Mechanistic Understanding of the Interactions of Cationic Conjugated Oligo- and Polyelectrolytes with Wild-type and Ampicillin-resistant Escherichia coli

An in-depth understanding of cell-drug binding modes and action mechanisms can potentially guide the future design of novel drugs and antimicrobial materials and help to combat antibiotic resistance. Light-harvesting π-conjugated molecules have been demonstrated for their antimicrobial effects, but their impact on bacterial outer cell envelope needs to be studied in detail. Here, we synthesized poly(phenylene) based model cationic conjugated oligo- (2QA-CCOE, 4QA-CCOE) and polyelectrolytes (CCPE), and systematically explored their interactions with the outer cell membrane of wild-type and ampicillin (amp)-resistant Gram-negative bacteria, Escherichia coli (E. coli). Incubation of the E. coli cells in CCOE/CCPE solution inhibited the subsequent bacterial growth in LB media. About 99% growth inhibition was achieved if amp-resistant E. coli was treated for ~3–5 min, 1 h and 6 h with 100 μM of CCPE, 4QA-CCOE, and 2QA-CCOE solutions, respectively. Interestingly, these CCPE and CCOEs inhibited the growth of both wild-type and amp-resistant E. coli to a similar extent. A large surface charge reversal of bacteria upon treatment with CCPE suggested the formation of a coating of CCPE on the outer surface of bacteria; while a low reversal of bacterial surface charge suggested intercalation of CCOEs within the lipid bilayer of bacteria.


Results and Discussion
Synthesis and optical properties of ccoes and ccpe. For this specific study, we synthesized two phenylene based model oligoelectrolytes each with 3 phenyl groups along the backbone, and two (2QA-CCOE) or four (4QA-CCOE) propoxy pendants terminated with cationic quaternary amine groups (Fig. 1). The CCOEs were synthesized via Suzuki coupling and subsequent quaternization reaction (using ethyl bromide). We also synthesized cationic conjugated polyelectrolyte (CCPE) with phenylene based backbone following the procedure described in literature 37 (Fig. 1). The step-by-step details of chemical synthesis and characterization to confirm the chemical structures of CCOEs and CCPE are discussed in supplementary information.
treatment of E. coli with ccpe/ccoe. The wild-type E. coli strain (DH10B) was transformed by appropriate plasmids to yield ampicillin-resistant (amp-resistant) E. coli strains (SSBIO002, see the experimental section, Supplementary Table S1 and other supporting information for details) 38,39 . Here, we treated E. coli suspension (wild-type and amp-resistant) with conjugated molecules in multiple ways. We first followed two-step processes where the cells were incubated with conjugated molecules first (either in phosphate-buffered saline (PBS) or Luria Bertani (LB) medium) and then transferred to LB media to allow the bacteria grow further (treat (PBS or LB), then grow (LB)). In addition, we followed a one-step process where the E. coli cells were treated and grown at the same time in LB media (treat and grow (LB)).  In the "treat (PBS), then grow (LB)" process, wild-type and amp-resistant E. coli were treated first by mixing bacterial suspension in 5 mM PBS buffer with 2QA-CCOE, 4QA-CCOE, and CCPE. We varied the concentration of 2QA-CCOE, 4QA-CCOE, and CCPE between 0-100 μM in PBS in order to study the effect of concentration of conjugated molecules on antimicrobial properties. The cells were treated with conjugated molecules for 0-6 h at 37 °C. It is to be noted that similar E. coli treatment in PBS buffer with conjugated molecules has been demonstrated in several reports 25,31,40 . The utility of treating the bacteria in LB-free liquid (i.e. PBS here) is that the bacteria do not grow further during the treatment and that allow us to elucidate the sole effect of conjugated molecules on a specific population of bacteria. For measuring the percent of dead cells during CCOE/CCPE treatment, we performed a live/dead assay using flow cytometry (Fig. 3). In this assay, SYTO9 and PI stains were used, where SYTO9 dye is able to stain both damaged and intact cells, while PI stains damaged cells only 41,42 . The percent (%) of dead cells in E. coli suspension after incubation with conjugated molecules were measured from this assay and varied between 9-14%. In addition, the absorbance of the bacterial suspension in PBS buffer varied insignificantly (absorbance at 600 nm ~0.45-0. 55, see Supplementary Fig. S1) as a function of time of incubation in CCOEs or CCPE during this step. The UV/Vis supported the findings of live/dead assay (Fig. 3) and the fact that CCPE and CCOEs did not kill wild-type or amp-resistant E. coli predominantly in PBS buffer (during the treatment period). While cell lysis or death 20,43,44 did not happen during this CCOE/CCPE treatment step in PBS buffer, there were prominent interactions between conjugated molecules and the bacterial outer cell envelope during this step. The effect of these interactions was evident in the subsequent step in which these bacterial cells were allowed to grow in LB media and percent growth inhibition was calculated (Fig. 4).

Measurement of % growth inhibition.
The CCOE/CCPE treated E. coli cells, when allowed to grow in LB media ("treat (PBS), then grow (LB)" process), the bacterial growth was efficiently inhibited (Fig. 4). The observed % growth Inhibition of E. coli (calculated using Eq. (1) as explained in Materials and Methods section) was consistently higher when E. coli was treated with CCPE in PBS before allowing it to grow in LB media (Fig. 4). Even when E. coli (both wild-type (Fig. 4a) and amp-resistant (Fig. 4c)) were treated with 50 μM CCPE for a few minutes (~3-5 min) in PBS, bacteria grew less than untreated bacteria and led to 40% growth inhibition. This indicated an instant action of CCPE on E. coli in PBS buffer. The % growth inhibition was 80% (Fig. 4b, wild-type) or above (Fig. 4d, amp-resistant) when CCPE concentration in treatment solution was increased from 50 (Fig. 4a,c) to 100 μM (Fig. 4b,d). 2QA-CCOE and 4QA-CCOE were also able to inhibit the growth of E. coli, but needed longer treatment to generate the inhibitory effect similar to CCPE. We observed 95% growth inhibition when E. coli cells were treated with 50 μM 4QA-CCOE or CCPE for 3 h. However, to achieve this level of growth inhibition with 2QA-CCOE, doubling the incubation time was required (6 h) (Fig. 4a,c). A higher concentration of conjugated molecules (100 μM, Fig. 4b,d) helped to reach ~99% growth inhibition faster (within ~3-5 min, 1 h and 6 h with CCPE, 4QA-CCOE and 2QA-CCOE, respectively for amp-resistant E. coli (Fig. 4d)). In addition, both CCOEs and CCPE were able to inhibit the growth of antibiotic-resistant bacteria to a similar extent as wild-type strains. The fact that the bacteria were not dying during the treatment step with CCOE/ CCPE in PBS buffer, but not growing significantly when transferred to LB media evoked interest. The primary speculations were that it is either a straightforward effect of CCOE/CCPE attachment to bacteria which did not let the bacteria grow/divide further or an effect of PBS buffer during the treatment stage which negatively affected the metabolism of bacteria. Our control experiments showed that if the bacteria cells were kept in PBS buffer (without CCOE/CCPE) and later transferred to grow in LB media, the cells continued to grow in LB media. This suggests that possibly PBS buffer is not significantly affecting the metabolism. If something was decreasing the metabolism, it was likely the attachment and interaction of CCOE/CCPE with bacterial outer membrane. Zeta potential measurements (shown in a later section) will provide evidence of attachment of CCOEs/CCPE with bacterial cells.  Table S2). Through calculation using maximum CCPE adsorption by bacteria ( Supplementary Fig. S4), we found that the number of CCPE chains available in both of these treatment systems were much higher than that needed to effectively coat all the bacterial cells present. For example, by using 100 μM CCPE in the treatment system, we offered 1.06×10 15 chains of CCPE/mL which is ~5 times higher than what was needed for achieving the most efficient coating of the cells. This suggested that not having enough CCPE (to coat bacteria) was not the root cause for the poor growth inhibition in these two treatment processes. One thing in common between "treat (LB), then grow (LB)" process, and, "treat and grow (LB)" process was: in both cases, CCOE/CCPE chains were in LB media. We thus hypothesized that the poor antimicrobial activity of CCOEs and CCPEs in LB media was due to high ionic strength of the LB media (171 mM) which led to charge screening of both bacteria and cationic conjugated molecules and minimized their electrostatic interactions. The hypothesis was supported by the zeta potential reported for E. coli cells in solution with various ionic strengths (discussed in detail in supporting information). Based on all these observations for different modes of treatment and culture, "treat (PBS), then grow (LB)" process ( Fig. 4) appeared to be the most effective mode to attain bacterial growth inhibition.
The growth inhibition data for E. coli ("treat (PBS), then grow (LB)" process) was further supported by colony-forming unit (CFU) reduction assays done on agar plates ( Fig. 5 and Supplementary Fig. S5 (wild-type), S6 (amp-resistant)). Briefly, 5 μL of CCOEs/CCPE-treated bacterial suspension (after 10 4 times dilution in 5 mM PBS) was spread over each agar plate and allowed to grow for 15 h (37 °C). % CFU reduction was calculated according to equation (2) as explained in Materials and Methods section. As the concentration of CCPE or 4QA-CCOE in the treatment system increased, the E. coli growth on agar plates decreased gradually (i.e. % CFU reduction increased) ( Fig. 5 and Supplementary Figs. S5 and S6). CCPE-treated E. coli showed the highest % CFU reduction, irrespective of CCPE concentration. Using a treatment system containing only 30 μM CCPE, ~90% CFU reduction was achieved which reached up to ~100% when CCPE concentration was ≥50 μM (Fig. 5). This suggested that CCPE can prevent the growth of wild-type E. coli similar to ampicillin if CCPE concentration is ≥50 μM (Supplementary Fig. S7). On the other hand, using 30 μM 2QA-CCOE and 4QA-CCOE in a treatment system, ≤50% CFU reduction of wild-type E. coli (Fig. 5, left) was achieved. In the case of CCOEs, % CFU reduction approached to values close to that of CCPE at higher concentration of CCOEs. This CFU reduction data supported that CCPE based treatment is faster and more effective to inhibit the growth of E. coli than those based on CCOEs with lower concentration of the conjugated molecule. Also, 2QA-CCOE, unlike 4QA-CCOE and CCPE, led to anomalous growth and random values of CFU units rather than a gradual decrease ( Fig. 5 and Supplementary Figs. S5 and S6) with the increase in 2QA-CCOE concentration. This could be attributed to the lower water solubility of 2QA-CCOE (~1 mg/mL) as compared to 4QA-CCOE, and CCPE (~20 mg/mL for both). The numbers of cationic quaternary amine groups for every 3 benzene rings along backbone were 2, 3 and 4 for 2QA-CCOE, CCPE and 4QA-CCOE, respectively. The poor hydrophilicity made it difficult for 2QA-CCOE to consistently interact with the net negatively charged outer cell envelope of E. coli. On the other hand, a large number (~57) of RUs 37 , being wired in a chain, with a suitable charge density made CCPE more favorable to attach to the E. coli surface and inhibit the growth of E. coli more efficiently.
Zeta potential of ccoe/ccpe treated E. coli. During the treatment process, the zeta potential also significantly changed as shown in Fig. 6. The untreated wild-type and amp-resistant E. coli both were net negatively charged with zeta potential of ~−40 mV (Fig. 6) suggesting similar charged nature of this specific E. coli strain after acquiring ampicillin resistance. Treatment with CCPE solution resulted in a large change in zeta potential (+15 mV (wild type, Fig. 6a) or +20 mV (amp-resistant, Fig. 6b)) of bacterial surface. Such a large surface charge reversal was expected based on high % CCPE adsorption on the bacteria surface ( Supplementary Fig. S8). Interestingly, upon treatment with CCOEs, a much lower surface charge reversal of E. coli was observed (~−25 mV for both wild type and amp-resistant strains) at staining concentration ≥50 μM. Such a small change in zeta potential did not scale with the % mass adsorption, especially for 4QA-CCOE (about 65% mass adsorption of 4QA-CCOE ( Supplementary Fig. S8d) was accompanied by a change in zeta potential from −37 mV to −27 mV only (Fig. 6)). We, therefore, believe that there are factors other than mass adsorption which controls the zeta potential (or surface charge) of E. coli treated with CCOE. Such a small increase in the zeta potential of bacteria treated with CCOEs was reported by others 6,21,29,32,33 and attributed to the intercalation of CCOEs within the lipid bilayer of the outer membrane of E. coli. CCOEs with shorter chain length are more prone to membrane insertion 32 and if the intercalation buries the CCOE chains and their cationic groups within lipid bilayer (rather than exposing the charges of CCOEs on the bacterial surface), the surface charge of bacteria may not change significantly despite CCOE adsorption.
The calculated backbone lengths of both 2QA-CCOE and 4QA-CCOE (~1.39 nm) were smaller than the thickness of the lipopolysaccharide layer (~2.8 nm 45 , where the length of the core oligosaccharide 6,45,46 part is ~2.1 nm) of the outer cell membrane of E. coli. Lipid bilayer intercalation is thus more probable with CCOEs. On the other hand, the long-chain CCPEs (~57 RUs) may experience steric hindrance to penetrate into the lipid bilayer and peptidoglycan layer of E. coli 18,23 . Moreover, large charge reversals, similar to our CCPE coated bacteria, have been observed for cell surfaces coated with many other long-chain polyelectrolytes 47,48 . In fact, a similar reversal of surface charge was reported for poly(fluorene) based CCPE coated E. coli 30 . Whitten et al. 18,23 proposed that the biocidal activity of CCPEs may originate not just from their lipid membrane perturbation activity, but also from their interactions with the charged functional groups of lipid bilayer exposed to cell surfaces (contributed by zwitterionic phospholipid head groups; charged groups of outer and inner core oligosaccharides, and, Lipid A) 5,49 . Based on these pieces of evidence, we propose that long-chain CCPE (~50 nm based on length of RU~0.889 nm) are predominantly forming a coating on the outer cell envelope of E. coli. However, the likelihood of minor CCPE chain intercalation (alongside coating as a major mechanism) cannot be ruled out since complete charge reversal was not achieved.
Some of the already established models support our prediction of binding interactions and antimicrobial mechanisms 18,50 . Coating of bacterial outer cell envelope resembles the carpet model 50,51 since the antimicrobial molecules, when following this model, orient themselves parallel to the membrane surface and cover the bacterial surface, like a carpet (instead of inserting into the lipid membrane). On the other hand, CCOE intercalation  (2)). The data represents mean and standard deviation of 3 replicates for each treatment. www.nature.com/scientificreports www.nature.com/scientificreports/ can be explained by toroidal pore model 18 based on which the membrane bound antimicrobial molecules insert into the lipid bilayer and force the outer leaflet to bend continuously to fuse with the inner leaflet of bacterial membrane 18 .
We also examined the effect of incubation time on the zeta potential of treated bacterial cells. A longer treatment time (6 h) did not change the zeta potential of treated bacteria much (Fig. 6(c,d)). Using 100 μM CCPE, ~99% growth inhibition of amp-resistant E. coli was achieved within 1 h (Fig. 4) which reached to 100% in the next 5 h. Additionally, the change in % mass adsorption (98% (1 h, Supplementary Fig. S8b); 98% (6 h, Supplementary Fig. S8d)) and zeta potential (+20 mV (1 h, Fig. 6b); +18 mV (6 h, Fig. 6d)) were minor. The results suggested that the adsorption of CCPE by bacteria was almost complete within 1 h of treatment and sufficient to achieve high % growth inhibition. Interestingly, the zeta potential of 2QA-CCOE treated amp-resistant bacteria reached a plateau by 1 h (~−25 mV after 1 h (Fig. 6b); −28 mV after 6 h (Fig. 6d)), but the bacteria needed the full 6 h-treatment to achieve 98% growth inhibition (growth inhibition was 25% after 1 h treatment (Fig. 4d)). This suggested that growth inhibition of 2QA-CCOE was a time-dependent process. Rapid intercalation (within 100 s, based on epifluorescence micrograph experiments) of CCOEs within the bacterial membrane was reported by Hinks et al. 32 for oligo(phenylenevinylene) with 4 RUs and 4 quaternary amine groups. However, the timescales of their experimental (100 s) and simultaneous molecular dynamics simulation (200 ns) studies on membrane perturbation of Gram-negative bacteria were very small and an understanding of membrane perturbation over large time scale is required. The reason is: to achieve 99% growth inhibition using 2QA-CCOE, cells had to be treated with 2QA-CCOE for a longer time (6 h), but within the first 1 h of this treatment time, adsorption of conjugated molecules by bacteria was complete. These results strongly suggested that after 1 h treatment, the already-attached CCOEs took part in dynamic membrane perturbation processes which continued for the next 5 h of treatment. This led to time-dependent growth inhibition by CCOE treatment. Furthermore, there was no significant difference in the % mass adsorption of CCOE/CCPE on wild-type ( Supplementary Fig. S8a,c) and amp-resistant ( Supplementary Fig. S8b,d) E. coli and the resulting zeta potential (Fig. 6a,c vs 6b,d). These again supported that any specific type of conjugated molecule interacts with wild-type and amp-resistant E. coli (used in this study) in a similar manner.
Morphological changes upon treatment with ccoes/ccpe. Fluorescence microscopy images of green fluorescent protein (GFP)-incorporated amp-resistant E. coli suspension before and after treatment with  Fig. S9). While CCOE-treatment did not change the size of E. coli aggregates ( Supplementary Fig. S9a,b,c,e,f) and the cells were observed to spread out across the frame, bacteria started to aggregate significantly upon interaction with CCPE ( Supplementary Fig. S9d,g). The size of aggregates (~15-20 μM) in CCPE (30 μM) treated E. coli suspension found from fluorescence microscopy image (Supplementary Fig. S9d) was similar to what obtained from SEM image (Fig. 7h). While interpreting the SEM data (Figure 7), we have to keep in mind that we are looking at 3D aggregates as a 2D image. Even though these are 2D images, the contrast can help us to understand the height wise growth of the aggregation. For untreated (Fig. 7a,e) and 30 μM of 2QA-CCOE-treated E. coli cells (Fig. 7b,f), the cells were lying flat like a single layer and spread out throughout the frame. When the cells were treated with 30 μM CCPE, the net negatively charged cells started to stick with each other in a three-dimensional manner due to electrostatic complexation assisted by cationic CCPE (Fig. 7d,h). In this case, the aggregates grew significantly along the third dimension, i.e. perpendicular to the image plane as evident from the image contrast (Fig. 7d,h). When the cells were treated with 4QA-CCOE (30 μM, Fig. 7c,g), cells were aggregated, but the height wise growth was less than the cells treated with CCPE. Cationic CCPE chains, when coated the negatively charged E. coli cells, can attract other uncoated, negatively charged bacteria cells around and form larger aggregates (Fig. 7d,h). Intercalated CCOEs, on the other hand, are likely to conceal the positive charges of CCOEs and minimize the tendency to form bacterial aggregates (Fig. 7b,c,f,g).
Aggregates became larger and clearly grew in three dimensions when the concentration of CCPE in treatment solution increased from 0 μM to 100 μM (Fig. 8a-d for wild-type, 8e-h for amp-resistant E. coli) and seemed to be covered by CCPE molecules. No obvious cell wall rupture or cell lysis was predicted especially for CCOEs since the single cells in both treated (Fig. 7b,c,f,g) and untreated (Fig. 7a,e) E. coli suspensions looked smooth visually. Due to large aggregate formation, the individual cells could not be seen in the case of CCPE treated E. coli which made it difficult to comment about cell wall rupture/lysis based on SEM images. However, no GFP release from treated cells (Supplementary Fig. S9), low amount of dead cells based on live/dead assay (Fig. 3), and unchanged absorbance of treated bacteria (Supplementary Fig. S1) confirm no cell wall rupture/lysis upon treatment with CCPE. These results indicated that phenylene based CCOEs and CCPEs, unlike, phenylene ethynylene based ones 18,23 (showing bacteriolytic activity), acted on the bacterial outer cell membrane more gently, but inhibited further growth. Having said that, the mild alterations at the single-cell level (non-bacteriolytic) could not be differentiated based on the SEM images (due to strong aggregation after treatment). But it is highly likely that mechanical changes are happening to some extent on bacterial outer cell envelope due to these treatments (currently under investigation). Finally, no significant difference in morphology between untreated wild-type (Fig. 7a) and amp-resistant (Fig. 7e) E. coli was observed and so was observed for treated ones (Fig. 7). This was consistent with the indifferent growth inhibition and zeta potential of these two strains.
conclusions. Through synthesizing phenylene based model cationic conjugated oligo-(2QA-CCOE, 4QA-CCOE) and polyelectrolytes (CCPE), we studied the biophysical changes on the outer cell envelopes of wild-type and amp-resistant E. coli strains upon interactions with these conjugated molecules. These conjugated molecules inhibited the growth of both wild-type and amp-resistant E. coli to a similar extent. About 99% growth inhibition was achieved if amp-resistant E. coli was treated for ~3-5 min, 1 h, and 6 h in 100 μM of CCPE, 4QA-CCOE, and 2QA-CCOE solutions, respectively. This indicated CCPE was a faster growth inhibitor of bacteria compared to oligomeric CCOEs. The better inhibitory activity of CCPE could be attributed to an optimum balance of side-chain hydrophilicity and backbone hydrophobicity, good water solubility as well as large chain www.nature.com/scientificreports www.nature.com/scientificreports/ length which made charge interaction of CCPE with net negatively charged bacteria more facile. The large charge reversal of bacteria upon treatment with CCPE suggested that bacterial cells were coated with CCPE chains. On the other hand, the low surface charge reversal suggested that CCOEs were intercalating within the lipid bilayer of the outer membrane of E. coli.

Materials and Methods
Bacteria culture. To make liquid growth media for E. coli, solid LB (Miller, AMRESCO) (25 g) was dissolved in DI water (1 L); while to make agar plates, solid LB (25 g) and agar powder (15 g) were dissolved in DI water (1 L). In both cases, the liquid LB media was autoclaved. Ampicillin (100 μg/mL) was then added to the media as needed. The liquid media was transferred to petri dishes to obtain solid agar plates. The agar plates were stored in the cold room (4 °C) till use. For each experiment, the E. coli strains stored in 15% glycerol were streaked on LB agar plates. The plates were incubated overnight at 37 °C. A single colony was then transferred into 4 mL of liquid LB media in 14 mL BD Falcon round-bottom culture tubes and incubated overnight at 37 °C in a shaker-incubator (250 rpm). Overnight grown cells were harvested by centrifugation (4700 rpm) for 15 min and the pellets were washed twice in 4 mL of 5 mM PBS. The cells were resuspended in 5 mM PBS and the optical density of resuspended cells measured at 600 nm (OD 600 ) was then adjusted as needed using Genesys 10 S UV-Vis spectrophotometer (Thermo Fisher Scientific, Waltham, MA).
Live/dead assay. Live/dead assay of wild-type E. coli was done using LIVE/DEAD BacLight bacterial viability and counting kit (Thermo Fisher Scientific, Waltham, MA) and samples prepared as described in the protocol 52 . Briefly, wild-type E. coli cells were grown for 8 h and OD 600 was adjusted to 0.2 in 5 mM PBS. Cells were treated with CCOE/CCPE (10-50 μM) and incubated at 37 °C, 250 rpm for 1 h. Treated cells were harvested; centrifuged at 10000 × g for 3 minutes; pellets were washed twice with 0.85 wt % NaCl in water and resuspended in 1 mL of 0.85 wt% NaCl solution. Samples were then diluted in two steps: In the first step, the cells were diluted with appropriate volume of 0.85 wt% NaCl solution to achieve OD 600 of bacterial suspension as 0.125; this cell suspension was then further diluted to obtain 1 million cells/mL by adding 10 µL of cell suspension (OD 600 = 0.125) to 987 µL of 0.85% NaCl solution. Finally, 3 µL of mixed SYTO9 and PI (1:1 v/v) was added to the cell suspension and used for flow cytometry. Each treatment had 2 replicates. BD FACSAria II (Franklin Lakes, NJ) at Nebraska Center for Biotechnology at University of Nebraska-Lincoln was then used for flow cytometry studies. BD FACSDiva software was used to analyze the flow cytometry data and obtain % of damaged cells.
Absorbance measurement upon ccoe/ccpe treatment. CCOE/CCPE and cell suspension were added together in 5 mM PBS ([CCOE/CCPE] after adding to PBS = 0-100 μM (based on RUs); absorbance of E. coli at 600 nm ~0.5 (initially); final volume = 250 μL). The suspension was incubated for 0-6 h (37 °C, 250 rpm). The change in absorbance at 600 nm was measured using i3xplate reader (Molecular Devices, San Jose, CA). The study was performed for both wild-type (DH10B) and amp-resistant E. coli (SSBIO002) (as shown in Supplementary  Fig. S1)). This procedure was followed to replicate the data three times.

Measurement of percent (%) growth inhibition.
The growth inhibition was studied in three ways: "treat and grow (LB)" (1-step-process); "treat (PBS), then grow (LB)" (2-step-process); and "treat (LB), then grow (LB)" (2-step-process). For the 1-step process, the CCOE/CCPE and cell suspension were added together in LB media ([CCOE/CCPE] after adding to LB = 0-100 μM (based on RUs); OD 600, initial = 0.2; final volume = 250 μL) and the cells were allowed to grow here for 0-6 h (37 °C, 250 rpm). For the 2-step processes, the procedure was similar to www.nature.com/scientificreports www.nature.com/scientificreports/ what described in the literature 22 . At first, the E. coli strains (OD 600, initial = 0.2) were treated with CCOE/CCPE for 0-6 h (37 °C, 250 rpm) in PBS (for "treat (PBS), then grow (LB)" process) or LB (for "treat (LB), then grow (LB)" process). After that, 50 μL of this treated cell suspension was transferred to 200 μL of LB media in black 96-well plates and allowed to grow for 3 h (37 °C, 250 rpm). The percent (%) growth inhibition of E. coli in LB media was calculated using the Eq. (1): Each absorbance value was background subtracted where the background was the absorbance of the sample at time = 0 (i.e. just after adding to LB media). colony-forming unit (cfU) reduction assay. The protocol described in the literature was used for CFU reduction studies 25 . Briefly, the bacterial strains grown in liquid LB media were used and OD 600 was adjusted to 0.2 by dilution with 5 mM PBS buffer. 180 μL of this E. coli suspension was transferred to black 96-well plates to which appropriate volume of a type of conjugated molecule (i.e., 2QA-CCOE, 4QA-CCOE, or CCPE) was also added. The final concentration of CCOEs or CCPE in this pretreatment step varied between 0 and 100 μM (based on RUs). The treated bacterial cells were diluted 10 4 times using 5 mM PBS. 5 μL of this diluted bacterial suspension was then spread on top of the agar plates (3 replicates for each treatment) and allowed to grow in dark for 15 h at 37 °C. The number of CFUs was then counted and CFU reduction (%) was calculated using the Eq. (2): Plain LB and LB-supplemented with ampicillin (100 μg/mL) were used to grow the wild-type and amp-resistant strains, respectively. The CFUs for untreated amp-resistant E. coli grown on an agar plate with and without ampicillin were almost similar ( Supplementary Fig. S10). This confirmed the effectiveness of colony selection process and growth of amp-resistant E. coli only (i.e. no growth of wild-type E. coli). Also, the wild-type and amp-resistant E. coli were grown on agar plates made of LB supplemented with ampicillin ( Supplementary  Fig. S11). While the amp-resistant E. coli grew on agar plates, there was no colony formation for wild-type E. coli. This confirmed that just amp-resistant E. coli was able to grow. Zeta potential measurement. Both wild-type and amp-resistant E. coli samples were treated with CCOEs or CCPE (as mentioned in prior sections) after which the cells were harvested by centrifugation (4700 rpm, 15 min). The pellets were washed twice with DI water (2 ml per wash) and then resuspended in 2 mL DI water for subsequent zeta potential measurement using Zeta PALS Zeta Potential Analyzer (Brookhaven Holtsville, NY).
Scanning electron microscope (SeM) imaging. For the investigation of the morphology of untreated and CCPE/CCOE treated E. coli, scanning electron microscope (SEM; Nova NanoSEM450, FEI, Hillsboro, OR) was performed at Nebraska Center for Materials and Nanoscience. 900 μL of bacterial suspension (OD 600 = 0.2) and different concentrations of CCPE or CCOEs (0-100 μM) were mixed and incubated for 1 h in culture tubes (37 °C, 250 rpm). The treated cells were harvested by centrifugation (4700 rpm, 15 min) and pellets were washed twice (with 1 mL of DI water per wash) and resuspended in 1 mL DI water. 3 μL of this suspension was added on top of a silicon wafer (0.5 cm × 0.5 cm) and allowed to dry at room temperature for 1 h. Dried samples were fixed with 0.5 vol% glutaraldehyde in 5 mM PBS buffer and kept at room temperature for another hour. This was followed by a second fixing using 1 vol% glutaraldehyde in 5 mM PBS and samples were kept at room temperature for 4 h. Fixed cells were washed thrice with DI water and dehydrated sequentially using 20, 30, 50, 70, 90 and 100 vol% of ethanol in water. Finally, the specimens were coated with gold prior to SEM measurement.

Data availability
All data generated and analyzed in this study are included in main text or Additional Information.