02 is observed at 3.0–5.2 pH range. Overall, results of
these NMR measurements suggest that the conversion of 9 to
BV02 in solution may occur, and it is favored by high concentra-
tions and high pH values.
By taking into consideration the behavior of 9 in solution, we
decided to investigate its interaction with recombinant 14-3-3
r
protein at different pH values. The interaction between the two
molecules was tested by means of NMR spectroscopy, performing
tr-NOESY, and longitudinal relaxation rates experiments of the
ligand either in absence or in presence of sub-stoichiometry
amount (0.1 equiv) of the protein. As extensively reported for a
large number of small molecules,17–20 monitoring tr-NOEs is a
straightforward method for screening protein–ligand association.
As reported in Figure 3A, the NOESY spectrum of 9 alone shows
no cross-peaks, in agreement with the null NOEs effects typically
expected for a molecule of this size.6 On the contrary, the addition
The dependence of 9-BV02 conversion on concentration and pH
was further evaluated. Change in concentrations end up with not
identical kinetic behavior, as easily observed by comparing the
00
00
intensity of signals belonging to H3 and H7 of BV02 at different
hours for both 0.1 and 2.5 mM solutions of 9 (Supplementary infor-
mation, Fig. S1). At these conditions, the intensity of aromatic sig-
nals of BV02 is progressively increased to 70% and 81% for 0.1 and
2.5 mM solutions, respectively. However, it should be noted that
the bioactive concentration of 9 is in the micromolar range, as pre-
viously determined.8 Therefore, the effect of concentration in the
conversion of 9 to BV02 should be relevant only in concentrated
stock solutions, whereas it is expected not to affect biological
of 0.1 equiv of 14-3-3r at pH 3.2, ends up with the presence of
strong NOE correlations with the same sign as the diagonal signals
(Fig. 3B). NOEs cross-peaks involve all aromatic protons of 9, in
particular intense correlations between the phenyl and methyl

896
D. Valensin et al. / Bioorg. Med. Chem. Lett. 26 (2016) 894–898
Table 1
the only species in solution we performed NOESY spectra in pres-
ence of 0.1 equiv of 14-3-3- (Supplementary information, Fig. S3).
The obtained data show the absence of any NOEs cross-peaks,
strongly indicating that BV02 does not interact with 14-3-3-
Percentage of 9 and BV02 in solution at t = 0 and t = 24 h
r
% t = 0 h
% t = 24 h*
r
9
BV02
9
BV02
These findings also explain the data reported in Figure 3C and D,
where the partial conversion of 9 to BV02, occurring during the
time needed for NOESY acquisition (ꢀ22 h), reduces the concentra-
tion of 9 in solution thus affecting the sensitivity of the NMR
measurement.
pH 8.0
pH 6.2
pH 5.2
pH 3.2
100
100
100
100
0
0
0
0
42
62
89
92
58
38
11
8.0
*
00
00
The data were obtained by integrating H3 and H7 signals of 9 and BV02 at t = 0
t = 24 h. The sum of the two integrals was normalized to 100.
Interestingly, the interaction between 9 and 14-3-3
r proved to
slow down the conversion rate of 9 to BV02. In fact, comparing
NMR spectra of 9 alone with those recorded for the molecule in
protons are evident. This behavior is consistent with the fact that 9
interacts with 14-3-3-r. On the other hand, no or slight changes
were observed on NOESY spectra recorded at pH 6.2 and 8.0 as
shown in Figure 3C and D, respectively. By carefully looking at
the NMR 1D spectra of samples at pH 6.2 and 8.0, recorded after
NOESY spectra, and in agreement with results above, we realized
the binary complex with 14-3-3
that it is still possible to detect noticeable amounts of 9 in the
9/14-3-3 complex after 72 hrs, whereas the compound alone in
solution is fully converted to BV02 at that time (Fig. 4). The forma-
tion of the 9/14-3-3 complex most likely protects 9 from the
aqueous medium and slows down its conversion to BV02.
The interaction between 9 and 14-3-3 was also verified by
r recorded at pH 8.0, clearly shows
r
r
that 9 is partially converted to BV02 in presence of 14-3-3-
well. In addition, we decided to evaluate the ability of BV02 to
interact with 14-3-3- . Starting from solutions of 9 we waited
r as
r
relaxation rates measurements. Table 2 shows the selective (R1sel
)
and non-selective Rn1sel longitudinal relaxation rates of H7 and
r
for its full conversion to BV02, then once verified that BV02 was
00
NCH3 (H1 ) nuclei of 9 either in absence or in presence of 0.1 equiv
Figure 3. NOESY spectra of: (A) 9 0.1 mM pH 3.2; (B) 9 0.1 mM + 0.1 equiv of 14-3-3
r pH 3.2; (C) mixture of 9 and BV02 0.1 mM + 0.1 equiv of 14-3-3r pH 6.2; (D) mixture of
9 and BV02 0.1 mM + 0.1 equiv of 14-3-3
r pH 8.0.

D. Valensin et al. / Bioorg. Med. Chem. Lett. 26 (2016) 894–898
897
to increase the possibility to have the bioactive species 9 as pre-
dominant in the sample.
Acknowledgements
Authors are grateful to Christian Ottmann for kindly providing
the expression plasmid of full length 14-3-3r. D.V. thanks
MIUR-PRIN project 2010M2JARJ_004 for financial support. Project
number 6/10 - Development of new inhibitors of tyrosine kinases
and 14-3-3 proteins for the treatment of drug-resistant leukemia.
Art.11 of D.M. n.593 of 8th August 2000.
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmcl.2015.12.
066.
References and notes
1. Mori, M.; Botta, M. In Drug Design and Screening by In Silico Approaches, in
Trypanosomatid Diseases: Molecular Routes to Drug Discovery; Jäger, T., Koch, O.,
Flohé, L., Eds.; Wiley-VCH Verlag GmbH & Co. KGaA: Weinheim, Germany,
2013. http://dx.doi.org/10.1002/9783527670383.ch4.
2. Mori, M.; Cau, Y.; Vignaroli, G.; Laurenzana, I.; Caivano, A.; Vullo, D.; Supuran,
C. T.; Botta, M. ACS Chem. Biol. 1964, 2015, 10.
Figure 4. Superimposition of 1H NMR spectra of 9 alone (black line) or in complex
with 14-3-3 at pH 8.0 (red line) after 24 h (upper panel), and after 72 h (bottom
panel). NMR peaks indicative of the presence of 9 are highlighted by arrows in the
r
bottom panel.
3. Shu, Y. Z.; Johnson, B. M.; Yang, T. J. AAPS J. 2008, 10, 178.
4. Roberts, G. C. Drug Discov. Today 2000, 5, 230.
5. Corradi, V.; Mancini, M.; Manetti, F.; Petta, S.; Santucci, M. A.; Botta, M. Bioorg.
Med. Chem. Lett. 2010, 20, 6133.
6. Corradi, V.; Mancini, M.; Santucci, M. A.; Carlomagno, T.; Sanfelice, D.; Mori, M.;
Vignaroli, G.; Falchi, F.; Manetti, F.; Radi, M.; Botta, M. Bioorg. Med. Chem. Lett.
2011, 21, 6867.
7. Mancini, M.; Corradi, V.; Petta, S.; Barbieri, E.; Manetti, F.; Botta, M.; Santucci,
M. A. J. Pharmacol. Exp. Ther. 2011, 336, 596.
8. Mori, M.; Vignaroli, G.; Cau, Y.; Dinic, J.; Hill, R.; Rossi, M.; Colecchia, D.; Pesic,
M.; Link, W.; Chiariello, M.; Ottmann, C.; Botta, M. ChemMedChem 2014, 9, 973.
9. Mori, M.; Vignaroli, G.; Botta, M. Drug Discov. Today Technol. 2013, 10(4), e541–
7. http://dx.doi.org/10.1016/j.ddtec.2012.10.001.
of 14-3-3
consistent with a null NOE effect. On the other hand, the addition
of 14-3-3 causes a large increase of selective relaxation rates, in
agreement with a slowing down of the molecular tumbling of 9
upon protein interaction. This behavior is easily denoted by the
ratio between the two values, whose value is reduced by increasing
the molecular tumbling in solution.21–26
r
. For 9 alone, similar Rs1el and R1nsel values are measured,
r
10. Cau, Y.; Fiorillo, A.; Mori, M.; Ilari, A.; Botta, M.; Lalle, M. J. Chem. Inf. Model.
2015.
11. Zhao, J.; Meyerkord, C. L.; Du, Y. H.; Khuri, F. R.; Fu, H. A. Semin. Cell Dev. Biol.
2011, 22, 705.
Our detailed NMR investigation convincingly confirms that the
bioactive form of the well-known 14-3-3
r PPI inhibitor BV02 is
the ‘closed’ phthalimide form, namely 9.27 Moreover, our study
12. Ottmann, C. Bioorg. Med. Chem. 2013, 21, 4058.
highlights for the first time the direct interaction between this
molecule and its macromolecular target 14-3-3r, further confirm-
ing the mechanism of action at molecular level. 9 can undergo to
chemical transformation to BV02 in particular conditions such as
high concentrations or high pH values, and in time (full conversion
in 50 hrs in TRIS buffer). Contrarily to BV02, the interaction
13. Mancini, M.; Leo, E.; Takemaru, K.; Campi, V.; Castagnetti, F.; Soverini, S.; De
Benedittis, C.; Rosti, G.; Cavo, M.; Santucci, M. A.; Martinelli, G. PLoS ONE 2015,
10, e0131074.
14. Mohammad, D. K.; Nore, B. F.; Hussain, A.; Gustafsson, M. O.; Mohamed, A. J.;
Smith, C. I. Mol. Cell. Biol. 2013, 33, 3214.
15. Rana, A.; Ali, G. M.; Ali, S.; Khan, A.; Sabiha, B.; Malik, S.; Riaz, A. M.; Farooqi, A.
A. .J Cancer Res. Ther. 2013, 9, 3.
16. Aristizabal-Corrales, D.; Schwartz, S., Jr.; Ceron, J. Worm 2013, 2, e21834.
17. Ludwig, C.; Guenther, U. L. Front. Biosci. 2009, 14, 4565.
18. Ni, F. Prog. Nucl. Magn. Reson. Spectrosc. 1994, 26, 517.
19. Post, C. B. Curr. Opin. Struct. Biol. 2003, 13, 581.
20. Carlomagno, T. Annu. Rev. Biophys. Biomol. Struct. 2005, 34, 245.
21. Fielding, L. Curr. Top. Med. Chem. 2003, 3, 39.
between 9 and the recombinant 14-3-3
r was observed via NMR,
and was found to slow down the conversion of 9 to BV02, thus
allowing the detection of intact 9 in noticeable amounts also after
72 h.
In conclusion, scientists that are interested in using the refer-
ence 14-3-3 PPI inhibitor BV02 in chemical biology study should
be aware that the bioactive form is most likely the phthalimide
derivative 9. It is highly recommended to use always freshly pre-
pared samples of 9, to keep the molecule in solution for the short-
est time as possible, or to use semisolid culture media.7 Based on
the chemical path of Figure 1 and results of this work, pH values,
concentration and incubation time should be carefully monitored
22. Skinner, A. L.; Laurence, J. S. J. Pharm. Sci. 2008, 97, 4670.
23. Luo, R. S.; Liu, M. L.; Mao, X. A. Spectrochim. Acta, Part A 1897, 1999, 55.
24. Valensin, G.; Kushnir, T.; Navon, G. J. Magn. Reson. 1982, 46, 23.
25. Berti, F.; Gaggelli, E.; Valensin, G. Appl. Magn. Reson. 2000, 18, 289.
26. Delfini, M.; Gianferri, R.; Dubbini, V.; Manetti, C.; Gaggelli, E.; Valensin, G. J.
Magn. Reson. 2000, 144, 129.
27. Materials and methods. Small molecules were synthesized and characterized as
described previously.5,6 IUPAC names: 2-(1,5-dimethyl-3-oxo-2-phenyl-2,
3-dihydro-1H-pyrazol-4-yl)-1,3-dioxo-5-isoindolinecarboxylic acid (9) and 2-
[(1,5-dimethyl-3-oxo-2-phenyl-2,3-dihydro-1H-pyrazol-4-yl)carbamoyl] tereph-
thalic acid (BV02). The BL21(DE3)pLys strain of Escherichia coli was transformed
with the pPROEX HTb-14-3-3 plasmid. Bacteria were cultured in LB medium at
37 °C, overnight. IPTG 100 lM was added to the culture medium to stimulate
Table 2
protein expression. After 3 h at 37 °C, cells were collected by centrifugation
and incubated for 10 min in cold lysis buffer (50 mM Tris, 300 mM NaCl,
10 mM Imidazole, 100 mM PMSF, 0.05% Tween 20, 0.4% Triton X-100). Cells
were sonicated on ice in pulsed mode (10 pulse of 10 s each one), membranes
were collected by centrifugation at 4 °C for 10 min at 7500g and the
supernatant was finally recovered. To purify Hys-tag 14-3-3 proteins from
lysate, Nickel beads (His Mag Sepharose Ni) were added into supernatant,
incubated on see-saw rocker at 4 °C for 60 min and then collected at 2000g for
5 min. Beads were then washed and the protein was finally recovered from
beads adding elution buffer (50 mM ph 8.0 Tris, 300 mM NaCl, 250 mM
Non selective and selective relaxation enhancements of protons of 9 0.1 mM, pH 3.2
in the absence and in presence of 0.1 equiv of 14-3-3
Rn1sel Rs1el
9/14-3-3
r
Rn1sel/Rs1el
9/14-3-3r
9
r
9
9/14-3-3
r
9
H7
NCH3
0.28
0.90
0.70
1.12
0.32
0.93
2.04
2.22
0.87
0.97
0.34
0.50

898
D. Valensin et al. / Bioorg. Med. Chem. Lett. 26 (2016) 894–898
Imidazole, 100 mM PMSF, 0.5 % Tween 20). Imidazole was later removed
MLEV-17 mixing sequence. NOESY spectra were acquired with a mixing time
of 320 ms. Spin lattice relaxation rates were measured with inversion recovery
pulse sequences. The same sequence was also used to measure the single-
through Amicon Ultra Centrifugal Filters. NMR spectra were obtained at 14.1 T
with a Bruker Avance 600 Spectrometer operating at controlled temperature
( 0.2 K) and using a 5 mm SEI probe. Chemical shifts were referenced to
external 3-(Trimethylsilyl)propionic-2,2,3,3-d4 acid sodium salt (TMSP). TOCSY
and NOESY spectra were obtained by using standard pulse sequences. TOCSY
experiments were acquired with a total spin-locking time of 60 ms using a
selective relaxation rates by means of suitably shaped
p-pulses instead of the
usual non-selective -pulse. All rates were calculated by regression analysis of
p
the initial recovery curves of longitudinal magnetization components leading
to errors not larger than 3%.