Membrane Association Modes of Natural Anticancer Peptides: Mechanistic Details on Helicity, Orientation, and Surface Coverage

Abstract

Anticancer peptides (ACPs) could potentially offer many advantages over other cancer therapies. ACPs often target cell membranes, where their surface mechanism is coupled to a conformational change into helical structures. However, details on their binding are still unclear, which would be crucial to reach progress in connecting structural aspects to ACP action and to therapeutic developments. Here we investigated natural helical ACPs, Lasioglossin LL-III, Macropin 1, Temporin-La, FK-16, and LL-37, on model liposomes, and also on extracellular vesicles (EVs), with an outer leaflet composition similar to cancer cells. The combined simulations and experiments identified three distinct binding modes to the membranes. Firstly, a highly helical structure, lying mainly on the membrane surface; secondly, a similar, yet only partially helical structure with disordered regions; and thirdly, a helical monomeric form with a non-inserted perpendicular orientation relative to the membrane surface. The latter allows large swings of the helix while the N-terminal is anchored to the headgroup region. These results indicate that subtle differences in sequence and charge can result in altered binding modes. The first two modes could be part of the well-known carpet model mechanism, whereas the newly identified third mode could be an intermediate state, existing prior to membrane insertion.

Keywords: anticancer peptides; flow-linear dichroism; molecular dynamics; peptide conformation; spectroscopy.

Scheme 1

Chemical structures of the compounds used in the study. (A) Helical wheel diagram of the peptides was drawn with the software package Protein ORIGAMI [62]. The arrow indicates the hydrophobic face of the peptide. (B) model membranes built up of phosphatidylcholine (DOPC, PC), phosphatidylglycerol (DOPG, PG), and phosphatidylserine (DOPS, PS) were used throughout the study. Pure PC, PC/PS (80:20), and PC/PG (80:20) were used for mimicking spatial distribution of electrostatic features for neutral and negatively charged biomembranes, respectively. Note the zwitterionic but net neutral nature of PC, the single negative charge of PG, and the −/+/− charge distribution in the head-group of PS.

Figure 1

Peptide structural changes induced upon binding to model membranes assessed by CD spectroscopy. (A–E) Far-UV CD spectra in the presence and absence of the model membranes PC, PC/PG, and PC/PS. Spectra were collected at 635 µM for (A–D) and at 320 µM for (E) lipid concentration resulting in the same peptide-to-lipid ratio of 1:16. Insets show the calculated helix content. (F) Calculated peptide helix content with various lipid membranes. Helix content was calculated using the online tool (http://bestsel.elte.hu) [67].

Figure 2

Membrane interaction studied by infrared spectroscopy (IR). Representative infrared spectra in the lipid head-group region of (A) PC, (B) PC/PG, and (C) PC/PS membranes alone and upon interaction with the studied peptides. The analysed vibrational modes, namely ν_asym(PO_2⁻), ν_sym(PO_2⁻), and ν(N⁺-(CH₃)₃), refers to the asymmetric and symmetric phosphate stretching, and to the asymmetric choline stretching, respectively. Representative infrared spectra of the acyl CH₂ group region of (D) PC, (E) PC/PG, and (F) PC/PS membranes alone and upon interaction with the studied peptides. Peptide and lipid concentrations were 80 µM and 1.27 mM, respectively, for all the peptides except for LL-37 which was 40 µM and 0.635 mM, respectively, resulting in the same peptide-to-lipid ratio.

Figure 3

Peptide orientation upon binding to lipid membrane. (A,B) LD spectra of Lasio III (A) and Tempo-La (B) in the presence of the model membranes PC, PC/PG, and PC/PS. Spectra were collected at 1.27 mM lipid concentration. (C) Schematic representation of the transitions moment directions of an α-helical peptide secondary structure. (D,E) Fluorescence emission spectra of Lasio III (Trp fluorophore, excited at 295 nm) (D) and Tempo-La (Tyr fluorophore, excited at 275 nm) (E) in the presence and absence of model membranes. Peptide and lipid concentrations were 2 and 100 µM, respectively. (F) IR analysis of the tyrosine band of Tempo-La alone and in the presence of lipid membranes, showing the Tyr vibration band region. Second derivative is shown below and was used for peak identification. Peptide and lipid concentrations were 80 µM and 1.27 mM, respectively.

Figure 4

Size and morphology variations of the vesicles induced by the ACPs. (A) Correlation function by DLS for the studied ACPs in the presence of the lipid membrane PC/PS; peptide and lipid concentrations are 40 and 635 µM for all the peptides, except for LL-37, which is 20 and 320 µM, respectively. Freeze fractured TEM images of (B) PC/PS-alone; lipid concentration 2540 µM and (C) LL-37-PC/PS mixture; peptide and lipid concentration are 80 and 2540 µM, respectively.

Figure 5

Peptide helicity changes in the presence of model membranes and distance of the ACPs from the bilayer during the MD simulations. The helicity of the peptides in the presence of the model membranes (A) PC, (B) PC/PG and (C) PC/PS is defined as the number of the residues with α-helical structure divided by the total number of the residues minus the two terminal residues (for further information see the methods section). The distance of the peptides from the surface of the different model membranes (D) PC, (E) PC/PG and (F) PC/PS as a function of the simulation time.

Figure 6

Membrane orientation of selected ACPs in the MD simulations. (A) Lasio III in the presence of PC/PS bilayer. (B,C) Tempo-La in the presence of PC and PC/PS bilayer, respectively. (D) LL-37 in the presence of PC/PS bilayer. The backbone of the peptides is shown as green cartoon. The lipid head groups are represented by orange (PC) and magenta (PS) spheres, while the lipid alkyl chains are grey lines. The H-atoms are not shown for better clarity. Representative snapshots were taken at the end of the simulations (at 500 ns).

Figure 7

REV-ACP interactions studied by LD spectroscopy. The Soret band in the LD spectra of REVs is indicative on how ACPs interact with the surface of the vesicular bilayer. LD spectra of REVs upon addition of (A) LL-37, (B) FK-16, and (C) Tempo-La. (D) LD peak intensity at 420 nm (close symbols) and at 220 (open symbols) as a function of the ACPs. Data are normalized to values measured in the absence of peptides.

Scheme 2

Proposed membrane-association mechanism for the peptides used in the study. (I) Highly helical, occupying a higher space on the surface in line with the general carpet model observed for FK-16 and LL-37. (II) Partially helical, still mostly lying on the surface observed for Lasio III and Macro1. (III) Highly helical, but preferentially perpendicular to the surface, occupying smaller space identified for Tempo-La.