The antibiotic and DNA-transfecting peptide LAH4 selectively associates with, and disorders, anionic lipids in mixed membranes
Posted Jul 27 2009 11:53pm
A. James Mason*,,1, Amélie Martinez, Clemens Glaubitz, Olivier Danos, Antoine Kichler and Burkhard Bechinger* * Faculté de chimie, University Louis Pasteur/CNRS UMR717-LC3, Institut le Bel, Strasbourg, France; Généthon-CNRS UMR8115, Evry, France; and Centre for Biomolecular Magnetic Resonance and Institut für Biophysikalische Chemie, J. W. Goethe Universität, Frankfurt, Germany
1 Correspondence: Institut Isis, 8 rue Gaspard Monge, Strasbourg 67000, France. E-mail: email@example.com
The efficient delivery of nucleic acids to eukaryotic cells by nonviral means has potential applications from basic research to gene therapeutic approaches for congenital disorders, cancer and viral infection. It has been proposed previously that cationic DNA containing complexes enter the cell through disruption of the endosomal membrane during endocytosis. It has also been observed that transformed cell lines are more easily transfected than primary cell lines by the cationic amphipathic DNA vector peptide LAH4. We therefore sought to determine whether anionic lipids, found in eukaryotic cells and presented on the surface of tumorigenic cells, perform a role in LAH4 mediated DNA delivery.
1. Solid-state NMR methods show that LAH4 interacts preferentially with anionic lipids in mixed model membranes LAH4 is a 26 residue peptide which adopts an amphipathic -helical conformation when associated with lipid bilayers. It adopts a transmembrane orientation at pH 7.5 where lysine residues at each terminus act as hydrophilic anchors. At acidic pH it switches into an in-plane orientation concomitant with the protonation of its four histidine residues. LAH4 is capable of complexing DNA, associating with the cell surface membrane and then, when enveloped within an endosome, disrupts the endosomal membrane as the pH drops.
The first step in our study was to determine whether cationic LAH4, as expected, would interact preferentially with anionic lipids in model membranes. The 31P NMR isotropic chemical shift of phospholipids incorporated into lipid vesicles has been shown previously to be a sensitive reporter of surface charge with upfield shifts accompanying an increase in positive charges and vice versa. We followed the isotropic resonances of the bilayer phospholipids by 31P magic angle sample spinning (MAS) NMR of vesicles containing anionic [phosphatidylserine (PS) and phosphatidylglycerol (PG)] as well as zwitterionic [phosphatidylcholine (PC)] lipids. On the addition of the cationic peptide LAH4, the isotropic chemical shifts of the anionic lipids were observed to shift upfield by a greater magnitude than those of PC while the line widths, as determined by measuring the full peak width at half height, also showed far greater increase for anionic lipids. The changes in line width and chemical shift as well as the differences between anionic lipids and PC were more marked at acidic pH than at pH 7.5.
A second solid-state NMR method revealed that addition of LAH4, to lipid vesicles containing chain-deuterated PC as a reporter of structure and motions within the lipid acyl chain, caused only a very weak reduction of order within the chain region. The effect was attenuated when unlabeled PS was included in the vesicles in place of unlabeled PC. Taken together, these results show that LAH4 preferentially interacts with anionic lipids in model membranes and this interaction is stronger at acidic pH, which the peptide would experience in the endosome during transfection of the cationic DNA/peptide complex.
2. 2H NMR of chain-deuterated phosphatidylserine shows that LAH4 effectively disrupts the lipid acyl chains at acidic pH Following our observation that LAH4 interacts preferentially with anionic lipids, we sought to characterize this interaction further by using wide line 2H NMR to monitor the acyl chain order of vesicles containing chain-deuterated PS as a reporter. Incorporation of LAH4 did not significantly affect 2H NMR spectra of mixed lipid vesicles containing PC/PS/cholesterol at pH 7.5. However, when the samples were resuspended and buffered at pH 5 a sharp reduction in lipid acyl chain order is observed. This data shows that not only does LAH4 associate preferentially with anionic lipids but it is also capable of disrupting their lipid chains with a much greater efficiency than those of zwitterionic PC.
3. The disruption of the lipid acyl chains can be manipulated by altering the angles subtended by the positively charged histidine residues and this activity is mirrored by the in vitro transfection efficiency of the peptides against eukaryotic cells We investigated how the angle subtended by the histidine residues, positively charged at acidic pH, influences the interaction of LAH4 derivatives with anionic PS. The membrane disordering due to the presence of the peptide was assessed again using 2H NMR of mixed PC/PS/cholesterol vesicles with chain-deuterated PS as reporter. LAH4 peptide isomers (Fig. 1 ) with angles of 80° and 100° were shown to disrupt lipid chain order with the highest efficiency while those with angles of 60° and 180° were noticeably weaker (Fig. 2 A). The same peptides were then tested in vitro for their transfection efficiency against four eukaryotic cell lines. It was found that the transfection efficiency for the LAH4 isomers in all cell lines tested mirrored the chain disruption efficiency as shown by solid-state NMR (Fig. 2B ), hence demonstrating a close relationship (Fig. 2C ) between the peptide activity in model membranes and in live cells. The same dependency of transfection efficiency against charged angle was found for all cell lines, including MRC-5 cells which are primary lung fibroblasts. These healthy cells are unlikely to have anionic lipids presented in the external leaflet of the cell membrane and hence our study is in agreement with a model whereby anionic lipids are recruited from the internal leaflet by lipid flip-flop, peptide induced or otherwise, to combine and destabilize the cationic DNA/peptide complex.
Figure 1. Helical wheel diagrams for LAH4 (A) and isomers LAH4-L0 (B), LAH-L1 (C), LAH4-L2 (D), and LAH4-AL6 (E). The angles subtended by the positively charged helix face, at pH 5, for the peptides are 60° (L0), 80° (L1), 100° (L2 and LAH4), and 180° (AL6). The pdb files were generated using Insight II and the figures rendered using Raswin.
Figure 2. 2H order parameter profiles for peptide/POPS-d31 containing vesicles, calculated relative to POPS-d31 in peptide free vesicles (A) and a comparison of the relative transfection efficiency of the LAH4 vectors on four eukaryote cell lines (B). In the transfection experiments, increasing amounts of peptide were tested and only the formulation giving the highest luciferase activity is shown. Error bars represent the SEM. *P
< 0.05 compared with LAH4 for each cell type. The relationship between the transfection data and order parameter is shown (C) by comparing normalized transfection efficiencies and normalized disruption efficiencies calculated for the lowest five segments of the acyl chain.
CONCLUSIONS AND SIGNIFICANCE
The mechanism by which nonviral cationic DNA vectors deliver their cargo to eukaryotic cells is not well understood. What is known for LAH4 is that the peptide has the ability to complex with DNA in solution, when associated with a biological membrane the peptide can exist in surface oriented or transmembrane topologies, where the in-plane alignment results in a more powerful membrane disrupting ability, and that endosomal acidification plays an important part in the transfection process. It has also been suggested previously that anionic lipids may perform a role in destabilizing cationic complexes leading to release of DNA to the interior of the cell. Anionic lipids are normally sequestered to the internal leaflet of eukaryotic cell membranes and hence the outer leaflet of endosomal membranes and so lipid flip-flop is thought to be a prerequisite for this process in healthy cells. This is not an unreasonable hypothesis as other membrane active peptides have been shown to induce lipid flop. Certain cells that are tumorigenic or, for example, infected with HIV, have been shown to present anionic PS at the external membrane leaflet, which may cause enhanced adsorption of cationic complexes and, according to the hypothesis above, also enhances transfection efficiency since more anionic lipids would be present in the internal leaflet of the endosomal membrane.
Our data shows that LAH4 does indeed interact with anionic lipids and is capable of disrupting these lipids in model membranes. Furthermore, a close relationship was demonstrated between these effects, studied in model membranes, and the process of LAH4 mediated transfection in live cells. This supports an important role for anionic lipids during the endosomal acidification phase in transfection mediated by cationic compounds such as LAH4.
The dependence on the positively charged angle was the same for all cell lines including the primary cell line MRC-5 as discussed above. It is notable, however, MRC-5 V2 cells, which are the same lung fibroblasts transformed with sv40 virus, showed the same angle dependence but were overall much more susceptible to transfection by all the peptides tested in this work.
Our study may explain therefore why the transfection by cationic amphipathic peptides such as LAH4 is more effective for cells that display anionic lipids such as PS at the external membrane surface. This information and the knowledge that the relative transfection efficiency of different cell lines can be manipulated in the manner shown will help in the design of new, more efficient and specific transfectants.