Structural study of bioisosteric derivatives of 5-(1H-indol-3-yl)-benzotriazole and their ability to form chalcogen bonds

Recently, interest in the isosteric replacement of a nitrogen atom to selenium, sulfur or oxygen atoms has been highlighted in the design of potential inhibitors for cancer research. In this context, the structures of 5-(1H-indol-3-yl)-2,1,3-benzotriazole derivatives [5-(1H-indol-3-yl)-2,1,3-benzothiadiazole (bS, C14H9N3S) and 5-(1H-indol-3-yl)-2,1,3-benzoxadiazole (bO, C14H9N3O)], as well as a synthesis intermediate of the selenated bioisostere [5-[1-(benzensulfonyl)-1H-indol-3-yl]-2,1,3-benzoselenadiazole (p-bSe, C20H13N3O2SSe)] were determined using single-crystal X-ray diffraction (SCXRD) analyses.


Chemical context
Isosteric replacement is a common strategy in drug design to modulate the physicochemical properties of potential inhibitors. In 2021, Kozlova and co-workers (Kozlova et al., 2021a) highlighted a series of bioisosteric derivatives acting as potential new inhibitors of the protein hTDO2, a therapeutic target in cancer research. These new molecules differ in the replacement of the central atom of benzotriazole by an oxygen, sulfur or selenium atom ( Fig. 1). At this time, these inhibitors have not yet been crystallized or structurally char-acterized. In this context, the present work provides a structural characterization of the inhibitors described by Kozlova et al. (2021b) completed by ab initio calculations for their conformational characterization.
The contribution of an oxygen, a sulfur or a selenium atom instead of a nitrogen affects the ability of these inhibitors to participate in the formation of chalcogen bonds (Vogel et al., 2019). In particular, in these compounds, oxygen, sulfur and selenium atoms could act as chalcogen-bond donors. In recent years, the importance of chalcogen bonds in the stability and folding of proteins as well as their interaction with ligands has been highlighted by numerous investigations (Newberry & Raines, 2019;Kříž et al., 2018;Iwaoka et al., 2001;Iwaoka & Babe, 2015;Burling & Goldstein, 1992). In this article, the potential ability of the compounds to interact with aromatic groups, by chalcogen-interactions (Aakeroy et al., 2019), was revealed by the crystallization of 5-[1-(benzensulfonyl)-1Hindol-3-yl] À2,1,3-benzoselenadiazole. Therefore, the effect of bioisosteric replacement on the ability to form chalcogen bonds has been studied by ab initio calculated electrostatic potential maps. This interesting series could be the starting point for the study of the effect of chalcogen interaction on protein stability and affinity.

Supramolecular features
In the structure of p-bSe, a synthesis intermediate of the selenated bioisostere of 5-(1H-indol-3-yl)benzotriazole, the benzenesulfonyl contributes to the stabilization of the crystal Ellipsoid plots with atom labeling for (a) p-bSe (b) bS and (c) bO. Displacement ellipsoids are drawn at the 50% probability level.

Figure 3
Two perpendicular views (I. and II.) of the two mirror images observed in the crystal packings for (a) p-bSe (b) bS and (c) bO. Superposition of the mirror images with respect to the indole group. Torsion angles are AE168.3 (2) for p-bSe, AE36.9 (2) for bS and AE146.7 (2) for bO.

Quantum ab initio studies of the bioisosteric substitution effect
As mentioned previously, the different derivatives vary mainly in their ability to interact through chalcogen bonds. In order to characterize these differences in depth, quantum mechanics studies have been conducted. First, the presence of a -hole in the electron density was studied by means of electrostatic maps. Analysis indicates that the oxygen in benzoxadiazole . The difficulty in crystallizing bSe (while p-bSe crystallized readily in THF) could be explained by the absence of the protecting group (benzenesulfonyl) in bSe. Indeed, the benzenesulfonyl group in p-bSe is electron-rich and acts as a well-oriented chalcogen-bond acceptor in p-bSe. Secondly, in order to determine the effect of the substitution on the flexibility of the derivatives, conformational scans were performed around the torsion angle formed between the indole ring and the benzodiazole part (T1). As presented in   Conformational scans and associated torsion angles calculated with Gaussian16a (!B97XD, 6-31+G*).   Computed electrostatic potential (EPS) surfaces and associated molecular structures (a) bSe (b) p-bSe. The EPS color scale ranges from +8.160 volts to À5.440 volts.
[AE36.9 (2) ] and bO [AE146.7 (2) ] are consistent with the energy minima determined by ab initio calculations with a relative deviation lower than 10%. Although the bioisosteric character of the flexibility is retained, two differences are observed between the bSe molecule and the bS and bO molecules. The first one is the energy at a torsion angle of 180 , which is lower for the molecule of bSe while it is slightly higher at 0 . The second one is a small shift observed between the angle associated with the energy minima of the bSe molecule and the bS and bO molecules.
The same calculations were performed for the protected molecule (p-bSe). The energy profile associated with the rotation around T1 in p-bSe is similar to those determined for bS and bO. There is a small shift of the values of the angle corresponding to the energy minimum with respect to bSe. This shift may be due to the protecting group that causes steric hindrance in p-bSe. The energies corresponding to minima A and C are higher than the energies for minima B and D, involving two local minima and a preference in the conformers. The maxima of the energies between A/B and C/D are also lower than in the case of the other compounds, supporting this hypothesis. In the case of p-bSe, the torsion angle observed in the crystallographic structure [AE168.3 (2) ] corresponds to an energy maximum on the energy profile associated with the rotation around T1 determined by ab initio calculations. The quasi-planarity of the molecule observed in the crystallographic structure could be encouraged by the formation of chalcogen bonds stabilizing this conformation of p-bSe.  Table 2. The search for a benzothiadiazole fragment resulted in 34 hits. Chalcogen bonds but no chalcogen-interactions are observed in these hits. The search for a benzoxadiazole fragment resulted in 24 hits but no chalcogen interactions are observed with the oxygen atom of benzoxadiazole as a chalcogen donor.

Refinement
Crystal data, data collection and structure refinement details are summarized in Table 3. In all of the structures, hydrogen atoms were placed in calculated positions and refined using a riding model [C-H bond length of 0.93 Å , with U iso (H) = 1.2U eq (C)]. In the structure of 5-(1H-indol-3-yl)-2,1,3-benzothiadiazole, bS, the hydrogen on the nitrogen atom in the indole group was refined without constraint and the refined N-H distance is 0.78 (2) Å .

Quantum ab initio methodology
All the molecules investigated in the study (bS, bO, bSe, p-bSe) were optimized starting from the crystal coordinates using the density functional method (DFT) with the exchange-  Table 2 Chalcogen bonds (Å , ) observed in benzoselenadiazole fragments in the CSD.

CCDC refcode
Se label Atoms of the system SeÁ Á Ácentroid distance N-SeÁ Á Ácentroid angle QIBQUQ a Se1 C18-C23 3.3676 (7) 147.08 (4) QIBQUQ a Se1 C12-C17 3.0597 (8) 172.30 (5) (4) 164.3 (4) correlation functional !B97XD and the 6-31+G* basis set. Because we were not able to crystallize the bSe compound, this molecule was created by substitution of the sulfur atom for a selenium atom from the coordinates of the bS molecule. The optimizations were performed with Gaussian16a (Frisch et al., 2016) the in gas phase. The electrostatic potential was calculated from the SCF-type density and was sliced by making 80 cubic points evenly distributed on a rectangular grid automatically generated by Gaussian16a. The resulting maps were visualized using DrawMol (Liegeois, 2021). For the conformational scans, the optimized structures were analyzed using relaxed scans around the torsion angle (T1) formed between the indole ring and the benzodiazole part from 0 to 360 by steps of 20 . The resulting conformations close to an energy minimum were extracted and refined by a new optimization at the same level of approximation. The preparation of the input files, as well as the visualization of the results was performed with the DrawMol and DrawSpectrum suite of programs (Liegeois, 2021). The graphs were drawn with the program Prism from GraphPad (one-way ANOVA followed by Dunnetts multiple comparisons test, Prism version 8.0.0 for Windows; GraphPad, 2021).

Special details
Geometry. All esds (except the esd in the dihedral angle between two l.s. planes) are estimated using the full covariance matrix. The cell esds are taken into account individually in the estimation of esds in distances, angles and torsion angles; correlations between esds in cell parameters are only used when they are defined by crystal symmetry. An approximate (isotropic) treatment of cell esds is used for estimating esds involving l.s. planes. Refinement. Structures were solved by the dual-space method of ShelXT (Sheldrick, 2015a) within Olex 2 (Dolomanov et al., 2009). Structures were refined by the least-squares method implemented in SHELXL (Sheldrick, 2015b) within ShelXle (Hübschle et al., 2011). Structures and crystal packings were visualized using Mercury (Macrae et al., 2020).

5-[1-(Benzenesulfonyl)-1H-indol-3-yl]-2,1,3-benzothiadiazole (bS)
Crystal data Special details Geometry. All esds (except the esd in the dihedral angle between two l.s. planes) are estimated using the full covariance matrix. The cell esds are taken into account individually in the estimation of esds in distances, angles and torsion angles; correlations between esds in cell parameters are only used when they are defined by crystal symmetry. An approximate (isotropic) treatment of cell esds is used for estimating esds involving l.s. planes. Refinement. Structures were solved by the dual-space method of ShelXT (Sheldrick, 2015a) within Olex 2 (Dolomanov et al., 2009). Structures were refined by the least-squares method implemented in SHELXL (Sheldrick, 2015b) within ShelXle (Hübschle et al., 2011). Structures and crystal packings were visualized using Mercury (Macrae et al., 2020).

Fractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (Å 2 )
x y z U iso */U eq S1 1.08880 (6) 0.65005 ( where P = (F o 2 + 2F c 2 )/3 (Δ/σ) max < 0.001 Δρ max = 0.14 e Å −3 Δρ min = −0.16 e Å −3 Special details Geometry. All esds (except the esd in the dihedral angle between two l.s. planes) are estimated using the full covariance matrix. The cell esds are taken into account individually in the estimation of esds in distances, angles and torsion angles; correlations between esds in cell parameters are only used when they are defined by crystal symmetry. An approximate (isotropic) treatment of cell esds is used for estimating esds involving l.s. planes.