Subscriber access provided by UNIV OF CALIFORNIA SAN DIEGO LIBRARIES
Article
Elucidating the Bacterial Membrane Disruption Mechanism of Human #-Defensin 5: a Theoretical Study Sang Won Jung, Juho Lee, and Art E. Cho J. Phys. Chem. B, Just Accepted Manuscript • DOI: 10.1021/acs.jpcb.6b11806 • Publication Date (Web): 09 Jan 2017 Downloaded from http://pubs.acs.org on January 16, 2017
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.
The Journal of Physical Chemistry B 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 28
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 Journal of Physical Chemistry
Elucidating the Bacterial Membrane Disruption Mechanism of Human α-defensin 5: a Theoretical Study Sang Won Jung†#, Juho Lee‡#, and Art E. Cho†* †Department of Bioinformatics, Korea University, 2511 Sejong-ro, Jochiwon-eup, Sejong, 339700, Korea ‡Graduate School of Energy, Environment, Water, and Sustainability, Korea Advanced Institute of Science and Technology, 291 Deahak-ro, Yuseong-gu, Daejeon 305-701, Korea
* Corresponding author email:
[email protected] # Both authors contributed equally to this work.
1
ACS Paragon Plus Environment
The Journal of Physical Chemistry
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 2 of 28
ABSTRACT Human α-defensin 5 (HD5) is a broad-spectrum antibacterial peptide produced by small intestinal Paneth cells. Despite considerable experimental evidence for the correlation between bacterial membrane destruction and the antibacterial activity of HD5, its membrane disruption mechanism remains unclear. Using all-atom molecular dynamics (MD) simulations and molecular mechanics Poisson-Boltzmann surface area (MM-PBSA) analysis, we demonstrate the membrane disruption mechanism of HD5 based on the intrinsic binding of HD5 to Gramnegative (GN) bacterial inner-membrane. It was found that both monomer and dimer forms of HD5 bind to the surface of the GN membrane rather than embedding in the hydrophobic core region of the bilayer. Regardless of the form of HD5, the peptide orientated itself similarly on the membrane surface with an inward-pointing electric dipole moment and an outward-pointing hydrophobic dipole moment. We investigated its possible membrane disruption mechanisms and determined that anionic lipid clustering is a plausible mode of action for HD5. Relative binding free energy analysis revealed that electrostatic interactions play the major role in this mechanism. Our findings shed light on the biophysical phenomena of HD5-GN membrane binding and suggest a possible membrane disruption mechanism for HD5. This analysis of the fundamental binding properties of the monomeric HD5-GN membrane complex provides a useful guide for defensin-derived antibiotic design.
2
ACS Paragon Plus Environment
Page 3 of 28
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 Journal of Physical Chemistry
INTRODUCTION Antimicrobial peptides (AMPs) are a crucial source of innate immunity and represent a first line of defense against infection by a broad spectrum of pathogens including bacteria, fungi, and viruses.1 These peptides are produced by the leucocytes and epithelial cells lining the environmental interfaces, namely the gastrointestinal and urogenital tracts, airways, and skin.2, 3 Of the AMPs, human defensin 5 and 6 (HD5 and HD6) are particularly abundantly expressed in human Paneth cells, which are present in the small intestine.4 HD5 is an α-defensin comprised of 32 amino acids including three disulfide bonds in its oxidized form, with Cys3-Cys31, Cys5Cys20, and Cys10-Cys30 connectivity (Figure 1a). Due to its amphipathic character, which is derived from a cluster of basic and polar residues and a substantial proportion of hydrophobic residues (Figure 1b and 1c),5 HD5 shows antibacterial activity against various pathogens including Gram-negative (GN) and Gram-positive (GP) bacteria. Moreover, HD5 is a potent lectin and can neutralize bacterial exotoxins.6 These activities suggest that the many roles of HD5 include supporting the homeostatic balance between colonizing microbiota and innate immune protection from enteric pathogens.7, 8 The antibacterial mechanisms of defensins vary despite similar tertiary structures.9 It is known that the antibacterial mechanism of HD5 against GN bacteria, particularly E. coli, includes it initially traversing the outer membrane and then damaging the inner-membrane.10 Due to the disruption to the inner-membrane induced by the damage, HD5 can favorably enter the cytoplasm where its antibacterial mechanism continues through effecting as yet unknown intracellular targets. Thus, a description of the antibacterial mechanism of HD5 must include an 3
ACS Paragon Plus Environment
The Journal of Physical Chemistry
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 4 of 28
understanding of its membrane disruption mechanism. Here, we demonstrate the membrane disruption mechanism of HD5 based on the intrinsic binding of HD5 monomers and dimers to GN bacterial membranes using a series of allatom molecular dynamics (MD) simulations and the molecular mechanics Poisson-Boltzmann surface area (MM-PBSA) method following on our previous work.11 Although HD5 usually forms dimers under physiological conditions,12 to broadly apply mechanistic theory to defensinderived antibiotic design, an understanding of the fundamental binding properties of the HD5 monomer is necessary. The binding of both the monomer and dimer forms of HD5 to the GN membrane was demonstrated in terms of its location and orientation. After determining the optimized orientation of HD5 monomers and dimers on the surface of the GN membrane, the effects of HD5 on the membrane related to its membrane disruption mechanism were investigated. Moreover, the binding strength of HD5 to the GN bacterial membrane was quantified by analyzing the residual contributions of the monomer or dimer forms of HD5 to membrane binding. Our findings improve understanding of HD5-GN membrane binding and suggest a possible membrane disruption mechanism of HD5. Furthermore, analysis of the fundamental properties of HD5 monomer-GN membrane binding provides a useful guide for defensin-derived antibiotic design.
METHODS Simulation Setup
4
ACS Paragon Plus Environment
Page 5 of 28
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 Journal of Physical Chemistry
All simulations use a bacterial membrane mimic comprised of a 3:1 ratio of palmitoyloleoylphosphatidylethanol-amine (POPE) and palmitoyloleoylphosphatidylglycerol (POPG). The bacterial membrane model was generated using the input generator from the CHARMM-GUI website (http://www.charmm-gui.org/). The initial coordinates of both monomeric and dimeric HD5 were obtained from the X-ray structure (PDB ID: 1ZMP) (Figure 1).12 In all systems, the protonation state of the peptides was defined as that at pH 7, similar to experimental conditions.13 To obtain proper system samplings, each system, including either the monomer or the dimer, took place in triplicate with different initial HD5 orientations on the membrane surface. Detailed descriptions of how to prepare the models are provided in Figure S1 and S2. In all cases, the initial distance from the center of mass of HD5 to the membrane surface was around 20 Å. All reported simulations were run in the GROMACS 4.6.5 package.14 For the potential function of the system, we used the CHARMM36 force field15-18 without cMAP correction. The TIP3P water model was employed to generate explicit solvation conditions.19 Newton’s equations of motion were integrated using the leapfrog algorithm.20 A 2 fs integration time step was used, with the bonds between hydrogen atoms and any heavy atoms constrained to their equilibrium lengths using the LINCS algorithm.21 Periodic boundary conditions were employed. For both van der Waals (vdW) interactions and electrostatic interactions cutoffs of 1.2 nm were used. Long-range electrostatic interactions were calculated using the particle mesh Ewald method.22 The temperature was maintained at 323K using the Nosé-Hoover thermostat,23, 24 with a coupling time constant of 0.5 ps. The system box was allowed to fluctuate under 1 atm using the semi-isotropic Parrinello-Rahman barostat.25 Since HD5 is known to be sensitive to low salt 5
ACS Paragon Plus Environment
The Journal of Physical Chemistry
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 6 of 28
concentration, only a number of counter ions were added to neutralize the whole system. All systems were thoroughly minimized and then equilibrated for a total 600 ps including NVT and NPT with the Berendsen weak coupling method.26 The z-coordinates of the lipid atoms were restrained during the equilibration steps to restrict their motion to the x-y plane. After equilibration, the simulations were carried out for 300 ns under the NPT ensemble without any position restraints. The simulation trajectories were analyzed using various GROMACS tools. Table 1 provides the number of lipids, the number of counter ions, and the run time of each simulation system. Orientation angle The orientation angle between the surface normal vector and the hydrophobic (θ) or electric (ψ) dipole moments of HD5 was calculated as in our previous study.11 The hydrophobic dipole moment of the protein, Hm, is defined as
= (1)
where Hi indicates the hydrophobicity scale of residue i, which is the normalized “consensus” scale. N is the total number of residues, and si is a unit vector pointing from the alpha-carbon atom of residue i to the center of mass of the residue’s side chain. The protein’s electric dipole moment, µ, is calculated by following equation 2,
μ = (2)
6
ACS Paragon Plus Environment
Page 7 of 28
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 Journal of Physical Chemistry
where qi is the partial charge of atom i and xi indicates the position of atom i subtracted from the molecule’s center of mass. The hydrophobic (θ) and electric (ψ) orientation angles are illustrated for the monomer and dimer cases, respectively, in Figure S1 and S2. Relative binding free energy calculation To calculate the relative binding free energies of the HD5-bacterial membrane complexes, the molecular mechanics Poisson-Boltzmann surface area (MM-PBSA) method was carried out using g_mmpbsa code according to our previous protocol.11, 27 50 snapshots were extracted from the last 20 ns trajectory of each MD simulation. The binding free energy of each complex was computed as the difference: ∆ = − + !" # (3) where Gcomplex indicates the free energy of the HD5-membrane complex and Gpeptide and Gmembrane are the free energies of isolated HD5 and isolated membrane in solvent, respectively. When calculating the free energy, G, the entropy contribution of the protein was ignored since the binding energy is used here to determine the relative binding strength of each complex. = 〈&'' 〉 + 〈) 〉 (4) where 〈&'' 〉 is the average MM potential energy when using a CHARMM 36 force field in a vacuum, and 〈) 〉 is the average solvation free energy. For EMM, due to the single trajectory approach, only non-bonded interactions, including electrostatic and van der Waals interactions, were taken into account.
7
ACS Paragon Plus Environment
The Journal of Physical Chemistry
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 8 of 28
The solvation free energy, Gsol, is composed of two terms. ) = ) + ) (5) Gps is the polar solvation contribution calculated by solving the non-linear Poisson-Boltzmann (PB) equation. The values for the solute (pdie) and solvent (sdie) dielectric constants were chosen to be 7 and 80, respectively. The ionic strength was set to 0.05 M, which is the usual salt concentration used for experiments.5, 13 The non-polar solvation free energy, Gnps, was estimated by the solvent accessible surface area (SASA) using a water probe radius of 1.4 Å: ) = ,-.-. + / (6) where the constants γ and b were set to 0.00542 kcal/mol/Å2 and 0.92 kcal/mol, respectively.28
RESULTS AND DISCUSSION Adsorption orientation of HD5 on GN bacterial inner-membranes To demonstrate the membrane disruption mechanism of HD5 through molecular modeling, we first investigated how the monomeric and dimeric forms of HD5 adsorb on the inner-membrane of GN bacteria depending on various initial configurations (Figure S1 and S2). For the binding of the HD5-membrane complex, the distance between the center of mass (COM) of each form of HD5 molecule and the COM of the membranes was measured as function of time (Figure S3). Regardless of initial configurations and forms, HD5 was found to approach the bacterial membrane rapidly within the first 10 ns, after which the distance converges to about 2.5 nm by 8
ACS Paragon Plus Environment
Page 9 of 28
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 Journal of Physical Chemistry
the end of simulation, meaning that HD5 binds stably to the membrane surface. To further characterize the adsorption behavior, the density profiles along the bilayer normal (Z-axis) were analyzed. Figure 2 shows the density profiles of various components including protein and lipid subgroups (phosphate, hydrophilic, and hydrophobic). Both forms of HD5 adsorbed to a depth of around 2.2 ~ 2.4 nm which is a region close to the phosphate and hydrophilic groups of the lipid and also in partial contact with its hydrophobic group. The density distribution of the HD5 dimer (from 1.2 nm to 4 nm) was broader than that of the monomer (from 1.2 nm to 3.6 nm). However, in neither case were the peptides located at 1.2 nm where the hydrophobic core region starts. This finding indicates that both forms of HD5 bind to the GN bacterial membrane via the surface region of the bilayer rather than embedding in its hydrophobic core. We next demonstrate the orientation of HD5 on the membrane surface which is of great importance to understanding their binding interaction, and therefore directly related to the membrane disruption mechanism. Following the approach of our previous work that clarifies the adsorption orientation of human β-defensin 3 (hBD3), we describe the orientation of HD5 that gives the most energetically stable binding complex in Table 2 (model C for monomer and model E for dimer) by measuring the angle of its electric and hydrophobic dipole moment with respect to the membrane normal (Figure 3). In this description, an angle of 180° indicates that the dipole is antiparallel to the surface normal, which means pointing toward the membrane, an angle of 90° indicates that it is parallel to the surface itself, and an angle of 0° indicates that it is parallel to the surface normal, which means pointing away from the membrane. On the surface of the membrane, the electric and hydrophobic moments of the HD5 monomer were found to be in opposite directions to each other with angles of about 155° and 55°, respectively, regardless of 9
ACS Paragon Plus Environment
The Journal of Physical Chemistry
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 10 of 28
initial orientations (Figure 3a). In short, the electric dipole moment of the HD5 monomer pointed toward the membrane, while its hydrophobic dipole moment pointed away from the membrane (Figure 3b). Although the dipole moment distribution of HD5 dimer showed more clearly separated peaks at 165° for the electric and 20° for the hydrophobic dipole moment (Figure 3c and 3d) compared with that of HD5 monomer, overall patterns were similar suggesting that the electrostatic interaction between HD5 and the membrane dominates the binding of the HD5membrane complex. The active region of HD5 is populated with basic and polar residues including Arg25, Leu26, Tyr27, and Arg28. This region, in agreement with a previous study,13 was found to contact the membrane surface directly in both the monomer and the dimer forms. Membrane disruption mechanism of HD5 We analyze how HD5 disrupts the inner membranes of GN bacteria, based on the optimized orientations of HD5 on the membrane surface found by the simulations. It has already been suggested that the antibacterial activity of HD5 is mediated via its binding to specific target proteins rather than directly killing bacteria via forming pores in the membrane. However, the process of inner-membrane disruption, which allows HD5 to permeate into the cytoplasm of GN bacteria, is a significant part of its antibacterial activity. Zhao et. al. demonstrated via a molecular modeling approach that hBD3 disrupts the bacterial membrane through the clustering of anionic membrane lipids.29 Considering the shared structural properties of hBD3 and HD5, anionic lipid clustering is a promising candidate HD5 membrane disruption mechanism. Thus, we shed some light on the disruption mechanism in terms of changes in the distribution of the anionic lipid, POPG, near HD5. To be specific, the distribution of the oxygen of the hydroxyl 10
ACS Paragon Plus Environment
Page 11 of 28
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 Journal of Physical Chemistry
group of POPG, OPOPG, was used. Using a radial distribution function, the number of POPG lipids within a cut-off distance from the center of mass (COM) of HD5 was counted as function of time (Figure 4a). In order to determine the cut-off value, the radius of gyration, Rg, which contains information about the size and shape of a protein, was employed. The average Rg after equilibration was 0.84 nm and 1.13 nm for the monomer and dimer, respectively (Figure S4). Following a calculation based on the scheme shown in Figure 4b, the distance cut-off values for counting the POPG lipids within the projected area of HD5 were chosen to be 1.2 nm for the monomer and 1.6 nm for the dimer. As Figure 4a shows, the number of OPOPG clearly increased right after either form of HD5 bound to the membrane surface (at around 12 ns in the figure) and continued to increase until 30 ns in the dimer system. In the case of the dimer-membrane complex system the number continued to rise gradually until 150 ns by which time it had equilibrated. The trends of each peptide chain comprising the dimer form were similar and their sum gives a similar OPOPG distribution to that obtained with the whole dimer unit. In the case of the HD5 monomer, the number of OPOPG was increased by about 2 oxygens, which accords with the result for chain A of the HD5 dimer. In order to confirm the clustering effect, the density of POPG in the projected area and the surface density of POPG in the whole system were compared. The density of OPOPG in the projected area was about 1.33 num/nm2, which was denser than its surface density in the whole system, 0.83 num/nm2, suggesting that POPG lipids locally cluster at the HD5-GN membrane interface. Our findings suggest that the membrane disruption mechanism induced by HD5 clearly includes the clustering of anionic lipids and seems to rule out some other mechanisms. Of course, defensins have other known membrane disruption mechanisms such as 11
ACS Paragon Plus Environment
The Journal of Physical Chemistry
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 28
membrane thinning/thickening, and the “carpet-model”.9 However, there was no evidence found of a significant effect on membrane thickness. In fact, the distance between the phosphate groups of the upper and lower leaflets (the green peaks in Figure 2) obtained in the systems with either HD5 monomer or dimer was similar to the membrane thickness without HD5 (Figure 2). Previous studies suggested that the thinning effect might occur in the presence of a highly dense AMP cluster, and this requirement may explain why the membrane thinning effect was not clearly observed in our simulations.29 In the “carpet-model”, AMPs act in a detergent-like manner by binding to the bacterial membrane surface where the orientation of their hydrophobic residues contacts the hydrophobic core region to form toroidal aggregates.9, 29 However, our data shows that the active region, with its highly dense hydrophilic regions, contacts the GN membrane surface rather than the hydrophobic core and that the hydrophobic dipole moment even points away from the membrane. Thus, the “carpet-model” also seems inapplicable to HD5. Adsorption energy of HD5-GN membrane complex The binding free energies of the HD5-membrane complex were calculated to provide further insight into the interactions between HD5 and the inner-membranes of GN bacteria. The MMPBSA method, which has been successfully used in a previous study to elucidate the interaction between AMP and bacterial membrane11 was also adopted in this study. Since the purpose of the free energy calculation is to determine the relative strength of the binding energy of HD5 based on its orientation and oligomeric states, the entropy contribution of the peptide was ignored. The results of the relative binding free energy analysis are provided in Table 2. The polar free energy difference (∆Gpolar) is defined as the sum of the electrostatic interaction energy and the difference 12
ACS Paragon Plus Environment
Page 13 of 28
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 Journal of Physical Chemistry
in polar solvation energy (∆Gps). In the same manner, the nonpolar free energy (∆Gnonpolar) is the sum of the vdW interaction energy and the difference in nonpolar solvation energy (∆Gnps). The overall binding free energy is the sum of ∆Gpolar and ∆Gnonpolar. In the case of the monomeric HD5 systems, the strongest binding energy of -507.9 kJ/mol was obtained with model C. Regardless of the initial configuration of the models, the relative contributions of polar free energy and nonpolar free energy to HD5-membrane binding remained about 48% and 52%, respectively. Moreover, in all models the electrostatic interaction dominated at about -914.4 kJ/mol. This finding corresponds to the results of the HD5 binding orientation analysis, which showed that the electric dipole moment was a key factor in HD5-GN membrane binding. The contribution of the electrostatic interaction to binding free energy was similarly dominant for dimer-membrane binding, and much larger than double the electrostatic interaction energy of the HD5 monomer-GN membrane complex. This is because, unlike vdW interactions, even if the number of POPG molecules under each chain of the dimer is similar to the monomer as shown in Figure 4, the projected electrostatic influence ranges cover a larger area than for the monomer. Therefore, some POPG lipids have electrostatic interactions with both chains of the HD5 dimer, leading to an overall electrostatic interaction energy for the dimer which is much larger than double that of the monomer. Although there are no direct experimental values for comparison, a comparison with our previously calculated -292.2 kJ/mol binding energy of hBD3 to GN membrane,11 suggests that HD5 would bind the bacterial membrane favorably due strong electrostatic interactions. The free energy decomposition value of each HD5 residue was determined (Figure 5). 13
ACS Paragon Plus Environment
The Journal of Physical Chemistry
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 14 of 28
Following previous studies,11 we define important residues as those whose energy contribution value is higher than 10.46 kJ/mol. In both the monomer and dimer systems, the cationic residual contributions of particularly arginine dominated. Of the residues interacting with the membrane, mutagenesis studies have shown that mutations of Arg6, Arg25, and Arg28 lead to reduced antibacterial activity.30, 31 Moreover, characterizing the importance of the electropositive charges in HD5 for membrane destruction supports our simulation results.13 Although Leu26 and Tyr27 are known to be key residues in the antibacterial activity of HD5, they both showed only weak binding interactions with the membrane, suggesting that both residues may be important for binding intracellular targets in the cytoplasm rather than membrane disruption.
CONCLUSIONS In summary, all-atom MD simulations to demonstrate the adsorption and disruption mechanisms of HD5 on the inner membrane of GN bacteria and a detailed binding free energy analysis were carried out. First, the adsorption mechanisms of both monomeric and dimeric forms of HD5 on the membrane from various starting configurations were investigated. The results showed that HD5 stably bound to the membrane surface rather than embedding in the hydrophobic core region of the bilayer regardless of form or starting configuration. The distribution of electric and hydrophobic dipole moments obtained through orientation angle analyses revealed that the HD5 monomer had an inward-pointing electric dipole moment and an outward-pointing hydrophobic dipole moment. Radial distribution function analysis demonstrated how HD5 disrupted the membrane and clearly showed that POPG lipids, which are anionic, were clustered near HD5 14
ACS Paragon Plus Environment
Page 15 of 28
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 Journal of Physical Chemistry
throughout the simulations. Although there are other known membrane disruption mechanisms, anionic lipid clustering was determined to be a plausible mode of action for HD5. The binding free energies of the HD5-membrane complex were also analyzed, including the contribution of each residue. The results showed that, for both the monomer and dimer forms of HD5, electrostatic interactions rather than hydrophobic interactions dominated the binding. The binding energy obtained for the HD5 dimer was much more than double that of the HD5 monomer due to POPG lipids being subject to electrostatic interactions with both chains of the HD5 dimer. This suggests that the dimeric form is more favorable for membrane binding. The free energy decomposition analysis identified that most of the positively charged residues, particularly Arg, are key residues for HD5-membrane binding. Our findings improve understanding of the biophysical phenomena of HD5-GN bacterial inner membrane binding and suggest a possible membrane disruption mechanism of HD5. In addition, analysis of the fundamental binding properties of the HD5 monomer-GN membrane complex provides a useful guide for further defensin-derived antibiotic design. SUPPORTING INFORMATION Description of orientation angles of the HD5 monomer and dimer with respect to the surface 23) of the membrane (Figures S1-S2), distance of the HD5 monomer and dimer from the normal (1 membrane bilayer normal (z-axis) as a function of simulation time (Figure S3), the radius of gyration of an HD5 monomer and dimer as a function of simulation time (Figure S4). This information is available free of charge via the Internet at http://pubs.acs.org. ACKNOWLEDGEMENTS 15
ACS Paragon Plus Environment
The Journal of Physical Chemistry
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 16 of 28
This work was supported by NRF Grants No. 2013R1A2A2A01067638 and 2016M3A7B 4025405.
16
ACS Paragon Plus Environment
Page 17 of 28
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 Journal of Physical Chemistry
REFERENCES 1.
Nizet, V.; Ohtake, T.; Lauth, X.; Trowbridge, J.; Rudisill, J.; Dorschner, R. A.;
Pestonjamasp, V.; Piraino, J.; Huttner, K.; Gallo, R. L. Innate antimicrobial peptide protects the skin from invasive bacterial infection. Nature 2001, 414, 454-457. 2.
Selsted, M. E.; Ouellette, A. J. Mammalian defensins in the antimicrobial immune
response. Nat. Immunol. 2005, 6, 551-557. 3.
Salzman, N. H.; Hung, K.; Haribhai, D.; Chu, H.; Karlsson-Sjöberg, J.; Amir, E.;
Teggatz, P.; Barman, M.; Hayward, M.; Eastwood, D. Enteric defensins are essential regulators of intestinal microbial ecology. Nat. immunol. 2010, 11, 76-82. 4.
Jones, D. E.; Bevins, C. L. Paneth cells of the human small intestine express an
antimicrobial peptide gene. J. Biol. Chem. 1992, 267, 23216-23225. 5.
Wommack, A. J.; Robson, S. A.; Wanniarachchi, Y. A.; Wan, A.; Turner, C. J.; Wagner,
G.; Nolan, E. M. NMR solution structure and condition-dependent oligomerization of the antimicrobial peptide human defensin 5. Biochemistry 2012, 51, 9624-9637. 6.
Lehrer, R. I.; Jung, G.; Ruchala, P.; Andre, S.; Gabius, H. J.; Lu, W. Multivalent binding
of carbohydrates by the human alpha-defensin, HD5. J. Immunol. 2009, 183, 480-490. 7.
Bevins, C. L.; Salzman, N. H. Paneth cells, antimicrobial peptides and maintenance of
intestinal homeostasis. Nat. Rev. Microbiol. 2011, 9, 356-368. 8.
Clevers, H. C.; Bevins, C. L. Paneth cells: maestros of the small intestinal crypts. Annu.
Rev. Physiol. 2013, 75, 289-311. 9.
Nguyen, L. T.; Haney, E. F.; Vogel, H. J. The expanding scope of antimicrobial peptide
structures and their modes of action. Trends Biotechnol. 2011, 29, 464-472. 17
ACS Paragon Plus Environment
The Journal of Physical Chemistry
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
10.
Page 18 of 28
Chileveru, H. R.; Lim, S. A.; Chairatana, P.; Wommack, A. J.; Chiang, I. L.; Nolan, E. M.
Visualizing attack of Escherichia coli by the antimicrobial peptide human defensin 5. Biochemistry 2015, 54, 1767-1777. 11.
Lee, J.; Jung, S. W.; Cho, A. E. Molecular insights into the adsorption mechanism of
human β-defensin-3 on bacterial membranes. Langmuir 2016, 32, 1782-1790. 12.
Szyk, A.; Wu, Z.; Tucker, K.; Yang, D.; Lu, W.; Lubkowski, J. Crystal structures of
human alpha-defensins HNP4, HD5, and HD6. Protein Sci. 2006, 15, 2749-2760. 13.
Wang, C.; Shen, M.; Gohain, N.; Tolbert, W. D.; Chen, F.; Zhang, N.; Yang, K.; Wang,
A.; Su, Y.; Cheng, T.; Zhao, J.; Pazgier, M.; Wang, J. Design of a potent antibiotic peptide based on the active region of human defensin 5. J. Med. Chem. 2015, 58, 3083-3093. 14.
Pronk, S.; Páll, S.; Schulz, R.; Larsson, P.; Bjelkmar, P.; Apostolov, R.; Shirts, M. R.;
Smith, J. C.; Kasson, P. M.; van der Spoel, D. GROMACS 4.5: A high-throughput and highly parallel open source molecular simulation toolkit. Bioinformatics 2013, 29, 845-854. 15.
Vanommeslaeghe, K.; Hatcher, E.; Acharya, C.; Kundu, S.; Zhong, S.; Shim, J.; Darian,
E.; Guvench, O.; Lopes, P.; Vorobyov, I. CHARMM general force field: A force field for drug‐like molecules compatible with the CHARMM all‐atom additive biological force fields. J. Comput. Chem. 2010, 31, 671-690. 16.
Best, R. B.; Zhu, X.; Shim, J.; Lopes, P. E.; Mittal, J.; Feig, M.; MacKerell Jr, A. D.
Optimization of the additive CHARMM all-atom protein force field targeting improved sampling of the backbone ϕ, ψ and side-chain χ1 and χ2 dihedral angles. J. Chem. Theory Comput. 2012, 8, 3257-3273. 17.
MacKerell Jr, A. D.; Bashford, D.; Bellott, M.; Dunbrack Jr, R. L.; Evanseck, J. D.; Field, 18
ACS Paragon Plus Environment
Page 19 of 28
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 Journal of Physical Chemistry
M. J.; Fischer, S.; Gao, J.; Guo, H.; Ha, S. All-atom empirical potential for molecular modeling and dynamics studies of proteins. J. Phys. Chem. B 1998, 102, 3586-3616. 18.
Klauda, J. B.; Venable, R. M.; Freites, J. A.; O’Connor, J. W.; Tobias, D. J.; Mondragon-
Ramirez, C.; Vorobyov, I.; MacKerell Jr, A. D.; Pastor, R. W. Update of the CHARMM all-atom additive force field for lipids: validation on six lipid types. J. Phys. Chem. B 2010, 114, 78307843. 19.
Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey, R. W.; Klein, M. L.
Comparison of simple potential functions for simulating liquid water. J. Chem. Phys. 1983, 79, 926-935. 20.
Cuendet, M. A.; van Gunsteren, W. F. On the calculation of velocity-dependent
properties in molecular dynamics simulations using the leapfrog integration algorithm. J. Chem. Phys. 2007, 127, 184102. 21.
Hess, B. P-LINCS: A parallel linear constraint solver for molecular simulation. J. Chem.
Theory Comput. 2008, 4, 116-122. 22.
Darden, T.; York, D.; Pedersen, L. Particle mesh Ewald: An N⋅ log (N) method for
Ewald sums in large systems. J. Chem. Phys. 1993, 98, 10089-10092. 23.
Nosé, S. A unified formulation of the constant temperature molecular dynamics methods.
J. Chem. Phys. 1984, 81, 511-519. 24.
Hoover, W. G. Canonical dynamics: equilibrium phase-space distributions. Phys. Rev. A
1985, 31, 1695-1697. 25.
Parrinello, M.; Rahman, A. Polymorphic transitions in single crystals: A new molecular
dynamics method. J. Appl. Phys. 1981, 52, 7182-7190. 19
ACS Paragon Plus Environment
The Journal of Physical Chemistry
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
26.
Page 20 of 28
Berendsen, H. J.; Postma, J. v.; van Gunsteren, W. F.; DiNola, A.; Haak, J. Molecular
dynamics with coupling to an external bath. J. Chem. Phys. 1984, 81, 3684-3690. 27.
Kumari, R.; Kumar, R.; Lynn, A. g_mmpbsa-- A GROMACS tool for high-throughput
MM-PBSA calculations. J. Chem. Inf. Model. 2014, 54, 1951-1962. 28.
Sitkoff, D.; Sharp, K. A.; Honig, B. Accurate calculation of hydration free energies using
macroscopic solvent models. J. Phys. Chem. 1994, 98, 1978-1988. 29.
Zhao, X.; Yu, H.; Yang, L.; Li, Q.; Huang, X. Simulating the antimicrobial mechanism of
human beta-defensin-3 with coarse-grained molecular dynamics. J. Biomol. Struct. Dyn. 2015, 33, 2522-2529. 30.
Chapnik, N.; Levit, A.; Niv, M. Y.; Froy, O. Expression and structure/function
relationships of human defensin 5. Appl. Biochem. Biotechnol. 2012, 166, 1703-1710. 31.
de Leeuw, E.; Rajabi, M.; Zou, G.; Pazgier, M.; Lu, W. Selective arginines are important
for the antibacterial activity and host cell interaction of human alpha-defensin 5. FEBS Lett. 2009, 583, 2507-2512.
20
ACS Paragon Plus Environment
Page 21 of 28
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 Journal of Physical Chemistry
Table 1. Details of all the simulations
HD5 units
Model
Num. of lipidsa
Num. of Na+
Run Time
none
-
120
30
300 ns
A
120
26
300 ns
B
120
26
300 ns
C
120
26
300 ns
D
240
52
300 ns
E
240
52
300 ns
Monomer
Dimer a
The lipid composition of all systems was 3 PE : 1 PG.
21
ACS Paragon Plus Environment
The Journal of Physical Chemistry
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 22 of 28
Table 2. Relative binding free energy components of the HD5 monomer and dimer -membrane complexes Monomer
Dimer
Model A
Model B
Model C
average
Model D
Model E
average
vdW
-192.5
-202.6
-218.0
-204.4
-343.3
-445.1
-394.2
Electrostatic
-953.0
-702.0
-1088.1
-914.4
-2643.9
-2751.2
-2697.6
∆Gps
757.6
499.0
831.1
695.9
1616.6
1663.2
1639.9
-30.1
-29.6
-32.9
-30.9
-56.5
-68.0
-62.3
-195.4
-203.0
-257.0
-218.5
-1027.0
-1088.0
-1057.5
∆Gnonpolarb
-222.6
-232.0
-251.0
-235.2
-400.0
-513.0
-456.5
∆Gbinding
-417.9
-435.2
-507.9
-453.7
-1427.1
-1601.1
-1514.1
∆Gnps ∆Gpolar
a
a
All energy units are in kJ/mol. Indicating polar free energy difference, ∆Gpolar = Electrostatic + ∆Gps. Indicating nonpolar free energy difference, b∆Gnonpolar = vdW + ∆Gnps.
22
ACS Paragon Plus Environment
Page 23 of 28
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 Journal of Physical Chemistry
Figure 1. Primary sequence (a) and atomic structure of the monomer (b) and dimer (c) forms of HD5. In (a) Cys, Arg, and Glu are colored yellow, blue, and red, respectively. In (b) the positively and negatively charged residues are highlighted with blue and red, respectively. The Val19 and Glu21 residues involved in dimer formation are highlighted with white and red, respectively. The active region suggested by experimental studies is demarcated by red dashed lines.
23
ACS Paragon Plus Environment
The Journal of Physical Chemistry
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 24 of 28
Figure 2. Density profiles of each subgroup in the HD5 monomer-membrane system (a) and the HD5 dimer-membrane system (b) as a function of the position normal to the bilayer (the z-axis). The subgroups are the HD5 protein (blue), the hydrophilic region of the bilayer (red), the phosphate of the bilayer (green), and hydrophobic region of the bilayer (purple). The dotted lines indicate the center of the phosphate group in a normal GN bacterial membrane system. The profiles were averaged over the last 20 ns of the simulations. 24
ACS Paragon Plus Environment
Page 25 of 28
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 Journal of Physical Chemistry
Figure 3. Orientation distributions of the electric dipole (blue) and hydrophobic dipole (red) of the monomer (a) and dimer (b) forms of HD5 adsorbed on the membrane. The final simulation 23 and snapshots of the monomer (c) and dimer (d) forms of HD5 are presented. The blue arrow 6
23 indicate the directions of the electric and hydrophobic dipole moments, respectively. red arrow 7 The active region of HD5 is demarcated by the dotted lines.
25
ACS Paragon Plus Environment
The Journal of Physical Chemistry
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 26 of 28
Figure 4. Distribution of OPOPG from the center of HD5 (a). (b) is a schematic view describing how to define the cut-off covering the projected area according to the radius of gyration. Final snapshots of HD5 monomer (c) and dimer (d) – membrane systems. HD5 tends to create an anionic cluster of POPG (red) rather than POPE (gray). HD5 is colored in orange and cyan. The yellow semi-transparent region shows the anionic cluster near HD5. The radial distribution of POPG from the COM of HD5 is up to 1.7 nm for the monomer and 2 nm for the dimer. The black dotted line indicates the moment when the HD5 just bound to the membrane surface.
26
ACS Paragon Plus Environment
Page 27 of 28
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 Journal of Physical Chemistry
Figure 5. Binding free energy contribution of each residue of HD5 in the HD5 monomer (a) and dimer (b) – membrane complex.
27
ACS Paragon Plus Environment
The Journal of Physical Chemistry
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 28 of 28
TOC graphic
28
ACS Paragon Plus Environment