Structural and Molecular Packing study of Three New Amidophosphoric Acid Esters and Assessment of Their Inhibiting Activity Against SARS‐CoV‐2 by Molecular Docking

Study of the intermolecular interactions and packing features by Hirshfeld surface (HS) analysis

The supramolecular features of the structures 1–3 are studied by using Hirshfeld surface analysis,[ ¹³ , ³³ ] an easy‐to‐operate graphical tool to visually investigate various intermolecular interactions in molecular crystal structures. The graphical 3D Hirshfeld surfaces (HSs) mapped over the d _norm parameter and the derived 2D fingerprint plots (FPs) for 1–3 are illustrated in Figures 4 ​55–6 and S1, S2, S3. For 2, the HS mapped over the shape index property is shown in Figure 5. The intermolecular contacts and their contribution percentages are also reported in Table 3.

Open in a separate window

Figure 4

Views of the d _norm Hirshfeld surface maps (up) plotted on the molecule P1 of 1 in two orientations, introducing close contacts, along with the H…H and H…O/O…H fingerprint plots (down).

Open in a separate window

Figure 5

Views of the d _norm Hirshfeld (up) and shape index (down) surface maps plotted on the compound 2 in two orientations, introducing close contacts, along with the full fingerprint plot labeling features related to the main intermolecular contacts (down).

Open in a separate window

Figure 6

Views of the d _norm Hirshfeld surface maps (up) plotted on the compound 3 in two orientations, introducing close contacts, along with the H…H and H…O/O…H fingerprint plots (down).

The characteristic red spots near the oxygen atom of the P=O group and the hydrogen atom of the N−H unit in the Hirshfeld surface of the structures 1 and 2 confirm the previously mentioned N−H…O=P classical hydrogen bond. For the molecules P1 and P31 of 1, some little bright red spots on white regions are also seen which are related to H…H contacts (Table 2). The weak C−H…O=P hydrogen bond interactions in 1 and 2 (Table 2) are represented as white regions, while for 2, some of these interactions visible as little pale‐red spots stabilize the structure. In the case of 3, in the absence of the classical N−H…O=P hydrogen bonds, the C−H…O=P hydrogen bond interactions play the important role in the stabilization of the molecular packing and are revealed as red spots. Moreover, for 2 bearing the aromatic substituent pyridine ring, the red π‐hole features on the shape index surface (Figure 5) affirm the presence of π‐effect interactions of the C−H…π type ( Table 2).

In the FPs of 1 and 3, the contributions originate only from H…H and H…O/O…H interactions, where the packing is dominated by H…H contacts, accounting for as high as 76.0 % and 75.0 % of the packing forces in 1 and 3, respectively. The percentage contribution of H…O/O…H contacts to the packing is 23.3 % (for 1) and 24.2 % (for 3). In the case of 2, the largest contribution is found for the H…H contact (53.1 %), followed by the H…O/O…H (25.1 %), H…C/C…H (13.3 %) and H…N/N…H (8.0 %) interactions. The larger variety of contacts in 2, including H…H, H…O/O…H, H…C/C…H and H…N/N…H interactions, compared with 1 and 3 having only H…H and H…O/O…H interactions, can be attributed to the presence of the aromatic substituent aminopyridine which provides the π‐effect C−H…π interactions, i. e. H…C/C…H type contacts. Furthermore, this substituent bearing the nitrogen atom in aromatic ring gives rise to the H…N/N…H contacts.

The combined H…O/O…H contacts for 1 and 2 appear as large symmetrically sharp spikes with minimum values of d _i+d _e≈2.0 Å for 1 and 1.8 Å for 2, whereas these spikes are small for 3 (d _i+d _e≥2.4 Å). This result can be attributed to the presence of the classical N−H…O=P hydrogen bond interaction providing the lowest values of d _i+d _e in 1 and 2 versus the absence of such classical interaction in 3. However, all such close H…O/O…H contacts in 1–3 are at values less than the sum of the van der Waals (vdW) radii of the oxygen and hydrogen atoms (≈2.7 Å). Moreover, the H…H contact in 1 leads to a sharp spike on the diagonal plot reflecting some short H…H contacts with d _i+d _e≈2.0 Å which are less than 2×the vdW radius of hydrogen atom (≈2.2 Å).

The wings at the top left and bottom right of the plot of 2 are related to a series of H…C/C…H contacts, while the fraction of these contacts to the packing is 13.3 %. Other regions of the plot are occupied by the H…N/N…H contacts whose fraction to the total surface accounts for only 8.0 %.

Enrichment ratio (E_XY)

The propensity of pair atoms (X, Y) to form contacts can be assessed by calculating the enrichment ratios E _XY from the Hirshfeld surface fingerprint plots. For a high propensity the X…Y contact is recognized as favored when E>1 and for a low propensity the interaction is disfavored when E<1.[ ²⁴ , ²⁵ ] The enrichment ratios of the compounds 1–3 computed by the results derived from the HS analysis have been given in Table 3.

The results of the enrichment ratios analysis affirm the importance of O…H/H…O contacts including the classical N−H…O=P hydrogen bond interaction in the crystal packings of 1–3 for which these contacts are favored (E _OH>1). In contrast, the H…H contacts are disfavored in all three compounds with E _HH=0.91–0.99, where the lowest value of E _HH is found for 2. As is mentioned above, the presence of the aromatic substituent in the structure of 2 supports the formation of the C…H/H…C and N…H/H…N contacts inside the H…H and O…H/H…O. The existence of such extra contacts in 2 compared with 1 and 3 can underestimate the importance of the H…H contacts in this structure to be completely disfavored. This is in line with the obtained results of the C…H/H…C and N…H/H…N ratios for 2 which are larger than unity (Table 3) and turn out to be enriched.

Molecular docking study

In order to predict the intermolecular framework formed between the studied compounds 1–3 and the target macromolecular structure generating a binding model, a molecular docking approach was employed. For this study, the small molecules 1–3 are docked into the target protein 6M03 (Figure 7) to calculate the energetic values at the binding sites, where 6M03 is one of the important proteins which plays a key role in the propagation of the coronavirus. The simulations were done on each of structures selecting the best pose of the compound toward the receptor from the ranking of different poses by the highest negative binding energy. The best ranking binding energies were computed as −6.35, −5.98 and −5.47 kcal/mol for 1–3, respectively. The results show that all studied compounds give efficient binding energy with negative values. These obtained data confirm that the interaction of 1–3 with 6M03 can inhibit the activity of this protein and in result, could prevent the coronavirous function. Thus, amidophosphoric acid esters 1–3 are potential inhibitors of M^Pro of SARS‐CoV‐2 to be confirmed by in‐vivo experiments.

Open in a separate window

Figure 7

Left: Visual description of the 3‐D compounds 1–3 (up to down, respectively) in the best position relative to 6M03 showing the highest negative binding affinity; Right: The 2D representation maps of 1–3 with amino acid sites inside of the active pocket of 6M03.

Figures 7 and S4 display the various interactions between the compounds 1–3 and the structure of the 6M03 protein. For 1, hydrogen bonding (HB) interactions are formed between the hydrogen atom of NH group and the oxygen atom of the esteric ring with the amino acid residues of GLU A166. The five‐membered ring of amino substituent of 1 interacts with the MET A165, LEU A167 and PRO A168 binding sites. The carbon atoms of the esteric ring also interact with the amino acid residues of MET A165, MET A49, CYS A145, HIS A164 and HIS A41. For 2, the oxygen atom and the carbon atom of the CH₃ group of the esteric ring are interacting with CYS A145. An interaction is also found between the aromatic pyridine ring with MET A165. Other observed interactions are formed between the HIS A163 with the CH₃ groups of the esteric ring. In the case of 3, the obvious interactions are the HBs formed between the P=O oxygen atom with the LYS A5 and ARG A4 and between the esteric ring with the LEU A282, PHE A3, TRP A207 and PHE A291.

For a detailed investigation, we performed the docking simulations for the main protein of SARS‐CoV‐2 (6M03) on similar compounds 4–8 which have been previously reported[ ¹⁹ , ²⁶ , ²⁷ ] using the crystal structure data available in the Cambridge Structural Database (CSD). All these structures are in the family of the amidophosphoric acid ester with a same [OCH₂C(CH₃)₂CH₂O]P(O) segment, but with different amino substitutions (Scheme 2). The goal is to investigate the effect of the amino substitions on the biological acivity of these such structures. The molecular docking figures of 1–8 are shown in Figs. S4 and S5, and the related data reported in terms of binding energy (kcal/mole) are summarized in Table 4.

Open in a separate window

Scheme 2

Chemical structures of the selected compounds 1–8 for docking simulation.

The largest negative binding energies are found for compounds 1, 2, 4, 5 and 8, respectively, where for compounds 2 and 8, the aromatic substituent comes into play in the amino segment. Compound 2 bearing the nitrogen atom at the aromatic pyridine ring has a slightly higher binding energy of 0.08 kcal/mole compared to that of 8. The lowest negative binding energy is for compound 7 bearing the small aliphatic amino substituent with a different value of 1.22 kcal/mole compared to the highest value in 1. Thus one may conclude that the presence of the aromatic substituents in the amido segment of the studied amidophosphoric acid esters provides a more favorable condition for their biological activities than those bearing the aliphatic amido substituents. Similar results have been obtained in a molecular docking study of some phosphoric triamides, where compounds bearing the aromatic substituents illustrated more negative binding affinity energies compared to those having only aliphatic substituents. ^[14]

Moreover, a comparison of the various interactions between the compounds 1–8 (Figure S4 and S6) and the structure of the 6M03 protein illustrates that the phosphoryl group is involved in an interaction with some amino acid residues of 6M03 for more compounds. The methyl groups of the esteric ring of structures for all compounds are connected to some amino acid residues of protein by various interactions, and for those including the NH group, amino acid residues of the protein are taking part in interactions with the NH group. However, in the case of the binding energy values of compounds with or without the NH group, one cannot find a specific trend.

Syntheses and crystallization

The studied compounds were prepared in a similar method by the treatment of a solution of the commercial starting material ClP(O)[OCH₂C(CH₃)₂CH₂O] (2 mmol) in dry acetonitrile (20 ml) with a solution of related amine (4 mmol, cyclopentylamine, 2‐aminopyridine and pyrrolidine, respectively for 1–3) in the same solvent (5 ml) at 273 K. The resulted solution was stirred for about 4 h on the stirrer. Then, the solvent was evaporated and the product was washed with distilled water. Single crystals of all three compounds were grown by slow evaporation at room temperature from a solution of the related compound and mixture solvents of CH₃OH/DMF (with volume ratio 3 : 1).

Data for 1: IR (KBr, ῡ, cm ⁻¹ ): 3200 (N−H), 2966, 2893, 1454, 1387, 1309, 1225 (P=O), 1182, 1086, 1013, 949, 822 (P−N), 785, 624, 511, 473; ¹ H NMR (300.13 MHz, DMSO‐d ₆ , 300.0 K): δ=0.87 (s, 3H, CH₃), 1.01 (s, 3H, CH₃), 1.41 (m, 4H, CH₂), 1.58 (m, 2H, CH₂), 1.73 (m, 2H, CH₂), 3.40 (m, 1H, NH), 3.81 (m, 2H, CH₂), 3.97 (m, 2H, CH₂), 5.14 (m, 1H, CH); ¹³ C NMR (75.47 MHz, DMSO‐d ₆ , 300.0 K): δ=20.48 (s, 1 C, CH₃), 20.97 (s, 1 C, CH₃), 22.86 (s, 2 C, cyclopentyl), 31.44 (d, ³ J(C,P)=5.1 Hz, 1 C, C(CH₃)₂), 34.06 (d, ³ J(C,P)=6.0 Hz, 2 C, cyclopentyl), 52.64 (s, 1 C, cyclopentyl), 75.23 (d, ² J(C,P)=5.7 Hz, 2 C, 2CH₂); ³¹ P{¹H} NMR (121.49 MHz, DMSO‐d ₆ , 300.0 K): δ=4.25 (s).

Data for 2: IR (KBr, ῡ, cm ⁻¹ ): 3161 (N−H), 2968, 1612, 1597, 1475, 1309, 1251 (P=O), 1215, 1063, 1010, 943, 860, 833, 787 (P−N), 623, 511, 486; ¹ H NMR (300.13 MHz, DMSO‐d ₆ , 300.0 K): δ=0.85 (s, 3H, CH₃), 1.15 (s, 3H, CH₃), 3.90 (m, 2H, CH₂), 4.17 (m, 2H, CH₂), 4.20 (m, 1H, NH), 6.88 (m, 2H, Ar), 7.61 (m, 1H, Ar), 8.15 (m, 1H, Ar), 8.74 (s, 1H, Ar); ¹³ C NMR (75.47 MHz, DMSO‐d ₆ , 300.0 K): δ=19.97 (s, 1 C, CH₃), 21.45 (s, 1 C, CH₃), 31.73 (d, ³ J(C,P)=6.8 Hz, 1 C, C(CH₃)₂), 77.03 (d, ² J(C,P)=6.8 Hz, 2 C, 2CH₂), 110.98 (d, ³ J(C,P)=7.8 Hz, 1 C, Ar), 116.56 (s, 1 C, Ar), 138.07 (s, 1 C, Ar), 147.71 (s, 1 C, Ar), 154.19 (d, ² J(C,P)=2.6 Hz, 1 C, Ar); ³¹ P{¹H} NMR (121.49 MHz, DMSO‐d ₆ , 300.0 K): δ=–6.65 (s).

Data for 3: IR (KBr, ῡ, cm ⁻¹ ): 2986, 2846, 1477, 1348, 1267 (P=O), 1207, 1114, 1055, 1009, 970, 922, 874, 825 (P−N), 808, 635, 563, 490, 432; ¹ H NMR (300.13 MHz, (DMSO‐d ₆ , 300.0 K): δ=0.85 (s, 3H, CH₃), 1.07 (s, 3H, CH₃), 1.76 (m, 4H, CH₂ of pyrrolidinyl ring), 3.09 (m, 4H, CH₂), 3.79 (m, 2H, CH₂ of pyrrolidinyl ring), 4.06 (m, 2H, CH₂ of pyrrolidinyl ring); ¹³ C NMR (75.47 MHz, DMSO‐d ₆ , 300.0 K): δ=20.17 (s, 1 C, CH₃), 21.07 (s, 1 C, CH₃), 25.85 (d, ³ J(C,P)=9.3 Hz, 2 C, pyrrolidinyl ring), 31.46 (d, ³ J(C,P)=5.1 Hz, 1 C, C(CH₃)₃), 46.07 (d, ² J(C,P)=5.1 Hz, 2 C, pyrrolidinyl ring), 75.37 (d, ² J(C,P)=5.7 Hz, 2 C, 2CH₂); ³¹ P{¹H} NMR (121.49 MHz, DMSO‐d ₆ , 300.0 K): δ=3.36 (s).

Crystal structure determination

The details of the crystal data, data collection and structure refinements of 1–3 are given in Table 1. All three measurements were performed on an Rigaku Oxford Diffraction Gemini diffractometer with graphite monochromated Mo Kα radiation (λ=0.71073 Å) using a Sapphire3 CCD detector. Frame integration, scaling and absorption correction using redundant reflections was carried out with CrysAlisProint ^[29] The structures were solved using the charge‐flipping method from Superflip program ^[30] with default parameters described elsewhere. ^[31] The CRYSTALS program ^[32] and OLEX2 programs ^[33] were used for the structural refinements. The three structures were refined using two different approaches, one with spherical form factors and the other with aspherical form factors for so‐called Hirshfeld atom refinements (HARs) ^[34] in which electron densities are determined from molecular wavefunctions calculated by quantum mechanical methods and then partitioned into atomic Hirshfeld electron density functions. ^[35] The calculation of the wavefunctions on the PBE level with a def2‐TZVPP basis set and a high SCF convergence threshold were done with the ORCA program.[ ³⁶ , ³⁷ , ³⁸ ] The essential step of HAR is to perform a Fourier transformation of the Hirshfeld electron density functions to obtain the related aspherical atomic form factors that are used in the subsequent least‐squares refinements. The NoSpherA2 module in OLEX2 was used for the aspherical form factor refinements. ^[39]

For the refinements with spherical form factors H atoms were all located in a difference map, but those attached to carbon atoms were repositioned geometrically. The H atoms were initially refined with soft restraints on the bond lengths and angles to regularize their geometry (C−H in the range 0.93–0.98 Å and N−H in the range 0.86–0.89 Å) and U _iso(H) (in the range 1.2–1.5 times U _eq of the parent atom), after which the positions were refined with riding constraints. ^[40]

The structures refined with aspherical form factors were used as starting point for HAR. It was not possible to refine all hydrogen positions completely freely and for a number of these distance restraints set at their approximate neutrons values were used during the refinements. The atomic displacement parameters were refined isotropically. R _F improved in all cases slightly. The final HARed structures were used for the molecular docking studies as it is believed that these give more accurate binding energies than the structures refined with spherical form factors. ^[15] The molecular graphics were drawn by using the programs PLATON ^[41] and Mercury. ^[42]

Hirshfeld surface analysis

The software package Crystal Explorer 17.5 ^[43] was utilized to draw the colored graphical maps (Hirshfeld surfaces (HSs)) and plots (fingerprint plots (FPs)) by using the crystallographic information file (CIF) as input. Herein, the HSs are generated based on the d _norm distance, a symmetric function of distances from the internal (d _i) and external (d _e) atoms, and shape index.[ ⁴⁴ , ⁴⁵ ] The FPs are derived from HS as a 2D grid formed by (d _i, d _e) pairs to facilitate the study of intermolecular interactions. Moreover, the enrichment ratio (E) calculations resulted from the collected data from Hirshfeld surface analysis is employed for a detailed evaluation of the intermolecular contacts.[ ²⁵ , ⁴⁶ ]

Support of this investigation by Semnan University is gratefully acknowledged.