Exogenous fatty acid growth characteristics in E. coli
In Gram-negative bacteria, several studies have reported that supplementation of growth media with unsaturated fatty acids augments growth in logarithmic and stationary phases [1,2,4,19]. When E. coli was grown in CM9 in the presence of 300 µM of individual PUFAs, elevated growth (as compared with control) was observed over the course of 12 hours (Figure 1A). Between 6-12 h, three fatty acids (18:2, 18:3a, and 18:3g) correlated with statistically significant (p < 0.04) higher growth, as determined by a two-tailed, two sample equal variance t-test.
To evaluate the growth of E. coli on fatty acids as a sole carbon source, the bacteria were supplemented with 1mM of each fatty acid in M9 minimal media (no glucose) and growth was measured for 12 hours. Linoleic acid and docosahexaenoic acid caused noticeable growth above other fatty acids, while dihomo-gamma-linolenic acid caused appreciable growth in the latter hours only (Figure 1B). Viability and purity of cultures was confirmed by plating for colonial growth at hour 12. CFU determination was performed at hour 7. All fatty acid-supplemented samples yielded higher CFU/ml (Supplemental Table 1).
E. coli incorporates exogenous PUFAs into its phospholipids
The membrane phospholipid incorporation of exogenous fatty acids was analyzed first by thin-layer chromatography of extracted phospholipids from E. coli grown to logarithmic phase in the presence of 300 µM of each fatty acid. The migration of the major bacterial phospholipids phosphatidylethanolamine (PE), phosphatidyglycerol (PG), and cardiolipin (CL) provided little, if any, qualitative data regarding incorporation of PUFAs (Figure 2). The spots at the top of the plate reflect free fatty acids that were not washed from the bacteria during the Bligh and Dyer extraction and thus represent fatty acids associated with the bacteria and/or imported but not (yet) assimilated.
For confirmation of fatty acid incorporation into E. coli phospholipids, negative ionization UPLC/ESI-MS was performed on total lipid extracts. The 50 V sampling cone of the mass spectrometer is of sufficient energy to cause minor in-source fragmentation, thus allowing simultaneous observation of [M – H]– parent ions and their cone fragments in the mass spectra of each chromatographically resolved component. Since the most prominent cone fragments of phospholipids consist of the carboxylate substituents attached to sn-1 and sn-2 positions of the glycerol backbone (Hsu and Turk 2000; Hsu and Turk 2001), in-source cone fragmentation allows unambiguous determination of phospholipid composition by directly observing the incorporated fatty acid residues.
All PUFAs, with the exception of 22:6, were identified as being incorporated at the sn2 position of PE and PG. For example, phospholipids extracted from cell cultures grown in the presence of 20:3 (Figure 3) had chromatographic signals corresponding to PG(16:0/20:3), PG(18:1/20:3), PE(16:1/20:3), PE(16:0/20:3), and PE(18:1/20:3), whereas the control did not show such signals. The relative incorporation of all PUFAs, except 22:6, is portrayed in Supplemental Figure 1.
Membrane permeability is altered by exogenous fatty acids
To evaluate the ramifications of newly adopted phospholipid species in the E. coli membrane, we next examined the effect on membrane permeability. First, a crystal violet uptake assay indicated that arachidonic acid and docosahexaenoic acid elicited significant increases (50% and 25%, respectively) in permeability as compared to the control and other fatty acids (Figure 4A). A second evaluation of permeability was performed using ethidium bromide (EtBr) to assess both uptake and accumulation over time. The uptake assay (Figure 4B) largely correlated with the accumulation assay (Figure 4C), which measures the fluorescence intensity of live cells (increasing as EtBr binds to DNA). All PUFAs increased membrane permeability. The only fatty acids to display a different trend were 18:3g and 20:3, which excluded EtBr yet displayed lower accumulation (Figure 4B&C). Despite the differential uptake between dyes, these assays illustrate the varied and significant impact of exogenous fatty acids on membrane permeability.
Exogenous fatty acids impact antimicrobial peptide resistance
The observed alteration to membrane permeability led to an investigation of the impact of fatty acids on antimicrobial peptide susceptibility, a phenomenon previous documented for several Gram-negative bacteria [1-4]. For these experiments, two antimicrobials (polymyxin B and colistin) with mechanisms of action that target membrane bilayers via biophysical intercalation were chosen to compare with an antibiotic (ampicillin) that relies on protein-mediated uptake for activity. In the minimum inhibitory concentration (MIC) assays, bacteria were pre-adapted with fatty acids prior to the assay, where fatty acids were also made available at 300 µM during exposure to two-fold concentrations of each antimicrobial. The availability of three fatty acids (20:3, 20:4, and 22:6) increase the MIC of E. coli against polymyxin B and colistin (Figure 5A&B). Strikingly, arachidonic acid raised the MIC to polymyxin B 8-fold compared to the no fatty acid control. Eicosapentaenoic acid (20:5) was the only fatty acid to lower the MIC to polymyxin B. Both arachidonic acid and docosahexaenoic acid increased the MIC to colistin by 4-fold. Minimal differences were observed for the MIC of ampicillin, although 20:4 and 18:3g-treated cultures maintained better survival at lower antibiotic concentrations (Figure 5C).
Exogenous PUFAs affect phenotypes associated with virulence
Bacterial motility and biofilm formation are phenotypes known to influence pathogenesis. Since previous studies have identified fatty acid-induced alterations in other Gram-negative bacteria for these phenotypes, this study also investigated swimming motility and biofilm formation in E. coli. The availability of 300 µM fatty acid in a soft agar assay resulted in increased motility (»10%) with 18:2 and decreased motility (»10%) for the 20-carbon fatty acids tested (20:3, 20:4, 20:5) (Figure 6). More striking differences were observed for biofilm formation, where supplementation of 18:3a, 18:3g, 20:4, and 22:6 significantly increased the amount of biofilm when the assay was performed in CM9 minimal media (Figure 7). In particular, biofilm formation was doubled with 22:6 and tripled with 18:3g and 20:4. When the assay was performed in LB, 18:3g and 20:3 elicited significant increases in biofilm formation, whereas 18:3a, 20:4, and 20:5 decreased biofilm formation to a near identical degree. Collectively, the assessment of phenotypes associated with virulence revealed several exogenous fatty acid-mediated impacts on swimming motility and biofilm formation.