Synthesis of Thicolchicine-Based Conjugates: Investigation towards Bivalent Tubulin/Microtubules Binders

Article abstract of DOI:10.1002/cplu.201800497

Four different hybrid compounds have been efficiently synthesized by conjugation of deacetylthiocolchicine with pironetin-inspired derivatives. The modest bioactivity and the apparent absence of interaction with α-tubulin is explained by a posteriori in silico investigation, which suggests a relevant distance between the thiocolchicine binding site and the proper pocket on the α-tubulin. The modest activity on resistant cells suggested that the lipophilic nature of the linker used renders the resulting compounds better substrates for p-Gp efflux pumps. The study better clarifies the design of bivalent compounds that target hetero tubulin/microtubules.

Full text of DOI:10.1002/cplu.201800497

Full Papers
DOI: 10.1002/cplu.201800497
1
2
3
4
5
6
7
8
9
Synthesis of Thicolchicine-Based Conjugates: Investigation
towards Bivalent Tubulin/Microtubules Binders
Elisa Bonandi,[a] Francesca Foschi,[a] Cristina Marucci,[a] Federico Dapiaggi,[a] Maurizio Sironi,[a]
Stefano Pieraccini,[a] Michael S. Christodoulou,[a, b] Francisco de Asís Balaguer,[c]
J. Fernando Díaz,[c] Nace Zidar,[d] and Daniele Passarella*[a]
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Four different hybrid compounds have been efficiently synthe-
sized by conjugation of deacetylthiocolchicine with pironetin-
inspired derivatives. The modest bioactivity and the apparent
absence of interaction with α-tubulin is explained by a
posteriori in silico investigation, which suggests a relevant
distance between the thiocolchicine binding site and the
proper pocket on the α-tubulin. The modest activity on
resistant cells suggested that the lipophilic nature of the linker
used renders the resulting compounds better substrates for p-
Gp efflux pumps. The study better clarifies the design of
bivalent compounds that target hetero tubulin/microtubules.
Introduction
activities, which are regulated by tubulin post translational
modifications.[3] Depending on the mode of action, MT-target-
ing agents can be divided into two different families: MT-
stabilizing agents and MT-destabilizing agents. In the last
decade, both these classes have been successfully applied in
the treatment of several cancer types,[4] and the search of new
cytotoxic or neuroprotective agents is still of primary interest.
Thus, our interest in the obtainment of new molecules able to
target tubulin and MTs, prompted us to plan the synthesis of
new tubulin/microtubules-targeting bivalent compounds. Bi-
functional drugs, bearing chemical entities that are able to bind
to separate sites of the protein, recently attracted the attention
of the scientific community, thanks to their possible great
pharmacological potential. The idea of conjugating different
MTs binders into bivalent entities led to the obtainment of
novel antimitotic agents, including taxoid-colchicine,[5] vinca
alkaloid-taxoid hybrids,[6] colchicineÀ tubulizine,[7] colchicine-pir-
onetin,[8] podophyllotoxin-thiocolchicine and hybrids merging
histone deacetylase inhibitors with different tubulin binders.[9,10]
On one hand the use of too short or rigid linkers prevents the
interaction with the desired binding site and leads to steric
clashes. On the other hand, the use of a too long and flexible
linker may lead to solubility and cell internalization problems.[11]
As the linker seems to play a critical role tuning the bioactivity
of these class of molecules, we planned the synthesis of new
bivalent compounds, constituted by a well-known β-tubulin
binder (thiocolchicine) and a pironetin-inspired compound
(hybrid 1, Figure 1) connected by linkers of different types and
length, to investigate how their properties affect the binding
capability.
Microtubules (MTs) are highly dynamic structures, diffused in
eukaryotic cells, as well as in some bacteria.[1] They are
fundamental components of the cytoskeleton and they play
important roles in several essential cellular processes, such as
cell division, locomotion, intracellular transport and cell shape
definition. MTs are originated from the head-to-tail arrange-
ment of α,β-tubulin heterodimers, leading to the formation of
cylindrical assemblies. In physiological conditions, MTs are
characterized by alternating phases of growth and shrinkage, a
behaviour known as “dynamic instability”. This dynamism is
essential for MTs correct functionality, in particular for the
formation of the mitotic spindle during the mitosis process.[2]
The perturbation of this dynamic behaviour can result in
deleterious effects for the cells, halting the cell replication and
inducing cellular apoptosis. Therefore, MTs represent very
suitable drug target, for anticancer therapy and also for
neurodegenerative diseases treatment, as recently emerged. In
fact, MTs dynamic instability is essential in many neuronal
[a] Dr. E. Bonandi, Dr. F. Foschi, C. Marucci, Dr. F. Dapiaggi, Prof. M. Sironi,
Dr. S. Pieraccini, Dr. M. S. Christodoulou, Prof. D. Passarella
Dipartimento di Chimica
Università degli Studi di Milano
via Golgi 19, 20133 Milano (Italy)
E-mail: daniele.passarella@unimi.it
[b] Dr. M. S. Christodoulou
Dipartimento di Scienze Farmaceutiche
Università degli Studi di Milano
via Golgi 19, 20133 Milano (Italy)
[c] Dr. F. de Asís Balaguer, Dr. J. F. Díaz
Chemical and Physical Biology
Centro de Investigaciones Biológicas
Consejo Superior de Investigaciones Cientificas CIB–CSIC
Madrid 28040 (Spain)
Results and Discussion
[d] Dr. N. Zidar
Faculty of Pharmacy
University of Ljubljana
Askerceva cesta 7, 1000 Ljubljana (Slovenia)
For the first explorative studies a ten-atom all-carbon chain and
a 24-atom pseudo-peptidic chain were taken into account. As
active units, we choose thiocolchicine and hybrid compound 1
as models of β- and possible α-binder, respectively (Figure 1).
Supporting information for this article is available on the WWW under
https://doi.org/10.1002/cplu.201800497
ChemPlusChem 2019, 84, 98–102
98
© 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Figure 1. General scheme for the assembly of the conjugate compounds and
representation of the employed specific components.
°
Scheme 1. Reagents and conditions: a) 2 M HCl, MeOH, 85 C, 3 days; b)
sebacic acid, HATU, DIPEA, CH2Cl2, rt, 3 h; c) 1a, HATU, DIPEA, CH2Cl2, rt, 4 h.
Compounds 1 were recently designed in our laboratory, as the
merging of two natural compounds, pironetin and dumetor-
ine.[12] Pironetin is an α-tubulin binder, characterized by a δ-
unsaturated lactone, which serves as Michael acceptor for the
Cys316 residue inside α-tubulin binding site.[13] It is noteworthy
that while several β-tubulin binders are currently available, only
few molecules are known to bind α-tubulin.
The modification of pironetin structure, introducing the
piperidine ring typical of dumetorine, resulted in the obtain-
ment of an anchor point exploitable for the synthesis of
bivalent compounds.
We took into consideration three of the eight prepared
stereoisomers,[12] in particular 1a–c, for the synthesis of the new
conjugate compounds on the base of their efficacy in giving
stable and pure compounds. First, N-deacetyl-10-thiocolchicine
2, deriving by acidic hydrolysis of thiocolchicine,[14] was
converted into compound 3, exploiting a condensation with
sebacic acid. A strict control over the amount of sebacic acid
avoided the undesired reaction on both the carboxylic moieties.
Reaction of 3 with compound 1a in the presence of HATU and
DIPEA, afforded compound 4, characterized by a linker of 10
carbon atoms (Scheme 1).
The condensation reaction of compound 3 with 1b and 1c
gave unexpectedly complex mixture of compounds probably
due to the intrinsic instability of the desired compounds. The
synthesis of conjugate compounds bearing a longer pseudo-
peptidic linker was considered (Scheme 2). To this extent, the 10
carbon atoms chain of compound 3 was furtherly elongated,
through the introduction of the tert-butyl 2-aminoethylcarba-
mate fragment 5, obtained according to reported procedure.[15]
The cleavage of the Boc protecting group, upon treatment with
TFA led to compound 7. The obtained free amino group was
exploited for a further reaction with sebacic acid. Under these
conditions, the thiocolchicine derivative 8 was obtained as major
product. Final compounds 9a–c were accessed by condensation
reaction between 8 and the proper stereoisomer 1.
Scheme 2. Reagents and conditions: a) 5, HATU, DIPEA, CH2Cl2, rt, 3 h; b)
TFA, CH2Cl2, 0 C to rt, 1 h; c) sebacic acid, HATU, DIPEA, CH2Cl2, rt, 3 h; d) 1a–
°
c, HATU, DIPEA, CH2Cl2, rt, 3 h.
Biological Data
We submitted compounds 4 and 9a–c to preliminary biological
tests, aimed at the elucidation of their activity as tubulin/
microtubules targeting compounds. Firstly, they were tested on
wild-type and resistant ovarian cancer cells lines (A2780 and
A2780AD respectively). Cytotoxicity data (reported in Table 1)
seem to suggest that our compounds do not interact through
the pironetin-like unit. In fact, if this was the case, they would
have accumulated inside the cells, avoiding the P-glycoprotein
resistance like other MTs covalent binders usually do.[16]
Compound 4 is almost as toxic as N-deacetyl-10-thiocolchicine
2 in non-resistant cells but 100 times less toxic in resistant cells;
compounds 9a–c displayed even less toxicity, and also in this
case the trend is more marked on resistant cells. These results
ChemPlusChem 2019, 84, 98–102
www.chempluschem.org
99
© 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
compounds depends from the interaction of the thiocolchicine
moiety with tubulin. These results highlight a slightly difference
in activity between compound 4 and compounds 9a–c, differ-
ing in linker length. In particular, the plain inability of the
obtained hybrid compounds to exploit the dual interaction with
tubulin/microtubules can be inferred from the data. These
observations prompted us to investigate the binding mode of
these compounds through molecular modelling studies.
Table 1. Cytotoxicity data.
1
2
3
4
5
6
7
8
9
Compound
IC50 [nM][a]
R/S[b]
A2780
A2780AD
4
9a
9b
9c
11.8�1
252�7
205�15
196�0.2
13.6�2
7.8�1
4300�600
20500�500
19000�6000
46000�1500
663�23
364
81
93
235
49
colchicine
N-deacetyl-10-thiocolchicine
paclitaxel
podophyllotoxin
30.9�0.6
4
0.4�0.08
8�1
1400�200
10.1�0.7
3500
1.2
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Molecular Dynamics Simulations
[a] IC50 values determined in the parental ovarian carcinoma A2780 line
and the MDR P-glycoprotein-overexpressing ovarian carcinoma A2780/
AD10. IC50 values in nM were determined after two days’ exposure to drugs
using the MTT cell proliferation assay. Data are the mean � SE of at least
four independent experiments. [b] Ratio of IC50 (resistant cell line)/IC50
(parental cell line). Values are the calculated relative resistance of each
mutant cell line obtained by dividing the IC50 value of the resistant line by
the IC50 value of the parental line, A2780.
Molecular dynamics (MD) simulations are a useful tool to assess
the conformational stability of a ligand within the binding site.
MD simulations were performed on compounds 4, 9a,
colchicine and N-deacetyl-10-thiocolchicine in complex with a
tubulin α-β dimer. We chose the PDB 4O2B[17] (tubulin-
colchicine complex) as the starting structure for MD simulations.
This choice is justified by a higher resolution compared to any
tubulin-thiocolchicine solved structure.[11] Moreover colchicine
and thiocolchicine are very similar molecules and the conforma-
tion of the residues forming the binding site is largely super-
imposable. The experimental results suggested that ligands
interact with tubulin through the N-deacetyl-10-thiocolchicine
moiety. Accordingly, in the MD starting structures the N-
deacetyl-10-thiocolchicine nucleus of each ligand was placed
into the colchicine binding site while the pironetin moiety was
placed between α-β subunits (see the Supporting Information,
Figure S1). A 50 ns MD simulation for each complex was
performed and subsequently the root mean square deviation
(RMSD) were computed with respect to the staring structure for
each ligand after least square fit to protein Cα (Figure 3). From
these data it resulted that the longer is the linker, the greater is
the displacement of the ligands within the binding site.
prompted us to think that our bivalent compounds interact
with tubulin exploiting only the thiocolchicine, justifying in this
way the partially retained bioactivity. However, the hydrophobic
linker seems to disturb the interaction of pironetin with the
target, as well as render our compounds better substrates for P-
glycoprotein, as emerged considering that the activity de-
creases with the length of the linker. Furthermore, to confirm
that the retained antiproliferative activity of the obtained
compounds was actually related to their capacity to destabilize
tubulin, tubulin polymerization assay was performed.
Podophyllotoxin and N-deacetyl-10-thiocolchicine were also
tested, for comparative purposes. The results of the tubulin
polymerization assay are reported in Figure 2. As can be
deducted from the graphic, all the tested compounds show a
strong destabilizing effect, indicating that the toxicity observed
is exerted through inhibition of tubulin assembly. This is
coherent with the assumption that the partial activity of our
°
Figure 2. Time course of 25 μM tubulin polymerization at 37 C measured by
Figure 3. RMSD plot for the ligands. Colchicine and N-deacetyl-10-thiocolchi-
cine show a low and stationary RMSD value, while the linker of compounds
4 and 9a has a negative effect on the conformational stability of the
molecules within the binding site.
350 nm turbidimetry in the presence of DMSO vehicle (control) or 27.5 μM of
ligand. The assay was performed in triplicate and the graph represents the
mean of the three experiments.
ChemPlusChem 2019, 84, 98–102
www.chempluschem.org
100
© 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
On the other hand, colchicine and N-deacetyl-10-thiocolchi-
cine show a low and stationary RMSD trend. In Figure S2
(Supporting Information), the superimposition of the final
structure of the MD of compound 4 with the crystallographic
structure of tubulin-thiocolchinine complex can be compared
(PDB ID: 5LP6). It is possible to infer that the presence of the
linker has a displacing effect on the thiocolchinine nucleus,
tubulin/microtubules targeting compounds by linking one α-
with one β-binder is feasible and handle. In contrast this study
demonstrates that the aim to improve tubulin/microtubules
targeting by an α,β-bivalent binder sounds unlikely due to the
required dimension and the tricky relevance of the linker.
1
2
3
4
5
6
7
8
9
which can lead to a partial loss of its key hydrophobic Acknowledgements
interactions with the β-monomer.
Moreover, the root mean square fluctuations (RMSF) for
ligands with heavy atoms were considered (Figure 4). Once
again, it is evident that the presence of a long linker (compound
9a, red line) leads to a higher mobility of the thiocolchicine
moiety. This effect is less pronounced in the case of compound
4, which has a shorter linker. Colchicine and N-deacetyl-10-
thiocolchicine show the lowest RMSF values, suggesting that
any of these linkers has a negative impact on the ligands
binding mode.
D.P. expresses his gratitude to MAECI Italia-India Strategic Projects
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
2017-2019 MAE0105301 “Development of nature inspired bivalent
antitubulins as anticancer agents”. This work has been devel-
oped under the umbrella of COST Action CM1407, “Challenging
organic syntheses inspired by nature - from natural products
chemistry to drug discovery” that favoured the networking and
financed an STSM grant for N.Z.
Conflict of Interest
The authors declare no conflict of interest.
Keywords: bivalent binders · microtubules · pironetin · tubulin
binders · thiocolchicine
[1] M. Pilhofer, M. S. Ladinsky, A. W. McDowall, G. Petroni, G. J. Jensen, PLoS
Biol. 2011, 9, e1001213.
[2] a) T. Mitchison, M. Kirschner, Nature 1984, 312, 237; b) K. R. Brunden,
J. Q. Trojanowski, A. B. Smith, V. M. Lee, C. Ballatore, Bioorg. Med. Chem.
2014, 22, 5040.
[3] D. Cartelli, F. Casagrande, C. L. Busceti, D. Bucci, G. Molinaro, A.
Traficante, D. Passarella, E. Giavini; G. Pezzoli, G. Battaglia, G. Cappelletti,
Sci. Rep. 2013, 3, 1837.
[4] C. Dumontet, M. A. Jordan, Nat. Rev. Drug Discovery 2010, 9, 790.
[5] a) K. Bombuwala, T. Kinstle, V. Popik, S. O. Uppal, J. B. Olesen, J. Vina,
C. A. Heckman, J. Beilstein, J. Org. Chem. 2006, 2, 13; b) B. Danieli, A.
Giardini, G. Lesma, D. Passarella, A. Silvani, G. Appendino, A. Noncovich,
G. Fontana, E. Bombardelli, O. Sterner, Chem. Biodiversity 2004, 1, 327.
[6] D. Passarella, A. Giardini, B. Peretto, G. Fontana, A. Sacchetti, A. Silvani,
C. Ronchi, G. Cappelletti, D. Cartelli, J. Borlake, B. Danieli, Bioorg. Med.
Chem. 2008, 16, 6269
[7] Y. B. Malysheva, S. Combes, D. Allegro, V. Peyrot, P. Knochel, A.
Gavryushin, E. Andrei, A. Y. Fedorov, Bioorg. Med. Chem. 2012, 20, 4271.
[8] C. Vilanova, S. Díaz-Oltra, J. Murga, E. Falomir, M. Carda, M. Redondo-
Horcajo, J. F. Díaz, I. Barasoain, J. A. Marco, J. Med. Chem. 2014, 57,
10391.
[9] a) B. Danieli, A. Giardini, G. Lesma, D. Passarella, B. Peretto, A. Sacchetti,
A. Silvani, G. Pratesi, F. J. Zunino, Org. Chem. 2006, 71, 2848; b) G.
Cappelletti, D. Cartelli, B. Peretto, M. Ventura, M. Riccioli, F. Colombo,
J. S. Snaith, S. Borrelli, D. Passarella, Tetrahedron 2011, 67, 7354.
[10] D. Passarella, D. Comi, A. Vanossi, G. Paganini, F. Colombo, L. Ferrante,
V. Zuco, B. Danieli, F. J. Zunino, Bioorg. Med. Chem. Lett. 2009, 19, 6358.
[11] J. Marangon, M. S. Christodoulou, F. V. M. Casagrande, G. Tiana, L. Dalla
Via, A. Aliverti, D. Passarella, G. Cappelletti, S. Ricagno, Biochem. Biophys.
Res. Commun. 2016, 479, 48.
[12] C. Marucci, M. S. Christodoulou, S. Pieraccini, M. Sironi, F. Dapiaggi, D.
Cartelli, A. M. Calogero, G. Cappelletti, C. Vilanova, S. Gazzola, G.
Broggini, D. Passarella, Eur. J. Org. Chem. 2016, 2016, 2029
[13] a) A. E. Prota, J. Setter, A. B. Waight, K. Bargsten, J. Murga, J. F. Díaz,
M. O. Steinmetz, J. Mol. Biol. 2016, 428, 2981; b) J. H. Yang, Y. X. Wang,
T. J. Wang, J. Jiang, C. H. Botting, H. T. Liu, Q. Chen, J. L. Yang, J. H.
Naismith, X. F. Zhu, L. J. Chen, Nat. Commun. 2016, 7, 12103.
[14] Q. Shi, P. Verdier-Pinard, A. Brossi, E. Hamel, K. Lee, Bioorg. Med. Chem.
1997, 5, 2277.
Figure 4. RMSF plot for the ligands. The fluctuations are higher for
compound 9a, which has the longest linker.
Conclusion
In conclusion, four different conjugate compounds have been
efficiently synthetized by linking thiocolchicine nucleus with
pironetin-inspired derivatives. One of them presents 10 atom
linker and three of them present 24-atom (C10-N-C2-N-C10) linker.
The detected reduced bioactivity and the apparent absence of
interaction by the α-tubulin binder prompted us to employ an
in silico approach to evaluate the binding mode of the
compounds. The MD results suggest that a longer linker has a
displacing effect on the thiocolchinine nucleus, probably
resulting in a loss of activity. The observed diminished activity
on resistant cells whispers that the lipophilic nature of the linker
renders the described compounds, better substrates for p-Gp
efflux pumps. The design and the synthesis of conjugate
ChemPlusChem 2019, 84, 98–102
www.chempluschem.org
101
© 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
[15] J. P. Holland, V. Fisher, J. A. Hickin, J. M. Peach, Eur. J. Inorg. Chem. 2010,
48–58.
[17] A. E. Prota, F. Danel, F. Bachmann, K. Bargsten, R. M. Buey, J. Pohlmann,
1
2
3
4
5
6
7
8
S. Reinelt, H. Lane, M. O. Steinmetz, J. Mol. Biol. 2014, 426, 1848.
[16] a) R. M. Buey, E. Calvo, I. Barasoain, O. Pineda, M. C. Edler, R. Matesanz,
G. Cerezo, C. D. Vanderwal, B. W. Day, E. J Sorensen, J. A. Lopez, J. M.
Andreu, E. Hamel, F. J. Díaz, Nat. Chem. Biol. 2007, 3, 117; b) A. J. Marco,
J. Garcia-Pla, M. Carda, J. Murga, E. Falomir, C. Trigili, S. Notararigo, F. J.
Díaz, I. Barasoain, Eur. J. Med. Chem. 2011, 46, 1630; c) J. J. Field, B. Pera,
E. Calvo, A. Canales, D. Zurwerra, C. Trigili, J. Rodriguez-Salarichs, R.
Matesanz, A. Kanakkanthara, S. Wakefield, J. A. Singh, J. Jimenez-
Barbero, P. Northcote, J. H. Miller, J. A. Lopez, E. Hamel, I. Barasoain, K.
Altmann, F. J. Díaz, Chem. Biol. 2012, 19, 686 .
Manuscript received: September 27, 2018
Revised manuscript received: November 29, 2018
Version of record online: ■■■, ■■■■
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
ChemPlusChem 2019, 84, 98–102
www.chempluschem.org
102
© 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim