Supramolecular antibiotics: a strategy for conversion of broad-spectrum to narrow-spectrum antibiotics for Staphylococcus aureus

Thameez M. Koyasseril-Yehiya a, Alam García-Heredia b, Francesca Anson a, Poornima Rangadurai a, M. Sloan Siegrist *bcd and S. Thayumanavan *abde
aDepartment of Chemistry, University of Massachusetts, Amherst, Massachusetts 01003, USA. E-mail:
bMolecular and Cellular Biology Program, University of Massachusetts, Amherst, Massachusetts 01003, USA
cDepartment of Microbiology, University of Massachusetts, Amherst, Massachusetts 01003, USA. E-mail:
dModels to Medicine, Institute for Applied Life Sciences, University of Massachusetts, Amherst, Massachusetts 01003, USA
eThe Center for Bioactive Delivery, Institute for Applied Life Sciences, University of Massachusetts, Amherst, Massachusetts 01003, USA

Received 29th June 2020 , Accepted 24th September 2020

First published on 24th September 2020


The propensity of broad-spectrum antibiotics to indiscriminately kill both pathogenic and beneficial bacteria has a profound impact on the spread of resistance across multiple bacterial species. Alternative approaches that narrow antibacterial specificity towards desired pathogenic bacterial population are of great interest. Here, we report an enzyme-responsive antibiotic-loaded nanoassembly strategy for narrow delivery of otherwise broad-spectrum antibiotics. We specifically target Staphylococcus aureus (S. aureus), an important blood pathogen that secretes PC1 β-lactamases. Our nanoassemblies selectively eradicate S. aureus grown in vitro with other bacteria, highlighting its potential capability in targeting the desired pathogenic bacterial population.


The introduction of antibiotics is estimated to have saved millions of lives.1 A longstanding challenge with antibiotics, however, is the development of drugs that specifically target pathogenic bacteria, while leaving the commensal bacteria unharmed. Classical broad-spectrum antibiotics do not differentiate between pathogenic and beneficial bacteria, thus eliminating both populations.2 Additionally, indiscriminate use of antibiotics, mutations and horizontal gene transfer have led to the emergence of bacterial resistance.3 There is a pressing need for strategies that specifically target pathogenic bacteria.

Nanoassemblies have exhibited great potential in many biological applications, especially in drug delivery.4,5 Stimuli-responsive nanoassemblies have gained particular interest due to their programmed release of encapsulated cargoes in a spatiotemporal manner.6–8 Specifically, enzyme-responsive assemblies have attracted interest due to direct relevance to biological systems.9–16 In this context, microbial enzymes17–21 and toxins22,23 are well studied in the literature. However, systems that respond to bacterial enzymes secreted as part of resistance mechanism have been less explored.24

Herein, we report a nanoassembly that responds to enzymes secreted by pathogenic bacteria, specifically to enzymes that deactivate antibiotics. A common and important bacterial resistance mechanism involves the secretion of β-lactamase enzymes, which disable the pharmacophore of β-lactam class of drugs through a ring opening pathway.25,26 We envisaged a reactive supramolecular disassembly strategy, where the very mechanism that the bacteria use to disable these antibiotics could be utilized to unleash a differential antibacterial drug (Fig. 1).

image file: d0nr04886k-f1.tif
Fig. 1 Schematic representation of antibiotic-loaded nanoparticles undergoing disassembly upon binding with β-lactamase enzyme, allowing the release of encapsulated antibiotics.

In our strategy, we harness β-lactamase activity to trigger the release of one antibiotic and concurrently deactivate the enzyme to potentiate another. To this end, we repurposed our protein-responsive supramolecular disassembly platform, where specific ligand–protein interactions can be used to cause binding-induced disassembly of amphiphilic nanoassemblies.27,28 This is driven by the difference in the hydrophilic-lipophilic balance (HLB) of the constituent amphiphiles in its protein-bound and unbound forms. In the current format, we installed specific covalent inhibitors for β-lactamase (e.g. clavulanic acid)29 on the surface of the amphiphilic assemblies, with two key consequences. First, the covalent nature of the inhibitor irreversibly deactivates the enzyme. Second, the covalent interaction with the enzyme is equivalent to the ligand–protein binding event that alters the amphiphile's HLB and causes disassembly. Disassembly then allows non-covalently encapsulated antibiotic molecules to be released. Thus, the mechanism that normally deactivates β-lactam antibiotics would unleash the nanoassembly cargo, in this case, antibacterial drugs.

Although β-lactamase responsive polymeric vesicles based on a self-immolative process has been reported before,24 polymeric system can be responsive to multiple β-lactamases from a variety of bacteria. Extensive use of β-lactam antibiotics has selected for the increased prevalence of multiple β-lactamases across bacterial species.30 The diversity includes differences in turnover number (tn) towards specific substrates. We chose a low tn substrate, clavulanic acid (CA), to facilitate the irreversible substrate binding to the β-lactamase enzyme to induce reactive supramolecular disassembly process. This approach aids in specifically targeting enzymes with low turnover number towards the clavulanate substrate. We focused on the PC1 β-lactamase (PC1-βL) secreted by S. aureus, which has tn = 1 for CA.31 We hypothesized that β-lactamases with low tn towards CA could trigger the disassembly process by irreversible binding to the nanoassemblies. This offers to specifically target PC1-βL (tn = 1), an enzyme secreted by S. aureus. S. aureus is one of the most common causative agents of bacteremia, a blood infection with an annual incidence rate ranging from 20–50 cases per 100[thin space (1/6-em)]000 population.32

Results and discussion

Design and synthesis of enzyme responsive nanoassemblies

We designed and synthesized an oligomeric amphiphile with polyethylene glycol (PEG, hydrophilic) and aliphatic decyl (hydrophobic) moieties (Scheme 1). CA, the key β-lactamase responsive moiety was incorporated along with a PEG linker using strain promoted azide alkyne cycloaddition (SPAAC) reaction to yield 1-CA (see ESI). The PEG linker solvent-exposes the CA moiety for facile interaction with β-lactamase, when assembled in aqueous phase. β-Lactamase initiates nucleophilic attack of the β-lactam of clavulanate using active site serine to form an acyl-enzyme species.33 The consequential ring opening cascade yields a transient imine intermediate. Nucleophilic functionalities at the vicinity of the enzyme's active site then reacts with the newly generated electrophilic center to achieve irreversible covalent binding (Scheme 1). We hypothesized that, this binding of bulkier enzyme onto the clavulanate-decorated nanoassemblies could disrupt the HLB, leading to disassembly and concurrent release of the encapsulated antibiotics.
image file: d0nr04886k-s1.tif
Scheme 1 Chemical structure of clavulanate decorated amphiphilic oligomer, self-assembly to encapsulate rifampicin in its hydrophobic pocket and its mechanism of action in covalent inhibition of β-lactamase enzyme.

Self-assembly and PC1-βL responsive behavior of 1-CA

The self-assembly 1-CA and its enzyme-induced molecular release through disassembly were characterized. 1-CA amphiphiles were self-assembled using hydrophobicity-induced aggregation through co-solvent approach. A minimal amount of acetone was initially used to solubilize 1-CA amphiphiles, while later the desired amount of water was added in a dropwise manner. The solution was stirred overnight uncapped, allowing the acetone to evaporate over time, forcing 1-CA molecules to self-assemble to minimize its interfacial energy. We found that 1-CA self-assembles into large compound micelles (∼190 nm) that can sequester hydrophobic guest molecules in aqueous media (ESI, Fig. S1). Similar oligomeric amphiphiles have been characterized before to form large spherical compound micelles.34 To test if these nanoassemblies are enzyme-responsive, a 25 μM solution of Nile red encapsulated 1-CA was incubated with 50 μM of PC1-βL. ∼90% release of the encapsulated hydrophobic dye was observed from the 1-CA nanoassembly, while no significant release was observed in buffer without PC1-βL (Fig. 2a). ∼20% dye release observed in this control could be due to the liberation of small percentage of loosely bound dye molecules to the nanocarriers. The disassembly characteristics were further analyzed using dynamic light scattering (DLS) and transmission electron microscopy (TEM). From DLS, drastic change in size (190 nm to <10 nm) was observed upon PC1-βL incubation for 24 hours (Fig. 2b). TEM images further correlates with the observed enzyme-induced size change (Fig. 2c and d). TEM assembly sizes were substantially different from DLS, likely because of the requisite sample drying for TEM experiments. To validate the specificity of the clavulanate inhibitor binding to β-lactamase, non-specific enzymes, such as lysozyme and RNase were also tested. Neither discernible release of the hydrophobic guest molecules nor significant change in the size of 1-CA nanoassemblies were found, demonstrating the specificity of the enzyme-mediated disassembly (ESI, Fig. S2).
image file: d0nr04886k-f2.tif
Fig. 2 (a) Plot of % release of Nile red and (b) DLS profiles of 1-CA upon incubation with PC1 β-lactamase; TEM images of 1-CA (c) before and (d) after treatment with enzyme for 24 hours; (e) release profiles due to β-lactamase enzymes with varying turnover numbers towards clavulanate.

We next tested the specificity of 1-CA nanoassemblies with respect to β-lactamases from other bacterial species, including TEM-1 from E. coli and β-lactamases from B. subtilis, B. cereus and E. cloacae. No significant release of encapsulated dye was observed from 1-CA in presence of β-lactamase enzymes except for the PC1-βL (Fig. 2e and Table S1). This could be attributed to differences in enzyme tn. Although clavulanate is known to be an irreversible inhibitor, enzymes undergo numerous catalytic cycles per unit time depending on their turnover number before binding irreversibly to an inhibitor. The tn's of PC1-βL (S. aureus), TEM-1 (E. coli) and β-lactamase from B. cereus towards CA are 1, 160 and 16[thin space (1/6-em)]000 respectively.35–37 This implies that TEM-1 β-lactamase (E. coli) requires ∼160 clavulanate molecules to inactivate the enzyme, whereas PC1-βL (S. aureus) requires only one clavulanate molecule. To trigger disassembly, high fidelity in the irreversible enzymatic reaction is critical. If nearly only 1/160 clavulanate cleavage results in enzyme binding, then <1% of the amphiphile is removed from the assembly, which is insufficient for the enzyme-induced disassembly and molecular release processes. We thus envisaged that the low tn of PC1-βL in principle, could allow selective targeting of pathogenic S. aureus.

We then asked whether dye release from the nanoassemblies is indeed due to the covalent binding of PC1-βL to the clavulanate in 1-CA. To this end, we compared the products of PC1-βL binding to small molecule inhibitor CA and 1-CA independently, using ESI mass spectrometry. Without any inhibitor treatment, the enzyme showed peaks around ∼31 kDa. When incubated with either CA or 1-CA, peaks corresponding to the unmodified enzyme diminished and new mass adducts (Δ + 52, Δ + 70 and Δ + 88) appeared. These peaks correspond to the different inactivated enzyme species, specifically arising from the cleavage of imine/enamine intermediate (ESI, Fig. S4). Elimination reaction or the hydrolytic cleavage of the intermediate results in the enzyme modified with a propynyl derivative (Δ + 52), aldehyde derivative (Δ + 70) and the hydrated version of the aldehyde (Δ + 88).33,38,39 Since all these adducts are formed due to the reaction of the active site serine, followed by a cleavage reaction within the CA functionality, the adducts are similar for the interaction of PC1-βL with CA and 1-CA.

Antibiotic encapsulation and antibacterial studies

To evaluate whether our system would be effective in narrowing the broad-spectrum antibiotic to target S. aureus, a hydrophobic, broad-spectrum antibiotic rifampicin (RIF) was encapsulated into 1-CA nanoassemblies (1-CA-RIF). The amount of drug encapsulated inside the nanocarriers was quantified using UV-Vis spectroscopy and the encapsulation was found to be stable without significant release. We observed ∼90% release of RIF from 1-CA nanoassemblies upon incubation with PC1-βL, comparable to that of the hydrophobic dye observed in Fig. 2a (ESI, Fig. S5e and f). We then evaluated the minimum inhibitory concentrations (MIC) of free RIF and 1-CA-RIF. We chose bacterial species with divergent cell envelope composition and shape to test whether these characteristics could induce the non-specific disassembly of nanoparticles (Table 1). The MIC values of free RIF and 1-CA-RIF were comparable in most species. In S. aureus, however, the value for 1-CA-RIF was 1000-fold lower than that of free RIF. This could be due to the combined effect of clavulanate inhibitor that inactivates the β-lactamase enzyme and the released rifampicin, which has the bactericidal effect. Moreover, control experiment with either free RIF and 1-CA or free RIF and CA were found to influence rifampicin's susceptibility towards S. aureus but not for E. coli (ESI, Fig. S6). 1-CA-RIF, RIF + 1-CA and RIF + CA treatments exhibited similar bacterial inhibition in S. aureus. Correspondingly, as observed here, the combinatorial administration of β-lactam inhibitors and antibiotics has been shown to be more effective in treating β-lactamase producing bacteria.40 Furthermore, in accordance to the dye release experiments (Fig. 2e and Table S1), our system did not show major activity against E. coli and B. subtilis (Table 1). These data demonstrate that disassembly of nanoassemblies can liberate the encapsulated RIF and selectively inhibit S. aureus growth likely due to PC1-βL secretion.
Table 1 Minimum inhibitory concentration of free RIF versus1-CA-RIF
Bacterial species RIF (μg mL−1) 1-CA-RIF (μg mL−1)
a The MIC value measurement was limited by the loading capacity of RIF inside the 1-CA nanoassemblies. Maximum encapsulated RIF concentration tested was ∼6 μg mL−1.
S. aureus 10.42 ± 2.95 0.012 ± 0.002
E. coli 10.4 ± 3.68 >6.00a
M. smegmatis 1.95 1.48 ± 0.25
B. subtilis 0.24 0.368 ± 0.062
P. aeruginosa 7.8 0.738 ± 0.125
L. lactis 0.49 0.738 ± 0.125
C. crescentus 0.19 0.368 ± 0.062

Selective targeting of PC1-βL secreting S. aureus

Since the MIC of 1-CA-RIF for S. aureus was ∼30–120 fold lower than the MICs of other species, we became interested in testing whether we could specifically eliminate S. aureus. We used an AlamarBlue assay to quantify the viability of the bacterial species upon treatment with fixed amounts of either 1-CA or 1-CA-RIF. While the 1-CA nanoassemblies alone were nontoxic to all seven bacterial species, 1-CA-RIF selectively inhibited S. aureus growth (Fig. 3). These data demonstrate that the nanoassemblies selectively target S. aureus in a homogenous, single-species context.
image file: d0nr04886k-f3.tif
Fig. 3 Percent viability of bacterial strains using alamarBlue assay with the treatment of either 1-CA or 1-CA-RIF. 1% of individual bacteria (OD 0.6–0.8) were inoculated with media containing 1-CA or 1-CA-RIF and incubated overnight at 37 °C. ns, no statistically significant differences; ****p < 0.0001.

Next, we tested whether our approach would be able to target S. aureus in the presence of other bacteria. To test this, we established an E. coliS. aureus co-culture model in which organisms were grown separately, later mixed in a culture and challenged with 1-CA-RIF. E. coli and S. aureus were chosen for co-culture because they have similar doubling times and can easily be distinguished based on morphology using microscopy. We hypothesized that nanoassemblies would release the RIF at the surface of S. aureus cells due to the presence of PC1-βL but that assemblies would remain intact near the surface of E. coli cells. To visualize this, we relied on widely used bacterial viability co-staining method using SYTO-9 and propidium iodide (PI).41 Due to their differences in DNA-binding affinity and membrane permeability dependence, green signal (SYTO-9) cells are considered alive and red fluorescent signal (PI) cells as dead. We challenged growing cultures of independent S. aureus, E. coli and a coculture to 1-CA or 1-CA-RIF for 6 hours, followed by labelling the cells with SYTO-9 and PI to discriminate live and dead bacteria. Fluorescence microscopy imaging revealed that, both rod-shaped E. coli and rounded S. aureus exhibited green fluorescence without any treatment. However, 1-CA-RIF treatment resulted in red fluorescence signal from S. aureus cells and green signal from E. coli cells, implying that it can selectively kill the S. aureus bacterial population (ESI, Fig. S7a). Free RIF treatment resulted in substantial killing of both S. aureus and E. coli grown as monocultures (ESI, Fig. S7b and c). Altogether, data suggests that 1-CA-RIF can target S. aureus in both individual and complex cultures.

Additionally, blood agar media was used to distinguish the colonies produced by S. aureus (white) and E. coli (dark grey) due to their differences in hemolysis (Fig. 4a). Without any treatment, both bacterial strains formed their characteristic colonies. However, upon incubation with 1-CA-RIF, total clearance of S. aureus colonies was observed (Fig. 4b and c). Additionally, nanoassemblies alone did not exert any significant effect on either E. coli or S. aureus colonies (ESI, Fig. S8). Notably, 1-CA-RIF did not significantly impact E. coli viability even with high colony density (ESI, Fig. S9), suggesting its potential for discriminating bacterial species in complex cultures for delivering encapsulated cargoes. The detailed mechanism of the selective elimination in co-culture model is currently under investigation. We plan to investigate the selective inhibition and dissociation process by fluorescently labelling the nanoassemblies.

image file: d0nr04886k-f4.tif
Fig. 4 Co-culture experiments: (a) E. coli (*) & S. aureus (#) plated on blood agar media; E. coli and S. aureus cultures were grown independently and normalized to ∼0.5 OD. Cells were combined (1[thin space (1/6-em)]:[thin space (1/6-em)]8, E. coli: S. aureus) and incubated overnight with or without 1-CA-RIF. Dilutions were made in sterile PBS and plated on blood agar media. E. coliS. aureus co-culture on blood agar plate with (b) without any treatment and (c) with 1-CA-RIF.

Further, we aimed to confirm the role of active PC1-βL in the disassembly event. A competitive experiment with CA was performed, to test if the small molecule could rescue the viability in S. aureus by inactivating PC1-βL. Our approach was to first inactivate the PC1-βL using CA, followed by the treatment with 1-CA-RIF nanoassemblies. We hypothesized that, the nanoassemblies would be more stable due to the reduction in active PC1-βL enzyme. To test this, S. aureus was incubated with CA for 30 minutes followed by the addition of 1-CA-RIF. As expected, CA alone showed similar results as the untreated control, indicating that the CA itself does not have any bactericidal effect (ESI, Fig. S11). Comparatively, 1-CA-RIF alone exhibited bactericidal activity. In contrast, preincubation with CA increased the viability of S. aureus colonies three-fold when challenged with 1-CA-RIF. These results validate that the interaction between active PC1-βL with 1-CA-RIF is essential to promote antibiotic release and induce bactericidal effect.

Finally, we tested the activity of other β-lactamase irreversible inhibitors, such as sulbactam and tazobactam. Although the mechanism of action of these inhibitors are similar to clavulanate, sulbactam and tazobactam have sulfones instead of an enol functionality at the C-1 position of the five membered ring adjacent to the β-lactam ring. We were interested in evaluating the disassembly potency of the amphiphiles bearing sulbactam and tazobactam moieties, compared to the clavulanate system. To test this, amphiphiles with sulbactam and tazobactam were synthesized by reacting DBCO functionalized precursor amphiphiles with the corresponding azide modified sulbactam and tazobactam using SPAAC reaction to yield 1-SB-RIF and 1-TB-RIF respectively (see ESI for details). First, we were gratified to find that 1-SB-RIF and 1-TB-RIF assemblies also exhibit selective bactericidal effect to S. aureus, compared to E. coli (Fig. 5), although the MIC of the free drug itself is very similar for these two bacterial strains. Additionally, we also noted that 1-CA-RIF exhibited higher potency among the three nanoassemblies. We hypothesize that, this could be due to the presence of better leaving group (enol) functionality at C-1 position of CA than the sulfones present in sulbactam and tazobactam. Better leaving group ability would facilitate secondary ring opening, eventually leading to effective covalent enzyme inhibition and supramolecular disassembly.

image file: d0nr04886k-f5.tif
Fig. 5 (a) Chemical structures of sulbactam and tazobactam modified oligomers; (b) comparison of MIC values of different inhibitors modified oligomeric systems in S. aureus and E. coli. [x] The MIC value measurement was limited by the loading capacity of RIF inside the 1-CA nanoassemblies. Maximum encapsulated RIF concentration tested was ∼6 μg mL−1.


In summary, we designed a supramolecular platform for programmed delivery of a broad-spectrum antibiotic to a desired pathogenic bacterial population. Bacterial β-lactamases trigger the antibiotic release by covalently binding to the complementary functionalities embedded on the surface of nanoassemblies. Our approach was able to primarily target S. aureus in homogenous and heterogenous cultures in vitro, owing to their PC1 β-lactamase secretion with low tn towards clavulanate moiety. Finally, clavulanate modified oligomers exhibited higher potency compared to the tazobactam and sulbactam oligomers due to the presence of better leaving group in 1-CA. In the future, we intend to evaluate the efficacy of our system in targeting Methicillin-resistant S. aureus (MRSA) and S. aureus in biofilms. Our strategy offers narrow delivery of broad-spectrum antibiotics to specific pathogenic bacterial populations, which could be potentially utilized in the future as an alternative to regular broad-spectrum antibiotics administration and may help to mitigate the effects of drugs on commensal bacteria. Overall, the idea of targeting the low turnover enzymes greatly expands the repertoire of enzyme-responsive materials for biomedical applications.

Conflicts of interest

The authors declare no competing financial interests.


We thank the support from the National Institutes of Health, NIH (GM-136395), NIH (R21 AI144748) and NIH (DP2 AI138238). A. G.-H. was supported by Honors Fellowship from Universidad Autónoma de Nuevo León. F. A was supported by the UMass BTP Program (NIH T32 GM108556). We are grateful to Dr Stephen Eyles and Mass spectrometry core facility, Institute for Applied Life Sciences at UMass Amherst for Mass spectrometry analysis and discussions. We thank A. Heuck lab and P. Chien lab at UMass Amherst, for providing P. aeruginosa and C. crescentus bacterial strains, respectively.

Notes and references

  1. C. L. Ventola, P T, 2015, 40, 277–283 Search PubMed .
  2. C. Jernberg, S. Löfmark, C. Edlund and J. K. Jansson, ISME J., 2007, 1, 56–66 CrossRef CAS .
  3. S. B. Levy and B. Marshall, Nat. Med., 2004, 10, S122–S129 CrossRef CAS .
  4. A. Gupta, S. Mumtaz, C.-H. Li, I. Hussain and V. M. Rotello, Chem. Soc. Rev., 2019, 48, 415–427 RSC .
  5. O. C. Farokhzad and R. Langer, ACS Nano, 2009, 3, 16–20 CrossRef CAS .
  6. A. P. Blum, J. K. Kammeyer, A. M. Rush, C. E. Callmann, M. E. Hahn and N. C. Gianneschi, J. Am. Chem. Soc., 2015, 137, 2140–2154 CrossRef CAS .
  7. E. R. Gillies, T. B. Jonsson and J. M. J. Fréchet, J. Am. Chem. Soc., 2004, 126, 11936–11943 CrossRef CAS .
  8. J. Zhuang, M. R. Gordon, J. Ventura, L. Li and S. Thayumanavan, Chem. Soc. Rev., 2013, 42, 7421–7435 RSC .
  9. A. J. Harnoy, M. Buzhor, E. Tirosh, R. Shaharabani, R. Beck and R. J. Amir, Biomacromolecules, 2017, 18, 1218–1228 CrossRef CAS .
  10. J. Rao, C. Hottinger and A. Khan, J. Am. Chem. Soc., 2014, 136, 5872–5875 CrossRef CAS .
  11. J. Zhuang, H. Seçinti, B. Zhao and S. Thayumanavan, Angew. Chem., Int. Ed., 2018, 57, 7111–7115 CrossRef CAS .
  12. S. Samarajeewa, R. Shrestha, Y. Li and K. L. Wooley, J. Am. Chem. Soc., 2012, 134, 1235–1242 CrossRef CAS .
  13. M. Zelzer, S. J. Todd, A. R. Hirst, T. O. McDonald and R. V. Ulijn, Biomater. Sci., 2013, 1, 11–39 RSC .
  14. A. J. Harnoy, I. Rosenbaum, E. Tirosh, Y. Ebenstein, R. Shaharabani, R. Beck and R. J. Amir, J. Am. Chem. Soc., 2014, 136, 7531–7534 CrossRef CAS .
  15. M. A. Azagarsamy, P. Sokkalingam and S. Thayumanavan, J. Am. Chem. Soc., 2009, 131, 14184–14185 CrossRef CAS .
  16. V. Kumar, T. M. Koyasseril-Yehiya and S. Thayumanavan, in Molecular Assemblies: Characterization and Applications, ACS, 2020, vol. 7, pp. 95–107 Search PubMed .
  17. Y. Liu, H. J. Busscher, B. Zhao, Y. Li, Z. Zhang, H. C. van der Mei, Y. Ren and L. Shi, ACS Nano, 2016, 10, 4779–4789 CrossRef CAS .
  18. M.-H. Xiong, Y.-J. Li, Y. Bao, X.-Z. Yang, B. Hu and J. Wang, Adv. Mater., 2012, 24, 6175–6180 CrossRef CAS .
  19. L.-L. Li, J.-H. Xu, G.-B. Qi, X. Zhao, F. Yu and H. Wang, ACS Nano, 2014, 8, 4975–4983 CrossRef CAS .
  20. B. Hisey, P. J. Ragogna and E. R. Gillies, Biomacromolecules, 2017, 18, 914–923 CrossRef CAS .
  21. X. Ding, A. Wang, W. Tong and F.-J. Xu, Small, 2019, 15, 1900999 CrossRef .
  22. Y. Chen, Y. Zhang, M. Chen, J. Zhuang, R. H. Fang, W. Gao and L. Zhang, Small, 2019, 15, 1804994 CrossRef .
  23. D. Pornpattananangkul, L. Zhang, S. Olson, S. Aryal, M. Obonyo, K. Vecchio, C.-M. Huang and L. Zhang, J. Am. Chem. Soc., 2011, 133, 4132–4139 CrossRef CAS .
  24. Y. Li, G. Liu, X. Wang, J. Hu and S. Liu, Angew. Chem., Int. Ed., 2016, 55, 1760–1764 CrossRef CAS .
  25. R. J. Holt and G. T. Stewart, Microbiology, 1964, 36, 203–213 CrossRef CAS .
  26. M. Babic, A. M. Hujer and R. A. Bonomo, Drug Resist. Updates, 2006, 9, 142–156 CrossRef CAS .
  27. M. A. Azagarsamy, V. Yesilyurt and S. Thayumanavan, J. Am. Chem. Soc., 2010, 132, 4550–4551 CrossRef CAS .
  28. M. R. Molla, P. Prasad and S. Thayumanavan, J. Am. Chem. Soc., 2015, 137, 7286–7289 CrossRef CAS .
  29. S. M. Drawz and R. A. Bonomo, Clin. Microbiol. Rev., 2010, 23, 160–201 CrossRef CAS .
  30. I. Massova and S. Mobashery, Antimicrob. Agents Chemother., 1998, 42, 1–17 CrossRef CAS .
  31. M. H. Richmond, Biochem. J., 1963, 88, 452–459 CrossRef CAS .
  32. S. J. van Hal, S. O. Jensen, V. L. Vaska, B. A. Espedido, D. L. Paterson and I. B. Gosbell, Clin. Microbiol. Rev., 2012, 25, 362–386 CrossRef CAS .
  33. R. P. A. Brown, R. T. Aplin and C. J. Schofield, Biochemistry, 1996, 35, 12421–12432 CrossRef CAS .
  34. J. Gao, X. Liu, H. Secinti, Z. Jiang, O. Munkhbat, Y. Xu, X. Guo and S. Thayumanavan, Chem. – Eur. J., 2018, 24, 1789–1794 CrossRef CAS .
  35. K. Bush, C. Macalintal, B. A. Rasmussen, V. J. Lee and Y. Yang, Antimicrob. Agents Chemother., 1993, 37, 851–858 CrossRef CAS .
  36. E. B. Chaibi, J. Péduzzi, S. Farzaneh, M. Barthélémy, D. Sirot and R. Labia, Biochim. Biophys. Acta, Protein Struct. Mol. Enzymol., 1998, 1382, 38–46 CrossRef CAS .
  37. J. P. Durkin and T. Viswanatha, J. Antibiot., 1978, 31, 1162–1169 CrossRef CAS .
  38. Y. Yang, K. Janota, K. Tabei, N. Huang, M. M. Siegel, Y. I. Lin, B. A. Rasmussen and D. M. Shlaes, J. Biol. Chem., 2000, 275, 26674–26682 CAS .
  39. J. M. Thomson, A. M. Distler, F. Prati and R. A. Bonomo, J. Biol. Chem., 2006, 281, 26734–26744 CrossRef CAS .
  40. L. M. Bush and C. C. Johnson, Infect. Dis. Clin. North Am., 2000, 14, 409–433 CrossRef CAS .
  41. L. Boulos, M. Prévost, B. Barbeau, J. Coallier and R. Desjardins, J. Microbiol. Methods, 1999, 37, 77–86 CrossRef CAS .


Electronic supplementary information (ESI) available. See DOI: 10.1039/D0NR04886K
These authors contributed equally.

This journal is © The Royal Society of Chemistry 2020