Lipid Tails Modulate Antimicrobial Peptide Membrane Incorporation and Activity

Abstract

Membrane disrupting antimicrobial peptides (AMPs) are often amphipathic peptides that interact directly with lipid bilayers. AMPs are generally thought to interact mostly with lipid head groups, but it is less clear how the lipid alkyl chain length and saturation modulate interactions with membranes. Here, we used native mass spectrometry to measure the stoichiometry of three different AMPs—LL-37, indolicidin, and magainin-2—in lipid nanodiscs. We also measured the activity of these AMPs in unilamellar vesicle leakage assays. We found that LL-37 formed specific hexamer complexes but with different intermediates and affinities that depended on the bilayer thickness. LL-37 was also most active in lipid bilayers containing longer, unsaturated lipids. In contrast, indolicidin incorporated to a higher degree into more fluid lipid bilayers but was more active with bilayers with thinner, less fluid lipids. Finally, magainin-2 incorporated to a higher degree into bilayers with longer, unsaturated alkyl chains and showed more activity in these same conditions. Together, these data show that higher amounts of peptide incorporation generally led to higher activity and that AMPs tend to incorporate more into longer unsaturated lipid bilayers. However, the activity of AMPs was not always directly related to amount of peptide incorporated.

Graphical Abstract

1. Introduction

Membrane disrupting antimicrobial peptides (AMPs) are generally small, amphipathic peptides that show antimicrobial, antifungal, and antiviral activity by interacting with the lipid bilayer. AMPs are produced by a variety of organisms and act as an innate defense against infection.[1–3] However, the modes of action of AMPs are often not well understood. The human cathelicidin AMP, LL-37, has been shown to act by several different mechanisms. It may form a toroidal pore that permeabilizes the membrane,[4] cause massive disruption by globally destabilizing the cell membrane,[5] or act differently in different bilayers by forming pores in unsaturated lipid bilayers but peptide-lipid fibrils in saturated lipid bilayers.[6] Because the lipid bilayer plays a key role in each of these processes, a clearer understanding of AMP mechanisms will require a better understanding of their interactions with lipids.

Because many AMPs are cationic, one common paradigm in the field is that these AMPs are attracted to the anionic lipid head groups present in bacterial membranes,[7–11] but it is less clear how lipid tails affect their membrane interactions. Bacteria can gain resistance to AMPs by altering both parts of lipid structure,[12] including creating less anionic head groups[13,14] and changing the membrane fluidity by modulating the fatty acids, which demonstrates that lipid tails can play an important role in AMP activity and resistance.[15,16] In vitro studies have shown that LL-37 insertion depth into the bilayer is dependent on lipid packing.[4,17] In phosphatidylcholine (PC) bilayers, LL-37 activity is modulated by the alkyl chain length. With shorter PC alkyl chains, LL-37 can disrupt membranes to form lipid-peptide nanoparticles, but LL-37 penetrates and interdigitates into longer alkyl chain PC bilayers.[18] Using lipid mixtures containing the same alkyl chain, Sevcsik et al. demonstrated that LL-37-lipid interactions are a complex interplay between head group charge and alkyl chain packing.[19] For indolicidin, membrane interactions are also affected by a combination of lipid head group and membrane fluidity, but indolicidin appears to disrupt bilayers by causing thinning of the membrane.[20,21] In contrast, magainin-2 has been shown to form pores in membranes, [22–24] and its activity is dependent on the presence of PG lipids.[25] However, less is known about the effect of the lipid alkyl chain on magainin-2 incorporation and activity. Thus, more detailed studies are needed to understand how tails of different lipid alkyl chain lengths and saturations influence the incorporation of AMPs into membranes and how this incorporation correlates with AMP activity.

In previous studies, we developed a novel approach to measure the stoichiometry and specificity of AMP complexes in controlled lipid bilayer nanodiscs using native mass spectrometry (MS).[7,8] We discovered that AMPs generally incorporated more into nanodiscs containing dimyristoyl-phosphatidylglycerol (DMPG) lipids over dimyristoyl-phosphatidylcholine (DMPC), consistent with the anionic lipid head group attracting the cationic peptides.[7,8] Unlike other AMPs that were usually nonspecific, LL-37 formed specific oligomeric complexes in DMPG nanodiscs that seemed to indicate preferences for dimers, tetramers, and hexamers. Here, we explored the effect of the lipid alkyl chain length and saturation on AMP incorporation into phosphatidylglycerol (PG) lipid bilayers, focusing on the specificity of LL-37 complex formation and comparing these results with indolicidin and magainin-2. We also used liposome leakage experiments to determine if the change in incorporation would cause more disruption of the lipid bilayer.

2. Materials and Methods

2.1. Materials

1,2-dilauroyl-sn-glycero-3-phospho-(1′-rac-glycerol) (DLPG), 1,2-dimyristoyl-sn-glycero-3-phospho-(1′-rac-glycerol) (DMPG), 1,2-dipalmitoyl-sn-glycero-3-phospho-(1′-rac-glycerol) (DPPG), 1-palmitoyl-2-oleoyl-sn-glycero-3-phospho-(1′-rac-glycerol) (POPG), and 1-stearoyl-2-oleoyl-sn-glycero-3-phospho-(1′-rac-glycerol) (SOPG) lipids were purchased from Avanti Polar Lipids (Alabaster, AL). We chose these lipids because the masses were compatible with native MS for these peptides and because they model a range of different thickness (Table S1) and fluidities. However, they may not represent the physical properties of physiological bacterial membranes. Indolicidin, LL-37, and magainin-2 were purchased from Anaspec (Fremont, CA). All peptides were 95% or greater purity. Ammonium acetate, Amberlite XAD-2, sodium cholate, Triton X-100, and 5(6)-carboxyfluorescein were purchased from Sigma-Aldrich. Methanol (LC/MS grade) was purchased from Fischer Scientific.

2.2. Nanodisc Assembly

Nanodiscs were assembled as previously described.[26,27] Briefly, lipids dissolved in chloroform were dried under nitrogen and overnight under vacuum. Dried lipids were dissolved in 0.1 M sodium cholate. The lipid-cholate mixture was mixed with membrane scaffold protein, MSP1D1(−), which was purified from E. coli as previously described.[27,28] The lipid and MSP were mixed at a molar ratio of 100:1 for DLPG, DMPG, and DPPG lipids, and 80:1 for POPG and SOPG lipids, with a final cholate concentration of 20–25 mM. After incubating at the lipid phase transition temperature, Amberlite XAD-2 hydrophobic beads were added to the mixture to remove the cholate detergent. The nanodiscs were removed from the beads, filtered, and loaded onto a Superose 6 Increase 10/300 column to purify the nanodiscs. The running buffer was 0.2 M ammonium acetate, pH 6.8.

2.3. Mass Spectrometry Sample Preparation

Samples were prepared for native MS as previously described.[8] Briefly, nanodiscs were purified and diluted to 2.5 μM before being mixed with peptides at specific molar ratios. Nanodisc samples were prepared in 0.2 M ammonium acetate and 19 μL of 2.5 μM nanodiscs was mixed with 3 μL of peptide (dissolved in methanol) and 1.5 μL of 0.4 M imidazole. Peptides were mixed with nanodiscs at 3:1, 9:1, and 18:1 molar ratios of peptide:nanodisc and allowed to equilibrate for 5 minutes before measuring.

2.4. Native Mass Spectrometry

Native MS was performed on a Q-Exactive HF quadrupole-Orbitrap mass spectrometer with the Ultra-High Mass Range (UHMR) research modifications (Thermo Fisher Scientific, Bremen, Germany).[7] Samples were introduced into the mass spectrometer using nano-electrospray ionization with a capillary voltage of 1.1 kV and a capillary temperature of 200 °C. Samples were analyzed from 2,000–25,000 m/z at a resolution setting of 15,000 and a trapping gas setting of 7. Source fragmentation was set to 50 V to help aid in desolvation, and in source trapping voltage was set to 0 V.

2.5. Mass Spectrometry Data Analysis

Data analysis was performed as previously described.[7,8] Briefly, native mass spectra were deconvolved using UniDec and MetaUniDec.[29,30] Deconvolution settings were as follows: mass range of 20–200 kDa, charge range of 5–25, and peak full width half-maximum of 10. The lipid mass was used as the mass difference. Following deconvolution, macromolecular mass defect analysis was used to determine the number of peptides incorporated into the nanodiscs.

2.6. Liposome Assembly

Unilamellar vesicles (ULVs) were made as previously described.[31] Briefly, ^~10 mg of lipid in chloroform was dried under nitrogen followed by overnight in vacuum. 500 μL of 20 mM carboxyfluorescein and 200 mM ammonium acetate, pH 6.8, were added to the dried lipids. The solution was sonicated for 30 seconds, followed by 15 seconds of vortexing. Sonicating and vortexing were repeated three times. The sample was extruded (Avanti Polar Lipids, Alabaster, AL) using a 100 nm polycarbonate Track-Etched filter (GE Healthcare). The extruded ULVs were loaded onto a HiTrap desalting column (GE Healthcare) to remove any carboxyfluorescein that was not encapsulated and exchange the buffer to only 0.2 M ammonium acetate. The lipid content of the liposomes was quantified by measuring the total phosphorous concentration.[32,33]

2.7. Carboxyfluorescein Leakage Assay

Leakage assays were performed on a Horiba PTI Quantamaster 400 fluorometer using an excitation wavelength of 492 nm, emission wavelength of 515 nm, and slit width of 5 nm. To maintain the stability of DMPG liposomes, all experiments were performed at 10 °C. ULVs were first diluted to a total lipid concentration of 2 μM. 3 mL of this ULV sample was added to a cuvette and the fluorescence emission was measured for 5 mins. At 5 mins, peptide was added at a specific peptide:lipid ratio equal to the peptide:lipid ratio used for mass spectrometry measurements with nanodiscs. For example, a DMPG nanodisc contains an average of 160 lipids. To achieve the same 3:1 molar ratio as with nanodiscs we needed to add 1 peptide for every ~53 lipids in our solution. Using a lipid concentration of 2 μM, this means adding peptide to a final concentration of 38 nM to match our 3:1 peptide:nanodisc molar ratio. The number of lipids in a nanodisc varied for each lipid used in these studies and therefore, the amount of peptide added to the ULVs varied for each lipid used. After peptides were added to the ULVs, the measurement continued for 20 minutes. Finally, at 25 minutes, Triton X-100 was added to a final concentration of 0.1% to give the maximum fluorescence value. Controls with the addition of methanol alone (without dissolved AMPs) had a minimal effect on the release of carboxyfluorescein (~1-2%).

Each ULV experiment was repeated three times with a different ULV preparation. The results of each individual experiment were normalized to the maximum and minimum fluorescence values for that sample, and the mean of the normalized fluorescence was plotted with the standard error of the mean.

3. Results and Discussion

3.1. Effects of lipid alkyl chains on LL-37 incorporation and activity

Our goal was to determine how bilayer thickness and lipid saturation affected LL-37 incorporation and complex formation in PG bilayers. Thus, we prepared nanodiscs containing a single PG lipid with alkyl chains of different length and degrees of saturation. LL-37 was titrated into each nanodisc, allowed to incubate, and the complex was then characterized by native MS (Figure S1A–B). We analyzed and quantified the small mass shifts upon AMP addition using mass defect analysis (Figure S1C) to determine the relative abundance of each peptide stoichiometry in nanodiscs (Figure S1D).

Although we cannot distinguish between peptide association with the surface of the nanodisc versus embedding into the bilayer, we can distinguish formation of specific oligomers. Native MS only reports directly on the stoichiometry of incorporation (how many peptides are associated with the nanodisc), but we can infer formation of specific complexes when non-random stoichiometry distributions appear. For example, LL-37 in DMPG nanodiscs first incorporated with a stoichiometry of two, with very little nanodiscs containing one or three peptides at a 3:1 LL-37:nanodisc ratio (Figure 1C).[7] The non-random stoichiometries reveal specificity for dimer in the nanodisc. No evidence for LL-37 trimer was observed at any ratio in DMPG nanodiscs, and higher ratios show non-random distributions with a progressive increase from dimer to tetramer to hexamer, with a population with a stoichiometry of five that may be pentamers but cannot be conclusively assigned.[7] Incorporation of monomers or dimers would show a random Poisson distribution that is not observed with LL-37 but has been observed for other non-specific AMPs.[8]

Figure 1.

(A) Schematic of addition of LL-37 to nanodiscs. (B–E) Relative amounts of incorporation of different LL-37 stoichiometries in nanodiscs containing DLPG (B), DMPG (C), DPPG (D), or SOPG (E), at 3:1 (purple), 6:1 (blue), 9:1 (green), or 18:1 (grey) molar ratios of LL-37:nanodisc. Data for DMPG are adapted from Walker et al.[7] (C) Adapted with permission from ref. 8. Note: hypothetical structures are meant to illustrate the method and not propose a particular structure.

In contrast, when LL-37 was added to nanodiscs containing the short alkyl chain lipid, DLPG, LL-37 incorporated at similar levels to DMPG nanodiscs but lost most of its specificity in complex formation (Figure 1B). LL-37 showed strong preferences for specific stoichiometries in DMPG nanodiscs but had much broader distributions of stoichiometries in DLPG nanodiscs. A truncated version of LL-37, KR-12, also lacked specificity in DMPG,[8] which indicates that the interplay between the lipid and the peptide is important for formation of specific complexes. However, at the highest concentration tested, LL-37 in DLPG shows some specificity for tetrameric complexes, which agrees with the tetrameric crystal structure solved in the presence of n-dodecyl-phosphocholine (a 12:0 detergent).[34]

With longer alkyl chain lipids (DPPG and SOPG), the stoichiometry of LL-37 in nanodiscs was higher than observed with the shorter alkyl chain lipids at the same ratios, demonstrating an overall higher affinity for incorporation into thicker bilayers (Figure 1 and Figure 2). Unfortunately, overlap between the mass of six POPG molecules and LL-37 prevented interpretation of native MS data with POPG nanodiscs (Table S2), so no results are reported for this lipid combination. Higher molar ratios of LL-37 (18:1) that were stable in DMPG generally caused dissociation of the SOPG and DPPG nanodiscs. Because the highest stoichiometry observed in any lipid nanodisc was six, these data suggest that the hexamer is the highest oligomer allowed, and any further incorporation beyond this destabilizes the nanodisc.

Figure 2.

Average incorporation of peptides in nanodiscs containing different lipids at 3:1 (A, C, E) and 9:1 (B, D, F) ratios of peptide:nanodisc for LL-37 (A, B), indolicidin (C, D), and magainin-2 (E, F). Connected lines above bars indicate significance as determined by a T-test at the 95% confidence level. Solid, dashed, and dotted lines are used only to help differentiate between overlapping lines. Data for DMPG are from Walker et al.[7]

In stark contrast with nanodiscs with DLPG and DMPG, trimeric LL-37 is clearly the dominant species in DPPG and SOPG nanodiscs at a 3:1 ratio. After the formation of the trimeric LL-37, additional LL-37 at higher ratios causes progressively higher stoichiometries before ultimately settling on hexamer complexes. Higher stoichiometries beyond six were not observed, demonstrating the specific formation of hexamer complexes. It is not clear whether tetramer and pentamer intermediates are specific, but there is a clear specificity for trimer and hexamer.

Together, these data reveal a fascinating picture of how lipid alkyl chains modulate formation of LL-37 complexes in PG bilayers. Overall, LL-37 has a higher affinity for thicker membranes. The similar stoichiometries between DPPG and SOPG, which have similar bilayer thickness[35] but different fluidity, suggest that membrane fluidity and unsaturation are less important than bilayer thickness. Very thin DLPG bilayers largely disrupt the formation of specific complexes except for the tetramer. DMPG, DPPG, and SOPG all showed clear evidence for formation of specific hexamers, but different intermediate oligomers are observed. DMPG had dimer intermediates whereas DPPG and SOPG both had trimer intermediates. We expect that the dimer and trimers are likely too small to form pores and are thus likely pre-pore intermediate complexes.

As described above, a structure of tetrameric LL-37 has been published with antiparallel orientation of the monomers [34], and our results in DLPG support that tetramers are likely structures in thinner detergents or lipid bilayers. DMPG may promote hexamer formation in a similar antiparallel structure by simply adding an additional dimer building block, but we cannot comment on the structural arrangement from our data alone. For DPPG and SOPG, it is harder to imagine how antiparallel units assemble into hexamers via trimer intermediates. It is possible that a different structural arrangement is used with different tilt or bending angles of the peptide in thicker membranes. For example, MscL channels adjust their tilt angle in response to changes in membrane thickness,[36] and their oligomeric state is also affected by the lipid environment.[37,38] It is also possible that the trimer undergoes major rearrangement before forming a hexamer. Although prior SEC and cross-linking studies have found evidence of LL-37 hexamers,[10,34] we are not aware of any hexameric LL-37 structures, so we cannot speculate on the structural arrangement of hexamers or trimer intermediates. In any case, lipid alkyl chains play clear roles in modulating the overall membrane affinity, the specificity of complex formation, and the oligomeric intermediates present at lower concentrations.

3.2. Effects of lipid alkyl chains on LL-37 activity

We also tested the effects of lipid alkyl chains on the activity of LL-37 using a vesicle leakage assay with carboxyfluorescein-encapsulated ULVs made with either DMPG or SOPG.[31] Prior studies used PC,[6] mixtures of PC:PG[6,39,40], or natural lipids [41], so our goal was to compare the activity of LL-37 in different pure PG ULVs. Although we could not match the concentrations used for native MS with nanodiscs, we titrated AMPs at similar peptide:lipid ratios with ULVs, and we compared relative trends rather than specific ratios. The absolute concentrations of the AMPs varied with the lipid type used (to keep the peptide:lipid ratio constant), but the absolute concentrations for each AMP was between 30 and 300 nM. Release of carboxyfluorescein was determined by an increase in fluorescence due to lower self-quenching upon dilution as it escaped the ULV.

When LL-37 was added to ULVs at LL-37:lipid molar ratios equivalent to 3:1, 9:1, and 18:1 LL-37:nanodisc, SOPG ULVs showed significantly more leakage of carboxyfluorescein compared to ULVs containing DMPG (Figure 3). SOPG ULVs showed partial vesicle leakage at 3:1 and nearly complete leakage (compared to addition of Triton X-100) at 9:1 and 18:1 ratios. DMPG ULVs also showed partial vesicle leakage but less than SOPG ULVs, which agrees with the higher incorporation of LL-37 in SOPG bilayers observed by native MS. Although there are differences in absolute concentration, temperature, and lipid phase between the nanodisc and ULV experiments, the higher activity of LL-37 in SOPG over DMPG indicates that lipid alkyl chains can modulate AMP activity, with both higher incorporation and higher activity shown for lipids with longer alkyl chains. In agreement with this data, previous results with PC lipids indicate that LL-37 has a different mode of action with different lipids, with faster leakage in thicker and more fluid membranes.[6,39]

Figure 3.

ULV Leakage Assay with DMPG (A) or SOPG (B) ULVs with addition of 3:1, 9:1, or 18:1 ratios of LL-37:nanodisc. Fluorescence of carboxyfluorescein encapsulated LUVs was measured for 5 minutes. At 5 minutes, LL-37 was added at the molar ratios indicated. At 25 minutes, 0.1% Triton was added to determine maximum fluorescence. Results were normalized to highest and lowest fluorescence values, and the mean of the normalized fluorescence from 3 different ULV preparations was calculated. The shaded areas indicate the standard error of the mean for each stoichiometry.

Because 16:0 and 18:1 alkyl chains are the most abundant in E. coli membranes,[42] these data overall support that LL-37 evolved to target bacterial membranes and lipids with their most common fatty acids. Lipids with shorter alkyl chains (12:0 and 14:0) are less common in bacteria[43] and thus may not mimic the bacterial membrane well. Prior native MS has shown that LL-37 interacts more with nanodiscs that have a blend of bacterial head groups than with PG nanodiscs alone.[44] Furthermore, a recent study by Reid et al. showed that LL-37 is less susceptible to photochemical footprinting in nanodiscs made of E. coli lipid extract compared to nanodiscs made of DMPC or DMPG lipids, which indicates that LL-37 has a stronger affinity for this natural lipid mixture.[45] Thus, both lipid head groups and alkyl chains can influence LL-37 interactions with the membrane, and lipid compositions that more closely mimic bacterial membranes drive greater incorporation.

3.3. Effects of lipid alkyl chains on indolicidin and magainin-2 incorporation and activity

To place our results with LL-37 in context, we compared them with similar data from two different peptides, indolicidin and magainin-2. When indolicidin was titrated into PG nanodiscs containing lipids with different alkyl chains, more indolicidin was incorporated into nanodiscs containing an unsaturated lipid (POPG and SOPG), with saturated lipids all showing similar amounts of incorporation (Figure 2C–D, S2, and S3). There was no statistical difference in the average amount of peptide incorporation between POPG and SOPG. Interestingly, indolicidin showed the lowest level of incorporation in nanodiscs containing DPPG. This indicates that indolicidin incorporation is not controlled by bilayer thickness but instead by bilayer saturation/fluidity. Previously, Zhao et al. found that indolicidin will incorporate into bilayers that contain PG lipids or cholesterol, indicating that indolicidin incorporation is an interplay between membrane fluidity and lipid head group.[25] Prior native MS experiments showed minimal incorporation into PC nanodiscs, confirming the preference for PG lipids,[8] and these new results show greater incorporation into more fluid bilayers. Combined, these studies indicate that indolicidin prefers to incorporate in membranes with PG and more fluid bilayers.

We then compared the native MS nanodisc data to the ULV leakage assay. At low molar ratios (3:1), indolicidin caused significantly more leakage in DMPG ULVs than SOPG ULVs (Figure S6). If fluidity drove activity, we would expect indolicidin to have more activity in SOPG bilayers, but instead it is in the more rigid DMPG bilayers were indolicidin is more active. At an intermediate stoichiometry (9:1), indolicidin showed similar levels of carboxyfluorescein leakage after 20 minutes in both lipids. However, SOPG ULVs took longer to reach the same level of leakage while the leakage from DMPG ULVs was almost instantaneous (Figure S6). At the highest stoichiometry tested (18:1), there was fast and complete leakage in both DMPG and SOPG ULVs (Figure S6). This contrasted the higher stoichiometries observed in SOPG nanodiscs over DMPG nanodiscs (Figure 2C–D). Overall, indolicidin incorporated more into more fluid membranes but was more disruptive in less fluid membranes.

It may be that differences in temperature, concentration, lipid phase, or other experimental parameters cause this apparent discrepancy between incorporation stoichiometries measured by native MS and activity measured by ULV leakage. It could also be that the more specific indolicidin complexes observed by native MS in DMPG nanodiscs are more active than the less specific complexes in SOPG, which may only interact through nonspecific membrane association (Figure S2). Finally, it may be that indolicidin is disrupting the DMPG bilayers but either forming pores or acting as a carrier in the SOPG bilayers, as suggested by Rokitskaya et al.[46] In any case, the fact that indolicidin is relatively short[47] and may not be able to span the full thickness of the SOPG bilayer may help explain its lower activity for these thicker ULVs.

Like indolicidin, we previously found that the AMP magainin-2 showed minimal specificity in DMPG nanodiscs.[8] When lipids with different alkyl chains were used, magainin-2 incorporated more into nanodiscs containing lipids with longer and more unsaturated alkyl chains (Figure 2E–F, Figure S4 & S5). At a 9:1 ratio, the highest incorporation was into nanodiscs containing SOPG, followed by POPG, and then DPPG. This data indicates that magainin-2 prefers lipids with both longer and more unsaturated alkyl chains. Like indolicidin, magainin-2 generally showed little preference for specific complex formation (Figures S4 & S5).

The SOPG ULVs showed ^~100% carboxyfluorescein leakage after 20 minutes at all molar ratios of magainin-2 tested but steadily increasing rates at higher concentrations (Figure S7B). With DMPG ULVs, the amount of carboxyfluorescein leakage also increased with concentration but never went higher than ~50% leakage after 20 minutes (Figure S7A). For this peptide, the ULV leakage data shows a strong correlation with the incorporation data, indicating that magainin-2 has higher activity and higher incorporation in longer and more unsaturated lipid bilayers. Although we cannot directly determine whether fluidity or thickness played more important roles in affecting magainin-2 activity, prior studies showing little influence of cholesterol on magaining-2 interactions with membranes may point to thickness playing a more important role.[25] Interestingly, each of these three peptides was differently effected by lipid alkyl tail composition, revealing that their membrane incorporation and disruption is driven by a complex interplay between the unique peptide and different combinations of membrane properties, including charge, thickness, and fluidity.

4. Conclusions

Here, we investigated how the alkyl chain length and saturation of PG lipids affects the incorporation of the AMP LL-37 into lipid nanodiscs and how this incorporation affects the leakage of carboxyfluorescein from ULVs. Overall, LL-37 showed higher levels of incorporation and activity in lipid bilayers containing longer alkyl chains but was relatively unaffected by unsaturation. Interestingly, LL-37 formed specific hexamer complexes in all but DLPG but showed dimer intermediates in DMPG and trimer intermediates in DPPG and SOPG, showing that the lipid alkyl chains can also influence oligomeric intermediates. In contrast, indolicidin incorporated more in more fluid membranes but had higher leakage activity in less fluid membranes. Finally, magainin-2 showed both higher incorporation and a higher level of vesicle leakage in membranes with longer, unsaturated alkyl chains. Together, these results reveal the intricate and peptide-specific roles of lipid alkyl chains in AMP interactions with membranes, affecting the overall membrane affinity, the assembly of specific complexes, and the ability to disrupt lipid bilayers.