Michael Acceptors Tuned by the Pivotal Aromaticity of Histidine to Block COVID-19 Activity.

The question of whether COVID protease (SARS-CoV-2 Mpro) can be blocked by inhibitors has been examined, with a particularly successful performance exhibited by α-ketoamide derivative substrates like 13b of Hilgenfeld and co-workers (Zhang, L., et al. Science 2020, 368, 409-412). After the biological characterization, here density functional theory calculations explain not only how inhibitor 13b produces a thermodynamically favorable interaction but also how to reach it kinetically. The controversial and unprovable concept of aromaticity here enjoys being the agent that rationalizes the seemingly innocent role of histidine (His41 of Mpro). It has a hydrogen bond with the hydroxyl group and is the proton carrier of the thiol of Cys145 at almost zero energy cost that favors the interaction with the inhibitor that acts as a Michael acceptor.

The emergence of a new type of coronavirus is responsible for the most widespread pandemic of the 21st century in the western world. Even though the possibility that such a virus could generate a pandemic was randomly predicted by several doctors, and even Bill Gates from Microsoft in a TED talk in 2015, it was still unexpected. This problem of globalization must make us active agents in finding the first tools with which to fight the virus and then in developing vaccines to prevent it.[1,2] Currently, there are no targeted and effective therapeutic treatments for fighting this virus. Recent basic research, combining structure-assisted drug design, virtual drug screening, and high-throughput screening, led to the identification of new drugs that target the COVID-19 main protease SARS-CoV Mpro. This enzyme plays a pivotal role in mediating viral replication and transcription, and a solution might be a drug that monitors its activity. Specifically, Jiang, Rao, Yang, and co-workers identified a mechanism-based inhibitor,[3] labeled N3 (Scheme a), with an electrophilic carbon atom capable of interacting with the thiol group of the protease, upon determination of the crystal structure of COVID-19 virus Mpro in complex with the inhibitor. Next, through a combination of structure-based virtual and high-throughput screening, assays of >10000 compounds (from approved drugs to drug candidates in clinical trials) were performed to check the inhibitory effect of N3 on Mpro. The values of IC50 ranged from 0.67 to 21.4 μM. Almost at the same time, Hilgenfeld and co-workers presented similar results on other crystal structures (Figure ),[4] with an α-ketoamide inhibitor 13b (Scheme c). Promisingly, the pharmacokinetic characterization of the optimized inhibitor reveals pronounced pulmonary tropism and its suitability for inhalation administration.

Scheme 1

Lewis Structures of Inhibitors (a) N3, (b) 13a, and (c) 13b

Figure 1

Region around the sulfur of Cys145 of Mpro in the X-ray structure in space group C2.

All of the X-ray structures containing both SARS-CoV-2 Mpro and any inhibitor have some similarities[5] and interestingly suggest the formation of an oxyanion hole on a keto group of the inhibitor,[6−8] facilitating C–S bond formation after the action of the thiol group of the protease. Biologically, the projects converge, but chemically, the active center has not been analyzed in detail yet to understand the mechanism, which is the key to blocking the activity of the Mpro.

In particular, the work of Hilgenfeld characterized accurately by X-ray two types of space groups (C2 and P212121, which contains two different monomers, A and B) together with inhibitor 13b (Figure ). The available X-ray structures containing the inhibitors are golden nuggets for computational chemists, because they provide the detail for excluding conformational studies that would be unaffordable.

Nonetheless, from them, an in-depth analysis using precise quantum chemistry should guide future experiments.[9] Here density functional theory calculations model the area (10 Å) around the thiol group of Cys145, but including the whole inhibitor and main amino acids around it from Mpro (see the computational details in the Supporting Information). Inhibitor 13b establishes a significant number of H-bonds when inserted into the cavity where the thiol group is located and deprotonates it to form the corresponding C–S bond (Figure a). The insertion of 13b provides strong H-bonds, and they are thermodynamically crucial as they release 19.2 kcal/mol; C–S bond formation releases an additional 4.8 kcal/mol.

Figure 2

(a) Reaction of inhibitor 13b with Mpro. (b) Transition state (TS) assisting in C–S bond formation. (c) TS assisted by a water molecule in C–S bond formation. (d) Tautomers of the imidazole of histidine and its protonated conformation. TSs of proton transfer from the thiol of Mpro and His41 to the N atom in (e) α and (f) β. The main distances are in angstroms.

The direct reaction between the thiol group of Cys145 and the inhibitor may appear to be feasible, with C–S bond formation and proton transfer from sulfur to oxygen. For the sake of consistency, quantum mechanics calculations calculated an energy barrier of nearly 30 kcal/mol (Figure b), even assisted by a water molecule (Figure c) present in the medium. Actually, water molecules are present in the X-ray structure, and there is enough space to rotate them and even create a chain of two to three water molecules, to facilitate the protonation of the keto group that leads to the next C–S bond formation. Nevertheless, our body works at 36 °C, and this energy barrier could not be overcome. Experiments in the lab were performed at room temperature, thus completely against those computed energy barriers. Moreover, His41 in the latter scenario is simply a spectator, because it creates a strong H-bond between the NH group of His41 and the oxygen atom of the former keto group belonging to 13b that is attacked by the thiol of Cys145. Nevertheless, His41 can be even more active. Consequently, the capacity to deprotonate the thiol group by His41 was also checked, and computationally, its imidazole ring enjoys nearly free rotation [1–2 kcal/mol kinetically speaking (Figure d)]. Additional proof that this is not the right mechanism is that the deprotonation of the thiol group still requires at least 30 kcal/mol using the standard parametrized values in biology or via calculations (260–270 kcal/mol), missing here a strong base,[10] but using a histidine as an acceptor, only 20.9 kcal/mol. This energy cost almost disappears when this protonated imidazole does not dissociate, with rather little kinetic effort [4–6 kcal/mol depending on whether it deprotonated the N in α or β (Figure e,f)] because we can find a proton in either nitrogen atom. This is a priceless hint about how C–S bond formation could take place in a mild way.

The reason for the ambivalent character of the two partially protonated tautomers of the imidazole of the histidine,[11] as well as its cationic dually protonated one, is that the aromaticity persists in all three species (Figure d). In detail, moving from the charged biprotonated form to any of the neutral monoprotonated tautomers,[12] the aromaticity differences on the NICS index are meaningless. Actually, NICS(1) is even 0.2 ppm more negative, and thus more aromatic (see Table S2).[13] Moreover, in the X-ray structures of Hilgenfeld and co-workers,[4] the pairs of C–N bond distances are similar for both N atoms [for the X-ray structure in space group C2, the C–N distances are 1.383 and 1.335 Å for the N(H) atom closer to the hydroxyl and 1.325 and 1.386 Å for the more distant N]. This confirms that their nature is similar, with the corresponding N atom protonated or not. Actually, the facility of obtaining a proton by this ring is extremely helpful for the thiol group of Cys145 to deprotonate. This protonated imidazole together with any water molecule around then can create an oxyanion activating the keto group of 13b,[14,15] where the positive charge on the carbon aims for the negative charge on the deprotonated sulfur, facilitating C–S bond formation.

To perform the 1,2-addition, the protonation of the keto group of 13b from the cationic His41 is seemingly barrierless. However, the two moieties would have to be rearranged to be close enough to facilitate this proton transfer, which automatically leads to the consequent C–S bond formation by a 1,2-addition. Thus, dynamics calculations will be mandatory to understand the degree of flexibility of His41 to be close enough to Cys145. In addition, both transition states in panels e and f of Figure show that, even though in the model of His41 the proton transfer to the N in α to the linking atom of His41 by 2.5 kcal/mol does not reflect that in α, the flexibility to obtain it will be more difficult than in the β one. Moreover, the product is more stable by 5 kcal/mol with a proton in the nitrogen in α, and a H-bond with the hydroxyl group in the other.

The low degree of steric hindrance around His41 was demonstrated by steric maps of Cavallo et al. (Figure ),[16,17] which confirmed that it is the less sterically hindered region in the first sphere around the thiol group where the reactivity exists. Those maps also help to confirm why the activity of inhibitor 13a is considerably worse (see Scheme b), for which its cyclohexyl residue (instead of the cyclopropyl residue in 13b) makes more difficult the mandatory free rotation of His41 to perform as an acceptor and donor of protons sequentially, instead of the predicted steric clash of the pyridine ring with Gln189. Ongoing studies are trying to unveil in more detail which substituents of the inhibitor are important in terms of steric hindrance and chemical reactivity because inhibitors 13a and 13b display another potentially reactive keto group.[18,19]

Figure 3

Steric map around the sulfur of Cys145 of Mpro from the crystallographic structure (space group C2) of the protease Mpro. On the z axis, there is the sulfur atom of Cis45 and the carbon atom of the carbonyl of C13 is at the origin, while its oxygen atom is on the x–z plane (in angstroms).

Overall, the sequence of the interaction between the inhibitor and the protease is somewhat different than expected; i.e., first the lone pair of the unsaturated nitrogen of His41 gets the proton of the thiol group, followed by the concerted transfer of the other proton of the other N of this imidazole together with the favored formation of the C–S bond. More importantly, the explanation comes from a concept like aromaticity, with its simplicity but its unproven existence as an absolute observable. However, by definition aromaticity is proven as an observable via proposed indices of aromaticity.[20] Here it can explain how the 1,2-addition between inhibitor 13b that acts as a Michael acceptor[21,22] and Mpro can stop or decrease the activity of replication of COVID-19. The closest histidine to the thiol group of Mpro thus facilitates C–S bond formation that blocks its activity.