Synthesis and structure–activity relationship of novel thiazole aminoguanidines against MRSA and Escherichia coli

Abstract

Novel lead thiazole aminoguanidines exhibited strong activity against Gram-positive bacteria. The potential targets of these substances are undecaprenyl diphosphate synthase (UPPS) and undecaprenyl diphosphate phosphatase (UPPP). Here, we report the synthesis and antibacterial evaluation of a library of thiazole aminoguanidines analogues, wherein the rotatable bond is inserted between the C2 position of thiazole and hydrophobic group. The molecular flexibility is increased, and new analogues with strong activity against MRSA and E. coli are produced. The best compound 4i showed rapid sterilization and low tendency to induce bacterial resistance. The IC₅₀ of compound 4i to EcUPPS enzyme is 145 μmol L⁻¹ (58 μg mL⁻¹). Compound 4i can also inhibit and destroy bacterial biofilms. These thiazole aminoguanidines can be developed as potential therapeutic candidates in the future.

Increasing molecular flexibility improves the performance of thiazole aminoguanidines against drug-resistant bacteria and Gram-negative bacteria.

Introduction

The World Health Organization reported that infections with drug resistant pathogens contribute to almost 5 million deaths per year.¹ In the United States, there are more than 2.8 million cases of antibiotic-resistant bacterial infections every year, and more than 35 000 people die directly.² The surveillance of antimicrobial resistance in China in 2021 showed that among the 300 000 clinical isolates from patients infected with bacteria, Gram-positive bacteria accounted for 28.6% and Gram-negative bacteria accounted for 71.4%, and the detection rate of methicillin-resistant Staphylococcus aureus (MRSA) was 30%.³ Therefore, new chemical class antibacterial compounds need to be developed.⁴

Designing and developing new antibacterial agents for potential targets is an effective approach to combat the rise of bacterial drug resistance. Another strategy is to optimize the known lead compound and study the structure–activity relationship (SAR). Undecaprenyl diphosphate synthase (UPPS) and undecaprenyl diphosphate phosphatase (UPPP) are essential bacterial enzymes and participate in cell wall biosynthesis. UPPS and UPPP do not exist in the human genome, so they are attractive targets for developing antibiotics.^5,6 Thiazole aminoguanidine compounds are new antibacterial lead compounds discovered in recent years, which not only belong to a novel chemical class, but also act on new targets and have great research value.⁷ Lead compound 1 (see Fig. 1) consists of a thiazole central ring and two characteristic structures connected to the thiazole ring: an electrophilic characteristic structure (guanidine group) at the C5 terminal and a lipophilic characteristic structure (n-butylphenyl group) at the C2 terminal. The lead compound 1 showed obvious bactericidal activity against Gram-positive bacteria.^7–10 Bacterial cytological analysis showed that thiazole aminoguanidines work by inhibiting the biosynthesis of bacterial cell walls.^11,12 Further detection of the concentration change in bacterial soluble cell wall precursor (UDP-N-acetylmuramyl pentapeptide), target over-expression test, and enzyme inhibition test confirmed that thiazole aminoguanidines target UPPS and UPPP, exhibiting a double inhibitory effect.¹²

Fig. 1. Chemical structures of thiazole aminoguanidine lead compound 1 and its new design.

Unfortunately, thiazole aminoguanidines are mainly effective against Gram-positive bacteria;^13–15 however, very few are effective against Gram-negative bacteria,^16,17 which limits its future application. UPPS and UPPP are essential in both Gram-positive and Gram-negative bacteria.¹⁸ In theory, thiazole aminoguanidines can resist Gram-negative bacteria. The poor effect on Gram-negative bacteria may be because penetration of the outer membrane of Gram-negative bacteria and into the periplasmic space or bacterial cells by thiazole aminoguanidines is challenging. Bacterial porins form the main pathway for antibiotics to enter bacterial cells. Porin channels distributed on the outer membrane are like an hourglass, and the inner wall is filled with various amino acid side groups. The size of the central contraction area is small, and the positively and negatively charged side groups are arranged face to face to form an electrostatic field. When drug molecules enter the channel, the direction of the molecular electric dipole moment should be constantly adjusted to crawl forward, especially through the central contraction zone.¹⁹ Thus, the key principle underlying the process of thiazole aminoguanidines inhibiting Gram-negative bacteria is that they can pass through the pore protein of the outer membrane of Gram-negative bacteria after structural optimization.

By increasing the rotatable bond of thiazole aminoguanidines, the molecular pose and electric dipole moment orientation can be easily adjusted when they pass through the Gram-negative outer membrane pore protein to enter the bacterial cell. In this work, we report the development of novel molecular flexible modified analogues of thiazole aminoguanidines, which have strong activity against MRSA and Escherichia coli.

Results and discussion

Chemistry

The derivatives of thiazole aminoguanidine were constructed through a multi-step reaction according to the preparation route in Scheme 1. Initially, the thiosemicarbazones 2a–o were synthesized by Schiff's base reaction of substituted aromatic aldehydes with thiosemicarbazone.^20,21 Then, the intermediate compounds 3a–o were prepared by Hantzsch's thiazole synthesis between α-halogenated 1,3-diketone and thiosemicarbazones 2a–o.^22,23 Subsequently, the target compounds 4a–o were obtained by Schiff's base reaction between aminoguanidine hydrochloride and the intermediate compounds 3a–o. The structures of compounds 4a–o were confirmed by their HRMS and ¹H NMR spectroscopy (see Experimental section). Considering the convenient reaction conditions and wide substrate range, this method can be used to prepare more thiazole aminoguanidines.

Scheme 1. Synthetic route for thiazole aminoguanidines 4a–o. Reagents and conditions: (i) ethanol, reflux, 6 h; (ii) ethanol, reflux, 3 h; (iii) LiCl as a catalyst, ethanol, reflux, 24 h.

Antibacterial activity

The in vitro antibacterial data of thiazole aminoguanidines 4a–o and norfloxacin are depicted as the minimum inhibitory concentration (MIC, μg mL⁻¹) and minimum bactericidal concentration (MBC, μg mL⁻¹) in Table 1. Table 1 shows that most of the target compounds 4a–i and 4k–o can inhibit gram negative bacteria (E. coli) as expected, highlighting the beneficial role of molecular flexibility (A rotatable single bond is added between aromatic group and thiazole ring.) in optimizing the inhibition of gram negative bacteria activity.^24–26 Compounds 4a–i and 4k–o still have moderate to good activity against S. aureus and MRSA strains. Compounds 4e, 4g, 4i and 4m gave 8-fold strong activity (MIC = 8 μg mL⁻¹) in comparison to norfloxacin against MRSA (MIC = 64 μg mL⁻¹). In particular, compound 4l can obviously inhibit the growth of MRSA at a low concentration of 4 μg mL⁻¹, and this activity is 16-fold superior to that of norfloxacin. Compound 4j lost its inhibitory effect, which may be caused by too many hydrogen bond acceptors. A comparison of compounds 4a to 4h shows that the fluorine-substituted phenyl group (4e–g) is better than the hydroxyl-substituted phenyl group (4a–d), and the phenyl group is better than the thiazolyl group (4h) considering the presence of one aromatic substituent ring.

Minimum inhibitory concentrations and minimum bactericidal concentrations of thiazole aminoguanidines 4a–o against S. aureus, E. coli, and MRSA.

Compound	E. coliATCC 8739		S. aureusATCC 6538P		MRSAATCC 33591
	MIC (μg mL−1)	MBC/MIC	MIC (μg mL−1)	MBC/MIC	MIC (μg mL−1)	MBC/MIC
4a	64	1	32	1	64	1
4b	32	1	8	2	32	1
4c	64	1	32	2	128	>1
4d	32	1	32	1	64	1
4e	16	1	16	1	8	1
4f	64	2	16	2	16	1
4g	16	1	8	1	8	1
4h	64	2	64	>2	64	2
4i	8	1	2	2	8	1
4j	>128	—	>128	—	>128	—
4k	32	4	32	1	32	1
4l	16	4	4	1	4	1
4m	8	1	16	1	8	1
4n	128	>1	32	>1	32	1
4o	32	1	128	1	32	1
Norfloxacin	0.12	>8	0.25	1	64	>2
Lead 1	—	—	—	—	2.8a	—

^aIndicates the antibacterial effect of lead 1 on MRSA ATCC 43300.⁷

A comparison of compounds 4i to 4n showed that thiophene (4i) and fluorine substituted phenyl (4l and 4m) are better than nitrogen-containing and oxygen-containing substitutions (4j, 4k, and 4n) when a ring group is added to the aromatic substituent. Compound 4o shows that the increasing molecular flexibility must be controlled within a suitable range, and an excessive increase in molecular flexibility also leads to adverse effects. A comparison of compounds 4e–g with 4i, 4l, and 4m indicates that proper extension of the molecule is beneficial to enhance the antibacterial effect of thiazole aminoguanidines. Considering MIC and MBC comprehensively, 4i is taken as the optimal structure for further research.

Time-kill kinetic analysis

Compared with bacteriostatic agents, antibacterial agents with bactericidal activity are considered to be helpful to recover from infection faster and provide better clinical results. To further prove that compound 4i is a bactericidal agent, we next checked how rapidly compound 4i can kill MRSA. Using a standard time-kill assay, MRSA was exposed to compounds 4i and norfloxacin at 1, 2, and 4× MIC. Samples were collected at specific time points and transferred to Mueller Hinton agar (HBA) plates to determine the number of viable bacteria remaining after treatment. As shown in Fig. 2, compound 4i exhibited bactericidal activity at all tested concentrations. The time to achieve sterilization differs depending on the concentration of the test compound. For compound 4i, MRSA was completely eliminated after 4 h at 1× MIC, and after only 2 h at 2× MIC and 4× MIC. However, norfloxacin could not completely eradicate MRSA even after 24 h. These results indicate that norfloxacin can only inhibit MRSA, and compound 4i kills MRSA. Compound 4i exhibits rapid bactericidal activity, which is more conducive to reducing the emergence of bacterial resistance.

Fig. 2. Time-kill analysis of compound 4i and norfloxacin against MRSA at (a) MIC, (b) 2× MIC and (c) 4× MIC.

Bacterial resistance development

The low tendency of bacterial resistance development is an important criterion for developing and evaluating new antibacterial drugs. Therefore, the induction of bacterial resistance by compound 4i against MRSA was studied in laboratory. The MIC value of compound 4i against MRSA at the 10th generation was only increased two times compared to that at the first generation (Fig. 3). In contrast, under the same conditions, the MIC value of the reference drug norfloxacin increased four times. This result indicates that using compound 4i, it was difficult to induce resistance in MRSA.

Fig. 3. Resistance development of MRSA after repetitive treatments with compound 4i and norfloxacin.

Biofilm inhibition and eradication ability

Bacteria produce biofilms to protect themselves under stressful environments; thus, the formation of biofilms reduces the effectiveness of antibiotics against bacteria. Here, we evaluated the potential of compound 4i to impede the formation of MRSA biofilms by a crystal violet staining assay. As shown in Fig. 4a and b, after co-incubation of MRSA with 0.25× MIC and 0.5× MIC compound 4i, biofilm formation of MRSA was significantly decreased by 44% and 59%, respectively. These results indicate that compound 4i can effectively inhibit biofilm formation of MRSA. Additionally, the eradication activity of compound 4i on mature biofilms was further examined. The results showed that compound 4i could eradicate bacterial biofilms in a dose-dependent manners. At its MIC, compound 4i could eradicate 70% MRSA biofilms (Fig. 4c). Therefore, compound 4i can destroy adherent MRSA biofilms.

Fig. 4. (a) The photographs of the MRSA biofilms stained by crystal violet after treatment with compound 4i. (b) Inhibitory effects of compound 4i on biofilm formation of MRSA. (c) Effect of compounds 4i in eradicating an established MRSA biofilm. All experiments were performed in triplicate. ** Indicates statistically significant differences compared with the control (P < 0.01). *** Indicates statistically significant differences compared with the control (P < 0.001).

Safety evaluation

Biological safety is an essential indicator to evaluate the safety of new antibacterial agents. Therefore, we performed a hemolysis test and cytotoxicity assay (CCK-8) to verify the safety of the compound. Compound 4i showed a hemolysis rate less than 5% at its 2×, 4×, 6×, 8× and 10× MIC values (Table 2). This suggest that compound 4i has low hemolytic activity toward rabbit red blood cells. Compounds 4a–o also showed hemolysis rates of less than 5% at their 10× MIC values. However, the cytotoxicity assay showed that all compounds 4a–o were highly toxic to mouse fibroblast (L929) cells. At their MICs, the survival rate of L929 cells is equal to or lower than 80%. This may be due to the existence of the hydrazine structure, which leads to oxidative stress in cells.²⁷

Hemolysis rate of rabbit red blood cells treated with compound 4i at different concentrations.

Compound	2 MIC	4 MIC	6 MIC	8 MIC	10 MIC
4i	2.6%	2.1%	1.7%	1.3%	3.3%

EcUPPS inhibition assays

Based on the above test results of antibacterial activity, we selected compound 4i to determine its IC₅₀ for EcUPPS enzyme. The detection mechanism is shown in Scheme 2. The experimental results are shown in Fig. 5a. As the concentration of compound 4i increases, its inhibitory effect on EcUPPS also increases. The curve fitted by DoseResp shows that the IC₅₀ of compound 4i is about 145 μmol L⁻¹ (58 μg mL⁻¹). The experimental results show that compound 4i can inhibit the activity of EcUPPS enzyme to some extent. Subsequently, the enzyme activities of EcUPPS were detected under different concentrations of compound 4i (0, 32 and 64 μg mL⁻¹) and different farnesyl pyrophosphate (FPP) substrate concentrations (0.25–2 ng mL⁻¹), and the data were fitted by Michaelis–Menten equation to form a Lineweaver–Burk plot, and the results are shown in Fig. 5b. The K_m value of compound 4i with different concentrations on EcUPPS enzyme remained unchanged, while V_max decreased with the increase of inhibitor concentration, indicating that compound 4i was the noncompetitive inhibition of EcUPPS.

Scheme 2. Schematic diagram of detection mechanism.

Fig. 5. (a) Dose response curves for EcUPPS enzyme inhibition by compound 4i, (b) Lineweaver–Burk plots of compound 4i.

Molecular docking

To reveal the mechanism and molecular interaction involved, a computational study was performed to predict the binding posture of compound 4a–o with the target UPPS (PDB code: 4 h8e).²⁸ Compound 4i–o has a good spatial complementarity with the binding pocket, and the binding score was high (from −10.0 to −11.8 kcal mol⁻¹) (see Table S1†). The short molecular length of compound 4a–h leads to weak spatial complementarity and a low score (from −8.5 to −10.0 kcal mol⁻¹).

The score of compound 4i is −10.4 kcal mol⁻¹, and it has three hydrogen bonds, one π-stacking bond, and five hydrophobic interactions with the binding pocket. The guanidine group at the hydrophilic terminal of compound 4i forms hydrogen bonds with ASN81 of UPPS (see Fig. 6). The NH on the flexible part introduced in compound 4i forms hydrogen bonds with ASN35 and HIS50, which enhances the binding force between compound 4i and the target. In addition, because the bond of the flexible part rotates 126°, the hydrophobic terminal group in compound 4i is easily inserted into the hydrophobic part of the binding pocket, and the thiophene ring forms a π-stacking with PHE148, and forms hydrophobic interactions with MET54, ILE57, ALA76, PRO96, and ILE150. These strong binding interactions are consistent with the low MIC value of compound 4i. Although the docking/building score of compound 4l–n is higher than that of compound 4i, the interaction force is less than that of one or two hydrogen bonds because of the difference in the binding posture, which is consistent with the fact that their antibacterial abilities are slightly weaker than that of compound 4i.

Fig. 6. Docking diagram of UPPS (PDB code: 4 h8e) and compound 4i.

Physicochemical properties and ADME prediction

Compounds with drug-likeness are ideal drug candidates, and they may exhibit good bioavailability. Based on Lipinski's rule of 5, we analyzed the physicochemical properties of compounds 4a–o. As shown in Table 3, compounds 4e–o conform to the rule, and compounds 4a–d displayed only one violation of the rule, which they are deemed acceptable.²⁹ The analysis of octanol–water partition coefficient log P (clog P) and molecular dipole moment (Fig. S25†) indicates that the compounds exhibit good solubility. In fact, D₂O can be used as solvent to test their NMR spectra, which proves that their solubility is at least above 20 mg mL⁻¹. Molecular polar surface area (TPSA) and molecular volume analysis showed that all compounds exhibit good permeability.⁵ ADMET analysis showed that all compounds exhibit good absorption, distribution, and metabolism in human body. However, they can be rapidly cleared and there is a risk of ames mutagenesis (see Table S2†).

Calculation of the physicochemical properties of compounds 4a–o as per Lipinski's rule of 5, the topological polar surface area, and electric dipole.

Compound number	cLog P	MW	n ON	n OHNH	n rotb	TPSA (Å2)	Volume (Å3)	Electric dipole (Debye)
	≤5	≤500	≤10	≤5	≤10	≤150
4a	1.80	331	8	6	6	131.78	288.06	4.4297
4b	1.36	331	8	6	6	131.78	288.06	4.0835
4c	1.38	331	8	6	6	131.78	288.06	2.7749
4d	0.83	347	9	7	6	152.00	296.08	4.0858
4e	1.98	333	7	5	6	111.55	284.94	5.1801
4f	2.10	351	7	5	6	111.55	289.90	6.0435
4g	2.12	351	7	5	6	111.55	289.90	4.5945
4h	1.03	322	8	5	6	124.44	266.60	3.7774
4i	3.46	403	7	5	7	111.55	322.87	3.6849
4j	0.66	382	10	5	7	142.27	328.50	3.5375
4k	2.37	384	8	5	7	114.79	349.19	7.3491
4l	3.82	409	7	5	7	111.55	356.38	3.6243
4m	4.55	459	7	5	8	111.55	382.75	4.7959
4n	3.71	421	8	5	8	120.78	377.00	3.3697
4o	2.67	377	7	5	7	111.55	317.32	4.1120

Experimental

General chemistry

All reagents and solvents used were commercially available. Nuclear magnetic resonance (NMR) spectra were obtained using a Bruker Avance spectrometer (400 MHz). High-resolution mass spectra (HRMS) were recorded with a Bruker maXis impact instrument equipped with an electrospray ionization ion (ESI) source. Purity of final compounds was determined using one of the following methods. Method A (this method is used for all compounds, if not mentioned otherwise) on a Waters ACQuity Arc LC with Azalea C18 columns (250 × 4.6 mm, 5 μm) and detection was performed with a UV PDA at a 260 to 400 nm wavelength. Elution was carried out with the following gradient: water (solvent A), MeOH (solvent B), 20% B for 5 min, 20–40% B in 2 min, 40% B for 5 min, 40–20% B in 1 min, 20% B for 3 min, stop time 16 min, flow 0.8 mL min⁻¹. Method B on a Waters ACQuity Arc LC with Azalea C18 columns (250 × 4.6 mm, 5 μm) and detection was performed with a UV PDA at 260 to 400 nm wavelength. Elution was carried out with the following gradient: 20% B for 5 min, 20–60% B in 2 min, 60% B for 10 min, 60–20% B in 1 min, 20% B for 3 min, stop time 21 min, flow 0.8 mL min⁻¹.

All final compounds are ≥95% pure by HPLC analysis. A representative number of HPLC traces of compounds are included in the ESI,† Fig. S26 to Fig. S30.

General procedure of the synthesis of the thiosemicarbazones 2a–o

A solution of thiosemicarbazone (5 mmol) in ethanol (10 mL) was added dropwise to a solution of substituted aromatic aldehydes (5 mmol) in ethanol (20 mL). The reaction mixture was allowed to react at 70 °C for 6 h, and the precipitated solid was filtered, providing a yield of 90%.

General procedure of the synthesis of the intermediate compounds 3a–o

A solution of 3-chloroacetylacetone (2.2 mmol) in ethanol (10 mL) was added dropwise to a solution of compounds 2a–o (2 mmol) in ethanol (10 mL). The reaction mixture was allowed to react at 70 °C for 3 h, and the precipitated solid was filtered, giving a yield of 60–90%.

General procedure of the synthesis of the target compounds 4a–o

A solution of aminoguanidine hydrochloride (1.2 mmol) in ethanol (10 mL) was added dropwise to a solution of compounds 3a–o (1 mmol) and LiCl (0.36 mmol) in ethanol (40 mL). The reaction mixture was allowed to react at 70 °C for 24 h. The solvent was evaporated under reduced pressure. The crude product was purified by silica gel flash column using dichloromethane and methanol (5/1) as the elution solvent to obtain the pure products 4a–o. Yield: 40–70% (its ¹H NMR spectra are shown in Fig. S1–S15†).

(E)-2-(1-(2-(2-((E)-2-Hydroxybenzylidene)hydrazineyl)-4-methylthiazol-5-yl)ethylidene)hydrazine-1-carboximidamide (4a)

Yield: 60%. ¹H NMR (400 MHz, DMSO-d₆, δ): 11.40 ((s, 1H, –NH–), 10.81 (1H, –NH–, exchanged by D₂O), 10.45 (s, 1H, –OH), 8.40 (s, 1H, –CH N–), 7.70–7.61 (m, 4H, phenyl, –NH₂ and NH), 7.23 (t, J = 7.5 Hz, 1H, phenyl), 6.95 (d, J = 8.1 Hz, 1H, phenyl), 6.88 (t, J = 7.3 Hz, 1H, phenyl), 2.40 (s, 3H, –CH₃), 2.36 (s, 3H, –CH₃). (Fig. S1†) ¹³C NMR (150 MHz DMSO-d₆, δ): 166.33 (1C, thiazole), 156.50 (1C, benzene), 155.80 (2C, guanidyl and –N C<), 148.59 (1C, thiazole), 130.76 (1C, thiazole and –N CH–), 119.84 (1C, benzene), 119.37 (2C, benzene), 116.24 (2C, benzene), 16.98 (2C, methyl). (Fig. S16†) HRMS (ESI) m/z: [M + H]⁺ calcd for C₁₄H₁₈N₇OS, 332.1293; found 332.1300. HPLC λ (370 nm), t_ret = 13.062 min, purity > 95%.

(E)-2-(1-(2-(2-((E)-3-Hydroxybenzylidene)hydrazineyl)-4-methylthiazol-5-yl)ethylidene)hydrazine-1-carboximidamide (4b)

Yield: 60%. ¹H NMR (400 MHz, DMSO-d₆, δ): 10.95 (s, 1H, NH), 9.64 (s, 1H, OH), 8.00 (s, 1H, –CH N–), 7.50 (s, 4H, guanidyl), 7.23 (t, J = 7.6 Hz, 1H, phenyl), 7.11 (s, 1H, phenyl), 7.07 (d, J = 7.5 Hz, 1H, phenyl), 6.81 (d, J = 7.1 Hz, 1H, phenyl), 2.42 (s, 3H, CH₃), 2.34 (s, 3H, CH₃). (Fig. S2†) ¹³C NMR (150 MHz DMSO-d₆, δ): 166.19 (1C, thiazole), 157.73 (1C, benzene), 155.76 (2C, guanidyl and –N C<), 147.85 (1C, thiazole), 147.04 (1C, –N CH–), 134.85 (1C, thiazole), 129.83 (1C, benzene), 118.67 (1C, benzene), 118.41 (1C, benzene), 117.55 (1C, benzene), 114.01 (1C, benzene), 16.80 (2C, methyl). (Fig. S17†) HRMS (ESI) m/z: [M − H]⁻ calcd for C₁₄H₁₆N₇OS, 330.1137; found 330.1165. HPLC λ (360 nm), t_ret = 11.454 min, purity > 95%.

(E)-2-(1-(2-(2-((E)-4-Hydroxybenzylidene)hydrazineyl)-4-methylthiazol-5-yl)ethylidene)hydrazine-1-carboximidamide (4c)

Yield: 66%. ¹H NMR (400 MHz, DMSO-d₆, δ): 10.83 (s, 1H, NH), 9.90 (s, 1H, OH), 7.98 (s, 1H, –CH N–), 7.51 (d, J = 8.1 Hz, 2H, phenyl), 7.50 (s, 4H, guanidyl), 6.83 (d, J = 8.6 Hz, 2H, phenyl), 2.41 (s, 3H, CH₃), 2.33 (s, 3H, CH₃). (Fig. S3†) HRMS (ESI) m/z: [M − H]⁻ calcd for C₁₄H₁₆N₇OS, 330.1137; found 330.1151. HPLC λ (360 nm), t_ret = 11.263 min, purity > 95%.

(E)-2-(1-(2-(2-((E)-3,5-Dihydroxybenzylidene)hydrazineyl)-4-methylthiazol-5-yl)ethylidene)hydrazine-1-carboximidamide (4d)

Yield: 60%. ¹H NMR (400 MHz, DMSO-d₆, δ): 10.97 (s, 1H, NH), 9.44 (s, 2H, OH), 7.89 (s, 1H, –CH N–), 7.52 (s, 4H, guanidyl), 6.53 (s, 2H, phenyl), 6.27 (s, 1H, phenyl), 2.42 (s, 3H, CH₃), 2.34 (s, 3H, CH₃). (Fig. S4†) HRMS (ESI) m/z: [M − H]⁻ calcd for C₁₄H₁₆N₇O₂S, 346.1086; found 346.1095. HPLC λ (360 nm), t_ret = 10.983 min, purity > 95%.

(E)-2-(1-(2-(2-((E)-2-Fluorobenzylidene)hydrazineyl)-4-methylthiazol-5-yl)ethylidene)hydrazine-1-carboximidamide (4e)

Yield: 70%. ¹H NMR (400 MHz, DMSO-d₆, δ): 11.05 (s, 1H, –NH–), 10.81 (1H, –NH–, exchanged by D₂O), 8.28 (s, 1H, –CH N–), 7.89 (t, J = 7.6 Hz, 1H, phenyl), 7.55–7.44 (m, 4H, phenyl, –NH₂ and NH), 7.29 (t, J = 8.0 Hz, 2H, phenyl), 2.43 (s, 3H, –CH₃), 2.35 (s, 3H, –CH₃). (Fig. S5†) ¹³C NMR (150 MHz DMSO-d₆, δ): 166.28 (1C, thiazole), 161.34 (1C, benzene), 155.76 (2C, guanidyl and –N C<), 147.86 (1C, thiazole), 138.66 (1C, thiazole), 132.07 (1C, –N CH–), 125.99 (1C, benzene), 124.65 (1C, benzene), 119.06 (1C, benzene), 115.89 (1C, benzene), 115.76 (1C, benzene), 16.71 (2C, methyl). (Fig. S18†) HRMS (ESI) m/z: [M + H]⁺ calcd for C₁₄H₁₇FN₇S, 334.1250; found 334.1259. HPLC λ (360 nm), t_ret = 14.642 min, purity > 95%.

(E)-2-(1-(2-(2-((E)-2,3-Difluorobenzylidene)hydrazineyl)-4-methylthiazol-5-yl)ethylidene)hydrazine-1-carboximidamide (4f)

Yield: 40%. ¹H NMR (400 MHz, DMSO-d₆, δ): 11.04 (s, 1H, NH), 8.24 (s, 1H, –CH N–), 7.67 (t, J = 7.0 Hz, 1H, phenyl), 7.54 (s, 4H, guanidyl), 7.46 (d, J = 9.4 Hz, 1H, phenyl), 7.29 (d, J = 5.1 Hz, 1H, phenyl), 2.42 (s, 3H, CH₃), 2.35 (s, 3H, CH₃). (Fig. S6†) HRMS (ESI) m/z: [M + H]⁺ calcd for C₁₄H₁₆F₂N₇S, 352.1156; found 352.1165. HPLC λ (360 nm), t_ret = 14.880 min, purity > 95%, (method B).

(E)-2-(1-(2-(2-((E)-2,5-Difluorobenzylidene)hydrazineyl)-4-methylthiazol-5-yl)ethylidene)hydrazine-1-carboximidamide (4g)

Yield: 50%. ¹H NMR (400 MHz, DMSO-d₆, δ): 11.19 (s, 1H, NH), 8.23 (s, 1H, –CH N–), 7.59 (s, 5H, guanidyl and phenyl), 7.37–7.30 (m, 2H, phenyl), 2.43 (s, 3H, CH₃), 2.36 (s, 3H, CH₃). (Fig. S7†) HRMS (ESI) m/z: [M + H]⁺ calcd for C₁₄H₁₆F₂N₇S, 352.1156; found 352.1166. HPLC λ (360 nm), t_ret = 14.762 min, purity > 95%, (method B).

(E)-2-(1-(4-Methyl-2-(2-((E)-thiazol-5-ylmethylene)hydrazineyl)thiazol-5-yl)ethylidene)hydrazine-1-carboximidamide (4h)

Yield: 60%. ¹H NMR (400 MHz, DMSO-d₆, δ): 12.19 (s, 1H, –NH–), 10.98 (s, 1H, –NH–), 9.11 (s, 1H, thiazole), 8.36 (s, 1H, –CH N–), 8.18 (s, 1H, thiazole), 7.47 (s, 3H, –NH₂ and NH), 2.40 (s, 3H, –CH₃), 2.33 (s, 3H, –CH₃). (Fig. S8†) ¹³C NMR (150 MHz DMSO-d₆, δ): 166.36 (1C, thiazole), 155.86 (2C, guanidyl and –N C<), 155.40 (1C, thiazole), 148.66 (1C, thiazole), 144.08 (1C, thiazole), 139.51 (1C, –N CH–), 135.06 (1C, thiazole), 133.68 (1C, thiazole), 17.00 (2C, methyl). HRMS (ESI) m/z: [M + H]⁺ calcd for C₁₁H₁₅N₈S₂, 323.0861; found 323.0862. HPLC λ (370 nm), t_ret = 10.955 min, purity > 95%.

(E)-2-(1-(2-(2-((E)-[2,2′-Bithiophen]-5-ylmethylene)hydrazineyl)-4-methylthiazol-5-yl)ethylidene)hydrazine-1-carboximidamide (4i)

Yield: 60%. ¹H NMR (400 MHz, DMSO-d₆, δ): 10.99 (s, 1H, NH), 8.25 (s, 1H, –CH N–), 7.58 (d, J = 5.0 Hz, 1H, thienyl), 7.53 (s, 4H, guanidyl), 7.42 (s, 1H, thienyl), 7.34 (s, 1H, thienyl), 7.31 (s, 1H, thienyl), 7.13 (s, 1H, thienyl), 2.41 (s, 3H, CH₃), 2.34 (s, 3H, CH₃). (Fig. S9†) ¹³C NMR (150 MHz DMSO-d₆, δ): 165.12 (1C, thiazole), 155.62 (2C, guanidyl and –N C<), 147.45 (1C, thiazole), 141.82 (1C, –N CH–), 138.84 (1C, thiazole), 136.40 (1C, thiophene), 135.78 (1C, thiophene), 132.30 (1C, thiophene), 128.52 (1C, thiophene), 126.53 (1C, thiophene), 125.01 (1C, thiophene), 124.32 (1C, thiophene), 118.01 (1C, thiophene), 16.95 (2C, methyl). (Fig. S19†) HRMS (ESI) m/z: [M + H]⁺ calcd for C₁₆H₁₈N₇S₃, 404.0785; found 404.0780. HPLC λ (360 nm), t_ret = 11.890 min, purity > 95%.

(E)-2-(1-(2-(2-((E)-4-(1H-1,2,4-Triazol-1-yl)benzylidene)hydrazineyl)-4-methylthiazol-5-yl)ethylidene)hydrazine-1-carboximidamide (4j)

Yield: 40%. ¹H NMR (400 MHz, DMSO-d₆, δ): 11.17 (s, 1H, –C N NH–), 9.39 (s, 1H, triazole), 8.29 (s, 1H, triazole), 8.19 (s, 1H, –CH N–), 7.95 (d, J = 8.6 Hz, 2H, phenyl), 7.87 (d, J = 8.4 Hz, 2H, phenyl), 7.60 (s, 4H, guanidyl), 2.44 (s, 3H, –CH₃), 2.37 (s, 3H, –CH₃). (Fig. S10†) ¹³C NMR (150 MHz DMSO-d₆, δ): 168.98 (1C, thiazole), 155.63 (2C, guanidyl and –N C<), 152.95 (1C, thiazole), 152.50 (1C, triazole), 142.45 (2C, triazole and –N CH–), 137.34 (1C, benzene), 133.35 (1C, benzene), 132.09 (1C, thiazole), 128.06 (2C, benzene), 119.61 (2C, benzene), 16.75 (2C, methyl). (Fig. S20†) HRMS (ESI) m/z: [M + H]⁺ calcd for C₁₆H₁₉N₁₀S, 383.1514; found 383.1523. HPLC λ (360 nm), t_ret = 11.484 min, purity > 95%.

(E)-2-(1-(4-Methyl-2-(2-((E)-4-(pyrrolidin-1-yl)benzylidene)hydrazineyl)thiazol-5-yl)ethylidene)hydrazine-1-carboximidamide (4k)

Yield: 60%. ¹H NMR (400 MHz, DMSO-d₆, δ): 11.32 (s, 1H, –NH–), 10.81 (s, 1H, –NH–), 7.96 (s, 1H, –CH N–), 7.50–7.40 (m, 5H, phenyl, –NH₂ and =NH), 6.58 (d, J = 8.6 Hz, 2H, phenyl), 3.29 (s, 4H, pyrrolidine), 2.41 (s, 3H, –CH₃), 2.33 (s, 3H, –CH₃), 1.98 (s, 4H, pyrrolidine). (Fig. S11†) HRMS (ESI) m/z: [M + H]⁺ calcd for C₁₈H₂₅N₈S, 385.1922; found 385.1916. HPLC λ (350 nm), t_ret = 11.751 min, purity > 95%.

(E)-2-(1-(2-(2-((E)-(4′-Fluoro-[1,1′-biphenyl]-4-yl)methylene)hydrazineyl)-4-methylthiazol-5-yl)ethylidene)hydrazine-1-carboximidamide (4l)

Yield: 50%. ¹H NMR (400 MHz, DMSO-d₆, δ): 12.33 and 11.81 (1H, –NH–, exchanged by D₂O), 10.88 (s, 1H, –NH–), 8.13 (s, 1H, –CH N–), 7.79–7.73 (m, 6H, phenyl), 7.49 (s, 3H, –NH₂ and NH), 7.33 (t, J = 8.8 Hz, 2H, phenyl), 2.43 (s, 3H, –CH₃), 2.35 (s, 3H, –CH₃). (Fig. S12†) ¹³C NMR (150 MHz DMSO-d₆, δ): 162.94 (1C, thiazole), 161.31 (1C, benzene), 155.73 (2C, guanidyl and –N C<), 148.03 (1C, thiazole), 140.40 (1C, benzene), 135.76 (1C, thiazole), 132.91 (2C, benzene), 128.76 (1C, –N CH–), 128.71 (2C, benzene), 127.54 (2C, benzene), 126.98 (2C, benzene), 115.65 (1C, benzene), 115.51 (1C, benzene), 16.73 (2C, methyl). (Fig. S21†) HRMS (ESI) m/z: [M + H]⁺ calcd for C₂₀H₂₁FN₇S, 410.1563; found 410.1574. HPLC λ (300 nm), t_ret = 14.288 min, purity > 95%.

(E)-2-(1-(4-Methyl-2-(2-((E)-(4′-(trifluoromethyl)-[1,1′-biphenyl]-4-yl)methylene)hydrazineyl)thiazol-5-yl)ethylidene)hydrazine-1-carboximidamide (4m)

Yield: 50%. ¹H NMR (400 MHz, DMSO-d₆, δ): 12.33 and 11.81 (1H, –NH–), 10.89 (s, 1H, –NH–), 8.14 (s, 1H, –CH N–), 7.96 (d, J = 8.2 Hz, 2H, phenyl), 7.86–7.83 (m, 6H, phenyl), 7.49 (s, 3H, –NH₂ and NH), 2.43 (s, 3H, –CH₃), 2.35 (s, 3H, –CH₃). (Fig. S13†) ¹³C NMR (150 MHz DMSO-d₆, δ): 166.69 (1C, thiazole), 155.70 (2C, guanidyl and –N C<), 148.61 (1C, thiazole), 146.16 (1C, –N CH–), 143.37 (2C, benzene), 139.21 (2C, benzene), 134.43 (1C, thiazole), 127.40 (4C, benzene), 127.23 (2C, benzene), 125.84 (2C, benzene), 125.25 (1C, –CF₃), 16.89 (2C, methyl). (Fig. S22†) HRMS (ESI) m/z: [M + H]⁺ calcd for C₂₁H₂₁F₃N₇S, 460.1531; found 460.1547. HPLC λ (310 nm), t_ret = 11.247 min, purity > 95%.

(E)-2-(1-(2-(2-((E)-(4′-Methoxy-[1,1′-biphenyl]-4-yl)methylene)hydrazineyl)-4-methylthiazol-5-yl)ethylidene)hydrazine-1-carboximidamide (4n)

Yield: 60%. ¹H NMR (400 MHz, DMSO-d₆, δ): 12.33 and 11.81 (1H, –NH–, exchanged by D₂O), 10.68 (s, 1H, –NH–), 8.10 (s, 1H, –CH N–), 7.74–7.66 (m, 6H, phenyl), 7.43 (s, 3H, –NH₂ and NH), 7.05 (d, J = 8.6 Hz, 2H, phenyl), 3.82 (s, 3H, –CH₃), 2.43 (s, 3H, –CH₃), 2.33 (s, 3H, –CH₃). (Fig. S14†) ¹³C NMR (150 MHz DMSO-d₆, δ): 166.98 (1C, thiazole), 159.18 (1C, benzene), 155.60 (2C, guanidyl and –N C<), 149.05 (1C, thiazole), 143.81 (1C, –N CH–), 140.66 (1C, benzene), 132.58 (1C, thiazole), 131.55 (2C, benzene), 127.83 (2C, benzene), 127.27 (2C, benzene), 126.37 (2C, benzene), 114.58 (2C, benzene), 55.51 (1C, methyl), 17.96 (1C, methyl), 17.31 (1C, methyl). (Fig. S23†) HRMS (ESI) m/z: [M + H]⁺ calcd for C₂₁H₂₄N₇OS, 422.1763; found 422.1774. HPLC λ (310 nm), t_ret = 13.543 min, purity > 95%.

(E)-2-(1-(2-(2-((1E,2E)-3-(2,6-Difluorophenyl)allylidene)hydrazineyl)-4-methylthiazol-5-yl)ethylidene)hydrazine-1-carboximidamide (4o)

Yield: 60%. ¹H NMR (400 MHz, DMSO-d₆, δ): 11.93 (s, 1H, –NH–), 7.97 and 7.95 (s, 1H, –CH N–), 7.49–7.42 (m, 4H, phenyl, –NH₂ and NH), 7.18 (t, J = 8.7 Hz, 2H, phenyl), 6.96 (m, 2H, vinyl), 2.42 (s, 3H, –CH₃), 2.33 (s, 3H, –CH₃). (Fig. S15†) ¹³C NMR (150 MHz DMSO-d₆, δ): 166.04 (1C, thiazole), 161.05 (1C, benzene), 159.39 (1C, benzene), 155.67 (2C, guanidyl and –N C<), 147.99 (1C, thiazole), 131.19 (1C, –N CH–), 130.43 (2C, benzene and vinyl), 123.48 (1C, vinyl), 112.04 (1C, benzene), 111.87 (1C, benzene), 118.82 (1C, thiazole), 113.13 (1C, benzene), 16.61 (2C, methyl). (Fig. S24†) HRMS (ESI) m/z: [M + H]⁺ calcd for C₁₆H₁₈F₂N₇S, 378.1312; found 378.1314. HPLC λ (300 nm), t_ret = 10.959 min, purity > 95%.

Antibacterial activity assay

The MIC values of the compounds 4a–o against E. coli (ATCC 8739), S. aureus (ATCC 6538P), and MRSA (ATCC 33591) were determined using the standard broth microdilution method. Each bacterial strain was inoculated in a Mueller Hinton broth (MHB) medium. First, a 256 μg mL⁻¹ solution of the test compound in 1% DMSO was prepared as the mother solution, which was added into the first column of a 96-well plate (100 μL per well), and then serially diluted in the subsequent columns. Then, a 100 μL bacterial suspension (10⁵ CFU mL⁻¹) was added to each well. The final concentrations of the test compounds obtained were 128, 64, 32, 16, 8, 4, 2, 1, 0.5, 0.25, and 0.12 μg mL⁻¹. After culturing at 37 °C for 24 h, the OD₆₀₀ was measured using the microplate reader. The control group included the blank contrast and negative contrast, which contained 200 μL of media and 200 μL of bacterial solution, respectively. Norfloxacin was used as the positive control.

The minimum bactericidal concentration (MBC) of the test compounds was determined using 5 μL of cell suspension from wells (where the MIC was determined) onto respective Mueller Hinton agar (MHA) plates for 1× MIC, 2× MIC, 4× MIC, and 8× MIC ratios. Plates were incubated at 37 °C for 24 h before recording the MBC. After 24 h of incubation, the bacteria that did not form colonies were considered to have been killed.

Bactericidal time-kill assay

MRSA (∼10⁵ CFU mL⁻¹) in the logarithmic growth phase was exposed to compound 4i at 1×, 2×, and 4× MIC values. At specific time points (2, 4, 6, 8, and 24 h), 10 μL sample was diluted to the appropriate concentration, and then, 50 μL diluted solution was coated on the MHA plates. The plates were incubated for about 24 h at 37 °C, and the bacterial colonies on the plates were counted to calculate the bacterial concentration before dilution. The results were plotted to obtain the logarithmic growth curve.

Resistance development assay

MRSA was taken from the 0.5× MIC well and incubated for 24 h at 37 °C. Then, it's the new MIC value and 0.5× MIC value of compound 4i was measured. Next, the bacteria from the new 0.5× MIC well were incubated for 24 h at 37 °C, and then, its new MIC value and 0.5× MIC value were measured. This process was repeated for 10 generations. The MIC values of compound 4i against each generation of MRSA were determined.

Biofilm inhibition assay

As previously described,³⁰ biofilm formation was performed in a 96-well tissue culture treated plate (Corning #3603). MRSA was grown in nutrient broth (1.0% peptone, 0.3% beef extract powder, and 0.5% sodium chloride; pH = 7.4 ± 0.2). Briefly, MRSA (10⁷ CFU mL⁻¹, 100 μL) was inoculated into wells containing different concentrations of compound 4i (100 μL) and incubated for 24 h. The control group did not contain compound 4i. After incubation, the non-adhering contents (planktonic cells) were removed, and the wells were washed twice with sterile water. The biofilm was stained with 0.1% (w/v) crystal violet in water. After 15 min, excess dye was removed, the wells were washed twice with sterile water, and crystal violet was dissolved in 95% ethanol in water (200 μL). Then, the contents of the wells (100 μL) were transferred to a new 96-well clear bottom plate, and the absorbance (OD₅₉₅) was measured at 595 nm. Each experiment was repeated three times. The percentage content of biofilm was calculated using the following formula:%Biofilm viability = (Test OD₅₉₅/Control OD₅₉₅) × 100

Biofilm eradication assay

MRSA was grown in nutrient broth (1.0% peptone, 0.3% beef extract powder, and 0.5% sodium chloride; pH = 7.4 ± 0.2). The 96 well tissue culture treated plate (Corning #3603) containing bacterial suspension (10⁶ CFU mL⁻¹, 100 μL) was incubated under stationary conditions for approximately 24 h. After incubation, the non-adherent contents (planktonic cells) were removed, and the wells were washed once with physiological saline. Then, compound 4i with different concentrations was added into the wells containing the preformed bacterial biofilm and incubated for 24 h at 37 °C. Next, 100 μL culture medium was added as the control. After 24 h, the planktonic cells were removed and the wells were washed twice with sterile water. The biofilm was stained with 0.1% (w/v) crystal violet in water. After 15 min, excess dye was removed, the well was rinsed twice with sterile water, and crystal violet was dissolved in 95% ethanol in water (200 μL). The contents of the wells (100 μL) were transferred to a new 96 well clear bottom plate. The absorbance (OD₅₉₅) was measured at 595 nm. Each experiment was repeated three times. The percentage eradication of biofilm was calculated using the following formula:%Biofilm eradication = ((Control OD₅₉₅ − Test OD₅₉₅)/Control OD₅₉₅) × 100

Cytotoxicity and hemolysis assays

The mouse fibroblast L929 cells (Fenghbio Company, Changsha, Hunan, China) were selected for cytotoxicity assays. Above all, L929 cells were first resurrected and subcultured using DMEM to a concentration of 10⁵ cells per mL. These cells were added to a 96-well plate (100 μL per well) and cultured for 24 h at 37 °C with 5% CO₂. Then, 10 μL of various concentrations of the test compound (in triplicate; 1280, 640, 320, 160, 80, 40, 20, 10, 5, 2.5 μg mL⁻¹) was added to respective wells, and the plate was incubated for 24 h, followed by addition of 10 μL of CCK-8 solution to each well and 2 h incubation. Absorbance was measured using a microplate reader at 450 nm. The cell viability is expressed as a percentage of DMSO-treated control cell viability (average of triplicate wells ± standard deviation).

An in vitro hemolysis assay was performed with rabbit red blood cells (Hqbio Company, Guangzhou, Guangdong, China) as the subject. The 2% red blood cell suspension (0.625 mL) and normal saline solution (0.5 mL) were mixed with the tested sample solution (0.125 mL) at concentrations of 10× MIC, 8× MIC, 6× MIC, 4× MIC, and 2× MIC, and transferred into 2 mL centrifuge tubes, representing the test group. Additionally, the positive control group was made of 2% red blood cell suspension (0.625 mL) and distilled water (0.625 mL), whereas the negative control consisted of 2% red blood cell suspension (0.625 mL) and normal saline (0.625 mL). All samples were slightly shaken for mixing prior to incubation at 37 °C for 3 h. The supernatant of miscible liquid was collected to determine OD₅₄₅ with an ultraviolet spectrophotometer, the data were recorded, and the hemolysis rates were calculated formulas follows:Hemolysis rate = (OD_{test group} − OD_{negative control})/(OD_{positive control} − OD_{negative control}) × 100%.A value of >5% indicated hemolysis.

All experiments were carried out with the approval of Chinese government authorities.

EcUPPS inhibition assays

PET-28(a+) plasmid was cut with restriction enzymes QuickCut NcoI (TAKARA 1620) and QuickCut XhoI (TAKARA 1635). The target gene EcUPPS was amplified from E. coli (ATCC 8739) by fast PCR polymerase PrimeStar Max DNA Polymerase (Takara R045). Upstream primer was 5′-ttaagaaggagatataccatgttgtctgctactcaaccact-3′ (NcoI). Downstream primer was 5′-tggtggtggtggtgctcgagggctgtttcatcaccgggct-3′ (XhoI). ClonExpress Mix (VAZYME C115) was used to recombine the amplified target gene into the cut pET-28(a+) plasmid. The recombinant plasmid was transformed into E. coli BL21(DE3) chemically competent cell (VAZYME C504), and the transformed bacteria were screened with 30 μg mL⁻¹ kanamycin. The single colony successfully transformed was inoculated in 10 mL of LB liquid medium containing 30 μg mL⁻¹ kanamycin, and cultured at 37 °C and 180 r min⁻¹ for 6 hours. 1 mL was taken out for later use, and the remaining 9 mL was cultured at 16 °C and 180 r min⁻¹ with the final concentration of 1 mmol L⁻¹ IPTG induction. Take 1 mL of induced bacterial solution and use SDS-PAGE electrophoresis to verify whether the target protein EcUPPS is expressed. For the expressed strains, the strains were preserved and verified by sequencing with 1 mL of standby bacterial liquid set aside before. Inorganic pyrophosphatase (EcPPA) was prepared by the same method. Upstream primer was 5′- TTAAGAAGGAGATATACCATGagcttactcaacgtccctgc-3′ (NcoI). Downstream primer was 5′- TGGTGGTGGTGGTGCTCGAGtttattctttgcgcgctcga-3′ (XhoI).

Inoculate the successfully verified strain into 300 mL LB liquid medium containing 30 μg mL⁻¹ kanamycin, and cultivate it for 6 hours at 37 °C and 180 r min⁻¹. After the strain grows well, add IPTG with the final concentration of 1 mmol L⁻¹, and induce it for 16 hours at 16 °C and 180 r min⁻¹. The culture medium was centrifuged at 4 °C and 8000 × g for 3 min. The cells collected at low temperature was resuspended in lysate A (50 mmol L⁻¹ Tris-HCl, pH 8.0, 300 mmol L⁻¹ NaCl, 0.1% Triton X-100 and 0.01% defoamer), and the cells were broken by ultrasonic crusher. The lysate was centrifuged at 4 °C and 8000 × g for 3 minutes, and the supernatant was collected. The target protein in the supernatant was separated and purified by 1 ml Ni-NTA 6FF column. Impurities (10 mmol L⁻¹ Tris-HCl, pH 8.0, 300 mmol L⁻¹ NaCl, 50 mmol L⁻¹ imidazole and 10% glycerol) were removed by pre-washing solution B, and the target protein (10 mmol L⁻¹ Tris-HCl, pH 8.0, 20 mmol L⁻¹ KCl, 100 mmol L⁻¹ NaCl, 500 mol L⁻¹ imidazole and 10% glycerol) was eluted by eluent C, collect the target protein, and collect one tube every 500 μL (judging by the color development of Modified Bradford reagent (SANGON C100530), from color change to no color change, it means that the collection is completed, and usually 6 tubes of protein can be collected). The concentration of EcUPPS was obtained by measuring the absorbance at 595 nm with Modified Bradford reagent (SANGON C100530) (using BSA standard solution (SANGON C506031)). The purity was analyzed by SDS-PAGE electrophoresis.

Compound 4i was prepared as 10.24 mg mL⁻¹ stock solution in DMSO, and the stock solution was continuously diluted from 128 μg mL⁻¹ to 1 μg mL⁻¹ in 96-well plates with HEPES buffer (100 mM HEPES, 50 mM NaCl, 0.5 mM MgCl₂, 0.02% DDM (w/v), pH 7.5), with 100 μL per well. Add 8 μL substrate farnesyl pyrophosphate (FPP, 0.025 mg mL⁻¹), 6 μL isopentenyl pyrophosphate (IPP, 0.25 mg mL⁻¹) and 1 μL EcUPPS (0.2 mg mL⁻¹) to each well, and incubate at 37 °C for 30 min, then add 1 μL EcPPA (0.2 mg mL⁻¹), and continue to incubate at 37 °C for 30 min. After the reaction, 84 μL of water and 70 μL of malachite green phosphate determination reagent (Beyotime S0196) were added to each well, and incubated at 37 °C for 30 minutes. The absorbance at 630 nm was measured to calculate the released phosphate. Calculate IC₅₀ by using the DoseResp fitting method in OriginPro 2018. DMSO was the blank control. The release amount of phosphate is obtained from the phosphate standard curve. The activity data of EcUPPS enzyme were detected at different substrate FPP concentrations (0, 0.22, 0.29, 0.43 and 0.87 μmol L⁻¹) and the final concentration of compound 4i was 0, 64 and 128 μg mL⁻¹. The data were fitted by Michaelis–Menten equation to form a Lineweaver–Burk plot.

Molecular docking

Docking simulations were performed using AutoDock Vina software. The structures of compounds 4a–o were drawn by Chem3D software. Energy minimization was performed using the MM2 force field. The crystal structure of S. aureus undecaprenyl diphosphate synthase (UPPS) was downloaded from the protein data bank (PDB code: 4 h8e). The protein was pretreated by removing small molecules other than protein and adding hydrogen atoms. The binding pocket was designated near the original substrate farnesyl diphosphate. The flexible ligand-rigid receptor docking method was used to perform docking, and the binding energy was scored. The top-scoring pose was extracted, and the interaction force was analyzed and drawn using PyMOL software.

Conclusions

In the field of antibacterial drug development, small molecules can be improved to be used against drug-resistant bacteria and Gram-negative bacteria. In this study, thiazole aminoguanidines were developed as antibacterial compounds by increasing the molecular flexibility and they could simultaneously work against both drug-resistant bacteria and Gram-negative bacteria. With the increased molecular flexibility, thiazole aminoguanidine compounds can pass through the outer membrane of E. coli, and the hydrophobic group at the C2 terminal of thiazole can rotate at a certain angle so that it can be inserted into the binding pocket of the target UPPS. This modification enhances the antibacterial ability of the compounds. The prepared thiazole aminoguanidines can inhibit and destroy biofilms and exhibit superior antibacterial effect, which is beneficial to rapid sterilization and indicate a low tendency of bacterial resistance development. The low synthesis cost of these thiazole aminoguanidines highlights these analogs as potential therapeutic candidates for future development.

Author contributions

Conceptualization, funding acquisition, supervision, validation, and writing of the original draft: Ping Yang. Data curation, formal analysis, investigation, methodology, and visualization: Hui-Zhong Liu. Software: Ying-Si Wang. Formal analysis: Hong Qi, Ling-Ling Wang, Bei-Bei Wang. Writing – review and editing, Project administration and resources: Xiao-Bao Xie.

Conflicts of interest

There are no conflicts to declare.