Article pubs.acs.org/Langmuir
Interaction of Amphotericin B with Lipid Monolayers F. Foglia,† G. Fragneto,‡ L. A. Clifton,§ M. J. Lawrence,† and D. J. Barlow*,† †
Institute of Pharmaceutical Science, King’s College London, Franklin Wilkins Building, 150 Stamford Street, London SE1 9NH, UK Institut Laue-Langevin, 6 rue Jules Horowitz, BP 156, 38042 Grenoble, Cedex 9, France § ISIS Pulsed Neutron and Muon Source, Science and Technology Facilities Council, Rutherford Appleton Laboratory, Harwell, Oxfordshire OX11 0QX, UK ‡
ABSTRACT: Langmuir isotherm, neutron reflectivity, and Brewster angle microscopy experiments have been performed to study the interaction of amphotericin B (AmB) with monolayers prepared from 1-palmitoyl-2-oleoylphosphatidylcholine (POPC) and mixtures of this lipid with cholesterol or ergosterol to mimic mammalian and fungal cell membranes, respectively. Isotherm data show that AmB causes a more pronounced change in surface pressure in the POPC/ergosterol system than in the POPC and POPC/cholesterol systems, and its interaction with the POPC/ergosterol monolayer is also more rapid than with the POPC and POPC/cholesterol monolayers. Brewster angle microscopy shows that, in interaction with POPC monolayers, AmB causes the formation of small domains which shrink and disappear within a few minutes. The drug also causes domain formation in the POPC/ cholesterol and POPC/ergosterol monolayers; in the former case, these are formed more slowly than is seen with the POPC monolayers and are ultimately much smaller; in the latter case, they are formed rather more quickly and are more heterogeneous in size. Neutron reflectivity data show that the changes in monolayer structure following interaction with AmB are the same for all three systems studied: the data are consistent with the drug inserting into the monolayers with its macrocyclic ring intercalated among the lipid acyl chains and sterol ring systems, with its mycosamine moiety colocalizing with the sterol hydroxyl and POPC head groups. On the basis of these studies, it is concluded that AmB inserts in a similar manner into POPC, POPC/cholesterol, and POPC/ ergosterol monolayers but does so with differing kinetics and with the formation of quite different in-plane structures. The more rapid time scale for interaction of the drug with the POPC/ergosterol monolayer, its more pronounced effect on monolayer surface pressure, and its more marked changes as regards domain formation are all consistent with the drug’s selectivity for fungal vs mammalian cell membranes.
■
INTRODUCTION
transplant/implant related infections (arising from stem cell therapy, for example). The widely accepted (“textbook”) mechanism of action of AmB involves a change in the permeability of fungal cell membrane due to the formation of pores or channels following an interaction between AmB and the membrane. The ability of AmB to form pores is considered to be due to its amphiphilic nature, possessing as it does both a hydrophobic (polyene) and hydrophilic (polyhydroxy) side. The formation of the pores in the membrane is believed to lead to the indiscriminate passage of ions and other small molecules across the cell membrane, causing major disruption in the cell and, ultimately, cell death. It is important to note that the channels formed by AmB have been reported to occur in mammalian as well as fungal cell membranes and are the origin of AmB toxicity in the host. The membrane sterols, ergosterol and cholesterol, are thought to be important components of pore formation by AmB by being located between the individual staves in the barrel of the pore. Support for this so-called “sterol hypothesis” comes from a variety of sources including molecular dynamics simula-
Recent years have seen an increased resistance of pathogenic fungi to antifungal drug therapy,1−4 including increasing resistance to the mainstay of the antimycotic armory, amphotericin B (AmB, Figure 1). This increase in resistance has been accompanied by a sharp increase in the numbers of patients affected by fungal infections5,6 and in particular systemic infections in immuno-compromised patients6,7 and transplant- and implant-related infections due to mycotic biofilm development.8 As a consequence, there is an urgent need for the development of new antimycotic drugs. In order to achieve this goal, however, it is necessary to gain a detailed understanding of the molecular mechanisms of the currently available antifungal and the reason(s) for the reduced susceptibility of the resistant pathogens. AmB was first isolated in 1955 from Streptomyces nodosus (a filamentous bacterium). It is a nonaromatic polyene macrolide drug belonging to the heptaene class. Pharmaceutically AmB is a broad-spectrum antibiotic used to treat a range of different fungal (e.g., Aspergillosis, Candidiasis, Coccidioidomycosis, and Cryptococcosis) and protozoal infections (e.g., Leishmaniasis)9 and is particularly important in fighting deep-seated systemic mycoses in immuno-compromised patients (in HIV) or in © 2014 American Chemical Society
Received: May 14, 2014 Revised: July 4, 2014 Published: July 14, 2014 9147
dx.doi.org/10.1021/la501835p | Langmuir 2014, 30, 9147−9156
Langmuir
Article
Brewster angle microscopy (BAM), and specular neutron reflectivity (SNR) studies, performed on both phospholipid and phospholipid−sterol monolayers before and after their exposure to AmB.
■
MATERIALS AND METHODS
Materials. All chemicalsother than those indicated belowwere purchased from Sigma-Aldrich Ltd. (Gillingham, UK) and were used as received. Spectroscopic grade dimethyl sulfoxide (DMSO) and methanol were purchased from Fluka, U.K. Ltd. (Poole, UK). Water was spectroscopically pure and obtained either double distilled from a well-seasoned still or purified using a Millipore Milli-Q system to a resistivity of 18 MΩ cm. D2O (>99.7% purity) was obtained from Aldrich Ltd. (Poole, UK). Protiated palmitoyloleoylphosphatidylcholine (h-POPC) and palmitoyl chain deuterated d31-POPC were purchased from Avanti Polar Lipids (Alabaster, AL). Before use, the purity of the POPC and the sterols was confirmed by determining their surface pressure:area isotherm (data not shown) to ensure that no differences in isotherms were observed for the various isotopic forms of POPC and that the isotherms obtained for POPC and sterol isotherms agreed with those previously reported.21−23 Langmuir Surface Pressure Isotherms. Monolayers of POPC(/ sterol) were prepared by spreading dropwise using a Hamilton syringe 40 μL of a 1.25 mg mL−1 solution of POPC or 2:1 molar ratio POPC/ sterol mixture in 2:1% v/v chloroform:methanol onto the surface of clean, ultrapure water in a Nima Technologies 601 Langmuir trough (Nima Technologies, Coventry, UK). The Teflon trough had been scrupulously cleaned using chloroform, ethanol, and copious amounts of water. Further POPC(/sterol) monolayers were prepared containing AmB, made by spreading 40 μL of a 2.1% w/v chloroform:methanol solution of 1.25 mg mL−1 POPC or POPC/ sterol 2:1 molar ratio and 5 mol % AmB. For all Langmuir trough experiments, the volume of the aqueous subphase in the trough was kept constant at 350 mL. Once spread, the lipid monolayers were allowed to equilibrate for 10 min (to ensure the complete evaporation of the chloroform) before being compressed to determine the variation in surface pressure with area per molecule. The surface pressure area isotherms of various POPC-containing monolayers were measured at 295 ± 2 K in triplicate using a freshly prepared monolayer each time, with surface pressure being recorded using a Wilhelmy plate cut from Whatman No. 1 filter paper (Merck UK Ltd., Nottingham, UK), previously degreased by soaking in chloroformsuspended from a microbalance during the computer-controlled compression of the monolayer at a rate of 30 cm2 min−1. Kinetics of Interaction of Amphotericin with Lipid Monolayers. The POPC(/sterol) monolayers prepared as described above in the absence of AmB were compressed to a surface pressure of 25 mN m−1 at a rate of 30 cm2 min−1 and held at constant area and the surface pressure of the monolayers monitored to ensure it remained constant for periods of up to 40 min, the maximum time required to perform a specular neutron reflectivity experiment. Once a monolayer was deemed to be steady, 4 mL of the required strength solution of AmB in dimethyl sulfoxide (DMSO) was carefully injected into the subphase to achieve a POPC:AmB molar ratio of 0.05the same concentration as previously used in small-angle neutron scattering (SANS) experiments investigating the interaction of AmB with preformed membranes.24 The AmB concentration of the DMSO solution was determined by taking into consideration the amount of POPC(/sterol) spread on the surface of the subphase. Any effects of DMSO on the phospholipid(/sterol) monolayers were neglected, given that its concentration in the subphase was only 1.12% v/v. The time required for and extent of change in the surface pressure of the monolayers in the presence of AmB was monitored to establish the kinetics of the interaction. All experiments were performed in triplicate at 295 ± 2 K, using a freshly prepared monolayer each time. Brewster Angle Microscopy. The lateral morphology of the monolayers at the air−water interface was examined using a Brewster angle microscope (BAM) (Nanofilm_ep3bam, Accurion GmbH, Goettingen, Germany) equipped with a 532 nm laser and a p-
Figure 1. Chemical structure of amphotericin B.
tions,10,11 which suggest that the AmB/sterol complex comprises eight molecules of drug and eight sterol molecules.10 The chemotherapeutic applications of AmB are based on its ability to select between fungal and human cells by nature of the drug’s greater affinity for the fungal sterol, ergosterol, rather than cholesterol, the principal sterol in mammalian cell membranes.12 Despite this direct role for the sterol in the drug interaction with the cell membrane, it has also been suggested that the difference observed between AmB behavior in the fungal vs mammalian cell membrane arises because of differences in the changes the sterol induces in the physicochemical properties of the lipid bilayer of the membrane.13−15 Numerous experiments have been performed in an attempt to establish the details of the mechanism of action of AmB, but sadly the results are rather inconsistent. It has been variously suggested that a wide range of chemical and physical parameters including temperature, phospholipid phase, sterol structure, and the concentration of both sterols and drug are critical in determining the exact structure of the pores formed in the membrane.16−18 Despite all the research performed, however, there is still no direct structural evidence to support the formation of AmB channels in the cell membrane. For example, linear dichroism and FTIR experiments performed by Gagoś et al.19 seem to suggest that AmB binds horizontally or vertically in the lipid bilayer, depending on the nature of the sterol present, while recent neutron diffraction studies have indicated that the drug orients parallel with the bilayer normal and colocates with the membrane sterol in both phospholipid− cholesterol and phospholipid−ergosterol multilayers.20 As a means to securing an improved understanding of the mechanism of action of AmB and also to determine how the nature of the membrane sterol influences the drug’s effects on biomembranes, we here report Langmuir adsorption isotherms, 9148
dx.doi.org/10.1021/la501835p | Langmuir 2014, 30, 9147−9156
Langmuir
Article
Table 1. Scattering Lengths (∑b), Molecular Volumes, and Scattering Length Densities (ρ) of the Molecular Constituents of the Various Monolayers Studied component cholesterol ergosterol AmB
POPC
POPC plus 5 mol % AmB
POPC/chol
POPC/chol plus 5 mol % AmB
POPC/erg
POPC/erg plus 5 mol % AmB
whole molecule (h) whole molecule (h) whole molecule (h) headgroup (h) acyl chains (h) whole molecule (h/d31) headgroup (h) acyl chains (h/d31) whole molecule (h/d31) headgroup (h) acyl chains (h/d31) whole molecule (h/d31) headgroup (h) acyl chains (h/d31) whole molecule (h/d31) headgroup (h) acyl chains (h/d31) whole molecule (h/d31) headgroup (h) acyl chains (h/d31) whole molecule (h/d31) headgroup (h) acyl chains (h/d31)
polarizer and fitted with a 10× objective lens, mounted above a customized Langmuir trough (Nima Technology Ltd., Coventry, UK). The angle of incidence used was 53.15°, the Brewster angle for water. BAM images were recorded on a charge-coupled device (CCD) camera with the background subtracted, and the images were corrected geometrically using the nanofilm_ep3bam software; the size of each image is 535 μm × 430 μm. Monolayers, at 295 ± 2 K, were compressed (at a speed of 30 cm2 min−1) to a surface pressure of 30 mN m−1. The POPC/(sterol) monolayers were prepared by spreading 40 μL of POPC(/sterol) dispersion at a concentration of 1.25 mg mL−1 on the surface of the aqueous subphase contained in the customized NIMA Langmuir trough. In order to obtain a final molar ratio POPC(/sterol):AmB of 0.05, 4 mL of the required strength solution of AmB in DMSO was carefully injected into the subphase. Images were repeated using fresh monolayers to ensure reproducibility. Images of the monolayers were recorded before, immediately after, and 20 min after injection of the AmB in DMSO solution into the aqueous subphase. Specular Neutron Reflection at the Air−Liquid Interface. Specular neutron reflectivity (SNR) experiments were performed using the time-of-flight (TOF) reflectometer FIGARO at the Institute Laue Langevin (ILL), Grenoble, France. In the experiments, neutrons of wavelengths ranging between 2 and 30 Å and at two incident angles, namely 0.62° and 3.2°, were used giving a Q-range from ∼0.002 to ∼0.32 Å−1. Calibration of FIGARO was performed using a D2O subphase. Backgrounds were subtracted from the data by the simultaneous acquisition of off-specular data for each measurement on the area detector. Note that no off-specular scattering was observed for the monolayers under the conditions of the experiment. All SNR experiments were performed at a temperature of 295 ± 2 K. The SNR profiles of lipid and lipid/(sterol) monolayers were recorded for monolayers prepared using a Nima Technologies 601 Langmuir trough (Nima Technologies, Coventry, UK). In brief, the POPC/(sterol) monolayer was initially compressed to a surface pressure of 25 mN m−1 at a rate of 30 cm2 min−1 and then held at constant area with monitoring of surface pressure to ensure the formation of a stable monolayer. Once the stability of the monolayer was ensured, AmB in DMSO solution was injected into the aqueous subphase beneath the lipid monolayer, and any variation in surface
∑b/10−5/Å
mol vol/Å3
ρ/10−6/Å−2
13.2 27.3 147.3 67.7 75.0 33.4/356.1 60.0 −26.7/296.0 39.1/345.7 60.4 −21.6/284.9 27.3/253.2 60.0 −14.7/211.2 33.3/247.0 60.4 −10.2/204.4 31.5/257.5 60.0 −10.5/215.4 37.3/252.0 60.4 −6.2/208.4
655 658 983 229 754 1242 344 937 1229 338 923 1066 344 852 1062 338 847 1067 344 853 1063 338 848
0.20 0.41 1.50 2.96 0.99 0.27/2.87 1.74 −0.28/3.16 0.32/2.81 1.79 −0.23/3.07 0.26/2.38 1.74 −0.17/2.48 0.31/2.33 1.79 −0.12/2.41 0.30/2.41 1.74 −0.12/2.53 0.35/2.37 1.79 −0.07/2.46
pressure due to the adsorption of the AmB was recorded. When the interaction of AmB with the lipid monolayer was complete (i.e., no further change in surface pressure was recorded), the SNR of the monolayer was measured. The reflectivity of the AmB-containing monolayers was measured after compression of the monolayer to 25 mN m−1 and holding the monolayer at constant area to check that the monolayer was stable. Regardless of the method of preparation of the monolayer, SNR data were obtained using three different H/D contrasts: namely, h-lipid monolayer on D2O, d-lipid monolayer on D2O, and d-lipid monolayer on air contrast matched water (ACMW; a mixture of composition 92% H2O and 8% D2O by volume), in all cases with the AmB protiated. Analysis of the Specular Neutron Reflection Data. The scattering length density, ρ, of POPC and its component parts (Table 1) was computed by considering either the whole molecule using the measured density of the lipid (at 293 K) or, for the glycerophosphatidylcholine headgroup and acyl chain regions, the group volume increments. For monolayers containing AmB and/or sterol, an average molecular volume was calculated from the molar average of each component (assuming the lipid to be in the fluid phase; Table 1). Similar calculations were also performed for the POPC/(sterol) monolayers with incorporated AmB; for these systems the assumptions were made that the AmB macrocycle and cholestane/ ergostane moieties formed part of the hydrophobe layer, with the sterol hydroxyl and AmB mycosamine group included in the phospholipid headgroup layer (Figure 2). Model fits to the individual SNR profiles were initially obtained (manually) using the program AFIT.25 The results so obtained were then used to guide automated, constrained simultaneous fits performed using the Motofit software.26 In order to obtain realistic estimates for the thicknesses of the acyl chains and headgroup regions of the monolayers, the values of the lipid headgroup hydration and layer roughness were constrained within Motofit to the (physically reasonable) values optimized manually using A-Fit. For the pure POPC, POPC/chol, and POPC/erg systems, the levels of hydration within the headgroup layers were constrained as 23%, 33%, and 34%, respectively, with the addition of AmB inducing an increase of around 20−30%. All layer roughness values were constrained as 3 or 4 Å (the value of this parameter affecting only the quality of the fit to the 9149
dx.doi.org/10.1021/la501835p | Langmuir 2014, 30, 9147−9156
Langmuir
Article
AmB (data not shown). A comparison of the Langmuir isotherms recorded for the POPC−cholesterol and POPC− ergosterol mixtures with and without the addition of 5 mol % AmB shows that the inclusion of the AmB in the POPC/sterol dispersions causes a greater increase in area per molecule than was obtained from the addition of 5 mol % AmB to POPC (viz. ∼5 Å2). At 25 mN m−1, the changes in a0 are ∼10 and ∼15 Å2 for POPC/cholesterol and POPC/ergosterol monolayers, respectively. The effect of AmB on the ergosterol-containing system thus seems slightly more marked than its effect on the cholesterol-containing system. Effect of Amphotericin on Monolayer Surface Pressure. Following the injection of AmB beneath a POPC monolayer compressed to a surface pressure of 25 mN m−1 and held at constant area, at a calculated POPC:AmB molar ratio of 1:0.05, adsorption kinetic experiments (Figure 3, black line) suggest an
Figure 2. Cartoon showing the acyl chains and headgroup regions of the molecules used in calculating the scattering length densities (ρ) of the sublayers within the monolayer. Figure 3. Kinetics of the interaction of AmB (at a POPC:AmB molar ratio of 0.05) with monolayers of POPC without/with 30 mol % sterol; POPC (black line); POPC/cholesterol (blue line) and POPC/ ergosterol (red line) monolayers recorded at 295 ± 2 K and monitored by variation in surface pressurecommencing with a monolayer maintained at 25 mN m−1 (solid line). The surface area of the Langmuir trough during injection of AmB and subsequent interaction was kept constant (surface area of the trough recorded as dotted line). The injection of AmB into the subphase was made at t = 4000 s (see arrow). (All experiments were performed in triplicate to confirm reproducibility.)
measured reflectivity profiles in the region around QZ = 0.2 Å−1). The reported uncertainties on the various fitted parameters are given as the statistical errors derived from the variance−covariance matrix obtained in the Levenberg−Marquardt optimization.
■
RESULTS AND DISCUSSION Characterization of POPC-Containing Monolayers. Surface Pressure−Area Isotherms. The isotherms recorded for h- and d31-POPC (data not shown) exhibit no marked phase transition and are consistent with those reported by previous workers.21,22 At a surface pressure of 25 mN m−1as used in the SNR experimentsboth the h- and d-POPC monolayers exist in an expanded phase with an area per molecule (calculated from the isotherm) of a0 ∼71 Å2. When either 30 mol % cholesterol or ergosterol is added to the POPC monolayer (data not shown), there is a condensing effect induced by the sterols, and the molecules comprising the monolayer become more closely packed. As expected, the isotherms for the sterol-containing POPC monolayers (data not shown) show no discernible phase transition. According to Smaby et al.,27 the POPC/sterol monolayers pass almost immediately from a gaseous phase to an expanded phase, and from the data presented here this is seen to occur at a0 of ∼55 and ∼56 Å2 for POPC/cholesterol and POPC/ergosterol, respectively. Given the molar ratio of the POPC and sterol and the interfacial areas of the pure sterols (∼37 and ∼39 Å2 for cholesterol and ergosterol, respectively), the recorded isotherms thus indicate monolayer condensations of ∼9%. Langmuir surface pressure area isotherms were also recorded for POPC(/sterol) monolayers spread containing 5 mol %
interaction between the AmB and POPC with an increase in surface pressure commencing at ∼40 min (± 2 min) after AmB injection and a steady surface pressure recorded after 10 min (± 1 min) postinjection, the total increase in surface pressure amounting to ∼1 ± 0.2 mN m−1. In the comparable experiment performed using a mixed POPC−cholesterol monolayer compressed to 25 mN m−1, the changes seen in the Langmuir adsorption are only slightly more pronounced than those seen with the POPC monolayer (Figure 5, blue line). The addition of the AmB to the POPC/chol monolayer results in a small increase of surface pressure of around 2 ± 0.5 mN m−1. The time to the onset of the change in pressure is less than for POPC (26 ± 2 min vs 40 ± 2 min), but the time to equilibration is rather longer (33 ± 1 min vs 10 ± 1 min). An entirely different result is obtained when AmB is injected underneath a POPC−ergosterol monolayer (Figure 5, red line). Here, the addition of AmB leads to a much larger increase in surface pressure of around 7 ± 0.5 mN m−1. The change in 9150
dx.doi.org/10.1021/la501835p | Langmuir 2014, 30, 9147−9156
Langmuir
Article
Figure 4. BAM images of monolayers of POPC without/with sterol: POPC (left-most images: A, D, G), POPC/cholesterol (center images: B, E, H) and POPC/ergosterol (right-most images: C, F, I) monolayers (maintained at 25 mN m−1) before (top panel), immediately after (center panel), and 20 min after (lower panel) injection of AmB into the aqueous subphase (at a POPC:AmB molar ratio of 0.05). Arrows highlight domains.
m−1 (from 29 to 30 mN m−1); and (iii) in the last 15 ± 1 min
pressure commences immediately after the injection and is complete within 25 min. A close inspection of the changing isotherm shows that the interaction is triphasic: (i) in the first 5 ± 1 min following injection the surface pressure increases by ∼4 mN m−1 (from 25 to 29 mN m−1); (ii) in the second 5 ± 1 min after injection the surface pressure increases by ∼1 mN
after injection the surface pressure increases by ∼2 mN m−1 (from 30 to 32 mN m−1). It can be calculated, therefore, that most of the interaction between AmB and the monolayer occurs within the first 5 ± 1 min. 9151
dx.doi.org/10.1021/la501835p | Langmuir 2014, 30, 9147−9156
Langmuir
Article
Brewster Angle Microscopy. Brewster angle microscopy (BAM) images of the POPC/POPC:sterol monolayers were recorded to determine whether their exposure to AmB gives rise to significant changes in their in-plane structure. With the surface pressure maintained at 30 mN m−1, all of the various POPC and POPC/sterol monolayers prior to AmB addition appear entirely homogeneous (see Figure 4A−C). In the case of the POPC monolayer, the BAM images indicate that there are domains formed directly after the AmB injection which, when first formed, are homogeneously distributed throughout the monolayer and around 10 μm diameter (Figure 4D). Over time, however, these domains are seen to shrink and within 5 min following injection they disappear completely (Figure 4G). For the POPC/cholesterol monolayer, there are no discernible changes seen in the BAM images immediately after the AmB injection (Figure 4E), but small, rather illdefined domains (∼4 μm diameter) are visible after ∼20 min (Figure 4H). There are domains also formed following the addition of AmB to the POPC/ergosterol monolayers (Figure 4F), but in this case the domains are initially rather heterogeneous in size (ranging from 2 to 40 μm diameter; Figure 4F) but become more uniformly sized over time, at 20 min following AmB injection stabilizing to 4−8 μm diameter (Figure 4I). Comparison of these findings with those reported on the basis of related BAM studies reported by others (cf. refs 28−32) is made problematic because of differences in the experimental protocols followed and/or because of differences in the compositions of the monolayers studied. For the BAM experiments detailed here, it may be noted that the monolayers are studied at a surface pressure that approximates to that expected in a cell membrane and that the lipid−sterol molar ratios employed are appropriate given the compositions of fungal and human plasma membranes (vide infra) and ensure homogeneity within the parent monolayer (pre-AmB exposure). Specular Neutron Reflectivity. The simultaneous fits to the SNR data for the POPC monolayers at π = 25 mN m−1 (Figure 5) give acyl chain and headgroup layer thicknesses (Table 2) that are of the same order as those reported by Kučerka et al. for POPC multilayers, viz., 13 ± 0.3 Å vs 13.5 Å and 7.7 ± 0.5 Å vs 5.5 Å.33 The molecular interfacial area is determined as 71.7 Å2,21−23 and the level of hydration is calculated as ∼7 H2O per lipid headgroup. The total thickness of the monolayer is determined as 20.7 Å (Table 2), and this is comparable with the extended length of a POPC molecule in the fluid state (∼23 Å). Simple extrapolation from these data, however, gives a value of 2 × 20.7 = 41.4 Å for the thickness of a POPC bilayer. For comparison, we note that the bilayer thickness derived through modeling of the SANS data for POPC vesicles is significantly smaller, viz., 36.5 ± 0.3 Å.24 This difference between the modeled and extrapolated bilayer thickness is high (∼5 Å) and cannot be explained just by experimental errors and/or the difference in lateral surface pressure between the monolayer and bilayer; it is more likely attributable to an interdigitation of the acyl chains of the lipids in the opposing leaflets of the bilayera hypothesis that is consistent with the observations recorded following recent neutron diffraction experiments.20 Comparison between the models used in fitting the SNR data for the POPC and POPC−sterol monolayers shows that
Figure 5. SNR data and best fits for POPC Langmuir monolayers recorded at a surface pressure of 25 mN m−1 and at 295 ± 2 K. The dchain contrast in ACMW is reported in the upper panel; the d-chain contrast in D2O in the center panel; the h contrast in D2O in the lower panel. Qz is the neutron momentum transfer, in Å−1. The structural parameters used to obtain the best fit data are given in Table 2.
the addition of 30 mol % of cholesterol or ergosterol to the POPC dispersion (Figure 6 and Table 2) results in an increase in the thickness of the acyl chains layer of around 4 Å and reduces the number of water molecules interacting with the lipid head groups (from 7 down to ∼5 per headgroup, irrespective of the type of sterol present in the monolayer). As for the previous monolayer data, the SNR model fits can be compared with the results obtained in the analyses of SANS data for the corresponding POPC/sterol vesicles. The difference between the modeled (41.4 ± 0.3 Å24) and extrapolated (2 × 25.0 = 50.0 Å) bilayer thicknesses is again very high (∼9 Å) and cannot be explained just by experimental errors. As argued previously, this difference in bilayer thickness seems likely to arise because of an interdigitation of the POPC chains within the bilayer, here perhaps made more significant because of the presence of sterols. Interestingly, the interfacial molecular areas derived through fitting of the monolayer SNR data indicate that the addition of cholesterol/ergosterol causes a condensation of ∼19% (Table 2), which, although significantly greater than the value calculated on the basis of the Langmuir isotherm data, is nevertheless close to the figure (of ∼13%) reported by Smaby et al.27 Specular Neutron Reflectivity Studies of Amphotericin Effects on Monolayer Structure. Model fits to the SNR data recorded following the addition of AmB to the subphase under POPC monolayers (Figure 7 and Table 2) show that the drug causes no change in the thickness of the headgroup region but does result in a 2 Å increase in the thickness of the lipid chain region; the added AmB also increases the level of solvent in the headgroup layer (from 23% to 51%) and decreases the lipid interfacial molecular area (from 71.7 to 61.5 Å2). Model fits to the SNR profiles obtained following the addition of AmB to the subphase beneath POPC−cholesterol and POPC−ergosterol monolayers (Figure 8 and Table 2) show that the drug again causes no significant change in the thickness of the headgroup region but does induce a slight 9152
dx.doi.org/10.1021/la501835p | Langmuir 2014, 30, 9147−9156
Langmuir
Article
Table 2. Results of two layer simultaneous fits of SNR data obtained for monolayers of POPC, POPC/cholesterol 70:30 mol % and POPC/ergosterol 70:30 mol % monolayers at 25 mN·m−1 and 295 ± 2K before and after the injection of AmB (at a POPC:AmB molar ratio of 0.05) into the sub-phase as well as for POPC(/sterol) monolayers spread containing 5 mol % AmB. Scattering length densities of the chains and head layers of the monolayers were fixed as detailed in Table 1 pure lipid parameter thickness/Å
hydration/%
roughness/Å
area per molecule/Å2
chains POPC POPC/chol POPC/erg POPC POPC/chol POPC/erg POPC POPC/chol POPC/erg POPC POPC/chol POPC/erg
13.0 17.0 17.0 0 0 0 3 4 4 71.7 50.7 50.7
± 0.4 ± 0.6 ± 0.4
following AmB injection head 7.7 ± 0.5 8.0 ± 0.8 8.0 ± 0.6 23 ± 2 33 ± 2 34 ± 2 3 4 4
±1 ±1 ±1
increase in the thickness of the chain layer (of ∼3 Å) and also increases the level of headgroup hydration and decreases the lipid interfacial molecular area (from 50 to ∼44 Å2). These findings (for both the POPC and POPC/sterol systems) were derived on the basis of model fits performed on the assumption that, at equilibrium, all of the added AmB was assimilated from the subphase into the monolayer. The validity of such an assumption was, in each case, confirmed through model fits to the SNR data obtained following the spreading of a 95:5 mol % mixture of POPC (or POPC/sterol) and AmB, with the resulting mixed monolayer maintained at 25 mN m−1 (Table 2). Extrapolation of the increase in monolayer thickness caused by the incorporation of AmB (into either POPC or POPC/ sterol monolayers) would suggest an increase in thickness of the corresponding vesicle bilayers of ∼4−6 Å, but such increases are much greater than those that were calculated from the SANS data obtained for the corresponding POPC and POPC/sterol vesicles, where increases of ∼3 Å were reported.24 Such differences, however, are readily explained by the lower surface pressure in the monolayer compared with the bilayer (25 mN m−1 vs 30−35 mN m−1), which would then mean that the AmB could more readily (and more deeply) insert into the less-compressed monolayer, thereby causing a greater extension of the lipid hydrocarbon chains and resulting in a more marked increase in layer thickness.
chains 15.0 19.0 19.0 0 0 0 4 4 4 61.5 44.4 44.5
± 0.4 ± 1.1 ± 0.5
±1 ±1 ±1
head 8.3 ± 0.7 9.0 ± 1.1 9.0 ± 0.5 51 ± 2 74 ± 2 52 ± 2 3 4 4
AmB-containing monolayer chains 15.0 19.0 19.0 0 0 0 4 4 4 61.5 44.4 44.4
± 0.5 ± ± 0.7 ± 0.4
head 8.3 ± 1.0 9.0 ± 0.8 9.0 ± 0.6 51 ± 2 62 ± 2 57 ± 2 3 4 4
±1 ±1 ±1
components determined as C16:0 (24%) and C18:1 (26%).34 The plasma membranes of mammalian epidermal cells are likewise reported to have phospholipids as their principal lipid component (accounting for ∼53%35), with ∼39% having a phosphatidylcholine headgroup,35 and with C16:0 and C18:1 fatty acids accounting for 26% and 19% of their acyl chains, respectively.35 In the studies reported here, therefore, we accordingly chose to prepare monolayers containing 1palmitoyl-2-oleoylphosphatidylcholine (POPC). As to the mole fraction of membrane sterol, we note that figures of 31 and 29 mol % ergosterol are reported for Saccharomyces cerevisiae36 and Canidida albicans and that the level of cholesterol in human cell membranes is typically 30−50 mol %.37,38 In light of these observations and given the additional practical constraint that membranes containing levels of sterol exceeding 30 mol % exhibit significant lateral inhomogeneity, we here chose to use cholesterol:POPC and ergosterol:POPC monolayers containing 30 mol % of sterol. The susceptibility of fungi to AmB varies according to the species and strain of organism, but the recorded values for the minimum concentration giving a 50% inhibition of fungal growth (MIC50) are typically in the range 0.1−1 μM.39 From our previous studies of AmB interaction with mixed lipid:sterol vesicles we know too that the changes in membrane thickness caused by the addition of AmB are exactly the same for all concentrations of drug in the range 1 μM − 100 μM.24 As a means to maximize the statistical reliability of any differences in the monolayer structure resulting from AmB addition, therefore, we here elected to add AmB into the subphase beneath the monolayers so as to give a concentration in the bulk of 100 μM. On the assumption then that all of the added AmB associates with and/or inserts into the monolayer (an assumption that was validated through measurements made for the corresponding POPC:sterol:AmB monolayers), this corresponds to a POPC:AmB ratio in the monolayer of 16:1 and to a sterol:AmB ratio of 5.7:1. For comparison, we note that some previous workers have reported studies performed using membranes prepared from fully saturated phospholipids (cf. refs 28−30), while others have used monounsaturated chain lipids but employed lipid:sterol ratios of 10:1 or even higher and AmB concentrations which give sterol:drug ratios of 1:1 or more (cf. refs 40 and 41).
■
CONCLUSIONS In usingas hererelatively simple model systems to investigate the interactions of amphotericin (AmB) with cell membranes, it is clearly essential that the model(s) employed approximate the real systems as closely as possible. Specifically, therefore, it is necessary to ensure that the chemical composition of the model membranes is close to that found in the relevant biological membranes and that the level of drug used is appropriate given its potency in regard to antifungal activity. As regards membrane composition, there are two critical issues, namely, the choice of lipid and the choice of lipid:sterol ratio. For the plasma membranes of the mycelial form of the fungus Candida albicans, Marriott34 reports the predominant lipid type as glycerophospholipid, with 50% having a phosphatidylcholine headgroup, and the principal fatty acid 9153
dx.doi.org/10.1021/la501835p | Langmuir 2014, 30, 9147−9156
Langmuir
Article
Figure 7. SNR data and best fits for POPC Langmuir monolayers containing AmB 5 mol %; at a surface pressure of 25 mN m−1; profiles were recorded and at 295 ± 2 K. The analysis here presented has been performed using the ρ of a lipid mixture containing AmB 5 mol %. The d-chain contrast in ACMW is reported in the upper panel; the d-chain contrast in D2O in the center panel; the h contrast in D2O in the lower panel. Qz is the neutron momentum transfer, in Å−1. The structural parameters used to obtain the best fit are given in Table 2.
POPC/cholesterol and POPC/ergosterol monolayers, but in the former case these domains are formed more slowly than is seen with the POPC monolayer, and they are ultimately much smaller (∼4 μm), while in the latter case, they are formed rather more quickly and are more heterogeneous in size (ranging from 4 to 8 μm). These differences in the kinetics of AmB interaction with POPC/ergosterol and POPC/cholesterol monolayers are consistent with the observation made on the basis of our earlier neutron diffraction studies of these systems,20 wherein the more pronounced perturbation caused by the drug to the location of cholesterol within the membrane was taken to indicate a more favorable entropic contribution to the drug’s interaction with ergosterol-containing membranes and this too is consistent with the results obtained in recent MD studies.42 The specular neutron reflectivity data, however, show that the changes in structure following the interaction of amphotericin with POPC, POPC/cholesterol, and POPC/ ergosterol monolayers are all much the same. In each system, the model fits to the measured neutron reflectivity data suggest that the drug inserts into the monolayers so that its macrocyclic ring colocalizes with the lipid acyl chains and the sterol ring systems, with its mycosamine headgroup colocalizing with the sterol hydroxyl and POPC head groups. The drug’s insertion into the monolayers leads to an increase in thickness of the monolayer (of ∼4 Å), an increase in the level of the lipid headgroup hydration (of ∼20%), and a decrease in the lipid interfacial molecular area (by ∼10 Å2 in the case of the POPC monolayers and ∼5 Å2 in the case of the POPC/sterol monolayers). The proposed model of AmB’s insertion and orientation within the lipid−sterol monolayers is consistent with the results obtained in 2H NMR studies.43 On the basis of these various studies, therefore, it is concluded that amphotericin inserts in a similar manner into POPC, POPC/cholesterol, and POPC/ergosterol monolayers
Figure 6. SNR data and best fits for monolayers of POPC containing 30 mol % sterol (a) POPC/cholesterol and (b) POPC/ergosterol Langmuir monolayers recorded at a surface pressure of 25 mN m−1 and at 295 ± 2 K. The d-chain contrast in ACMW is reported in the upper panel; the d-chain contrast in D2O in the center panel; the h contrast in D2O in the lower panel. Qz is the neutron momentum transfer, in Å−1. The structural parameters used to obtain the best fit are given in Table 2.
The Langmuir isotherm experiments reported here show that amphotericin exhibits different effects on the different monolayers: it causes a much more pronounced change in surface pressure in the POPC/ergosterol system than in the POPC and POPC/cholesterol systems (+7 mN m−1 vs 1−2 mN m−1), and its interaction with the former is also much more rapid than with the latter, reaching equilibrium after ∼15 min as against 50−60 min. The Brewster angle microscopy studies show that the lipid and mixed lipid/sterol monolayers are homogeneous prior to their exposure to amphotericin but on addition of drug show transient and/or permanent changes in their in-plane structure. In the case of the pure POPC monolayer, the amphotericin is seen to cause the formation of small (10 μm) homogeneously distributed domains which shrink and disappear within 20 min. The amphotericin also causes the formation of domains in the 9154
dx.doi.org/10.1021/la501835p | Langmuir 2014, 30, 9147−9156
Langmuir
Article
but does so with differing kinetics and with the resultant formation of quite different in-plane structures. The more rapid time scale for the interaction of the drug with the POPC/ ergosterol monolayer, its more pronounced effect on the monolayer surface pressure, and its more marked changes as regards domain formation are all consistent with the selectivity of the drug for fungal vs mammalian cell membranes.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail
[email protected]; Tel +44 (0) 207 848 4827; Fax +44 (0) 207 848 4800 (D.J.B.). Present Address
F.F.: Department of Chemistry, Christopher Ingold Laboratories, University College London, 20 Gordon Street, London WC1H 0AJ, United Kingdom. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS
■
REFERENCES
F.F. gratefully acknowledges the support that was provided by an EPSRC “New Generation Facility Users” grant (EP/ F021291/1). The authors also acknowledge the provision of neutron beam time at the STFC ISIS facility, Didcot, UK, and at the Institut Laue-Langevin, Grenoble, France, together with the associated experimental support provided by staff at these facilities.
(1) Vandeputte, P.; Ferrari, S.; Coste, A. T. Antifungal resistance and new strategies to control fungal infections. Int. J. Microbiol. 2012; 2012, article 713687. (2) Kanafani, Z. A.; Perfect, J. R. Resistance to antifungal agents: mechanisms and clinical impact. Clin. Infect. Dis. 2008, 46, 120−8. (3) Vermeulen, E.; Lagrou, K.; Verweij, P. E. Resistance in Aspergillus fumigatus: a growing public health concern. Curr. Opin. Infect. Dis. 2013, 26, 493−500. (4) Huang, M.; Kao, K. C. Population dynamics and the evolution of antifungal drug resistance in Candida albicans. FEMS Microbiol. Lett. 2012, 333, 85−93. (5) Pappas, P. G.; Alexander, B. D.; Andes, D. R.; et al. Invasive fungal infections among organ transplant recipients: results of the transplant-associated infection surveillance network (TRANSNET). Clin. Infect. Dis. 2010, 50, 1101−1111. (6) Brown, G. D.; Denning, D. W.; Gow, N. A. R.; Levitz, S. M.; Netea, M. G.; White, T. C. Hidden killers: human fungal infections. Sci. Transl. Med. 2012, 4, 165rv13. (7) Nesher, L.; Rolston, K. V. I. The current spectrum of infection in cancer patients with chemotherapy related neutropenia. Infection 2014, 42, 5−13. (8) Ramage, G.; Rajendran, R.; Sherry, L.; Williams, C. Fungal biofilm resistance. Int. J. Microbiol. 2012, 2012, 528521. (9) British National Formulary, BMJ Publishing Group Ltd., London, UK, March 2014. (10) Baginski, M.; Resat, H.; McCammon, J. A. Molecular properties of amphotericin B membrane channel: A molecular dynamics simulation. Mol. Pharmacol. 1997, 52, 560−570. (11) Baginski, M.; Sternal, K.; Czub, J.; Borowski, E. Molecular modelling of membrane activity of amphotericin B, a polyene macrolide antifungal antibiotic. Acta Biochim. Pol. 2005, 52, 655−658. (12) Wizke, N. M.; Bittman, R. Dissociation kinetics and equilibrium binding properties of polyene antibiotic complexes with phospholipid/ sterol vesicles. Biochemistry 1984, 23, 1668−1674.
Figure 8. SNR data and best fits for monolayers of POPC containing 30 mol % sterol (a) POPC-cholesterol Langmuir monolayers containing AmB 5 mol % and (b) POPC−ergosterol Langmuir monolayers containing AmB 5 mol % recorded at a surface pressure of 25 mN m−1 and at 295 ± 2 K. The d-chain contrast in ACMW is reported in the upper panel; the d-chain contrast in D2O in the center panel; the h contrast in D2O in the lower panel. Qz is the neutron momentum transfer, in Å−1. The structural parameters used to obtain the best fit are given in Table 2. Inset (c) shows an overlay of the SNR data and best fits for d-POPC (squares), d-POPC-cholesterol (circles), and d-POPC-ergosterol (stars) with 5 mol % added AmB. 9155
dx.doi.org/10.1021/la501835p | Langmuir 2014, 30, 9147−9156
Langmuir
Article
(33) Kučerka, N.; Tristram-Nagle, S.; Nagle, J. F. Structure of fully hydrated fluid phase lipid bilayers with monounsaturated chains. J. Membr. Biol. 2005, 208, 193−202. (34) Marriott, M. S. Isolation and chemical characterization of the plasma membranes from the yeast and mycelial forms of Candida albicans. J. Gen. Microbiol. 1975, 86, 115−132. (35) Gray, G. M.; Yardley, H. J. Lipid compositions of cells isolated from pig, human and rat epidermis. J. Lipid Res. 1975, 16, 434−440. (36) Schneiter, R.; Brugger, B.; Sandhoff, R.; Zellnig, G.; Leber, A.; Lampl, M.; Athenstaedt, K.; Hrastnik, C.; Eder, S.; Daum, G.; Paltauf, F.; Wieland, F. T.; Kohlwein, S. D. Electrospray ionization tandem mass spectrometry (ESI-MS/MS) analysis of the lipid molecular species composition of yeast subcellular membranes reveals acyl chainbased sorting/remodeling of distinct molecular species en route to the plasma membrane. J. Cell Biol. 1999, 146, 741−754. (37) Boesze-Battaglia, K.; Schimmel, R. J. Cell membrane compositin and distribution: implications for cell and lessons learned from photoreceptors and platelets. J. Exp. Biol. 1997, 200, 2927−2936. (38) Hui, S. W.; Stewart, C. M.; Carpenter, M. P.; Stewart, T. P. Effects of cholesterol on lipid organization in human erythrocyte membrane. J. Cell Biol. 1980, 85, 283−291. (39) Amirkia, V. D.; Qiubao, P. The antimicrobial index: a comprehensive literature-based anti-microbial database and reference work. Bioinformation 2011, 5, 365−366. (40) Matsumori, N.; Tahara, K.; Yamamoto, H.; Morooka, A.; Doi, M.; Oishi, T.; Murata, M. Direct interaction between amphotericin B and ergosterol in lipid bilayers as revealed by 2H NMR spectroscopy. J. Am. Chem. Soc. 2009, 131, 11855−11860. (41) Anderson, T. M.; Clay, M. C.; Cioffi, A.; Diaz, K. A.; Hisao, G. S.; Tuttle, M. D.; Nieuwkoop, A. J.; Comellas, G.; Maryum, N.; Wang, S.; Eno, B. E.; Wildeman, E. L.; Gonen, T.; Rienstra, C. M.; Burke, M. D. Amphotericin forms an extramembranous and fungicidal sterol sponge. Nat. Chem. Biol. 2014, 10, 400−406. (42) Neumann, A.; Baginiski, M.; Czub, J. How do sterols determine the antifungal activity of amphotericin B? Free energy of binding between the drug and its membrane targets. J. Am. Chem. Soc. 2010, 132, 18266−18272. (43) Matsumori, N.; Sawada, Y.; Murata, M. Mycosamine orientation of amphotericin B controlling interaction with ergosterol: steroldependent activity of conformation-restricted derivatives with an amino-carbonyl bridge. J. Am. Chem. Soc. 2005, 127, 10667−10675.
(13) Bolard, J. How do the polyene macrolide antibiotics affect the cellular membrane-properties. Biochim. Biophys. Acta 1986, 864, 275− 304. (14) Hsu-Chen, C. C.; Feingold, D. S. Polyene antibiotic action on lecithin liposomes - effect of cholesterol and fatty acyl chains. Biochem. Biophys. Res. Commun. 1973, 51, 972−978. (15) Czub, J.; Baginski, M. Modulation of amphotericin B membrane interaction by cholesterol and ergosterols - A molecular dynamics study. J. Phys. Chem. B 2006, 110, 16743−16753. (16) Brutyan, R. A.; McPhie, P. On the one-sided action of amphotericin B on lipid bilayer membranes. J. Gen. Physiol. 1996, 107, 69−78. (17) Cotero, B. V.; Rebolledo-Antunez, S.; Ortega-Blake, I. On the role of sterol in the formation of the amphotericin B channel. Biochim. Biophys. Acta 1998, 1375, 43−51. (18) Cohen, B. E. A sequential mechanism for the formation of aqueous channels by amphotericin-b in liposomes - the effect of sterols and phospholipid-composition. Biochim. Biophys. Acta 1992, 1108, 49−58. (19) Gagoś, M.; Gabrielska, J.; Dalla Serra, M.; Gruszecki, W. I. Binding of antibiotic amphotericin B to lipid membranes: monomolecular layer technique and linear dichroism-FTIR studies. Mol. Membr. Biol. 2005, 22, 433−442. (20) Foglia, F.; Lawrence, M. J.; Demè, B.; Fragneto, G.; Barlow, D. J. Neutron diffraction studies of the interaction between amphotericin B and lipid-sterol model membranes. Sci. Rep. 2012, 2, 778. (21) Weis, I.; Wezel, P. B.; Bähr, G.; Schwarz, G. Equations of state for POPX lipids at the air/water interface. A comprehensive study. Chem. Phys. Lipids 2000, 105, 1−8. (22) Gramlich, G.; Zhang, J.; Winterhalter, M. A long-lived amphiphilic fluorescent probe studied in POPC air-water monolayer and solution bilayer systems. Chem. Phys. Lipids 2001, 113, 1−9. (23) Mansour, H. M.; Zografi, G. Relationships between equilibrium spreading pressure and phase equilibria of phospholipid bilayers and monolayers at the air-water interface. Langmuir 2007, 23, 3809−3819. (24) Foglia, F.; Drake, A. F.; Terry, A. E.; Rogers, S. E.; Lawrence, M. J.; Barlow, D. J. Small-angle neutron scattering studies of the effects of amphotericin B on phospholipid and phospholipid-sterol membrane structure. Biochim. Biophys, Acta 2011, 1808, 1574−1580. (25) http://physchem.ox.ac.uk/∼rkt/afit.html. (26) Nelson, A. Co-refinement of multiple-contrast neutron/X-ray reflectivity data using MOTOFIT. J. Appl. Crystallogr. 2006, 39, 273− 276. (27) Smaby, J. M.; Momsen, M. M.; Brockman, H. L.; Brown, R. E. Phosphatidylcholine acyl unsaturation modulates the decrease in interfacial elasticity induced by cholesterol. Biophys, J. 1997, 73, 1492− 1505. (28) Minoñes, J.; Minoñes, J., Jr.; Conde, O.; Rodriguez Patino, J. M.; Dynarowicz-Latka, P. Mixed monolayers of amphotericin Bdipalmitoyl phosphatidyl choline: Study of complex formation. Langmuir 2002, 18, 2817−2827. (29) Minoñes, J., Jr.; Conde, O.; Minoñes, J.; Rodriguez Patino, J. M.; Seoane, R. Amphotericin B-dipalmitoyl phosphatidyl glycerol interactions responsible for the reduced toxicity of liposomal formulations: A monolayer study. Langmuir 2002, 18, 8601−8608. (30) Dynarowicz-Latka, P.; Seoane, R.; Minoñes, J., Jr.; Velo, M.; Minoñes, J. Study of penetration of amphotericin B into cholesterol or ergosterol containing dipalmitoyl phosphatidylcholine Langmuir monolayers. Colloids Surf., B 2003, 27, 249−263. (31) Dynarowicz-Latka, P.; Minoñes, J., Jr.; Conde, O.; Casas, M.; Iribarnegaray, E. BAM studies on the penetration of amphotericin B into lipid mixed monolayers of cellular membranes. Appl. Surf. Sci. 2005, 246, 334−341. (32) Minones, J., Jr.; Conde, O.; Dynarowic-Latka, P.; Casas, M. Penetration of amphotericin B into DOPC monolayers containing sterols of cellular membranes. Colloids Surf., A 2005, 270-271, 129− 137. 9156
dx.doi.org/10.1021/la501835p | Langmuir 2014, 30, 9147−9156
