Druggable targets from coronaviruses for designing new antiviral drugs.

Severe respiratory infections were highlighted in the SARS-CoV outbreak in 2002, as well as MERS-CoV, in 2012. Recently, the novel CoV (COVID-19) has led to severe respiratory damage to humans and deaths in Asia, Europe, and Americas, which allowed the WHO to declare the pandemic state. Notwithstanding all impacts caused by Coronaviruses, it is evident that the development of new antiviral agents is an unmet need. In this review, we provide a complete compilation of all potential antiviral agents targeting macromolecular structures from these Coronaviruses (Coronaviridae), providing a medicinal chemistry viewpoint that could be useful for designing new therapeutic agents.

Introduction

Coronaviruses are enveloped viruses grouped into the Coronaviridae family (subfamily: Coronavirinae) and order Nidovirales. The subfamily has four genera, Alpha-, Beta-, Gamma- and - Deltacoronavirus. Coronaviruses are known to infect bats, rodents, and can be transmitted to other mammals (Alpha- and Betacoronavirus) and birds (Gamma- and Deltacoronavirus). Alpha- and Betacoronavirus can infect humans and cause mild or severe respiratory syndromes - Human Coronaviruses (HCoV). Two severe respiratory syndromes epidemics were previously reported, being Severe Acute Respiratory Syndrome-coronavirus (SARS-CoV) in 2002, and Middle East Respiratory Syndrome-coronavirus (MERS-CoV) in 2012. In December 2019, in Wuhan, China, new cases of severe pneumonia were reported, which was later associated with the emergence of a novel coronavirus, initially called 2019-nCoV (2019 novel coronavirus). Taxonomic and phylogenetic analyzes of it pointed similarities with the SARS-CoV prototype, renaming it as SARS-CoV-2. Gradually, new cases have been reported in Asia, Europe, and Americas, quickly reaching a global scale and leading the World Health Organization (WHO) to declare the pandemic situation in March 2020.

Medicinal chemistry strategies aim to develop new candidates to progress through several standard levels in the drug development process, in which apply a broad variety of tools involving numerous related disciplines. Normally, these levels include (i) the selection of a biological target of medicinal interest and identifying initial hits and/or lead compounds; (ii) compounds’ design and development of an efficient structure–activity relationship (SAR) analysis in the lead-optimization process; (iii) application of in silico methods; (iv) screening of synthesized compounds performing in vitro biological assays; and (v) pharmacokinetics and in vivo studies. In recent years, the term “one-target-one-disease” has been growing, mainly due to the emergence of several diseases with high complexity. In this context, the discovery, biological evaluation, and optimization of active small molecules is the core of the drug discovery process. So far, this approach has allowed to discover potential lead compounds from natural and synthetic sources, as well as combinatorial compounds libraries. These last have been applied for both drug lead discovery and optimization,11, 12, 13 since several methods have been able to search diverse chemical libraries containing a wide range of natural, synthetic small-molecules (organics), and (non)peptidomimetic analogs. Notwithstanding these facts, computer-aided drug design (CADD) uses in silico methods that were developed to facilitate the identification of new active molecules by applying rigorous steps in drug discovery and development workflows, such as the creation of virtual libraries, molecular docking, and in silico screening.15, 16, 17 Additionally, it quickly identifies potential targets, optimizes the research, and avoids unnecessary expenses.18, 19 Among the CADD approaches, structure-based drug design (SBDD) is a very effective alternative since it allows that drug-like compounds to be designed based on the structure of a macromolecular target.20, 21 In SBDD, virtual chemical libraries containing millions of molecules and structural databases are utilized to verify the ability of ligands to interact with the target. Then, a potential drug-like is designed and synthesized considering its biological target in order to ensure that the new molecule acts specifically on the target chosen. Still, ADMET (absorption, distribution, metabolism, excretion, and toxicity) filters can be applied to enhance the probability of obtaining hits or leads with good drug-like properties.23, 24 Along with in silico methods increase, drug repurposing has emerged as an interesting alternative to re-discover (repurpose) known drug against new diseases. Considering this, Kumar et al. (2019) and Pillaiyar et al. (2020) have reported perspectives about the recent advances and challenges in drug repurposing for different diseases, including dengue fever, cancer, and central nervous system (CNS) disorders. By using drug repurposing, several recent studies have used this approach to identify potentially effective treatments for SARS-CoV-2 (COVID-19) pandemic.27, 28, 29, 30, 31, 32, 33, 34, 35, 36

In addition, clinical trials with adult patients applying different COVID-19 therapeutic strategies are being carried out around the world, as previously reviewed by Wang and coworkers (2020). The most common treatments under antiviral research used drugs such as remdesivir, ribavirin, favipiravir, kaletra (a compound preparation with lopinavir and ritonavir), chloroquine-hydroxychloroquine, interferon-α, and arbidol. Despite researchers’ efforts, no treatment or drugs has been really effective, pointing to a necessary search for new drugs, vaccines or peptide inhibitors. Recently, an important review study containing inhibitors for SARS-, MERS- and others human coronaviruses has been reported. In general, these compounds block the membrane fusion stage or the co-receptor interaction, providing valuable tools in the race against SARS-CoV-2.

Deeming the aforementioned background, in this review, we initially provide information concerning biological aspects from Coronaviruses (CoV), such as the viral cycle of infection and proteins’ organization. Thereafter, a deep, recent, and complete compilation of all the most relevant researches involving the development of potential antiviral agents against macromolecular targets from SARS-CoV, MERS-CoV, SARS-CoV-2 (COVID-19), and HCoV-229E, as well. Hereafter, we discussed these works in detail, including chemical structures (from natural and synthetic origins), viral strains, and molecular docking studies (showing covalent, hydrophobic, and hydrogen-bonding interactions). Subsequently, we provide a table compiling all inhibitors targeting CoV biological structures found in the literature. Finally, we believe that this ‘piece of knowledge’ represents a novel useful scientific source for designing new inhibitors as promising agents to fight against Coronavirus outbreaks.

Human coronaviruses (HCoV)

Alpha- and Betacoronavirus can infect the human respiratory tract, ranges from mild to more serious illness that can lead to death. Previously, six coronaviruses had been identified as capable of infecting humans, being HCov-OC43, HCoV-229E, HCoV-NL63, HKU1, SARS-CoV, and MERS-CoV.4, 39, 40, 41, 42, 43 The first epidemic of coronavirus took place between 2002 and 2003, caused by SARS-CoV, where 8,096 people contracted the virus and had an outcome of 774 deaths, exhibiting a 9% mortality rate.43, 44 Ten years later, MERS-CoV emerges in 2012, and the cases have progressed slowly and, despite the epidemic situation regressed, cases are still reported. By November 2019, a total of 2,494 cases and 858 deaths were reported, reaching a mortality rate of 35%.

It is considered that coronaviruses transmission origin can occur from a reservoir host to an intermediate host until finally infect humans. Genetic sequencing analyzes pointed out the origin of SARS-CoV, MERS-CoV, HCoV-NL63, and HCoV-229E from bats, while HCoV-OC43 and HKU1 are originated from rodents.46, 47, 48, 49, 50, 51, 52

The seventh coronavirus to infect humans was identified in December 2019, named SARS-CoV-2, causing a severe acute respiratory syndrome (SARS) and spreading the current pandemic scenario. The infection was namely COVID-19 (Coronavirus Disease 2019) and there was reported 13,150,645 cases and 574,464 deaths worldwide on July 15, 2020. There are no vaccines to either HCoV, therefore, this review will focus on the three types of Betacoronavirus that have demonstrated major impacts on global health (SARS-CoV, MERS-CoV, and SARS-CoV-2) and their possible targets in medicinal chemistry viewpoint.

SARS-CoV, MERS-CoV, and SARS-CoV-2

The first coronavirus epidemic emerged from an outbreak in Guangdong province, China, in November 2002 caused by SARS-CoV and was controlled in mid of July 2003, by isolating infected people.3, 43 At the time, it was characterized the transmission through direct contact with infected people or with contaminated fomites related to droplets and aerosols released by sick individuals. The virus infects airway epithelial cells that can worsen and become a pneumonia characterized as a SARS. Normally, the infection is characterized by fever, dyspnea, lymphopenia and, in some rare cases, diarrhea and other gastrointestinal symptoms, which may occur the active replication of the virus in the small and large intestine, which suggests a possible fecal-oral transmission.43, 54 Also, a prolonged clotting time and elevated liver enzymes were observed. Still, SARS-CoV is able to infect macrophages and dendritic cells, inducing the pro-inflammatory cytokines and chemokines releasing, which play an important role in the disease progression.55, 56 However, the exact mechanism involved in pulmonary consolidation and severity remains uncertain.

After the SARS-CoV epidemic in 2002–2003, MERS-CoV appeared in Jeddah, Saudi Arabia in 2012, causing infections in the lower respiratory system similar to those caused by SARS-CoV, and also with a similar transmission.4, 58, 59, 60, 61 Phylogenetic analyzes concluded that the MERS-CoV genome has almost identical proximity to the virus found in dromedaries, an animal widely used in the Middle East and, therefore, has direct contact with humans.62, 63, 64 Symptoms are also similar to those of SARS-CoV, and in more severe cases, there is progression to an acute syndrome of respiratory distress, septic shock, multiple organ failure, and death.4, 59, 60, 61 MERS-CoV induces the activation of the innate immune response, through the interferon I pathway, and as a mode of evading the immune system, the virus produces proteins that block the signaling of IFN and NF-κB pathways thus increasing viral replication and pathogenicity.65, 66, 67

Seven years later, in December 2019, a new coronavirus caused severe cases of pneumonia in the city of Wuhan, China, and triggered the second pandemic of the 21st century, after the swine flu caused by the Influenza A (H1N1) virus in 2009. SARS-CoV-2 (named after phylogenetic analyzes) has been the aim of molecular, immunological and pharmacological studies to understand the viral mechanisms, and what treatments would be viable, in addition to the rational development of vaccines. The virus transmission mode is person-to-person through respiratory droplets, as well as contact with contaminated fomites, in which the main symptoms are equal to SARS-CoV and MERS-CoV. The secretion of cytokines in COVID-19 has also been reported, as well as in SARS and MERS, and may lead to the so-called “Cytokine Storm Syndrome” presenting high-levels of analytes such as IL-6, MCP-1, VEGF, and IL-8, among others, contributing to hypotension and pulmonary dysfunction.

There are still no specific antivirals or effective vaccines for coronaviruses. Then, the only effective methods to control the disease are to avoid contamination promoting social distancing, reinforcement of hygiene procedures, and masks utilization. As COVID-19 is an ongoing pandemic caused by a highly contagious virus, the search for new effective antiviral drugs is extremely essential.

Viral structure and genome organization of HCoV

Virion organization revealed by cryo-electron microscopy (Cryo-EM) showed coronaviruses as spherical viruses with approximately 80 to 125 nm with a particular feature of protruding oligomers (Spike protein) projected on its surface, presenting a crown-like appearance and giving rise to its name.1, 70 Three other proteins make up the viral structure, namely membrane protein (M), envelope protein (E), and nucleocapsid protein (N). Within the viral membrane, the helical nucleocapsid, composed by the N protein, can induce a cell cycle delay, assists viral RNA during replication, and be complexed with the genetic material (Fig. 1 A).1, 71

Fig. 1

A. Schematic representation of the HCoVs viral structure and genome organization. HCoVs are spherical viruses, with oligomeric Spike proteins (S) that protrude on their surface, and interact with the cell receptors. The surface has also the lipid bilayer from the host cell with the membrane (M), the envelope (E) proteins, and the (+) ssRNA packed with the nucleocapsid protein (N) on a helical shape. B. Organization of the SARS-CoV, MERS-CoV, and SARS-CoV-2 genome. The HCoVs genome is a (+) ssRNA of ~30 kb encoding non- and structural proteins. The 5′ end encodes the ORFs 1a and 1b giving rise to the pp1a and pp1ab, which is cleaved into non-structural proteins. The 3′ end encodes the structural proteins that will be assembled into the viral particle. The figure shows a schematic representation of the SARS-CoV, MERS-CoV and SARS-CoV-2 genome, with an emphasis on structural and accessory proteins, highlighting the similarities and differences among them. Figure based on complete genome sequences from Genbank: AY274119.3, NC_019843.3, and NC_045512.2.

The coronavirus (CoV) genome is a single-stranded, positive-sense RNA of approximately 30 kb, with a 5́cap and a 3́poly (A) tail (Fig. 1B). The 5′ end shows the replicase gene encoding two Open Reading Frames (ORFs) 1a and 1b with an overlapping. The ORF1ab encode two polyproteins named pp1a (ORF1a) and pp1ab (ORF1a and ORF1b).1, 57 These polyproteins are then proteolytically cleaved, generating 16 non-structural proteins (nsP’s). The 3′ end of the genome encodes structural and accessory proteins interspersed within them. Both ends have untranslated regions (UTRs) that play a role in RNA viral transcription, replication, and synthesis. The 5′ end UTR has stem-loop structures that are necessary for these processes.1, 57

The replication cycle of HCoV

The initial contact with the host organism occurs through mucous membranes from the nose, mouth, and eyes. The coronaviruses bind the RBD domain from Spike protein (S) to the host cell, recognizing the human angiotensin-converting enzyme 2 (ACE2) receptor in infections by SARS-CoV and SARS-CoV-2,72, 73 and the dipeptidyl peptidase 4 (DPP4) receptor in infections caused by MERS-CoV74 (Fig. 2 ). This interaction ends up causing a conformational rearrangement resulting in the fusion between the viral membrane and the host cell membrane, after the cleavage of S protein at its S2 site and then releasing the nucleocapsid within the cell cytoplasm.1, 57, 75 Once inside the cell, the 5′ end from the RNA has the ORFs 1a and 1b translated into polyproteins pp1a and pp1ab, processed later into the nsP1-11 and nsP1-16, respectively.1, 57, 76 To encode the two polyproteins, the virus uses a stop codon sequence (5′-UUUAAAC-3′) and a pseudoknot RNA that causes a −1 ribosomal frameshifting upstream of the ORF 1a stop codon, allowing the continuous reading to ORF1b. Thus, the pseudoknot structure is unrolled until it finds the stop sequence in ORF1a, preventing continuous ribosome elongation, extending the conversion to ORF 1b and resulting in the translation of pp1ab.1, 57, 77

Fig. 2

Replication cycle of HCoV. The virus enters the host cell by binding protein S to the ACE2 receptor (SARS-CoV and SARS-CoV-2) or the DPP4 receptor (MERS-CoV), leading the viral membrane fusion with the cell membrane host (1). Fusion occurs because of S protein cleavage, allowing entry through the endosomal pathway (2). The viral RNA is released into the cell cytoplasm (3) and the pp1a and pp1ab polyproteins (4) are translated, which will be cleaved by the proteases of the RTC complex (5), synthesizing the (+) and (−) RNAs (6). From (−) gRNA, sgRNAs will be discontinuously transcribed until act as mRNAs (7) and finally being translated into the structural proteins (8) that will be transferred to the ERGIC intermediate complex (9). The N proteins then bind to the (+) gRNA in the cytoplasm and are assembled with structural proteins in the ERGIC complex (10). Vesicles containing virion are transported to the plasma membrane and released via exocytosis (11) in the extracellular space, thus infecting neighboring cells.

The ORF1a from SARS-CoV, SARS-CoV-2, and MERS-CoV has the PLpro (papain-like protease) domain within the nsP3, responsible for the processing of mature replicase proteins. Also, ORF1a presents another domain composed of a serine-type protease, 3CLpro (chymotrypsin-like picornavirus 3C-like cysteine protease or main protease (Mpro)) within nsP5.78, 79, 80 While PLpro does the cleavage at the limits between nsP1-2, nsP2-3, and nsP3-4, 3CLpro is responsible for 11 other remaining cleavage events.1, 57, 78 The nsP’s are organized in the replicase-transcriptase complex (RTC) and have other domains and enzymatic functions. For example, in ORF1b there are nsP’s encoding the RNA polymerase-dependent RNA domain (RdRp), RNA helicase, exoribonuclease, and a ribose dependent on 2′-O-methyltransferase activity.1, 57 All of these nsP’s with specific activities contribute to the construction of an environment that enables the synthesis of RNA, the replication and transcription of subgenomic RNAs (sgRNAs), and genomic RNAs (gRNA)57, 81 (Fig. 2).

The replication initiates with the synthesis of a negative polarity viral RNA that produces both gRNA and sgRNAs, which works as mRNA for the translation of structural and accessory proteins. From the RTC complex, there is the transcription of subgenomic mRNAs with 3′-coterminal end and gRNA, which have in common a leader sequence derived from the 5′ end of the genome. From there, the proteins are translated, and E and M proteins are translocated to the Endoplasmic Reticulum-Golgi intermediate compartment (ERGIC), where it is believed to be the site of assembly, budding, and transport of virions.1, 57, 82 The nucleocapsid proteins pack the genomic RNA to form the nucleocapsids on helical structures. The N protein then interacts with the E and M proteins, producing the virus-like particles (VLPs) that incorporate the S proteins. Mature virions are transported to the cell surface inside vesicles and are released into the extracellular space, allowing the interaction and infection of neighboring cells (Fig. 2).1, 57

Druggable targets from HCoVs

Spike glycoprotein (also named as S protein)

The CoV spike protein (S) is a type I transmembrane glycoprotein playing roles in viral attachment, fusion, and cell entry as well as tissue tropism. S proteins assemble into trimers protruding from the viral surface forming the crown-like appearance. It has two functional subunits named S1 (responsible for binding to the cell receptors) and S2 (acts in the fusion of the viral and cellular membranes).83, 84, 85, 86, 87

The S1 subunit forms the protein globular head and presents the receptor-binding domain (RBD) responsible for virus interaction with the host cellular receptors (e.g. ACE2 for SARS-CoV and SARS-CoV-2 or DPP4 for MERS-CoV),88, 73, 74 thus allowing the viral entry into the targeted cell. The receptor-binding S1 subunit of CoV S proteins has two domains, the N-terminal and C-terminal domains, either them can act as RBDs. The Cryo-EM structure of SARS-CoV-2 S trimer was performed87, 89 and a furin cleavage site at the S1/S2 boundary was identified which is known to be processed during biogenesis.

Recently, it was identified that RBD in SARS-CoV-2 S protein can bound strongly to both human and bat ACE2 receptors, and interestingly a higher binding affinity was detected for SARS-CoV-2 RBD than to SARS-CoV.89, 91 The crystal structure of the SARS-CoV-2 RBD bounded to the ACE2 was elucidated and is nearly identical to that of the SARS-CoV RBD. The SARS-CoV-2 entry requires cellular proteins to process S protein helping the infection. In this way, S protein is primed by cellular serine protease TMPRSS2 and a furin preactivation of the SARS-CoV-2 S protein-enhanced virus entry into some cell types. Similarly to the SARS-CoV, the cellular serine protease TMPRSS2 was pointed as a possible antiviral target for entry step of SARS-CoV-2.

The C-terminal S coronavirus S2 subunit forms a stalk-like structure anchored in the membrane during the virus fusion to the host cell membrane and presents heptad repeats (HRs) motifs. SARS-CoV and SARS-CoV-2 have an 89.8% sequence identity in their S2 subunits. After the binding of RBD at S1 subunit to the ACE2 receptor on the host cell, the membrane distal 1 (HR1) and membrane-proximal 2 (HR2) heptad repeat domains in the S2 subunit interact with each other, forming a six-helix bundle (6-HB) fusion core thus bringing viral and cellular membranes into proximity for fusion and infection.97, 98 Interestingly, it was shown that SARS-CoV-2 has a much higher capacity of membrane fusion than SARS-CoV.

S protein inhibitors

The attachment of SARS-CoV spike (S) protein to cellular Angiotensin-Converting Enzyme 2 (ACE2) is the first step in SARS-CoV infection. Several reports indicated that blocking the S protein attachment to cellular receptors could prevent the virus entry. Thus, the SARS-CoV S protein becomes an attractive target for the development of anti-SARS drugs.

Small-peptide-derived structures of HR regions from the SARS-CoV S protein have been shown to inhibit SARS-CoV infection by interfering in SARS-CoV fusion upon target cells.100, 101, 102, 103 Encouraged by these results, Ho et al. (2006) synthesized 14 small synthetic peptide derivatives from the S protein and screened all of them against the S protein attachment to the ACE2, and upon the S-protein-pseudotyped retrovirus infectivity, as well. Among these, the peptide SP-10 (residues 668–679 - STSQKSIVAYTM) (Fig. 3 ) was identified as the most promising protein–protein interaction (PPI) inhibitor, exhibiting 90% inhibition at 10 nmol concentration and blocking this interaction in a dose-dependent manner, with an IC50 value of 0.0018 μM (1.8 nM). The inhibitory effect of SP-10 on SARS-CoV S protein and Vero E6 cells was analyzed by Biotinylated Enzyme-Linked Immunosorbent Assay (ELISA), Immunofluorescence assay (IFA), and S-protein-pseudotyped retrovirus infectivity. Vero E6 cells were treated with BSA, Biotin-labeled S protein or SP-10/biotin-labeled S protein mixture. Moreover, the results indicated that SP-10 was capable of blocking the attachment of S protein to Vero cells. Additionally, it was investigated the possible domains involved in receptor interactions to identify biologically active peptides using peptide-scanning method, which involved overlapping peptides of 12 residues covering additional amino acids on both the N- and C-terminals from the SP-10 structure. The inhibitory effect of small overlapping peptides on the SARS-CoV S protein and ACE2 interaction was analyzed by competitive biotinylated ELISA. The receptor-binding regions from the SARS-CoV S protein have been defined in different studies.104, 105, 106 It was suggested that the region comprising residues 660–683 from the SARS-CoV S protein may interacts with ACE2. Therefore, SP-10 was the first small-peptide designed as a PPI inhibitor targeting SARS-CoV S protein attachment to the ACE2. According to these authors, it could be developed as an anti-SARS-CoV agent for the treatment of SARS-CoV infection.

Fig. 3

Chemical structure of SP-10 (peptide sequence: STSQKSIVAYTM), a small-peptide-derived from the SARS-CoV S protein.

Finally, another peptide-derived inhibitor (comprising the sequence: EEQAKTFLDKFNHEAEDLFYQSSGLGKGDFR) of SARS-CoV S protein is shown in Table 1 .

Table 1

Inhibitors targeting Coronaviruses found in the literature.

Structure	Source	Organism	IC50	Target	PDB - Interactions/(H-bond)	Ref.
Spike (S) protein
EEQAKTFLDKFNHEAEDLFYQSSGLGKGDFR	Synthetic	SARS-CoV	0.1 µM	S protein	Not revealeda	209
Chymotrypsin-like cysteine protease (3CLpro)
	Synthetic	SARS-CoV	57 µM	3CLpro	(PDB: 4TWW) - Cys145, His41, Met49, Met165, Asp187, His163(H), Phe140, Leu141, and Glu166	210
	Synthetic	SARS-CoV	30 µM	3CLpro	(PDB: 3AW1) - Cys145(H) and Gln189(H)	211
	Synthetic	SARS-CoV	5.8 µM	3CLpro	(PDB: 2ALV) - Gly143(H), Ser144(H), Cys145(H), His163(H), His41(H), Leu27, Met49, and Gln189	212
	Synthetic	HCoV-NL63	1.08 µM	3CLpro	(PDB: 6FV2) - His41(H), *Gly142(H), Phe139(H), and Glu166(H).	213
	Synthetic	SARS-CoV	0.24 µM	3CLpro	(PDB: 5N5O) - Met49, Met165, Asp187, Gln189(H), and Thr190	213
	Synthetic	MERS-CoV	1.7 µM	3CLpro	(PDB: 4RSP) - Glu166(H), Glu169(H), Gln192(H), Cys148 (Covalent), and His41	214
		SARS-CoV	0.2 µM	3CLpro	Not revealeda	214
	Synthetic	SARS-CoV	0.51 µM	3CLpro	(PDB: 2AMD) - Phe140, Leu141, His163, Met165, Glu166, His172, Asn142(H), Gly143(H), and Cys145(H)	215
	Synthetic	MERS-CoV	0.4 µM	3CLpro	(PDB: 5WKL) - Cys148, Gln192(H), *Gln167(H), Glu169(H), His41(H), *His166(H), and *Phe143(H)	216
		SARS-CoV	5.1 µM	3CLpro	Not revealeda	216
	Synthetic	MERS-CoV	0.6 µM	3CLpro	Not revealeda	216
		SARS-CoV	2.1 µM	3CLpro	Not revealeda	216
	Synthetic	SARS-CoV	26 µM	3CLpro	(PDB: 4TWW) - Cys145, His163(H), Phe140, Leu141, and Glu166	217
	Synthetic	SARS-CoV	95 µM	3CLpro	Not revealeda	218
	Synthetic	MERS-CoV	0.28 µM	3CLpro	(PDB: 1UK3) - Cys148 (Covalent), *His166(H), and His175(H)	219, 220
	Synthetic	SARS-CoV	4.6 µM	3CLpro	(PDB: 3VB5) - Cys145(Covalent), Pro168, Thr190(H), Glu166(H), Gln189(H), His164(H), Phe140(H), and His163(H)	221
	Synthetic	SARS-CoV	0.23 µM	3CLpro	Not revealeda	222
	Synthetic	SARS-CoV	6.1 µM	3CLpro	(PDB: 1UK4) - Glu166(H), Gly143(H), Cys145(H), Met49, and Gln189	223
	Synthetic	SARS-CoV	24.1 µM	3CLpro	Not revealeda	224
	Natural	SARS-CoV	8.3 µM	3CLpro	(PDB: 2Z3E) - Leu141(H), His163(H), Val186(H), Gln189(H), and Gln192(H),	225
	Natural	SARS-CoV	8.3 µM	3CLpro	Not revealeda	226
	Synthetic	SARS-CoV	1.0 µM	3CLpro	Not revealeda	227
	Synthetic	SARS-CoV	10.0 µM	3CLpro	(PDB: 1UK4) - His41, Met49, Phe140, Gly143(H), Cys145(H), His163, Met165, Glu166(H), Pro168, and Gln189(H)	228
	Synthetic	SARS-CoV	1.7 µM	3CLpro	Not revealeda	229
	Synthetic	SARS-CoV	10.0 µM	3CLpro	(PDB: 1WOF) - Cys145, Ser144(H), His163(H), His164(H), Glu166(H), and Gln189(H)	230
	Synthetic	SARS-CoV	0.74 µM	3CLpro	(PDB: 1WOF) - Cys145, Ser144(H), His163(H), His164(H), Glu166(H), and Gln189(H)	231
	Synthetic	SARS-CoV	5.0 µM	3CLpro	(PDB: 1UK4) - Glu166(H), His163(H), Cys145(H), Ser144(H), Asn142(H), Phe140(H), Thr26(H), and His41	232
	Synthetic	SARS-CoV	0.5 µM	3CLpro	Not revealeda	233
	Synthetic	SARS-CoV	3.0 µM	3CLpro	Not revealeda	234
	Natural	SARS-CoV	10.0 µM	3CLpro	Not revealeda	235
	Synthetic	SARS-CoV	13.0 µM	3CLpro	Not revealeda	236
	Synthetic	SARS-CoV	8.9 µM	3CLpro	Not revealeda	237
		SARS-CoV	2.5 µM	3CLpro	Not revealeda	238
	Synthetic	SARS-CoV	17.2 µM	3CLpro	Not revealeda	239
	Synthetic	SARS-CoV	1.5 µM	3CLpro	Not revealeda	240
	Natural	SARS-CoV	2.7 µM	3CLpro	(PDB: 2ZU5) - Thr190(H), His163(H), Ser144(H), His41(H), and Cys145(H)	241
	Synthetic	SARS-CoV	13.9 µM	3CLpro	Not revealeda	242
	Synthetic	SARS-CoV	1.04 µM	3CLpro	(PDB: 1UK4) - Met49, Leu141, Asn142, Gly143(H), Ser144, Cys145(H), Met165, Arg188, Gln189, and Gln192	243
	Synthetic	SARS-CoV	0.95 µM	3CLpro	(PDB: 1UK4) - Gly143(H), Cys145(H), His41(H), His164, Phe140, His163, and Met165	244
	Synthetic	SARS-CoV	0.3 µM	3CLpro	(PDB: 1UK4) - Met49, His41(H), Pro39, Leu27, Cys145, His164, Met165, Leu167, Gln192, and Gln189	245
	Synthetic	SARS-CoV	0.06 µM	3CLpro	(PDB: 2A5A) - Glu189, Met49, His41, Cys145, Gly143(H), His163(H), Phe140, His172, and Glu166	246
	Synthetic	SARS-CoV	0.051 µM	3CLpro	Not revealeda	247
	Synthetic	SARS-CoV	0.098 µM	3CLpro	(PDB: 3ATW) - Thr90(H), Glu166(H), His163(H), Phe140, Leu141, Asn142, His41, Met49, Met165, Asp187, and Cys145(H).	248
	Natural	SARS-CoV	27.5 µM	3CLpro	(PDB: 4WY3) - Gln166(H), Leu167(H), Phe140(H), Gln189(H), Asn142(H), Thr26(H), and Thr24(H).	249
	Natural	SARS-CoV	11.4 µM	3CLpro	(PDB: 2ZU5) - His163(H), Ser144(H), and Cys145(H)	250
	Synthetic	MERS-CoV	5.8 µM	3CLpro	(PDB: 4YLU) - His41, Ser147, Glu169(H), Leu170, Val193, and Gln195	212
	Synthetic	SARS-CoV2	0.67 µM	3CLpro	(PDB: 6Y2F) - His41(H), Gly143(H), Cys145(H), Ser144(H), Phe140(H), Glu166(H), and His163(H)	127
		SARS-CoV	0.90 µM	3CLpro	Not revealeda	127
		MERS-CoV	0.58 µM	3CLpro	Not revealeda	127
	Synthetic	SARS-CoV	9.6 µM	3CLpro	(PDB: 1UK4) - Glu166(H), Gly143, Gln192, Met49, Arg188, and Gln189	251
Papain-like protease (PLpro)
	Synthetic	SARS-CoV	(a) 0.46 µM and (b) 1.3 µM	PLpro	(PDB: 3E9S) - Tyr269, Gln270, and Asp165	252
	Synthetic	SARS-CoV	14.2 µM	PLpro	(PDB: 4M0W) - Cys271	253
		MERS-CoV	22.7 µM	PLpro	Not revealeda	253
		SARS-CoV	5.0 µM	PLpro	Not revealeda	254
	Synthetic	MERS-CoV	6.6 µM	PLpro	(PDB: 4RF1) - Tyr279(H), Ser167(H), Pro163, Asp164, Asp165, *Gly248, *Thr249, Pro250, *Phe269, Glu273, Ala275, Val276, Gly277, and Thr308	255
	Natural	SARS-CoV	0.8 µM	PLpro	Not revealeda	256
	Natural	SARS-CoV	5.0 µM	PLpro	Not revealeda	257
	Synthetic	SARS-CoV	(a) 0.39 µM and (b) 0.35 µM	PLpro	(PDB: 3 MJ5) - Cys112, *Leu163, Asp165, Pro248, Pro249, Tyr265, Tyr269, Gln270, Tyr274, Thr302, and Asp303	258
	Natural	SARS-CoV	1.2 µM	PLpro	(PDB: 3 MJ5) - His176(H) and His172(H)	250
	Synthetic	SARS-CoV	8.45 µM	PLpro	(PDB: 2FE8) - His290, Asp287, His273, Ala289, Lys106(H), Cys112(Covalent), and Trp107(H).	259b
		SARS-CoV2	2.26 µM	PLpro	(PDB: 6W9C) - *Leu289, Ala288, Asp286, Lys105, Tyr268, *Trp106, His272, and *Cys111(Covalent).	259b
NTPase/Helicase (nsP13)
	Synthetic	SARS-CoV	6 µM	nsP13	Not revealeda	260
			12 µM50 µM5.9 µM	Y277AnsP13K508AnsP13WTnsP13	(homology model) Tyr277, Arg507, and Lys508.	261
	Synthetic	MERS-CoV	2.5 µM	nsP13	(homology model) Tyr7, Tyr159, Arg163, and Tyr171.	261, 262
([Bi(L5) (H2O) (ClO4)3], L5 =	Synthetic	SARS-CoV	5.0 µM	nsP13	Not revealeda	263
	Synthetic	SARS-CoV	3.0 µM	nsP13	Not revealeda	264
	Synthetic	SARS-CoV	11.0 µM	nsP13	Not revealeda	265
Guanine-N7-Methyltransferase (also named as N7-MTase or nsP14)
	Synthetic	SARS-CoV	0.6 µM	nsP14	(PDB: 5C8T) - Trp385, Phe401, Tyr420, Phe426, Phe506, Cys387, Pro335, Val290, Trp292, Ile332, Phe367, Ala353, Val389, Asn386 (H), Arg310 (H), Gly333 (H), Ile338 (H), Lys336 (H), His424 (H), and Asp352 (H).	266

*: Divergent residue labeling due to the difference between the SARS-CoV, SARS-CoV-2, and MERS-CoV proteins’ structures. a: The authors have not performed in silico studies. b: Preprint study published online at BioRxiv, doi: 10.1101/2020.05.17.100768, available at: https://www.biorxiv.org/content/10.1101/2020.05.17.100768v1.

Envelope protein (also named as E protein)

The envelope protein (E) is the smallest structural protein that plays roles in the viral assembly, budding, virion release, and pathogenesis.82, 107, 108 Its structure has a short hydrophilic N-terminal, a large hydrophobic (TMD), and a hydrophilic C-terminal domain.82, 109 For the SARS-CoV, E protein was mainly detected in the endoplasmic reticulum-Golgi intermediate compartment (ERGIC), and only a small portion is incorporated into the virion envelope.

The C-terminal SARS-CoV E domain interacts with the PALS1, a tight junction-associated protein, and alters tight junction formation and epithelial morphogenesis in mammalian cells. Besides, the SARS-CoV E protein forms ion conductive pores in lipid bilayers, exhibiting an important function in the virus-host interaction. Also, it was shown that the activity of SARS-CoV E protein ion channel contributes to virus pathogenesis and it was required for inflammasome activation in infected-mice.

The SARS-CoV E protein-induced apoptosis in Jurkat T-cells and interacted with antiapoptotic protein BcL-xL suggesting a novel mechanism of T-cell apoptosis observed in SARS infection. In another study, it was reported that the E protein PDZ-binding motif involved in protein–protein interactions has a major role in SARS-CoV virulence using the cellular protein syntenin as a mediator of p38 MAPK induced inflammation.

E protein inhibitors

Expressed in SARS- and MERS-CoV, and most recently described in SARS-CoV-2, envelope (E) protein is a small membrane protein responsible for forming ion channels and is directly related to the viral infection, replication, dissemination through host tissues and modulation of the immune response.82, 117 Notwithstanding these functions, E protein has been considered as an important target in the development of selective inhibitors against CoV.

Among the classes of active compounds against E protein, acylguanidines (AG) standout for having several antiviral potential agents, where BIT225 is the first derivative from this class in clinical trials against hepatitis C (HCV) and immunodeficiency type 1 (HIV1) viruses, as well as AG derivatives from zanamivir and oseltamivir, potentially actives against influenza infections. Concerning this chemical class, Wilson et al. (2006) performed a study involving the hexamethylene amiloride (HMA, Fig. 4 ), an acylguanidine analogous to the amiloride drug, where a peptide corresponding to the E protein sequences from three coronaviruses in the GenBank database was synthetized. Thereafter, it was observed that the HMA significantly inhibited the conductance of HCoV-229E E protein ion channel (−70 mV) at 100 µM concentration. The authors reported its inhibitory activity with an EC50 value of 1.34 µM upon the HCoV-229E replication, as well as from the mouse hepatitis virus (MHV), with an EC50 value of 3.91 µM.

Fig. 4

Hexamethylene amiloride (HMA) inhibitor of E protein from HCoV-229E. Hydrogen-bonding interactions are shown as blue dotted lines.

Posteriorly, Pervushin et al. (2009) synthesized a peptide (sequence: E8TGTLIVNSVLLFLAFVVFLLVTLAILTALR-NH2) corresponding to the transmembrane domain of SARS-CoV E protein to evaluate the binding mode with the HMA. By using Nuclear Overhauser Effect (NOE) in NMR analysis, it was revealed that HMA has two binding sites in the SARS-CoV E protein, being one near to the Asp15 residue and another at Arg38. Besides, they also revealed that HMA nitrogen 5 is found in the protonated state (NH+) and bound to the ion channel, in which it is stabilized by hydrogen-bonding interactions with Asp15 and Arg38 amino acid residues, as a signal at 10.7 ppm (for Asp15) in the 1H NMR spectrum.

Chymotrypsin-like picornavirus 3C-like protease (3CLpro) (also named as main protease (M pro), or 3C)

The 3CLpro (also called Mpro) is a protease (corresponding to nsP5) highly conserved in HCoVs, and as well as PLpro, is responsible for the cleavage of the polyproteins pp1a and pp1ab playing an essential role in the viral replication.123, 124 This protease cleaves the polyprotein at 11 conserved sites involving the amino acid sequence Leu-Gln/Ser-Ala-Gly, which cleavage site is between the Gln and Ser. Although the cleavage pattern of 3CLpro seems to be conserved in SARS-CoV and SARS-CoV-2,123, 125 additional mechanisms are required to the polyprotein processing in the MERS-CoV, such as dimerization of the substrate-binding site to obtain a mature dimer.

The coronavirus 3CLpro was studied, revealing three domains. The N-terminal 3CLpro I and II domains are antiparallel β-sheets with 6 to 13 β-ribbons forming a chymotrypsin-like structure.123, 125 The catalytic binding site is located between these two domains and a loop structure links domain II to the C-terminal domain III, comprising a dyad of conserved residues (cysteine and histidine) both in SARS-CoV and SARS-CoV-2.123, 127 Domain III consists of a globular cluster of five helices associated with the protein dimerization due interactions between these domains of each monomer. Also, domain III has been associated with the proteolytic activity of 3CLpro.123, 125, 126

3CLpro inhibitors

Ghosh et al. (2008) reported the design and biological evaluation of a series of 5-chloropyridine ester-derived inhibitors against chymotrypsin-like protease (3CLpro) from SARS-CoV. This protease is crucial for viral replication, along with papain-like protease (PLpro), constituting important targets in the search for selective and effective antiviral inhibitors.129, 130, 131 Deeming this information, 11 chloropyridine ester derivatives were synthesized and biologically evaluated. Then, the indole-derived (Fig. 5 ) was found to be a covalent inhibitor, confirmed by MALDI-TOF analysis. Thereafter, it demonstrated to be the most promising inhibitor against SARS-CoV 3CLpro, with an IC50 value of 0.03 µM. Furthermore, it was verified that this compound presented activity against infected cells with the virus, exhibiting an EC50 value of 6.9 µM.

Fig. 5

Promising covalent inhibitor of 3CLpro from SARS-CoV. Covalent and hydrogen-bonding interactions are shown as red and blue dotted lines, respectively.

Computational studies of molecular docking by using GOLD® v. 3.2 software were carried out involving this target (PDB ID: 2HOB). It was suggested that the critical residues for the interaction between the indole-derived inhibitor and the protease involve a carbonyl group at position 4. Finally, the authors described that the distance between the carbon dioxide and the sulfur atom from the Cys145 residue prior to the nucleophilic attack is about 2.8 Å. Also, the distance between the nitrogen from the chloropyridinyl ring (as a H-acceptor group) and His163 residue is about 2.4 Å. Furthermore, the carbonyl oxygen atom is located between three backbone nitrogens, forming three hydrogen-bonding interactions with Cys145 (NH···O, 2.3 Å), Ser144 (NH···O, 2.4 Å), and Gly143 (NH···O, 2.8 Å) residues, increasing the stability of the complex and favoring the nucleophilic attack by Cys145 residue (Fig. 5). In this context, Ghosh et al. (2008) reported that in the covalently modified state after the aforementioned nucleophilic attack (PDB ID: 2V6N), the indole ring is placed into the S1 pocket, where the chloropyridinyl ring was before the nucleophilic attack, suggesting a dynamic enzymatic reorganization. Still, heterocycles indole and imidazole (from the His41 residue) are within a distance of 4 Å, suggesting that these rings have their initial position modified by the covalent reaction progress, reinforcing the aforementioned enzymatic reorganization.

Finally, diverse 3CLpro inhibitors from CoV have been reported in the literature and they can be found in Table 1.

Papain-like protease (also named as PLpro)

The PLpro is a cysteine protease encoded by nsP3 protein that recognizes the consensus cleavage sequence LXGG at the amino-terminal of replicase products in the cell cytosol during the virus replication. The tetrapeptide motif is found between nsP1/2, nsP2/3, and nsP3/4 proteins releasing the nsP1-2-3 from the viral polyprotein.78, 132

Additionally, the PLpro enzymatic activity is also characterized as the deubiquitinating activity of cellular proteins, possibly providing a favoring environment for intracellular viral replication. PLpro molecular structure resembles deubiquitinating enzymes (DUBs) responsible for cleaving both Ubiquitin and UBL-ISG15, which are important cellular modifiers that are covalently attached to target proteins by a peptide bond. Like DUBs, PLpro acts on this isopeptide bond cleaving viral proteins during the viral replication process.132, 133, 134

Also, it is known that due to the deubiquitinating activity, PLpro can inhibit the production of cytokines and chemokines having a significant role in the innate immune response against viral infection antagonizing IFN pathway.

The molecular structure and function of PLpro from SARS-CoV, MERS-CoV, and SARS-CoV-2 is very similar, with slight conformational differences. The monomer consists of four domains: ubiquitin-like (UBL), the thumb, the palm, and the fingers.136, 137 Previously, some studies showed that PLpro domains are conserved in all coronaviruses, and the same was observed with SARS-CoV-2. The catalytic core has a Zn-ribbon domain with four cysteine residues, demonstrated in HCoV-229E to be important for enzyme activity. The PLpro hydrophobic domain, important to the processing of the nsP3/4 site, is glycosylated, mediating an intracellular membrane association. It is hypothesized that the anchoring of PLpro may be important to the assemble of replication complex during the virus replication.

PLpro inhibitors

Ghosh et al. (2010) developed a new lead compound belonging to the class of piperidine carboxamide from a chemical library using high-throughput screening (HTS). The compound 6577871 (Fig. 6 ) exhibited an IC50 value of 59 μM. Also, studies involving lead optimization and structure–activity relationship (SAR) analyses were performed to provide improved inhibitors targeting PLpro and antiviral activity against infected Vero E6 cells with SARS-CoV. The (S)-Me (a) (IC50 = 0.56 μM for PLpro; and EC50 = 9.1 μM for antiviral assays) and (R)-Me (b) (IC50 = 0.32 μM for PLpro; and EC50 = 9.1 μM for antiviral assays) derivatives were the most potent compounds in this series (Fig. 6), showing equivalent enzymatic inhibition and antiviral activity. From the structure–activity relationship (SAR) analysis, the authors established that the introduction of a benzodioxolane ring led to more strong inhibitory activity than its precursor compound (6577871). Also, 1-(naphthyl)ethylamides are more potent than 2-naphthyl derivatives, and it was observed a preference for (R)-methyl enantiomers. The X-ray structure of complex for b-PLpro from SARS-CoV was determined in a 2.6 Å resolution and a model for a-PLpro complex revealed molecular features associated with interactions, in which it was shown a unique binding mode into the PLpro structure. In silico studies demonstrated that the activity of this series of compounds is not associated with stereoisomerism, in contrast to other compounds already synthesized. Superposing the structure of compound a on the crystal structure of the b-PLpro complex, the authors observed that there is an inversion of the piperidine ring between compounds a and b binding modes, which allows the naphthyl rings of both isomers to be accommodated at the active site, in a very similar orientation. Finally, the flexible piperidine ring acts as a spacer (or linker) group that enables the carboxamide (CONH) from both a and b compounds to perform hydrogen-bonding interactions with the backbone carbonyl oxygen at the Tyr269 residue, thereby retaining the potency of both enantiomers.

Fig. 6

Chemical structures of piperidine carboxamide analogs with activity against PLpro from SARS-CoV. Hydrogen-bonding interaction is shown as a blue dotted line. Bumping collision is represented by a yellow-ray for the compound a.

The crystal structure of the b-PLpro complex confirmed the presence of water molecules conserved between the apoenzyme and compound b. Water molecules conserved into the pocket, located between residues Asp165, Asp303, and Thr302, preventing naphthyl rings from occupying this pocket, while water nearby residues such as Leu163 and Lys158 prevent the forthcoming of the benzodioxolane ring to residue Lys158. The naphthyl ring is placed into the hydrophobic pocket formed by Pro248, Pro249, Tyr265, Tyr269, and Thr302 amino acid residues. Based on these interactions, it was observed that a gem-dimethyl substitution at the same position of methyl from the a or b inhibitors decreases the freedom degree around the carbon atom and locks the compound in a conformation in which one methyl group exhibits a bumping collision with the side chain of Asp165 residue when compared to the (R)-Me substituted analog (b). Therefore, SAR analysis, systematic modification guided by ligands-PLpro complex X-ray crystallography, and subsequent molecular modeling resulted in a potent inhibitor (b), with an IC50 value of 320 nM and antiviral activity with an EC50 value of 9.1 μM upon SARS-CoV-infected Vero E6 cells.

Lastly, other PLpro inhibitors from CoV have been reported in the literature and these compounds are shown in Table 1.

RNA-dependent RNA polymerase (also named as nsP12)

The enzymatic complex of HCoVs presents an RNA-dependent RNA polymerase (RdRp, also named nsP12) as a catalytic subunit for the synthesis of a negative-sense RNA, gRNA and sgRNA being a central component of coronaviral replication and transcription machinery. This protein has two domains: the N-terminal with nidovirus RdRp-associated nucleotidyltransferase (NiRAN) activity and the C-terminal canonic RdRp domain, connected by an interface domain.141, 142, 143

The RdRp adopts the conserved structure of viral polymerase family, of a cupped right hand composed of the fingers, palm and thumb subdomains in the C-terminal region, catalyzing the viral RNA during the replication and transcription cycle. The RdRp catalyzes the formation of phosphodiester bonds between ribonucleotides from an RNA template in the presence of divalent metal ions and thus synthesizing the negative-sense RNA, gRNA, and sgRNA.

It was showed that SARS-coronavirus RdRp (nsP12) needs to interact with nsP7 and nsP8 as a tripartite complex to activate its capability to replicate long RNA. Recently, the Cryo-EM structure of SARS-CoV-2 nsP12 in complex with the cofactors nsP7 and nsP8 and a newly β-hairpin domain at its N-terminal region was reported. For the MERS-CoV, it was demonstrated that nsP8 and nsP12 form an active complex.

RdRp (nsP12) promising inhibitors

Since RdRp (nsP12) is involved in the replication-transcription process in CoV, as well as in the production of genomic RNA, this target represents an interesting alternative for the development of new drug candidates in search of an effective pharmacotherapy against SARS-CoV-2.148, 149, 150, 151 Thus, recent researches have been developed, especially involving in silico methods, as well as drug repurposing,131, 152, 153 in order to inhibit the viral replication cycle and, consequently, reducing the infection.154, 155

Recently, in a study performed by Choudhury et al. (2020), molecular docking was used to analyze 30 FDA-approved drugs at the active site of the RdRp enzyme (PDB ID: 6M71) from SARS-CoV-2. Then, the authors demonstrated that the drugs chlorhexidine and remdesivir (Fig. 7 ) presented the best docking score values, being −132.84 and −114.46, respectively. Based on these values, the authors conclude that chlorhexidine was considered as the best inhibitor in the study, while the remdesivir was found to be the best among the listed antiviral drugs.

Fig. 7

Promising FDA-approved drugs virtually screened against RdRp from SARS-CoV-2.

Posteriorly, Pokhrel et al. (2020) performed a virtual screening of 1,930 FDA-approved drugs forward RdRp enzyme (PDB ID: 6M71) from SARS-CoV-2, utilizing the Chemoinformatic Tools and Databases, and also the AutoDock Vina to perform docking (by using ensemble, rigid, and flexible methods). In this study, the authors observed that quinupristin (Fig. 8 ), an antimicrobial active against Gram-positive bacteria, was the most promising ligand in all three types of docking studies, exhibiting a docking score of −12.3 Kcal/mol. Moreover, dactinomycin and sirolimus antibiotics (Fig. 8) also demonstrated encouraging results, with scores of −12.2 Kcal/mol for both, where these results could be used as a basis for the search for RdRp from SARS-CoV-2 pandemic.

Fig. 8

Promising antibiotic inhibitors against RdRp from SARS-CoV-2.

Searching for repurposing drugs against RdRp, Beg et al. (2020) performed a study based on molecular homology of RdRp from SARS-CoV-2 (PDB ID: 6NUR), using the SWISS-MODEL server, as well as molecular docking in the AutoDock Tools software. Then, the authors analyzed 74 different antiviral drugs, including HCV, HIV, human cytomegalovirus (HCMV), herpes simplex virus (HSV), human papillomavirus (HPV), varicella-zoster virus (VZV), and influenza virus. It was observed that the anti-HIV nelfinavir (−8.8 Kcal/mol), raltegravir (−8.7 Kcal/mol), and delavirdine (−8.5 Kcal/mol) drugs (Fig. 9 ), as well as anti-HCV paritaprevir (−10 Kcal/mol), beclabuvir (−10 Kcal/mol), and ledipasvir (−10 Kcal/mol) (Fig. 9) drugs were to be the most promising molecules against SARS-CoV-2 RdRp.

Fig. 9

Anti-HIV and -HCV promising drugs screened against SARS-CoV-2 RdRp.

Among anti-HIV drugs, the main amino acid residues involved in hydrogen-bonding interactions were identified as Ala576, Asn582, and Ser650 for delavirdine; Arg444, Asp509, Asp514, and Arg515 for nelfinavir; and Asp343, Arg444, Asp514, Thr571, Ser573, and Asn582 for raltegravir. Concerning anti-HCV drugs, hydrogen-bonding interactions were observed with Ser392, Asn398, and Asn434 for beclabuvir; Lys391, Arg446, Arg460, and Asp514 for ledipasvir; and Ser392, Asn398, and Lys402 for paritaprevir.

Finally, the aforementioned studies reported the search for RdRp inhibitors by using techniques of drug repurposing, in which promising compounds have been identified that could be useful against SARS-CoV-2 in the future. Furthermore, this compilation will serve as a basis to accelerate the search for promising and effective compounds to inhibit RdRp from SARS-CoV-2.

NTPase/Helicase (also named as nsP13)

Helicases are classified into six superfamilies (SFs) and those from the Nidovirales order belong to the SF1 superfamily and the Upf1 family (SF1B), characterized by moving in the 5′ → 3′ direction along the nucleic acid chain.143, 159 In coronaviruses, the SF1 helicase domain (HEL1) is located in the C-terminal region of the nsP13, from a cleavage product of pp1ab replicase. The N-terminal region of nsP13 contains a multinuclear zinc-binding domain (ZBD), which is one of the most conserved domains of the Nidovirales order. Structurally, the CoV HEL1 domain is formed by two RecA-like domains, named 1A and 2A.143, 160

For the SARS-CoV nsP13, both RNA and DNA duplex-unwinding activities were detected allowing the efficient strand separation of extended regions of double-stranded RNA and DNA in a 5′-to-3′direction. Besides, both (deoxy)nucleoside triphosphatase (dNTPase) and RNA-5′-triphosphatase activities were also reported being the SARS-CoV nsP13. Subcellular localization showed nsP13 on membranes from the endoplasmic reticulum, apparently the site of SARS-CoV RNA synthesis. The biochemical characterization of MERS-CoV nsP13 showed a dsRNA-unwinding activity with the helicase requiring more ATP for the optimal unwinding of RNA substrates with short 5′ loading strands. Recently, the NTPase and RNA helicase activities were reported for SARS-CoV-2 nsP13 and that bismuth salts can inhibit both these activities in a dose-dependent manner. There is speculation that SARS-CoV nsP13 may has a function in RNA capping because of the demonstrated RNA triphosphatase TPase activity, although further studies are necessary.

NTPase/Helicase (nsP13) inhibitors

Keum & Jeong (2012) evaluated the viral helicase as a potential target for developing chemical inhibitors against SARS-CoV. In this study, they analyzed the full genome sequence from SARS-CoV and verified that 2/3 of the SARS-CoV genome consists of viral replicase genes, which encode 16 non-structural proteins (nsP’s). Among these, RNA-dependent RNA-polymerase (named as nsP12 or RdRp) and NTPase/helicase (nsP13) are very important for viral replication. The SARS-CoV NTPase/helicase consists of 601 amino acids obtained from the replicase region,166, 167 and analyses of the amino acid sequence suggest that the helicase is divided into two separate domains. The first is a metal-binding domain (MBD) located at the N-terminus and, the second, a helicase domain (Hel). The SARS-CoV helicase uses ATP and dATP as preferred energy sources.161, 169 The authors conducted in vitro biochemical experiments to find natural compounds that might suppress either the DNA unwinding or the ATPase activity of the SARS-CoV NTPase/helicase. It was observed that none of the 64 compounds evaluated interfered with the DNA unwinding activity of the nsP13 protein. However, the flavonoid myricetin (IC50 = 2.71 µM) and flavone scutellarein (IC50 = 0.86 µM) strongly inhibited the ATPase activity of nsP13 from the SARS-CoV (Fig. 10 ). Nevertheless, other investigations are still required since it is unclear how myricetin and scutellarein suppress the ATPase activity.

Fig. 10

Natural compounds able to inhibit the NTPase/helicase from SARS-CoV.

Finally, other inhibitors of NTPase/helicase (nsP13) from CoV are displayed in Table 1.

Guanine-N-Methyltransferase (also named as N-MTase or nsP14)

The nsP14 has both exoribonuclease (ExoN) and guanine-N-methyltransferase (N-MTase) activities and both roles are important for viral replication and transcription. Coronaviruses replicate in the cytoplasm and have evolved strategies to cap their RNAs, with several nsP’s involved in this process. The SARS-CoV nsP14 was identified as a cap guanine-N-methyltransferase (N-MTase) producing the cap-0 structure (m7GpppN), from the family of S-adenosyl--methionine (SAM)-dependent methyltransferases. Several critical residues for the nsP14 methyltransferase activity on GTP were identified including F73, R84, W86, R310, D331, G333, P335, Y368, C414, and C416. The SARS-CoV nsP14 C-terminal N-MTase domain has a noncanonical MTase fold with a β-sheet insertion and a peripheral zinc finger.

Interestingly, the association of nsP10 with nsP14 stimulates the ExoN activity while does not affect the N-MTase activity.164, 173 In this way, one molecule of nsP10 interacts with N-terminal ExoN domain of nsP14 to stabilize it and stimulate its activity, and the presence of two zinc fingers are crucial for this nsP14 function.

In contrast with nsP14 MTase, SARS-CoV nsP16 also presented MTase methylation activity, in a sequence-specific manner to m7GpppA-RNA. Crystal structure showed an nsP16/nsP10 complex, where nsP16 binds to substrates m7GpppA-RNA and SAM cofactor with the assistance of nsP10.

N7-MTase (nsP14) inhibitors

Natural products from plants, animals, and microorganisms have been used to treat diseases since the beginning of human life.174, 175 Among diverse bacteria of medicinal interest, Streptomyces species (Gram-positive bacteria) have been widely reported as chassis organisms suitable for the development of bioactive molecules,176, 177, 178 constituting the main resource of antibiotics for clinical use.179, 180

Sinefungin (Fig. 11 ) constitutes an important amino acid-containing nucleoside, obtained from Streptomyces griseolus and S. incarnatus species, structurally related to S-adenosyl-methionine (SAM) and S-adenosyl-L-homocysteine (SAH), thus constituting an anticancer agent against metastasis in lung and primary breast tumors.182, 183, 184 Recently, promising antiviral activity against the Zika virus (ZIKV),185, 186 and Feline herpesvirus type 1 (FHV-1). In this context, Sun et al. (2014) described the activity of this natural product against SARS-CoV, exhibiting a promising activity at 1.68 µM concentration. Moreover, it acts as an inhibitor of the enzyme guanine-N 7-methyltransferase, which has been identified as a promising target for the development of new antiviral drugs against SARS-CoV since it inhibits the RNA-processing enzyme.

Fig. 11

Sinefungin, a pan-inhibitor against SAM-dependent methyltransferases.

The aforementioned authors identified this inhibitor by developing a genetic system as a high-throughput enzymatic activity assay platform, involving 3,000 natural product extracts as candidates for N 7-MTases inhibitors. These extracts were obtained from the natural microbial product library at Hubei Biopesticide Engineering Research Center (HBERC), in which secondary metabolites were isolated using classical methods. Additionally, it was concluded that sinefungin represents a broad-spectrum inhibitor, since it has shown activity upon N 7-MTase enzymes from different CoV species, such as murine coronavirus (MHV), transmissible gastroenteritis coronavirus (TGEV), and infectious bronchitis virus (IBV), with IC50 values of 1.89, 1.67, and 1.9 µM, respectively. Unfortunately, it does not constitute an ideal viral inhibitor since it has a poor selectivity index (SI) of 0.32, due to its results towards human N 7-MTase (IC50 = 0.55 nM) and SARS-CoV N 7-MTase (IC50 = 1.68 µM). In contrast, it contributes as a potential scaffold for designing new selective N-MTases inhibitors, in which it could be also used against SARS-CoV-2.

In a study developed by Aouadi et al. (2017) was described a high-throughput N 7-MTase assay based on Homogenous Time-Resolved Fluorescence (HTRF®), where they identified 20 potential compounds that inhibit SARS-CoV N 7-MTase. Therefore, 2,000 compounds obtained from the Prestwick Chemical Library®, including 1,280 small molecules (FDA- and EMA-approved drugs), 320 natural products, and 400 pyridazine-derived compounds were also screened in this study. Among these compounds, 20 best molecules were identified as potential antiviral agents. Lastly, the natural product, a polyphenol, demonstrated to be the most promising inhibitor targeting the SARS-CoV nsP14 (Fig. 12 ), with an IC50 value of 0.019 µM.

Fig. 12

Most promising inhibitor against guanine-N7-methyltransferase from SARS-CoV.

Furthermore, the authors also reported that sinefungin has shown inhibition <0.05 µM against human mRNA cap guanine-Nmethyltransferase (RNMT), which characterizes it has poor selectivity because of its low selectivity index (SI) value (<2.63). This low SI value could be associated with the structural homology between the SARS-CoV N 7-MTase and human RNMT forms, which limits the possibility of discovering specific inhibitors for this viral target.

Finally, another inhibitor of SARS-CoV NMTase is presented in Table 1.

Nucleocapsid protein (also named as N protein)

The nucleocapsid protein (N) is a multifunctional RNA-binding protein having pivotal roles in the viral replication cycle as viral core formation, viral assembly, virus budding/envelope formation, and mRNA replication/gRNA synthesis. This protein binds and packs the viral RNA genome into a helical nucleocapsid structure called ribonucleoprotein (RNP) complex. Also, the coronavirus N protein has several roles in the cellular response including chaperone activity, cell cycle regulation, stress response, viral pathogenesis/immune system interference, and signal transduction.

N protein structure has an N-terminal RNA-binding domain (NTD) responsible for RNA binding, a central Ser/Arg(SR)-rich linker intrinsically disordered, and a C-terminal dimerization domain (CTD).190, 191 The N-terminal domain of the SARS-CoV N protein has five-strand β-sheet and single-stranded RNAs bind to the protein surface at the junction between positively charged β-hairpin and the core structure. The crystallographic structure of the SARS-CoV N protein CTD showed a dimer with extensive interactions between the two subunits thus suggesting that the N protein is not stable in the monomeric form.

Recently, the crystallographic structure of the N-terminal RNA binding domain of SARS-CoV-2 N protein was determined and a unique potential hydrophobic RNA binding pocket alongside the β-sheet core was detected. Also, the biochemical characterization of expressed SARS-CoV-2 N protein detected the protein as a large dimer in solution by CTD-CTD interaction.

N Protein inhibitors

Among the CoV structural proteins, the nucleocapsid (N) protein seems to be a promising target for designing therapeutic drugs with antiviral activity, since it plays essential function on the virus particle assembly. The mouse CoV N protein is a major pathological marker in the host cell, in which could lead to the upregulation of proinflammatory cytokines, block innate immune response, and also cause induce host cell apoptosis. Some studies have reported the utilization of polyphenolic compounds with therapeutic properties, acting as antioxidants and anticancer, as well as antiviral and antimicrobial agents.197, 198, 199, 200 Both flavonoids (−)-catechin gallate and (−)-gallocatechin gallate demonstrate medicinal interest, due to their antioxidant effects with possible application for different disorders, including cancer. Also, (−)-catechin gallate has demonstrated promising effects in the treatment of HIV,203, 204 as well as against prostatic, brain, and other types of cancers.

Deeming these natural compounds, Roh et al. (2012) developed a new approach for the inhibitor screening forwards SARS-CoV N protein by using a quantum dots-conjugated oligonucleotide system with broad application in imaging analysis on a biochip. From a high-throughput screening (HTS) strategy using an optical nanoparticle-based RNA oligonucleotide, it was identified the inhibitory effects of (−)-catechin gallate and (−)-gallocatechin gallate (Fig. 13 ) upon SARS-CoV N protein. The biological evaluation revealed that both flavonoids exhibit high inhibition activity in a concentrated manner against SARS-CoV N protein. These compounds, at 0.005 µg/mL or more, concentration-dependently attenuated the binding affinity as evidenced by quantum dots-RNA oligonucleotide on the biochip. This concentration was found to be the IC50 value for both natural products. Regarding this, the authors have concluded that a new property for (−)-catechin gallate and (−)-gallocatechin gallate has been identified, and also their approach was found to be promising for drug discovery. Besides, no information about molecular interactions from in silico studies were provided by the authors. Finally, the authors’ technique could be mainly characterized by low-cost, high-sensibility, fast result, low labor-effort, compatibility for miniaturization, allowing which it could be applicable for discovering inhibitors in screenings against other types of diseases.

Fig. 13

Flavonoids with activity against SARS-CoV N protein.

Based on Table 1, it is observed that Han et al. (2006) have developed a peptide sequence (containing 31 amino acids) capable of inhibiting the S protein from SARS-CoV, with an IC50 value of 0.1 µM. In general, it is possible to verify that 3CLpro inhibitors are more broadly studied in the medicinal chemistry of CoV, followed by PLpro inhibitors. In this context, we performed a complete analysis of the interactions associated with 3CLpro inhibitors found in the literature (IC50 values ranging from 97 to 0.051 µM). Then, it was revealed that hydrophobic contacts are present in 49.09% of all interactions. Additionally, hydrogen-bonding interactions are observed for 49.54% of compounds, being the Cys145 (11.1%), His163 (12.96%), and Glu166 (14.81%) residues most present in interactions between ligands and SARS-CoV 3CLpro. In some studies, the Glu166 residue has been replaced with His166 residue,216, 219 because the authors have used MERS-CoV 3CLpro (PDB ID: 4RSP) and SARS-like human β-Coronaviruses 2c EMC/2012 (HCoV-EMC), respectively. In general, MERS-CoV 3CLpro is also a chymotrypsin-like cysteine protease that has a catalytic dyad constituted of Cys148 and His41, along with an extended binding site.219, 267, 268, 269, 270 This target exhibits a rigorous primary substrate specificity for a P 1 Gln residue. Also, a strong affinity for a P 2 Leu residue has been identified. Finally, P 3 residue side chain is oriented towards the hydrophobic P 4 residue, such as Ala.216, 271 In contrast, some studies have suggested that SARS-CoV 3CLpro has a catalytic dyad Cys145-His41, similarly to the SARS-CoV-2 (Cys145 and His41), HCoV 3CLpro (Cys144 and His41), and TGEV 3CLpro (Cys145 and His41).125, 272 Some works have suggested that cysteine residue from catalytic dyad could act as a nucleophile, being Cys145 from SARS-CoV221 and SARS-CoV-2 3CLpro, and Cys148 from MERS-CoV 3CLpro.214, 219 Then, considering all interactions reported for 3CLpro inhibitors into Table 1, it was built a graphic containing the frequency of residues (Fig. 14 A). This type of graphic has been used by our research team to identify fingerprints of residues from a cysteine protease. Regarding the Fig. 14A, it is possible to observe that eight amino acids are more frequently associated with higher than 30% of interactions between ligands and 3CLpro target, being His41 (51.4%), Met49 (31.4%), Phe140 (40%), Cys145 (60%), His163 (51.4%), Met165 (34.2%), Glu166 (48.5%), and Gln189 (40%) residues from protomer A. Moreover, the cysteine (Cys145-148) residue performs covalent bonding modes in only 1.37% of interactions. All these residues were used to build a tridimensional active site structure (Fig. 14B) in complex with the inhibitor N3 (Fig. 14C), containing only the most important residues, as previously discussed. It was verified that all these amino acids are comprised into binding pocket 1 from SARS-CoV-2 3CLpro, as described by Shi et al. (2020). Notwithstanding these observations, it is evident that most of the active compounds found in the literature preferably bind into the pocket 1. Recently, Harcourt et al. (2004) verified that the PLpro from SARS-CoV-2 is a crucial viral enzyme and a potential target for designing inhibitors. Additionally, SARS-CoV PLpro has a catalytic triad constituted of Cys112, His273, and Asp287. Also, the side chain of Trp107 that is placed into the oxyanion hole participates in the stabilization of the negatively charged tetrahedral transition state from the intermediate structure.132, 258, 276 In the SARS-CoV-2 PLpro structure, the catalytic triad is composed of Cys111, His272, and Asp286 residues. Besides, the auxiliary amino acid for stabilization of the negatively charged intermediate is Trp106 residue. However, this stabilizing function could be shared with other amino acids, such as Tyr268, Ala289, and Leu298 residues. Lastly, these different residues labeling is associated with the X-ray crystallographic structure used by the authors (PDB ID: 6W9C). Regarding the PLpro inhibitors (IC50 values ranging from 24.4 to 0.35 µM) in Table 1, these compounds commonly interact with Asp165, Tyr269 (from SARS-CoV), Gln270, Cys111 (SARS-CoV-2), and Cys112 (from SARS-CoV). Helicases are proteins responsible for catalyzing duplex oligonucleotides separation (double-stranded RNA-dsRNA) into single strands (ssRNA), using energy from ATP hydrolysis. Previously, a triazole-derived inhibitor, named SSYA10-001, was found to be capable of inhibiting nsP13 Y277A, nsP13 K508A, and WT SARS-CoV nsP13, exhibiting IC50 values of 12, 50, and 5.9 µM, respectively. Also, it was revealed that SSYA10-001 interacts with Tyr277, Arg507, and Lys208 residues from SARS- and MERS-CoV nsP13 targets.260, 261 Still, concerning NTPase/helicase (nsP13) inhibitors (IC50 values varying from 11 to 2.5 µM), a triazole derivative recently synthesized by Zaher et al. (2020) exhibited activity against MERS-CoV nsP13 (IC50 = 2.5 µM), where it was investigated by using in silico studies. As a result, it was observed that there are only hydrophobic interactions involving Tyr7, Arg163, Tyr159, and Tyr171 amino acid residues from MERS-CoV nsP13 (PDB ID: 5WWP). N and 2′-O-methylation of the viral RNA cap are critical steps for viral infections since their inhibition leads to a decrease in the synthesis of viral proteins and helps the virus elimination by stimulation of the immune response. Then, MTases are considered as attractive targets for drug design and discovery,188, 189 where the Phe426 residue is considered as the largest influence in nsP14 activity. S-adenosyl methionine (SAM) and RNA binding sites of NMTase nsP14 present 95% sequence identity, when comparing SARS-CoV with SARS-CoV-2 organisms. In this context, nucleoside SAM analogs could target nsP14 from both viruses. A potent SARS-CoV nsP14 inhibitor was synthesized by Ahmed-Belkacem et al. (2020), in which it demonstrated an IC50 value of 0.6 µM. Additionally, it was verified that this compound is able to perform seven hydrogen-bonding interactions, with Arg310, Gly333, Lys336, Ile338, Asp352, Asn386, and His424.

Fig. 14

Frequency of residues present in ligand-3CLpro complexes (A) found in the literature and compound N3 in complex with eight most frequent amino acid residues of the pocket 1 (B) from SARS-CoV 3CLpro (PDB ID: 2HOB). In B), covalent Cys145 residue is shown as blue color, exhibiting its connectivity to Michael’s acceptor moiety. Finally, the most frequent amino acid residues from pocket 1 are demonstrated in a molecular surface overview (in orange color). In C), the Michael’s acceptor N3, a peptide-derived, which has been explored in different works, where it has been identified as a promising inhibitor of 3CLpro from MERS-CoV (IC50 = 0.28 µM), SARS-CoV-2 (EC50 = 16.77 µM), GS-WT12 (Ki = 9.0 ± 0.8 µM), WT-GPH6 (Ki = 2.3 ± 0.1 µM), and WT (Ki = 1.9 ± 0.1 µM) modified proteases..

Conclusion

Concerning all aspects aforementioned in this manuscript, it is evident that coronaviruses are viruses which account for millions of human illness worldwide, being the SARS-, MERS-, and SARS-CoV-2 the main viruses of medical interest. Currently, there are no vaccines or FDA-approved drugs to specifically treat these viral infectious diseases. Then, the development of researches involving identification and/or synthesis of inhibitors emerges as an essential approach for drug discovery, since selective antivirals targeting macromolecules from HCoVs are an unmet need. In this context, the medicinal chemistry provides tools for the development of more selective inhibitors by using computer-aided drug design (CADD) approaches, such as structure-based drug design (SBDD), in which molecular characteristics from the macromolecular target (including interactions with ligands) should be considered during the rational design of more potent and selective inhibitors. In this scenario, classical medicinal chemistry along with SBDD have resulted in promising new antiviral agents targeting HCoVs’ macromolecules, providing us important aspects about the current druggable targets from these viruses which could be used to design desirable inhibitors.

In general, it was verified that natural products have demonstrated great results in different studies, such as hexamethylene amiloride targeting E protein from HCoV-229E (IC50 = 1.34 µM, and hydrogen-bonding interactions with Asp15 and Arg38 residues), flavone scutellarein targeting nsP13 from SARS-CoV (IC50 = 0.86 µM), sinefungin targeting nsP14 from SARS-CoV (IC50 = 1.68 µM), and the flavonoids (or phenolic compounds) (−)-catechin gallate and (−)-gallocatechin gallate targeting N protein from SARS-CoV (both with IC50 = 0.005 µg/mL). Additionally, small synthetic peptides have exhibited activity against the S protein from SARS-CoV, where the best peptide sequence (STSQKSIVAYTM) displayed an IC50 value of 1.8 nM, suggesting this compound as an excellent inhibitor. However, it was not provided information about the molecular interactions by using in silico methods. Concerning 3CLpro inhibitors, it was verified that the covalent interaction between the cysteine residue from the catalytic dyad and ligands are frequently observed. The best 3CLpro, a synthetic indole analog, was able to covalently interact with Cys145 and, via hydrogen-bonding, with Gly143, Ser144, and His163 residues, at distances ranging from 2.3 to 2.8 Å. In the SARS-CoV 3CLpro inhibition assay, it demonstrated an IC50 value of 0.03 µM. Moreover, the best PLpro inhibitor, a synthetic piperidine carboxamide analog, was not able to interact with Cys residues from the catalytic triad from this target. However, it presented an IC50 value of 320 nM against SARS-CoV PLpro. Also, it was verified that this piperidine analog shows a bumping collision with the Asp165 amino acid residue when an S-methyl substituent is present. Finally, a drug repurposing study was capable of suggesting chlorhexidine (an antimicrobial drug) and remdesivir (an antiviral drug) have been highlighted as promising agents targeting nsP12 from SARS-CoV-2. However, none inhibition assay was performed upon isolated nsP12. Still, anti-HIV and HCV drugs have been also suggested as promising candidates against SARS-CoV-2 nsP12, such as nelfinavir, raltegravir, delavirdine, beclabuvir, ledipasvir, and paritaprevir.

3CLpro inhibitors have been broadly studied, allowing us to generate a %frequency graphic showing the most relevant residues observed during the ligand-target complex formation. From this, it was possible verified that these compounds preferably interact via hydrogen-bonding interactions (49.54%), where the Glu166 residue is the most frequent. Additionally, it was observed that Cys145 (from SARS-CoV) and Cys148 (MERS-CoV) residues together are verified in only 1.37% of interactions. In general, it was verified that the 3CLpro inhibitors are more frequently found into the pocket 1, being the residues His41 (51.4%), Met49 (31.4%), Phe140 (40%), Cys145 (60%), His163 (51.4%), Met165 (34.2%), Glu166 (48.5%), and Gln189 (40%) more commonly found in interactions with this macromolecular target.

Finally, we believe that the search for inhibitors targeting HCoVs macromolecular structures is in continuous development, where several studies have been released, mainly targeting 3CLpro and PLpro. However, the other targets should be considered in the drug design of new antiviral compounds. This manuscript represents a great collection of relevant studies targeting HCoV in order to develop new drugs, which could be used for guiding future studies focused on the development of inhibitors to fight against coronaviruses, including SARS-CoV-2 (COVID-19). Furthermore, it is possible suggest that the classical medicinal chemistry strategies are insufficient to quickly discover new drugs against this virus. Thus, new methodologies and technologies should be exploited to identify novel anti-SARS-CoV-2 drug candidates in a time- and cost-effective manner. Then, it is time to review the various drug discovery methods involving computer-aided protocols to seek trends and identify promising new avenues of anti-SARS-CoV-2 drug discovery.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.