Search

The recent discoveries of serious artifacts associated with the use of cell fixation in studies of the cellular uptake of cell-penetrating peptides (CPPs) have prompted a reevaluation of the current understanding of peptide-mediated cellular delivery. Following a report on the differential cellular uptake of a number of penetratin analogs in unfixed cells, we here investigate their membrane translocation abilities in large and giant unilamellar vesicles (LUVs and GUVs, respectively). Surprisingly, in contrast to the behavior in living cells, all peptides readily entered the giant vesicles (>1 microm) as proved by confocal microscopy, while none of them could cross the membranes of LUVs (100 nm). For determination of the location of the peptides in the LUVs, a new concept was introduced, based on sensitive resonance energy transfer (RET) measurements of the enhanced fluorescence of acceptor fluorophores present solely in the inner leaflet. An easily adopted method to prepare such asymmetrically labeled liposomes is described. The membrane insertion depths of the tryptophan moieties of the peptides were determined by use of brominated lipids and found to be very similar for all of the peptides studied. We also demonstrate that infrared spectroscopy on the lipid carbonyl stretch vibration peak is a convenient technique to determine phospholipid concentration.

Reports on serious artifacts associated with the use of cell fixation in studies of the cellular uptake of cell-penetrating peptides, also denoted protein transduction domains, have demonstrated the need for a reevaluation of the current understanding of peptide-mediated cellular delivery of large, hydrophilic molecules. In a recent study on the internalization in unfixed cells of penetratin and its analogues in which tryptophans are substituted for phenylalanines (Pen2W2F), lysines for arginines (PenArg), and arginines for lysines (PenLys), we revealed large dissimilarities in cell interactions among the peptides [Thorén et al. (2003) Biochem. Biophys. Res. Commun. 307, 100-107]. We here investigated possible correlations with their respective affinities for the lipid membranes of large unilamellar vesicles. The variations found in membrane affinity correlated qualitatively with differences in hydrophobicity among the peptides but were by far too small to account for the striking differences in cell membrane binding. Interestingly, we found that the inclusion of a small fraction of lipids conjugated to poly(ethylene glycol) (PEG) in the vesicles both stabilized the vesicle dispersion against peptide-induced aggregation and, furthermore, enhanced the binding of the peptides to the membrane. By use of PEG-conjugated lipids, it could be shown that vesicle aggregation drives an alpha-helix to beta-sheet conformational transition for these peptides. A similar transition was discovered at submicellar concentrations of sodium dodecyl sulfate in aqueous solution for all peptides except PenLys. Finally, significant changes of the contributions to CD spectra from aromatic residues due to their insertion into the membrane were observed.

The binding of penetratin, a peptide that has been found useful for cellular delivery of large hydrophilic molecules, to negatively charged vesicles was investigated. The surface charge density of the vesicles was varied by mixing zwitterionic dioleoylphosphatidylcholine (DOPC) and negatively charged dioleoylphosphatidylglycerol (DOPG) at various molar ratios. The extent of membrane association was quantified from tryptophan emission spectra recorded during titration of peptide solution with liposomes. A singular value decomposition of the spectral data demonstrated unambiguously that two species, assigned as peptide free in solution and membrane-bound peptide, respectively, account for the spectral data of the titration series. Binding isotherms were then constructed by least-squares projection of the titration spectra on reference spectra of free and membrane-bound peptide. A model based on the Gouy-Chapman theory in combination with a two-state surface partition equilibrium, separating the electrostatic and the hydrophobic contributions to the binding free energy, was found to be in excellent agreement with the experimental data. Using this model, a surface partition constant of approximately 80 M(-)(1) was obtained for the nonelectrostatic contribution to the binding of penetratin irrespective of the fraction of negatively charged lipids in the membrane, indicating that the hydrophobic interactions are independent of the surface charge density. In accordance with this, circular dichroism measurements showed that the secondary structure of membrane-associated penetratin is independent of the DOPC/DOPG ratio. Experiments using vesicles with entrapped carboxyfluorescein showed that penetratin does not form membrane pores. Studies of the cationic peptide penetratin are complicated by extensive adsorption to surfaces of quartz and plastics. By modification of the quartz cell walls with the cationic polymer poly(ethylenimine), the peptide adsorption was reduced to a tolerable level. The data analysis method used for construction of the binding isotherms eliminated errors emanating from the remaining peptide adsorption, which otherwise would prevent a proper quantification of the binding.