Racemic-Benzimidazolyl Pentafluorobenzyl Sulfoxide

Abstract

As a part of our research on the presence of conglomerates among the aryl benzyl sulfoxides, racemic-benzimidazolyl pentafluorobenzyl sulfoxide was synthesised, and its crystal structure was determined by a single crystal X-ray diffraction experiment. The main interactions building up the crystal structure were recognised and compared with those of similar compounds. Since the crystal structures of racemic and enantiopure benzimidazolyl pentafluorobenzyl sulfoxides are different, the presence of a conglomerate is excluded in the present case.

1. Introduction

Heterocyclic moieties are present in more than 85% of the biologically active compounds [1]. Among them, nitrogen-containing heterocycles are the most frequent [1] and have great relevance in medicinal chemistry [1]. For example, imidazole rings are present in anti-bacterial, anti-inflammatory, anti-viral, and anti-tumour agents, as well as in many other drugs [2]. Intermediates containing the benzimidazole ring have received recent attention as anti-tumour agents [3], even if the most famous benzimidazole-containing drugs are a family of structurally related compounds used against gastric diseases, among which the omeprazole is the most famous and the most successful from a commercial point of view [4,5,6].

Omeprazole is basically a benzimidazolyl sulfoxide and exists in two different enantiomers, due to the chirality connected with the sulfinyl group. Nowadays, the (S)-omeprazole is preferred and sold worldwide as treatment for gastric diseases [4,5,6].

The demand for a single enantiomer drug has prompted many researchers to propose asymmetric syntheses of this drug [7]. The first successful process leading to (S)-omeprazole was proposed by Astra-Zeneca [8,9]. This process is asymmetric oxidation of the corresponding sulfide with hydroperoxides in the presence of a complex between titanium and (S,S)-diethyl tartrate [8,9]. The process resembles the original work by Kagan on the asymmetric oxidation of sulfides in the presence of a complex between titanium and diethyl tartrate [10,11], but it shows significant differences [8,9]. Notwithstanding the research performed by many groups in the world, this process is still valid and used in the industrial production of the drug [7]. The main alternatives to the Astra-Zeneca procedure (different catalysts or bio-technologically processes) were critically discussed and evaluated [7] and no concrete alternative to asymmetric oxidation has emerged [7].

A different approach was reported by Kellogg et al. [12]. In fact, they found a conglomerate [13] in a salt connected with the omeprazole. Conglomerates are “racemic mixtures of two enantiomers with each crystal being made up of a single enantiomer” [13]. Conglomerates are very interesting from an industrial point of view [13] because, at least in principle, a separation of enantiomers could be set up without resorting to asymmetric synthesis [14].

Considering the enormous interest in the (S)-omeprazole business, it is surprising that the work of Kellogg et al. has been overlooked, and that the investigation on the presence of conglomerates in the family of these highly profitable compounds has not been conducted in depth.

2. Results

In our work on the asymmetric synthesis of sulfoxides [15,16], we extended our highly successful, easy and straightforward synthesis of aryl benzyl sulfoxides [15,16] to benzimidazolyl sulfoxides [17], models of omeprazole-like molecules. Our synthetic work was accompanied by investigations on the crystal structures of the sulfoxides [18,19,20], through which we provided original contributions [20,21,22].

In our crystallographic studies, we found that the crystal structure of the racemic-benzimidazolyl benzyl sulfoxide 1 (Figure 1) reported by Loginova et al. [23] is identical to that of the enantiopure (R)-1 that was synthesised by us [17].

A close inspection of the crystallographic report revealed that the authors did not realise that their “racemic” compound was composed of single enantiomer crystals, solved in a Sohncke space group, as observed for chiral compounds; therefore it is a conglomerate. Likewise, we compared the crystal structure of the racemic form of the N-methyl derivative of 1, that is the 2-benzylsulfinyl-1-methyl-1H-Benzo[d]imidazole 2 (Figure 1), obtained by us [17], with the crystal structure of the enantiopure (S)-2 compound synthesised in different research [24].

We found that the crystal structure of our “racemic”-2 refers only to the (R)-enantiomer, solved in a Sohncke space group, thus showing that we had isolated crystals composed of only the (R)-enantiomer, and thus a conglomerate [17]. In our research, we found two further conglomerates between aryl benzyl sulfoxides not bearing heterocyclic moieties [22]. It appears that a good number of conglomerates, compounds that are usually considered rare, are hidden among this type of sulfoxide.

Currently, predictions of the crystals that give rise to conglomerates are impractical. Even a simple listing of hidden conglomerates in the Cambridge Structural Database (CSD) cannot be performed automatically, but must be supplemented by a manual search [25,26]. New developments in this field will only be achieved if new further crystal structures of both racemic and enantiopure pairs are studied and a dedicated database of conglomerates is built [25].

At this point, we consider it of interest to test another benzimidazolyl sulfoxide deriving from our research, to check the eventual presence of conglomerates. The sulfoxide under investigation is the benzimidazolyl pentafluorobenzyl sulfoxide 3, already synthesised in our work [17]. We compared many racemic and enantiopure aryl benzyl sulfoxides [22], but we have not found a single case of conglomerate in which the pentafluorophenyl group is present.

Racemic-benzimidazolyl pentafluorobenzyl sulfoxide 3 (Figure 1) was synthesised [17] by standard m-CPBA oxidation. The mp of the racemic compound was 189–190 °C, similar to the value of the mp of the (R)-3 (190–192 °C) already reported [17].

At this point, we decided to determine the crystal structure of racemic-benzimidazolyl pentafluorobenzyl sulfoxide 3 and to compare it with the crystal structure of (R)-benzimidazolyl pentafluorobenzyl sulfoxide 3 determined by us (absolute structure parameter 0.065) [17].

The crystal structure of racemic-benzimidazolyl pentafluorobenzyl sulfoxide 3 was determined by a single crystal X-ray diffraction experiment (see in the Section 3). An ORTEP plot is represented in Figure 2. Crystal data and structure refinement parameters are collected in Appendix A of the present paper. The packing plot is represented as Figure S1 in the Supplementary Materials Section.

The analysis of the crystal structure of 3 was accompanied also by the energetic calculations performed with the aid of CrystalExplorer21 [27,28] (see Section 3) in accordance with the approach that we have already applied in our recent articles [17,22]. The outputs of the energetic calculations performed are collected in the Supplementary Materials in Tables S1 and S2.

The energetic calculations yield a lattice energy for the racemic-3 of −171.2 KJ/mol, almost 10 KJ/mole more stable than the calculation performed for (R)-3 (−161.8 KJ/mol [17]), as it occurs frequently in the comparison between racemic and enantiopure compounds [22,29]. On the other hand, differences in the melting points cannot be used as a test to infer information on the lattice energies [29].

From a comparison of the two crystal structures, it is evident that the crystal structure of racemic-3 is different from the crystal structure of (R)-3. Thus, racemic-3 is not a conglomerate. As confirmation, the space group of (R)-3 is P2₁, which is a Sohncke space group; in the case of racemic-3, the space group is P2₁/c, a space group for racemic compounds, in which the (R)- and (S)-enantiomers are connected by inversion centres or glides.

The main feature of compounds 1 and 3 is the strong hydrogen bonding [30] due to the amino hydrogen atom acting as a donor. In fact, in the case of the crystal structure of N-methylated sulfoxide 2 [17], in which there is no amino hydrogen, the crystal structure is very different from those of 1 and 3.

It is interesting to observe that, in the case of compound 1, the amino hydrogen atom interacts with the nitrogen atom of a different molecule, this atom behaving as the acceptor. On the other hand, in the cases of racemic-3 and (R)-3, the amino hydrogen atom interacts with the sulfinyl oxygen atom, which behaves as the acceptor. In

Table 1. Angle and distances between atoms involved in the strong hydrogen bonding.

Compound	Angle N1-H1N-N1 (°)	H1N⋯N1 (Å)	N1⋯N1 (Å)	H1N-N1 (Å)
(R)-1	162(5)	2.07	2.885	0.84
	Angle N1-H1N-O1 (°)	H1N⋯O1 (Å)	N1⋯O1 (Å)	H1N-N1 (Å)
(R)-3	173(4)	2.05	2.820	0.77
rac-3	166(1)	2.01(2)	2.810(2)	0.82(2)

, the main characteristics of this strong hydrogen bonding are reported for compounds (R)-1, (R)-3 and racemic-3.

As can be seen, the characteristics of the strong hydrogen bonding in compounds in Table 1 are similar, notwithstanding their being fluorinated or not fluorinated, or racemic or enantiopure [17,22].

Based on the inspection of the calculations performed with the CrystalExplorer21 program (Supplementary Material, Tables S1 and S2), it is evident that the largest contribution to the lattice energy is due to the energy calculated for the interaction between the central molecule, with the closest molecule having a different chirality connected by this strong hydrogen bonding (Supplementary Material, Tables S1 and S2, first rows). Moreover, the nature of this interaction is clearly identified as strong hydrogen bonding due to the very high contribution of the electronic energy to the total energy (Supporting Information, Tables S1 and S2, first rows).

As represented in Figure 3, two of these strong hydrogen bondings between the amino hydrogen atoms with the sulfinyl oxygen atoms yield a head-to-tail assembly of two different enantiomers.

The pattern is indicated as R²₂(10) (a ring of 10 terms with two donors and two acceptors), according to the classification [31].

Another relevant interaction that contributes to the stabilisation of the crystal structure of racemic-3 is the stacking between the pentafluorophenyl group and the benzimidazolyl moiety, previously also discussed for the crystal structure of (R)-3 [17].

A further structural aspect that can be inferred from the present research is the choice between the anti- or gauche-conformation around the aryl carbon/sulfur/methylene carbon/aryl carbon sequence [22]. The non-fluorinated sulfoxide 1 is in a gauche-conformation (torsion angle 61.9(4)° [17]). On the other hand, the fluorinated sulfoxide 3 reported in this paper is in an anti-conformation (torsion angle 166(1)°). The switch gauche/anti between fluorinated and non-fluorinated aryl benzyl sulfoxides occurs frequently and was reported and interpreted in our recent paper [22]. The comprehension of different conformations could be a valuable tool in crystal structure prediction [32].

In summary, we have reported new interesting features, with a special focus on the crystallographic aspects, of the racemic-benzimidazolyl pentafluorobenzyl sulfoxide 3. For the case under investigation, the possibility of finding a further conglomerate in this family of sulfoxides is excluded. Moreover, from the results in the present and in the past research [22], the pentafluorophenyl moiety does not appear to favour the formation of conglomerates.

3. Materials and Methods

3.1. X-Ray Diffraction Experiments

A suitable single crystal was mounted on a Bruker D8 Venture automatic diffractometer (Karlsruhe, Germany) using monochromated Cu Ka radiation at a 219 K temperature. Structures were solved by direct methods (SHELXS) [33] and refined by full-matrix least-squares methods on F² for all reflections (SHELXL-2016) [33]. The intensities were corrected for Lorentz and polarization effects as well as for absorption [34]. Crystal data and structure refinement information are detailed in Appendix A.

Non-hydrogen atoms were refined anisotropically. Hydrogen atoms were placed in calculated positions with isotropic displacement parameters fixed at 1.2 times the U_eq of the corresponding carbon atoms. Amino hydrogen atoms were located by means of difference Fourier synthesis.

The crystallographic .cif file was deposited in the Cambridge Structural Database (CSD) with the number n. 2536423. These data can be obtained free of charge via http://www.ccdc.cam.ac.uk/structures/ (or from the CCDC, 12 Union Road, Cambridge CB2 1EZ, UK; Fax: +44 1223 336033; E-mail: deposit@ccdc.cam.ac.uk).

3.2. Molecular Pairwise and Lattice Energy Calculations

The CrystalExplorer21 program [27,28] was used to perform the lattice energy calculations starting from the coordinates recorded in the crystallographic .cif file. The procedure is based on the selection of a central molecule and the building up of a network of adjacent molecules within a 10 Å radius. The interaction energy of each molecule of the network with the central species is calculated [27,28]. Four different contributions (electronic, polarization, dispersion and repulsion energies) are determined for each interaction. Then, each contribution is suitably weighted and summed up to yield the energy value of each interaction [27,28]. Finally, all these energies are summed up to obtain the lattice energy. In the present case, in which Z′ = 2, the calculations are repeated for each molecule and the value of the lattice energy is evaluated as the average of the two calculations.