Influence of temperature for the azide displacement in benzodiazepine derivatives: Experimental and DFT study of competing SN1, SN2 and double SN2 reaction pathways

Article abstract of DOI:10.1016/j.tetlet.2021.152937

Molecules carrying the azido functionality are widely used as intermediates in medicinal chemistry. Pyrrolobenzodiazepines are molecules with a broad range of interesting pharmacological properties. Due to the wide importance of azido-PBD as a key interme

Full text of DOI:10.1016/j.tetlet.2021.152937

Tetrahedron Letters 68 (2021) 152937
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Influence of temperature for the azide displacement in benzodiazepine
derivatives: Experimental and DFT study of competing SN1, SN2 and
double SN2 reaction pathways
Tereza Cristina Santos Evangelista a, Maicon Delarmelina b,c, Dinesh Addla a, Rafael A. Allão d,
Carlos Roland Kaiser a, José Walkimar de M. Carneiro b, Floriano Paes Silva-Jr e, Sabrina Baptista Ferreira a,
⇑
a LaSOPB – Laboratório de Síntese Orgânica e Prospecção Biológica, Instituto de Química, Universidade Federal do Rio de Janeiro, Rio de Janeiro, RJ 21941-909, Brazil
b Instituto de Qu´ımica, Universidade Federal Fluminense, Niterói, Rio de Janeiro 24020-141, Brazil
c School of Chemistry, Cardiff University, Main Building, Park Place, Cardiff CF10 3AT, United Kingdom
d Instituto de Química, Universidade Federal do Rio de Janeiro, Rio de Janeiro, RJ 21941-909, Brazil
e LaBECFar – Laboratório de Bioquímica Experimental e Computacional de Fármacos, Instituto Oswaldo Cruz, Fundação Oswaldo Cruz, Rio de Janeiro, RJ 21.040-360, Brazil
a r t i c l e i n f o
a b s t r a c t
Article history:
Molecules carrying the azido functionality are widely used as intermediates in medicinal chemistry.
Pyrrolobenzodiazepines are molecules with a broad range of interesting pharmacological properties.
Due to the wide importance of azido-PBD as a key intermediate in many synthetic pathways, in this work
we have investigated the effect of a variety of experimental parameters on the mechanism of nucleophilic
substitution reaction for introducing an azide group into the PBD scaffold. This study was carried out by
combined experimental and DFT approaches, showing that SN2 and double SN2 pathways are possible
and highly dependent on the reaction temperature.
Received 7 November 2020
Revised 30 January 2021
Accepted 11 February 2021
Available online 23 February 2021
Keywords:
Azide
Pyrrolobenzodiazepines
Nucleophilic substitution reactions
Diastereoisomers
Temperature
Ó 2021 Elsevier Ltd. All rights reserved.
DFT
Introduction
11a, which provides a three-dimensional conformation that allows
the molecules to fit perfectly in the minor groove of DNA [7].
Pyrrolo[2,1–c][1,4]benzodiazepines (PBD) (1) are a class of nat-
urally occurring molecules that can be isolated from Streptomyces
bacteria and that exhibit a broad spectrum of biological activities,
featuring its antitumor properties as DNA minor groove binding
agents. The first PBD molecule isolated was anthramycin (2) [1].
After that, a number of PBD molecules have been isolated from nat-
ural sources and efforts have been applied to synthesize new PBD
derivatives that also presents pharmacological and biological prop-
erties, including anticancer, binding to GABAA receptor and forma-
tion of antibody-drug conjugates (ADCs) [2–6].
PBD molecules consist of a tricyclic structure, including an aro-
matic A-ring, a diazepine B-ring and a pyrrolidine C-ring (3), differ-
ing from one another specially by substitutions on the A and C
rings, and by the degree of unsaturation of the pyrrolidine ring
(Fig. 1). All naturally occurring PBD have S configuration at C-
Concerning the structure activity relationships regarding PBD
derivatives as anticancer agents, studies have shown that modifi-
cations at the substituents of the C-2 position can significantly
alter the DNA-binding ability and the cytotoxicity of the molecules.
For example, investigations involving anthramycin showed that
the removal of its C-2 side chain led to a strong decline of its
DNA binding capacity [8,9]. This effect is due to the location along
the minor groove of the extended side chains at C-2. Once accom-
modated along the minor groove, the groups of the side chain can
interact with the minor groove floor through hydrogen bonds and
Van der Waals interactions [1]. Those interactions stabilize the
adduct and enhance the DNA binding affinity leading to higher
cytotoxic potency. The comprehension of these interactions has
helped researchers to design novel PBD analogues including mono-
mers and dimers with modifications at the C-2 position, with
enhanced DNA binding ability and cytotoxicity (Fig. 1) [2,3].
Therefore, we hypothesized that the addition of a second phar-
macophore nucleus as a substituent to the C-2 position in the PBD
scaffold, would entail increased DNA binding affinity and therefore
⇑
Corresponding author.
E-mail address: sabrinab@iq.ufrj.br (S.B. Ferreira).
https://doi.org/10.1016/j.tetlet.2021.152937
0040-4039/Ó 2021 Elsevier Ltd. All rights reserved.

Tereza Cristina Santos Evangelista, M. Delarmelina, D. Addla et al.
Tetrahedron Letters 68 (2021) 152937
afford the azido-PDB derivative with inversion of configuration at
the nucleophilic center.
As presented in Scheme 1, in the first attempt to obtain the
azido-PBD derivative, we followed the procedure published by
Schultz and collaborators [17] where the reaction was carried out
at 55 °C in DMF for 36 h. The product obtained was recrystallized
from acetone, affording colorless crystals. The azide displacement
went exclusively with inversion of configuration at C-2 leading to
one diastereomer as confirmed by NMR elucidation and crystal
structure. In order to optimize the reaction, we decided to apply
different conditions to the azide displacement, as described by
Rault and collaborators [18] that performed the reaction at
100 °C in DMF for 5 h. However, it was not observed a 100% inver-
sion of configuration at C-2 under these conditions, leading to a
mixture of the two possible azido-PBD diastereomers. Such a result
could imply that a change in mechanism is occurring at different
temperatures.
In the interest of understanding this phenomenon, different
reaction conditions were investigated, including the effect of sol-
vents, leaving groups, temperature and reaction times, as pre-
sented in Table 1. The reactions were performed using the same
ratio between the reactants (1 equiv.) and NaN3 (5 equiv.) accom-
panied by TLC (eluent: ethyl acetate) until all the reactant was con-
sumed. The results obtained for the different conditions are
presented in Table 1, including the yield and the diastereomeric
ratios obtained in the cases where the reactions led to the mixture.
The ratio was assigned by 1H NMR analysis, as shown in Fig. 2.
Concerning the solvent choices, the reactions were carried out
in DMF and formamide. DMF is a polar aprotic solvent, which
can improve the efficiency of the nucleophile in an SN2-type reac-
tion while formamide, besides being a polar solvent, also show the
ability to make hydrogen bonds, which is not usually ideal in an
SN2-type reaction. However, it was not observed significant
changes concerning the diastereomeric ratio obtained for the dif-
ferent reactions, indicating that the solvent is not important in
changing the mechanistic pathway in the case of azido-PBD forma-
tion. The only difference we could observe between the two sol-
vents is the isolated yield, which is higher when using DMF for
all reaction conditions (1–3 and 7–9). The reaction time also do
not seem to have an impact on the mechanism of the reaction,
although as expected longer reaction times were observed for
lower temperatures.
Fig. 1. The pyrrolobenzodiazepine (PBD) scaffold and the structure of pharmaco-
logically active PBDs.
pharmacological activity. Triazoles [10], tetrazoles [11], amides
[12,13] and its derivatives have shown anticancer activity when
linked with another pharmacophore function. These hybrid mole-
cules, containing two or more pharmacophore units, can potential-
ize the anticancer effect of the PBD core, reduce side effects and
overcome other drawbacks associated with PBD derivatives.
Results and discussion
To obtain the triazole and tetrazole heterocycles mentioned
above, a suitable key intermediate is the azido group. The azido
functionality has been successfully applied in carbohydrate chem-
istry [14], synthesis of heterocycles and peptide chemistry [15].
The azido group can be obtained via different reactions, such as
substitution or addition via the insertion of the N3 group, diazoti-
zation, diazo transfer, cleavage of triazines and azide rearrange-
ment. Among those, the most common method applied for the
synthesis of alkyl azides is the classic nucleophilic substitution
reaction [16].
The reaction was performed at different temperatures based on
previously reported works [16,17]. As discussed before, when the
reaction was carried out at 55 °C, just one product was observed,
and it was the one corresponding to a 100% inversion of configura-
tion at C-2. The same result was observed when applying different
leaving groups and solvents, confirming a substitution by direct
displacement (SN2) mechanism (Entries 3, 6, 9 and 12 in Table 1).
However, once the temperature was increased to 100 °C, the for-
mation of a mixture of diastereomers started to be observed, where
the main diastereomer formed was still 9b. The same results were
obtained for all the reactions performed at 100 °C, regardless of
which leaving group or solvent was used, with diastereomeric
ratios similar to one another (Entries 2, 5, 8 and 11). When the
reactions were performed at 120 °C, the mixture of diastereomers
was observed for all the entries. However, the diastereomeric ratio
of the products changed
In the present work, the azido-PDB derivative was synthesized
according to Scheme 1, starting with a condensation reaction
between isatoic anhydride and trans-hydroxy-L-proline, followed
by the conversion of the hydroxyl functionality into a better leav-
ing group. After that, the desired SN2 reaction with NaN3 would
drastically, leading to the isomer 9a being the major one for en-
tries 1, 7 and 10, except for entry 4, where the ratio is close to
50:50.
X-ray crystallography was employed to unequivocally charac-
terize stereochemistry and conformation of one of the obtained
azido-PBD diastereomers. A single crystal of compound 9b was
obtained by slow evaporation of acetone from the solution of pure
diastereomer. 9b crystallizes in the P21 space group containing one
Scheme 1. Synthesis of the azido-PBD from isatoic anhydride and trans-hydroxy-L-
proline.
2

Tereza Cristina Santos Evangelista, M. Delarmelina, D. Addla et al.
Tetrahedron Letters 68 (2021) 152937
Table 1
Evaluation of different reaction conditions for the azide displacement reaction.
Entry
LGa
Temperature
RTb
Solvent
1H NMR ratio Isomer 9a: 9b
Yield (%)
1
2
3
4
5
6
7
8
Tosyl
Tosyl
Tosyl
Tosyl
Tosyl
Tosyl
Mesyl
Mesyl
Mesyl
Mesyl
Mesyl
Mesyl
120 °C
100 °C
55 °C
120 °C
100 °C
55 °C
120 °C
100 °C
55 °C
120 °C
100 °C
55 °C
12 h
5 h
36 h
12 h
5 h
36 h
12 h
5 h
36 h
12 h
5 h
DMF
DMF
DMF
Formamide
Formamide
Formamide
DMF
DMF
DMF
Mixture (59:41)
Mixture (9:91)
(0:100)
Mixture (49:51)
Mixture (15:85)
(0:100)
Mixture (63:37)
Mixture (36:64)
(0:100)
Mixture (65:35)
Mixture (9:91)
1 diastereoisomer
94
85
85
29
45
55
82
62
47
50
46
50
9
10
11
12
Formamide
Formamide
Formamide
36 h
a
Leaving Group.
Reaction time.
b
independent molecule per unit cell. The asymmetric unit is shown
in Fig. 3, and the stereochemistry of C2 confirms that the reaction
occurred through SN2 mechanism. The crystal packing is stabilized
by intermolecular hydrogen bonds between oxygen atom O1 and
amine group (N2-H7). Each molecule shows intermolecular inter-
‘‘double” SN2, as shown in Fig. 4. As observed during the experi-
mental investigations, the SN2 pathway (Fig. 4) presented the low-
est free energy barrier from 8b (27.2 kcal molÀ1), which clearly
justify the exclusive formation of product 9b at low reaction tem-
peratures. However, with increasing temperature, a mixture of
products was observed which raised the hypothesis of a possible
change in the observed mechanisms, either via a SN1 mechanism
or via the so-called ‘‘double” SN2 as proposed by Szabó and Czakó
[18,19].
Firstly, we investigated the energy cost to form a carbocation
species in the reaction mixture by elimination of the substituent
group at C-2 (either OMsÀ or N3À). As described in previous works,
computing the energy barriers for formation of such intermediate
is, in general, a good approximation to estimate the rate constant
actions with two other adjacent molecules generating
a
supramolecular chain along the crystallographic direction b (see
Fig. S1 for details).
At the final stage of this investigation, DFT calculations have
been applied to provide deeper insight into the effect of tempera-
ture over the alternative reaction mechanisms for OMsÀ displace-
ment by azide in the investigated benzodiazepine derivatives.
Using 8b as substrate for this investigation, the energy profile of
three competing reaction channels were computed, SN1, SN2 and
Fig. 2. 1H NMR spectra from aliphatic region used for the determination of the diastereomeric ratio. (CDCl3, 400.13 MHz).
3

Tereza Cristina Santos Evangelista, M. Delarmelina, D. Addla et al.
Tetrahedron Letters 68 (2021) 152937
On the other hand, the alternative ‘‘double” SN2 process pre-
sents much lower free energy barrier than that to form the above-
mentioned carbocation intermediate. Starting from intermediate
9b, which has relative free energy around À21.0 kcal molÀ1 com-
pared to 8b, the activation Gibbs free energy to form the product
with a second inversion via a second SN2 reaction is 37.6 kcal molÀ1
.
Although this is a relatively high energy, requiring heat to be sur-
mounted, it is lower than the activation Gibbs free energy for the
SN1 process. As a result, the most favorable process identified here
for OMsÀ substitution in 8b is an initial SN2 reaction, with a Gibbs
free energy barrier of 27.2 kcal molÀ1, followed by a ‘‘double” SN2
step presenting a free energy barrier of 37.6 kcal molÀ1. These
results are consistent with the experimental observation of enan-
tioselective formation of 9a and 9b at distinct temperatures. For-
mation of 9b occurs with the lowest energy barrier. As this is a
stable intermediate, its lifetime is enough to dissipate energy. At
low temperature this is the main product formed. Increasing the
temperature 9b can be converted into 9a by reaction with excess
NÀ3 in the reaction mixture. Furthermore, the slightly higher stabil-
ity of 9a regarding 9b is also consistent with the observation of
increased amount of 9a when longer reaction time is used. This
observation was investigated experimentally by treating diastere-
omerically pure 9b with sodium azide in DMF at 120 °C overnight.
This resulted in 63:37 9a:9b ratio, confirming the preference for a
‘‘double” SN2 pathway at higher temperatures.
Fig. 3. ORTEP view of the asymmetric unit of 9b with thermal ellipsoids drawn at
30% of probability.
for conventional SN1 reactions [20]. As it can be seen in Fig. 4, such
carbocation would be a highly energetic intermediate, presenting
relative free energy higher than 45 kcal molÀ1 in comparison to
the isolated reactants. The possible stabilization of 10 by interac-
tions with nucleophile and leaving group was also tested by the
construction of a [OMsÀ 10 N3À] complex, however the resulting
decrease in energy was smaller than 2 kcal molÀ1 when the anions
were kept 3.5 Å away from the carbocationic center.
Fig. 4. Computed energy profile for alternative SN1, SN2, and ‘‘double” SN2 reaction pathways for OMsÀ displacement by azide in 8b.
4

Tereza Cristina Santos Evangelista, M. Delarmelina, D. Addla et al.
Tetrahedron Letters 68 (2021) 152937
Conclusions
Appendix A. Supplementary data
This study showed that small changes in temperature led to
retention of configuration in the azide displacement of PBD deriva-
tives. This is in contrast to previously reported literature, which
describes 100% inversion of configuration at the substitution cen-
ter. 1D and 2D-NMR techniques were used to identify products
and follow the changes in the reaction over time. Furthermore,
X-ray crystallography was employed to unequivocally characterize
stereochemistry and conformation of one of the obtained azido-
PBD diastereomers. While products with inversion of configuration
are the commonly expected outcome for this reaction, we have
shown that small changes in the reaction temperature can lead
to the diastereomer with retention of configuration. Results of
DFT calculations have provided a theoretical insight into the effect
of the temperature over the reaction mechanism, showing that a
‘‘double” SN2 mechanism takes place at higher temperatures,
which was confirmed experimentally. Our results can be useful
when evaluating the pharmacological profile of molecules pre-
pared from key azido-PBD intermediates, where molecules with
high enantiomeric purity are required. In addition, the ability to
obtain both diastereomers is also an advantage, once it allows for
evaluating their pharmacological properties, which can vary
considerably.
Supplementary data (detailed experimental procedures, charac-
terization data (1H, 13C NMR and HRMS), copies of spectra and fur-
ther information concerning DFT studies and X-ray are provided.
CCDC 2041317 contains the supplementary crystallographic data
as CIF file for this paper. This data can be obtained free of charge
from The Cambridge Crystallographic Data Centre via www.ccdc.-
cam.ac.uk/structures) to this article can be found online at
https://doi.org/10.1016/j.tetlet.2021.152937.
References
[1] J. Mantaj, P.J.M. Jackson, K.M. Rahman, D.E. Thurston, Angew. Chem. Int. Ed. 56
(2017) 462–488.
[2] M.S.K. Sutherland, R.B. Walter, S.C. Jeffrey, P.J. Burke, C. Yu, H. Kostner, I. Stone,
M.C. Ryan, D. Sussman, R.P. Lyon, W. Zeng, K.H. Harrington, K. Klussman, L.
Westendorf, D. Meyer, I.D. Bernstein, P.D. Senter, R.B. Dennis, J.G. Drachman, A.
McEarchem, Blood 122 (2013) 1455–1463.
[3] A. Mieczkowski, M. Psurski, M. Baginski, B. Bieszczad, M. Mroczkowska, M.
Wilczek, J. Czajkowska, D. Trzybinski, K. Wozniak, J. Wietrzyk, Bioorg. Med.
Chem. Lett. 28 (2018) 618–625.
[4] A.C. Araújo, A.P. Rauter, F. Nicotra, C. Airoldi, B. Costa, L. Cipolla, J. Med. Chem.
54 (2011) 1266–1275.
[5] G.C. Kemp, A.C. Tiberghien, N.V. Patel, F. D’Hooge, S.M. Nilapwar, L.R. Adams, S.
Corbett, D.G. Williams, J.A. Hartley, P.W. Howard, Bioorg. Med. Chem. Lett. 27
(2017) 1154–1158.
[6] A.C. Tiberghien, J.N. Levy, L.A. Masterson, N.V. Patel, L.R. Adams, S. Cobert, D.G.
Williams, J.A. Hartley, P.W. Howard, ACS Med. Chem. Lett. 7 (2016) 983–987.
[7] J.A. Hartley, Expert. Opin. Investig. Drugs 20 (2011) 733–744.
[8] M.L. Kopka, D.S. Goodsell, I. Baikalov, K. Grzeskowiak, D. Cascio, R.E. Dickerson,
Biochemistry 33 (1994) 13593–13610.
Declaration of Competing Interest
[9] S.B. Hoewitz, S.C. Chang, A.P. Grollman, A.B. Borkovec, Science 174 (1971) 159–
161.
[10] Z. Xu, S.-J. Xhao, Y. Liu, Eur. J. Med. Chem. (2019) 111700.
[11] L. Lauko, P.H.J. Kouwer, A.E. Rowan, Anti-Cancer Agents Med. Chem. 17 (2017)
1856–1868.
[12] Y. Lu, Z. Wang, C.-M. Lib, J. Chen, J.T. Dalton, W. Li, D.D. Miller, Bioorg. Med.
Chem. 18 (2010) 477–495.
The authors declare that they have no known competing finan-
cial interests or personal relationships that could have appeared
to influence the work reported in this paper.
[13] T. Sreelatha, S. Kandhasamy, R. Dinesh, S. Shruthy, S. Shweta, D. Mukesh, D.
Karunagaran, R. Balaji, N. Mathivanan, Bioorg. Med. Chem. Lett. 24 (2014)
3647–3651.
[14] S.B. Ferreira, A.C.R. Sodero, M.F.C. Cardoso, E.S. Lima, C.R. Kaiser, F.P. Silva-Jr, V.
F. Ferreira, J. Med. Chem. 53 (2010) 2364–2375.
Acknowledgements
[15] I.F. Silva, P.R.C. Martins, E.G. Silva, S.B. Ferreira, V.F. Ferreira, K.R.C. Costa, M.C.
Vasconcellos, E.S. Lima, F.C. Da Silva, Med. Chem. 9 (2013) 1085–1090.
[16] Y. Bi, P.G. Schultz, Bioorg. Med. Chem. Lett. 6 (1996) 2299–2300.
[17] A.C. Gillard, M. Alkhader, S. Rault, Heterocycl. Commun. 2 (1996) 409–414.
[18] I. Szabo, G. Czako, Nature Commun. 6 (2015) 5972.
[19] J. Xie, W.L. Hase, Science 352 (2016) 32–33.
The authors acknowledge LDRx-UFF for using its crystallo-
graphic facilities. The authors would like to acknowledge CNPq
(grants no. 309095/2018-2 and 306836/2017-3), FAPERJ (grant
no. E-26/202.743/2019). T.C.S.E. would like to acknowledge Coor-
denação de Aperfeiçoamento de Pessoal de Nível Superior – Brasil
(CAPES) for a Ph.D. fellowship.
[20] N. Streidl, B. Denegri, O. Kronja, H.A. Mayr, Acc. Chem. Res. 43 (2010) 1537–
1549.
5