English

• Acinetobacter baumannii (A. baumannii) is a nosocomial infection pathogen that predominantly infects the most vulnerable hospitalized patients [1, 2]. They survive in adverse environmental conditions and spread in the hospital settings, especially in intensive care units (ICUs) [3]. Mortality caused by A. baumannii was reported to be 26% in hospital settings and 43% in ICUs [1]. In addition, most clinical A. baumannii isolates have a common ability of natural transformation [4]. They show resistance to all clinical antibiotics used currently and exhibit a wide array of drug resistance mechanisms [5, 6]. Due to its abilities of rapidly acquiring drug resistance and resisting all commonly prescribed antibacterial drugs, carbapenem-resistant A. baumannii is classified at a level of "urgent" pathogen that causes the greatest threat to human health by World Health Organization (WHO) [7].

  Currently, owing to the shortage of new antibiotics, there has been an increase in the use of colistin as the last-line therapy to treat infections caused by MDR Gram-negative bacteria [8]. Colistin, a cationic cyclic lipopeptides produced by the soil bacterium Bacillus polymyxa [9], was introduced into the clinic in the 1950s, however, its state of clinical use rapidly waned in the 1970s since it would result in severe renal and neurological toxicity [10]. Furthermore, numerous publications have suggested that the risk of renal toxicity was related to the total doze of colistin [11-14]. In light of this, it is necessary to find new strategies to decrease the required dosage of colistin in order to extend its use.

  The combination of antibiotics has been explored [15-18]. Specifically, many research works considered colistin as a membrane-active adjuvant that sensitizes other antibiotics for the better antimicrobial activity [19-22]. Nevertheless, the approach is available only for a limited number of effective antibiotics. As an alternative strategy, antibiotic adjuvants, which are compounds with low antimicrobial activity, can be utilized to inhibit the development of resistance and enhance the effects of antibiotics. Several research efforts have been devoted to discovering novel adjuvants that enhance the potency of colistin and ultimately allow for reducing dosing [23-25]. Herein, our group focused on the construction of a diversity-oriented 2-aminothiazolyl compound library and the discovery of potent antibiotic adjuvants against multidrug-resistant Gram-negative bacteria [26, 27]. Currently, we are interested in discovering small molecule adjuvants that can potentiate the activity of colistin against A. baumannii. Toward this end, the diversity-oriented 2-aminothiazolyl compound library was screened. Fortunately, compound 1, a 2-aminothiazopiperidine derivative, was found and exhibited effective synergism with colistin (Fig. 1A).

  Figure 1

  

  Figure 1.  (A) Library screening for colistin potentiation. (B) Structural optimization for hit 1.

  DownLoad: Full-Size Img PowerPoint

  The small molecule itself has no bactericidal effect on ATCC19606 [minimum inhibitory concentration (MIC) of 128 µg/mL]; however, it reduced the MIC of colistin by 8-fold at 32 µg/mL, from 0.5 µg/mL to 0.0625 µg/mL, while at 2 µg/mL, it decreased the MIC of colistin by 4-fold. According to the result, we identified compound 1 as a potential hit and organized further structural optimization. The structure of compound 1 was divided into four parts as shown in Fig. 1B: part 1, the pyrazolyl head (highlighted blue); part 2, the linker group (highlighted orange); part 3, the 2-aminothiazopiperidine core (highlighted red) and part 4, the aryl tail (highlighted purple).

  Initially, we focused on the pyrazolyl head in the structure of compound 1 and synthesized the first-generation library, which was rapidly assembled by using the approach shown in Scheme 1.

  Scheme 1

  

  Scheme 1.  General synthetic route for first-generation library: (i) KSCN, acetone, r.t. for 1 h. Yield: 80%. (ii) 4-Trifluoromethyl-aniline, EA, reflux for 1 h at 80 ℃. Yield: 98%. (iii) 2 mol/L NaOH, EtOH, reflux for 1 h at 80 ℃. Yield: 95%. (iv) tert-Butyl 4-(2-bromoacetyl)piperidine-1-carboxylate. TEA, EtOH, reflux for 1 h at 80 ℃. Yield: 54%. (v) 4 mol/L HCl/dioxane, DCM, r.t. for 30 min. (vi) Commercially available carboxylic acid, HATU, DIPEA, DCM, r.t. for 24 h. Yield: 32%–63%.

  DownLoad: Full-Size Img PowerPoint

  Benzoyl isothiocyanate 3 was prepared by reacting benzoyl chloride 2 with potassium thiocyanate at room temperature; and then by reacting benzoyl isothiocyanate 3 with 4-trifluoromethylaniline under reflux condition to obtain compounds 4. The thiourea 5 was obtained by the debenzoylation of compounds 4. The resulting thiourea was dissolved in ethanol and reacted with tert-butyl 4-(2-bromoacetyl)piperidine-1-carboxylate under reflux conditions affording intermediates 6 that was served as the key building block. To complete the construction of the library, the Boc group was deprotected by 4 mol/L HCl in dioxane and then the generated products were reacted with commercially available carboxylic acid derivatives to obtain 2-aminothiazoyl piperidine derivatives 7a-7t.

  After that, the ability of each analogue to enhance colistin efficacy against ATCC19606 was investigated. The standalone toxicity against ATCC19606 of each analogue of the first-generation library was tested individually by determining the MIC at first. Among the 20 analogs, only three analogs exhibited MIC values of 64 µg/mL (7b, 7n and 7o), while the MIC values of the rest of the analogs were higher than 128 µg/mL. Next, the MIC of colistin against ATCC19606 was determined in presence of each compound at 32 µg/mL, 8 µg/mL and 2 µg/mL. The results were shown in Fig. 2 and Table S1 (Supporting information).

  Figure 2

  

  Figure 2.  First-generation library 7a-7t potentiation of colistin against ATCC19606. Synergistic effect on colistin at the concentration of (A) 32 µg/mL, (B) 8 µg/mL, (C) 2 µg/mL, (D) FICI of the first-generation library.

  DownLoad: Full-Size Img PowerPoint

  Compound 7a bearing non-substituted phenyl and 7b-7k bearing substituted phenyl presented a comparison between the effects of replacing pyrazolyl and aromatic ring. Most of these compounds exhibited a decrease in synergistic activity by comparing with compound 1. Subsequently, compounds 7l-7o were obtained by replacing pyrazolyl with non-aromatic groups. When the concentrations of three compounds (7l, 7n and 7o) were at 32 µg/mL, the MIC of colistin reduced more than 16-fold. However, compared with compound 1, these compounds displayed a diminishment in adjuvant activity at 8 µg/mL and 2 µg/mL, which indicated that the pyrazole ring was significant for synergistic activity. Next, the influence of substituent on pyrazolyl ring of compound 1 was investigated (7p-7t). The synergistic activity of compound 7t, which contained a 3-ethyl-4-chlorosubstituted pyrazole ring, was the best among these analogs, returning the colistin MIC value to 0.0156 µg/mL (32-fold reduction) at 32 µg/mL and 0.0625 µg/mL (8-fold reduction) at 8 µg/mL.

  Fractional inhibitory concentration index (FICI), which could be used to evaluate the degree of potentiation when two inhibitory substances combined [28], was selected to further verify the synergistic effect of the compound 7a-7t with colistin. The synergistic activities of all compounds in the first-generation library were shown in Fig. 2. Although the FICI values of the compound 7j and the compound 7t were relatively low, the compound 7t had a better synergistic effect at the concentration of 32 µg/mL, so we chose the compound 7t for further structural modification.

  The second-generation library was constructed based on the structure of compound 7t. The 3-ethyl-4-chlorosubstituted pyrazole ring was fixed while various functional groups were introduced into the aryl tail. The compounds of second-generation library were synthesized by using the synthetic route outlined in Scheme 2. The reaction of benzoyl isothiocyanate 3 with commercially available corresponding aniline derivatives followed by debenzoylation afforded the thioureas 9. The intermediates reacted with tert-butyl 4-(2-bromoacetyl)piperidine-1-carboxylate to afford intermediates 10. Hydrochloric acid mediated deprotection followed by condensation with 4-chloro-3-ethyl-1-methyl-1H-pyrazole-5-carboxylic acid delivered compounds 11a-11t.

  Scheme 2

  

  Scheme 2.  General synthetic route for second-generation library: (i) KSCN, acetone, r.t. for 1 h. Yield: 80%. (ii) Commercially available corresponding aniline, EA, reflux for 1 h at 80 ℃. Yield: 98%. (iii) 2 mol/L NaOH, EtOH, reflux for 1 h at 80 ℃. Yield: 95%. (iv) tert-Butyl 4-(2-bromoacetyl)piperidine-1-carboxylate, TEA, EtOH, reflux for 1 h at 80 ℃. Yield: 54%. (v) 4 mol/L HCl/dioxane, DCM, r.t. for 30 min. (vi) 4-Chloro-3-ethyl-1-methyl-1H-pyrazole-5-carboxylic acid, HATU, DIPEA, DCM, r.t. for 24 h. Yield: 32%–65%.

  DownLoad: Full-Size Img PowerPoint

  The standalone toxicity against ATCC19606 of compounds 11a-11t was also tested individually by determining the MIC. Among the 20 compounds, only two compounds retained a MIC of 64 µg/mL (11b and 11s), while the MICs of the rest of the compounds were higher than 128 µg/mL. And then the MIC of colistin against ATCC19606 was measured in the presence of each analogue at the same concentration as the first library. The results were shown in Fig. 3 and Table S2 (Supporting information).

  Figure 3

  

  Figure 3.  Second-generation library 11a-11t potentiation of colistin against ATCC19606. Synergistic effect on colistin at the concentration of (A) 32 µg/mL, (B) 8 µg/mL, (C) 2 µg/mL, (D) FICI of the second-generation library.

  DownLoad: Full-Size Img PowerPoint

  Compound 11a, without substituent on phenyl ring, exhibited worse synergistic activity compared with compound 7t (4-CF₃). Compounds 11b-11g, which incorporated a 4-monosubstiution pattern, allowed a comparison of the effects of replacing -CF₃ with -Me, -F, -Cl, -OMe, -CN and -acetyl. Compound 11b (4-Me) and Compound 11f (4-CN) only exhibited 8-fold reduction in the MIC of colistin at 32 µg/mL, while compounds 11d (4-Cl) and 11g (4-acetyl) also exhibited moderate activity, returning the colistin MIC to 0.0313 µg/mL at 32 µg/mL (16-fold reduction). Compound 11c (4-F) and compound 11e (4-OMe) were more active and exhibited 32-fold reduction, which were similar to that of compound 7t. The 2-chloro and 3-chloro analogues 11i and 11k exhibited moderate adjuvant activity, decreasing colistin MIC 16-fold at 32 µg/mL. Moreover, both analogues 11h (2-CF₃) and 11j (3-CF₃) reduced colistin MIC against ATCC19606 by 32-fold at 32 µg/mL. Especially, compound 11j bearing 3-trifluoromethyl exhibited the highest activity, lowering the colistin MIC by 16-fold at a concentration of 8 µg/mL. The results suggested that the presence of a trifluoromethyl substituent imparted better adjuvant activity to the 2-aminothiazoyl piperidine scaffold. Since 3-trifluoromethyl appeared to be a promising substituent for adjuvant activity, it was fixed and the double substitution pattern (11l-11r) was evaluated subsequently. When the dose was respectively reduced to 8 µg/mL, all of these compounds exhibited a reduction in synergistic activity compared with compound 11j. Therefore, it can be seen from this that the trifluoromethyl-containing double substitution pattern had a negative effect on the adjuvant activity.

  In addition, the synergistic activities of pyridine analogues 11s and 11t were also evaluated. Compound 11s exhibited improved adjuvant activity at 32 µg/mL compared with 7t, reducing colistin MIC against ATCC19606 by 64-fold. However, at 8 µg/mL, 11s displayed slightly lower synergistic activity compared with 11j. Compound 11t exhibited worse adjuvant activity at 32 µg/mL (16-fold reduction). In general, compound 11j, which contained a 3-trifluoromethyl on phenyl ring, was still the most active compound among compounds 11a-11t.

  We comprehensively analyzed the synergistic effect of the second-generation library, and the corresponding results were shown in Fig. 3. At the concentrations of 2 µg/mL and 8 µg/mL, the compound 11j had the best synergistic activity; compared with the compound 11s, although the compound 11j had a slightly lower synergistic effect at the concentration of 32 µg/mL, its FICI value was also lower. Therefore, we then chose the compound 11j to continue research (Figs. 3A-C). Meanwhile, we found that the compounds of the second-generation library had evident synergistic activities, except compounds 11f, 11l, 11q (FICI > 0.5), which further proved that the skeleton of these compounds had the potential to explore synergistic activity.

  In order to explore the effect of compound 11j on colistin activity over time, time-kill curves for combination of compound 11j and colistin against ATCC19606 were constructed. A. baumannii strain ATCC19606 was cultivated in the presence of compound 11j, colistin, or the combination of the two, and samples were taken and plated at 0, 2, 4, 8, 12 and 24 h time points (Fig. 4A).

  Figure 4

  

  Figure 4.  (A) Time-kill curves for combination of 11j and colistin against ATCC19606. (B) The schedule shows the experimental design. ATCC19606 were incubated probes at 37 ℃ for 30 min. Four aliquots of cultures were then treated with 11j (4 µg/mL), colistin (0.5 µg/mL) and both together. (C) Dynamic curves of the permeability of the outer membrane probed with NPN.

  DownLoad: Full-Size Img PowerPoint

  Colistin was used at its active concentration (0.5 µg/mL) and the control was tested at the same condition without drug. Bacterial growth was inhibited at the 4 h time point by colistin alone at 0.5 µg/mL, but bacterial regrowth of ATCC19606 was observed within 4–24 h and the CFU (colony forming unit) approached a similar amount to the control at the 24 h time point. Meanwhile, bacterial growth of ATCC19606 was unaffected when treated with 4 µg/mL compound 11j alone, and its growth curve was similar to the control. When the bacteria were cultivated in the presence of colistin and different detected concentrations of compound 11j (1 µg/mL, 2 µg/mL and 4 µg/mL, respectively), a similar trend to that of colistin alone for bacteriostatic growth inhibition at early time points (up to 4 h) was exhibited. However, 4 h later, bacterial growth of ATCC19606 was inhibited. When 1 µg/mL and 2 µg/mL compound 11j were combined with 0.5 µg/mL of colistin, a complete killing was observed at the 24 h and 12 h time points, respectively. Further, as the dose of compound 11j was added up to 4 µg/mL, the effect of 11j in combination with colistin on bacterial growth was the most evident; also, a complete killing of bacteria within 8 h was displayed and no regrowth was observed at the 24 h time point. Based on the above, the conclusion that compound 11j was able to potentiate colistin against a strain of A. baumannii was drawn.

  After the synergistic activity of synthesized compounds was verified, we preliminarily probed into the action mechanism of 11j. Outer membrane is the main target of colistin and the major barrier of antibiotic uptake, so we examined the disruption of bacterial outer membrane as a possible mechanism. Compared with the simultaneous addition of compound 11j and colistin, bacterial viability was tested through the sequential addition of sole compound 11j and colistin (Fig. 4B). When the mixture of colistin and compound 11j was added to the probe-labeled bacteria, we found that the fluorescence value increased rapidly and was higher than that of the colistin and compound 11j alone group. When compound 11j was added to the bacteria pretreated with colistin, it could be found that the fluorescence value increased rapidly, which indicated that the drug caused serious damage to the outer membrane of bacteria. Nevertheless, when colistin was added to the bacteria pretreated with 11j, the fluorescence value did not change significantly. It suggested that compound 11j can assist colistin to destroy the outer membrane of bacteria (Fig. 4C).

  Predicting the absorption, distribution, metabolism, excretion, and toxicity (ADMET) of candidate drugs has become an indispensable part of the drug discovery process. Therefore, the discovery studio software was used to predict the pharmacokinetics and the toxicity of the synthesized compounds (Tables S3-S5 in Supporting information). Furthermore, we found that all the synthesized compounds showed the satisfactory results on CYP2D6 liver, which indicated that the candidate drug were non-inhibitors of CYP2D6 and the synthesized compounds were well metabolized in Phase-I metabolism. In light of the toxicity prediction analysis, the second generation of 2-aminothiazoyl piperidine was less toxic overall.

  As these compounds demonstrated powerful synergistic potentials and promising applications in developing adjuvant, 3D-QSAR model employing comparative molecular similarity index analysis (CoMSIA) was established to expound the structure−activity relationship and provide useful information for the future design of the molecular structure [29, 30]. The CoMSIA analysis was performed with SYBYL X 2.1.1. Thirty target molecules were selected as the training set for CoMSIA, while the left molecules were employed as the testing set (Table S6 in Supporting information). We firstly assembled the 3D structures of all molecules by Chem 3D and then optimized their conformation with the "Minizine" function employing Tripos force field. Next, compound 11j with the greatest synergistic activity was selected as the template and all of these molecules were aligned. The alignment pattern was shown in Fig. 5A. Relevant model parameters and details were shown in the supplementary file.

  Figure 5

  

  Figure 5.  (A) Molecular alignment of all target compounds. (B) Steric fields, with green and yellow polyhedra indicating the regions where steric bulk would enhance and reduce the activity. (C) Electrostatic fields, with blue and red polyhedra indicating the regions where positive and negative charges would enhance activity. (D) Hydrophobic fields, with yellow and gray polyhedra indicating the regions where hydrophobicity and hydrophilicity would enhance activity. (E) H-bond donor fields, with cyan contours indicating the regions where H-bond donor groups would enhance activity. (F) H-bond acceptor fields, with red contours indicating the regions where H-bond acceptor groups would enhance activity.

  DownLoad: Full-Size Img PowerPoint

  The fabricated contour maps of the CoMSIA model are presented in Fig. 5. The steric field pattern (Fig. 5B) indicated that the bulky groups at the methylene located in the yellow region were unfavorable for the synergistic activity. By contrast, the green area suggested that the introduction of bulky groups at the 4- or 5-position of the benzene ring would improve the bioactivity. For the electrostatic field map (Fig. 5C), the blue area suggested that the introduction of electron-withdrawing groups enhanced the bioactivity, which verified that the chlorine at the 5-position of the pyrazole ring played a part in enhancing the biological activity of compound 11j (5-Cl, 0.125). On the contrary, the red portion located in the 4-position of benzene ring implied that an electron-donating group was preferred to the bioactivity, which was verified from the bioassay result of compounds 11e (4-OCH₃, 0.1875), 11d (4-Cl, 0.3125) and 11f (4-CN, 0.5625). For the CoMSIA hydrophobic field (Fig. 5D), the yellow pattern with hydrophobic groups at this area was found to enhance the bioactivity by observing the yellow color surrounding the methyl group. Thus, this group was indeed devoted to the bioactivity. By contrast, the gray area showed that the introduction of hydrophilic group would enhance the bioactivity. From the cyan pattern drawn by observing the H-bond donor field map (Fig. 5E), it can be seen that providing hydrogen bonds in this region could increase the activity. The H-bond acceptor field pattern (Fig. 5F) represented that the reception of hydrogen bonding in the red region could ameliorate the synergistic efficiency. In view of the foregoing, the COMSIA model provides an insight into the rational design of fresh bioactive compounds as adjuvants.

  According to the frontier molecular orbital theory, the HOMO and LUMO play crucial roles in the electric properties and determine the way that the molecule interacts with other species. The HOMO represents the ability to donate an electron, whereas LUMO represents the ability to obtain an electron [31, 32]. Energy gap between HOMO and LUMO gives an idea about the kinetic stability and conductivity as well. A density functional theory (DFT) study was carried out by using the DFT-B3LYP/6-31G(d, p) method. The DFT-derived calculation of E_HOMO, E_LUMO, and energy gap (ΔE) between HOMO and LUMO was calculated by the Multiwfn program. The energies of HOMO and LUMO and the value of ΔE for the selected target compounds 11j, 7l and 7t were shown in Fig. 6. The positive molecular orbitals were symbolized with red, and the negative molecular orbitals were symbolized with green. The HOMO and LUMO orbitals are mostly the π-antibonding type molecular orbitals of the compounds. The energy gaps between the HOMO and LUMO of 11j, 7l and 7t are 4.7808, 4.7654, and 4.7479 eV, respectively. Compared with 7l and 7t, 11j has a higher energy gap but lower LUMO energy. From the point of view of the DFT calculation, compound 11j may be more stable than compounds 7l and 7t.

  Figure 6

  

  Figure 6.  Frontier molecular orbitals of selected target compounds 11j, 7l and 7t.

  DownLoad: Full-Size Img PowerPoint

  Frontier orbital energy can also provide useful information about the biological mechanism of the molecules and target receptor [33]. The blue and green parts represent the cloud density of frontier orbitals. The frontier molecular orbitals are located on the main groups in which atoms can easily bind with the receptor [34]. As shown in Fig. 6, the HOMO and LUMO of the compounds 11j, 7l and 7t are concentrated on the same 2-aminothiazole phenyl moiety. The results suggest that the 2-aminothiazole phenyl moiety in all molecules may have a positive effect on the adjuvant activity through hydrophobic interactions or π−π conjugations.

  In summary, following an initial screen of our diversity-oriented 2-aminothiazolyl compound library, compound 1 was identified as a hit to potentiate colistin against a strain of A. baumannii. 20 analogues of the first-generation library were synthesized and evaluated for their effect on the synergistic activity for colistin. From this library, compound 7t was identified as the most active molecule. 20 analogs of the second-generation library based on 7t were synthesized and similarly evaluated their synergistic effect. Compound 11j exhibited the best adjuvant activity, decreasing the MIC of colistin by 16-fold at 8 µg/mL. Time–kill curves indicated that 11j had significant adjuvant activity and the predicted ADMET analysis showed that 2-aminothiazoyl piperidine derivatives had good drug-likeness and acceptable physicochemical properties. Furthermore, the membrane permeability experiment suggested that compound 11j can assist colistin to destroy the outer membrane of bacteria. Finally, 3D-QSAR models and DFT calculations were established to direct the following discovery of higher active adjuvants. Besides, further studies on structural optimization and synergistic mechanism are currently in progress.

  Declaration of competing interest

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

  Acknowledgments

  This work was financially supported by the National Natural Science Foundation of China (No. 81872775). The authors thank Lu Zhang and Hongchen Yang for the preparation of intermediates and revision of the manuscript.

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2022.107948.

  
  
• 1. [1]

     C. Greene, G. Vadlamudi, D. Newton, B. Foxman, C. Xi, Am. J. Infect. Control. 44 (2016) e65-e71. doi: 10.1016/j.ajic.2015.12.012

  2. [2]

     T. Demirdal, U.S. Sari, S.A. Nemli, Ann. Clin. Microbiol. Antimicrob. 15 (2016) 11-17. doi: 10.1186/s12941-016-0123-7

  3. [3]

     O. Ayobami, N. Willrich, T. Harder, et al. Emerg. Microbes Infect. 8 (2019) 1747-1759. doi: 10.1080/22221751.2019.1698273

  4. [4]

     S. Domingues, N. Rosário, Â. Cândido, et al. Microorganisms 7 (2019) 30-37. doi: 10.3390/microorganisms7020030

  5. [5]

     V. Manchanda, S. Sanchaita, N. Singh, J. Global Infect. Dis. 2 (2010) 291-304. doi: 10.4103/0974-777X.68538

  6. [6]

     S. Siricilla, K. Mitachi, J. Yang, et al. J. Med. Chem. 60 (2017) 2869-2878. doi: 10.1021/acs.jmedchem.6b01805

  7. [7]

     Global priority list of antibiotic-resistant bacteria to guide research, discovery and development of new antibiotics, WHO, 2017.

  8. [8]

     J. Li, R.L. Nation, Clin. Infect. Dis. 43 (2006) 663-664. doi: 10.1086/506571

  9. [9]

     T. Velkov, P.E. Thompson, R.L. Nation, J. Li, J. Med. Chem. 53 (2010) 1898-1916. doi: 10.1021/jm900999h

  10. [10]

      H. Spapen, R. Jacobs, V.V. Gorp, J. Troubleyn, P.M. Honoré, Ann. Intensive Care 1 (2011) 14-21. doi: 10.1186/2110-5820-1-14

  11. [11]

      J.D. Hartzell, R. Neff, J. Ake, et al. Clin. Infect. Dis. 48 (2009) 1724-1728. doi: 10.1086/599225

  12. [12]

      C.J. Kubin, T.M. Ellman, V. Phadke, et al. J. Infect. 65 (2012) 80-87. doi: 10.1016/j.jinf.2012.01.015

  13. [13]

      D.S. Akajagbor, S.L. Wilson, K.D. Shere-Wolfe, et al. Clin. Infect. Dis. 57 (2013) 1300-1303. doi: 10.1093/cid/cit453

  14. [14]

      M.H. Rigatto, T.F. Behle, D.R. Falci, et al. J. Antimicrob. Chemother. 70 (2015) 1552-1557. doi: 10.1093/jac/dku561

  15. [15]

      X.Y. Zheng, Q. Cao, Q. Cao, et al. Chin. Chem. Lett. 31 (2020) 413–417. doi: 10.1016/j.cclet.2019.07.063

  16. [16]

      M. Tyers, G.D. Wright, Nat. Rev. Microbiol. 17 (2019) 141-155. doi: 10.1038/s41579-018-0141-x

  17. [17]

      A.R. Brochado, A. Telzerow, J. Bobonis, et al. Nature 559 (2018) 259-263. doi: 10.1038/s41586-018-0278-9

  18. [18]

      C. Liu, H. Shen, S. Wang, et al. Chin. Chem. Lett. 29 (2018) 1824–1828. doi: 10.1016/j.cclet.2018.10.025

  19. [19]

      M.D. Mandler, V. Baidin, J. Lee, et al. J. Am. Chem. Soc. 140 (2018) 6749-6753. doi: 10.1021/jacs.8b02283

  20. [20]

      C.R. MacNair, J.M. Stokes, L.A. Carfrae, et al. Nat. Commun. 9 (2018) 458. doi: 10.1038/s41467-018-02875-z

  21. [21]

      S. French, M. Farha, M.J. Ellis, et al. ACS Infect. Dis. 6 (2020) 1405-1412. doi: 10.1021/acsinfecdis.9b00159

  22. [22]

      M. Vaara, Molecules 24 (2019) 249-256. doi: 10.3390/molecules24020249

  23. [23]

      B.M. Minrovic, D. Jung, R.J. Melander, C. Melander, ACS Infect. Dis. 4 (2018) 1368-1376. doi: 10.1021/acsinfecdis.8b00103

  24. [24]

      J.M. Stokes, C.R. MacNair, B. Ilyas, et al. Nat. Microbiol. 2 (2017) 17028. doi: 10.1038/nmicrobiol.2017.28

  25. [25]

      G.Q. Song, W.M. Wang, Z.S. Li, Y. Wang, J.G. Wang, Chin. Chem. Lett. 29 (2018) 899–902. doi: 10.1016/j.cclet.2017.09.035

  26. [26]

      R. Wang, S. Hou, X. Dong, et al. MedChemComm 8 (2017), 2060-2066. doi: 10.1039/C7MD00354D

  27. [27]

      P. Liu, H. Shen, X. Shao, Z. Li, X. Xu, Synlett 25 (2014) 2797-2801. doi: 10.1055/s-0034-1379100

  28. [28]

      G.B. Elion, S. Singer, G.H. Hitchings, J. Biol. Chem. 208 (1954) 477-488. doi: 10.1016/S0021-9258(18)65573-5

  29. [29]

      Y.L. Wang, F. Wang, X.X. Shi, et al. Brief Bioinform. 22 (2021) bbaa276. doi: 10.1093/bib/bbaa276

  30. [30]

      F.F. Huang, H. Ling, J. Li, et al. Chin. Chem. Lett. 31 (2020) 141–144. doi: 10.1016/j.cclet.2019.03.054

  31. [31]

      C.S. Yu, Q. Wang, J. Bajsa-Hirschel, et al. J. Agri. Food. Chem. 69 (2021) 6423–6430. doi: 10.1021/acs.jafc.0c07538

  32. [32]

      J.M. Jin, J.F. Chen, H.F. Wang, P.J. Hua, Chin. Chem. Lett. 30 (2019) 618–623. doi: 10.4163/jnh.2019.52.6.618

  33. [33]

      Z.H. Shen, Z.H. Sun, J.J. Becnel, et al. Lett. Drug Des. Discov. 15 (2018) 951–956. doi: 10.2174/1570180815666180102141640

  34. [34]

      X.H. Liu, L. Pan, C.X. Tan, et al. Pestic. Biochem. Physiol. 101 (2011) 143–147. doi: 10.1016/j.pestbp.2011.08.006

• 

  Figure 1  (A) Library screening for colistin potentiation. (B) Structural optimization for hit 1.

  下载: 全尺寸图片 幻灯片
  

  Scheme 1  General synthetic route for first-generation library: (i) KSCN, acetone, r.t. for 1 h. Yield: 80%. (ii) 4-Trifluoromethyl-aniline, EA, reflux for 1 h at 80 ℃. Yield: 98%. (iii) 2 mol/L NaOH, EtOH, reflux for 1 h at 80 ℃. Yield: 95%. (iv) tert-Butyl 4-(2-bromoacetyl)piperidine-1-carboxylate. TEA, EtOH, reflux for 1 h at 80 ℃. Yield: 54%. (v) 4 mol/L HCl/dioxane, DCM, r.t. for 30 min. (vi) Commercially available carboxylic acid, HATU, DIPEA, DCM, r.t. for 24 h. Yield: 32%–63%.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  First-generation library 7a-7t potentiation of colistin against ATCC19606. Synergistic effect on colistin at the concentration of (A) 32 µg/mL, (B) 8 µg/mL, (C) 2 µg/mL, (D) FICI of the first-generation library.

  下载: 全尺寸图片 幻灯片
  

  Scheme 2  General synthetic route for second-generation library: (i) KSCN, acetone, r.t. for 1 h. Yield: 80%. (ii) Commercially available corresponding aniline, EA, reflux for 1 h at 80 ℃. Yield: 98%. (iii) 2 mol/L NaOH, EtOH, reflux for 1 h at 80 ℃. Yield: 95%. (iv) tert-Butyl 4-(2-bromoacetyl)piperidine-1-carboxylate, TEA, EtOH, reflux for 1 h at 80 ℃. Yield: 54%. (v) 4 mol/L HCl/dioxane, DCM, r.t. for 30 min. (vi) 4-Chloro-3-ethyl-1-methyl-1H-pyrazole-5-carboxylic acid, HATU, DIPEA, DCM, r.t. for 24 h. Yield: 32%–65%.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  Second-generation library 11a-11t potentiation of colistin against ATCC19606. Synergistic effect on colistin at the concentration of (A) 32 µg/mL, (B) 8 µg/mL, (C) 2 µg/mL, (D) FICI of the second-generation library.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  (A) Time-kill curves for combination of 11j and colistin against ATCC19606. (B) The schedule shows the experimental design. ATCC19606 were incubated probes at 37 ℃ for 30 min. Four aliquots of cultures were then treated with 11j (4 µg/mL), colistin (0.5 µg/mL) and both together. (C) Dynamic curves of the permeability of the outer membrane probed with NPN.

  下载: 全尺寸图片 幻灯片
  

  Figure 5  (A) Molecular alignment of all target compounds. (B) Steric fields, with green and yellow polyhedra indicating the regions where steric bulk would enhance and reduce the activity. (C) Electrostatic fields, with blue and red polyhedra indicating the regions where positive and negative charges would enhance activity. (D) Hydrophobic fields, with yellow and gray polyhedra indicating the regions where hydrophobicity and hydrophilicity would enhance activity. (E) H-bond donor fields, with cyan contours indicating the regions where H-bond donor groups would enhance activity. (F) H-bond acceptor fields, with red contours indicating the regions where H-bond acceptor groups would enhance activity.

  下载: 全尺寸图片 幻灯片
  

  Figure 6  Frontier molecular orbitals of selected target compounds 11j, 7l and 7t.

  下载: 全尺寸图片 幻灯片
•