Nanomechanical studies of the interaction of antimicrobial peptides with bacterial cells

© 2016 Dr. Anna MularskiAntimicrobial peptides (AMPs) are promising therapeutic alternatives to conventional antibiotics. Many AMPs are membrane-active but their mode of action in killing bacteria or in inhibiting their growth remains elusive. Recent studies indicate the mechanism of action depends on peptide structure and lipid components of the bacterial cell membrane. To date, most of these studies have been conducted with synthetic membrane systems, which neglect the possible role of bacterial surface structures in these interactions. Here, we use atomic force microscopy (AFM) to study the interactions of three different AMPs on living Klebsiella pneumoniae bacterial cells. K. pneumoniae are recognized as a serious cause of nosocomial infections and a source of shared antimicrobial resistance. Typically, K. pneumoniae assemble a polysaccharide layer (known as the capsule) on the outside of the cell envelope, taking the form of a hydrated polyelectrolyte network that can grow up to several hundred nanometers thick, and can provide protection from environmental stresses, including desiccation, detergents, antibiotics and host immune defenses. It also plays a role in bacterial adherence to surfaces (such as medical devices) through the formation of biofilms. In situ biophysical measurements were performed to understand how the antimicrobial modulate various biophysical behaviours of individual bacteria, including the turgor pressure, cell wall elasticity, and bacterial capsule thickness and organisation. Firstly, the interactions of a melittin-derived lytic peptide with live K. pneumoniae cells were studied revealing that exposure to the peptide had a significant effect on the turgor pressure and Young’s modulus of the cell. The turgor pressure increased upon peptide addition followed by a later decrease, suggesting that cell lysis occurred and pressure was lost through destruction of the cell envelope. The Young’s modulus also increased, indicating that interaction with the peptide increased the rigidity of the cell wall. The bacterial capsule appeared unaffected by exposure to the peptide and microbiological assays determined that the presence of the capsule conferred no advantage to the wild-type over capsule deficient cells when exposed to the peptide. These findings suggest that even though the long-range electrostatic attraction between the peptide and capsule may first attract the peptide to the bacterial cell, it is the drive of the peptide to associate with the membrane that dominates once the peptide approaches the bacterial outer envelope. The methodology developed was then applied to study the interaction of the polypeptide antibiotic, colistin sulfate, with both wild-type and colistin-resistant K. pneumoniae. The capsular polysaccharides of wild-type cells were rearranged after exposure to colistin, particularly at lysis inducing concentrations, where the capsule appears to unravel. Colistin resistant mutant cells did not display the same behaviour. Microbiological assays showed that the presence of the capsule confers no advantage for wild-type over the capsule deficient cells. Genetic sequencing of the colistin resistant mutant showed that it’s ten-fold resistance to colistin is conferred by a reduction in net negative charge of the lipopolysaccharide molecules embedded in the outer membrane. Negative charge reduction could hinder the initial electrostatic binding of colistin to the outer membrane and the displacement of the divalent cations that function to bridge adjacent LPS molecules throughout the capsular polysaccharide network. Retention of the cross-linking divalent cations may explain why capsule thickness measurements did not increase in the colistin-resistant strain after colistin exposure. These results contrast with those for other K. pneumoniae strains that suggest the capsule confers colistin resistance. Finally, the interaction of caerin 1.1 with K. pneumoniae cells was studied. Caerin is a peptide similar in structure and size to melittin. As with the melittin derived peptide, the capsule of K. pneumoniae AJ218 was unchanged by exposure to caerin, indicating once again that the ionic interaction of the positively charged peptide with the negatively charged capsular polysaccharide is not a critical component of AMP interaction with K. pneumoniae AJ218 cells, the presence of a capsule conferring no advantage to wild-type over capsule-deficient cells when exposed to the AMP caerin. Time-resolved AFM images revealed that the antimicrobial peptide (AMP) caerin 1.1 caused localised defects in the cell walls of lysed Klebsiella pneumoniae cells, corroborating a pore-forming mechanism of action. The defects continued to grow during the AFM experiment, in corroboration with large holes that were visualised by scanning electron microscopy. Defects in cytoplasmic membranes were visualised by cryo-EM using the same peptide concentration as in the AFM experiments. At three times the minimum inhibitory concentration of caerin, ‘pores’ were apparent in the outer membrane. The three studies presented here demonstrate that AFM measurements can be used to track changes in bacterial cells as they lyse due to peptide interaction. In all three studies, the presence capsular polysaccharides conferred no advantage to wild-type over capsule-deficient K. pneumoniae, suggesting ionic interaction between the peptides and bacteria-bound capsular polysaccharides is not a key component of the interaction with the K. pneumoniae AJ218. Finally, AFM and cryo-EM data have shown these three peptides operate via different mechanisms of action with K. pneumoniae