5042
Langmuir 2006, 22, 5042-5050
Effect of Peptide Length on the Interaction between Consensus Peptides and DOPC/DOPA Bilayers Lovisa Ringstad,*,† Artur Schmidtchen,‡ and Martin Malmsten† Department of Pharmacy, Uppsala UniVersity, P.O. Box 580, SE-751 23 Uppsala, Sweden, and Department of Dermatology and Venereology, Lund UniVersity, Biomedical Center, SE-221 84 Lund, Sweden ReceiVed February 1, 2006. In Final Form: March 19, 2006 The effect of peptide length and electrostatics on the interaction between Cardin motif peptides and lipid membranes was investigated for (AKKARA)n (n ) 1-4) and (ARKAAKKA)n (n ) 1-3) peptides (A, K, and R refer to alanine, lysine, and arginine, respectively) by fluorescence spectroscopy, circular dichroism, ellipsometry, z potential, and photon correlation spectroscopy measurements. The effect of the peptides regarding leakage induction of both zwitterionic and anionic liposomes increased with increasing peptide length, as did the peptide-induced killing of Enterococcus faecalis and Bacillus subtilis bacteria. The peptides, characterized by a random coil conformation both in buffer and when attached to the liposomes (helix content less than 20%), displayed an increased adsorption with increasing peptide length, and plateau adsorption for the longest peptides corresponded to 1 peptide per 65 and 17 lipid molecules for zwitterionic and anionic membranes, respectively. Control experiments with uncharged peptide analogues as well as experiments at high excess electrolyte concentration showed that peptide charges are important both for peptide adsorption and leakage induction. These observations, together with observations of the liposome z potential at different peptide additions as well as a comparison between the results for zwitterionic and anionic liposomes, suggest that electrostatically affected local packing effects are crucial for the action of these peptides, although pore formation such as that observed for many AMPs cannot be excluded at present.
Introduction As a result of accelerating problems with multidrug resistance in bacteria but also as a consequence of an increasing understanding of the biological role of these substances, there is growing interest in antimicrobial peptides (AMPs). Presently, some 800 AMPs have been identified in various biological contexts.1-7 Although AMPs are multifunctional substances affecting bacteria in many different ways, including the inhibition of cell wall synthesis, the inhibition of DNA, RNA, and protein synthesis, and the inhibition of enzymatic activity,8-10 the main mode of action of AMPs seems to be the disruption of bacterial membranes.3-5,7 The rapid bacterial breakdown in combination with the bacterial membrane being the main target minimizes the development of resistance.7 The walls of bacteria are complex structures, including two lipid membranes in the case of Gram-negative bacteria (one for Gram-positive bacteria), peptidoglycan envelopes for both types of bacteria, lipopolysacharides in the case of Gram-negative bacteria, the presence of proteins in the lipid membranes, and varying lipid compositions between different bacteria.11 Despite that, investigations on the interaction between AMPs and model lipid systems offer some opportunities for gaining improved * Corresponding author. E-mail:
[email protected]. † Uppsala University. ‡ Lund University. (1) Beisswenger, C.; Bals, R. Curr. Protein Pept. Sci. 2005, 6, 255-264. (2) Boman, H. G. J. Intern. Med. 2003, 254, 197-215. (3) Papo, N.; Shai, Y. Peptides 2003, 24, 1693-1703. (4) Shai, Y. Biopolymers 2002, 66, 236-248. (5) Toke, O. Biopolymers 2005, 80, 717-35. (6) Tossi, A. http://www.bbcm.univ.trieste.it/∼tossi/pag2.htm. (7) Zasloff, M. Nature 2002, 415, 389-395. (8) Brodgen, K. A. Nat. ReV. Microbiol. 2005, 3, 238-250. (9) Lohner, K.; Blondelle, S. E. Comb. Chem. High Throughput Screening 2005, 8, 241-256. (10) Otvos, L. J. Pept. Sci. 2005, 11, 697-706. (11) Madigan, M. T.; Martinko, J. M.; Parker, J. Brock Biology of Microorganisms, 8th ed.; Prentice Hall: Upper Saddle River, NJ, 1997.
understanding of the mode of action of AMPs as well as how peptide properties such as sequence, charge and charge distribution, conformation, hydrophobicity and hydrophobicity distribution, and peptide length affect the interaction of AMPs with lipid membranes, bacteria, and cells.3,8,9,12 Heparin binding consensus sequences XBBXBX and XBBBXXBX (herein denoted as Cardin peptides), where X represents a hydrophobic or uncharged amino acid and B represents a basic amino acid, were initially described by Cardin and Weintraub in 1989.13 The heparin affinity of these peptides is related to the choices and positions of amino acids X and B as well as to the number of repeats of the consensus sequences, where increasing the number of repeats increase the affinity.14 The Cardin peptides also exert molecular-weight-dependent antimicrobial activity against both Gram-negative and Grampositive bacteria as well as fungus.15 An attractive feature of any consensus structure such as the Cardin motifs is that peptide length is varied simply by increasing the number of repeat sequences. This means that the peptide length can be varied essentially without affecting properties such as the charge, charge distribution, hydrophobicity, and hydrophobicity distribution. Such repeat sequences are therefore ideal for investigating the effect of peptide length on their interaction with bacteria and lipid model systems, particularly if the peptide secondary structure is independent of peptide length. In the present study, we therefore aim to gain further knowledge of the previously observed effect of the molecular weight of Cardin peptides on their antimicrobial activity by use of well-defined phospholipid model membranes. Additionally, the influence of peptide secondary structure, adsorption, and electrostatic interac(12) Jelinek, R.; Kolusheva, S. Curr. Protein Pept. Sci. 2005, 6, 103-114. (13) Cardin, A. D.; Weintraub, H. J. R. Arteriosclerosis 1989, 9, 21-32. (14) Verrecchio, A.; Germann, M. W.; Schick, B. P.; Kung, B.; Twardowski, T.; San Antonio, J. D. J. Biol. Chem. 2000, 275, 7701-7707. (15) Andersson, E.; Rydengård, V.; Sonesson, A.; Mo¨rgelin, M.; Bjo¨rck, L.; Schmidtchen, A. Eur. J. Biochem. 2004, 271, 1219-1226.
10.1021/la060317y CCC: $33.50 © 2006 American Chemical Society Published on Web 04/22/2006
Peptide-DOPC/DOPA Bilayer Interactions
Langmuir, Vol. 22, No. 11, 2006 5043
Table 1. Structure and Properties of the Peptides Studieda peptide
a
sequence
molecular weight (M)
isoelectric point
AKK6 AKK12 AKK18 AKK24
AKKARA AKKARAAKKARA AKKARAAKKARAAKKARA AKKARAAKKARAAKKARAAKKARA
644 1270 1895 2521
11.17 12.03 12.32 12.49
ARK8 ARK16 ARK24
ARKKAAKA ARKKAAKAARKKAAKA ARKKAAKAARKKAAKAARKKAAKA
843 1668 2493
11.26 12.05 12.33
AHH24:1 AHH24:2
AHHAHAAHHAHAAHHAHAAHHAHA AHHHAAHAAHHHAAHAAHHHAAHA
2516.6 2516.6
7.43 7.43
A, K, R, and H refer to alanine, lysine, arginine, and histidine, respectively.
tions on membrane disruption was investigated by using a method combining fluorescence spectroscopy, circular dichroism, ellipsometry, z potential, and photon correlation spectroscopy measurements. Furthermore, the results obtained for model membranes were compared with the bactericidal activity of the peptides in order to gain information on the relevance of studies with these types of model membranes. Experimental Section Materials. The peptides (AKK6, AKK12, AKK18, AKK24, ARK8, ARK16, ARK24, AHH24:1, and AHH24:2; Table 1) were synthesized by Innovagen (Lund, Sweden). The purity of these peptides was confirmed by MALDI-TOF MS analysis (Voyager, Applied Biosystems) to be >95% in all cases. 1,2-dioleoyl-snglycero-3-phoshocholine (DOPC) and 1,2-dioleoyl-sn-glycero-3phosphate (DOPA) (monosodium salt) were both obtained from Avanti Polar Lipids (Alabaster, AL) and were of >99% purity. Cholesterol (>99% purity), n-dodecyl-β-D-maltoside (DDM) (g98% purity), poly-L-lysine (Mw ) 59 and 170 kDa), and Triton X-100 were from Sigma-Aldrich (St. Louis, MO). 5(6)-carboxyfluorescein (99% purity) was from Acros Organics (New Jersey). Other chemicals used were of analytical grade. Water used was from a Millipore Milli-Q Plus 185 ultrapure water system. All measurements were performed at 37 °C at pH 7.4 in 10 mM tris-HCl buffer containing 5 mM glucose, unless stated otherwise. Surfaces. Silica surfaces were prepared from polished silicon slides (Okmetic, Finland) by thermal oxidation to give an oxide layer thickness of 30 nm. The oxidized slides were cleaned in 25% NH4OH, 30% H2O2, and H2O (1:1:5 by volume) and then in 32% HCl, 30% H2O2, and H2O (1:1:5, by volume), both at 80 °C for 5 min. This procedure resulted in surfaces with an advancing contact angle of less than 10°. The surfaces were then kept in 95% ethanol until use. Just prior to use, the surfaces were plasma cleaned at 18 W in low-pressure residual air (0.2 mbar) (Harrick Plasma Cleaner, PDC-32G, Harrick Scientific Corp.) for 5 min. Microorganisms. Enterococcus faecalis 2374 isolate was obtained from a patient with a chronic venous ulcer, and Bacillus subtilis ATCC isolate 6633 was from the Department of Bacteriology, Lund University Hospital. Viable Count Analysis. E. faecalis bacteria were grown to midlogarithmic phase in Todd-Hewitt (TH) medium. Bacteria were washed and diluted in 10 mM tris at pH 7.4 containing 5 mM glucose. Bacteria (50 µL, 2 × 106 bacteria/mL) were incubated at 37 °C for 2 h with the synthetic peptide at the indicated concentrations. To quantify the bactericidal activity, serial dilutions of the incubation mixture were plated onto TH agar, followed by incubation at 37 °C overnight, and the number of colony-forming units (cfu) was determined. Radial Diffusion Assay. Radial diffusion assays (RDA) were performed essentially as described earlier.16 Briefly, bacteria (B. subtilis) were grown to mid-logarithmic phase in 10 mL of fullstrength (3% w/v) trypticase soy broth (TSB) (Becton-Dickinson, Cockeysville, MD). The microorganisms were washed once with 10 (16) Lehrer, R. I.; Rosenman, M.; Harwig, S. S.; Jackson, R.; Eisenhauer, P. J. Immunol. Methods 1991, 137, 167-73.
mM tris at pH 7.4. Bacterial cfu ( 4 × 106) were then added to 5 mL of the underlay agarose gel consisting of 0.03% (w/v) TSB, 1% (w/v) low-electroendosmosis-type (low-EEO) agarose (Sigma, St. Louis, MO), and a final concentration of 0.02% (v/v) Tween 20 (Sigma). The underlay was poured into a Ø 85 mm Petri dish. After the agarose solidified, 4-mm-diameter wells were punched, and 6 µL of the test sample was added to each well. Plates were incubated at 37 °C for 3 h to allow peptide diffusion. The underlay gel was then covered with 5 mL of molten overlay (6% TSB and 1% lowEEO agarose in dH2O). The antimicrobial activity of a peptide was visualized as a clear zone around each well after 18-24 h of incubation at 37 °C. Liposome Preparation. The liposomes investigated in the present study were either zwitterionic (DOPC/cholesterol 60/40 mol/mol) or anionic (DOPC/DOPA/cholesterol 30/30/40 mol/mol). Dry lipid films were formed on the walls of a glass flask by dissolving phospholipid(s) and cholesterol in chloroform followed by evaporating under vacuum for 45 min at 60 °C and subsequently in a vacuum oven (Lab-line) at 30 inHg and room temperature overnight. The lipids were then resuspended in 100 mM 5(6)-carboxyflourescein (CF) in buffer, and the solution was subjected to eight freeze-thaw cycles by freezing in liquid nitrogen and heating to 60 °C while vortexing. Unilamellar liposomes were produced by multiple extrusions through 100 nm polycarbonate membranes mounted in a LipoFast Basic extruder (Avestin, Germany). The liposomes were separated from untrapped CF by running the sample on a Sephadex G-25 column (Amersham Biosciences, Sweden) with tris buffer as the eluent. Leakage Assay. In the liposome leakage assay, self-quenching of CF was used. Thus, at 100 mM CF is self-quenched, and the recorded fluorescence intensity is low. On leakage from the liposomes, released CF is dequenched, and hence fluorescence increases. The CF release was determined by monitoring the emitted fluorescence at 520 nm from a liposome dispersion (10 mM lipid in 10 mM tris at pH 7.4). An absolute leakage scale is obtained by disrupting the liposomes at the end of the experiment through the addition of 0.05 wt % Triton X-100, thereby causing 100% release and dequenching of CF. A Spex fluorolog 1680 0.22m double spectrometer (Instruments S. A. Group) was used for the liposome leakage assay. All measurements were made in at least duplicate. Circular Dichroism Spectroscopy (CD). The CD spectra of the peptides were measured on a Jasco J-810 spectropolarimeter (Jasco, Japan). The measurements were performed at 37 °C in a 10 mm quartz cuvette under stirring at a peptide concentration of 10 µM. The effect of liposomes on the peptide secondary structure was monitored at a lipid concentration of 100 µM in the range of 200250 nm. The only peptide conformations observed under the conditions investigated were R-helix and random coil. The calculation of the ratio between these conformations has been described elsewhere.17 In brief, the fraction of the peptide in R-helical conformation, XR, can be calculated from (17) Sjo¨gren, H.; Ulvenlund, S. Biophys. Chem. 2005, 116, 11-21.
5044 Langmuir, Vol. 22, No. 11, 2006 XR )
A - Ac AR - A c
Ringstad et al. (1)
where A is the recorded CD signal at 225 nm and AR and Ac are the CD signals at 225 nm for a reference peptide in 100% R-helix conformation and 100% random coil conformation, respectively. R-Helix (100%) and random coil (100%) references were obtained from 0.133 mM (monomer concentration) poly-L-lysine (Mw ) 79 kDa) in 0.1 M NaOH and 0.1 M HCl, respectively.17,18 To account for the instrumental differences between measurements, the background value (detected at 250 nm, where no peptide signal is present) was subtracted. Signals from the bulk solution were also corrected for. Ellipsometry. Peptide adsorption to supported lipid bilayers was studied by in situ null ellipsometry. The instrument used was an Optrel Multiskop ellipsometer (Optrel, Germany) equipped with a 100 mW argon laser at 532 nm, and the angle of incidence was 67.66°. All measurements were carried out in a 5 mL cuvette under stirring (300 rpm). The instrument measures the change in state of the polarization of light reflected by the substrate by monitoring the optical angels ψ and ∆. From these parameters, the mean refractive index (n) and layer thickness (d) of the adsorbed layer were calculated numerically by optical layer models.19-21 From the thickness and refractive index, the adsorbed amount (Γ) was calculated according to22 Γ)
(n - n0) d dn/dc
(2)
where dn/dc is the refractive index increment and n0 is the refractive index of the bulk solution. The refractive index increment used was 0.154 cm3/g.23,24 Corrections were routinely made for changes in the bulk refractive index caused by changes in temperature and excess electrolyte concentration. Zwitterionic bilayers were deposited by coadsorption from a mixed micellar solution containing 60/40 mol/mol DOPC/cholesterol and DDM, as described in detail previously.23 In brief, mixed micellar solutions of DOPC, cholesterol, and DDM were formed by adding 19 mM DDM in water to dry lipid films of DOPC and cholesterol followed by stirring overnight. The concentrations of the constituents in the solution were then 97.3 mol % DDM, 1.6 mol % DOPC, and 1.1 mol % cholesterol. The micellar solution was added to the cuvette, and the resulting adsorption was monitored as a function of time. When the adsorption had stabilized, the cuvette was rinsed with Milli-Q water (5 mL/min until the adsorption had stabilized, Figure 1a) to remove mixed micelles from solution and surfactant from the substrate. By repeating this procedure and subsequently lowering the concentration of the micellar solution, stable, densely packed bilayers are formed, with properties similar to those of the bulk lamellar structures of the lipids.23,25 As shown in Figure 1a, the bilayer thickness and area per molecule agree quite well with findings reported in the literature from neutron reflectivity.25,26 Bilayer buildup was performed at 25 °C. Because the mixed-micelle approach resulted in subbilayer and patchy adsorption in the case of anionic lipid mixtures, liposome adsorption was used for the generation of supported lipid bilayers in this case. DOPC/DOPA/cholesterol liposomes (30/30/40 mol/ (18) Greenfield, N. Biochemistry 1969, 8, 4108-4116. (19) Azzam, R. M. A.; Bashara, N. M. Ellipsometry and Polarized Light; North-Holland Publishing Company: Amsterdam, 1989. (20) Landgren, M.; Jo¨nsson, B. J. Phys. Chem. 1993, 97, 1656-1660. (21) Malmsten, M. J. Colloid Interface Sci. 1994, 166, 333-342. (22) de Feijter, J. A.; Benjamins, J.; Veer, F. A. Biopolymers 1978, 17, 17591772. (23) Tiberg, F.; Harwigsson, I.; Malmsten, M. Eur. Biophys. J. 2000, 29, 196-203. (24) Vacklin, H. P.; Tiberg, F.; Fragneto, G.; Thomas, R. K. Biochemistry 2005, 44, 2811-21. (25) Vacklin, H. P.; Tiberg, F.; Thomas, R. K. Biochim. Biophys. Acta 2005, 1668, 17-24. (26) Vacklin, H. P.; Tiberg, F.; Fragneto, G.; Thomas, R. K. Langmuir 2005, 21, 2827-2837.
Figure 1. (a) Buildup of a DOPC/cholesterol bilayer on silica. The thickness (0) and adsorbed amount (b) are shown. Initially, 200 µL of a DOPC/cholesterol/DDM solution was added. The following additions of DOPC/cholesterol/DDM solution are given, with arrows indicating rinsing with Milli-Q water at 5 mL/min. Average values of the properties of the final bilayer for all measurements: surface excess, 3.8 ( 0.2 mg/m2; thickness, 51 ( 9 Å; and refractive index, 1.47 ( 0.026. (b) Buildup of a DOPC/DOPA/cholesterol bilayer on silica. The thickness (0) and adsorbed amount (b) are shown. The added amount of liposome solution with a lipid concentration of 6.7 mM is given, with the arrow indicating rinsing. Average values of the properties of the final bilayer for all measurements: surface excess, 3.7 ( 0.2 mg/m2; thickness, 40 ( 10 Å; and refractive index, 1.51 ( 0.044. mol/mol) to be deposited were prepared by freeze-thaw cycles as described above, with the exception that the dried lipid films were resuspended in tris buffer only. To avoid the adsorption of peptide directly onto the silica substrate through any defects in the supported lipid layer and also to promote deposition and bilayer formation for the anionic liposomes, poly-L-lysine (Mw ) 170 kDa) was preadsorbed from water prior to lipid addition to an amount equal to 0.045 ( 0.01 mg/m2. This preadsorption was followed by the removal of nonadsorbed poly-L-lysine by rinsing with water at 5 mL/min for approximately 20 min. Water was replaced by buffer also containing 150 mM NaCl, and liposomes in buffer were then added to a lipid concentration of 20 µM, followed by rinsing with buffer (5 mL/min for approximately 15 min) when the liposome adsorption had stabilized. As can be seen in Figure 1b, this resulted in only very marginal desorption. The final layer formed had properties similar to those that were formed by zwitterionic bilayers and suggests that a layer fairly close to a complete bilayer is formed (Figure 1b). Again, bilayer buildup was performed at 25 °C. (The fact that marginal polylysine preadsorption was used for the anionic but not for the zwitterionic system is of minor importance, the main point being that the end-result lipid bilayer is essentially defect free in both cases, hence allowing AMP adsorption directly onto the underlying silica surface through bilayer defects to be safely discarded.) After lipid bilayer formation, the temperature was raised, and water was gently replaced by buffer at a rate of 5 mL/min over a period of 30 min. When the values of ψ and ∆ were stable, normally
Peptide-DOPC/DOPA Bilayer Interactions
Langmuir, Vol. 22, No. 11, 2006 5045
Table 2. LD90 Values (in µM) of AKK, ARK, and AHH Peptides for E. faecalisa peptide
LD90
AKK24 AKK18 AKK12 AKK6 ARK24 ARK16 ARK8 AHH24.1 AHH24:2
0.6 3 >100 >100 0.6 30 >100 >100 >100
a The values represent the minimal concentration of peptide required to kill at least 90% of the bacteria. Results were inferred from viable count analyses in 10 mM tris buffer using peptide concentrations of 0.03-100 µM.
after 40 min, peptide was added to a concentration of 0.01 µM. The adsorption was then monitored for 1 h followed by three subsequent peptide additions to 0.1, 1, and 10 µM, and in all cases, we monitored the adsorption for 1 h. All measurements were made in at least duplicate. Zeta Potential. The z potential of the liposomes with and without added peptides was determined by measuring the electrophoretic mobility using a Zetasizer Nano ZS (Malvern Instruments Ltd., U.K.). Measurements were performed at 25 °C (because of air bubbles precluding measurements at 37 °C) in disposable Zeta cells (Malvern Instruments Ltd., U.K.). The lipid concentrations used were 40 µM for zwitterionic liposomes and 20 µM for anionic liposomes. These concentrations were chosen in order to obtain a count rate that was high enough for the measurements. Peptide-to-lipid ratios were chosen to be identical to those used in the leakage experiments. All measurements were made in at least triplicate 1 h after sample preparation. (Although this would be possible and interesting, we have not performed z potential measurements on bacteria and monitored the effects of AMP addition on the bacteria z potential. This is partially motivated by the focus of the present article being the lipid model systems, with bacteria data included only to demonstrate the correlation between bacteria data and liposome results, and partially by the experimental difficulties in obtaining clear results for gram-negative bacteria, in particular, where the thick lipopolysaccharide outer leaflet may greatly affect the plane of shear and where regiospecific AMP binding to this leaflet can effectively “hide” AMP binding. Both effects preclude quantification.) Liposome Size Measurement. The liposome diameter was measured by photon correlation spectroscopy using a Zetasizer Nano ZS (Malvern Instruments Ltd., U.K.). All measurements were made at a 173° scattering angle and 37 °C in disposable plastic cuvettes. The lipid concentration was 40 µM for zwitterionic liposomes and 20 µM for anionic liposomes. All measurements were made in triplicate 30 min after sample preparation.
Figure 2. Antimicrobial activity of the peptides against B. subtilis determined by RDA at a peptide concentration of 100 µM. The higher the diameter (d), the higher the bacterial growth inhibition.
Results Bactericidal Effects. Viable count and RDA analysis regarding the effect of different peptides for Gram-positive B. subtillis and E. faecalis, respectively (Table 2 and Figure 2) show that the longer peptides (AKK24 and ARK24) are efficient at inhibiting bacterial growth and low concentrations are needed to kill bacteria, whereas significantly higher concentrations are needed for the shorter peptides. Reducing the peptide charge by substituting all arginine and lysine residues by histidine (uncharged at this pH) as well as increasing the excess electrolyte concentration by 150 mM NaCl eliminates the antibacterial effect of the longer peptides. Liposome Leakage. The molecular weight of the peptide is of great importance for peptide-induced liposome leakage. Thus, AKK6, AKK12, and AKK18 all induce less than 25% leakage of CF from zwitterionic liposomes even at peptide concentrations as high as 10 µM, whereas AKK24 induces 60% leakage at 0.1
Figure 3. (a) CF leakage from zwitterionic (filled symbols/solid lines) and anionic (open symbols/dashed lines) liposomes induced by AKK6 (bO), AKK12 (90), AKK18 (()), and AKK24 (24). (b) CF leakage from zwitterionic (filled symbols/solid lines) and anionic (open symbols/dashed lines) liposomes induced by ARK8 (bO), ARK16 (90), and ARK24 (()).
µM (Figure 3a). Similar trends were observed for ARK8, ARK16, and ARK24 (Figure 3b), although ARK24 is less potent than AKK24. Leakage from anionic liposomes follows the same peptide length dependence, although quantitatively there are some
5046 Langmuir, Vol. 22, No. 11, 2006
Ringstad et al.
Figure 4. Comparison of leakage induction for highly charged AKK24/ARK24 at low ionic strength (10 mM tris, light grey bars), the same peptides at high ionic strength (10 mM Tris, 150 mM NaCl, dark grey bars), and for uncharged AHH24:1/AHH24:2 analogues (black bars). Results are shown for zwitterionic (a, c) and anionic (b, d) liposomes. Table 3. Helix Content of the Peptides in Solution and in the Presence of Zwitterionic (DOPC/Cholesterol) and Anionic (DOPC/DOPA/Cholesterol) Liposomes peptide
buffer (%)
zwitterionic liposomes (%)
anionic liposomes (%)
AKK6 AKK12 AKK18 AKK24 AKK24a
6 4 5 8 14
6 5 7 10 13
6 5 6 14 17
ARK8 ARK16 ARK24 ARK24a
6 6 7 13
6 7 10 14
6 5 17 14
AHH24:1 AHH24:2
7 12
7 11
7 12
a
Measured in buffer containing 150 mM NaCl.
differences in leakage induction between the zwitterionic and anionic liposomes (Figures 3). For the two liposome compositions studied, increasing the excess electrolyte concentration resulted in a strongly reduced leakage induction by both AKK24 and ARK24. Additionally, substituting all positively charged lysine and arginine residues by neutral histidine drastically decreased liposome leakage induction compared to that for AKK24 and ARK24 for both zwitterionic and anionic liposomes (Figure 4). Circular Dichroism. For all peptides investigated, the R-helix content is low, and the main conformation in buffer as well as in the presence of zwitterionic and anionic liposomes is random coil (Table 3). On the basis of the peptide adsorption results (see below), we expect most peptides to be associated with the liposomes in the presence of the latter. However, to check for the contribution to the observed CD signal from unbound peptide
we performed control experiments where the liposome concentration was doubled. The effect of such a doubling of liposome concentration on the determined helix content was within the measurement uncertainty (about 2%, not shown). Hence, we conclude that the signal contribution to the CD results from unbound peptide is negligible. The secondary structure is not markedly affected by the peptide length, although the presence of anionic liposomes gives rise to an increased R-helix content for the longer peptides. As expected from helix-coil theory,27 the addition of salt increases the R-helix content of AKK24 and ARK24 in buffer but only to a minor degree (Table 3). Interestingly, substituting lysine and arginine by histidine also has little influence on the R-helix content, and the helical content of AHH24:1 and AHH24:2 was not affected by presence of liposomes. Ellipsometry. Adsorption to zwitterionic and anionic supported bilayers is highly dependent on the peptide length, as can be seen by comparing both AKK6 with AKK24 (Figures 5) and ARK8 with ARK24 (Figure 6). The adsorption to zwitterionic bilayers increases with increasing peptide concentration in a similar manner as the leakage for both AKK24 and ARK24. Saturation adsorption was found to be 51 ( 1 and 49 ( 19 nmol/m2 for AKK24 and ARK24, respectively, corresponding to phospholipidto-peptide ratios of 65 and 68, respectively. Little or no difference in adsorption from buffer was observed for zwitterionic bilayers in the absence and presence of 150 mM NaCl for either AKK24 or ARK24. When the charge on the peptide was eliminated by amino acid substitution (AHH24:1 and AHH24:2), however, the adsorbed amount decreased compared to that for the corresponding charge analogues (Figures 5a and 6a). (27) Nilsson, S.; Zhang, W. Macromolecules 1990, 23, 5234-5239.
Peptide-DOPC/DOPA Bilayer Interactions
Langmuir, Vol. 22, No. 11, 2006 5047
Figure 5. Effect of peptide length (AKK6 compared to AKK24) and electrostatic interactions (AKK24 compared to AHH24:1, i.e., with reduced charge, and AKK24 at high electrolyte concentration) on the adsorption to zwitterionic (a) and anionic (b) bilayers measured by ellipsometry for AKK6, AKK24, and AHH24:1. *Measured in buffer containing 150 mM NaCl. (c) Adsorption of AKK24 to silica coated by 0.045 ( 0.01 mg/m2 poly-L-lysine.
The adsorption of AKK24 and ARK24 to anionic bilayers (Figures 5b and 6b) yields adsorbed amounts approximately 4 times higher than those for the zwitterionic bilayers (Γmax ) 194 ( 9 nmol/m2 for AKK24 and 181 ( 31 mg/m2 for ARK24, corresponding to about 17 and 18 phospholipid molecules per peptide molecule, respectively); otherwise, the adsorption isotherms are qualitatively similar to those observed for the zwitterionic bilayers for both AKK24 and ARK24. In contrast
to the results obtained for zwitterionic bilayers, however, the presence of 150 mM NaCl decreases the adsorption of both AKK24 and ARK24 compared to that in the absence of 150 mM NaCl. As with the zwitterionic membranes, peptide charge elimination gave rise to reduced adsorption, but in the case of anionic bilayers, the adsorption profile was more similar to the profile obtained when measurements were made in 150 mM NaCl (Figures 5b and 6b). To estimate peptide adsorption in
5048 Langmuir, Vol. 22, No. 11, 2006
Ringstad et al.
Figure 6. Effect on peptide length (ARK8 compared to ARK24) and electrostatic interactions (ARK24 compared to AHH24:2, i.e., with reduced charge, and ARK24 in high electrolyte concentration) on the adsorption to zwitterionic (a) and anionic (b) bilayers measured by ellipsometry for ARK8, ARK24, and AHH24:2. *Measured in buffer containing 150 mM NaCl.
residual bilayer defects, peptide adsorption was also measured at silica coated with poly-L-lysine preadsorbed to 0.045 ( 0.01 mg/m2 under identical conditions to those used in the deposition of the anionic bilayer. The results for AKK24 presented in Figure 5c show that the adsorption to the substrate is low, possibly occurring through adsorption at openings in the incomplete polylysine adsorbed layer (background adsorption for AKK24 at bare silica under these conditions is 138 nmol/m2), and that defect adsorption can be excluded as a complicating factor for the systems investigated. Peptide concentration is expressed in absolute terms throughout the article. In the ellipsometry experiments, translation into peptide/lipid ratios is not trivial because of the unknown deposition of the lipid onto the quartz cuvelle walls. Assuming this to be equal to that of the silica surface, however, a peptide concentration in the cuvette of 1 µM corresponds to a P/L ratio of approximately 0.4. Zeta Potential. The z potential of the zwitterionic and anionic liposomes without added peptide was -7 ( 1 and -27 ( 8 mV, respectively (Figure 7). Although the addition of AKK6 did not markedly affect the z potential of the liposomes, the addition of AKK24 increased the z potential of both anionic and zwitterionic liposomes with increasing peptide concentration and even caused a charge reversal in the case of anionic liposomes at a peptide-
Figure 7. z potential for zwitterionic (a) and anionic (b) liposomes in the presence of AKK6 (O) and AKK24 (b). The z potential of the liposomes without added peptides is also shown (().
to-lipid ratio of 1. Similar trends were observed for ARK8 and ARK24 (results not shown). Liposome Size. To evaluate whether peptide-induced liposome leakage correlated with liposome flocculation and coalescence, the liposome size was monitored (Figure 8). Without added peptides, the liposome size was 141 ( 3 and 143 ( 15 nm for zwitterionic and anionic liposomes, respectively. The addition of peptides, largely irrespectively of peptide length, increased the size of the liposomes by a rather marginal 9 to 10% for the zwitterionic liposomes and 12 to 13% for the anionic liposomes.
Peptide-DOPC/DOPA Bilayer Interactions
Figure 8. Liposome size presented as the Z-average diameter for zwitterionic liposomes (black bars) and anionic liposomes (grey bars) without added peptide and with AKK6 and AKK24 at a peptideto-lipid ratio of 1.
Discussion RDA and viable count experiments, leakage measurements, and ellipsometry/z potential measurements show that increasing the peptide length increases the bactericidal activity, leakage induction, and peptide adsorption, respectively. Together, these results indicate that peptide-length-dependent adsorption strongly contributes to the length-dependent leakage induction in liposomes and suggest that peptide density at the lipid bilayer promotes the killing of bacteria by these peptides. A correlation between peptide length and antibacterial activity has also been observed for other antimicrobial peptides.28 The R-helix content, which in many previous studies has been shown to be induced by the presence of membranes and is of great importance to the antimicrobial activity of a number of peptides,5,7,29-33 seems not to be a significant factor in membrane lysis by the presently investigated peptides. Thus, for all peptides investigated the R-helix content is low, as has also been found previously for Cardin motif peptides interacting with heparin.14 Accordingly, the dominating conformation for these peptides is the random coil both in the absence and presence of zwitterionic and anionic liposomes. Furthermore, reducing the peptide charge density by substituting arginine and lysine by histidine or screening electrostatic interactions by the addition of 150 mM NaCl affects the peptide helix content in a way that is uncorrelated or weakly anticorrelated to the effect of these changes on liposome leakage. These findings suggests that lysis mechanisms based on ordered peptide assemblies spanning the lipid bilayer are less likely, although it is not possible on the basis of the present data to exclude the possibility of such mechanisms not requiring high helix contents. However, liposome leakage and peptide adsorption seem to be more strongly correlated. Not only does the adsorption increase with increasing peptide length for both peptide series investigated for the zwitterionic as well as for the anionic lipid systems, but it also decreases with decreasing peptide charge density and, in the case of the anionic lipid system, with increasing excess (28) Deslouches, B.; Phadke, S. M.; Lazarevic, V.; Cascio, M.; Islam, K.; Montelaro, R. C.; Mietzner, T. A. Antimicro. Agent Chemother. 2004, 49, 316322. (29) Dathe, M.; Meyer, J.; Beyermann, M.; Maul, B.; Hoischen, C.; Bienert, M. Biochim. Biophys. Acta 2002, 1558, 171-186. (30) Matsuzaki, K. Biochim. Biophys. Acta 1998, 1376, 391-400. (31) Oren, Z.; Lerman, J. C.; Gudmundsson, G. H.; Agerberth, B.; Shai, Y. Biochem J. 1999, 341, 501-513. (32) Pouny Y, S. Y. Biochemistry 1992, 31, 9482-9490. (33) Tossi, A.; Sandri, L.; Giangaspero, A. Biopolymers 2000, 55, 4-30.
Langmuir, Vol. 22, No. 11, 2006 5049
electrolyte concentration (Figures 5 and 6). These effects correlate well with decreased liposome leakage and, interestingly, also with decreased bacterial killing. Regarding the zwitterionic lipid system and excess electrolyte concentration, the correlation between leakage and adsorption is less straightforward because the adsorption not is significantly affected in this case. Peptides adsorbing at lipid membranes to an extent such as that found in the present investigation may deteriorate bilayer stability in several ways. First, simply diluting the lipid molecules a peptide not penetrating into the membrane interior allows for conformational relaxation of the aliphatic tails of the phospholipids, which results in bilayer thinning, as has been discussed for a number of peptides.34-37 It has been shown that this effect can also be observed in supported bilayers, where the bilayer thickness is reduced after peptide addition and remains constant when the peptide concentration is increased.38 In the systems presently investigated, however, the adsorbed layer thickness remained essentially unaffected by peptide adsorption within experimental uncertainty. Given the relatively large uncertainty ((3-7 Å) as well as the absence of a thickness increase following peptide adsorption, however, membrane thinning effects cannot be conclusively excluded at this point. Another possible mechanism of leakage induction could be the extensive peptide adsorption causing the buildup of a global electrostatic potential of the bilayer that would eventually destabilize the zero-curvature self-assembly, analogous to the effect of charged surfactants on zwitterionic liposomes.39,40 Although z potential measurements displayed charge reversal in the case of the anionic liposome, this did not occur in the case of the zwitterionic liposome. Quantitatively, the magnitude of the z potential in the absence of peptides in absolute terms is, in fact, higher for both the zwitterionic and the anionic systems, demonstrating that global electrostatic potential buildup is not the mechanism behind the adsorption-induced leakage (Figure 7). Additionally, liposome aggregation and/or coalescence resulting from peptide adsorption and reduced electrostatic repulsion between liposomes could be an explanation for peptide-induced liposome leakage, as could peptide-induced liposome-to-micelle transitions.41 Although we observe some minor increase in liposome size at high concentration of AKK24, at least for the anionic lipid system, we observe similar effects for the lowadsorbing AKK6 that were found not to affect the z potential in any appreciable manner (Figures 7 and 8). Given this as well as the relatively broad size distribution of the anionic liposomes and resulting uncertainties in the particle size determination, we infer that pepide-induced liposome aggregation and coalescence are unlikely to be reasons for liposome leakage. Also, on the basis of the absence of a major size decrease normally observed for liposome-to-micelle transitions,40 this mechanism of peptideinduced liposome leakage can be excluded. Although peptide adsorption at the lipid membranes seems to be a key process determining much of the defect formation in the systems investigated, it is not the only determinant. For (34) Chen, F.-Y.; Lee, M.-T.; Huang, H. W. Biophys J. 2003, 84, 3751-3758. (35) He, K.; Ludtke, S. J.; Heller, W. T.; Huang, H. W. Biophys. J. 1996, 71, 2669-2679. (36) Heller, W. T.; Waring, A. J.; Lehrer, R. I.; Harroun, T. A.; Weiss, T. M.; Yang, L.; Huang, H. W. Biochemistry 2000, 39, 139-145. (37) Ludtke, S.; He, K.; Huang, H. Biochemistry 1995, 34, 16764-16769. (38) Mecke, A.; Lee, D.-K.; Ramamoorthy, A.; Orr, B. G.; Banaszak Holl, M. M. Biophys. J. 2005, 89, 4043-4050. (39) Edwards, K.; Gustafsson, J.; Almgren, M.; Karlsson, G. J. Colloid Interface Sci. 1993, 161, 299-309. (40) Gustafsson, J.; Ora¨dd, G.; Almgren, M. Langmuir 1997, 13, 6956-6963. (41) Edwards, K.; Almgren, M.; Bellare, J.; Brown, W. Langmuir 1989, 5, 473-478.
5050 Langmuir, Vol. 22, No. 11, 2006
example and as mentioned above, the adsorption at high salt concentration is essentially as high as that at low salt concentration in the case of the zwitterionic system and only marginally lower for the anionic system, despite the liposome leakage being essentially zero at the high salt concentration for both liposome compositions (Figures 5 and 6). The adsorption at high salt concentration of ARK24 and AKK24 is also remarkably different for the zwitterionic system, despite the physicochemical properties of these peptides being very similar. Moreover, there is considerable adsorption, although significantly lower than that of the charged analogues, of the uncharged AHH peptide variants despite the leakage again being essentially undetectable. These findings show that nonelectrostatic peptide-lipid interactions contribute to the peptide adsorption driving force, which is not unexpected given the many nonpolar and heterocyclic peptide residues in the ARK/AKK and AHH peptides, respectively. It also shows that peptide adsorption is a necessary but not sufficient requirement for membrane defect formation and leakage induction, which suggest that membrane disruption possibly occurs in a two-step manner where the peptide is first attached to the membrane followed by lysis. The second step is affected by the extent of peptide penetration into the lipid layer.42 Together, the leakage and adsorption data imply that local electrostatic interactions (κ-1 ) 7.8 Å in 150 mM NaCl) play an important role in the defect formation caused by these peptides. Liposome z potential measurements showed that the zwitterionic liposomes had a negative z potential that was approximately 4 times smaller than the z potential for anionic liposomes (Figure 7), a ratio that is also seen in the peptide adsorption to the different bilayers. This, in turn, further indicates that electrostatic interactions are a major driving force for peptide adsorption. An interesting finding, however, is that leakage induction by AKK24, and to a smaller extent also by ARK24, is greater for the zwitterionic than for the anionic liposomes and also occurs at lower peptide concentrations for the zwitterionic liposomes. It is therefore less than straightforward to couple the differences regarding leakage induction for the zwitterionic and the anionic system to differences in adsorption because the charge-induced adsorbed amounts of AKK24 and ARK24 were approximately equal for the two bilayers (after correction for the bilayer z potential). Similar trends have been observed for ranacyclins, another group of AMPs with random coil structure in the presence of anionic and zwitterionic membranes.43 A possible explanation could be the lower negative surface charge of the zwitterionic membranes allowing the peptide to access the membrane interior to a larger extent in the case of zwitterionic membranes, whereas (42) Huang, H. W. Biochemistry 2000, 39, 8347-8352. (43) Mangoni, M. L.; Papo, N.; Mignogna, G.; Andreu, D.; Shai, Y.; Barra, D.; Simmaco, M. Biochemistry 2003, 42, 14023-14035.
Ringstad et al.
for anionic membranes the stronger electrostatic interactions can tether the peptide in the headgroup area to a greater extent.44 Alternatively, the difference between the zwitterionic and the anionic membranes can be explained by differences in headgroup area for the two phospholipids used. When the bilayers contain lipids with smaller headgroups, such as PA when effectively neutralized by the oppositely charged peptide, there is more space available for peptide adsorption in the bilayer, which implies that higher amounts of peptide are necessary to cause membrane thinning followed by membrane disruption.45 However, further investigations on the effect of peptide charge and local packing effects for the systems used in this study will be needed to support this mechanism hypothesis.
Conclusions Peptides of the Cardin motifs (AKKARA)n and (ARKAAKKA)n display peptide-length-dependent leakage inductions in liposomes, which to a significant extent are determined by peptide adsorption at the lipid membrane. This molecular-weightdependent adsorption, in turn, stems from the entropy penalty associated with the adsorption and is a generic phenomenon displayed by polymers, proteins, and other macromolecules. However, adsorption is not the only determinant for defect formation. Instead, charges on the peptide are essential for leakage induction. Although the detailed mechanism by which the adsorption of these structurally disordered peptides results in defect formation is unknown at present, the results point toward a combined effect of membrane thinning and local packing defects caused by the peptide charges, although pore formation in the sense traditionally seen for many AMP systems cannot be excluded at present. Acknowledgment. This work was supported by grants from the Swedish Research Council (projects 13471 and 621-20034022), the Royal Physiographic Society in Lund, the WelanderFinsen, Thelma-Zoegas, Groschinsky, Crafoord, Åhlen, Alfred O ¨ sterlund, Lundgrens, Lions, and Kock Foundations, The Swedish Government Funds for Clinical Research, and DermaGen AB. We also thank Ms. Lotta Wahlberg, Ms. Mina Davoudi, and Dr. Jonny Eriksson for expert technical assistance, Dr. Stefan Ulvenlund for valuable discussions, Professor Lennart Bergstro¨m for putting the z-potential and photon correlation spectroscopy instrumentation at our disposal, and Jovice Boon-Sing Ng for assistance with the z-potential measurements. LA060317Y (44) Dathe, M.; Schumann, M.; Wieprecht, T.; Winkler, A.; Beyermann, M.; Krause, E.; Matsuzaki, K.; Murase, O.; Bienert, M. Biochemistry 1996, 35, 1261212622. (45) Lee, M.-T.; Hung, W.-C.; Chen, F.-Y.; Huang, H. W. Biophys. J. 2005, 89, 4006-4016.
