Design, synthesis and biological evaluation of 2-aminoquinazolin-4(3H)-one derivatives as potential SARS-CoV-2 and MERS-CoV treatments.

Despite the rising threat of fatal coronaviruses, there are no general proven effective antivirals to treat them. 2-Aminoquinazolin-4(3H)-one derivatives were newly designed, synthesized, and investigated to show the inhibitory effects on SARS-CoV-2 and MERS-CoV. Among the synthesized derivatives, 7-chloro-2-((3,5-dichlorophenyl)amino)quinazolin-4(3H)-one (9g) and 2-((3,5-dichlorophenyl)amino)-5-hydroxyquinazolin-4 (3H)-one (11e) showed the most potent anti-SARS-CoV-2 activities (IC50 < 0.25 μM) and anti-MERS-CoV activities (IC50 < 1.1 μM) with no cytotoxicity (CC50 > 25 μM). In addition, both compounds showed acceptable results in metabolic stabilities, hERG binding affinities, CYP inhibitions, and preliminary PK studies.

Coronavirus is a single positive-stranded RNA virus that was discovered in 1960 while looking for a new cold virus that infects the upper respiratory tract.1, 2 The clinical significance was relatively low because of its weak symptoms. However, with the outbreak of SARS-CoV in 2003, its clinical significance has received new attention.3, 4 SARS-CoV spread to 4 countries, with 8,422 confirmed cases and 916 deaths, with its mortality rate of 11%, raising the public health awarness.4, 5 In 2012, a new outbreak of MERS-CoV occurred and spread to 27 countries by January 2020, resulting in 2519 confirmed cases and 866 deaths, with a mortality rate of 34%. Most recently, COVID-19, which was caused by SARS-CoV-2 outbreak in Wuhan, China, in December 2019, has spread worldwide, causing 40 million confirmed cases and 1 million deaths. Eventually, on March 11, 2020, COVID-19 was declared as third pandemic after the 1968 Hong Kong flu and 2009 influenza by the WHO.8, 9

For treatments of MERS-CoV, it is generally recommended to use drugs such as interferon, immunomodulatory factor, and antiviral drugs such as ribavirin or lopinavir. However, there are reports that interferon and ribavirin may cause side effects such as poor bone marrow function, anemia, and virus mutations. In addition, although the monkey model showed a therapeutic effect, it did not show a great effect in actual clinical trials, requiring the development of a safer and more efficient MERS treatment.

In the case of COVID-19, Remdesivir, which received urgent approval, as well as Nafamostat, and Hydroxychloroquine, are considered as promising therapeutic candidates. However, due to side effects and low clinical effects, safer and more effective treatments need to be developed.

Our research for novel antivirals inhibiting these fatal coronaviruses started with a biochemical high content screening (HCS) of a library containing 200,000 compounds from Korea Chemical Bank. In the past few years, we reported inhibitors of MERS-CoV such as 2-phenylchroman-4-one derivatives, 3-acyl-2-phenylamino-1,4-dihydroquinolin-4-(1H)-one derivatives, and 4-aniline-6-amino-quinazoline derivatives as well as inhibitors of SARS-CoV-2 such as cyclic sulfonamide derivatives.

In this study, the core scaffold of 3-acyl-2-phenylamino-1,4-dihydroquinolin-4-(1H)-one of our lead compound (1) was changed to 2-amino-quinazolin-4(3H)-one (Fig. 1 ) . Because the aromatic rings substituted with halogen groups or electron withdrawing groups of dihydroquinolinones 1 were important for activity, 2-amino-quinazolin-4(3H)-ones with aromatic rings fixed at the similar positions were designed and evaluated to find compound 2 as a hit (IC50 = 2.6 μM for SARS-CoV-2). 2-Amino-quinazolin-4(3H)-one derivatives have been known to possess high activities as inhibitors of aldose reductase, KATP channel opener, anti-cancer agents, and anti-hyperglycemic agents,21d Here we report on the synthesis and biological effects of 2-amino-quinazolin-4(3H)-one derivatives.

Fig. 1

Our new design of anti-coronavirus Scaffold.

A series of 2-amino-quinazolin-4(3H)-one analogues were synthesized as shown in Scheme 1 . Anthranilic acids 3 were treated with urea by heating 150 °C for 20 h to afford quinazolinediones 4. Dichloroquinazolines 5 were prepared by reacting quinazolinediones 4 with POCl3 in the presence of trimethylamine. Treatment of dichloroquinazolines 5 with 2 N NaOH led to 2-chloro-4(3H)-quinazolinones 6. 2-Chloroquinazolin-4(3H)-ones 6 and anilines 7 with various substituents were heated in DMF at 85 °C for 16 h to obtain 2-amino-quinazolin-4(3H)-ones 8.

Scheme 1

Synthesis of 2-aminoquinazolin-4(3H)-ones derivatives. Reagents and conditions: (a) Urea, 150 °C, 20 h; (b) POCl3, TEA, 115 °C, 17 h; (c) 2 N NaOH, rt, 20 h; (d) DMF, 85 °C, 16 h.

The anti-SARS-CoV-2 and anti-MERS-CoV activities of the synthesized compounds were evaluated by immunofluorescent assay in Vero cells.16, 20 In this study, Vero cells were stained using antibodies targeting spike protein for MERS-CoV and nucleocapsid protein for SARS-CoV-2, and the infection rate was measured by imaging the infected Vero cells through microscope.

First, we began structure–activity relationship (SAR) studies for anti-SARS-CoV-2 activities by varying anilines at 2 position of quinazolinone ring of compound 2 (Table 1 ). For anilines with one fluoride atom, compound 9b (IC50 = 1.4 μM) bearing aniline substituted with fluoride at meta-position showed good effect, whereas compounds with ortho-fluoroaniline (9a) or para-fluoroaniline (9c) showed no inhibitory effect. In the same way, compound with anilines having two fluoride atoms at ortho and para position (9d) had no activity, while compound 9e with 3,5-difluoroaniline (IC50 = 0.24 μM), which has two fluorides at double meta-position, showed a highly increased activity because of the synergic effect of fluoride at meta-position. Changing 2,4-fluoroaniline and 3,5-difluoroaniline to 2,4-chloroaniline (9f) and 3,5-dichloroaniline (9g) showed similar activities (IC50 : >25 and 0.23 μM, respectively). Next, we investigated the substituent effects at meta-position of anilines (9h-9m). Electron-donating groups such as methoxy (9h), hydroxy (9i), and dihydroxy (9j) decreased anti-SARS-CoV-2 activities (IC50 > 11 μM). Electron-withdrawing groups such as cyano (9k) and trifluoromethyl (9l) showed much better effects than those of electron-donating groups (IC50 = 1.7 and 0.68 μM, respectively), while methylester (9m) decreased anti-SARS-CoV-2 activities (IC50 = 8.8 μM). Then, anti-SARS-CoV-2 effects of aliphatic amine substituents such as cyclohexylmethyl amine (9n), isopropyl amine (9o), n-butyl amine (9p), and piperidinyl amine (9q) were evaluated to exhibit decreased potency (IC50 > 14 μM).

Table 1

Lead optimization of 2-anilino groups.

Compound	R	SARS-CoV-2
		IC50a(μM)	CC50b(μM)	SIc
2	2,3,4-F3-Ph-NH-	2.6	>25	9.4
9a	2-F-Ph-NH-	>25	>25	1
9b	3-F-Ph-NH-	1.4	>25	18
9c	4-F-Ph-NH-	>25	>25	1
9d	2,4-F2-Ph-NH-	>25	>25	1
9e	3,5-F2-Ph-NH-	0.24	18	74
9f	2,4-Cl2-Ph-NH-	>25	>25	1
9g	3,5-Cl2-Ph-NH-	0.23	>25	110
9h	3-CH3O-Ph-NH-	11	>25	2.3
9i	3-HO-Ph-NH-	>25	>25	1
9j	3,5-(HO)2-Ph-NH-	>25	>25	1
9k	3-CN-Ph-NH-	1.7	>25	15
9l	3-CF3-Ph-NH-	0.68	>25	37
9m	3-CH3O(O = )C-Ph-NH-	8.8	>25	2.8
9n	Cyclohexylmethyl-NH-	14	>25	1.7
9o	i-Pr-NH-	>25	>25	1
9p	n-Bu-NH-	>25	>25	1
9q	piperidinyl-N-	>25	>25	1

IC50 and CC50 were derived from the results of at least two dependent experiment in Vero cells infected with SARS-CoV-2

SI(selective index) = CC50/IC50 for inhibiting SARS-CoV-2 infection.

In the next phase of optimization, we evaluated substituent effects of 5 to 8 positions of quinazolinone ring, having fixed with 3,5-difluoroaniline or 3,5-dichloroaniline at 2 position (Table 2 ). Compounds with electron-withdrawing groups, such as 7-trifluoromethyl (10a and 11a) and 7-nitro (10b and 11b), showed high binding affinities (0.20–0.51 μM). In the case of compounds with electron-donating groups, variations of inhibitory effects were shown. Compounds with 5-hydroxy (10e and 11e), 8-hydroxy (10f and 11f), 5,8-dichloro (10i and 11i), and 7,8-dichloro (10j and 11j) showed high binding affinities (0.15–1.6 μM). 7-Amino (10c and 11c), 5-methoxy (10d and 11d), 7-hydroxy (10g and 11g), and 6,8-dimethyl (10h and 11h) derivatives showed no inhibitory effects. The 7-N-substituted quinazolinone compound 11k displayed no anti-SARS-CoV-2 activity.

Table 2

SAR studies of 2-anilinoquinazolin-4(3H)-one derivatives

Compound	R	SARS-CoV-2
		IC50a(μM)	CC50b(μM)	SIc
10a	7-CF3	0.20	7.6	38
10b	7-NO2	0.41	>25	61
10c	7-NH2	>25	>25	1
10d	5-CH3O	>25	>25	1
10e	5-OH	0.47	>25	54
10f	8-OH	1.6	>25	15
10g	7-OH	>25	>25	1
10h	6,8-(CH3)2	>25	>25	1
10i	5,8-Cl2	0.33	>25	76
10j	7,8-Cl2	0.25	>25	98
11a	7-CF3	0.21	7.1	34
11b	7-NO2	0.51	>25	49
11c	7-NH2	>25	>25	1
11d	5-CH3O	>25	>25	1
11e	5-OH	0.15	>25	168
11f	8-OH	0.37	>25	65
11g	7-OH	>25	>25	1
11h	6,8-(CH3)2	>25	>25	1
11i	5,8-Cl2	0.35	>25	71
11j	7,8-Cl2	0.24	>25	103
11k	7-Morpholinyl	>25	>25	1

IC50 and CC50 were derived from the results of at least two dependent experiment in Vero cells infected with SARS-CoV-2

SI(selective index) = CC50/IC50 for inhibiting SARS-CoV-2 infection.

With the compounds having potent activities toward SARS-CoV-2, we tested anti-MERS-CoV activities (Table 3 ). All compounds except 10b and 11b showed good antiviral activities (IC50 = 0.39–3.1 μM). It seems that our quinazolinone compounds are potent broad spectrum coronavirus inhibitors. In particular, all the above compounds except 10a and 11a displayed no obvious cytotoxicity (CC50 > 25 μM).

Table 3

Anti-MERS-CoV activity of 2-Aminoquinazolin-4(3H)-ones Derivatives

Compound	MERS-CoV			SARS-CoV-2
	IC50a(μM)	CC50b(μM)	SIc	IC50d(μM)
9e	0.39	>25	65	0.24
9g	0.93	>25	27	0.23
10a	0.73	>25	35	0.20
11a	0.96	7.3	7.6	0.21
10b	8.5	>25	2.9	0.41
11b	>25	>25	1.0	0.51
10e	1.1	>25	22	0.47
11e	1.02	>25	25	0.15
10f	3.1	>25	8.0	1.64
11f	1.7	>25	14	0.37
10i	0.79	>25	31	0.33
11i	0.62	>25	42	0.35
10j	0.83	>25	30	0.25
11j	0.88	>25	31	0.24

IC50 and CC50 were derived from the results of at least two dependent experiment in Vero cells infected with MERS-CoV

SI(selective index) = CC50/IC50 for inhibiting MERS-CoV infection.

IC50 were derived from the results of at least two dependent experiment in Vero cells infected with SARS-CoV-2

We selected the compounds for the purpose of further pharmacological investigations on SARS-CoV-2. In the case of electron-withdrawing substituents in the aromatic ring of quinazolinone, compound 9g with chloro substituent at position 7 was selected because it has the highest SI (1 1 0). And we chose compound 11e with electron-donating substituent (hydroxyl) at 5 position because it is the most active on SARS-CoV-2 and has the highest SI (1 6 8). Compound 9g and 11e were further evaluated for their microsomal stabilities, cytotoxicities, human ether a-go-go (hERG) bindings, plasma protein bindings (PPB) and CYP inhibitions (Table 4, Table 5 ). First, the results of microsomal stability show that 9g is stable in mouse, rat, and human and 11e is microsomally more unstable than 9g, which seems to be due to the presence of –OH in the structure. In the cytotoxicity assay, 9g and 11e showed a little toxicity in HFL-1 (CC50 = 7.5 and 9.6 μM), but did not show toxicity in the rest of the cells (Table 4).

Table 4

Result of Microsomal stability and cytotoxicity of 9g and 11e

Compound	MSa			Cytotoxicity (μM)b
	ma	ra	ha	HFL-1	L929	NIH 3 T3	CHO-K1
9g	112	105	99	7.5	31	>100	11
11e	42	54	55	9.6	27	60	28

% of remaining after 30 min

Cell information. HFL-1: human embryonic lung cell line, L929: NCTC clone 929, mouse fibroblast cell line, NIH 3 T3 : mouse embryonic fibroblast cell line, CHO-K1 : Chinese hamster ovary cell line.

Table 5

Result of hERG, PPB, CYP inhibition of 9g and 11e

Compound	hERG(10 μM)	PPB (5 μM)		CYP inhibition (10 μM)
		r	h	1A2	2C9	2C19	2D6	3A4
9g	38	100	100	15	18	<1	20	32
11e	28	97	100	65	34	12	46	65

In hERG channel inhibition assay, it was found that both compounds 9g and 11e did not interact with hERG channel. In the PPB assay, both compounds 9g and 11e showed high binding rates (97–100%). The results of CYP inhibition assay show that 9g showed little inhibition of the CYP enzyme and 11e displayed some inhibition (>65% CYP inhibition at 10 μM) in CYP1A2 and CYP3A4.

The preliminary pharmacokinetic properties of 9g and 11e were investigated by intravenous (i.v.) and oral (p.o) routes in rats, with 2 mg/kg and 10 mg/kg, respectively (Table 6 ). The oral bioavailability of 9g and 11e were 15.6% and 7.8%, respectively, partly because the clearance of compound 11e is higher than that of 9g. The PK profile of 9g seemed to be acceptable for the discovery of anti-coronavirus drugs.

Table 6

Rat pharmacokinetic study of 9g and 11e

Compounds	9g		11e
Parametersa	I.V., 2 mg/kg	P.O., 10 mg/kg	I.V., 5 mg/kg	P.O., 10 mg/kg
Tmax (h)	NAb	2.5	NA	1.0
Cmax (μg/h)	NA	0.9	NA	0.06
T1/2 (h)	2.2	5.7	9.2	42
AUC (μg·h/mL)	8.5	6.6	6.5	1.01
CL (L/h/kg)	0.25	NA	0.67	NA
Vss(L/Kg)	1.6	NA	6.8	NA
Ft(%)	NA	15.6	NA	7.8

All results are the mean of experiments using three rats. b NA: not applicable

In conclusion, we designed and developed 2-aminoquinazolin-4(3H)-one derivatives as potent inhibitors against both SARS-CoV-2 and MERS-CoV. Among them, 7-chloro-2-((3,5-dichlorophenyl)amino)quinazolin-4(3H)-one (9g) and 2-((3,5-dichlorophenyl)amino)-5-hydroxyquinazolin-4(3H)-one (11e) were considered as new drug candidates because both have high anti-SARS-CoV-2 activities [(9g, IC50 = 0.23 μM), (11e, IC50 = 0.15 μM)] and anti-MERS-CoV activities [(9g, IC50 = 0.93 μM), (11e, IC50 = 1.02 μM)] with no cytotoxicity (CC50 > 25 μM). Our two lead compounds also showed good microsomal stabilities, relatively low cytotoxicities, low hERG binding affinities and CYP inhibitions. The PK profile of 9g seemed to be acceptable for the discovery of antivirals. 2-Aminoquinazolinone derivatives were found to be a promising new scaffold against coronaviruses and further optimizations to increase pharmacokinetic profiles are currently underway.

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.