Molecular Mechanism of Uptake of Cationic Photoantimicrobial

Article Cite This: J. Phys. Chem. B XXXX, XXX, XXX−XXX

pubs.acs.org/JPCB

Molecular Mechanism of Uptake of Cationic Photoantimicrobial Phthalocyanine across Bacterial Membranes Revealed by Molecular Dynamics Simulations Philipp S. Orekhov,*,†,‡,∇,○ Ekaterina G. Kholina,‡,∇ Marine E. Bozdaganyan,‡,∥ Alexey M. Nesterenko,§ Ilya B. Kovalenko,‡,∥,⊥,# and Marina G. Strakhovskaya‡,∥ †

Moscow Institute of Physics and Technology, Dolgoprudny 141700, Russia Department of Biology and §Belozersky Institute of Physico-Chemical Biology, Lomonosov Moscow State University, Moscow 119234, Russia ∥ Federal Research and Clinical Center of Specialized Medical Care and Medical Technologies, Federal Medical and Biological Agency of Russia, Moscow 115682, Russia ⊥ Astrakhan State University, Astrakhan 414056, Russia # Scientific and Technological Center of Unique Instrumentation of the Russian Academy of Sciences, Moscow 117342, Russia ○ Sechenov University, Trubetskaya 8-2, Moscow 119991, Russia ‡

S Supporting Information *

ABSTRACT: Phthalocyanines are aromatic macrocyclic compounds, which are structurally related to porphyrins. In clinical practice, phthalocyanines are used in fluorescence imaging and photodynamic therapy of cancer and noncancer lesions. Certain forms of the substituted polycationic metallophthalocyanines have been previously shown to be active in photodynamic inactivation of both Gram-negative and Grampositive bacteria; one of them is zinc octakis(cholinyl)phthalocyanine (ZnPcChol8+). However, the molecular details of how these compounds translocate across bacterial membranes still remain unclear. In the present work, we have developed a coarse-grained (CG) molecular model of ZnPcChol8+ within the framework of the popular MARTINI CG force field. The obtained model was used to probe the solvation behavior of phthalocyanine molecules, which agreed with experimental results. Subsequently, it was used to investigate the molecular details of interactions between phthalocyanines and membranes of various compositions. The results demonstrate that ZnPcChol8+ has high affinity to both the inner and the outer model membranes of Gram-negative bacteria, although this species does not show noticeable affinity to the 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphatidylcholine membrane. Furthermore, we found out that the process of ZnPcChol8+ penetration toward the center of the outer bacterial membrane is energetically favorable and leads to its overall disturbance and formation of the aqueous pore. Such intramembrane localization of ZnPcChol8+ suggests their twofold cytotoxic effect on bacterial cells: (1) via induction of lipid peroxidation by enhanced production of reactive oxygen species (i.e., photodynamic toxicity); (2) via rendering the bacterial membrane more permeable for additional Pc molecules as well as other compounds. We also found that the kinetics of penetration depends on the presence of phospholipid defects in the lipopolysaccharide leaflet of the outer membrane and the type of counterions, which stabilize it. Thus, the results of our simulations provide a detailed molecular view of ZnPcChol8+ “self-promoted uptake”, the pathway previously proposed for some small molecules crossing the outer bacterial membrane.

■

INTRODUCTION Antibiotic resistance is one of the biggest threats to global health, food security, and development today.1 The spread of severe infections caused by multidrug resistant bacteria has impact on healthcare, economy, and the community, as it leads to increased morbidity and mortality, prolonged length of hospital stay for patients, and higher medical costs.2,3 Reviewed in recent years, a wide range of strategies to combat drug resistance include new effective antibiotics, antimicrobial peptides, bacteriophages, enzymes, quorum sensing inhibitors, © XXXX American Chemical Society

reactive oxygen, photo-antimicrobials, and combined treatments.4−10 Antimicrobial photodynamic inactivation (aPDI) is a promising approach to eradicate bacterial pathogens in terms of water11 and blood disinfection12 and photodynamic therapy of localized infections.13 It utilizes photosensitizer (PS), Received: November 28, 2017 Revised: February 17, 2018 Published: March 19, 2018 A

DOI: 10.1021/acs.jpcb.7b11707 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B

render the outer bacterial membrane more penetrable is missing up to the present time. In this article, we present the first coarse-grained (CG) molecular model of polycationic ZnPcs with the cholinyl substituents compatible with the popular MARTINI force field. The latter has been recently extended to LPSs,30,31 making the investigations of permeation mechanisms of various antimicrobial compounds through the outer bacterial membrane feasible with molecular dynamics simulations. The developed MARTINI model of zinc octakis(cholinyl)phthalocyanine (ZnPcChol8+) was further used to examine behavior of polycationic ZnPcs in water solution and their interactions with bilayers composed of 1-palmitoyl-2-oleoyl-sn-glycero-3phosphatidylcholine (POPC), 1-palmitoyl-2-oleoyl-sn-glycero3-phosphatidylethanolamine (POPE)/1-palmitoyl-2-oleoyl-snglycero-3-phosphatidylglycerol (POPG) and LPS/dipalmitoylphosphatidyl-ethanolamine (DPPE), i.e., the models of eukaryotic and prokaryotic membranes. The results provide a molecular view of the ZnPcChol8+ penetration across different membranes as well as the particular mode by which polycationic phthalocyanines destabilize the outer membrane of Gram-negative bacteria.

molecular oxygen, and visible light of the appropriate wavelengths, absorbed by PS, to produce reactive oxygen species that kill targeted pathogen.14 Between two groups of bacteria, Gram-positive and Gramnegative, the latter are less sensitive for photosensitization due to more complicated structure of their cell walls that protect sensitive cellular targets from external agents.15 The cell wall of Gram-negative bacteria contains additional asymmetric bilayer outer membrane, with the inner leaflet composed mainly of phospholipids and the outer leaflet rich with lipopolysaccharides (LPSs). LPSs are amphiphilic molecules which include three domains: lipid A, inner and outer core oligosaccharides, and O-specific polysaccharide (O-antigen). LPSs bear a negative charge, which originates from phosphate, pyrophosphate, and carboxyl groups mainly located in the inner core region.16 Some of the O-antigen polysaccharides may also be acidic due to the presence of hexuronic acids, nonulosonic acids, and acidic nonsugar components, such as lactic, glyceric, pyruvic acids, amino acids, or phosphate.17 The LPS layer with many negative charges distributed inside it represents an effective permeability barrier against various external hydrophobic or anionic molecules. At the same time, negatively charged groups of LPSs attract different inorganic and organic cations. Divalent cations (Ca2+ and Mg2+) bind to phosphate groups of the inner core and form cross-links between the LPS molecules, which stabilizes the outer membrane and is essential for its integrity.16 Phthalocyanines (Pcs) are macrocyclic compounds composed of four isoindole units linked by aza nitrogen atoms with intensive absorption at far red−near infrared wavelengths.18 Pcs readily chelate a variety of metals in the center of tetrapyrrole rings forming metallo-organic macrocyclic coordination complexes, metallophthalocyanines. Among the latter, ZnPcs with long triplet lifetimes, high photostability, and ability to produce 1 Δg singlet oxygen under photoexcitation are considered as potent PSs in photodynamic therapy of tumors.19−21 However, low solubility of uncharged nonsubstituted ZnPcs in water and their tendency to form aggregates with reduced singlet oxygen yields are significant disadvantages of these dyes as antimicrobial PSs. Peripheral charged substituents prevent aggregation, greatly improve compatibility of ZnPcs with biological fluids and aqueous media, and increase singlet oxygen yield generation that reaches 0.60−0.65 values for octasubstituted compounds.22 Addition of peripheral substituents with the positively charged groups to ZnPcs results in compounds with the enhanced affinity to negatively charged units of bacterial cell walls via electrostatic interactions. Consequently, the cationic substituted ZnPcs induce photodynamic inactivation of both Gram-positive and Gram-negative bacteria.23−28 Despite the evident ability of cationic Pcs to photoinactivate the Gram-negative bacteria and to span their outer membrane, the molecular details of this process remain elusive. The molecular weight of these compounds does not allow them to follow the porin pathway. As an alternative, an idea of the “selfpromoted uptake” was suggested for cationic ZnPcs.24 According to this mechanism, the competitive displacement of the divalent cations (e.g., Mg2+ and Ca2+), which cross-links LPSs of the outer bacterial membrane by positively charged ZnPcs, results in distortion of the LPS leaflet of the outer membrane29 and consequent increase of the outer membrane permeability for ZnPcs themselves and other compounds as well. However, the comprehensive understanding of how ZnPcs

■

MATERIALS AND METHODS QM Calculations. The density functional theory calculations were performed with the hybrid functional B3LYP5, as implemented in PC GAMESS/Firefly.32 Geometry optimization calculations were carried out for the ground state molecule. At the first stage, we performed geometry optimization using the SBKJC basis set with energy-consistent pseudopotential (ECP).33 Final optimization was performed using cc-PVDZ basis set for all atoms but Zn.34 cc-pVDZ-PP was considered for valence electrons of Zn atom, as well as ECP10MDF core potential (Stuttgart/Cologne ECP) was exploited.35 All optimizations were performed in delocalized coordinates implemented in Firefly, version 8.32 Geometry optimizations of molecules of aa−bb type were performed within the C2 symmetry point group. Meso- and ab-type molecules were calculated within C4 symmetry group. Calculations of ab initio electrostatic potential were performed at points of the Connolly surface36 for equilibrium geometry with the protocol used for detailed energy minimization. Atomic partial charges were fitted to reproduce the ab initio electrostatic potential. Atomic partial charge calculations were performed using the RESP algorithm.37 Fukui functions representing a reactivity for electrophilic attack were calculated as a half of highest occupied molecular orbital density of a molecule. All-Atom Simulations. Topology for the ZnPcChol8+ was created using the ATB web-service38 and the GROMOS96 54a7 force field.39 The atomic partial charges were obtained as described above. A single ZnPcChol8+ molecule was placed in the water box (SPC water model)40 with the dimensions of 4.55 nm × 4.55 nm × 2.91 nm and an appropriate number of Na+/Cl− ions, which assured the ionic strength of 0.15 M and zero total charge. The system, which consisted of 5894 atoms, was subsequently simulated for 50 ns using Gromacs 5.1.4 software package.41 The temperature (303.15 K) and pressure (1 bar) were controlled by means of the Nosé−Hoover and the Parrinello−Rahman algorithms, respectively; the integration time step was 2 fs; and the Verlet cutoff scheme42 and particle mesh Ewald (PME) 43 were used for the nonbonded interactions with the cutoff value set to 1.2 nm. The coordinates of the Pc molecule were saved every 10 ps, B


Article


were adjusted on every ith iteration according to the following formulae

resulting in 5000 frames, which were further used for parameterization of the coarse-grained MARTINI model. CG Mapping of ZnPcChol8+. The all-atom model of ZnPcChol8+ was mapped to the CG representation (Figure 1)

l0i = l0i − 1 + 0.5·(l AA − l CG)

(1)

θ0i = θ0i − 1 + 0.5·(θ AA − θ CG)

(2)

⎛ σ AA ⎞ ki = 0.5·ki − 1·⎜ CG ⎟ ⎝σ ⎠

(3)

where lAA/θAA are mean values of the AA distributions for a bond/angle term; lCG/θCG are mean values of the CG distributions for a bond/angle term at the current iteration; i−1 i−1 li−1 are equilibrium bond length/angle and force 0 /θ0 and k constant values used at the previous iteration; li0/θi0 and ki are equilibrium bond length/angle and force constant values adjusted at the current iteration, and σAA and σCG are standard deviations of the AA distribution and the CG distribution at the current iteration. The histograms were considered matched if they overlapped by not less than 80%. We used standard MARTINI Lennard-Jones bead types for all particles and the corresponding MARTINI 2.2P interaction matrix.44,47 Only the beads corresponding to the choline groups, the metal center, and the four beads coordinating the metal center were modeled charged (+1, −1, and +0.25, respectively). Simulation Systems. The equilibrated AA model of the asymmetric outer bacterial membrane of Pseudomonas aeruginosa PAO1 was taken from48 and converted to the CG representation according to the reported MARTINI topology of Ra LPS (“rough”, i.e., lacking the O-antigen chains)31 and the standard MARTINI topology for DPPE. The obtained model consisted of 72 LPS in one leaflet and 180 DPPE molecules in another one, as well as 288 Ca2+ ions bound with the core moiety of LPS. The bilayer was further solvated with the polarizable CG water molecules.47 It was previously established experimentally49 as well as using the CG simulations50 that phospholipids and LPS tend to segregate forming separate domains. To reproduce in our simulations the formation of distinct lipid domains of adequate sizes, we have used an artificially increased concentration of DPPE. Although our model was the mixed outer bacterial membrane with the DPPE phospholipid present in both leaflets, we built and equilibrated a bilayer system containing 25% (molar) of Ra LPS and 75% DPPE in the outermost leaflet. The model inner bacterial membrane comprised 150 POPE molecules and 50 POPG. It was prepared and solvated using the insane.py script.51 The model eukaryotic membrane comprised 200 POPC molecules, and it was prepared in the same way. Additionally, an appropriate number of Na/Cl ions was placed into all simulation boxes to achieve the desired ionic strength (0.15 M) and to neutralize systems. Coarse-Grained Simulations. All simulations were performed using Gromacs 5.1.4.41 For each system, energy minimization using the steep descent algorithm was followed by equilibration simulation at constant pressure and temperature maintained by means of the Parrinello−Rahman barostat (time constant = 12.0 ps, compressibility = 3 × 10−4 bar−1 according to recommendations for complex membrane systems52) and the V-rescale thermostat, respectively. The PME method was used for the long-range electrostatics, and the relative electrostatic screening had a value of 2.5 since the polarizable water model was used. The time step of 20 fs was

Figure 1. Structural formula of ZnPcChol8+ molecule with the CG beads schematically mapped onto it. The beads are captioned with their names (as they appear in the topology), whereas the selected MARTINI bead types are shown in the legend.

following the guidelines used for the chemically related species (chlorophyll A and heme b) in ref 44: (1) CG beads consist of four heavy atoms, on average (an exception was made for the zinc-metal center, which was mapped to a separate bead); (2) CG beads should unite specific chemical groups; (3) the topology should account the molecular symmetry, and identical valent terms should be described by the same parameters; and (4) ringlike chemical patterns should contain at least three beads. The CG bead types of the macrocyclic region were assigned by analogy with those used for the model of chlorophyll A44 and for the side chain of phenylalanine in the standard MARTINI amino acid library. Two beads were used to represent each choline substituent: charged Q0 bead (used for choline of the phosphocholine-containing lipids in the standard MARTINI lipidome library) and polar SP1 (by analogy with the serine side chain although with the reduced size). The eventually chosen mapping scheme is largely determined by the correspondence between the distinct chemical groups, which are present in the Pc, and those already parameterized within the MARTINI force field, e.g., heme and choline groups.45,46 This allowed to adapt nonbonded parameters from the existing MARTINI topologies. Force-Field Parameters of the CG Model of ZnPcChol8+. The force-field parameters for the bonded interactions were iteratively optimized such that the bond and angle histograms obtained from a series of short (10 ns) CG MD simulations gradually converged to the target distributions derived from the accessory all-atom MD trajectory mapped to the CG resolution. To achieve it, the equilibrium length (angle) and the force constant values of a given term C



■

used for all simulations. The Verlet pair-lists cutoff scheme was used with the neighbor list updated every 10 steps. Periodic boundary conditions were used in all simulations. The water pore was characterized by calculation of the average area occupied by the water particles (assumed as spheres with the radius of CG beads = 0.26 nm) situated in membrane slices perpendicular to the membrane normal (thickness of each slice equaled 0.25 nm). Only the vicinity of the Pc molecule (cylinder centered at Zn of Pc with the radius = 2 nm) was taken into account in this analysis (Figure S8). The list of molecular systems (including their sizes and compositions) and performed simulations is provided in Table S4. The area per lipid was calculated for all systems and compared with experimental data53,54 and MD simulations55,56 (provided in Table S4) to control that the membranes were equilibrated. Calculation of the Free-Energy Profiles. The free-energy profiles of translocation of the ZnPcChol8+ molecule across bilayers were estimated using the umbrella-sampling (US) approach. First, the pulling simulations were performed, in which the ZnPcChol8+ center-of-mass (COM) was steered with the constant velocity (0.01 nm/ns) from the aqueous environment to the center of the hydrophobic region of the bilayer (with the force constant of 10 000 kJ/mol nm2). Then, for each system, the initial US configurations (windows) were picked from the pulling trajectories with the reaction coordinate (distance between the ZnPcChol8+ center-of-mass and the center of the hydrophobic region of the bilayer along the bilayer normal) decreasing from 7.0 nm (the LPS/DPPE system) or from 6.0 nm (the POPE:POPG system, the POPC system) to 0.0 nm with the 0.2 nm step. Subsequently, the US simulations were run for 400 ns in each US window, with the ZnPcChol8+ center-of-mass restraint at the window center by a harmonic potential (force constant of 1000 kJ/mol nm2). To obtain the free-energy profiles, the series of US simulations were combined by means of the weighted histogram analysis method, as implemented in the Gromacs software suite.57 Only the last 200 ns of each US trajectory were used for the freeenergy profile estimation, whereas the first 200 ns were retained for equilibration. Estimation of Aggregation Ability of Differently Charged ZnPc Species. The propensity of the ZnPc species with different charges (0, +4, and +8) to form aggregates in water solution was estimated from CG simulations (the concentration of the Pc species in the simulations was ∼8 mM) and from the experimental electronic absorption spectra. The pairwise matrix of minimal distances between centers-ofmass (COM) of Pc molecules was computed at every frame of MD trajectory. Then, the calculated matrix was clustered and Pc molecules were considered to belong to the same cluster when the distance between COM of any two molecules was less than 0.8 nm. To estimate the fraction of Pc molecules that form an aggregate of a given size, the number of clusters was counted for each frame of the MD trajectory. The electronic absorption spectra were measured on the Hewlett-Packard 8453 spectrophotometer. The fraction of light absorbed by the sample was calculated by integrating the overlapping of the transmission spectra of light filters and the absorption spectrum of Pc.22 The concentration of the Pc species was 5 mM.

Article

RESULTS AND DISCUSSION

Prediction of the ZnPcChol Isomers. Because of the features of chemical synthesis, the ZnPcChol8+ phthalocyanine appears as a mixture of isomers with the choline substituents at different positions of the macrocyclic core.58 However, the development of a biologically relevant model suitable for further extensive molecular simulations required us to determine the most plausible structure of ZnPcChol8+, which is likely prevailed among the products of synthesis. Addition of choline is performed in two stages. The first stage is electrophilic attack of the aromatic ring with chloromethylation agent that occurs eight times consequently. Choline next substitutes chlorine atom in the resulting molecule. Thus, to estimate possible positions of the choline residues in ZnPcChol8+, we estimated the most probable positions of chloromethyl addition. It is well known that the probability of electrophilic substitution can be calculated used the Fukui functions mapped onto the given molecular topology. Higher values of the f− Fukui function correspond to sites with higher probability of the electrophilic substitution. According to the calculated Fukui functions (see Figure S1), probability of the electrophilic attack to the isoindole rings is higher at positions a and a′ compared with positions b and b′. Taking into account these results as well as the steric effects and the electrostatic repulsion of the chloromethyl residues, we suggested that the first four substitutions will most probably take place at the a positions of the isoindole rings, implying a plausible isomeric form of the ZnPcChol4+ species. Following the analogous steric and electrostatic considerations, we propose that the next four substituents will be attached at the b′ positions with the highest probability on average resulting in their even distribution throughout the macrocycle. Thus, for the further investigations, we have chosen the meso isomer of ZnPcChol8+. It should represent one of the most abundant forms appearing upon the chemical synthesis, and hereby it should satisfactorily reflect the general properties of ZnPcChol8+. Parameters of the CG Model. The CG model of ZnPcs was developed within the MARTINI force-field framework, which achieved broad popularity over the last years59 as a common tool for the coarse-grained simulations of biomolecular systems. This is particularly due to a wide chemical space it covers60 and the straightforward parameterization scheme61 for those chemical compounds, which are not present in the standard library. The MARTINI force field is especially useful for probing permeation of various compounds across biological membranes.62 The set of CG parameters was obtained via the iterative procedure described in the Methods on the basis of the auxiliary all-atom simulation. The final CG model of ZnPcChol8+ consists of 62 beads (refer to Figure 1 for mapping). Since the macrocycle structure is very rigid and highly packed, even when using the compact S-type CG MARTINI beads, it experiences a heightened strain. To reduce this strain, we use the nonbonded exclusions between all of the beads forming the macrocycle. Also, we replaced several bonded terms with the constraints for the sake of stability and included several dihedral terms to the final topology to maintain planarity of the macrocycle. The obtained CG topology allows stable simulations with a time step of up to 20 fs. D


Article


Figure 2. Distributions of the bonded terms from the AA simulation (blue histogram) and from the CG simulation with the final set of optimized parameters (red histogram). Bonds (a)−(g) correspond to bond ids 1−7 in Table S1.

Figure 3. Distributions of the angle terms from the AA simulation (blue histogram) and from the CG simulation with the final set of optimized parameters (red histogram). Angles (a)−(f) correspond to angle id 1−6 in Table S3.

Aggregation of the Neutral and Charged Pc Species in Water. The tendency of differently charged Pc species to aggregate in water was estimated both experimentally and from the simulations. Experimental electronic absorbance spectra were obtained for water solutions of ZnPcChol8+ and ZnPcChol4+ species. The neutral ZnPcChol0 species are virtually insoluble in water, and the spectrum of their water solution could not be thus measured. The spectrum of ZnPcChol4+ has two maxima in the Q band region, ∼630 and ∼680 nm, whereas only the long wavelength maximum persists in the case of ZnPcChol8+ species. Previously, the Q band of the electronic absorbance spectra with the maximum near 680−690 nm was attributed to the monomeric form, whereas the appearance of the short wavelength peaks indicates the formation of aggregated forms, dimers (630−640 nm), or oligomers (600−630 nm).63,64 Upon the assumption that the solution of polycationic

Histograms for the angles and bonds from the AA simulation and the CG simulation with the final set of optimized parameters are shown on Figures 2 and 3. Most of the distributions derived from the CG simulation fit well the corresponding distributions from the mapped AA trajectory. Root-mean-square deviations of the mean and standard deviation for seven bonds were 0.05 and ±0.05 Å, respectively; for six angles, −0.82 ± 1.77°. We were not able to avoid an appearance of several multimodal distributions with the chosen mapping (Figures 2f,g and 3c,e). We tried to get rid of them by introduction of additional dihedral potentials, but it sufficiently impaired the stability of the simulations, so we preferred to stay with the developed model. Ultimately, the CG distributions were broad enough to cover all of the AA peaks in these cases, preventing undesirable neglect of any plausible configurations. The full topology along with the structure file in Gromacs format are provided in the Supporting Information. E


Article


repulsion between them and a huge number of the associated water particles. The results of molecular dynamics simulations appear to be in very good agreement with the experimental investigations of the aggregation ability of substituted phthalocyanines, indicating a sufficient degree of accuracy of the developed CG topologies. Our results show that addition of the positively charged choline substituents gradually decreases aggregability of Pcs with the octacationic Pc derivative showing no tendency to aggregate at all. Since aggregability is an important factor of cytotoxicity and bioavailability of Pc compounds,66 such simulations may serve as a simple and efficient way for the in silico prediction of their tendency to aggregate in solutions. Interaction of the Pc with Various Model Membranes. We have chosen the octacationic ZnPcChol8+ derivative to examine interactions of Pc with different types of biological membranes, as this species exhibits the greatest bactericidal effect among the substituted metallophthalocyanines.22,28 Three types of simplistic models of biological membranes were used in the simulations: (1) the symmetrical POPE:POPG bilayer as a model of the inner membrane of Gramnegative bacteria,67,68 (2) the asymmetrical “rough” LPS// DPPE bilayer as a model of the outer membrane of the clinically relevant Gram-negative bacteria P. aeruginosa,69 and (3) the symmetric POPC bilayer as a model of the eukaryotic membrane. Pc Binding. We performed equilibrium MD simulations to explore early steps of binding of Pc to the biological membranes with different composition. The initial adsorption process of 10 ZnPcChol8+ molecules was tracked over 1 μs long simulations. In the beginning of all simulations, the Pc molecules were distributed randomly in the aqueous phase. In the case of the LPS//DPPE and the POPE:POPG systems, the Pc molecules were readily adsorbed at the LPS leaflet of the former bilayer and at both leaflets of the latter one within the first 100 ns of the corresponding simulations (Figures 5 and S9a,c). For the LPS//DPPE system, we did not observe binding of Pc to the phospholipid (DPPE) monolayer (Figures 5 and S9a). Unlike these two membrane systems, the Pc molecules did not show any affinity to the POPC bilayer during the entire simulation time (Figures 5 and S9b). Moreover, they almost never appeared at the distances smaller than 3.5 nm from the bilayer center, i.e., at about 1.5 nm from the average position of the lipid phosphate groups, which is due to the presence of the hydration shell formed by the water particles around the Pc molecules in solution, as can be envisioned from the solvent radial distribution function provided in Figure S2. The binding of Pc at the LPS//DPPE bilayer should proceed as a multistep process due to the necessity for Pc to exchange with bivalent ions,70 which stabilize the LPS leaflet,71 and in the above-mentioned 1 μs long simulation, we could barely monitor this process entirely. Thus, we extended the simulation up to 60 μs and increased the simulation temperature up to 350 K to improve sampling and accelerate kinetics. Indeed, in such a simulation, we could observe a stepwise penetration of a single Pc molecule into the core region of LPS beginning at approximately 25 μs, which followed the initial rapid adsorption of all of the Pc molecules at the membrane surface (see Figure S3a−c). The total distance traversed by this Pc molecule from the average position of the Pc molecules adsorbed at the surface reached 1.2 nm at the end of the simulation. The migration of

substituted metallophthalocyanines contains solely monomers and dimers, the following formula for the fraction of dimers, α, was proposed22

α=

ελm − ελ ελm − ελd

d where ελ, εm λ , and ελ are the molar extinction coefficients at a wavelength λ for substituted phthalocyanine and its monomer and dimer forms (per phthalocyanine molecule in the latter case), respectively.22 According to this formula, the fraction of dimers in the ZnPcChol4+ solution was estimated as ∼0.6, whereas for ZnPcChol8+, the lack of the short wavelength maximum in the absorption spectrum points to the absence of any oligomeric aggregates in water. In accordance with the experimental results, the CG simulations of three Pc species with different number of charged substituents (0, 4, and 8 choline groups) demonstrate decrease of the aggregation ability with increase of the total Pc charge. The neutral Pc molecules are highly hydrophobic and start to aggregate in water shortly after the beginning of simulation. They form predominantly trimeric and tetrameric sandwichlike aggregates within the first 100 ns and remain in such oligomeric forms until the rest of the simulation (Figure 4b,c). The

Figure 4. Aggregation of differently charged Pc species. (a) Experimental electronic absorption spectra of differently charged Pc species dissolved in water. (b) Representative snapshots of an individual molecule of Pc8+ species, a dimer formed by the Pc4+ species and a tetrameric aggregate formed by the neutral Pc (from top to bottom). (c) Clustering of Pc molecules during 500 ns long CG simulations in water (10 molecules were simulated in each box). Color of a cell corresponds to the fraction of Pc molecules forming an aggregate of a given size (from monomer to tetramer).

observed aggregates resemble phthalocyanine nanopillars, which were recently found experimentally.65 The simulations indicate that the aggregated forms of ZnPcChol4+ are represented by dimers at least at the selected concentration. We found that the Pc4+ molecules existed in water solution in two forms: monomeric and dimeric, with a slight prevalence of dimers over monomers (α = 0.4−0.6), which is consistent with the experimentally estimated degree of dimerization. Meanwhile, the Pc8+ species remain in the monomeric form over the whole time of simulation due to the strong electrostatic F


Article


Figure 5. Density profiles along the bilayer normal for membranes of various compositions and ZnPcChol8+ molecules estimated from the adsorption simulations. (a) For the asymmetric LPS//DPPE bilayer, the profiles of Pc are shown for the initial adsorption simulation of 1 μs (solid line) and the last 15 μs of the extended simulation performed at 350 K (dashed line): deeper penetration of ZnPcChol8+ is designated with the arrow. (b) The LPS//DPPE bilayer with a phospholipid defect incorporated into its LPS leaflet (calculated for the last 15 μs). (c, d) The symmetric POPE:POPG and POPC bilayers (for the initial adsorption simulations of 1 μs). The boundaries of the hydrophobic regions are shown with the dashed lines (mean positions of SYB/XYA beads of LPS, which belong to the inner core; GLY (glycerol) beads of POPE, POPG, and POPC); the inter-monolayer plane is indicated by the bold black line.

Since the extended simulations of the Pc binding were run at the elevated temperatures, we roughly estimated the first passage time for the initial step of the Pc insertion into the LPS leaflet at the physiological temperature using the empirical van’t Hoff law (i.e., the reaction rate increases 2- to 4-fold for each 10 °C rise of temperature), which leads to the values of 1−25 ms for the pure LPS//DPPE system with Ca2+ counterions, 0.1−5 ms for the pure LPS//DPPE system with Na+ counterions, and 0.05−2 ms for the LPS:DPPE//DPPE system incorporating the phospholipid defect. It is important to note that in the present study, we have investigated interactions of ZnPcChol8+ with the LPS//DPPE bilayers consisting of the “rough” type of LPS, i.e., without Oantigen. The presence of O-antigens can prevent the Pc adsorption to the bacterial outer membrane. At the same time, “rough” LPS species appear often in the outer bacterial membranes, especially those of Pseudomonas sp. Most of the P. aeruginosa strains along with the full O-specific antigen contain the homopolymeric common polysaccharide antigen, OSAcapped, and uncapped forms of LPS,78 whereas P. aeruginosa PAO1 completely loses its ability to synthesize O-specific antigen78,79 at elevated temperatures. Also, the loss of the ability to produce O-antigen is common for some clinically relevant strains, e.g., the chronic mucoid strains of P. aeruginosa infecting the lungs of patients with cystic fibrosis,80 the “rough” strain of Pseudomonas putida, which was shown to be better adapted to the mixed-species biofilm environment comparing with the wild type strains.81 Therefore, the study of the Pc adsorption and penetration through membranes with the “rough” forms of P. aeruginosa LPS is a necessary first step

Pc inside the LPS core occurred as a transient jump, implying that this process is retarded by exchange with the bound ions. To demonstrate the essential role of bivalent ions for the kinetics of Pc penetration through the LPS membrane, we have carried out an additional simulation of the identical LPS// DPPE bilayer system together with 10 Pc molecules but with all Ca2+ ions replaced by an appropriate number of Na+ ions. In this case, we observed the overall acceleration of the penetration rate: the Pc molecules started to enter the core region just within the first 5 μs, with the total of three Pc specimen passed 1.4−2.0 nm by the end of the 30 μs long simulation (Figure S3d,e). The acceleration of Pc penetration was accompanied by more than 2-fold increase of the lateral diffusion of LPS (Table S5). Altogether, these observations are in a good agreement with the previously reported data.72,73 Even higher speeding up of binding was achieved when we used the LPS//DPPE bilayer with a phospholipid defect incorporated into its LPS leaflet. Similar defects are present in vivo74 along with other imperfections (e.g., porin proteins75), which break the homogeneity of the LPS monolayer of the outer membrane. We mimicked them by adding an admixture of DPPE phospholipids into the outer LPS leaflet of the modeled system following the approach from.76 DPPE and LPS molecules tended to separate from each other during the simulation (Figure S4) in agreement with the experimental studies.16,74,77 The first Pc molecules incorporated into the LPS domain just in 2−3 μs after the beginning of the simulation (Figure S3f,g), whereas more Pc molecules penetrated deeper (Figure 5b) inside the LPS domain than in the previous simulations with the pure LPS leaflet. G


Article

The Journal of Physical Chemistry B for understanding the mechanisms of photodynamic inactivation of P. aeruginosa and the sensitivity of clinical isolates and biofilm components to photosensitization depending on the LPS structure. Free-Energy Profiles of the Membrane Crossing. Despite the clear tendency for ZnPcChol8+ to penetrate into the LPS leaflet, this process occurs at a time scale of milliseconds at room temperature, which makes it impossible to trace it completely in the equilibrium simulations of attainable length. To overcome this limitation and to estimate the thermodynamic parameters of the translocation of ZnPcChol8+ across membranes, we used the enhanced sampling (umbrellasampling, US) MD simulations, which have proven themselves a powerful tool for revealing molecular details of interactions of small molecules with biological membranes.76,82,83 Here, the free-energy profiles of translocation of ZnPcChol8+ across three types of membranes were obtained: LPS//DPPE, POPE:POPG, and POPC. The free-energy profiles for POPE:POPG and POPC systems (Figure 6a) reveal that ZnPcChol8+ faces substantial free-energy barriers for membrane crossing in both cases. However, the behavior of ZnPcChol8+ near the water− membrane interfaces is different: for the POPE:POPG bilayer, the profile features a minimum of about 15 kT at around 1.4 nm from the membrane center, which appears due to electrostatic attraction of the oppositely charged headgroups of POPG and the Pc molecule and corresponds to the adsorption process observed in the equilibrium simulation. There is no free-energy minimum observed in the case of POPC: instead, the free-energy profile obtained for this system has a shallow maximum of ∼2 kT at 3.2 nm, which corresponds to destruction of the hydration shell around ZnPcChol8+ while it approaches to the bilayer. Indeed, in the corresponding equilibrium simulation (Figures 5d and S9), we did not observe any ZnPcChol8+ molecules closer than 3.5 nm to the center of the POPC bilayer. In contrast to the bilayers composed of phospholipids, the process of translocation of ZnPcChol8+ inside the LPS//DPPE membrane is energetically favorable (Figure 6a−d). The global minimum of the free-energy profile is situated just at 0.5 nm from the center of the hydrophobic slab of bilayer, which indicates that ZnPcChol8+ can spontaneously intrude toward the center of LPS//DPPE membrane, leading to its serious distortion and formation of a water pore piercing the whole bilayer. The formation of the continuous and stable water pore was taking place when the separation between ZnPcChol8+ and the membrane center was equal to or smaller than 1.8 nm, as indicated by the density profiles for water calculated across the membrane (Figure S5) as well as the calculated pore radius (Figure S8). This process is coupled to a complementary displacement of the polar moieties of phospholipids, which protrude toward the Pc molecule from the opposite side of the bilayer and form a hydrophilic coating of the pore (Figure S6). Similar bulges of the lipid headgroups were previously observed upon penetration of individual ions inside the hydrophobic region of membrane, and they appear due to the charge−dipole interactions of ions with the headgroups.84 The ZnPcChol8+ molecule serves as a bridge linking polar moieties of the LPSs and the phospholipids via adaptation of the almost perpendicular with respect to the bilayer plane orientation (Figure S7). Such orientation owes to the rigid macrocyclic scaffold of Pc, and might be relevant to the membrane

Figure 6. (a) Free-energy profiles of the ZnPcChol8+ species translocation from the aqueous solution toward the center of the hydrophobic membrane slab. (b−d) Representative snapshots of the Pc8+ immersed in the LPS//DPPE asymmetric bilayer, with the individual figures corresponding to the global minimum of the freeenergy profile (b) and two local minima (c, d). DPPE and LPS are shown as translucent cyan and pink sticks, respectively; beads corresponding to the negatively charged moieties of LPS, the choline and phosphate groups of DPPE as pink, ocher, and blue balls; Pc8+ as blue/gray/red spheres; Na+, Ca2+, and Cl− ions as cyan, yellow, and red balls; and water as the blue surface, respectively.

penetration mechanisms of other antimicrobial agents such as cyclic peptides. The free-energy profile for ZnPcChol8+ penetration inside the LPS//DPPE membrane also features two local minima at 2.7 and 1.4 nm. These minima correspond to the gaps between the average positions of the charged beads of the LPS molecules (compare Figure S3b,c). The escapes from these regions are likely associated with the overall distortion of the LPS leaflet and break of the connections between neighboring LPS molecules bridged by Ca2+ ions, leading to local rises of the free energy and a stepwise character of ZnPcChol 8+ penetration. Remarkably, the described above mechanism of ZnPcChol8+ uptake shares similarities with that proposed before for cyclic peptide antibiotics polymyxins.85,86 In the Langmuir monolayer assay, polymyxins showed low affinity for neutral lipids (phosphatidylcholine and phosphatidylethanolamine) but bound to LPS, PG, and cardiolipin monolayers. We also found that the kinetics of ZnPcChol8+ penetration is H


Article

The Journal of Physical Chemistry B significantly governed by the type of counterions stabilizing the LPS leaflet of the outer membrane. Substitution of bivalent counterions by monovalent ones speeds up the ZnPcChol8+ translocation toward the center of membrane. Similarly, for polymyxins, the binding to LPS monolayers was inhibited by Mg2+ ions, indicating antagonistic interactions of organic polycations and inorganic ones. “Self-Promoted” Uptake. We also detected a nonzero density of the sodium ions inside the membrane, which appeared upon the formation of pore (Figure S6). This observation confirms that the emerged transmembrane hydrophobic pore can serve as a pathway for passage of small hydrophilic compounds and provides a mechanistic view of the previously postulated “self-promoted uptake”.87,71 The latter mechanism states that the displacement of bivalent cations stabilizing the LPS leaflet of the outer membrane by the polycationic species disrupts integrity of the outer membrane either due to release of LPS into the aqueous phase resulting in the emergence of voids or due to distortion of the outer membrane by the bulky polycationic species.24 Here, we demonstrate that even a single polycationic Pc can render the outer bacterial membrane permeable for charged species. To examine whether the pore, which appears upon insertion of a single ZnPcChol8+ inside the membrane, can indeed facilitate further penetration of ZnPcChol8+ molecules according to the “self-promoted” scheme, we have carried out an additional 10 μs simulation, starting from the conformation corresponding to the global minimum of the free-energy profile (the final frame of the corresponding umbrella-sampling trajectory similar to the one shown in Figure 6b), with 10 extra ZnPcChol8+ molecules placed randomly in the aqueous region of the simulation box. In course of the simulation, all of the bulk ZnPcChol8+ molecules became adsorbed to the LPS leaflet of the membrane within the first 5 μs (Figure 7a,b). Subsequently, one of the Pcs was translocating toward the membrane center over the following 2.5 μs, meanwhile changing its orientation from predominantly parallel to the membrane surface to perpendicular (Figure 7c,d). During this penetration process, the ZnPcChol8+ molecule followed the contour of the preformed pore, which evidences that even local defects in the LPS leaflet can speed up penetration of hydrophilic species similar to larger phospholipid defects as demonstrated above.

Figure 7. “Self-promoted” uptake of ZnPcChol8+ species. (a−d) Snapshots of the uptake process at different time points (simulation time shown for each snapshot). DPPE and LPS are shown as translucent cyan and pink sticks, respectively; negatively charged moieties of LPS as pink balls, the choline and phosphate groups of DPPE as ocher and blue balls; Pc8+ as blue/gray/red spheres; Ca2+ ions as yellow balls; and water and bulk ions are not shown for the sake of clarity.

■

CONCLUSIONS In the present study, we developed an extension of the popular MARTINI coarse-grained force field for computer simulations of the choline-substituted Zn(II) phthalocyanine species, which correctly reproduces experimental data on their aggregability in water. This model was further used to simulate interactions of these potent antimicrobial agents with various biological membranes. The molecular dynamics simulations revealed that ZnPcChol8+ has high affinity to two model prokaryotic membranes imitating the outer and the inner membranes of Gram-negative bacteria but it does not show noticeable affinity to the zwitterionic POPC membrane mimicking the eukaryotic membrane. Moreover, the free-energy profiles of the ZnPcChol8+ translocation across these bilayers estimated by means of enhanced sampling simulations suggest that these Pc species can spontaneously translocate inside the outer bacterial membrane, inducing their distortion and formation of pores. The latter can provide a pathway for the self-promoted

penetration of the succeeding Pc molecules as well as of other compounds. We also found that the kinetics of penetration is significantly governed by the type of counterions stabilizing the LPS leaflet of the outer membrane and by the presence of phospholipid defects in it. Both substitution of bivalent counterions by monovalent ones and addition of phospholipids into the LPS leaflet speed up the ZnPcChol8+ translocation toward the center of membrane. Thus, our simulations shed light on the molecular mechanism of ZnPcChol8+ “self-promoted uptake”, which can have common features with the penetration pathways of other polycationic antimicrobial compounds, stimulating the rational design of novel antibacterial agents and improvement of the I


Article


for New Antimicrobials. Expert Opin. Pharmacother. 2014, 16, 159− 177. (4) Burrowes, B.; Harper, D. R.; Anderson, J.; McConville, M.; Enright, M. C. Bacteriophage Therapy: Potential Uses in the Control of Antibiotic-Resistant Pathogens. Expert Rev. Anti-Infect. Ther. 2011, 9, 775−785. (5) Worthington, R. J.; Melander, C. Combination Approaches to Combat Multidrug-Resistant Bacteria. Trends Biotechnol. 2013, 31, 177−184. (6) Brooks, B. D.; Brooks, A. E. Therapeutic Strategies to Combat Antibiotic Resistance. Adv. Drug Delivery Rev. 2014, 78, 14−27. (7) Scutera, S.; Zucca, M.; Savoia, D. Novel Approaches for the Design and Discovery of Quorum-Sensing Inhibitors. Expert Opin. Drug Discov. 2014, 9, 353−366. (8) Mahlapuu, M.; Håkansson, J.; Ringstad, L.; Bjö rn, C. Antimicrobial Peptides: An Emerging Category of Therapeutic Agents. Front. Cell. Infect. Microbiol. 2016, 6, 194. (9) Dryden, M. Reactive Oxygen Therapy: A Novel Antimicrobial. Int. J. Antimicrob. Agents 2018, 299. (10) Wainwright, M.; Maisch, T.; Nonell, S.; Plaetzer, K.; Almeida, A.; Tegos, G. P.; Hamblin, M. R. Photoantimicrobials-are We Afraid of the Light? Lancet Infect. Dis. 2017, 17, e49−e55. (11) Jemli, M.; Alouini, Z.; Sabbahi, S.; Gueddari, M. Destruction of Fecal Bacteria in Wastewater by Three Photosensitizers. J. Environ. Monit. 2002, 4, 511−516. (12) Marciel, L.; Teles, L.; Moreira, B.; Pacheco, M.; Lourenço, L. M.; Neves, M. G.; Tomé, J. P.; Faustino, M. A.; Almeida, A. An Effective and Potentially Safe Blood Disinfection Protocol Using Tetrapyrrolic Photosensitizers. Future Med. Chem. 2017, 9, 365−379. (13) Kharkwal, G. B.; Sharma, S. K.; Huang, Y.-Y.; Dai, T.; Hamblin, M. R. Photodynamic Therapy for Infections: Clinical Applications. Lasers Surg. Med. 2011, 43, 755−767. (14) Hamblin, M. R. Antimicrobial Photodynamic Inactivation: A Bright New Technique to Kill Resistant Microbes. Curr. Opin. Microbiol. 2016, 33, 67−73. (15) Jori, G.; Fabris, C.; Soncin, M.; Ferro, S.; Coppellotti, O.; Dei, D.; Fantetti, L.; Chiti, G.; Roncucci, G. Photodynamic Therapy in the Treatment of Microbial Infections: Basic Principles and Perspective Applications. Lasers Surg. Med. 2006, 38, 468−481. (16) Nikaido, H. Molecular Basis of Bacterial Outer Membrane Permeability Revisited. Microbiol. Mol. Biol. Rev. 2003, 67, 593−656. (17) Knirel, Y. A. Structure of O-Antigens. Bact. Lipopolysaccharides 2011, 41−115. (18) Fukuda, T.; Kobayashi, N. UV-Visible Absorption Spectroscopic Properties of Phthalocyanines and Related Macrocycles. Handb. Porphyrin Sci. 2010, 1−644. (19) Spikes, J. D. Photosensitizing Properties of Porphyrins in Model Cell Systems. Porphyrins Tumor Photother. 1984, 51−60. (20) Valduga, G.; Nonell, S.; Reddi, E.; Jori, G.; Braslavsky, S. E. The Production of Singlet Molecular Oxygen by Zinc(II) Phthalocyanine in Ethanol and in Unilamellar Vesicles. CHEMICAL Quenching and Phosphorescence Studies. Photochem. Photobiol. 1988, 48, 1−5. (21) Sibata, M. N.; Tedesco, A. C.; Marchetti, J. M. Photophysicals and Photochemicals Studies of zinc(II) Phthalocyanine in Long Time Circulation Micelles for Photodynamic Therapy Use. Eur. J. Pharm. Sci. 2004, 23, 131−138. (22) Makarov, D. A.; Kuznetsova, N. A.; Yuzhakova, O. A.; Savvina, L. P.; Kaliya, O. L.; Lukyanets, E. A.; Negrimovskii, V. M.; Strakhovskaya, M. G. Effects of the Degree of Substitution on the Physicochemical Properties and Photodynamic Activity of Zinc and Aluminum Phthalocyanine Polycations. Russ. J. Phys. Chem. A 2009, 83, 1044−1050. (23) Minnock, A.; Vernon, D. I.; Schofield, J.; Griffiths, J.; Parish, J. H.; Brown, S. T. Photoinactivation of Bacteria. Use of a Cationic Water-Soluble Zinc Phthalocyanine to Photoinactivate Both GramNegative and Gram-Positive Bacteria. J. Photochem. Photobiol. B 1996, 32, 159−164. (24) Minnock, A.; Vernon, D. I.; Schofield, J.; Griffiths, J.; Parish, J. H.; Brown, S. B. Mechanism of Uptake of a Cationic Water-Soluble

existing ones. At the same time, the developed coarse-grained model can be useful for molecular dynamics investigations of phthalocyanines and their derivatives in molecular pharmacology and other fields, including material science and nanotechnology.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpcb.7b11707. Optimized parameters for valent bonds, bond constraints, and valent angles of the MARTINI model (Tables S1−S3); details for all performed CG simulations (Table S4); lateral diffusion coefficients (Table S5); Fukui functions for mono-, di-, and trisubstituted Pc molecule (Figure S1); radial distribution function of water particles (Figure S2); equilibrium dynamics of the ZnPcChol8+ penetration (Figure S3); snapshot of the mixed LPS/DPPE (Figure S4); number density profiles of water (Figure S5) and various components of the Pc8+/LPS/DPPE system (Figure S6); angle between the Pc8+ and the LPS/DPPE membrane (Figure S7); average radius of water pore (Figure S8); adsorption of ZnPcChol8+ molecules (Figure S9) (PDF) Coarse-grained MARTINI topology of ZnPcChol8+ (TXT) Coarse-grained structure of ZnPcChol8+ (PDB)

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Tel: +7 968 4633964. Fax: +7 495 4084254. ORCID

Philipp S. Orekhov: 0000-0003-4078-4762 Author Contributions ∇

P.S.O. and E.G.K. contributed equally.

Notes

The authors declare no competing financial interest. The research is carried out using the equipment of the shared research facilities of HPC computing resources at Lomonosov Moscow State University.

■

ACKNOWLEDGMENTS The work was partially supported by the RFBR grant no. 16-3401047 to M.E.B. E.G.K. was supported by the Foundation for Assistance to Small Innovative Enterprises in the framework of the Program “UMNIK” grant no. 9343GU/2015.

■

ABBREVIATIONS: LPS, lipopolysaccharide; Pc, phthalocyanine; ZnPc, zinc phthalocyanine; ZnPcChol 8+ , zinc octakis(cholinyl)phthalocyanine; CG, coarse-grained

■

REFERENCES

(1) WHO. Antibiotic Resistance; WHO, 2017. (2) Maragakis, L. L.; Perencevich, E. N.; Cosgrove, S. E. Clinical and Economic Burden of Antimicrobial Resistance. Expert Rev. Anti.-Infect. Ther. 2008, 6, 751−763. (3) Thabit, A. K.; Crandon, J. L.; Nicolau, D. P. Antimicrobial Resistance: Impact on Clinical and Economic Outcomes and the Need J


Article


(44) de Jong, D. H.; Liguori, N.; van den Berg, T.; Arnarez, C.; Periole, X.; Marrink, S. J. Atomistic and Coarse Grain Topologies for the Cofactors Associated with the Photosystem II Core Complex. J. Phys. Chem. B 2015, 119, 7791−7803. (45) Ramírez, C. L.; Petruk, A.; Bringas, M.; Estrin, D. A.; Roitberg, A. E.; Marti, M. A.; Capece, L. Coarse-Grained Simulations of Heme Proteins: Validation and Study of Large Conformational Transitions. J. Chem. Theory Comput. 2016, 12, 3390−3397. (46) Hills, R. D., Jr.; Lu, L.; Voth, G. A. Multiscale Coarse-Graining of the Protein Energy Landscape. PLoS Comput. Biol. 2010, 6, No. e1000827. (47) Yesylevskyy, S. O.; Schäfer, L. V.; Sengupta, D.; Marrink, S. J. Polarizable Water Model for the Coarse-Grained MARTINI Force Field. PLoS Comput. Biol. 2010, 6, No. e1000810. (48) Kirschner, K. N.; Lins, R. D.; Maass, A.; Soares, T. A. A GlycamBased Force Field for Simulations of Lipopolysaccharide Membranes: Parametrization and Validation. J. Chem. Theory Comput. 2012, 8, 4719−4731. (49) Kubiak, J.; Brewer, J.; Hansen, S.; Bagatolli, L. A. Lipid Lateral Organization on Giant Unilamellar Vesicles Containing Lipopolysaccharides. Biophys. J. 2011, 100, 978−986. (50) Hsu, P.-C.; Jefferies, D.; Khalid, S. Molecular Dynamics Simulations Predict the Pathways via Which Pristine Fullerenes Penetrate Bacterial Membranes. J. Phys. Chem. B 2016, 120, 11170− 11179. (51) Wassenaar, T. A.; Ingólfsson, H. I.; Böckmann, R. A.; Peter Tieleman, D.; Marrink, S. J. Computational Lipidomics Withinsane: A Versatile Tool for Generating Custom Membranes for Molecular Simulations. J. Chem. Theory Comput. 2015, 11, 2144−2155. (52) de Jong, D. H.; Baoukina, S.; Ingólfsson, H. I.; Marrink, S. J. Martini Straight: Boosting Performance Using a Shorter Cutoff and GPUs. Comput. Phys. Commun. 2016, 199, 1−7. (53) Kučerka, N.; Nieh, M.-P.; Katsaras, J. Fluid Phase Lipid Areas and Bilayer Thicknesses of Commonly Used Phosphatidylcholines as a Function of Temperature. Biochim. Biophys. Acta, Biomembr. 2011, 1808, 2761−2771. (54) Petrache, H. I.; Dodd, S. W.; Brown, M. F. Area per Lipid and Acyl Length Distributions in Fluid Phosphatidylcholines Determined by 2H NMR Spectroscopy. Biophys. J. 2000, 79, 3172−3192. (55) Ma, H.; Irudayanathan, F. J.; Jiang, W.; Nangia, S. Simulating Gram-Negative Bacterial Outer Membrane: A Coarse Grain Model. J. Phys. Chem. B 2015, 119, 14668−14682. (56) Van Oosten, B.; Harroun, T. A. A MARTINI Extension for Pseudomonas aeruginosa PAO1 Lipopolysaccharide. J. Mol. Graphics Modell. 2016, 63, 125−133. (57) Hub, J. S.; de Groot, B. L.; van der Spoel, D. g_wham-A Free Weighted Histogram Analysis Implementation Including Robust Error and Autocorrelation Estimates. J. Chem. Theory Comput. 2010, 6, 3713−3720. (58) Makarov, D. A.; Yuzhakova, O. A.; Slivka, L. K.; Kuznetsova, N. A.; Negrimovsky, V. M.; Kaliya, O. L.; Lukyanets, E. A. Cationic Zn and Al Phthalocyanines: Synthesis, Spectroscopy and Photosensitizing Properties. J. Porphyrins Phthalocyanines 2007, 11, 586−595. (59) Marrink, S. J.; Peter Tieleman, D. Perspective on the Martini Model. Chem. Soc. Rev. 2013, 42, 6801. (60) Bereau, T.; Kremer, K. Automated Parametrization of the Coarse-Grained Martini Force Field for Small Organic Molecules. J. Chem. Theory Comput. 2015, 11, 2783−2791. (61) Marrink, S. J.; Risselada, H. J.; Yefimov, S.; Tieleman, D. P.; de Vries, A. H. The MARTINI Force Field: Coarse Grained Model for Biomolecular Simulations. J. Phys. Chem. B 2007, 111, 7812−7824. (62) Menichetti, R.; Kanekal, K. H.; Kremer, K.; Bereau, T. In Silico Screening of Drug-Membrane Thermodynamics Reveals Linear Relations between Bulk Partitioning and the Potential of Mean Force. J. Chem. Phys. 2017, 147, No. 125101. (63) Alarcón, E.; Edwards, A. M.; Garcia, A. M.; Muñoz, M.; Aspée, A.; Borsarelli, C. D.; Lissi, E. A. Photophysics and Photochemistry of Zinc Phthalocyanine/bovine Serum Albumin Adducts. Photochem. Photobiol. Sci. 2009, 8, 255−263.

Pyridinium Zinc Phthalocyanine across the Outer Membrane of Escherichia coli. Antimicrob. Agents Chemother. 2000, 44, 522−527. (25) Soncin, M.; Fabris, C.; Busetti, A.; Dei, D.; Nistri, D.; Roncucci, G.; Jori, G. Approaches to Selectivity in the Zn(ii)-phthalocyaninePhotosensitized Inactivation of Wild-Type and Antibiotic-Resistant Staphylococcus Aureus. Photochem. Photobiol. Sci. 2002, 1, 815−819. (26) Mantareva, V.; Kussovski, V.; Angelov, I.; Borisova, E.; Avramov, L.; Schnurpfeil, G.; Wöhrle, D. Photodynamic Activity of WaterSoluble Phthalocyanine zinc(II) Complexes against Pathogenic Microorganisms. Bioorg. Med. Chem. 2007, 15, 4829−4835. (27) Spesia, M. B.; Caminos, D. A.; Pons, P.; Durantini, E. N. Mechanistic Insight of the Photodynamic Inactivation of Escherichia coli by a Tetracationic zinc(II) Phthalocyanine Derivative. Photodiagn. Photodyn. Ther. 2009, 6, 52−61. (28) Strakhovskaya, M. G.; Antonenko, Y. N.; Pashkovskaya, A. A.; Kotova, E. A.; Kireev, V.; Zhukhovitsky, V. G.; Kuznetsova, N. A.; Yuzhakova, O. A.; Negrimovsky, V. M.; Rubin, A. B. Electrostatic Binding of Substituted Metal Phthalocyanines to Enterobacterial Cells: Its Role in Photodynamic Inactivation. Biochemistry 2009, 74, 1305− 1314. (29) Lam, N. H.; Ha, B.-Y. Surface-Lattice Model Describes Electrostatic Interactions of Ions and Polycations with Bacterial Lipopolysaccharides: Ion Valence and Polycation’s Excluded Area. Langmuir 2014, 30, 13631−13640. (30) Ma, H.; Irudayanathan, F. J.; Jiang, W.; Nangia, S. Simulating Gram-Negative Bacterial Outer Membrane: A Coarse Grain Model. J. Phys. Chem. B 2015, 119, 14668−14682. (31) Van Oosten, B.; Harroun, T. A. A MARTINI Extension for Pseudomonas aeruginosa PAO1 Lipopolysaccharide. J. Mol. Graph. Model. 2016, 63, 125−133. (32) Granovsky, A. A. Firefly (Formerly PC GAMESS) Home Page, 2017. http://classic.chem.msu.su/gran/gamess/index.html (accessed Oct 22, 2017). (33) Stevens, W. J.; Basch, H.; Krauss, M. Compact Effective Potentials and Efficient Shared-exponent Basis Sets for the First- and Second-row Atoms. J. Chem. Phys. 1984, 81, 6026−6033. (34) Dunning, T. H., Jr. Gaussian Basis Sets for Use in Correlated Molecular Calculations. I. The Atoms Boron through Neon and Hydrogen. J. Chem. Phys. 1989, 90, 1007−1023. (35) Burkatzki, M.; Filippi, C.; Dolg, M. Energy-Consistent SmallCore Pseudopotentials for 3d-Transition Metals Adapted to Quantum Monte Carlo Calculations. J. Chem. Phys. 2008, 129, No. 164115. (36) Connolly, M. L. Solvent-Accessible Surfaces of Proteins and Nucleic Acids. Science 1983, 221, 709−713. (37) Cieplak, P.; Cornell, W. D.; Bayly, C.; Kollman, P. A. Application of the Multimolecule and Multiconformational RESP Methodology to Biopolymers: Charge Derivation for DNA, RNA, and Proteins. J. Comput. Chem. 1995, 16, 1357−1377. (38) Malde, A. K.; Zuo, L.; Breeze, M.; Stroet, M.; Poger, D.; Nair, P. C.; Oostenbrink, C.; Mark, A. E. An Automated Force Field Topology Builder (ATB) and Repository: Version 1.0. J. Chem. Theory Comput. 2011, 7, 4026−4037. (39) Schmid, N.; Eichenberger, A. P.; Choutko, A.; Riniker, S.; Winger, M.; Mark, A. E.; van Gunsteren, W. F. Definition and Testing of the GROMOS Force-Field Versions 54A7 and 54B7. Eur. Biophys. J. 2011, 40, 843−856. (40) Fuhrmans, M.; Sanders, B. P.; Marrink, S.-J.; de Vries, A. H. Effects of Bundling on the Properties of the SPC Water Model. Theor. Chem. Acc. 2010, 125, 335−344. (41) Abraham, M. J.; Murtola, T.; Schulz, R.; Páll, S.; Smith, J. C.; Hess, B.; Lindahl, E. GROMACS: High Performance Molecular Simulations through Multi-Level Parallelism from Laptops to Supercomputers. SoftwareX 2015, 1−2, 19−25. (42) Páll, S.; Hess, B. A Flexible Algorithm for Calculating Pair Interactions on SIMD Architectures. Comput. Phys. Commun. 2013, 184, 2641−2650. (43) Essmann, U.; Perera, L.; Berkowitz, M. L.; Darden, T.; Lee, H.; Pedersen, L. G. A Smooth Particle Mesh Ewald Method. J. Chem. Phys. 1995, 103, 8577−8593. K


Article

The Journal of Physical Chemistry B (64) Kliaber; Nackiewicz. Synthesis and Selected Properties of Metallo and Metal-Free 2,3,9,10,16,17,23,24 -octacarboxyphthalocyanines. ARKIVOC 2015, 2015, 269. (65) Kawaguchi, T.; Okamura, S.; Togashi, T.; Harada, W.; Hirahara, M.; Miyake, R.; Haga, M.-A.; Ishida, T.; Kurihara, M.; Kanaizuka, K. Potential Tuning of Nanoarchitectures Based on Phthalocyanine Nanopillars: Construction of Effective Photocurrent Generation Systems. ACS Appl. Mater. Interfaces 2015, 7, 19098−19103. (66) Liu, J.-Y.; Wang, C.; Zhu, C.-H.; Zhang, Z.-H.; Xue, J.-P. Preparation and In Vitro Photodynamic Activity of Glucosylated Zinc(II) Phthalocyanines as Underlying Targeting Photosensitizers. Molecules 2017, 22, 845. (67) Mukherjee, S.; Kar, R. K.; Nanga, R. P. R.; Mroue, K. H.; Ramamoorthy, A.; Bhunia, A. Accelerated Molecular Dynamics Simulation Analysis of MSI-594 in a Lipid Bilayer. Phys. Chem. Chem. Phys. 2017, 19, 19289−19299. (68) Amos, S.-B. T. A.; Vermeer, L. S.; Ferguson, P. M.; Kozlowska, J.; Davy, M.; Bui, T. T.; Drake, A. F.; Lorenz, C. D.; James Mason, A. Antimicrobial Peptide Potency Is Facilitated by Greater Conformational Flexibility When Binding to Gram-Negative Bacterial Inner Membranes. Sci. Rep. 2016, 6, No. 37639. (69) Dias, R. P.; da Hora, G. C. A.; Ramstedt, M.; Soares, T. A. Outer Membrane Remodeling: The Structural Dynamics and Electrostatics of Rough Lipopolysaccharide Chemotypes. J. Chem. Theory Comput. 2014, 10, 2488−2497. (70) George, S.; Hamblin, M. R.; Kishen, A. Uptake Pathways of Anionic and Cationic Photosensitizers into Bacteria. Photochem. Photobiol. Sci. 2009, 8, 788−795. (71) Hancock, R. E. Alterations in Outer Membrane Permeability. Annu. Rev. Microbiol. 1984, 38, 237−264. (72) Nascimento, A., Jr.; Pontes, F. J. S.; Lins, R. D.; Soares, T. A. Hydration, Ionic Valence and Cross-Linking Propensities of Cations Determine the Stability of Lipopolysaccharide (LPS) Membranes. Chem. Commun. 2014, 50, 231−233. (73) Pontes, F. J. S.; Rusu, V. H.; Soares, T. A.; Lins, R. D. The Effect of Temperature, Cations, and Number of Acyl Chains on the Lamellar to Non-Lamellar Transition in Lipid-A Membranes: A Microscopic View. J. Chem. Theory Comput. 2012, 8, 3830−3838. (74) Nikaido, H.; Vaara, M. Molecular Basis of Bacterial Outer Membrane Permeability. Microbiol. Rev. 1985, 49, 1−32. (75) Delcour, A. H. Outer Membrane Permeability and Antibiotic Resistance. Biochim. Biophys. Acta 2009, 1794, 808−816. (76) Hsu, P.-C.; Jefferies, D.; Khalid, S. Molecular Dynamics Simulations Predict the Pathways via Which Pristine Fullerenes Penetrate Bacterial Membranes. J. Phys. Chem. B 2016, 120, 11170− 11179. (77) Kubiak, J.; Brewer, J.; Hansen, S.; Bagatolli, L. A. Lipid Lateral Organization on Giant Unilamellar Vesicles Containing Lipopolysaccharides. Biophys. J. 2011, 100, 978−986. (78) Lam, J. S.; Taylor, V. L.; Islam, S. T.; Hao, Y.; Kocíncová, D. Genetic and Functional Diversity of Pseudomonas aeruginosa Lipopolysaccharide. Front. Microbiol. 2011, 2, 118. (79) Makin, S. A.; Beveridge, T. J. Pseudomonas aeruginosa PAO1 Ceases to Express Serotype-Specific Lipopolysaccharide at 45 Degrees C. J. Bacteriol. 1996, 178, 3350−3352. (80) Maldonado, R. F.; Sá-Correia, I.; Valvano, M. A. Lipopolysaccharide Modification in Gram-Negative Bacteria during Chronic Infection. FEMS Microbiol. Rev. 2016, 40, 480−493. (81) Hansen, S. K.; Rainey, P. B.; Haagensen, J. A. J.; Molin, S. Evolution of Species Interactions in a Biofilm Community. Nature 2007, 445, 533−536. (82) Bozdaganyan, M. E.; Orekhov, P. S.; Shaytan, A. K.; Shaitan, K. V. Comparative Computational Study of Interaction of C60-Fullerene and Tris-Malonyl-C60-Fullerene Isomers with Lipid Bilayer: Relation to Their Antioxidant Effect. PLoS One 2014, 9, No. e102487. (83) Carpenter, T. S.; Kirshner, D. A.; Lau, E. Y.; Wong, S. E.; Nilmeier, J. P.; Lightstone, F. C. A Method to Predict Blood-Brain Barrier Permeability of Drug-like Compounds Using Molecular Dynamics Simulations. Biophys. J. 2014, 107, 630−641.

(84) Tepper, H. L.; Voth, G. A. Mechanisms of Passive Ion Permeation through Lipid Bilayers: Insights from Simulations. J. Phys. Chem. B 2006, 110, 21327−21337. (85) Zhang, L.; Dhillon, P.; Yan, H.; Farmer, S.; Hancock, R. E. Interactions of Bacterial Cationic Peptide Antibiotics with Outer and Cytoplasmic Membranes of Pseudomonas aeruginosa. Antimicrob. Agents Chemother. 2000, 44, 3317−3321. (86) Berglund, N. A.; Piggot, T. J.; Jefferies, D.; Sessions, R. B.; Bond, P. J.; Khalid, S. Interaction of the Antimicrobial Peptide Polymyxin B1 with Both Membranes of E. coli: A Molecular Dynamics Study. PLoS Comput. Biol. 2015, 11, No. e1004180. (87) Falla, T. J.; Nedra Karunaratne, D.; Hancock, R. E. W. Mode of Action of the Antimicrobial Peptide Indolicidin. J. Biol. Chem. 1996, 271, 19298−19303.

L


Molecular Mechanism of Uptake of Cationic Photoantimicrobial

Recommend Documents