Tuning cell surface charge in E. coli with conjugated oligoelectrolytes

Conjugated oligoelectrolytes intercalate into and associate with membranes, thereby changing the surface charge of microbes, as determined by zeta potential measurements.


Introduction
Although the manipulation of microbial cell properties offers the potential for harnessing and tuning the abilities of microorganisms, it remains a signicant challenge due to the aqueous environment and overall structural complexity and diversity. 1Genetic engineering, while effective, is limited to materials the cell itself is capable of producing.Synthetic materials and molecular systems offer possible functionalities that are not encountered in nature.With this in mind, conjugated oligoelectrolytes (COEs) are synthetic molecules generally characterized by 3-5 p-conjugated repeat units (RUs) equipped with pendant ionic groups to impart solubility in polar media.4][15][16][17] As such, a variety of COEs have found utility in bioimaging [18][19][20][21][22][23][24] and biological detection schemes [25][26][27][28][29][30] of their own.
A distinct subset of COEs, and that used in this contribution, is distinguished by ionic functionalities tethered at the two terminal ends of a phenylenevinylene sequence.These bolaamphiphilic structures, in particular DSBN+ and DSSN+ (Chart 1), have been shown to spontaneously intercalate into lipid bilayers with a concomitant increase in uorescence quantum yield. 14Polarized confocal microscopy has been used to demonstrate a preferential alignment of the COE's molecular long axis relative to the membrane plane.They have also been implicated in boosting the performance of a variety of microbial electronic devices 31,32 employing organisms ranging from yeast, 14 to E. coli 33,34 and Shewanella, 35,36 and even naturally occurring bacteria in wastewater. 37Although the exact mechanism of their action is still unclear, 36,38,39 it is thought that the COEs' ability to intercalate into microbial membranes is paramount for linking intracellular metabolism to extracellular electrodes in these devices. 40hile the lipid membrane intercalation of COEs is welldocumented, other biological interactions of COEs and their consequences have not yet been studied.Previously we showed that an anionic COE analogous to DSSN+ was prevented from incorporating into E. coli membranes most likely due to electrostatic repulsion from the innate negative surface charge of the cells. 40These negative charges occur mostly as ionized carboxyl and phosphate groups that are part of lipopolysaccharide (LPS) macromolecules composing the outer leaet of most Gram-negative bacteria. 41,42][45] Furthermore, in studies concerning the effects of COEs on biological systems, COE concentrations are chosen in the low micromolar regime with no consideration given to the total number of cells; the amount of COE that associates with each cell and that which is le in solution remains to be quantied.With this purpose, we compare 8 COEs varying in molecular length and core substitutions for their association with E. coli and effect on cell zeta potential, nding a remarkable length dependence on these properties.

Chemical structures
The chemical structures of the COEs used in this study are shown in Chart 1; their syntheses have been described in the literature. 14,34,46Their basic structure can be described by 3-5 phenylenevinylene repeat units (RUs) anked on both ends by either an amine (COE1 series) or two meta-positioned alkoxy (COE2 series) linkages carrying trimethylammonium iodide terminated hexyl chains.Tetrauorine substitution of the center phenylene ring of the 3-RU molecules offers variance of the central hydrophobic core to determine its role, if any, in cell association and cell surface charge.

Confocal microscopy
In order to rst visualize how each COE interacts with E. coli, we exploited the photoluminescent p-conjugated core of the molecules for uorescence microscopy.Cells were stained with 10 mM solutions of COE for 1 hour and imaged with a laser scanning confocal microscope, the results of which are shown in Fig. 1 (bright eld images are shown in Fig. S1 †).As anticipated based on the bolaamphiphilic structure shared by the molecules, all COEs display an emission pattern around the edges of cells consistent with membrane intercalation.In this regard, the substitution of alkoxy pendant linkages for amine or the addition of 4 uorine atoms to the center phenyl ring of the 3-RU COEs provides no discernable difference in terms of observable cell localization in E. coli.

Cell association studies
Taking advantage of the strong visible light absorbing properties provided by the conjugated core of the molecules, 14 the amount of each COE that associates with E. coli in solution was quantied.Full experimental details are given in the Experimental section.Briey, cells (OD 600 nm ¼ 1.0) were stained in different concentrations of COE ranging from 1-40 mM for 1 hour in 50 mM phosphate buffered saline (PBS) solutions.All concentrations of COE used in this study were less than the critical aggregation concentration (CAC) reported for DSBN+, which is at 0.51 mM. 47A staining time of 1 hour was found sufficient to establish equilibrium within these experimental conditions (Fig. S2 †).The cells were then centrifuged and the supernatant analysed by UV-vis absorption to determine the amount of COE le in solution (i.e.not associated with the pelleted cells).This method is illustrated in Fig. 2 for 10 mM and 20 mM DSSN+.Comparing the control spectra of the solutions containing just DSSN+ in PBS (solid lines) to the spectra of the supernatants resulting from cell staining, one observes that at 10 mM, no discernable DSSN+ is le in solution, meaning that all COE has associated with the cells.In contrast, at 20 mM a signicant absorption is observed indicating that some DSSN+ remains in the solution and did not associate with the E. coli.
In subsequent experiments, the amount of COE associated with cells using the UV-vis absorption method was quantied by subtracting the absorbance of the supernatant of stained and centrifuged E. coli at a wavelength of 420 nm (COE1 series) or 380 nm (COE2 series) from control samples that did not contain cells.Fig. 3 shows the trends in COE/cell association for the unuorinated COEs at different staining concentrations normalized to 1 OD 600 nm of cells.Interestingly, at concentrations between 1-15 mM for all 6 COEs, 100% association is  observed resulting in a linear increase in COE association with increasing staining concentration, reaching $15 nmol/OD 600 nm associated at 15 mM staining concentration.Looking at the COE1 series in Fig. 3A, at concentrations >15 mM the 4-and 5-RU COEs, DSSN+ and COE1-5C, reach a maximum association of $20 AE 0.4 nmol/OD 600 nm and $25 AE 1.0 nmol/OD 600 nm respectively.In contrast, the 3 RU COE, DSBN+, does not reach a plateau and attains a maximum association of 34 AE 0.2 nmol/OD 600 nm at 40 mM staining concentration.It should be noted that for COEs with minimum inhibitory concentration (MIC) data published (DSBN+ and DSSN+), the MICs (normalized to cell count) required to reduce growth of E. coli are 2 orders of magnitude higher than the concentrations used in this study. 13,48Moreover, it is pointed out in another study, where cytotoxicity tests on E. coli with 20 mM of all COE1 series, that no toxicity phenomena is observed in colony forming units (CFUs). 34 similar trend is observed for the COE2 series in Fig. 3B with maximum associations of 37 AE 2.4, 18 AE 0.5 20 AE 1.1 nmol/ OD 600 nm for the 3-, 4-and 5-RU COEs, respectively.When comparing the two series of COEs, the 3-RU COE2 series shows slightly greater maximum association than the 3-RU COE1 series DSBN+, suggesting that the structural modication afforded by the alkoxy pendant linkages provide a modest advantage in this respect.However, the comparison between the 4-and 5-RU COEs shows a slightly higher maximum association in the COE1 series than the COE2 series.Interestingly, previous cytotoxicity tests on E. coli with 20 mM of all COE2 series showed no toxicity for COE2-3C, while COE2-4C and COE2-5C have demonstrated a $30% loss in CFUs than controls.34 Regardless of series type, there is a clear dependence of COE association with E. coli on molecular length: the amount able to associate with cells for the 4-RU and 5-RU COEs plateaus within the concentration range tested, while the 3-RU COEs do not.
On the secondary y-axes in Fig. 3 are the estimated number of COE molecules associated per cell at each staining concentration, with 1 OD 600 nm corresponding to a concentration of 10 9 cells per mL. 49With this estimate, it can be seen that maximum COE associations observed in these experiments are greater than 10 7 molecules per cell for 4-and 5-RU COEs and greater than 2 Â 10 7 for both 3-RU COEs.When comparing these numbers to an estimate of the number of lipids per E. coli cell 50 of $2.2 Â 10 7 one can see that the 4-and 5-RU COEs would approach a 1 : 1 lipid : COE ratio in cells and the 3-RU COEs surpass this threshold at the 40 mM staining concentration.As discussed in the Introduction, much evidence has been presented that COEs intercalate into microbial membranes, and up until this point, this has been the only interaction considered.With ratios at or above 1 : 1 lipid : COE per cell, which would be morphologically impossible, it is obvious that not all of the associated COE is intercalating into lipid bilayers.A plausible hypothesis is that some COE is associating with the outside of the E. coli, which, with its net negative charge, 51 is a likely candidate for electrostatic interaction with positively charged molecules. 45,52,53ta potential measurements In order to determine the effect of COE association on cell surface charge, stained E. coli cells from the previous experiment were washed and resuspended in PBS buffer for zeta potential measurements, 51 the results of which are shown in Fig. 4. Unstained cells were found to have an average zeta potential of about À16 mV under these conditions, indicating a net negative charge, as expected. 52The cells stained with COE1 series follow a trend of increasing zeta potential to more positive values as the staining concentration increases.Maximum zeta potential values of À13.1 AE 0.4, À7.8 AE 1.1, and À2.4 AE 0.4 mV are reached for the 3-, 4-, and 5-RU COEs, respectively, trending more positive with increasing molecular length.In addition, zeta potential values reect the association trends observed in Fig. 3A, in that the 4-and 5-RU COEs reach a plateau at a staining concentration around the same concentration that the cell association for these COEs plateaus.Despite having the highest maximum cell association of the COE1 series, the 3-RU COE causes the least change in zeta potential but reects the association trend in Fig. 3A in that the zeta potential does not appear to plateau in the concentration range tested.The effect on E. coli zeta potential of the COE2 series is shown in Fig. 4B.The COE2 series displays a similar length dependence with maximum zeta potential values for E. coli of À13.8 AE 1.1, À11.3 AE 0.6, and À2.9 AE 0.7, observed for COE2-3C, COE2-4C and COE2-5C, respectively.Cells stained by the 3and 4-RU COE2 molecules display noticeably less positive zeta potential values than their COE1 counterparts but ultimately a similar trend follows in that cells stained by longer COEs result in more positive zeta potential values.Ultimately the change from amine to alkoxy linked pendant groups has only a minor inuence on the COE zeta potential effects as a whole.
Rather than observing charge reversal towards high positive values as is seen with cells being coated with positively charged polyelectrolytes, 43,44,52 the trend towards charge neutralization in this experiment suggests that not many of the COE positive charges are extending beyond the LPS.COEs are much smaller in size than polyelectrolytes and easily intercalate into lipid membranes and perhaps also 'interdigitate' with the oligomeric sugars that form the core of LPS rather than coating the outside cells.In fact, this non-lipid portion of LPS in E. coli K12 is estimated to be $2.1 nm in length. 54,55This length is slightly longer than the 3-RU phenylenevinylene core and slightly shorter than the 4-RU conjugated core, which are estimated to be 1.8 nm and 2.4 nm respectively.With the 5-RU core estimated to be around 3 nm, one can begin to rationalize the length scales with the zeta potential results.More specically, the 4-and 5-RU COEs have a greater chance of spanning the full length or even extending past the outermost LPS units than do the 3-RU COEs, possibly explaining the molecular length dependence of the zeta potential results.

Fluorinated derivatives
Lastly, cell association and zeta potential experiments were carried out with the uorine-substituted 3-RU COEs (4FCOEs), the results of which are plotted with the unsubstituted counterparts for comparison and are shown in Fig. 5. Cell association for the 4FCOEs (Fig. 5A) is largely indistinguishable from their unsubstituted counterparts until staining concentrations of $25-40 mM, at which point the 4FCOEs associate slightly less.At the highest staining concentration tested (40 mM), there were approximately 2.0 (AE0.06)Â 10 7 and 2.4 (AE0.02)Â 10 7  molecules associated per cell for 4F-DSBN+ and 4F-COE2-3C, respectively.These values are 23% and less than for DSBN+ and COE2-3C, respectively.A possible explanation for this deviation at higher staining concentrations is the polar-hydrophobic nature of uorinated compounds, 56 making these molecules less likely to aggregate in the lipid membrane due to interactions between the cationic pendant groups and the uorinated core. 13Being less likely to aggregate or pack closely would result in less overall cell association.It is worth noting, however, that aggregation of COEs in a lipid membrane has yet to be experimentally proven.
The zeta potential of E. coli stained with the 4FCOEs (Fig. 5B) follows the same trend as the unuorinated COEs, in that a gradual increase in zeta potential is observed as staining concentration increases.Cells stained with 4F-DSBN+ reach a more positive maximum (À14.8AE 0.6 mV) than those stained with 4F-COE2-3C (À15.4AE 0.6 mV), with both maxima being slightly less positive than the corresponding unuorinated COEs at À13.1 AE 0.4 mV and À13.8 AE 1.1 mV, respectively.Ultimately, uorine substitution of the center ring of 3-RU COEs has minimal inuence on cell association and zeta potential of stained E. coli.

Conclusions
In conclusion, 8 COEs varying in length and substitutions to the aromatic core have been compared in terms of their association with E. coli and their effect on cell zeta potential.Confocal microscopy showed patterns consistent with lipid membrane association for all COEs.At low staining concentrations (<20 mM) nearly 100% of COE in solution associates with cells, leaving none remaining in the supernatant of centrifuged samples.At higher concentrations, 3-RU COEs continue to associate while 4-and 5-RU COEs plateau, reaching a maximum association that cannot be overcome by adding more COE to the staining solution.The 3-RU COEs associate past a 1 : 1 lipid : COE ratio while the 4-and 5-RU COEs approach it, which is morphologically impossible and indicative of cellular association not exclusive to membrane intercalation.Cells stained with COEs generally showed more positive zeta potential values with increasing staining concentration, indicating a neutralization of anionic charges of the LPS by the cationic charges of the COEs.Additionally, more positive zeta potential values were observed for longer COEs suggesting that they are able to extend beyond the negatively charged molecular constructs of the E. coli LPS.The other structural variations presented here, namely amine vs. alkoxy pendant linkages and uorination of the aromatic core, proved less important than molecular length, as they had minimal effects on cell association and zeta potential, when compared to analogues with the same number of repeat units.These changes alter the photophysical properties of the molecules and thus increase the number of COEs available for applications in bioimaging 19,20,[57][58][59] and optoelectronics. 60,61[64][65][66][67][68] Experimental Materials All materials were used as received and purchased from Sigma-Aldrich or Fisher Scientic unless otherwise noted.

Cell staining for microscopy
Before staining, E. coli was rinsed twice from the growth medium with phosphate buffered saline (PBS) containing the following: 45.7 mM NaCl, 0.9 mM KCl, 3.3 mM Na 2 HPO 4 and 0.6 mM KH 2 PO 4 at pH 7.4.0.5 mL of OD 600 ¼ 0.9 cells were stained with 10 mM COE for 1 hour in the dark at room temperature before rinsing twice.Samples were then resuspended in 100 mL of PBS and 5 mL were dropped onto a clean glass slide and a cover slip placed on top.Cover slips were sealed with clear nail polish and all samples were imaged within 2 hours.

Confocal microscopy
All images were obtained via laser scanning confocal microscopy using an Olympus FluoView 1000S spectral scanning microscope equipped with a 60 Â 1.30 silicon oil immersion lens.A 405 nm laser was used as the excitation source.For the COE1 series, emission was collected from 480 nm-580 nm.For the COE2 series, emission was collected from 410 nm-510 nm.All images were processed using ImageJ.

COE cell association experiments
E. coli cells at OD 600 nm ¼ 1.0 were stained in clear 96-well plates (BD Biosciences, San Jose, CA) at 20 C for 1 hour in the dark with shaking.Total volume of each sample was 200 mL and samples were measured in triplicate.Aer centrifugation of the plate (3500 rpm, 4 minutes), 100 mL of supernatant was transferred to a clean well for UV-vis absorption with a Tecan M220 Innite Pro plate reader (Tecan, Männedorf, Switzerland).Absorbance was measured at 420 nm for COE1 series and 380 nm for COE2 series molecules.Control samples with no cells were treated the same and their absorbance values represented the total COE from which the supernatant values were subtracted to give the amount associated with cells.

Zeta potential measurements
Stained, twice-rinsed cells were resuspended in PBS to their original OD 600 nm ¼ 1.0.100 mL of each sample was diluted into 900 mL PBS for zeta potential measurements on a Malvern Zetasizer Nano ZS (Malvern Instruments, Malvern, U.K.) at 20 C. Data points given are an average of 4 biological replicates with 3 measurements each.

Fig. 1
Fig. 1 Laser scanning confocal micrographs of E. coli stained with 10 mM COE in PBS for 1 hour.Top row is COE1 series (green), bottom row is COE2 series (blue).Excitation wavelength was 405 nm for all images.5 mm scale bar is the same for all images.

Fig. 2
Fig.2UV-vis absorption of 20 mM (blue, solid) and 10 mM (red, solid) DSSN+ in PBS.After staining E. coli (OD 600 nm ¼ 0.9) for 1 hour with these concentrations of DSSN+, the cells are centrifuged and the DSSN+ remaining in the supernatant (dashed lines) is measured in order to determine how much COE associates with cells.

Fig. 3
Fig.3COE associated with E. coli cells as a function of staining concentration for (A) COE1 series and (B) COE2 series molecules.The amount of COE associated was calculated by subtracting the absorbance at 420 nm (COE1) or 380 nm (COE2) of the supernatant after centrifugation from that of a control staining solution with no cells.Approximate number of cells assuming 1 OD 600 nm ¼ 10 9 cells per mL.

Fig. 4
Fig. 4 Zeta potential measurements of E. coli cells as a function of COE staining concentration for (A) COE1 series and (B) COE2 series.Dashed line represents the zeta potential of unstained E. coli.

Fig. 5
Fig. 5 Comparing 3-ring COEs with fluorine substitution (dashed lines, open symbols) and without (solid lines, closed symbols).(A) COE associated with E. coli as a function of staining concentration.Approximate number of cells assuming 1 OD 600 nm ¼ 10 9 cells per mL.(B) Zeta potential measurements of stained E. coli as a function of COE staining concentration.Black dashed line represents the zeta potential of unstained E. coli.