Theoretical Insight into Non-Covalent Complexes of Closo-Borate Anions [B_nH_n−1X]^y− with Glycine

Abstract

Non-covalent contacts play a significant role in binding between fragments in supramolecular assemblies. Understanding the non-covalent binding capabilities of closo-borate anions and their derivatives is a significant research challenge, due to their ability to interact with biomolecules. The present work was focused on the theoretical study of non-covalent complexes between glycine and closo-borate anions [B_nH_n−1X]^y− (X = H, NH₃, OH, SH, F; n = 10, 12; y = 1, 2). The main binding patterns between glycine and cluster systems were defined, and the effect of the exo-polyhedral substituent on the stability of non-covalent complexes was analysed. Complexes based on ammonium and hydroxy derivatives of closo-borate anions [B_nH_n−1X]^y− (X = NH₃, OH; n = 10, 12; y = 1, 2) were the most stable among all the derivatives considered. The findings of this work can be applied to the design of non-covalent complexes of closo-borate systems with biomolecules.

1. Introduction

Non-covalent interactions (NCIs) are one of the cornerstones of modern chemistry and related sciences [1,2]. Their formation is a driving force in numerous chemical and biological processes and is essential for the development and maintenance of life on Earth [3,4,5]. Even though some NCIs have low energy values, these contacts determine the structures of many complex systems, such as supramolecular assemblies, biomolecules, metal–organic frameworks, and many others [6,7,8,9,10]. To date, various types of non-covalent contacts have been studied. The most studied and best-known NCIs are hydrogen bonds of the general type DnH···Ac (Dn = donor, Ac = acceptor) [11,12,13,14]. Atoms such as nitrogen, oxygen and fluorine can function as both donors and acceptors [15,16,17,18]. The range of energy values of hydrogen bonds varies from 0.5 to more than 40 kcal/mol [19,20]. In addition to inorganic systems, hydrogen bonds play a crucial role in the structuring of biomolecules. In particular, they define the secondary structure of proteins and impact significantly upon the binding of small molecules to the active sites of proteins [21,22].

Closo-borate anions [B_nH_n]²⁻ (n = 6–12) and related compounds such as carboranes and metallaboranes contain negatively charged exo-polyhedral hydrogen atoms. The unique structure of these systems enables the formation of hydrogen bonds and other NCIs, such as halogen bonding and BH···π contacts, with various organic and organoelement systems [23,24,25,26,27,28].

The interactions of [B₁₀H₁₀]²⁻ and [B₁₂H₁₂]²⁻ with various alcohols were described in the work of Shubina et al. [29], where the main type of interactions was OH···HB contacts. The strength of these contacts for closo-borate anions was less than that of the [BH₄]⁻ anion and decreased from [B₁₀H₁₀]²⁻ to [B₁₂H₁₂]²⁻. Non-covalent complexes of the polyhedral carborane CB₁₁H₁₂ with several amino acids, such as glycine, serine, phenylalanine, glutamic acid, lysine, and arginine, were calculated in vacuo [30]. It was shown that the dominant interaction between carboranes and biomolecules was the formation of dihydrogen bonds characterised by a short distance between hydrogen atoms (as close as 1.8 Å). Average energy values of such contacts were in the range of 4.2 to 5.8 kcal/mol. The molecular assemblies of heteroboranes with Ala–Gly–Ala–Ala tetrapeptides were calculated [31]. In this work, calculations were performed using the implicit solvent model, and the effect of solvation on the stability of non-covalent complexes was described.

The ability of NCI formation determines the application of cluster-based compounds as drugs against various diseases [32,33,34]. A current example of the application of boron clusters is Boron Neutron Capture Therapy (BNCT) [35,36,37,38,39,40]. For effective treatment, the selective accumulation of boron clusters in tumour cells is required. L-type amino acid transporter 1 (LAT1) is a critical molecular target in BNCT therapy [41,42,43,44,45]. In the work of Y. Zhang et al., assemblies of LAT1 with several conjugates of boron clusters with amino acids were studied by means of docking modelling [44]. Data analysis revealed that the introduction of one of the conjugates of carborane with an amino acid into the protein resulted in the formation of hydrogen bonds between nitrogen and oxygen atoms and tyrosine 254 and asparagine 258, respectively. In several works, metallacarborane derivatives have been shown to be effective inhibitors of HIV-1 protease [46,47]. Applying the hybrid quantum mechanics/molecular mechanics (QM/MM) calculations, the binding of a metallacarborane-based inhibitor to HIV-1 protease was interpreted through atomistic and energy details [48]. Combining X-ray crystallography and quantum–chemical calculations has demonstrated the significant role of B–H⋯π hydrogen bonding in inhibiting carbonic anhydrase II by carborane sulfonamides [49].

As has been described, much theoretical work has been carried out on the NCIs of carborane and metallaborane systems with bioorganic systems [30,31,44,48,50], but the literature on closo-borate [B_nH_n]²⁻ (n = 6–12) systems remains limited. Thus, one of the motivations of the present study was to redress the balance. The focus of this study has been on molecular complexes of [B_nH_n−1X]^y− (X = H, NH₃, OH, SH, F; n = 10, 12; y = 1, 2) with glycine, with an assessment being made of the effect of the exo-polyhedral substituent on the cluster’s propensity to form non-covalent complexes. Glycine is the simplest amino acid, but its zwitterion form contains negatively charged carboxyl COO⁻ and positively charged ammonium NH₃⁺ groups. The preferable site of binding with boron clusters was identified, and several isomers of non-covalent complexes were localised. The authors determined the nature of NCIs between glycine and boron clusters and evaluated the stability of the resulting molecular assemblies.

2. Computational Details

All calculations of molecular complexes of [B_nH_n−1X]^y− (X = H, NH₃, OH, SH, F; n = 10, 12; y = 1, 2) with glycine were carried out using the ORCA 6.0.1 program package [51]. The initial geometry pre-optimisation and conformational scanning of key geometrical parameters, specifically the cluster–glycine distance and the rotation of the NH₃⁺ and COO⁻ functional groups of the amino acid, were performed at the r²SCAN-3c level of theory. Further geometry optimisation of the most energetically favourable isomers was performed at the ωB97X-D4rev/def2-TZVPP level of theory [52,53,54]. Range-separated DFT functionals combined with large triple-zeta basis sets enabled the calculation of the electronic structure and energy aspects of non-covalent molecular complexes to a high level of accuracy [55,56,57]. Tight criteria of SCF convergence (Tight SCF) were used for the calculations, which were performed using the RIJCOSX approximation with the def2/J auxiliary basis set [54,58,59]. Solvent effects were considered using the Solvation Model based on Density (SMD) [60], and water was chosen as the solvent. As glycine exists as a zwitterion under physiological conditions, it was used in this form for the current work. The singlet spin state was the ground state for all systems investigated. Only equatorial isomers of closo-decaborate anions [B₁₀H₉X]^y− (X = H, NH₃, OH, SH, F; y = 1, 2) were considered. Equatorial isomers are more energetically stable and easier to obtain than apical systems. Consequently, they are more extensively described in the literature [61,62,63,64,65]. Symmetry operations were not applied during the geometry optimisation procedure. Hessian matrices were calculated analytically for all model structures to prove the location of the correct minima on potential energy surfaces. No imaginary frequencies were found in any case. For an accurate evaluation of the electronic structures of the systems under consideration, additional single-point calculations at the Domain-based Local Pair Natural Orbital Coupled-Cluster with Single, Double, and perturbative Triples excitation method DLPNO-CCSD(T)/def2-TZVPP level were made [66,67,68]. Topological analysis of the NCI electron density distribution within the Quantum Theory of Atoms in Molecules (QTAIM) formalism [69,70,71], developed by Bader, was carried out using the Multiwfn program (version 3.8) [72]. A brief description of the main electron density descriptors is provided in the ESI. Energy values of NCIs were estimated based on the Espinosa–Lecomte correlation (Equation (1)):

$$E_{\mathit{NCI}}=-{\displaystyle \frac{1}{2}}V\left(r_{\mathit{bcp}}\right)$$

(1)

where V(r_bcp)—potential energy density; r_bcp—radius vector of coordinates of bond critical point; and ½ coefficient is in volume units (ų). For convenience, all energy values of NCIs are presented in kcal/mol (1 hartree = 627.509 kcal/mol).

The stability constant of the complex based on closo-borate anions and its derivatives with glycine was calculated according to the following equation (Equation (2)):

$$K={\displaystyle \frac{a_{\mathit{complex}}}{a_{\mathit{closo}\text{-}\mathit{borate}} \times { a}_{\mathit{glycine}}}}=e^{-{\displaystyle \frac{∆G}{RT}}}$$

(2)

where a_complex—activity of molecular complex of closo-borate anion and glycine; a_closo-borate—activity of closo-borate anion; a_glycine—activity of glycine; ∆G—free Gibbs energy of the complexation reaction, kcal/mol; R—universal gas constant, 0.001987 kcal/K × mol; and T = 298.15 K.

The interaction energy, E_int, is the difference between the energy of the complex and the sum of the energies of interacting units with their geometries taken from the geometry of the complex. For an accurate estimate of E_int (Equation (3)), the full counterpoise correction technique of estimating basis set superposition errors (BSSE) was implemented [73].

$$E_{\mathit{int}}={∆E}_{A-B}-{∆E}_A+{∆E}_B$$

(3)

E_bin considers the deformation of these units resulting from the complexation (Equation (4)).

$$E_{\mathit{bin}}=E_{\mathit{int}}+E_{\mathit{def}}$$

(4)

where E_int—interaction energy, and E_def—the energy needed to distort the fragments from their equilibrium configuration to the interacting geometry.

The bonding interactions between boron clusters and glycine were analysed by applying the Extended Transition State–Natural Orbitals for Chemical Valence decomposition scheme (ETS-NOCV) [74,75,76]. These calculations were carried out using the ORCA 6.1.0 program package. The ETS method decomposes the energy associated with bond formation into four components (Equation (5)):

$${∆E}_{\mathit{int}}={∆E}_{\mathit{elstat}}+{∆E}_{\mathit{Pauli}}+{∆E}_{\mathit{orb}}+{∆E}_{\mathit{disp}}$$

(5)

where ∆E_elstat—electrostatic interaction between molecular species (closo-borate anion and glycine); ∆E_Pauli—exchange repulsion between molecular species; ∆E_orb—interaction between the occupied and unoccupied orbitals of two molecular species; and ∆E_disp—dispersion contribution to the interaction between molecular species.

The deformation density shows the electron density rearrangement taking place upon the complex formation [75]. The visualisation of the optimised structures and deformation density was carried out using the ChemCraft program (version 1.8) [77]. XYZ files with coordinates of all optimised structures can be found in the attached zip folder.

3. Results

3.1. Complex of [B_nH_n]²⁻ Clusters with Glycine

Characterisation of the molecular complexes of [B_nH_n]²⁻ (n = 10, 12) anions with glycine began with a consideration of the geometric parameters of the intermolecular contacts between amino acid and cluster anions. The ammonium group NH₃ was positively charged and attracted to the negatively charged atoms of the boron cluster. Thus, the contact between the atoms of the NH₃ group and the cluster cage was the shortest. In the case of [B₁₀H₁₀]²⁻, the length of the contact between the ammonium hydrogen atom and the apical hydrogen atom of the cluster was equal to 2.14 Å. The spacing between the amine hydrogen atom of the NH₃ group and the apical boron atom was equal to 2.35 Å (Figure 1). As the distance between the NH₃-group and the apical hydrogen atom was the shortest, it was expected that the study would find a bonding path between the exo-polyhedral hydrogen atom and the hydrogen atom of the NH₃-group by QTAIM analysis. However, the bonding path occurred between the hydrogen atom of the NH₃-group and the apical boron atom of the cluster cage (Figure S1). Similar results were obtained in the case of the DLPNO-CCSD(T) calculation. Applying the Espinosa–Lecomte correlation (Equation (1)), the energy of the NH⋯B interaction was found to be equal to 2.6 kcal/mol (2.8 kcal/mol for DLPNO-CCSD(T)). In addition, contacts between the CH₂ group of glycine and the cluster cage were observed. The energy value of these contacts was equal to 0.9 kcal/mol, both for the ωB97X-D4rev and DLPNO-CCSD(T) calculations (Table S1).

In the case of the [B₁₂H₁₂]²⁻ anion, one of the ammonium hydrogen atoms had two practically equal contacts with two neighbouring cluster hydrogens. The lengths of these contacts were equal to 2.15 and 2.18 Å. For the [B₁₂H₁₂]²⁻ anion, there was a difference between the QTAIM results and those of the ωB97X-D4rev and DLPNO-CCSD(T) levels of theory. Based on the DFT calculation, the hydrogen of the ammonium group had only one contact with the hydrogen atom of the cluster cage. DLPNO-CCSD(T) calculations revealed that the NH₃ hydrogen atom formed two bonding paths with two different hydrogen atoms. Since the lengths of two NH⋯HB contacts were almost equivalent, one can expect the occurrence of two bonding paths between the NH₃ group and the boron cluster. Thus, the results of the DLPNO-CCSD(T) appeared to be more reasonable. The DLPNO-CCSD(T) values of energy of NH⋯BH contacts lie in the range of 1.8 to 1.9 kcal/mol. Based on the values of electron density descriptors at bcp and the value of non-covalent interaction energy, it can be concluded that the NH⋯B contact was stronger in the case of the [B₁₀H₁₀]²⁻ complex than the NH⋯HB contacts in the case of the [B₁₂H₁₂]²⁻ assembly with glycine. In the case of carborane-based molecular assemblies with biomolecules, it was revealed that the energy of NH⋯HB contacts was in the range of 2.1 to 5.8 kcal/mol [30]. Thus, the NH⋯HB and NH⋯B contacts found in the present work had energies similar to those in analogous carborane-based systems. In addition to the NH⋯HB contacts, the CH⋯HB contact was established. As in the case of the glycine complex with [B₁₀H₁₀]²⁻, the CH⋯HB contact was weaker than the NH⋯HB contacts, and the energy of this contact was equal to 0.8 kcal/mol.

The value of the interaction energy E_int of complex [B₁₀H₁₀]²⁻ with glycine was found to be greater (by modulus) than analogous values for the [B₁₂H₁₂]²⁻ assembly in both the ωB97X-D4rev and DLPNO-CCSD(T) calculations (Table S9). In the case of the ωB97X-D4rev calculation, the difference in values between complexes was equal to 3 kcal/mol, whereas in the case of DLPNO-CCSD(T), this difference was equal to 1.1 kcal/mol. The value of binding energy E_bin was also greater in the case of the [B₁₀H₁₀]²⁻ complex for both computing approaches. In the case of ωB97X-D4rev, the difference was equal to 0.9 kcal/mol, and for DLPNO-CCSD(T), the difference was equal to 0.8 kcal/mol. Despite the negative values of E_int and E_bin, the Gibbs energy (∆G) values for the complexation reactions were positive due to the positive values of entropy (∆S) (Table S9). The ∆G for the [B₁₀H₁₀]²⁻ complex was lower by 1.0 and 1.1 kcal/mol than for the [B₁₂H₁₂]²⁻ complex according to the ωB97X-D4rev and DLPNO-CCSD(T) approaches, respectively. Using the ∆G values, the stability constants of the molecular assemblies were estimated (Table S10). For the [B₁₀H₁₀]²⁻ complex, the stability constant was equal to 3.1 × 10⁻⁵ at the DLPNO-CCSD(T) level of theory (4.8 × 10⁻⁶ for the DFT calculation). For the [B₁₂H₁₂]²⁻ complex, the stability constant was equal to 4.7 × 10⁻⁶ at the DLPNO-CCSD(T) level of theory (8.8 × 10⁻⁷ for the DFT calculation).

The ETS-NOCV approach was used to implement the decomposition of bond energy to its main components (Table S11). The overall bonding energy, E_int, between [B₁₀H₁₀]²⁻ and glycine was equal to −11.4 kcal/mol. Electrostatic energy contributed the most to the overall bonding energy. The value of this component was equal to −44.5 kcal/mol, whereas the value of orbital energy was equal to −5.4 kcal/mol. The value of bonding energy between [B₁₂H₁₂]²⁻ and glycine was less than in the case of the analogous complex based on [B₁₀H₁₀]²⁻. This value was equal to −7.6 kcal/mol. In the case of the [B₁₂H₁₂]²⁻ complex, the contribution of electrostatic energy was more significant than in the system based on [B₁₀H₁₀]²⁻, and was equal to −59.6 kcal/mol. The value of orbital energy was equal to −3.9 kcal/mol.

From the graphical representation of deformation density contributions, it can also be concluded that the electron density donor was not a single cluster atom (boron or hydrogen), but a facet containing atoms of boron and hydrogen (Figure 2, indicated with the blue colour). This explains the rather unusual results of the QTAIM analysis, in which the bonding path linked the hydrogen atom of the ammonium group and the boron atom of the cluster. The same conclusion can be drawn for the complex of glycine and the [B₁₂H₁₂]²⁻ cluster.

Thus, the complex of the [B₁₀H₁₀]²⁻ anion with glycine had a stronger bond between the cluster and amino acid than that of the system based on the [B₁₂H₁₂]²⁻ cluster. This conclusion is based on energy evaluations, QTAIM, and ETS-NOCV analysis. The most precise description of the bonding pattern between cluster anions and glycine was obtained by the ETS-NOCV approach, where the hydrogen atom of the glycine ammonium group was found to interact with the whole facet of the boron cluster.

3.2. Complex of [B_nH_n−1NH₃]⁻ Clusters with Glycine

The second type of NCI featured in this study belonged to the complexes of ammonia derivatives [B_nH_n−1NH₃]⁻ (n = 10, 12) with glycine. Positively charged hydrogen atoms of the exo-polyhedral cluster ammonium group formed a short contact with the negatively charged oxygen atom of the carboxyl group of glycine. The length of this contact was equal to 1.84 Å in the case of the [B₁₀H₉NH₃]⁻ anion, and 1.82 Å in the case of the 12-vertex anion [B₁₂H₁₁NH₃]⁻ (Figure 3). In addition to COO_gly⋯HN_cluster contacts, the hydrogen atoms of the glycine ammonium group formed short interactions with the atoms of the cluster. For [B₁₀H₉NH₃]⁻, the NH_gly⋯HB length was equal to 2.15 Å, and NH_gly⋯B was equal to 2.38 Å. The glycine ammonium hydrogen atom formed two equal NCIs with two exo-polyhedral hydrogen atoms of the unsubstituted closo-dodecaborate anion [B₁₂H₁₂]²⁻, while in the case of [B₁₂H₁₁NH₃]⁻, one of the NH_gly⋯HB contacts was equal to 2.11 Å and another was equal to 2.35 Å.

The bonding path between the hydrogen atom of the NH₃- group and the apical boron atom of the cluster cage in the glycine complex with [B₁₀H₉NH₃]⁻ was also inspected. Based on the Espinosa–Lecomte correlation (Equation (1)), the energy of such a contact was equal to 2.4 kcal/mol (Table S2). In the case of the [B₁₂H₁₁NH₃]⁻ cluster complex, both the DFT and DLPNO-CCSD(T) approaches detected two bond paths between the H atom of the glycine NH₃-group and two exo-polyhedral hydrogen atoms (Figure S2). These contacts differed not only by their lengths but also by their descriptors of electron density. The energies of the contacts differed by 0.5 kcal/mol using both the ωB97X-D4rev and DLPNO-CCSD(T) approaches.

COO_gly⋯HN_cluster interactions were stronger than NH_gly⋯B ([B₁₀H₉NH₃]⁻) and NH_gly⋯HB ([B₁₂H₁₁NH₃]⁻) interactions. The values of the main descriptors of electron density ρ(r_bcp) of COO_gly⋯HN_cluster interactions were greater than the analogous values for NH_gly⋯B and NH_gly⋯HB. According to the Espinosa–Lecomte correlation (Equation (1)), the DFT values of COO_gly⋯HN_cluster interactions lie in the range of 9.2 to 9.8 kcal/mol (9.5 to 10.1 kcal/mol with the DLPNO-CCSD(T) approach). The interaction within the complex of the [B₁₂H₁₁NH₃]⁻ anion COO_gly⋯HN_cluster had a greater energy value (9.8 kcal/mol for ωB97X-D4rev and 10.1 kcal/mol for DLPNO-CCSD(T)) than the analogous complex of the [B₁₀H₉NH₃]⁻ anion with glycine. According to Jeffrey’s classification, biologically significant N–H⋯OC hydrogen bonds have energies in the range of 4 to 15 kcal/mol [78]. The values of the COO_gly⋯HN_cluster contacts obtained in the present work can be attributed to moderate-strength interactions. The energy of weak CH⋯HB and COO⋯HB interactions lies in the range of 0.7 to 1.0 kcal/mol using both the DFT and DLPNO-CCSD(T) approaches.

The complex based on the closo-decaborate anion [B₁₀H₉NH₃]⁻ had greater values of E_int than the closo-dodecaborate derivative [B₁₂H₁₁NH₃]⁻, similar to the unsubstituted boron clusters (Table S9). However, the difference between the E_int values of the two complexes was equal to 0.1 kcal/mol, according to the ωB97X-D4rev calculation. The difference was equal to 0.8 kcal/mol when using the DLPNO-CCSD(T) calculation. The difference in energies E_bin was equal to 0.2 kcal/mol in the case of the DFT approach and 1.1 kcal/mol when using DLPNO-CCSD(T). As in the case of molecular complexes based on [B_nH_n]²⁻, the ∆G values for complex formation were positive (Table S9). The ∆G of the [B₁₀H₉NH₃]⁻ complex was lower by 1.3 and 1.2 kcal/mol than for the [B₁₂H₁₁NH₃]⁻ complex, according to the ωB97X-D4rev and DLPNO-CCSD(T) calculations, respectively. For the [B₁₀H₉NH₃]⁻ complex, the stability constant was equal to 3.6 × 10⁻³ at the DLPNO-CCSD(T) theory level (1.0 × 10⁻³ for DFT calculation, Table S10). For the [B₁₂H₁₁NH₃]⁻ complex, the stability constant was lower by an order of magnitude and equal to 5.0 × 10⁻⁴ at the DLPNO-CCSD(T) theory level (1.1 × 10⁻⁴ for DFT calculation).

The ETS-NOCV approach revealed that the E_int of the [B₁₀H₉NH₃]⁻ complex with glycine was 3.8 kcal/mol higher than the E_int of the [B₁₂H₁₁NH₃]⁻ complex (Table S11). Electrostatic interaction was predominant for the binding of the boron cluster and glycine in both complexes. The value of this interaction type was higher in the [B₁₀H₉NH₃]⁻ complex with glycine. An analogous trend was observed for orbital interaction.

A graphical representation of deformation density contributions demonstrated that the main contribution to the formation of complexes between ammonium derivatives of closo-borate anions and amino acid was provided by the COO_gly⋯HN_cluster interaction (Figure 4).

Thus, the findings of the study suggest that glycine and the [B₁₀H₉NH₃]⁻ anion were bonded more firmly than the complex based on the [B₁₂H₁₁NH₃]⁻ anion. The interaction between the oxygen atom of the carbonyl group and the hydrogen atom of the ammonium group contributed the most to the binding between the clusters and the amino acid. In addition, interactions between hydrogen atoms of the NH₃- group and the atoms of the cluster cage were shown to make a moderate contribution to the interactions between the clusters and glycine.

3.3. Complex of [B_nH_n−1OH]²⁻ Clusters with Glycine

For complexes of [B_nH_n−1OH]²⁻ clusters with glycine, several isomers were localised. These isomers were formed due to the existence of different binding modes between cluster anions and the amino acid. Glycine could bind to the boron cluster by the interaction NH···O between the hydrogen atom of the glycine ammonium group and the oxygen atom of the exo-polyhedral OH-group. A variant on this mechanism was the interaction CO···HO between the oxygen atom of the glycine carboxyl group and the hydrogen of the exo-polyhedral OH-group. Subsequently, the NH···O type of structure will be referred to as Iso_1, while CO···HO structures will be designated as Iso_2, and the third variant, which is based on the presence of both types of interactions (NH···O and CO···HO), will be designated as Iso_3. Iso_1 was most energetically favourable for the complex based on [B₁₀H₉OH]²⁻, although the values of electronic energy of all found isomers vary by less than 1 kcal/mol (0.4 kcal/mol DLPNO-CCSD(T)). Iso_3 was the most energetically stable for the glycine complex with the [B₁₂H₁₁OH]²⁻ anion. As for closo-decaborate derivatives, the values of electronic energy for all considered isomers also differed slightly from each other.

For Iso_1 systems, the lengths of the NH···O interactions were practically identical for both [B_nH_n−1OH]²⁻ complexes. For the assembly based on the [B₁₀H₉OH]²⁻ anion, the value of a given contact was longer, and equal to 1.72 Å. For the closo-dodecaborate anion, the analogous value was equal to 1.70 Å (Figure 5). Also, for Iso_1 systems, the complex of the [B₁₂H₁₁OH]²⁻ anion had greater values for the main descriptors of electron density (by modulus) at the bcp of NH···O interactions (Table S3). For example, the difference in electron density ρ(r_bcp) was equal to 0.0019 e/ų (0.0018 e/ų for DLPNO-CCSD(T)). Based on the Espinosa–Lecomte correlation (Equation (1)), the energy value of the NH···O interaction for the [B₁₂H₁₁OH]²⁻ complex was 0.7 kcal/mol higher than for the [B₁₀H₉OH]²⁻ complex. The energies of NH···O contacts obtained at the ωB97X-D4rev theory level lay within the interval of 13.7 to 14.4 kcal/mol (13.9 to 14.9 kcal/mol for DLPNO-CCSD(T)). As mentioned above, the energies of biologically significant N–H⋯OC hydrogen bonds are in the range from 4 to 15 kcal/mol. Thus, the NH···O contacts found in this work can be attributed to strong interactions.

In addition, another hydrogen atom of the ammonium group formed a contact with the exo-polyhedral hydrogen atom (Figure S3). This contact was observed for both [B_nH_n−1OH]²⁻ complexes. The lengths of this contact lie in the interval of 2.39 to 2.40 Å. The values of the NH···HB contact energy lie in the interval of 1.0 to 1.1 kcal/mol. The hydrogen atom of the glycine CH₂-group also had short contacts with the exo-polyhedral hydrogen atoms. The length of the CH···HB contact was 2.49 Å in the case of [B₁₀H₉OH]²⁻. For the [B₁₂H₁₁OH]²⁻ anion complex, the H atom of the glycine group formed two contacts with cluster atoms. The length of one contact was equal to 2.63 Å, while the other contact was 2.71 Å in length. For the CH···HB interaction, the energy value lay in the interval of 0.6 to 0.8 kcal/mol.

CO···HO interactions in Iso_2 exhibited significant differences between complexes based on closo-decaborate and closo-dodecaborate derivatives. In the case of the [B₁₀H₉OH]²⁻ complex, this interaction had a length that was equal to 1.94 Å, whereas for the [B₁₂H₁₁OH]²⁻ anion, the analogous interaction was shorter and equal to 1.87 Å (Figure 6). These contacts had significant differences between the main descriptors of electron density (Table S4). For example, for the complex based on the [B₁₀H₉OH]²⁻ anion, the value of the CO···HO interaction ρ(r_bcp) was equal to 0.0247 e/ų (0.0236 e/ų DLPNO-CCSD(T)), whereas the analogous value for the [B₁₂H₁₁OH]²⁻ complex was equal to 0.0296 e/ų (0.0282 e/ų DLPNO-CCSD(T)). The difference between the energies of such contacts was 1.7 kcal/mol for both ωB97X-D4rev and DLPNO-CCSD(T) cases. The DFT values of the CO···HO contacts lie in the range of 6.6 to 8.4 kcal/mol (6.8 to 8.5 kcal/mol for DLPNO-CCSD(T)). CO···HO contacts were investigated most thoroughly in the case of malonaldehyde and its derivatives. The values of energy for this NCI lie within the interval of 9.7 to 20.6 kcal/mol [79,80]. Thus, the CO···HO interactions in the complexes of [B_nH_n−1OH]²⁻ clusters with glycine were weaker than those in malonaldehyde and its derivatives. However, it is worth noting that the hydrogen bond in malonaldehyde is intramolecular, and this type of NCI is typically stronger than intermolecular ones.

In addition, the short contacts between the hydrogen atom of the glycine ammonium group and the atoms of the cluster cage were inspected for Iso_2 (Figure S4). In the case of the Iso_2 complexes of [B_nH_n−1OH]²⁻, the NH···B contacts were much longer than the analogous contacts for [B₁₀H₁₀]²⁻ and [B₁₀H₉NH₃]⁻ systems, and lie in the range of 2.62 to 2.70 Å. Thus, for the [B₁₀H₉OH]²⁻ complex, short NH···HB contacts between the hydrogen atom of the NH₃ group and the two exo-polyhedral hydrogens were observed and confirmed by means of QTAIM analysis. The lengths of this contact lie in the range of 2.20 to 2.21 Å. The energy values of these contacts were 1.5 kcal/mol using the ωB97X-D4rev calculation and 1.6 kcal/mol using the DLPNO-CCSD(T) approach. Similar results were obtained for the complex based on the [B₁₂H₁₁OH]²⁻ anion. The lengths of NH···HB contacts lie in the range of 2.18 to 2.20 Å, while the energy value was 1.5 kcal/mol according to the ωB97X-D4rev calculation (1.6 kcal/mol DLPNO-CCSD(T)), as in the case of the [B₁₀H₉OH]²⁻ complex.

For one of the hydrogen atoms of the CH₂-group, a CH···HB contact was observed. This type of interaction was established for both types of cluster complexes. The lengths of this contact were equal to 2.23 and 2.48 Å for the [B₁₀H₉OH]²⁻ and [B₁₂H₁₁OH]²⁻ complexes, respectively. The energy values lie within the interval of 1.2 to 1.4 kcal/mol. O···HB contacts were also observed, with the energy of such contacts lying in the interval of 0.6 to 1.1 kcal/mol.

For Iso_3 systems, both NH···O and CO···HO contacts were observed (Figure 7 and Figure S5). These contacts were longer than those in systems containing only one type of interaction (NH···O or CO···HO). NH···O contacts lie in the range of 1.74 to 1.76 Å, while the CO···HO contacts were in the range of 1.94 to 1.98 Å. NH···O contacts had greater values for electron density descriptors than CO···HO contacts. For example, the values of electron density at bcp lie in the range of 0.0399 to 0.0417 e/ų and 0.0239 to 0.0257 e/ų for NH···O and CO···HO contacts, respectively (Table S5). In both cases, the shortest contacts were observed for the [B₁₂H₁₁OH]²⁻ complex. Also, for the system based on the closo-dodecaborate anion, these contacts had larger values of electron density descriptors. The difference in the values of electron density, ρ(r_bcp), corresponding to NH···O interaction, was 0.0018 e/ų (0.0017 e/ų DLPNO-CCSD(T)). A similar difference in values was obtained for ρ(r_bcp), corresponding to CO···HO contact (0.0018 e/ų using the DFT approach and 0.0017 e/ų using DLPNO-CCSD(T)). The values of energy interaction were also greater in the case of the [B₁₂H₁₁OH]²⁻ complex. The difference was 0.7 kcal/mol with both the ωB97X-D4rev and the DLPNO-CCSD(T) calculations. It is noteworthy that the energy values of the NH···O and CO···HO interactions in Iso_3 systems were less than for the analogous contacts in Iso_1 and Iso_2.

Further analysis revealed additional weak interactions, beyond the NH···O and CO···HO contacts, between the cluster cage of [B_nH_n−1OH]²⁻ and the amino acid. NH···HB, N···HB, and C···HB interactions were established using QTAIM analysis. The energy of such interactions lay in the range of 0.6 to 1.1 kcal/mol and 0.6 to 1.2 kcal/mol, with ωB97X-D4rev and DLPNO-CCSD(T) calculations, respectively.

For complexes of [B₁₀H₉OH]²⁻ with glycine, the highest values of E_int and E_bin were calculated for the Iso_1 system, whereas the Iso_3 system had the lowest values (Table S9). In the case of [B₁₂H₁₁OH]²⁻, the highest values were obtained for Iso_3 systems, while the lowest values belonged to Iso_2 systems. Iso_1 and Iso_2 system complexes based on [B₁₀H₉OH]²⁻ had greater values of E_int and E_bin than [B₁₂H₁₁OH]²⁻ systems. The highest values of E_int and E_bin were equal for the Iso_1 complex of the [B₁₀H₉OH]²⁻ anion and the Iso_3 complex of [B₁₂H₁₁OH]²⁻, according to the DFT approach. The DLPNO-CCSD(T) calculation produced the highest values of E_int and E_bin for the Iso_3 complex of [B₁₂H₁₁OH]²⁻ only.

The values of ∆G for complex formation were positive (Table S9). For the complexes of [B₁₀H₉OH]²⁻ with glycine, the lowest positive value of ∆G was observed for the Iso_1 system at both the ωB97X-D4rev and DLPNO-CCSD(T) theory levels. For Iso_1 molecular complexes, the stability constant was equal to 2.0 × 10⁻⁴ (ωB97X-D4rev) and 3.1 × 10⁻⁴ (DLPNO-CCSD(T), Table S10). In the case of [B₁₂H₁₁OH]²⁻ complexes, the lowest positive value of ∆G was observed for Iso_3 and Iso_2 complexes at the DFT and DLPNO-CCSD(T) theory levels, respectively.

The calculated values of E_int and E_bin showed good agreement with the data obtained using the ETS-NOCV approach (Table S11). The highest value of total energy corresponded to the interaction of glycine complexes with [B₁₀H₉OH]²⁻ in the Iso_1 system and for the [B₁₂H₁₁OH]²⁻ complex in the Iso_3 system. The main contribution to binding between the clusters and amino acid was attributed to electrostatic energy, as it was for [B_nH_n]²⁻ and [B_nH_n−1NH₃]⁻ complexes. Orbital interaction had much less impact on binding between the cluster systems and glycine. The graphical representation of deformation density contributions was most informative in the case of Iso_3 systems, indicating that NH···O contacts provided the greatest contribution to binding between the boron cluster and glycine (Figure 8).

Therefore, three possible types of binding were determined for the complexes of [B_nH_n−1OH]²⁻ with glycine. The most energetically favourable system for [B₁₀H₉OH]²⁻ was the complex with an NH···O interaction, whereas for [B₁₂H₁₁OH]²⁻, it was the complex stabilised by both NH···O and CO···HO contacts. Furthermore, NH···O contacts facilitated stronger binding between the cluster anion and the amino acid than CO···HO interactions.

3.4. Complex of [B_nH_n−1SH]²⁻ Clusters with Glycine

Several main types of coordination were possible for the [B_nH_n−1SH]²⁻ glycine complexes, which were similar to their [B_nH_n−1OH]²⁻ analogues. Glycine could be bound to the cluster derivative by the NH···S interaction or the CO···HS contact. The isomer (Iso_1) in the presence of the NH···S contact was slightly more energetically favourable than the isomer (Iso_2) with a CO···HS contact for the derivative [B₁₀H₉SH]²⁻. The difference between the values of total electronic energy was 0.3 kcal/mol according to the DFT approach and only 0.02 kcal/mol when using the DLPNO-CCSD(T) approach. The energy difference between isomers was more significant and was equal to 2.3 kcal/mol (ωB97X-D4rev) and 1.9 kcal/mol (DLPNO-CCSD(T)) for the complexes of derivative [B₁₂H₁₁SH]²⁻.

The glycine complex of [B₁₀H₉SH]²⁻ had a shorter bond length than the [B₁₂H₁₁SH]²⁻ analogues (2.28 Å vs. 2.33 Å) in the case of Iso_1 systems (Figure 9). The values of the main electronic descriptors of NH···S bcp for the [B₁₀H₉SH]²⁻ complex were higher than the analogous values for the [B₁₂H₁₁SH]²⁻ complex (Table S6). The difference between the values of electron density was equal to 0.0027 e/ų (ωB97X-D4rev) and 0.0025 e/ų (DLPNO-CCSD(T)). The difference between the interaction energies was 0.7 kcal/mol, according to the both DFT and DLPNO-CCSD(T) calculations. It is worth noting that the NH···S interactions were much weaker than the NH···O interactions in [B_nH_n−1OH]²⁻ complexes. The values of NH···S contact energy lay in the range of 4.5 to 5.2 kcal/mol (ωB97X-D4rev) and 4.7 to 5.4 kcal/mol (DLPNO-CCSD(T)), while the values of NH···O contact energy lay in the range of 13.7 to 14.4 kcal/mol (ωB97X-D4rev) and 13.9 to 14.6 kcal/mol (DLPNO-CCSD(T)). According to the literature, the S-mediated NCIs have energy values in the range of −4.5 to −5.5 kcal/mol [81]. Therefore, the data obtained in the present work was in good agreement with previously published data. In addition to NH···S contacts, some weak NH···HB, N···HB, O···HB, and C···HB interactions were examined (Figure S6). The energy of these contacts lay in the range of 0.6 to 1.1 kcal/mol (ωB97X-D4rev) and 0.6 to 1.2 kcal/mol (DLPNO-CCSD(T)).

Iso_2 systems were formed due to non-covalent contact between the oxygen atom of the carboxyl group of glycine and the hydrogen atom of the cluster exo-polyhedral substituent (Figure 10). It is worth noting that CO···HS contacts of Iso_2 systems were much longer than the NH···S interactions of Iso_1 systems. For the [B₁₂H₁₁SH]²⁻ system, the length of the O···HS contacts was the shortest, and equal to 2.61 Å. The length of this contact was 0.11 Å longer for the [B₁₀H₉SH]²⁻ complex. The energy values of this contact were significantly lower than the O···HO contacts of Iso-2 systems of [B_nH_n−1OH]²⁻ anions (Table S7). For the complex of [B₁₂H₁₁SH]²⁻, the energy value was equal to 1.1 kcal/mol (ωB97X-D4rev) and 1.2 kcal/mol (DLPNO-CCSD(T)). For the [B₁₀H₉SH]²⁻ complex, this value was equal to 0.9 kcal/mol (ωB97X-D4rev) and 1.0 kcal/mol (DLPNO-CCSD(T)). Thus, the CO···HS interactions in the complexes of [B_nH_n−1SH]²⁻ clusters with glycine can be attributed to weak hydrogen bonds. Using QTAIM analysis, the NH···B contact was established for the complex of derivative [B₁₀H₉SH]²⁻ (Figure S7). The length of this contact was 2.34 Å, and the energy value was 2.6 kcal/mol. In the case of the [B₁₂H₁₁SH]²⁻ anion, the hydrogen atom of the ammonium group formed a contact with the hydrogen atom of the cluster cage. The length of this contact was 2.13 Å. The energy value of the given interaction was less than that for the NH···B contact of the [B₁₀H₉SH]²⁻ complex with glycine. This value was equal to 1.2 kcal/mol (ωB97X-D4rev) and 1.3 kcal/mol (DLPNO-CCSD(T)). Weak non-covalent interactions of the types NH···HB, O···HB, and C···HB were observed in the Iso_3 systems. Their calculated energies ranged from 0.7 to 1.1 kcal/mol (ωB97X-D4rev) and from 0.6 to 1.2 kcal/mol (DLPNO-CCSD(T)).

For non-covalent complexes of [B_nH_n−1SH]²⁻ with glycine, several trends in E_int and E_bin were revealed. First, these values were less (by modulus) than the analogous values for [B_nH_n−1OH]²⁻ systems (Table S9). This observation held for both DFT and DLPNO-CCSD(T) calculation approaches and is in good agreement with the fact that the NH···S and CO···HS contacts were weaker than the NH···O and CO···HO interactions. The Iso_1 system had higher values (by modulus) of E_int and E_bin than the Iso_2 system for the [B₁₂H₁₁SH]²⁻ complex, and this was found with both the ωB97X-D4rev and DLPNO-CCSD(T) calculations. In the case of [B₁₀H₉SH]²⁻, this was established only when using the DFT results, whereas higher values between E_int and E_bin were identified for Iso_2 according to DLPNO-CCSD(T) calculations. However, the difference between the values was only 0.1 kcal/mol. Complexes based on the [B₁₀H₉SH]²⁻ anion had higher values of E_int than analogous systems of the [B₁₂H₁₁SH]²⁻ anion. For E_bin, the Iso_1 isomer of the [B₁₂H₁₁SH]²⁻ complex had a higher value than that of the [B₁₀H₉SH]²⁻ system.

The ∆G values for complex formation were positive (Table S9). For the complexes based on [B₁₀H₉SH]²⁻, the Iso_2 isomer had the lowest values of ∆G energies at both the ωB97X-D4rev and DLPNO-CCSD(T) theory levels. The [B₁₂H₁₁SH]²⁻ systems had an opposite trend, and Iso_1 was more thermodynamically stable than Iso_2. The [B₁₂H₁₁SH]²⁻ complexes had lower values of ∆G than the [B₁₀H₉SH]²⁻ ones.

For the systems described previously, electrostatic interactions provided the main contribution to binding between the clusters and amino acids, according to ETS-NOCV data (Table S11). The highest values of electrostatic and orbital interaction energies were revealed for the complex of [B₁₀H₉SH]²⁻. It is noteworthy that the [B₁₂H₁₁SH]²⁻ complex was the only system where the dispersion interaction energy exceeded the orbital interaction energy. For Iso_2 systems, the main contribution to binding between clusters and glycine was attributed to the interaction between the hydrogen atom of the ammonium group and atoms of the cluster cage, according to the graphical representation (Figure 11).

Thus, based on thermodynamic data, QTAIM, and ETS-NOCV analysis, the NH···S contacts between cluster derivatives [B_nH_n−1SH]²⁻ and glycine promoted stronger binding than O···HS contacts. Moreover, NH···HB or NH···B contacts significantly contributed to the association between the cluster and the amino acid for Iso_2 complexes. [B₁₀H₉SH]²⁻ formed a more stable complex with glycine than [B₁₂H₁₁SH]²⁻.

3.5. Complex with [B_nH_n−1F]²⁻ Clusters with Glycine

The interaction between [B_nH_n−1F]²⁻ derivatives and glycine occurred through NH···F contacts. Since these complexes with glycine exhibited only one binding mode, a single isomer for each type of cluster ([B₁₀H₉F]²⁻ and [B₁₂H₁₁F]²⁻) was considered (Figure 12). The lengths of NH···F contacts were almost equal for closo-decaborate and closo-dodecaborate fluorine derivatives and lay in the range of 1.82 to 1.83 Å. QTAIM analysis revealed that the main electron density descriptors at the bcp of the NH···F contact had larger magnitudes in the [B₁₀H₉F]²⁻ complex than in the [B₁₂H₁₁F]²⁻ analogue (Table S8). For example, the difference in the value of ρ(r_bcp) was 0.0014 e/ų (ωB97X-D4rev) and 0.0012 e/ų (DLPNO-CCSD(T)). The energies of the NH···F interactions were quite high, lying in the range of 8.4 to 9.0 kcal/mol (ωB97X-D4rev) and 8.5 to 9.0 kcal/mol (DLPNO-CCSD(T)). The energy of the NH···F contact was 0.5 kcal/mol higher for the [B₁₀H₉F]²⁻ anion than for the [B₁₂H₁₁F]²⁻ system. The literature data indicated that the energy values for X–H···F interactions span a wide range, from 0.3 to 11.5 kcal/mol [82,83,84]. Thus, according to the reported data, the NH···F contacts in the complexes of [B_nH_n−1F]²⁻ clusters with glycine can be attributed to strong interactions. In addition, some weak CH···HB, O···HB, and N···HB interactions were examined (Figure S8). The energy of such contacts lay in the range of 0.4 to 0.9 kcal/mol, with both DFT and DLPNO-CCSD(T) calculations.

The values of E_int and E_bin of the closo-decaborate derivative [B₁₀H₉F]²⁻ were higher than in the case of the closo-dodecaborate derivative [B₁₂H₁₁F]²⁻. The difference between the E_int and E_bin values of [B₁₀H₉F]²⁻ and [B₁₂H₁₁F]²⁻ complexes was in the range of 1.1 to 1.5 kcal/mol, using both ωB97X-D4rev and DLPNO-CCSD(T) levels of theory (Table S9). The [B₁₀H₉F]²⁻ complex had a lower positive value of ∆G than the [B₁₂H₁₁F]²⁻ analogue according to both DFT and DLPNO-CCSD(T) approaches (Table S9). For the [B₁₀H₉F]²⁻ complex, the stability constant was equal to 1.2 × 10⁻⁵ at the DLPNO-CCSD(T) theory level (4.8 × 10⁻⁶ for the DFT calculation, Table S10).

Based on the ETS-NOCV approach, the value of E_int of the [B₁₀H₉F]²⁻ complex was higher than in the case of the [B₁₂H₁₁F]²⁻ anion analogue (Table S11). The difference was 4.3 kcal/mol. The [B₁₀H₉F]²⁻ complex also exhibited higher orbital and electrostatic interaction energies (by absolute value). The graphical representation of deformation density contributions proved that the main contribution to binding between the cluster anion and glycine was provided by NH···F contact (Figure 13).

Thus, the combined thermodynamic, QTAIM, and ETS-NOCV data suggested that the [B₁₀H₉F]²⁻ anion formed complexes bound with glycine that were stronger than those of the [B₁₂H₁₁F]²⁻ system.

3.6. General Trends

General patterns within the non-covalent interactions between glycine and closo-borate anions and their derivatives [B_nH_n−1X]^y− (X = H, NH₃, OH, SH, F; n = 10, 12; y = 1, 2) were analysed. NH···B and NH···HB contacts played a crucial role in the binding between the cluster anion and glycine for unsubstituted [B_nH_n]²⁻ systems. XH···HB contacts (X = C, N, O) are well-known types of contact, according to the literature [23,24,48,49], whereas NH···B contacts have not been mentioned previously. However, based on ETS-NOCV data, these contacts could be attributed to multicentre interactions between the hydrogen atom of the NH₃ group and atoms of the boron cage facet. For [B_nH_n−1SH]²⁻ based complexes (Iso_2 systems), these contacts also contributed the most to binding between cluster and amino acid species. These interactions were stronger for closo-decaborate systems than for closo-dodecaborate systems. When compared with the literature data, OH···HB interactions had a similar range of energy values (1.8 to 4.4 kcal/mol) [29]. However, the strongest OH···HB contacts had higher energy values than NH···B and NH···HB contacts.

For the substituted derivatives [B_nH_n−1X]^y− (X = SH, NH₃, OH, F; y = 1, 2), NH···X (X = O, S, F), and CO···HX (X = N, O, S), contacts had a significant impact on binding between the cluster fragment and glycine. These contacts had higher energy values than NH···B and NH···HB contacts. Most of these contacts could be attributed to moderate or strong interactions compared with analogous types of interactions in organic and biological systems [78,81]. NH···X contacts were stronger than CO···HX (X = N, O, S) contacts. In general, NH_gly···X_cluster and CO_gly···HX_cluster interactions exhibited greater strength in closo-dodecaborate derivatives than in closo-decaborate analogues. The NH···O interaction for the [B₁₂H₁₁OH]²⁻ Iso_1 complex had the highest energy value (14.4 kcal/mol DT and 14.6 kcal/mol DLPNO-CCSD(T)) among all the non-covalent contacts described.

General trends of the energy aspects of [B_nH_n−1X]^y− complexes with glycine were analysed. First, a good agreement between the DFT and DLPNO-CCSD(T) calculations of E_int and E_bin was found. In addition, the values of E_int obtained using the ETS-NOCV scheme strongly correlated with the DFT and DLPNO-CCSD(T) calculations of E_int and E_bin (Figure 14).

The correlation between the energy of non-covalent interactions and the overall E_int and E_bin values was not so clear and not so strong. A correlation analysis was carried out on the values of the strongest interactions and the total sum of the energy values of NCI. The highest correlation coefficients with NCI descriptors were obtained for the total sum of the NCI energy values. E_int and E_bin were higher for complexes of closo-decaborate anion and its derivatives [B₁₀H₉X]^y−, whereas the energies of most valuable non-covalent interactions were higher for complexes of the closo-dodecaborate anion and its derivatives [B₁₂H₁₁X]^y−. The reason for this discrepancy was found in the analysis of the ETS-NOCV data. The calculated values for Pauli repulsion and solvation energy were more positive for closo-dodecaborate complexes, and these positive contributions counteracted the strong attraction of these complexes.

The attachment of –SH and –F substituents to the closo-borate cage encouraged the energy values of E_int and E_bin to exceed those of the unsubstituted clusters. However, the increase in E_int and E_bin was less pronounced than that of the complex with –NH₃ and –OH substituted cluster derivatives. The [B₁₀H₉NH₃]⁻ complex exhibited the highest E_int and E_bin values among all systems described.

Despite the negative values of E_int and E_bin, all studied complexes had positive ∆G values. This was due to a large unfavourable entropy change (–T∆S), which outweighed the favourable enthalpy change. However, a lot of molecular complexes have positive values of ∆G, and these systems have been actively studied both theoretically and experimentally [85,86,87,88,89,90,91]. The ∆G values obtained using both the DFT and DLPNO-CCSD(T) approaches had good correlation with E_int and E_bin. The correlation coefficients lie in the range of 0.83 to 0.95. The systems based on the derivatives of the closo-decaborate anion had lower ∆G values than the assemblies of the closo-dodecaborate anion, and, consequently, higher stability constants. The assemblies of the hydroxy and ammonium derivatives had lower ∆G values when compared with other derivatives and the unsubstituted closo-borates. The complex of ammonium derivative [B₁₀H₉NH₃]⁻ had the lowest ∆G value and the highest stability constant (3.6 × 10⁻³ according to DLPNO-CCSD(T) approach). The obtained value indicates that a noticeable amount of the complex between closo-decaborate anions [B_nH_n−1NH₃]⁻ and glycine exists in its intact form and can be detected by experimental methods.

4. Conclusions

The non-covalent complexes between glycine and closo-borate anions and their derivatives [B_nH_n−1X]^y− (X = H, NH₃, OH, SH, F; n = 10, 12; y = 1, 2) were investigated using several approaches. First of all, rather unusual NH···B contacts were observed for complexes based on closo-decaborate derivatives. The introduction of an exo-polyhedral substituent to the closo-borate cluster cage promoted stronger binding with glycine than was seen with unsubstituted closo-borate anions [B_nH_n]²⁻. This binding occurred due to the formation of non-covalent NH···X (X = O, S, F) and CO···HX (X = O, S) contacts. NH···O contacts for [B_nH_n−1OH]²⁻ Iso_1 systems were the strongest of all the molecular complexes considered. Complexes based on ammonium and hydroxy derivatives of closo-borate anion [B_nH_n−1X]^y− (X = NH₃, OH) were the most stable among all the derivatives considered. The complex of [B₁₀H₉NH₃]⁻ with glycine was the most energetically favourable.

Thus, the quantum chemical calculations described in this paper can be applied for the construction of closo-borate anion molecular complexes with biomolecules, in particular with proteins. First of all, the use of closo-decaborate anion derivatives facilitated stronger binding with biomolecules. In addition, it is suggested that the introduction of exo-polyhedral ammonium groups into the closo-borate could promote strong binding between the cluster anion and the target biomolecule. Consequently, the presence of hydrogen bond acceptor groups in a biomolecule promotes strong binding to [B_nH_n−1NH₃]⁻ anions. An alternative to the ammonium derivatives is the hydroxy derivative of closo-borate anions [B_nH_n−1OH]²⁻. An advantage of these systems is their ability to form strong NCIs with both hydrogen bond donors and acceptors.

Future work will extend upon the results presented in this manuscript. Further research will involve calculations of molecular complexes with different amino acid types (featuring charged, uncharged, and hydrophobic side chains), followed by studies of complexes with oligopeptides and proteins. Upon completing these investigations, we will be able to formulate more precise design principles for constructing non-covalent complexes of closo-borate systems with various biomolecules. Predictions of the cluster’s binding mode with biomolecules, derived from our calculations, could be applied to the development of potential drugs for BNCT, as well as antiviral and antimicrobial agents.