Mechanisms of Selective Antimicrobial Activity of Gaegurin 4

Gaegurin 4 (GGN4), an antimicrobial peptide isolated from a Korean frog, is five times more potent against Gram-positive than Gram-negative bacteria, but has little hemolytic activity. To understand the mechanism of such cell selectivity, we examined GGN4-induced K+ efflux from target cells, and membrane conductances in planar lipid bilayers. The K+ efflux from Gram-positive M. luteus (2.5microgram/ml) was faster and larger than that from Gram-negative E. coli (75microgram/ml), while that from RBC was negligible even at higher concentration (100microgram/ml). GGN4 induced larger conductances in the planar bilayers which were formed with lipids extracted from Gram-positive B. subtilis than in those from E. coli (p<0.01), however, the effects of GGN4 were not selective in the bilayers formed with lipids from E. coli and red blood cells. Addition of an acidic phospholipid, phosphatidylserine to planar bilayers increased the GGN4-induced membrane conductance (p<0.05), but addition of phosphatidylcholine or cholesterol reduced it (p<0.05). Transmission electron microscopy revealed that GGN4 induced pore-like damages in M. luteus and dis-layering damages on the outer wall of E. coli. Taken together, the present results indicate that the selectivity of GGN4 toward Gram-positive over Gram-negative bacteria is due to negative surface charges, and interaction of GGN4 with outer walls. The selectivity toward bacteria over RBC is due to the presence of phosphatidylcholine and cholesterol, and the trans-bilayer lipid asymmetry in RBC. The results suggest that design of selective antimicrobial peptides should be based on the composition and topology of membrane lipids in the target cells.

Increased activity of large conductance Ca2+-activated K+ channels in negatively-charged lipid membranes

The effects of membrane surface charge originated from lipid head groups on ion channels were tested by analyzing the activity of single large conductance Ca2+-activated K+ (maxi K) channel from rat skeletal muscle. The conductances and open-state probability (Po) of single maxi K channels were compared in three types of planar lipid bilayers formed from a neutral phosphatidyledianolamine (PE) or two negatively-charged phospholipids, phosphatidylserine (PS) and phosphatidylinositol (PI). Under symmetrical KCl concentrations (3 apprx 1,000 mM), single channel conductances of maxi K channels in charged membranes were 1.1 apprx 1.7 times larger than those in PE membranes, and the differences were more pronounced at the lower ionic strength. The average slope conductances at 100 mM KCl were 251 +/- 9.9, 360 +/- 8.7 and 356 +/- 12.4 (mean +/- SEM) pS in PE, PS and PI membranes respectively. The potentials at which Po was 1/2, appeared to have shifted left by 40 mV along voltage axis in the membranes formed with PS or PI. Such shift was consistently seen at pCa 5, 4.5, 4 and 3.5. Estimation of the effect of surface charge from these data indicated that maxi K channels sensed the surface potentials at a distance of 8 apprx 9 ANG from the membrane surface. In addition, similar insulation distance (7 apprx 9 ANG) of channel mouth from the bilayer surface charge was predicted by a 3-barrier-2-site model of energy profile for the permeation of K+ ions. In conclusion, despite the differences in structure and fluidity of phospholipids in bilayers, the activities of maxi K channels in two charged membranes composed of PS or PI were strikingly similar and larger than those in bilayers of PE. These results suggest that the enhancement of conductance and Po of maxi channels is mostly due to negative charges in the phospholipid head groups.

Dual action of d-tubocurarine on large-conductance Ca2+-activated K+ channels from rat brain reconstituted into planar lipid bilayer

Using the planar lipid bilayer method, we investigated the effect of d-tubocurarine (dTC) on the extracellular side of large-conductance Ca2+-activated K+ channel from rat brain. When the initial open probability (Po) of the channel was relatively high, dTC decreased channel activity in a concentration dependent manner. In contrast, when the initial Po was lower, sub-micro molar dTC increased channel activity by destabilizing the closed states of the channel. Further addition of dTC up to micro molar range decreased channel activity. This dual effect of dTC implicates that there exist at least two different binding sites for dTC.