Ferrocenyl Hydroxy Methyl Coupled Caffeine Synthesis, Density Functional Theory, ADME Prediction and Docking Analysis of It's Binding Interactions to Coronavirus Main Protease

The COVID-19 pandemic has made the world aware of how crucial the development of cost-effective and scalable antiviral drugs is. Here we report the synthesis of caffeine-based 8-[(ferrocenyl)(hydroxy)methyl]-1,3,7-trimethyl-3,7-dihydro-1H-purine-2,6-dione (FHC), and its use as an inhibitor of protease (M pro ), an essential enzyme for SARS-CoV-2 viral replication. FHC was modelled through density functional theory to get an insight of its properties, and fully characterized through conventional techniques. Its activity against M pro was investigated using a molecular docking approach, showing excellent binding affinity to the catalytic dyad of His41 and Cys145 and the active sites of M pro with energies score ranging from −6.7 to −7.0 kcal/mol. The affinity of conformers to bind to the active pocket was 44%. Based on a detailed investigation, it appears that FHC has a safe ADME profile, and that it could be a potential inhibitor for M pro of SARS-CoV-2.


INTRODUCTION
Naturally abundant caffeine (1,3,7trimethylxanthine) is the most widely consumed psychostimulant drug worldwide.Due to its many therapeutic properties, caffeine is considered as a unique compound in the field of medicinal chemistry.The attention of the scientific community towards caffeine is due to its ability to comply with the rule of five, including the moderate lipophilicity (experimental logP−0.01,other sources logP−0.07),5] Another study reported the inhibition of the viral protein synthesis by caffeine and the suppression of the plaque formation of HSV-1 by interfering with the cell-to-cell transmission of the virus. 15It was found that Poliovirus and HSV-1 replication are sensitive to caffeine, suggesting that caffeine can indeed inhibit both DNA and RNA viruses. 160][21][22][23][24][25] The unique structure of Fc, featuring an iron metal sandwiched between two (Cp) rings leads to the electron donating ability of Fc which results in several chemical properties such as the ease of reversable oxidation to ferricenium ions, the high reactivity in electrophilic substitution reactions and the ease of functionalization. 26,27 c also shows lipophilic and non-toxic properties, good solubility in most of common organic solvents, stability in aerobic and aqueous media, good air and thermal stability up to 400. 288][39] The fact that Fc groups possess lipophilicity facilitates its penetration through cellular and nuclear membranes. 24RS-CoV-2 is a beta coronavirus of group 2B with over 70% genetic sequence similarity to that of SARS-CoV-1.7,40 SARS-CoV-2 has shown a higher infection rate and a more extended incubation period, as compared to previous coronaviruses such as Middle East respiratory syndrome coronavirus.41 Thus, it has become crucial for medicinal chemists to develop strategic research plans to produce anti-viral drugs for COVID-19 treatment.There are two enzymes that control proteolysis, the coronavirus main protease (M pro ) and the papain-like protease (PL pro ). 42 Snce the Mpro enzyme, also called 3CL pro , is essential to the viral replication and infection process, it has been considered as an ideal target for antiviral therapy.[43][44][45] Our previous study showed the ability of caffeine and its derivatives to function as inhibitor against M pro with binding affinity score ranging from -4.9 to -8.6 kcal/mol.18 Among the investigated drugs, caffeine showed a high binding affinity via hydrogen bonding to Glu 166 and Cys 145, with a relatively high binding percentage (67%) and low RMSD value.Further, the inhibitory activity of M pro with 42 pyrimidonic pharmaceuticals were also investigated.46 The study revealed that the high molecular weight drugs interact poorly to the active site of M pro , whereas the low molecular weight drugs interact with poor binding energy.Taking this balance into account, we designed and sinthesized a new biologically active caffeine containing drug and studied its potential as a therapeutic candidate against M pro of SARS-CoV-2 by molecular docking studies.Our design combines caffeine with the well-known biologically active Fc moiety.

Computational details
Density functional theory (DFT) calculations were performed using Gaussian 09 software suite. 47he system was geometrically optimized, using the B3LYP functional, and a mixed basis set, which includes the 631++G(d,p) for every atom except Fe, for which a LanL2DZ basis set was employed.A further frequency calculation was run to ensure no negative IR frequencies were present.The NMR spectrum was simulated in water using the mixed basis set: 6311G+(2d,p)/LanL2DZ. TMS, calculated at the same level of accuracy was used as a reference.A TD-DFT calculation, using a CAM-B3LYP functional and the same basis set employed for the ground state calculation, was performed to determine the frontier molecular orbitals of FHC and to determine the UV-Vis absorption spectrum in water.

Molecular docking
The M pro crystal structure was obtained from the Protein Data Bank (PDB ID: 6Y2E).The water molecules were removed, missing hydrogens atoms were added, and minimized to its lower energy conformation. 46The net charge was calculated using antechamber, and its energy was minimized using Molecular Modelling Toolkit plugin UCSF Chimera.The AutoDock Vina has been used to perform docking with grid box size of (29.08, 58.58, 57.04) Å, cantered at (−16.19× −26.14 × 17.46) Å. UCSF Chimera was used for visualization of bonds interactions and images processing. 46

Synthesis
The synthesis of the target compound FHC is shown in Scheme 1. FHC was synthesised via the reaction of commercially available compound ferrocenecarboxaldehyde with caffeine in the presence of an excess amount of sodium hydroxide (NaOH).NaOH was aimed to deprotonate the caffeine creating the anionic derivative which, in turn, attacks the electrophilic carbonyl of ferrocenecarboxaldehyde. FHC was obtained in 65% yield and was characterized using FT-IR, HR-MS and 1 H-NMR techniques.Since FHC is only soluble in water, 1 H-NMR spectrum (Fig. 1) was recorded in D 2 O.The spectrum shows two main broad singlets peaks at chemical shifts of 2.66 and 4.77 ppm.The peak at 2.66 ppm was assigned to the protons of the methyl groups while the peak at 4.77 ppm was assigned to the protons of cyclopentadienyl (Cp) of the Fc which are overlapping with the solvent residue peak.The main important indication of the successful substitution is the disappearance of the caffeine aromatic peak which was reported to appear at 7.58 ppm. 48Although the hydroxyl proton (-OH) was not confirmed by 1 H-NMR as expected due to proton exchange with D 2 O, FT-IR spectrum (Fig. 2) shows the presence of a broad stretch band from 2904-3672 cm −1 which was assigned to the OH group.TOF-MS using positive and negative ionization mode of electrospray is shown in Fig. 3.The TOF-MS spectrum indicates the presence of the weak parent molecular peak with M of m/z 408.3736, as well as M + using the positive mode (Fig. 3(a)) while the negative mode MS spectra shows a strong M -of m/z 407.1982 (Fig. 3(b)).The UV-Vis absorption spectrum (Fig. 4) shows a wide absorption band in the UV region from 283 nm to 400 nm with a maximum absorption l max at 291 nm.The optical band gap of FHC is 2.6 eV which was calculated from λl onset (481 nm).

Theoretical studies
In order to gain insight into the electronic structural properties of FHC, Density Functional Theory (DFT) calculations were conducted.The geometrically optimized structure of FHC (Fig. 5) was calculated using the B3LYP functional, and a mixed basis set, which includes the 631++G(d,p) for every atom except Fe, for which a LanL2DZ basis set was employed.A further frequency calculation was run to ensure the absence of negative IR frequencies.As shown in Fig. 5, the caffeine and the Cp units of the ferrocene lay on orthogonal planes (dihedral angle of 89.3 o ), likely due to the presence of the sp3 carbon that breaks the conjugation between the two units.
A TD-DFT calculation, at the same accuracy level, was performed to determine the frontier molecular orbitals of FHC.Both, the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) lay over the caffeine unit (Fig. 6) Their values are −6.16eV for the HOMO and −1.53 eV for the LUMO, corresponding to a calculated energy gap of 4.63 eV.The NMR spectrum was simulated in water using the mixed basis set: 6311G+(2d,p)/LanL2DZ. TMS, calculated at the same level of accuracy was used as a reference.Due to the rigidity introduced by the geometry optimization, the model does not take into account the rotational averaging of chemically equivalent protons, which could be represented by different chemical shifts.For this reason, their chemical shifts were averaged and reported in Table 1, with atom labels reported as indicated in Figure 7. involving the HOMO and other higher unoccupied molecular orbitals, but also a deeper occupied orbital and the LUMO.Interestingly, all the contributions to the absorption arise from transitions involving orbitals that lay over the caffeine unit.The close approximation of the calculated UV-Vis spectrum of FHC to the experimental one (Fig. 4), which shows a l max of 302 nm in solution and 291 nm in thin film, gives an idea of the robustness of the parameters chosen for the DFT calculation.It can be seen that the predicted chemical shifts can be grouped into two areas, the first is in the range of 2.22-2.94ppm while the second from 3.15-5.02ppm.This is in a close approximation to what has been obtained experimentally (Fig. 1).A TD-DFT calculation, at the same accuracy level, was performed to determine the UV-Vis absorption spectrum of FHC in water.The spectrum is shown in Fig. 8, which reports two maxima of absorption (l max ) at 229 nm and 293 nm.As summarized in Fig. 9, the band at higher wavelengths (green arrow) is due to the HOMO-LUMO transition.Its onset is around 350 nm, corresponding to an optical band gap of 3.5 eV.The band at lower wavelengths is the result of a combination of multiple transitions (purple arrows)

ADME Predictions
It has become indispensable to predict the absorption, distribution, metabolism, and excretion (ADME) and drug-likeness profiles of a candidate at the early stages of drug discovery, saving considerably in terms of time and expense. 48The most terminated candidate drugs in clinical phases are for safety and toxicity reasons rather than due to their efficacy.The behavior of FHC was investigated using two web servers, Swiss ADME . 48As a result of the FHC, an 82.05 topological polar surface area (TPSA) was obtained, corresponding to acceptance ranges of 50.16 to 110.87.Table 2 also showed their lipophilicity and water solubility, which were all within acceptable limits.Lipinski's rule of five states that candidates' violation values should not exceed one; The FHC does not violate any drug likeness rules and has passed all tests (Table 3).Based on a permanent BBB test, FHC does not appear to have any adverse effects on the central nervous system.Moreover, it does not inhibit any essential enzymes in the cytochrome system.In addition, FHC has a high value of GI absorption.These data from Swiss ADME suggest that FHC has an ideal pharmacokinetic profile.

Molecular docking
0] The binding sites between FHC and target main protease (M pro ) were revealed through blind docking to shed unbiased light on the interaction by searching the entire receptor.In M pro , proton transfer from the thiol of Cys 145 to the imidazole ring of His 41 activates the catalytic dyad.2][53] FHC has shown an excellent binding affinity to the active pocket of M pro with a score ranging from −6.7 to −7.0 kcal/mol.The binding percentage of the conformers docked to the active site was 44%.FHC conformers formed hydrogen bonds to ASN 142 and GLU 166 with a distance range of 2.118 Å to 2.444 Å.
Figure 10 shows the interaction of FHC to the active site of M pro .Our results show that the binding score energy of caffeine linked to ferrocenyl hydroxy methyl has significantly improved compared to our previously reported study on non-substituted caffeine which showed a binding score energy of −5.6 ± 0.30 kcal/mol. 18The aforementioned residue plays a key role in the inhibition of the protease activity.Therefore, these results suggest that FHC might prevent the formation of non-structural proteins that are essential for viral replication and, therefore, the formation of new virions.

CONCLUSION
This work presents the synthesis, D F T, a n d A D M E p r e d i c t i o n s o f a n ove l ferrocenyl caffeine derivative FHC, which was successfully synthesized by coupling ferrocenecarboxaldehyde to caffeine.The target compound was characterized using FT-IR, NMR, and HR-MS, and UV-Vis.A DFT calculation was carried out to gain insight into electronic and structural properties as well as the HOMO-LUMO interactions and the simulation of 1 H-NMR and UV-Vis spectra.The experimental and theoretical data show a good agreement.In a molecular docking study, FHC shows excellent binding affinity to M pro active residues with shallow binding score energy.An in-depth investigation and discussion of FHC's interactions with M pro suggest that this agent has a safe ADME profile and could be a potential inhibitor for M pro of SARS-CoV-2.

Fig. 10 .
Fig. 10.The FHC conformers docked with M pro , with particular attention to HIS 41, CYS 145, ASN142, and GLU 166.Nitrogen atoms are blue, oxygen atoms are red, and hydrocarbons are cyan in FHC.Blue lines depict hydrogen bonds; yellow lines represent van der Waals forces