The application of high electric fields to cells or tissues permeabilizes the cell membrane and is thought to produce aqueous-filled pores in the lipid bilayer (Crowley, 1973; Dimitrov, 1984; Glaser et al., 1988; Needham and Hochmuth, 1989; Teissié et al., 1999; Zimmerman, 1996; Zimmermann et al., 1976). This process, first observed for planar bilayer lipid membranes (Abidor et al., 1979; Benz et al., 1979), is referred to as membrane breakdown, electropermeabilization, or electroporation (Tsong, 1991; Weaver, 1995). It finds today numerous applications since, under certain conditions, it is reversible and hence permits efficient transmembrane transfer of small molecules (Teissié, 2002). Electroporation is routinely used in molecular biology and biotechnology and has recently found applications in medicine (Golzio et al., 2002; Harrison et al., 1998; Lee et al., 1992; Lundqvist et al., 1998; Mir et al., 1995; Neumann et al., 1982; Nishi et al., 1996). The method is also efficient for transdermal drug delivery and the transport of drugs, oligonucleotides, antibodies, and plasmids across cell membranes (Neumann et al., 1999; Prausnitz et al., 1993; Suzuki et al., 1998; Tsong, 1983, 1987).
Electroporation is witnessed when the lipid membrane is subject to transmembrane (TM) potentials of the order of a few hundred millivolts. The electroporation threshold depends on the composition of the bilayer. It can be modified by addition of amphiphilic surfactants. For instance, addition of polaxomer, a triblock anionic copolymer, increases the electroporation threshold and facilitates the membrane resealing, a property that is beneficial for minimizing electrical tissue injuries (Schmolka, 1994). In opposition, the presence of polyoxyethylene (CnEm) surfactants lowers the electroporation threshold. These may therefore be used as additives in biotechnological applications such as transdermal drug delivery to avoid thermal tissue injuries due to application of high electrical shocks (Lee and Kolodney, 1987). The intrinsic properties of the lipid membrane and its constituents may influence the electroporation threshold. Whereas cholesterol increases the electroporation threshold (Needham and Hochmuth, 1989), lysophosphatidylcholine has an opposite effect (Chernomordik et al., 1987). Membrane proteins may also influence the stability of the membrane under an external electric field. Because of their interactions with lipids, integral membranes are shown, for instance, to modulate the bilayer resealing as it has been demonstrated for gramicidin (Troiano et al., 1999).
To date, the molecular processes involved in membrane electroporation are still poorly known. The aim of our study is to bring about a detailed molecular level picture of the phenomena, using molecular dynamics (MD) simulations. In very recent investigations, several key aspects of the electropermeabilization process were revealed from multinanoseconds MD simulations of lipid bilayers. It was shown that under a high electric field, 0.5 V.nm–1 and above, pore formation can be induced in bilayers on a nanosecond timescale (Tieleman et al., 2003). In a more detailed and recent study (Tieleman, 2004), it was found that for a large enough system, multiple pores with sizes up to 10 nm form independently. The simulations have evidenced that the electroporation process takes place in two stages. First, water molecules organized in single file like wires penetrate the hydrophobic core of the bilayer. This water penetration is apparently favored by local defects in the lipid headgroup region. Then, the water wires grow in length and expand into water-filled pores. These pores are stabilized by lipid headgroups that migrate from the membrane-water interface to the middle of the bilayer. It is suggested that water wires formation, the precursor to full electroporation, is driven by local field gradients at the water-lipid interface.
According to Tieleman's investigation, qualitatively, the pore formation does not seem to depend on the nature of the lipid headgroup. In fact, his MD simulations show that pores form even in the case of an octane layer sandwiched between water layers, i.e., in the absence of polar headgroups. This is consistent with experimental evidence on planar membranes of phosphatidylcholine and phosphatidylserine lipids (Diederich et al., 1998) that suggests that the rupture behavior, viz., membrane breakdown voltage and rupture kinetics are almost independent of the charges carried by the lipid headgroups. Similarly both previous simulations and experiments suggest that the electroporation process is independent of the ionic strength of the medium surrounding the membrane.
Here, after presenting rather similar results obtained independently by us, we address several key questions that remain open: 1), Do we observe resealing of the pores when the electric field is switched off? What is then the sequence of events? 2), What effect has the presence of a transmembrane channel on the process? and 3) What is the likely sequence of events that take place for translocation of a DNA plasmid placed near a membrane interface when the system is subject to the electric field?
To do so, we performed MD simulations of a bare bilayer, a bilayer containing a peptide nanotube channel, and a model membrane with a peripheral DNA double strand. In all cases, the applied voltage induced formation of water channels across the membrane that are stabilized by hydrophilic pores formed by participating lipid headgroups and acyl chains. The peptide channel is shown to stabilize the membrane due to its strong interaction with nearby lipids. The DNA strand migrates to the membrane interior only after electroporation of the bilayer. Interestingly, switching off the external transmembrane potential allows for complete resealing and reconstitution of the bilayer.