Subscriber access provided by NEW YORK UNIV
Article
Membrane sterols modulate the binding mode of amphotericin B without affecting its affinity for a lipid bilayer Anna Neumann, Mi#osz Wieczór, Joanna Zieli#ska, Maciej Baginski, and Jacek Czub Langmuir, Just Accepted Manuscript • DOI: 10.1021/acs.langmuir.5b04433 • Publication Date (Web): 23 Mar 2016 Downloaded from http://pubs.acs.org on March 24, 2016
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Langmuir is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 34
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Langmuir
Membrane sterols modulate the binding mode of amphotericin B without affecting its affinity for a lipid bilayer Anna Neumann,† Milosz Wieczor,‡ Joanna Zielinska,¶ Maciej Baginski,† and Jacek Czub∗,‡ Department of Pharmaceutical Technology and Biochemistry, Gdansk University of Technology, Gdansk, Poland, Department of Physical Chemistry, Gdansk University of Technology, Gdansk, Poland, and Department of Pharmaceutical Chemistry, Medical University of Gdansk, Gdansk, Poland E-mail:
[email protected] Abstract Membrane-active antibiotics are known to selectively target certain pathogens based on cell membrane properties, such as fluidity, lipid ordering and phase behavior. These are in turn modulated by the composition of a lipid bilayer, and in particular by the presence and type of membrane sterols. Amphotericin B (AmB), the golden standard of antifungal treatment, exhibits higher activity towards ergosterol-rich fungal membranes, which permits its use against systemic mycoses; however, the selectivity for fungal membranes is far from satisfactory leading to severe side effects. Despite decades of ∗
To whom correspondence should be addressed Department of Pharmaceutical Technology and Biochemistry, Gdansk University of Technology ‡ Department of Physical Chemistry, Gdansk University of Technology ¶ Department of Pharmaceutical Chemistry, Medical University of Gdansk †
1
ACS Paragon Plus Environment
Langmuir
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
research, no consensus has emerged on the origin of AmB specificity for fungal cells and its actual mode of action at the molecular level. Previously, it has been proposed that the specific action of AmB is related to differences in its affinity for membranes of different composition. In this work, we investigate this relationship by employing molecular dynamics simulations to compare the free energy of insertion of AmB into three types of membranes: a pure DMPC bilayer, and DMPC bilayers containing 30% of cholesterol or ergosterol. We analyze the orientation of AmB molecules within the bilayer in order to unambiguously establish their membrane binding mode, and relate the orientational freedom to the sterol-dependent tightness of lipid packing. Our results strongly indicate that the membrane insertion of AmB proceeds virtually to completion independent of membrane type, and hence the higher toxicity against fungal membranes may rather result from differences in subsequent oligomerization in the membrane and assembly of monomers into functional transmembrane pores. In particular, the latter could be facilitated by sterol-induced ordering of AmB molecules along the membrane normal, revealed by our free energy profiles. Moreover – in contrast to certain claims – we find no stable binding mode corresponding to the horizontal adsorption of AmB on the membrane surface.
Introduction Lipid bilayer properties, such as composition, thickness, fluidity and phase behavior, are known to modulate the activity of membrane-active agents, and therefore might be exploited to selectively target certain pathogens. 1,2 One of the major differences in the plasma membrane composition between organisms from different taxonomic groups involves the presence and type of membrane sterols. An example of a drug whose activity is sterol-dependent is amphotericin B (AmB; see Figure 1A), a polyene macrolide antibiotic, which – at therapeutically relevant concentrations – disrupts sterol-containing cell membranes of eukaryotic cells while at the same time being essentially inactive against sterol-free bacterial membranes. 3
2
ACS Paragon Plus Environment
Page 2 of 34
Page 3 of 34
Langmuir
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 ACS Paragon Plus Environment
Langmuir
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
the development of novel antimycotic agents. 13,14 To develop strategies to rationally design semisynthetic AmB derivatives with a markedly improved toxicity profile, it is fundamental to understand the mechanism of action of AmB in molecular detail. This, however, has proven to be a challenging task, and despite decades of effort many aspects of the AmB’s mode of action at the molecular level remain unclear. 15,16 AmB is considered to be a membrane-active agent, as it has been shown to disrupt the integrity of the cell membrane causing leakage of ions and other cellular components, eventually leading to cell death. 17,18 The commonly accepted view is that the increase in the membrane conductance is a result of sterol-dependent oligomerization of the amphiphilic AmB molecules into trans-membrane barrel-stave pores or other permeabilizing structures. 17,19 However, the precise nature of the AmB-induced permeabilization, including the mechanism of membrane insertion and pore formation, as well as the molecular basis for the role of sterols in dictating the drug’s selectivity, have thus far remained largely elusive. It has been suggested that the selective toxicity of AmB against fungal and parasitic cells may be related to the plasma membrane pool of the antibiotic molecules, available to oligomerize into a functional pore. 20–22 According to this model, the affinity of AmB for ergosterol-rich membranes is slightly higher than for membranes with cholesterol and much higher than for fluid membranes containing no sterols, as would be consistent with the degree of permeabilization. These varying affinities could originate from differences in global, sterol-modulated properties of lipid bilayers, such as acyl chain ordering and hydrophobic thickness, 23,24 or from the fact that AmB binds specifically to ergosterol and, less strongly, to cholesterol. 15,25–31 Notably, the difference in affinity should be most pronounced in case of AmB monomers, as this form of the drug is known to prevail at submicromolar, therapeutical concentrations. 17 Alternatively, even if AmB molecules are incorporated into various types of plasma membrane with similar affinities, the properties of the lipid environment or specific interactions with sterols may dictate the organization and dynamics of the membrane-bound antibiotic
4
ACS Paragon Plus Environment
Page 4 of 34
Page 5 of 34
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Langmuir
molecules and, therefore, affect the oligomerization process or the structure and stability of the transmembrane pore. In particular, the orientation of rigid and elongated AmB molecules within the membrane and their tilt-angle fluctuations are considered to be one of such differentiating features, as they are expected to vary markedly between various membranes. 32,33 In general, membrane-bound AmB molecules were suggested to assume two distinct orientations: i) roughly parallel to the bilayer normal, with the polar head (i.e., mycosamine and the carbxyl group) located at the membrane/water interface and the lactone ring buried within the membrane hydrocarbon core, 33–37 or ii) perpendicular to the bilayer normal with the hydrophobic polyene portion facing the membrane interior and the hydrophilic polyol chain exposed to the aqueous phase. 38–40 Both of these membrane binding modes might be relevant for the AmB mechanism of action. Indeed, on the one hand, the parallel orientation seems to be consistent with pore formation scenario as it should facilitate self-association into permeabilizing structures. On the other, the horizontal binding mode at the lipid bilayer surface was suggested to be responsible for extracting ergosterol from the fungal plasma membrane in the recently proposed sterol-depleting mechanism. 41 From this perspective, determining how the equilibrium between the two binding modes and the extent of orientational fluctuations depend on the lipid composition is crucial for the molecular-level understanding of the AmB membrane activity. Here, we used molecular dynamics simulations to determine the membrane binding mode of AmB and examine its affinity for lipid bilayers of different sterol composition. To this end, we computed the free energy profiles in three different membrane systems (ergosterol- and cholesterol-containing DMPC bilayers and sterol-free pure DMPC bilayer) along two reaction coordinates: one describing the transfer of a single AmB molecule from the aqueous phase to the membrane, and the other corresponding to the tilt of a single AmB molecule within a lipid bilayer. The results indicate that it is favorable for AmB to insert into all studied types of membranes, as the major driving force for insertion is hydrophobic. Furthermore, we show that in all three systems, the equilibrium orientation of AmB is, on average, parallel
5
ACS Paragon Plus Environment
Langmuir
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
to the membrane normal, with the polar head located at the bilayer/water interface and the lactone ring buried in the hydrocarbon core. The bilayer packing, however, modulates the extent to which the insertion depth and tilt of a membrane-bound AmB monomer can fluctuate. While sterol-free membranes allow the antibiotic molecule to tilt away from this vertical orientation and penetrate deeper into the membrane, or even span it entirely, AmB monomers in the sterol-containing bilayers remain oriented parallel to the bilayer normal and located within a single membrane leaflet. Notably, no free energy minimum can be associated with the horizontal location of AmB on the bilayer/water interface, contrary to what was suggested in several previous studies using DPPC-sterol monolayers, 38 as well as DPPC liposomes 40 and SDS micelles. 39
Methods Molecular model We studied (a) the insertion of an AmB molecule into lipid bilayers and (b) the orientation of a membrane-embedded AmB molecule with respect to the bilayer normal. In both cases, we used lipid bilayers of three different compositions: (i) a dimyristoylphosphatidylcholine (DMPC) membrane, (ii) a DMPC membrane with ∼30 mol % of ergosterol (Erg) and (iii) a DMPC membrane with ∼30 mol % of cholesterol (Cho). These systems will be referred to as DMPC/0, DMPC/Erg and DMPC/Cho, respectively. The initial structures of the bilayers were prepared using equilibrated models taken from our previous studies, for both (a) 27 and (b). 42 Originally, these bilayers were obtained from a model DMPC bilayer, where lipid molecules were placed on a grid in the xy-plane to form a liquid-crystalline lattice, and a desired number of randomly chosen DMPC molecules were substituted with sterols prior to energy minimization and a 50-ns period of NPT equilibration. Systems used in the simulations in (a) contained 1 membrane-embedded AmB molecule, 63 DMPC lipids (31 and 32 molecules per leaflet), 28 sterol molecules (in DMPC/Erg and DMPC/Cho) and 6
ACS Paragon Plus Environment
Page 6 of 34
Page 7 of 34
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Langmuir
4020 water molecules. As the surface area per molecule for AmB is roughly equivalent to that for DMPC, with values of 0.55–0.63 nm2 and 0.60–0.62 nm2 , respectively, 43,44 the AmBcontaining leaflet of the bilayer contained one DMPC molecule less than the other one. To prepare the starting structures for the simulations in (b), one of the AmB molecules was removed from the previously used systems where the AmB dimer were considered, 42 and one DMPC molecule was removed from the other leaflet in order to balance the membrane surface area. Thus, the structures contained 1 membrane-embedded AmB molecule, 97 DMPC lipids (48 and 49 molecules per leaflet), 44 sterol molecules (in DMPC/Erg and DMPC/Cho) and 3256 water molecules.
Molecular dynamics simulations All the simulations presented here were conducted using NAMD2.8. 45 The all-atom CHARMM force field was used. For both DMPC and sterol molecules, CHARMM36 was applied 46 so that the MD simulations could be performed in tensionless isothermal-isobaric ensamble (NPT). For the antibiotic molecule, we used the well validated CHARMM22-based set of parameters, 26,27,42,47,48 in which charges were obtained by fitting to the quantum-mechanical electrostatic potential (see Table S1 and SI file for a complete set of AmB parameters). The three-site TIP3P model consistent with CHARMM was used for water. In all simulations, the pressure was kept constant at 1 bar using the Langevin piston method. 49 The temperature was maintained at 310 K by means of Langevin dynamics. Longrange electrostatic interactions were calculated using the Particle Mesh Ewald method 50 with the real-space cutoff of 10 Å and the Lennard-Jones interactions were smoothly switched off between 11 and 12 Å. The lengths of all covalent bonds involving hydrogen atoms were constrained to equilibrium values using the SHAKE method. The SETTLE algorithm was applied to constrain the geometry of water molecules. The constraints permitted a time step ∆t = 2 fs to be used to integrate the equations of motion in the velocity Verlet algorithm.
7
ACS Paragon Plus Environment
Langmuir
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Free Energy Calculations (a) The free energy profiles for the insertion of an AmB molecule into the three considered membrane environments were obtained using the adaptive biasing force (ABF) method implemented in NAMD. 51–53 The reaction coordinate for this process was defined as the distance of AmB from the bilayer midplane, i.e., the distance between the center of mass (COM) of the lipid bilayer and the COM of AmB molecule projected on the z-axis (z-distance; see panel B in Figure 1). The pathway of the insertion process, 0.0 ≤ z ≤ 42.0 Å, was divided into 37 windows of uneven sizes to enable the key region of the membrane/water interface to be sampled more thoroughly with narrower windows: the spacing used was 1 Å in the ranges of 0–11, 15–23 and 31–35 Å, and 2 Å in the ranges of 11–15, 23–31 and 35–41 Å. The initial structures for each window were extracted from the steered MD simulations, in which the AmB molecule, initially embedded in the membrane at z ≈ 10–15 Å (depending on the system), 27 was slowly pulled (over a 100 ns timescale) either inside the membrane until z = 0 Å or outside the membrane until it was completely surrounded by water at z = 42 Å. Before the actual ABF simulations, the systems were allowed to relax for 10 ns with the z-distance harmonically restrained around the center of a given window. For each of these windows at least 400 ns of MD trajectory was generated. Instantaneous values of the force acting along z-distance were accrued in 0.1-Å-wide bins. The free energy profile, G(z), was recovered by integration of the mean force, and the insertion free energy was calculated as △G0 = −kB T ln P . Here, P is the corresponding partition coefficient, computed as R ze 1 exp(−G(z)/kB T )dz, where the upper integration limit ze was selected to correspond ze −z0 z0 to the maximum z-distance at which AmB and the bilayers still interact directly (ze = 27.5, 27.0 and 26.0 Å for the DMPC/Erg, DMPC/Cho and DMPC/0 systems, respectively). (b) The free energy profiles along the angular coordinate, θ, between the axis determined by the rigid and elongated AmB molecule and the bilayer normal (see panel C in Figure 1) in the three considered membranes were obtained using the umbrella sampling (US) method. 54 The AmB axis was defined as a vector connecting the COMs of two groups of atoms, marked 8
ACS Paragon Plus Environment
Page 8 of 34
Page 9 of 34
Langmuir
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 ACS Paragon Plus Environment
Langmuir
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ferences in affinity for cell membranes of different composition, we calculated free energy profiles for the insertion of an AmB molecule into three lipid bilayers: DMPC with 30% mol ergosterol (DMPC/Erg), DMPC with 30% mol cholesterol (DMPC/Cho) and a sterol-free, pure DMPC one (DMPC/0). The resulting profiles obtained for these three membranes are shown in Figure 2, and all proved to be well converged, as shown in Figure S1. Figure 2 clearly shows that a single AmB molecule inserts favorably into all three studied membranes: Erg- and Cho-containing DMPC bilayers as well as sterol-free DMPC membrane. For the membrane-bound state, i.e., when AmB’s center of mass (COM) is located within the bilayer leaflet (as defined by the average distance of the DMPC phosphate groups from the bilayer midplane, at z-distance less than 22.5, 22.0 and 21.0 Å for Erg-, Chocontaining and sterol-free membranes, respectively; see vertical lines in Figure 2), the free energy assumes negative values, relative to 0 kcal/mol arbitrarily chosen for the membraneunbound configurations. In fact, membrane-bound configurations are up to 14–15 kcal/mol more favorable than unbound ones. Importantly, especially in light of conflicting findings reported in recent monolayer studies, 37,38 there is no minimum for AmB located at the bilayer/water interface, which indicates that AmB monomers do not tend to accumulate on the membrane surface but rather penetrate into the lipid bilayer. Still, the exact location of AmB molecules can be to some degree affected by specific experimental conditions, such as protonation state or concentration of AmB, the use of charged lipids, the choice of solid substrate for Langmuir-Blodgett films, or micelle size, so that experimental findings need to be interpreted with the specific setup in mind. Certain previous studies also suggested that differences in the affinity of AmB for membranes of different composition may be responsible for the selective toxicity of this drug. 20–22 Accordingly, the fungicidal activity of AmB would stem from a higher concentration of the antibiotic achieved in the fungal membranes compared to that in mammalian membranes. This is in contrast with our estimate of the free energy of transfer of AmB from the water phase to the outer membrane leaflet, ∆G0 . Although the obtained values suggest that the
10
ACS Paragon Plus Environment
Page 10 of 34
Page 11 of 34
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Langmuir
insertion of AmB to Erg-rich membranes is slightly more favorable (−14.5 kcal/mol) than to Cho-rich and sterol-free membranes (−13.8 and −13.7 kcal/mol, respectively), these differences seem not to be significant enough to support the above hypothesis and to yield the selectivity required for therapeutic applications. Further, these results show that the insertion of AmB into sterol-free membranes proceeds virtually to completion, whilst it is known that in this environment, the membrane-permeabilizing activity of AmB is greatly reduced. We note that the time scale of our simulations might not have allowed to fully capture the formation of specific AmB-ergosterol complexes, which – as determined previously 27 – could provide an additional thermodynamical driving force of up to 2 kcal/mol in DMPC/Erg. Even in such a case, though, AmB would be expected to penetrate all membranes without significant differences in affinity in the therapeutic range of concentrations. Thus, the presence of sterols does not seem to significantly affect the partitioning of AmB between the membrane and aqueous phase, implying that AmB oligomerization in the bilayer and the subsequent formation of a transmembrane pore depend largely on specific AmB-sterol interactions and/or modulation of membrane properties by sterol molecules. This impact of sterols on the behavior of membrane-bound AmB is clearly visible in the free energy profiles in Figure 2, which indicate that the binding mode of AmB is quite similar in the DMPC/Erg and DMPC/Cho systems but differs markedly in DMPC/0. Most importantly, while in the presence of sterols minima are narrow and well-defined, in pure DMPC the free energy basin is substantially broader, implying that sterols strongly affect the transverse mobility of a membrane-bound AmB molecule. Indeed, it is well known that sterols enhance the ordering of lipid hydrocarbon chains and consequently increase the bilayer thickness. 55–57 As a result, in the presence of sterols, AmB occupies only a single favorable vertical location, corresponding to the deep free energy minimum (−15.5 kcal/mol) at 12.5 Å. Hence in the sterol-containing bilayers with the thickness of ca. 43–44 Å, 55 the 25 Å-long AmB molecule spans only one leaflet, with the polar head located at the bilayer/water interface (see Figure S2) and the lactone ring buried within the membrane hydrocarbon
11
ACS Paragon Plus Environment
Langmuir
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 12 of 34
core. In the less ordered and thinner (
