Carnosine to Combat Novel Coronavirus (nCoV, COVID-19): Molecular Docking and Modeling to Co-crystallized Host Angiotensin-Converting Enzyme 2 (ACE2) and Viral Spike Protein

Aims: Angiotensin-converting enzyme 2 (ACE2) plays an important role in the entry of coronaviruses into host cells. This paper described how carnosine, a naturally occurring supplement, can be an effective drug candidate for coronavirus disease (COVID-19) on the basis of molecular docking and modeling to host ACE2 co-crystallized with COVID-19 spike protein. Methods: First, the starting point was ACE2 inhibitors and their structure-activity relationship (SAR). Next, chemical similarity (or diversity) and PubMed searches made it possible to repurpose and assess approved or experimental drugs for COVID-19. In parallel, at all stages, authors performed bioactivity scoring to assess potential repurposed inhibitors at ACE2. Finally, investigators performed molecular docking and modeling of the identified drug candidate to host ACE2 cocrystallized with COVID-19 spike protein. Results: Carnosine emerged as the best known drug candidate to match ACE2 inhibitor structure. Preliminary docking was more optimal to ACE2 than the known typical angiotensin-converting enzyme 1 (ACE1) inhibitor (enalapril) and quite comparable to known or presumed ACE2 inhibitors. Viral spike protein elements binding to ACE2 were retained in the best carnosine pose in SwissDock at 1.75 Angstroms. Out of the three main areas of attachment expected to the co-crystallized protein structure, carnosine bind with higher affinity to two compared to the known ACE2 active site. LibDock score was 92.40 for site 3, 90.88 for site 1, and inside the active site 85.49. Conclusion: Carnosine has promising inhibitory interactions with host ACE2 co-crystallized with COVID-19 spike protein and hence could offer potential mitigating effect against current COVID-19 pandemic.

. Overview of the study methodology This led us to the scaffold general structure of ACE2 inhibitors that has been reported by Dales and Torres ( Figure 2) [15,19]. We considered the simplest chemical structure possible where R 1 and R 3 groups are replaced with hydrogen atoms, and R 2 is dropped ( Figure 2).
Step 1: Obtain the general structure of ACE2 inhibitors and SAR from a general Google search Step 3: Perform a similarity search on the general structure of

ACE2 inhibitors on ChemSpider and DrugBank
Step 4: Calculate and compare the bioactivity and similarity scores on molinspiration and ChemMine for identified similar and the general structure of ACE2 inhibitors Step 6: Perform PubMed searches on identified structures and their COVID-19 research Step 7: Proof of concept will follow in subsequent experimental validation in vitro and in animal models Step 5: Perform a molecular docking for carnosine using SwissDock and LibDock from BioVia ® protocols Step 2: Calculate and compare the bioactivity and similarity scores on molinspiration and ChemMine for eight ACEI and the general structure of ACE2 inhibitors Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 23 October 2020 doi:10.20944/preprints202010.0486.v1 Next, as angiotensin-converting enzyme inhibitors (ACEI) work on the first known member in this enzyme family we randomly selected eight and compared to the general structure of ACE2 inhibitor. This step could enable us to distinguish if any ACEI would preferentially be more cross active to ACE2 (Figure 3). Evaluated ACEI included captopril, enalapril, ramipril, trandolapril, perindopril, benazepril, fosinopril, and temocapril. However, a full discussion of the relationship of ACE1 to ACE2 is beyond the scope of the current work.  Figure 3. Structures for the eight selected angiotensin-converting enzyme inhibitors Moreover, molinspiration bioactivity score calculator was utilized to predict the activity of the identified drug candidates at the six important ligands for ACE2; namely, 1: GPCR ligand, 2: ion channel modulator, 3: kinase inhibitor, 4: nuclear receptor ligand, 5: protease inhibitor, and 6: enzyme inhibitor [20]. On the other hand, ChemMine similarity scoring was used to check how close the molecule to the general ACE2 inhibitor scaffold structure was [21]. Therefore, the higher the following similarity metrics; AP Tanimoto, MCS Tanimoto, MCS Min, and MCS Max, the closer are the drug to the general scaffold ACE2 inhibitor. ChemSpider and DrugBank enabled practitioner to look for new compounds similar to the scaffold general structure of ACE2 inhibitors [22,23]. Going ahead of ourselves, carnosine was found to be the matching known ligand to the general scaffold ACE2 inhibitor. Bioactivity scoring was repeatedly run on every new potential or reference molecule identified. Absolute difference between the binding scores of each molecule and the general scaffold structure was calculated for each of the six ligands for ACE2 and summed up into a final total Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 23 October 2020 doi:10.20944/preprints202010.0486.v1 number. Hence, the lower this absolute summed total the closer supposedly the identified drug to the general scaffold structure.
Preliminary molecular docking was performed using SwissDock server as described by other research groups [24]. ACE2 code in SwissDock server is 6M0J. Note this is the co-crystallized structure of ACE2 with COVID-19 spike protein. Subsequently, reference compounds were used during docking to ACE2 including ACE1 inhibitor, enalapril and melatonin ( Figure 4). While enalapril is an ACEI, melatonin has shown good docking results at ACE2 (Figure 4) [25]. Binding modes of carnosine were visually presented with UCSF chimera version 1.14. To compare docking results of the various drugs both the lowest estimated Gibbs free ΔG energy for cluster 0 first elements and the summary of all binding modes were considered. protein was deposited in protein data bank (PDB) under the code (2AJF) [26]. It was retrieved from (PDB) and prepared using "prepare protein" protocol in discovery studio 2020 from Biovia®. The prepare protein parameters were left as default and this protocol will standardize atom names, insert missing atoms in residues and remove alternate conformations. Also remove water and ligand molecules, Insert missing loop regions based on either SEQRES data or user specified loop definitions, optimize short and medium size loop regions with the LOOPER algorithm, minimize the remaining loop regions and Calculate the pK and protonate the structure.
Analysis of the crystal structure which is a representation to the SARS-CoV-2 spike receptorbinding domain bound to the ACE2 receptor. It is revealed that there are three main areas that the two proteins come in proximity. Herein, none is clear which is more important than the other, so trying to prevent attachment of the two proteins will surly abort the entry of COVID to the host cell exploiting ACE2.
Within this work, carnosine will be docked at the three points of attachment to see its orientation and the pattern of binding in addition to docking the carnosine inside the active site of ACE2. The docking software is "LibDock protocol" from Discovery Studio 2020. The parameters were set to be as default except the "Docking preferences" was changed to "high quality" and the minimize algorithm was changed to "Smart minimizer" which performs 1000 steps of Steepest Descent with a RMS gradient tolerance of 3, followed by Conjugate Gradient minimization. The main amino acid in each of the three sites was used as a reference point for choosing the docking sphere and it was given Moreover, searches on PubMed were performed for carnosine and salicyl-carnosine using the following strategy: Proof of concept for this study should follow with in vitro or animal model experimentation and that will be our subsequent future step.

Results & Discussion
As detailed in the methods section, the general Google search led us to the general scaffold structure of ACE2 inhibitors from Dales and Torres ( Figure 2). The simplest form of this structure when R 1 and R 3 groups are replaced with hydrogen atoms, and R 2 is dropped served as our reference in all subsequent bioactivity scoring and chemical similarity searches.

ACE Inhibitors and COVID-19
Clinicians are still in doubt about whether to continue or discontinue ACEI in COVID-19 patients, but the consensus currently is that there is no evidence that it is harmful to continue ACEI [27].
Comparing the molinspiration bioactivity scores for the selected ACEI has shown none to be similar to the general structure of ACE2 inhibitors from Dales and Torres (Table 1). However, on the basis of total absolute difference from the scaffold general structure of the various ligands, one would predict that enalapril could be the best ACEI to use for patients with COVID-19 (i.e. lowest absolute difference) followed possibly by ramipril (Table 2). ACEI similarity with the scaffold general structure was the highest for enalapril considering AP Tanimoto and MCS Min scores (Table 2). Although it would vary with the similarity scoring method, combining molinspiration bioactivity scores and similarity search, one would expect that enalapril is the best ACEI to use or continue in patients with COVID-19.   Table 1 and  [34]. Back to preliminary docking, there were a total of 42 clusters and 253 elements all having favorable binding. Carnosine has a more favorable full fitness to ACE2 (-3416.80 kcal/mol) than enalapril (-3281.90 kcal/mol) and melatonin (-3365.10 kcal/mol). An estimated ΔG for carnosine (-6.29 kcal/mol) was a bit lower than enalapril (-7.42 kcal/mol) but only slightly less than melatonin (-6.57 kcal/mol). Figure 5 shows the distribution of clusters (elements) for a full fitness and an estimated ΔG for the three drugs; carnosine, enalapril, and melatonin. It can be easily inferred that overall carnosine had an overall full fitness and estimated ΔG much better than both enalapril and melatonin for all clusters combined. All comparisons were statistically significant (P value less than 0.05). Carnosine best pose showed parts of the viral spike protein ligand binding with ACE2 retained at 1.75 Å, and as result, both protein-protein interactions are vulnerable to inhibitory actions by carnosine in this model. Clearly, carnosine is anticipated to be an inhibitor of the protein-protein or at least a good starting point to design potent such ACE2 inhibitors. Based on a previous assessment of other drugs including most ACEI including ramipril, one can infer that carnosine is a better inhibitor of ACE2 than chloroquine and hydroxychloroquine, both initially implicated as good drug qualifiers for COVID-19 [35]. Moreover, results of preliminary docking of carnosine to ACE2 were comparable to those of melatonin, yet another candidate drug for COVID-19. However, carnosine figures of preliminary docking are in contrast slightly less than other identified ACE2 inhibitors by Chikhale et al such as withanoside X with an estimated ΔG (-7.07 kcal/mol) and ashwagandhanolide (-6.50 kcal/mol) [36].
Nevertheless, what makes carnosine probably stand out is that it is a widely commercialized supplement and hence can easily be studied or mobilized in the fight against COVID-19.

Detailed Molecular Docking and Modeling
It is expected that there are two possible mechanisms in which carnosine acts as an inhibitor of COVID-19. First, via its inhibition through binding to the active site especially chelating zinc atom. And second, through preventing the interaction between COVID-19 spike protein with ACE2. Based on these assumptions, our research group has conducted a docking study to evaluate the binding pattern of carnosine to the ACE2 active site and another study to investigate the binding of carnosine to the protein-protein interaction sites between the two proteins. Figure 6 shows the two proteins interaction surface based on the crystal structure 2AJF from PDB. Clearly, there are three major attachment points between the two proteins. So, we docked carnosine molecule to the three areas on the surface of ACE2 to evaluate its binding pattern and its docking scores (Figure 7). The protein was prepared using "Prepare protein" protocol, and then the prepared protein used by assigning the three major sites to be docked by selecting the key amino acids in these areas. LibDock protocol from Discovery Studio 2020 was utilized to perform the docking and the high quality option was used to get the best results from the docking process.  On the other hand, the docking of carnosine inside the active site was done to investigate the binding pattern and its score. The LibDock score was 85.49 and as Figure 9 shows In summary, carnosine seems to prefer sites 3 and 1 more so than ACE2 active site while at the same token performs essential interactions of ACE2 inhibitors. As a result, it is expected that it will be an important supplement to take to the next level of evaluation for COVID-19 and that is in vitro kit testing and animal modeling.

Literature Evaluation
PubMed search strategy yielded only three publications with commentaries on using carnosine and salicyl-carnosine, suggested for patients with COVID-19 [37][38][39]. In these papers, carnosine and Dengue and Zika viruses [40]. To further highlight the potential of carnosine, it has been shown it is also active in ameliorating lung injury associated with the swine flu [41]. Finally, the detailed molecular docking and modeling showed that carnosine to prefer two binding sites at the proteinprotein interaction surface while at the same time performing essential interactions at the ACE2 active site.

Limitations
There are several important limitations for this chemical analysis. First, similarity in the chemical structure may sometimes fail to translate into clinical similarity. Yet, the fact that the general ACE2 inhibitors structure is quite close to carnosine could at least be the starting point to make new more matching lead molecules. Second, the scaffold general structure and the structures of carnosine, salicyl-carnosine, and histidine have several chiral centers. However, this can be simply taken into consideration while synthesizing or testing such molecules in further studies. Finally, even the recommended scaffold and other structures by Dales and Torres may fail in the pre-clinical and clinical phases of research. Nevertheless, the urgency to find quick answers for current COVID-19 pandemic is far from being a luxury anyway, and only real experimental validation in vitro, animal models, and patients can prove or disprove these drugs for the management of SARS-CoV-2.

Conclusions
This is the first, to our best knowledge, preliminary and detailed molecular docking and modeling study that shows carnosine has probably excellent fit with an inhibitor activity against

Funding
This research received no external funding.