4-Trifluoromethyl bithiazoles as broad-spectrum antimicrobial agents for virus-related bacterial infections or co-infections

Abstract

Respiratory tract infections involving a variety of microorganisms such as viruses, bacteria, and fungi are a prominent cause of morbidity and mortality globally, exacerbating various pre-existing respiratory and non-respiratory conditions. Moreover, the ability of bacteria and viruses to coexist might impact the development and severity of lung infections, promoting bacterial colonization and subsequent disease exacerbation. Secondary bacterial infections following viral infections represent a complex challenge to be overcome from a therapeutic point of view. We report herein our efforts in the development of new bithiazole derivatives showing broad-spectrum antimicrobial activity against both viruses and bacteria. A series of 4-trifluoromethyl bithiazole analogues was synthesized and screened against selected viruses (hRVA16, EVD68, and ZIKV) and a panel of Gram-positive and Gram-negative bacteria. Among them, two promising broad-spectrum antimicrobial compounds (8a and 8j) have been identified: both compounds showed low micromolar activity against all tested viruses, 8a showed synergistic activity against E. coli and A. baumannii in the presence of a subinhibitory concentration of colistin, while 8j showed a broader spectrum of activity against Gram-positive and Gram-negative bacteria. Activity against antibiotic-resistant clinical isolates is also reported. Given the ever-increasing need to adequately address viral and bacterial infections or co-infections, this study paves the way for the development of new agents with broad antimicrobial properties and synergistic activity with common antivirals and antibacterials.

Is magic trifluoromethyl a thing? Replacing 4-CH₃ with 4-CF₃ in bithiazoles, allowed to identify broad antimicrobial agents active against multiple viruses and also against Gram-positive/negative bacteria.

Introduction

Each year, an estimated 3 to 5 million fatalities are ascribed to respiratory tract infections (RTIs), significantly impacting public health and representing a significant socioeconomic challenge. RTIs affect both children and adults and are a prominent cause of morbidity and mortality globally, also due to the onset of antibiotic resistance.^1,2 A diverse range of microorganisms, such as viruses, bacteria, and fungi, are responsible for causing RTIs. Rhinoviruses (RVs) are responsible for more than one-half of upper respiratory tract infections (URTI). Even though most RV infections are benign, self-limiting cold-like illnesses, they have also been found as the causal agent of severe pneumonia in frail patients. RVs, along with enteroviruses (EVs), are positive-sense single-stranded RNA viruses and belong to the EV genus within the Picornaviridae family. The EV genus includes 12 species, classified into three species of human RV (RV-A to RV-C), four human EV (EV-A to EV-D), and five EV species that infect only animals.³ Except in rare cases, the tropism of RVs is limited to upper respiratory airways, whereas EVs can infect a wide variety of cells, resulting in very different clinical manifestations, including respiratory infections with RV-like symptoms.^4,5 A special case is represented by EV-D68, which is normally associated to respiratory infections. However, in 2014 it caused an outbreak in the USA with high morbidity in children and neurological symptoms.⁶ Furthermore, although typical RV and EV infections are commonly linked to mild illnesses, they have the potential to exacerbate or worsen various pre-existing respiratory and non-respiratory conditions. RV infections have been associated with more severe exacerbation of chronic respiratory diseases, such as asthma, chronic obstructive pulmonary disease (COPD), and cystic fibrosis (CF) as well as with higher levels of inflammatory biomarkers.⁷ In CF, an autosomal recessive disorder affecting almost 89 000 individuals worldwide, the progression of the disease is strongly connected to respiratory infections and consequent airway inflammation, being the principal causes of pulmonary exacerbation and death.^8,9 Unfortunately, no drugs for the treatment of Enterovirus genus infections have been approved to date, and considering all the aspects mentioned above, there is an urgent need for new drug candidates.

In addition, it is becoming widely understood that the ability of bacteria and viruses to coexist might impact the development and severity of infections and generally of lung diseases.¹⁰ According to several studies, viral respiratory infections change the lung environment, promoting bacterial colonization and subsequent disease exacerbation, especially promoted by Pseudomonas aeruginosa and Staphylococcus aureus.^11,12 The mechanisms underlying bacterial superinfections or virus-induced secondary bacterial infections of the respiratory tract are (i) the impairment of mucociliary clearance due to excessive mucus production, the reduction of ciliary beat frequency, uncoordinated ciliary movement and a reduction in the number of ciliated cells;^13,14 (ii) enhanced receptor availability for bacterial binding, by the increased expression of host receptors (among others, the platelet-activating factor receptor or PAFR)^15–17 or by viral structures serving as coupling agents for bacteria to host tissue;¹⁸ (iii) immunological aberrances, such as reduced expression and responsiveness of PRRs^19,20 and impaired immune cell functions;²¹ (iv) changes of the microenvironment, as the increase of nutrient availability for bacterial growth.^22,23 Moreover, viral–bacterial coinfections could enhance the severity of RV symptoms, hospitalization, and mortality.²⁴ A proper immune response is necessary to eliminate invading pathogens and treat disease, but a too vigorous response may cause immunopathology and exacerbate symptoms.²⁵ From a therapeutic point of view, virus-induced bacterial superinfections pose a complex challenge for healthcare professionals: the administration of antibiotic therapy becomes a common practice with patients hospitalized with viral infections to prevent potentially life-threatening complications. However, this essential clinical intervention, aimed at combating bacterial over-infections, is not without repercussions. The overuse or misuse of antibiotics in these circumstances may contribute significantly to the development of antimicrobial resistance (AMR) or represent an increased risk factor for patients to be co-infected with multi- or ultra-drug-resistant bacteria whose prevalence could be significant in the hospital setting, thus jeopardizing the effectiveness of antibiotics and rendering previously treatable bacterial infections untreatable. In light of the alarming surge in antibacterial resistance and the increased threat of virus-related secondary bacterial infections,^26–29 the search for new strategies to directly or indirectly combat both viral and bacterial infections is of particular interest. It is therefore intriguing to explore whether antiviral compounds might also demonstrate an additional antibacterial effect or synergistic efficacy with existing antibiotics. This inquiry could prove valuable in addressing more effectively viral–bacterial co-infections and possibly preventing or mitigating the severity of secondary bacterial infections. The rational design of such multifunctional drug candidates is a very complex endeavour and there is a growing awareness indicating that the reductionist target-based drug discovery often leads to molecules unable to exert their effect on more complex biological systems (e.g. cell, animal model) despite promising data on target engagement.³⁰ A phenotypic drug discovery approach based on antimicrobial chemotypes seems therefore a reasonable approach for the early identification of new hits with the desired polypharmacological effect. Based on these premises, the bithiazole chemotype represents a versatile and privileged scaffold for a phenotypic campaign aimed at the identification of new broad-spectrum antimicrobial agents. Several bithiazole derivatives, both natural and synthetic, with different mechanisms of action, have indeed been reported in the literature for their antibacterial properties.^31–33

In the past few years, our research group has also thoroughly investigated the pharmacological versatility of the bithiazole chemotype, which can be finely tuned by specific functionalization of its privileged structure (Fig. 1).^34–37 Starting from Corr4a, a cystic fibrosis corrector active on the CFTR-F508del mutation, structural modification of the bithiazole scaffold led to the identification of MR250 as a potent inhibitor of the lipid kinase PI4KIIIβ (PI4KB). MR250 exhibited broad-spectrum antiviral activity against the Picornaviridae family while exerting no impact on bacterial growth. We then observed that a simple replacement of the C4-methyl group with a trifluoromethyl (MR335 and MR459) led to a significant reduction in activity against PI4KB and a concurrent acquisition of antibacterial properties. Intrigued by the effect of this point modification (4-CH₃vs. 4-CF₃) on the polypharmacological activity of these bithiazole derivatives, we have analysed the effect of the most promising antibacterial compound MR459 against a few viruses belonging to different families (entry 1, Table 1). It is noteworthy that despite MR459 having no effect on PI4KB, it is able to inhibit the replication of all tested viruses (hRVA16, EVD68, ZIKV) in the low micromolar range.

Fig. 1. Multidimensional pharmacological space of representative bithiazole derivatives and general structure of the target molecules. MR compounds have been reported as antiviral^34–36 or antibacterial agents.³⁷ Corr4a is a known type II corrector of the most common CFTR mutation in cystic fibrosis (CFTR-F508del). Antiviral activity against human rhinoviruses (HRV), minimal inhibitory concentration (MIC) vs. S. aureus and inhibitory activity of PI4KB are shown.

Activity of newly synthesized compounds versus PI4KB, virus replication (hRVA16, EVD68 and ZIKV) and Gram-positive/-negative organisms.

Entry	Cpd	PI4KIIIβ IC50a	Antiviral activity IC50b	Cytotoxicity CC50c	MIC/MBCd (μg mL−1)				MAC (μg mL−1) in the presence of 0.5 × MIC colistine
		[μM]	[μM]	[μM]	Bsu	Efa	Spy	Sau	Eco	Kpn	Aba	Pae
1f	MR-250	0.09	9.70 (hRV2)	36.3 (HeLa)	>128	>128	128	>128	>64	>64	>64	>64
			0.42 (EV71)	32.1 (RD)
			17.0 (CVB3)	110.7 (Vero)
2	MR-459	NA	1.82 (hRVA16)	46.0 (HeLa)	4/8	8/>32	4/4	8/32	4	4	4	>64
			8.37 (EVD68)
				>50 (VeroE6)
			11.47 (ZIKV)
3	8a	NA	0.49 (hRVA16)	>50 (HeLa)	>32	>32	>32	>32	2	64	8	>64
				35.9 (VeroE6)
			1.11 (EVD68)
			4.62 (ZIKV)
4	8b	NA	NA (hRVA16)	>50 (HeLa)	>32	>32	>32	>32	>64	>64	>64	>64
				>50 (VeroE6)
			12.29 (EVD68)
			NA (ZIKV)
5	8c	NA	NA (hRVA16)	>50 (HeLa)	>32	>32	>32	>32	>64	>64	>64	>64
			NA (EVD68)	>50 (VeroE6)
			NA (ZIKV)
6	8d	3.97	9.77 (hRVA16)	>50 (HeLa)	32/>32	>32	16/>32	>32	>64	>64	>64	>64
				>50 (VeroE6)
			10.04 (EVD68)
			NA (ZIKV)
7	8e	NA	NA (hRVA16)	>50 (HeLa)	16/>32	16/>32	4/32	16/32	>64	>64	>64	>64
				>50 (VeroE6)
			NA (EVD68)
			NA (ZIKV)
8	8f	NA	0.25 (hRVA16)	47.3 (HeLa)	>32	>32	>32	>32	>64	>64	>64	>64
				>50 (VeroE6)
			2.01 (EVD68)
			6.13 (ZIKV)
9	8g	4.06	9.52 (hRVA16)	>50 (HeLa)	16/16	32/>32	4/4	16/<32	>64	>64	>64	>64
			4.18 (EVD68)	>50 (VeroE6)
			NA (ZIKV)
10	8h	NA	2.44 (RVA16)	22.9 (HeLa)	>32	>32	>32	>32	>64	>64	>64	>64
			NA (EVD68)	>50 (VeroE6)
			NA (ZIKV)
11	8i	1.24	NA (hRVA16)	>50 (HeLa)	4/32	2/>32	2/>32	>32	>64	>64	>64	>64
				>50 (VeroE6)
			NA (EVD68)
			NA (ZIKV)
12	8j	NA	9.02 (hRVA16)	31.8 (HeLa)	8/16	8/16	8/8	32/>32	2	2	4	>64
			6.74 (EV68)
			6.90 (ZIKV)
				46.0 (VeroE6)
13	10	NA	10.37 (hRVA16)	>50 (HeLa)	>32	>32	>32	>32	32	>64	16	>64
				>50 (VeroE6)
			5.74 (EV68)
			8.52 (ZIKV)

^aValues are the mean of at least three independent experiments.

^bIC₅₀: half-maximal inhibitory concentration calculated in a cell-based assay using live virus; each value is the mean of three experiments.

^cCC₅₀: half-maximal cytotoxic concentration.

^dThe minimum inhibitory concentration (MIC) was determined on the following Gram-positive bacteria: Bacillus subtilis ATCC 6633 (Bsu), E. faecalis ATCC 29212 (Efa), S. pyogenes ATCC 12344 (Spy), and S. aureus ATCC 25923 (Sau); the minimum bactericidal concentration (MBC) value was determined only for compounds showing MIC values ≤64 μg mL⁻¹;

^eNone of the compounds exhibited direct-acting antibacterial activity on the tested Gram-negative strains (MIC, >64 μg mL⁻¹). Therefore, the minimum antibacterial concentration (MAC) was determined on the following Gram-negative bacteria: E. coli CCUG^T (Eco), K. pneumoniae ATCC 13833 (Kpn), A. baumannii ATCC 17978 (Aba), and P. aeruginosa ATCC 27853 (Pae); experiments were conducted in the presence of a fixed subinhibitory concentration (0.12 μg mL⁻¹ for E. coli and K. pneumoniae; 0.25 μg mL⁻¹ for A. baumannii and P. aeruginosa) of colistin.

^fData from ref. 35 and 37.

To further explore the role of the 4-CF₃ substitution on the polypharmacological effect of the bithiazole chemotype, we report herein the synthesis and biological evaluation of novel 4-trifluoromethyl bithiazole derivatives as potential broad-spectrum antimicrobial agents.

Results and discussion

Considering the importance of the 4-CF₃ group for the antibacterial activity of the bithiazole chemotype, we decided to synthesize a series of analogues of MR459 having the general structure depicted in Fig. 1 to define the molecular requirements needed to exert a broad antimicrobial effect against multiple viruses and bacteria. We planned a screening workflow encompassing PI4KB, selected viruses from the Picornaviridae (hRVA16, EVD68) and Flaviviridae (ZIKV) families, and a panel of representative Gram-positive and Gram-negative bacteria. Moreover, given the lack of activity of MR459 against PI4KB, we hypothesize that these 4-CF₃ bithiazoles could have a different mechanism of action in comparison with 4-CH₃ bithiazole antivirals targeting PI4KB. Hence, additional investigations were undertaken to explore the potential mechanism of action. In all planned analogues of MR459, we thus maintained the 4-CF₃ on the thiazole ring, introducing different functional groups in R₁/R₂, or different heteroatoms in X/Y in an attempt to improve both the antiviral and the antibacterial properties. The first compound in our set (8a) was obtained by the only replacement in position R₁ of the tert-butyl group with a bulkier phenyl group: this modification had previously been found by us to be well tolerated and effective for antiviral activity against EV.³⁴ Then, our attention was directed toward addressing the primary challenge encountered in our previous research – the poor diffusion through the outer membrane, as underscored by the need for a subinhibitory concentration of colistin to make MR459 active against Gram-negative bacteria. A second set of compounds (8b–8e), characterized by a siderophore moiety on the right part of the molecules, was thus synthesized to improve their internalization through the outer membrane, exploiting the so-called Trojan horse strategy. The idea arises from the bacterial attitude of surviving during the infection by secreting siderophores to increase the intracellular concentration of ferric ions.^38,39 Another set of molecules (8f–8i) was planned as a merging of the pharmacophore features of the antiviral MR250 (carboxy moiety on the right phenyl ring) and those of the antibacterial MR459 (CF₃ moiety on both the thiazole and the right phenyl ring). Finally, two point modifications on MR459 led to compounds 8j and 10 by introducing an additional nitrogen atom to promote the outer membrane permeation according to the eNTRy rules.^40–44

The synthesis of some key intermediates was necessary to obtain our final compounds: the α-bromoacetyl derivatives 5a and 5b, the thioureas 7a–7f, and the guanidine 9. For the synthesis of the α-bromoacetyl derivatives 5a and 5b, the commercial 1,1,1-trifluoro-2,4-pentadione 1 was reacted with hydroxy(tosyloxy)iodobenzene in dry acetonitrile at 80 °C for 2 h under an argon atmosphere to afford the intermediate 2 that was directly reacted with thiourea to obtain after 5 h the 1-(2-amino-4-(trifluoromethyl)thiazol-5-yl)ethan-1-one 3. Intermediate 3 was reacted with pivaloyl chloride or benzoyl chloride and then submitted to α-bromination to afford the key intermediates 5a and 5b (Scheme 1).

Scheme 1. Reagents and conditions: (i) hydroxy(tosyloxy)iodobenzene, dry CH₃CN, 80 °C, 2 h; (ii) thiourea, dry CH₃CN, 80 °C, 5 h; (iii) R₁COCl, pyridine, dry THF, 0 °C reflux, o.n.; (iv) HBr 48%, Br₂, 1,4-dioxane, 20 °C, 48 h.

Thioureas 7a–7f were synthesized by the reaction of the corresponding amine 6a–6e with benzoyl isothiocyanate, followed by deprotection with NaOMe to obtain the free thioureas. Considering the thiourea 7d, a further protective step of the carboxylic acid moiety into a methyl ester is necessary to obtain the target compounds in good yields (Scheme 2). Indeed, the direct reactions between the thiourea bearing the carboxylic acid moiety with the α-bromoacetyl derivatives 5a and 5b had previously led to a complex mixture of compounds from which it was unlikely to obtain the pure products. Finally, the target compounds were obtained by reacting the intermediates 5a and 5b with the different substituted thioureas 7a–7f in refluxing EtOH. The derivatives 8g and 8i bearing the free carboxylic acid moiety were simply prepared by the hydrolysis of the ester (Scheme 2). A further modification was based on the scaffold hopping approach, replacing the bithiazole scaffold of MR459 with a thiazole–imidazole scaffold, which also introduced an additional nitrogen atom that may facilitate the outer membrane permeation. Initially, we synthesized the 1-(3-(trifluoromethyl)phenyl)guanidine 9 through the reaction between the substituted aniline and cyanamide. Subsequently, compound 9 was subjected to a reaction with intermediate 5a, ultimately yielding the desired final compound 10 (Scheme 3). All synthesized compounds were screened for their direct-acting and synergistic antibacterial activity against selected Gram-positive and Gram-negative (see Experimental section for details) organisms, for their efficacy against PI4KB and for their inhibitory activity (IC₅₀) against selected viruses. In particular, we decided to test all compounds against hRVA16 and EVD68, belonging to the Picornaviridae family, and against ZIKV, belonging to the Flaviviridae family, in order to examine the pan-antiviral activity of these compounds. Compound 8a showed an improved broad-spectrum antiviral activity in the low micromolar range, while maintaining the absence of activity on PI4KB as compound MR-459. Moreover, even if it did not show any direct-acting antibacterial activity on the tested strains (MIC, >64 μg mL⁻¹), it resulted to be active against E. coli and Acinetobacter baumannii in the presence of a subinhibitory concentrations of colistin, with a noticeable synergistic activity (MAC values, 2 and 4 μg mL⁻¹, respectively). Among the first set of compounds 8b–8e, none showed direct-acting activity on Gram-negative strains (MIC, >64 μg mL⁻¹), but compound 8d, characterized by the siderophore moiety, is slightly active as an antiviral and antibacterial agent against Gram-positive strains; compound 8e has an interesting antibacterial activity against Gram-positive bacteria, especially against S. pyogenes and S. aureus (MIC values, 4 μg mL⁻¹ and 16 μg mL⁻¹, respectively), being bactericidal on these organisms at 32 μg mL⁻¹. Compounds 8f–8i, bearing both the carboxy and the trifluoromethyl moieties on the right phenyl ring, could overall be considered as interesting antivirals, especially 8f, also active against ZIKV, and 8h, both characterized by the methyl ester group. Otherwise, 8g and 8i, featuring the acid moiety, are good antibacterial agents against the whole panel of Gram-positive bacteria, although devoid of any activity (both direct-acting and synergistic) on Gram-negative organisms. These two compounds appear particularly active against S. pyogenes without any noticeable cytotoxic activity on human cell lines, and therefore a potentially promising selectivity. Finally, compounds 8j and 10, characterized by an additional nitrogen in different positions, can be considered interesting antivirals. In addition, 8j exhibits an interesting antibacterial activity against Gram-positive (especially Enterococcus faecalis and Streptococcus pyogenes, against which it is bactericidal at 8–16 μg mL⁻¹). Compound 8j, although devoid of any direct-acting activity on Gram-negative strains (MIC, >64 μg mL⁻¹), showed a noticeable synergistic activity in the presence of a subinhibitory concentration of colistin, a membrane-interacting antibiotic with well-documented permeabilizing properties (Table 1). Remarkably, low MAC values (2–4 μg mL⁻¹) were measured with clinically relevant species, such as Escherichia coli, Klebsiella pneumoniae and A. baumannii, but not P. aeruginosa. These data show that the lack of direct-acting activity of these compounds may rely, at least in part, on the lack of sufficient accumulation in the bacterial cell due to the poor diffusion through the outer membrane. Considering the interesting antibacterial activity exhibited by some compounds, the parent compound MR459, together with compounds 8a, 8e, 8g, 8i and 8j, were further tested on contemporary clinical isolates with well-characterised antibiotic resistance phenotypes (Table 2). Interestingly, the parent compound MR459 retained a good antibacterial activity on MRSA and vancomycin-resistant enterococci strains, although with a modest bactericidal activity. It did not show, as expected, any direct-acting antibacterial activity on Gram-negative isolates, but did exhibit a remarkable synergistic activity with colistin, with MAC values ranging from 0.03 to 2 μg mL⁻¹, except on the extensively drug-resistant K. pneumoniae isolate. Surprisingly, the growth of the MDR colistin-resistant P. aeruginosa isolate could be inhibited with a combination of MR459/colistin (2/2 μg mL⁻¹), somewhat in contrast with the lack of synergistic activity observed on the type strain (ATCC 27853), and potentially suggesting a high fitness cost of resistance in the clinical isolate. The MDR A. baumannii was highly susceptible to MR459 in combination with colistin, and a strikingly low concentration of MR459 (0.03 μg mL⁻¹) was able to revert colistin resistance. Unfortunately, none of the other compounds showed an improved activity on these antibiotic-resistant isolates, with the exception of compounds 8a and 8j, whose activity against the MDR colistin-resistant A. baumannii clinical isolate is greatly enhanced in the presence of a subinhibitory concentration of colistin.

Scheme 2. Reagents and conditions: (i) BzNCS, dry DCM, 25 °C, 1–2 h; (ii) NaOMe, dry MeOH, 25 °C, 1–8 h; (iii) H₂SO₄, dry MeOH, 15 h; (iv) 5a–5b, EtOH, reflux, 1–4 h; (v) LiOH, THF/MeOH, 60 °C, 4 h.

Scheme 3. Reagents and conditions: (i) cyanamide, HNO₃, EtOH/H₂O, reflux, 48 h; (ii) 5a, TEA, EtOH, reflux, 4 h.

Antibacterial activity of selected compounds against antibiotic-resistant clinical isolates.

Strain	Resistance phenotype	MIC/MBC or MACa (μg mL−1)
		MR-459	8a	8e	8g	8i	8j
Gram-positive isolatesb
S. aureus ATCC 43300	MRSA	4/32	—c	32/>64	8/>64	>64	16/>64
S. warneri SI-5/2011	Oxacillin-resistant	4/64	—	64/64	16/>64	>64	>64
S. haemolyticus SI-6/2011	Oxacillin-resistant	>64	—	32/64	16/>64	32/>64	>64
S. epidermidis SI-1266	Linezolid-resistant	16/>64	—	32/64	16/64	32/>64	>64
E. faecium SI-759	Vancomycin- and teicoplanin-resistant	2/4	—	16/64	16/>64	64/>64	64/>64
E. faecium SI-1831	Vancomycin-resistant	2/>64	—	32/>64	16/>64	>64	>64
E. faecium SI-3856B	Vancomycin-resistant	4/>64	—	32/>64	32/>64	>64	>64
Gram-negative isolatesd
E. coli SI-63	Colistin-resistant, mcr-positive	0.5	1	—	—	—	1
K. pneumoniae BO04	XDR, colistin-resistant	>64	>64	—	—	—	>64
A. baumannii N50	MDR, carbapenem- and colistin-resistant	0.03	0.06	—	—	—	0.06
P. aeruginosa SI-270	MDR, colistin-resistant	2	16				64

^aAll experiments were performed in triplicate.

^bThe minimum inhibitory concentration (MIC) and minimum bactericidal concentration (MBC, determined only for compounds showing MIC values ≤64 μg mL⁻¹) were determined for Gram-positive isolates.

^c—, not determined.

^dNone of the compounds exhibited direct-acting antibacterial activity on the tested Gram-negative strains (MIC, >64 μg mL⁻¹). Therefore, the minimum antibacterial concentration (MAC) was determined in the presence of a fixed subinhibitory concentration (2 μg mL⁻¹) of colistin.

Analyzing the comprehensive dataset presented in Table 1, a notable distinction emerges when comparing 4-trifluoromethyl bithiazoles to the previously published 4-methyl bithiazoles,^34–36 such as MR250. It becomes evident that the antiviral activity of 4-trifluoromethyl bithiazoles cannot be ascribed to PI4KB inhibition, contrary to the observed effects in the case of MR250. To gather more information, we selected the highly promising broad-antimicrobial candidate 8a for an in-depth investigation into its antiviral mechanism against hRVA16. This choice was motivated by its superior potency against this particular virus. Time-of-addition assays were performed by adding the compounds before, during or after infection and quantifying the viral RNA intracellularly and the viral infectious particles in the supernatant. The results, shown in Fig. 2, indicate that the tested compound is mostly active when added post-infection, suggesting an inhibition of replication or of late stages of infection, given that both RNA intracellular levels (Fig. 2A) and the release of infectious viruses are affected (Fig. 2B). Further studies were conducted in order to verify the mechanism of action. By growing hRVA16 in the presence of increasing concentrations of 8a or rupintrivir, a known inhibitor of hRV 3C protease, the selective pressure was increased. After 7 passages it was possible to select for resistance against rupintrivir, verified both with a shift in the EC₅₀ (Table 3) and through sequencing of the 3C protease which evidenced the presence of two previously reported mutations: T130A and I160M (Fig. S1A and B†).⁴⁵ We observed a shift as well in the EC₅₀ of 8a, and upon sequencing we identified a mutation in the 3A protein (I61T) (Fig. S1A†). Further analysis is needed to identify the role of the mutation; however, 3A is a protein involved in the formation of the replication organelles, and the mutation suggests an involvement of compound 8a in the replication of the virus, as supported as well by the results of Fig. 2. To verify if the resistance to 8a or rupintrivir conferred cross-resistance to other antivirals active against PI4KB, enviroxime, a known inhibitor, was tested against the resistant viruses.⁴⁶ The results are shown in Table 3 and reveal the lack of appearance of cross-resistance, confirming the different targets of the antiviral molecules. Moreover, considering the appearance of resistance against 8a, a possible solution could be to administer it in combination with other antivirals with known activity. Therefore, we tested 8a and rupintrivir or 8a and enviroxime for synergistic activity. The results showed generally an additive effect, with some areas of synergistic effect (Fig. 3). The combination of drugs can therefore have a double advantage: lower doses of drugs are needed to achieve inhibition and reduction of risk of resistance

Fig. 2. Time of addition on hRVA16. The compound 8a (5 μM) was added to cells before, during or at different times post-infection. The infectivity was evaluated by cell lysis and quantification of intracellular viral RNA (A) or through titration of the supernatant collected at 24 hpi (B). The results are the mean and SEM of three independent experiments.

Antiviral activity against resistant viruses.

Cpd	EC50 (μM) vs. RVA16 wt	EC50 (μM) vs. RUP p7	EC50 (μM) vs.8a p7
Enviroxime	0.0499	0.0271	0.0278
8a	0.491	1.04	4.322
Rupintrivir (RUP)	0.00204	0.0144	0.00478

Fig. 3. Synergistic effect. 8a and rupintrivir (A)and 8a and enviroxime (B) were added with a matrix of concentrations on HeLa cells 1 hpi with hRVA16. The infectivity was evaluated 48 hpi. The synergism was calculated using SynergyFinder with the ZIP synergy score. Results are from 2 independent experiments.

Conclusions

The aim of the present work was to identify broad antimicrobial agents for potential applications in the treatment of co-infections or to mitigate the onset of secondary bacterial infections induced by viruses. Indeed, a preliminary screening on previously identified antibacterial compounds revealed the noteworthy antiviral properties of a 4-CF₃ bithiazole derivative (MR459). Despite its lack of activity against PI4KB, the target typically inhibited by 4-CH₃ bithiazole antivirals, MR459 demonstrated the ability to inhibit the replication of hRV16, EVD68, and ZIKV. The replacement of a CH₃ moiety with a CF₃ at position C4 of the bithiazole scaffold was solely accountable for the loss of activity against PI4KB, while retaining antiviral properties and concurrently gaining a new antibacterial effect. To further analyse the biologically relevant chemical space around the 4-CF₃ bithiazole chemotype, we synthesized a series of 4-CF₃ bithiazole analogues to be screened against PI4KB, selected viruses (hRVA16, EVD68, and ZIKV) and a panel of Gram-positive and Gram-negative bacteria.

From this screening, we identified an interesting broad-spectrum antiviral (8f), with low micromolar and sub-micromolar activity against the above mentioned viruses, and two interesting broad antimicrobial candidates (8a and 8j). The latter compounds showed in fact low micromolar activity against all tested viruses (with no activity against PI4KB) and encouraging antibacterial activity: while 8a showed synergistic activity on some Gram-negative bacteria (E. coli and A. baumannii) in the presence of a subinhibitory concentration of colistin, 8j showed a broader spectrum of activity, including direct-acting bactericidal activity on Gram-positive and synergistic activity against relevant Gram-negative bacteria with overall lower MIC, MBC and MAC values. Interestingly, the parent compound MR459 and, to some extent, compounds 8a and 8j, were also active on antibiotic-resistant clinical isolates. Moreover, due to the lack of correlation between PI4KB inhibition and antiviral activity of these compounds, we investigated the mechanism of action of the most potent 8a against hRVA16. The results showed an inhibition of replication and the lack of cross-resistance between enviroxime, a known PI4KB inhibitor. 8a also showed additive and synergistic effects in combination with either rupintrivir or enviroxime, underscoring a different mechanism of action with respect to common 3C protease and PIKB inhibitors. Considering the urgent need to combat efficiently viral and bacterial infections or co-infections, this work paves the way for the development of new agents with broad antimicrobial properties. The polypharmacological effect of these agents holds the potential for enhanced therapeutic efficacy compared to the separate administration of antiviral and antibacterial drugs. Further studies are ongoing to improve the potency of this class of molecules and better understand their mechanism of action. Future work will be directed in selecting resistance in the presence of the combination of molecules and on the efficacy on more relevant respiratory tissue models. These additional studies may lead to the development of broad-antimicrobial agents for the treatment of acute infections/co-infections, also preventing or mitigating the severity of secondary bacterial infections.

Experimental section

General

All commercially available chemicals were purchased from Fisher Scientifics or Fluorochem and, unless otherwise noted, used without any previous purification. Solvents used for work-up and purification procedures were of technical grade. TLC was carried out using Merck TLC plates (silica gel on Al foils, SUPELCO Analytical). Where indicated, products were purified by silica gel flash chromatography on columns packed with Merck Geduran Si 60 (40–63 μm). ¹H and ¹³C NMR spectra were recorded on BRUKER AVANCE 300 MHz and BRUKER AVANCE 400 MHz spectrometers. Chemical shifts (δ scale) are reported in parts per million relative to TMS. ¹H NMR spectra are reported in this order: multiplicity and number of protons; signals were characterized as s (singlet), d (doublet), dd (doublet of doublets), ddd (doublet of doublet of doublets), t (triplet), m (multiplet), bs (broad signal). Elemental analyses were performed on a FlashSmart CHNS analyzer (Thermo Fisher) with gas-chromatographic separation. All final compounds were >95% pure as determined by elemental analysis (within 0.4% of the theoretical values). Low-resolution mass spectrometry measurements were performed on an Agilent InfinityLab LC/MSD iQ, Single Quadropole analyzer and are reported in the form of m/z.

Chemistry

General procedure for the synthesis of intermediate 1-(2-amino-4-(trifluoromethyl)thiazol-5-yl)ethan-1-one 3

To a solution of 1,1,1-trifluoropentane-2,4-dione 1 (0.607 mL; 5.00 mmol) in anhydrous acetonitrile (10 ml), hydroxy(tosyloxy)-iodobenzene (2535 mg; 6.00 mmol) was added and the mixture was refluxed for 90 min under an argon atmosphere. The mixture was cooled to room temperature (r.t.) and thiourea (457 mg; 6.00 mmol) was added. The mixture was refluxed for 12 h under an argon atmosphere. The solvent was concentrated under reduced pressure, and the crude was purified by flash chromatography using hexane/ethyl acetate 80 : 20 as eluent.

Yield: 70%. ¹H NMR (400 MHz, DMSO-d₆): δ 2.39 (s, 3H); 8.28 (s, 2H).

General procedure for the synthesis of intermediates 4a and 4b

1-(2-Amino-4-(trifluoromethyl)thiazol-5-yl)ethan-1-one 3 (690 mg; 3,3 mmol) was suspended in anhydrous THF (6 mL) and the mixture was cooled to 0 °C. Pyridine (0.6 mL) was added, followed by a dropwise addition of pivaloyl chloride (1.02 mL; 8.3 mmol) for compound 4a and of benzoyl chloride (1.15 mL; 9.9 mmol) for compound 4b. The reaction mixture was allowed to warm to r.t. and heated at reflux o.n. after cooling, H₂O and EtOAc were added, and the aqueous phase was extracted three times with ethyl acetate. The combined organic layers were washed three times with a saturated aqueous NH₄Cl solution and two times with brine, dried over Na₂SO₄ and concentrated under reduced pressure. The intermediate 4a was used for the next step without further purification, while 4b was purified by flash chromatography using petroleum ether/ethyl acetate from 95 : 5 to 80 : 20 as eluent.

N-(5-Acetyl-4-(trifluoromethyl)thiazol-2-yl)pivalamide 4a

Yield: 98%. ¹H NMR (400 MHz, DMSO-d₆): δ 1.26 (s, 9H); 2.50 (s, 3H); 12.67 (s, 1H).

N-(5-Acetyl-4-(trifluoromethyl)thiazol-2-yl)benzamide 4b

Yield: 52%. ¹H NMR (400 MHz, CDCl₃): δ 2.68 (s, 3H); 7.52 (t, 1H, J = 7.6 Hz); 7.61 (m, 2H); 8.09 (d, 1H, J = 7.7 Hz); 8.16 (d, 1H, J = 7.7 Hz), 10.96 (s, 1H).

General procedure for the synthesis of intermediates 5a and 5b

A solution of Br₂ (35 μL; 0.68 mmol) in 1,4-dioxane (1 mL) was added dropwise to a stirred solution of intermediates 4a and 4b (0.68 mmol) in dioxane (8 mL) and HBr 48% (0.5 mL). The mixture was stirred at r.t. o.n. after neutralization of HBr by addition of saturated aqueous NaHCO₃ solution. The mixture was extracted three times with dichloromethane, then dried over Na₂SO₄ and concentrated under vacuum. The crude was used for the next step without further purification.

N-(5-(2-Bromoacetyl)-4-(trifluoromethyl)thiazol-2-yl)pivalamide 5a

Yield: 90%. ¹H NMR (400 MHz, DMSO-d₆): δ 1.27 (s, 9H); 4.73 (s, 2H); 12.80 (s, 1H).

N-(5-(2-Bromoacetyl)-4-(trifluoromethyl)thiazol-2-yl)benzamide 5b

Yield: 95%. ¹H NMR (400 MHz, CDCl₃): δ 4.40 (s, 2H); 7.61 (m, 2H); 7.72 (m, 1H); 7.99 (m, 2H); 9.79 (s, 1H).

General procedure for the synthesis of intermediates 7a–7d and 7f

(A) Benzoyl isothiocyanate (0.439 mL; 3.26 mmol) was added dropwise to a solution of the appropriate commercial amine (3.26 mmol) in dichloromethane (10 mL). The mixture was stirred at room temperature for 90 min for compound 7a and for 12 h for 7b–7d and 7f. The crude was used in the next step without further purification. (B) Sodium (178 mg; 7.74 mmol) was added to anhydrous MeOH (3 mL) under an argon atmosphere and stirred until a transparent solution was obtained, where NaOMe was formed. This solution was added dropwise to a solution of the proper intermediate (2.58 mmol) in anhydrous MeOH (10 mL). The mixture was stirred under an argon atmosphere at room temperature for 60 min for 7a, for 5 h for 7b, for 12 h for 7c and 7d and for 90 min for 7f. The solvent was evaporated under vacuum. The crude was purified by flash chromatography using the proper eluent: petroleum ether/ethyl acetate 70 : 30 for 7a, dichloromethane/methanol from 98 : 2 + 1% formic acid to 90 : 10 + 1% formic acid for 7b and 7c, 70 : 30 petroleum ether/ethyl acetate + 1% formic acid for 7d and hexane/ethyl acetate from 90 : 10 to 80 : 20 for 7f.

1-(3-(Trifluoromethyl)phenyl)thiourea 7a

Yield: 90%. ¹H NMR (400 MHz, CDCl₃): δ 6.50 (m, 2H); 7.51 (m, 1H); 7.57 (m, 3H); 9.08 (s, 1H).

5-Thioureidoisophthalic acid 7b

Yield: 99%. ¹H NMR (400 MHz, DMSO-d₆): δ 7.71 (s, 2H); 8.19 (m, 1H); 8.29 (m, 2H); 10.05 (s, 1H); 13.24 (s, 2H).

2-Hydroxy-5-thioureidobenzoic acid 7c

Yield: 73%. ¹H NMR (400 MHz, DMSO-d₆): δ 6.84 (d, 1H, J = 8.7 Hz); 7.4 (m, 1H); 7.67 (m, 1H); 8.15 (s, 1H); 9.51 (s, 1H).

3-Thioureido-5-(trifluoromethyl)benzoic acid 7d

Yield: 92%. ¹H NMR (400 MHz, CDCl₃): δ 7.36 (m, 1H); 7.41 (m, 1H); 7.61 (m, 1H).

1-(6-(Trifluoromethyl)pyridin-2-yl)thiourea 7f

Yield: 20%. ¹H NMR (400 MHz, DMSO-d₆): δ 7.50 (d, 1H, J = 8.4 Hz); 7.56 (d, 1H, J = 7.5 Hz); 8.05 (t, 1H, J = 8.0 Hz); 9.16 (s, 1H); 9.93 (s, 1H); 10.94 (s, 1H).

General procedure for the synthesis of intermediate methyl 3-thioureido-5-(trifluoromethyl)benzoate 7e

To a solution of 3-thioureido-5-(trifluoromethyl)benzoic acid 7d in dry MeOH, 2 drops of H₂SO₄ were added. The reaction was stirred at reflux for 12 h. The solvent was removed under vacuum. The crude was purified by flash chromatography using dichloromethane/ethyl acetate 95 : 5 as eluent.

Yield: 32%. ¹H NMR (400 MHz, CDCl₃): δ 4.00 (s, 3H); 6.17 (s, 2H); 7.79 (m, 1H); 8.16 (m, 1H); 8.23 (m, 1H).

General procedure for the synthesis of compounds 8a–8f, 8h, and 8j

A solution of the proper intermediate 5a and 5b (0.15 mmol) and the proper thiourea 7a–7c, 7e, and 7f (0.15 mmol) in ethanol (2.5 mL) was heated at reflux for 1 h. After cooling, compounds 8c and 8e were obtained after filtration of the suspension. For compounds 8a, 8b, 8d, 8f, 8h, and 8j, water and ethyl acetate were added and the aqueous phase was extracted three times with ethyl acetate. The combined organic phases were washed with brine, dried over Na₂SO₄ and concentrated under vacuum. The crude was purified by flash chromatography using the proper eluent: hexane/ethyl acetate 96 : 4 for 8a, dichloromethane/methanol 98 : 2 + 1% formic acid for 8b, dichloromethane/ethyl acetate 90 : 10 + 1% formic acid for 8d, dichloromethane/ethyl acetate 98 : 2 for 8f and 8h and hexane/ethyl acetate 90 : 10 for 8j.

N-(4′-(Trifluoromethyl)-2-((3-(trifluoromethyl)phenyl)amino)-[4,5′-bithiazol]-2′-yl)benzamide 8a

Yield: 50%. ¹H NMR (400 MHz, CD₃OD): δ 7.14 (s, 1H); 7.28 (m, 1H); 7.54 (t, 1H, J = 8.2 Hz); 7.58 (m, 2H); 7.68 (m, 1H); 7.94 (m, 1H); 8.07 (m, 2H); 8.16 (m, 1H). ¹³C NMR (100 MHz, CD₃OD) δ 107.38; 113.30; 117.53; 119.97; 122.92; 125.62; 127.75; 128.50; 129.40; 130.88; 131.20; 131.83; 132.65; 133.02; 139.45; 141.56; 157.19; 162.92; 166.37. [M + H]⁺: 515.0 m/z.

5-((2′-Pivalamido-4′-(trifluoromethyl)-[4,5′-bithiazol]-2-yl)amino)isophthalic acid 8b

Yield: 49%. ¹H NMR (400 MHz, DMSO-d₆): δ 1.27 (s, 9H); 7.24 (s, 1H); 8.09 (t, 1H, J = 1.6 Hz); 8.47 (d, 2H, J = 1.5 Hz); 10.77 (s, 1H); 12.30 (s, 1H); 13.21 (s, 1H). ¹³C NMR (100 MHz, DMSO-d₆) δ 26.93; 39.34; 92.32; 109.45; 121.69; 123.32; 128.46; 132.67; 139.00; 141.70; 145.32; 158.03; 163.21; 167.09; 177.92. [M + H]⁺: 515.0 m/z.

5-((2′-Benzamido-4′-(trifluoromethyl)-[4,5′-bithiazol]-2-yl)amino)isophthalic acid 8c

Yield: 65%. ¹H NMR (400 MHz, DMSO-d₆): δ 7.29 (s, 1H); 7.58 (t, 2H, J = 7.6 Hz); 7.68 (t, 1H, J = 7.3 Hz); 8.10 (m, 1H); 8.15 (m, 2H); 8.49 (m, 2H); 10.79 (s, 1H); 13.19 (s, 2H). ¹³C NMR (100 MHz, DMSO-d₆) δ 109.49; 120.54; 121.67; 123.23; 123.40; 128.83; 129.15; 129.95; 131.81; 132.30; 132.81; 133.48; 138.93; 141.66; 158.08; 163.23; 166.13; 167.18. [M + H]⁺: 535.0 m/z.

2-Hydroxy-5-((2′-pivalamido-4′-(trifluoromethyl)-[4,5′-bithiazol]-2-yl)amino)benzoic acid 8d

Yield: 35%. ¹H NMR (400 MHz, DMSO-d₆): δ 1.27 (s, 9H); 6.96 (d, 1H, J = 8.9 Hz); 7.10 (s, 1H); 7.83 (dd, 1H, J = 8.9, 2.9 Hz); 8.03 (d, 1H, J = 2.8 Hz); 10.28 (s, 1H); 10.98 (s, 1H); 12.27 (s, 1H). ¹³C NMR (100 MHz, DMSO-d₆): δ 26.92; 39.33; 107.98; 113.28; 118.01; 118.87; 120.60; 123.29; 126.26; 128.84; 133.33; 138.90; 156.68; 157.87; 164.13; 172.17; 177.97. [M + H]⁺: 487.1 m/z.

5-((2′-Benzamido-4′-(trifluoromethyl)-[4,5′-bithiazol]-2-yl)amino)-2-hydroxybenzoic acid 8e

Yield: 63%. ¹H NMR (400 MHz, DMSO-d₆): δ 6.98 (d, 1H, J = 8.9 Hz); 7.17 (s, 1H); 7.58 (t, 2H, J = 1.3 Hz); 7.68 (m, 1H); 7.81 (dd, 1H, J = 9.0, 2.9 Hz); 8.10 (d, 1H, J = 2.8 Hz); 8.16 (m, 2H); 10.31 (s, 1H); 11.00 (s, 1H); 13.1039 (s, 1H). ¹³C NMR (100 MHz, DMSO-d₆) δ 108.09; 113.24; 118.03; 118.83; 126.29; 128.83; 129.16; 131.74; 133.30; 133.51; 138.79; 156.70; 157.79; 164.10; 166.10; 172.19. [M + H]⁺: 507.0 m/z.

Methyl 3-((2′-pivalamido-4′-(trifluoromethyl)-[4,5′-bithiazol]-2-yl)amino)-5-(trifluoromethyl)benzoate 8f

Yield: 61%. ¹H NMR (400 MHz, CDCl₃): δ 1.37 (s, 9H); 4.04 (s, 3H); 7.03 (s, 1H); 7.64 (s, 1H); 8.01 (m, 1H); 8.08 (m, 1H); 8.47 (m, 1H); 9.03 (s, 1H). ¹³C NMR (100 MHz, CDCl₃) δ 27.15; 39.17; 52.76; j107.91; 118.15; 119.92; 120.37; 121.35; 122.06; 122.62; 124.77; 129.00; 132.26; 139.84; 140.71; 156.70; 162.04; 165.70; 176.38. [M + H]⁺: 553.1 m/z.

Methyl 3-((2′-benzamido-4′-(trifluoromethyl)-[4,5′-bithiazol]-2-yl)amino)-5-(trifluoromethyl) benzoate 8h

Yield: 34%. ¹H NMR (400 MHz, CD₃OD): δ 4.05 (s, 3H); 7.17 (s, 1H); 7.58 (m, 2H); 7.65 (m, 1H); 7.97 (m, 1H); 8.06 (m, 2H); 8.31 (m, 1H); 8.72 (m, 1H). ¹³C NMR (100 MHz, DMSO-d₆) δ 53.15; 110.06; 117.26; 118.03; 121.21; 122.77; 123.70; 128.82; 129.04; 129.19; 129.74; 131.11; 131.72; 132.04; 133.54; 139.06; 142.38; 157.74; 162.83; 165.49; 166.18. [M + H]⁺: 573.0 m/z.

N-(4′-(Trifluoromethyl)-2-((6-(trifluoromethyl)pyridin-2-yl)amino)-[4,5′-bithiazol]-2′-yl)pivalamide 8j

Yield: 54%. ¹H NMR (400 MHz, CDCl₃): δ 1.29 (m, 9H); 6.79 (d, 1H, J = 8.32 Hz); 7.19 (s, 1H); 7.21 (d, 1H, J = 7.3 Hz); 7.62 (t, 1H, J = 7.9 Hz); 9.21 (s, 1H); 10.47 (s, 1H). ¹³C NMR (100 MHz, CDCl₃): δ 26.04; 38.12; 111.43; 111.88; 112.33; 118.97; 121.70; 128.27; 132.04; 136.10; 137.73; 144.59; 150.03; 155.98; 159.30; 175.48. ¹⁹F NMR (376 MHz, CDCl₃): δ −68.26; −60.47. [M + H]⁺: 496.0 m/z.

General procedure for the synthesis of compounds 8g and 8i

To a solution of 8f and 8h (0.068 mmol) in tetrahydrofuran/methanol 1 : 1 (2 mL), a solution of LiOH (0.200 mL) was added and the mixture was heated at 60 °C for 4 h. After cooling to r.t., the solvent was removed. The crude was purified by flash chromatography using the proper eluent: dichloromethane/methanol from 98 : 2 to 95 : 5 for 8g and dichloromethane/methanol from 98 : 2 + 1% formic acid to 95 : 5 + 1% formic acid for 8i.

3-((2′-Pivalamido-4′-(trifluoromethyl)-[4,5′-bithiazol]-2-yl)amino)-5-(trifluoromethyl)benzoic acid 8g

Yield: 89%. ¹H NMR (400 MHz, DMSO-d₆): δ 1.27 (s, 9H); 7.28 (s, 1H); 7.79 (s, 1H); 8.11 (s, 1H); 8.70 (s, 1H); 11.00 (s, 1H). ¹³C NMR (100 MHz, DMSO-d₆) δ 26.92; 29.47; 109.51; 115.03; 118.58; 121.81; 126.25; 128.26; 132.04; 139.12; 140.11; 141.77; 148.38; 149.64; 157.66; 163.17; 170.21; 177.96. [M + H]⁺: 539.0 m/z.

3-((2′-Benzamido-4′-(trifluoromethyl)-[4,5′-bithiazol]-2-yl)amino)-5-(trifluoromethyl)benzoic acid 8i

Yield: 99%. ¹H NMR (400 MHz, DMSO-d₆): δ 7.34 (s, 1H); 7.56 (m, 2H); 7.67 (t, 1H, J = 7.4 Hz); 7.77 (s, 1H); 8.15 (m, 4H); 8.75 (s, 1H); 10.97 (s, 1H); 13.16 (s, 1H). ¹³C NMR (100 MHz, DMSO-d₆) δ 109.50; 112.33; 118.50; 121.80; 123.23; 126.05; 128.81; 129.08; 130.23; 133.23; 139.31; 141.90; 148.54; 150.60; 160.54; 163.07; 167.61; 171.38; 186.39. [M + H]⁺: 559.0 m/z.

General procedure for the synthesis of compound 1-(3-(trifluoromethyl)phenyl)guanidine 9

Nitric acid 69% (0.138 mL, 3.11 mmol) was added dropwise to a stirred solution of 3-(trifluoromethyl)aniline (0.387 mL, 3.11 mmol) in ethanol (5 mL). A solution of cyanamide (653 mg, 15.5 mmol) in water (1 mL) was added. The mixture was refluxed for 12 h. After cooling to r.t., the solvent was evaporated under reduced pressure. NaHCO₃ and ethyl acetate were added and the aqueous phase was extracted three times with ethyl acetate; the combined organic layers were concentrated under reduced pressure.

Yield: 100%. ¹H NMR (400 MHz, DMSO-d₆): δ 5.33 (s, 1H); 6.61 (s, 2H); 7.02 (m, 2H); 7.08 (d, 1H, J = 7.8 Hz); 7.36 (t, 1H, J = 7.8 Hz).

General procedure for the synthesis of compound N-(4-(trifluoromethyl)-5-(2-((3-(trifluoromethyl)phenyl)amino)-1H-imidazol-4-yl)thiazol-2-yl)pivalamide 10

A solution of N-(5-(2-bromoacetyl)-4-(trifluoromethyl)thiazol-2-yl)pivalamide 5a (50 mg, 0.134 mmol), 1-(3-(trifluoromethyl)phenyl)guanidine 9 (33 mg, 0.161 mmol) and triethylamine (0.037 mL, 0.268 mmol) in ethanol (4 mL) was heated at reflux for 4 h. After cooling, the solvent was removed under vacuum. NH₄Cl and ethyl acetate were added and the aqueous phase was extracted with ethyl acetate three times. The combined organic phases were washed with brine, dried over Na₂SO₄ and concentrated under vacuum. The crude was purified by flash chromatography using hexane/ethyl acetate 80 : 20 as eluent.

Yield: 21%. ¹H NMR (400 MHz, CDCl₃): δ 1.34 (s, 9H); 4.94 (s, 1H); 7.06 (s, 1H); 7.71 (m, 3H); 7.76 (s, 1H); 9.09 (s, 1H). ¹³C NMR (100 MHz, CDCl₃) δ 26.12; 38.10; 113.71; 120.77; 121.87; 123.57; 124.15; 125.97; 127.13; 128.06; 129.86; 131.57; 131.90; 135.76; 146.52; 154.92; 175.16. [M + H]⁺: 478.2 m/z.

In vitro kinase inhibition assays

Recombinant full length, HIS6-tagged PI4KIIIβ was purchased from ProQinase (Germany). Reactions were performed in 15 μl at 30 °C for 10 min using 25 μM ATP, 50 μM PI:3PS and 0.4 ng μL⁻¹ PI4KIIIβ in 250 mM HEPES NaOH pH 7.5, 15 mM MnCl₂, 5 mM EGTA, 500 mM NaCl, 0.15% CHAPS, 10 mM DTT, and 10% DMSO. To avoid protein adsorption to the plastic surface, protein low-binding tubes were used. ADP-Glo kinase assay (Promega) was then used to detect kinase activity according to the manufacturer's instruction with minor modifications. In detail, reactions were transferred to white 384-well plates and stopped by adding 10 μl of ADP-Glo reagent (Promega) for 50 min at room temperature. 20 μL of detection reagent (Promega) was then added for 30 min and the reaction was read using a GloMax Discover microplate reader (Promega). Data were plotted using GraphPad Prism 5.0. ID₅₀ values were obtained according to eqn (1):

v = V/{1 + (I/ID₅₀)} 1

where v is the measured reaction velocity, V is the apparent maximal velocity in the absence of inhibitor, I is the inhibitor concentration, and ID₅₀ is the 50% inhibitory dose.

Antiviral assays

RVA16 and EVD68 were a kind gift from the laboratory of Caroline Tapparel at the University of Geneva, and ZIKV from the Labor Spiez. Antiviral activity was determined on HeLa cells grown at 33 °C for 48 h for RVA16 and EVD68, and on VeroE6 cells grown at 37 °C for 72 h for ZIKV. The infection was evaluated through immunostaining with primary antibody directed against viral proteins (rhinovirus VP3 monoclonal antibody G47A, Enterovirus pan-monoclonal antibody L66J, and Flavivirus E protein 4G2) and a secondary antibody conjugated with HRP. Compounds were evaluated in parallel for cytotoxicity on HeLa and VeroE6 cells by MTT assay.

Time-of-addition assays were carried out by infecting HeLa cells at MOI 0.1 while adding the compound 8b at 5 μM either 1 h before infection, during the infection or at 1, 2, 4 and 6 h post infection. The supernatant was collected 24 h after infection and titrated on HeLa cells while the cells were lysed, and the RNA was extracted and subjected to specific qPCR for RV (primer forward 5′-AGCCTGCGTGGCKGCC-3′ and 5′-CYNAGCCNTGCGTGG-3′, primer reverse 5′-GAA ACA CGG ACA CCC AAA GTA GT-3′, probe CTC CGG CCC CTG AAT GYG GCT AA).

Selection of resistance was carried out by infecting HeLa cells with RVA16 and treating the cells with rupintrivir or 8b. The supernatants were collected when the cells displayed extensive cytopathic effect (3 to 6 dpi) and clarified. Each sample was titrated on Hela cells. For the following passages, cells were infected with the virus of the previous corresponding passage (MOI 0.01) and the concentration of the compound was doubled when possible. After 7 passages the RNA was extracted and sequenced. The EC₅₀ and CC₅₀ values for inhibition curves were calculated by regression analysis using the program GraphPad Prism version 9.1 to fit a variable slope sigmoidal dose–response curve. The synergism data were generated with SynergyFinder.

Antibacterial assays

Bacterial strains, including representatives of both Gram-positive and Gram-negative bacteria, were obtained from the ATCC or CCUG. Compounds were resuspended in DMSO at a final concentration of 50 mg mL⁻¹ and subsequently diluted in the culture medium. Minimum inhibitory concentrations (MICs) and minimum bactericidal concentrations (MBCs) of the compounds were determined in triplicate in Mueller–Hinton medium using the microdilution broth method and a bacterial inoculum of 5 × 10⁴ CFU per well, as recommended by the Clinical Laboratory Standards Institute (CLSI).⁴⁷ MICs and MBCs were recorded after 18 h of incubation at 35 ± 2 °C. The minimum antibacterial concentration (MAC) of the compounds was determined in the presence of a fixed and subinhibitory (corresponding to 0.5 × MIC) concentration of colistin (0.12 μg mL⁻¹ for E. coli CCUGT and K. pneumoniae ATCC 13833; 0.25 μg mL⁻¹ for A. baumannii ATCC 17978 and P. aeruginosa ATCC 27853).⁴⁸ Selected compounds were also tested on clinical isolates with well-characterized resistance phenotypes (Table 2) that were present in our collection and were described elsewhere.⁴⁹ Similarly, both the MIC and MAC values were determined with colistin-resistant Gram-negative isolates (colistin MIC values for E. coli SI-63, K. pneumoniae BO04, A. baumannii N50 and P. aeruginosa SI-270; 4, 256, 16 and 128 μg mL⁻¹, respectively), but using a fixed concentration of colistin equal to 2 μg mL⁻¹, i.e. corresponding to the EUCAST resistance breakpoint for Enterobacterales and A. baumannii.

Author contributions

The manuscript was written through the contributions of all authors. All authors have given approval to the final version of the manuscript.

Conflicts of interest

The authors declare no conflict of interest.