a computer simulation study

School of Chemistry and Chemical Engineering, Guangdong Provincial Key Lab for Green. Chemical Product Technology, South China University of Technolog...
7 downloads 5 Views 4MB Size
Download PDF
Article Cite This: Langmuir 2017, 33, 14480−14489

pubs.acs.org/Langmuir

Understanding the Cellular Uptake of pH-Responsive Zwitterionic Gold Nanoparticles: A Computer Simulation Study Xuebo Quan, Daohui Zhao, Libo Li, and Jian Zhou* School of Chemistry and Chemical Engineering, Guangdong Provincial Key Lab for Green Chemical Product Technology, South China University of Technology, Guangzhou 510640, P. R. China ABSTRACT: Surface functionalization of nanoparticles (NPs) with stealth polymers (e.g., hydrophilic and zwitterionic polymers) has become a common strategy to resist nonspecific protein adsorption recently. Understanding the role of surface decoration on NP−biomembrane interactions is of great significance to promote the application of NPs in biomedical fields. Herein, using coarse-grained molecular dynamics (CGMD) simulations, we investigate the interactions between stealth polymer-coated gold nanoparticles (AuNPs) and lipid membranes. The results show that AuNPs grafted with zwitterionic polymers can more easily approach the membrane surface than those coated with hydrophilic poly(ethylene glycol) (PEG), which can be explained by the weak dipole−dipole interaction between them. For zwitterionic AuNPs which can undergo pH-dependent charge conversion, different interaction modes which depend on the polymer protonation degree are found. When the protonation degree is low, the particles just adsorb on the membrane surface; at moderate protonation degrees, the particles can directly translocate across the lipid membrane through a transient hydrophilic pore formed on the membrane surface; the particles are fully wrapped by the curved lipid membrane at high protonation degrees, which may lead to endocytosis. Finally, the effect of polymer chain length on the cellular uptake of zwitterionic polymer-coated AuNPs is considered. The results demonstrate that longer polymer chain length will block the translocation of AuNPs across the lipid membrane when the protonation degree is not high; however, it can improve the transmembrane efficiency of AuNPs at high protonation degrees. We expect that these findings are of immediate interest to the design and synthesis of pH-responsive nanomaterials based on zwitterionic polymers and can prompt their further applications in the field of biomedicine.

1. INTRODUCTION Nowadays, nanoparticles (NPs) have been widely applied for numerous biomedical applications, such as bioimaging, biosensing, drug delivery, diagnostics, and therapy.1−4 However, there exits one major obstacle to their in vivo applications, that is, nonspecific protein adsorption. It is well-known that there are plenty of plasma proteins in the blood, which can adsorb on the NP surfaces through various interactions (e.g., hydrophobic and electrostatic interactions), thus leading to the formation of a protein corona.5−7 This may initialize the body’s immune response and result in the clearance of NPs out of the blood, which will limit the application of NPs in biological and medicine fields. To address this issue, a common strategy is to decorate NPs with hydrophilic and uncharged polymers such as poly(ethylene glycol) (PEG), a process known as “PEGylation”. It is believed that the good antifouling property of PEG is derived from the combination of a steric repulsion and a hydration layer via hydrogen bonding around molecular chains. Pelaz et al.8 investigated the effect of PEG modification on the adsorption of proteins on NPs using fluorescence correlation spectroscopy. For NPs without a PEG shell, a thickness increase of ∼3 nm was found, which corresponds to a human serum albumin adsorption monolayer; however, this thickness increase dropped 50% for PEGylated NPs. Walkey et al.9 demonstrated that the adsorption of serum proteins on gold nanoparticles (AuNPs) can be © 2017 American Chemical Society

controlled by PEG grafting density; a higher PEG density will lead to the reduction of total serum protein adsorption. Zwitterionic polymers, which contain both positively and negatively charged groups but maintain overall charge neutrality, are demonstrated to be more excellent over conventional hydrophilic polymers to prevent nonspecific protein adsorption through the hydration layer formed by ionic solvation and thus have attracted much attention in recent years.10−15 Tatumi et al.16 demonstrated that AuNPs modified with a zwitterionic ligand consisting of an imidazolium cation and a sulfonate anion show great stability in aqueous solutions with high protein concentration. Yang et al.17 modified AuNPs with zwitterionic poly(carboxybetaine acrylamide) (polyCBAA); they found that these functionalized AuNPs are very stable in undiluted human blood serum. Besides, they also demonstrated that polyCBAAcoated AuNPs exhibit superior antifouling performance over AuNPs with PEG coatings. Although these stealth polymer coatings can greatly resist the adsorption of serum proteins on NPs, which will prolong the circulation time of NPs in vivo, it will also affect the interactions between NPs and cell membranes and may even bring some side effects to the cellular uptake of NPs. For example, Pelaz et al.8 Received: October 11, 2017 Revised: November 14, 2017 Published: November 22, 2017 14480

DOI: 10.1021/acs.langmuir.7b03544 Langmuir 2017, 33, 14480−14489

Article

Langmuir

In this work, using the CGMD simulation method based on the MARTINI force field, we investigated the interactions between pH-responsive zwitterionic polymer-grafted AuNPs (Zwitt-AuNPs) and lipid membranes. As a comparison, the AuNPs functionalized with PEG (PEG-AuNPs) were also considered. The translocation mechanism of Zwitt-AuNPs across lipid membranes at different pH environments was explored. Besides, we also discussed the effect of polymer chain length on the cellular uptake of Zwitt-AuNPs.

demonstrated that PEGylated NPs display a pronounced reduction of cellular uptake with respect to bare NPs. Dai et al.18 found that the PEG coating can interfere with the binding of targeting ligands on the surface of NPs to its corresponding cellular receptor in some cases. Ding et al.19 reported that the decoration of hydrophilic and zwitterionic polymers onto the NP surface may not only block the insertion of hydrophobic NPs but also weaken the adsorption of positively charged NPs onto the membrane and thus will decrease the translocation efficiency of NPs through cell membranes. In recent years, to enhance the selective accumulation of NPs at tumor tissues, some types of pH-responsive zwitterionic polymers have been designed, which possess high biocompatibility and pH sensitivity simultaneously.20−22 These polymers are nearly electroneutral at normal tissue pH and become positively charged at tumor tissue pH, which can dramatically enhance the electrostatic interaction between polymer-coated NPs and negatively charged membranes, thus the cellular uptake efficiency of NPs could be greatly improved.21 Meanwhile, the translocation of charged NPs may disrupt the membrane structure and thus induce cell death. Therefore, it is urgent and of great significance to deeply explore the interactions between the pH-responsive zwitterionic polymer-coated NPs and cell membranes to promote their reasonable usage in biomedical applications. As indicated by some recent reviews, keeping the balance between the delivery efficiency and toxicity of NPs should be carefully considered.23,24 Despite the fact that great experimental progresses have been made in the past few years, nevertheless, limited by the available experimental technologies, the interactions between polymerfunctionalized NPs and cellular membranes are far from being completely understood. Computer simulation, as a powerful complementary approach to experiments, can be used to systematically probe the molecular mechanism of these interactions. However, due to the complexity of the systems and the limitation of computing resources, it is prohibitive to study this kind of research system by the all-atom molecular dynamics (AAMD) simulation method, even though more accurate information can be obtained. Hence, coarse-grained molecular dynamics (CGMD) simulations could be more appropriate for this complex system. This method generally groups several atoms or units into one interaction site to reduce the number of degrees of freedom, which has usually been adopted to speed up the simulation process to obtain both molecular-level structural details and dynamics on larger spatial scales and longer time scales. Recently, some simulation works about the effects of polymer modification on the interactions between NPs and biomembranes have been reported.19,25−28 For instance, Li et al.27 investigated the interactions between PEGlated NPs with the lipid membrane via CGMD simulations. They found that the PEG chain length and the NP core size play a crucial role in the shielding effect of stealth PEGylated nanocarriers. Liang29 studied the penetration of polymer-grafted NPs with different chain lengths and grafting densities through a lipid bilayer using self-consistent field theory. Through dissipative particle dynamics (DPD) simulations, Ding and coworkers19 investigated the cellular uptake of NPs functionalized with stealth polymers, and they found that this surface decoration can reduce the translocation efficiency of NPs. Besides, they also studied the cellular uptake of NPs under different pH environments by introducing a pH-sensitive polymer,30 and the results showed that the NPs can be taken up by cell membrane at lower and higher pH, while the cellular uptake will be blocked for pH in the middle range.

2. COMPUTATION METHODS 2.1. Force Fields. The standard MARTINI CG force field developed by Marrink et al. was used in this work.31 In the framework of this CG model, small groups of atoms are united as a single interaction site to reduce the number of degrees of freedom, which can yield both molecular-level structural details and dynamics on larger spatial scales and longer time scales. In general, four heavy atoms are mapped into one CG particle, except for the case of ring structures (such as benzene, cholesterol, etc.), where a two or three to one mapping scheme is adopted. Based on the differences of the chemical nature of the underlying structure, four main types of interaction sites are considered: polar (P), nonpolar (N), apolar (C), and charged (Q), each of which has a number of sublevels to distinguish the hydrogen-bonding capabilities (d = donor, a = acceptor, da = both, and 0 = none) or the degree of polarity (1−5, from low polarity to high polarity). It is well-known that this CG model can well reproduce several structural and dynamic properties of phospholipid bilayers, and it has been used to study the interactions between nanomaterials and lipid membranes in recent years.32−36 It should be noted that the effective time that the system has gone through is four times longer than the simulation sampling time because of the smoothed energy barrier in the MARTINI force field.31 2.2. CG Models. 2.2.1. Polymer-Functionalized AuNPs. The AuNP was fabricated by arranging 147 Au beads on a face centered cubic (fcc) lattice into a cuboctahedral shape, with the diameter of about 2.5 nm. The SC4 particle type in the MARTINI force field was assigned for Au beads. The bond length between neighboring Au beads was set to 0.408 nm, the same as the lattice constant of the Au atom. To make the AuNP move as a rigid body, a large force constant (15 000 kJ mol−1 nm−2) of the harmonic bonding potential was adopted, as described in our previous works.36−38 The stealth polymers covalently decorated on the AuNP surface are composed of several connected beads (the number of the beads can be varied), and the length of polymers is denoted by the number (N) of beads. For hydrophilic PEG, the beads are all hydrophilic and noncharged (type SN0), other related force field parameters were taken from the work of Lee et al.39 For zwitterionic polymer, the ending two beads are charged, where the last bead carries one positive charge (type Qd, + 1e) and the penultimate bead carries one negative charge (type Qa, −1e), the remaining beads are the same as that of PEG.21 In the present work, the zwitterionic polymer can undergo a pH-dependent zwitterionic-to-cationic charge conversion, and thus the negatively charged beads will be protonated and become charge neutral at specific pH values. There are a total of 100 polymer chains evenly grafted on the AuNP surface, the grafting density is reasonable according to the experimental range.21 2.2.2. Lipid Membrane. Dipalmitoylphosphatidylcholine (DPPC) and dipalmitoylphosphatidylglycerole (DPPG), two typical lipids in the MARTINI force field, were used to build the lipid membrane in the present work. Both of them contain 12 coarse-grained beads, but DPPG has a glycerol group instead of a choline group, which confers a negative net charge on the lipid. To model the DPPG lipid, the bead type Q0 in the headgroup of the DPPC lipid was changed to P4. The outer leaflet of the membrane is composed of 500 DPPC and the inner leaflet consists of 400 DPPC and 100 DPPG, which can mimic the natural membrane of cells more accurately.40 This simple membrane model has been widely adopted in the study of NP−biomembrane interactions.32,36,41−43 All models used in this work are illustrated in Figure 1. 2.3. Simulation Details. A simulation box with dimensions of 18 × 18 × 32 nm3 was created. The pre-established lipid membrane was then 14481

DOI: 10.1021/acs.langmuir.7b03544 Langmuir 2017, 33, 14480−14489

Article

Langmuir

Figure 1. Schematic illustration of the models of (a) lipid, (b) AuNP decorated with hydrophilic PEGs, and (c) AuNP decorated with pHresponsive zwitterionic polymers. The lipid headgroups are shown in blue, lipid tails in silver, gold core in yellow, PEG in green, zwitterionic polymer in blue, and zwitterionic polymer after protonation in magenta. equilibrated in the water by a 400 ns MD simulation. After that, the polymer-grafted AuNPs were placed about 2 nm above the membrane surface. Enough sodium ions were subsequently added to keep electroneutrality. The systems were first minimized by the steepestdescent method to remove inappropriate geometry or steric overlap. Then, the polymer-grafted AuNPs and the lipid membrane were constrained to equilibrate water and ions for 100 ns. Finally, the constraint was removed, and a 400 ns MD simulation was performed for each system to achieve equilibration. All simulations were performed under the NPT ensemble. The temperature was controlled to 310 K via the V-rescale thermostat with a relaxation time of 1 ps. The pressure was coupled to 1 bar by using an isotropic Berendsen barostat with a time constant of 5 ps. A cutoff distance of 1.2 nm was used for van der Waals (vdW) interactions. To provide a more realistic description of the interaction between charged NPs and lipid membranes, the particle mesh Ewald (PME) method44 with a cutoff radius of 1.2 nm was adopted to treat electrostatic interaction. The dielectric constant is set to 15 to compensate for the neglect of explicit polarization of the MARTINI water model. A time step of 20 fs was used for integration, and the trajectory was saved every 100 ps for analysis. The potential of mean force (PMF) profiles of PEG-AuNPs and Zwitt-AuNPs adsorbed on the membrane surface were determined by the umbrella sampling technique45 and the weighted histogram analysis method.46 Besides, we also computed the PMF of Zwitt-AuNPs entering into the lipid membrane core at different protonation degrees. The starting configurations with a 0.2 nm step size were generated by an external force for acceleration, and the biased simulation was done by a harmonic potential with a force constant of 1000 kJ mol−1 nm−2. Each window was simulated up to 400 ns. The first 200 ns of each run was discarded as equilibration, and the last 200 ns was chosen for data analysis. All simulations were performed by using the GROMACS 4.5.4 package.47 The Visual Molecular Dynamics (VMD) software was implemented for structure visualization.48

Figure 2. Time sequence of representative snapshots of interactions between two different types of AuNPs and lipid membranes. (a) The PEG-AuNP and (b) the Zwitt-AuNP. Water molecules are not displayed for clarity.

membrane (Figure 2a). For the Zwitt-AuNP (unprotonated), there exists dipole−dipole interactions between the grafted zwitterionic polymers and the zwitterionic DPPC lipids. Therefore, the Zwitt-AuNP can weakly adsorb on the outer membrane surface at last (Figure 2b). In both cases, the penetration of AuNPs into the membrane core was not observed. However, previous studies have demonstrated that hydrophobic NPs can spontaneously penetrate into lipid membranes.49−51 For example, Gupta et al.49 explored the penetration of AuNPs through the model skin lipid membrane by computer simulations. They demonstrated that AuNPs with different sizes can both disrupt the bilayer packing and enter the interior of the bilayer rapidly. Thus, the results obtained here indicate that the stealth polymer modification can block the penetration of AuNPs into lipid membranes. This is consistent with the results observed by Ding et al. through DPD simulations.19 To further understand the interaction mechanisms, we calculated the PMFs of PEG-AuNPs and Zwitt-AuNPs adsorbed on the membrane surface as a function of the center-of-mass separation distance between AuNPs and lipid membranes. As revealed in Figure 3, the PMF keeps growing as the PEG-AuNP approaches the outer membrane surface, indicating that there is a free energy barrier (about 94 kcal/mol) for spontaneous adhesion of the PEG-AuNP onto the lipid membrane due to the hydrophilic repulsion. For the Zwitt-AuNP, the PMF first

3. RESULTS AND DISCUSSION The interactions between two stealth polymer-modified AuNPs (PEG-AuNPs and Zwitt-AuNPs) and lipid membranes were first studied by CGMD simulations for a comparison. Then, we mainly focused on the effects of protonation degree and polymer chain length on the cellular uptake of Zwitt-AuNPs. Unless otherwise specified, the polymer chain length is N = 5. Results are shown in Figures 2−13. 3.1. Interactions between PEG-AuNPs and ZwittAuNPs with Lipid Membranes. As shown in Figure 2, the interactions between these two types of polymer-grafted AuNPs and lipid membranes are remarkably different. For the PEGAuNP, since the PEG coating is hydrophilic, there is no attractive interaction between the PEG-AuNP and the hydrophilic lipid head groups. Moreover, the hydration shell formed around the hydrophilic PEG coating and the zwitterionic DPPC lipids could produce a hydrophilic repulsion between the PEG-AuNP and the lipid membrane. Thus, the PEG-AuNP fluctuates randomly in water and cannot adsorb on the outer membrane surface throughout the simulation, even leaves far away from the lipid

Figure 3. PMF curves for PEG-AuNP and Zwitt-AuNP adsorbed on lipid membranes. The inset is the snapshots of Zwitt-AuNP at denoted distances, z = 6 (a), 4.5 (b), 3 nm (c). 14482

DOI: 10.1021/acs.langmuir.7b03544 Langmuir 2017, 33, 14480−14489

Article

Langmuir decreases as the distance becomes smaller, and there is a local minimal free energy value on the position close to the outer leaflet surface, which indicates that the Zwitt-AuNP can spontaneously approach the outer membrane surface because of the dipole−dipole interaction. After then, the PMF increases, and a free energy barrier (about 17 kcal/mol) must be overcome if the Zwitt-AuNP wants to tightly stick to the lipid membrane, which can be explained by the existence of electrostatic repulsion between the positive groups of zwitterionic polymers and the positive choline groups of lipids. The PMF calculations clearly explain the results shown in Figure 2. From the above analysis, we can know that it is easier for the Zwitt-AuNP to approach the outer membrane surface when compared with the PEG-AuNP; thus, it can create a more favorable prerequisite for the coming cellular uptake. For instance, the acidic pH condition around the cancer cell membranes could make the zwitterionic polymers protonated, which can enhance the cellular uptake of Zwitt-AuNPs, as discussed in section 3.2. Zhou et al.52 also demonstrated that the covalent conjugation of zwitterionic phosphorylcholine onto gold nanorods leads to enhanced cellular uptake when compared with gold nanorods decorated with PEG. This experimental result can be explained by the above analysis to some extent. In addition, it has been proved that NPs decorated with zwitterionic polymers can better resist the adsorption of plasma proteins than those coated with PEG when circulating in the blood.17 Therefore, the NPs decorated with zwitterionic polymers will present more attractive prospects in biomedical applications. 3.2. Effect of Protonation Degree on the Interactions between Zwitt-AuNPs and Lipid Membranes. In recent years, some pH-sensitive zwitterionic polymers have been reported,20,21 which can promote the cellular uptake of NPs, especially for cancer tissue. Therefore, much higher tumor therapy efficiency could be obtained in the case of NPs decorated by zwitterionic polymers. However, the underlying transmembrane mechanism is still not well understood. In this section, we will discuss the effect of pH environment on the interactions between Zwitt-AuNPs and lipid membranes. For simplicity, the relationship between the protonation degree (α) of pH-sensitive groups with external pH is given by the Henderson−Hasselbalch formula as follows: α=

Figure 4. Time sequence of the center-of-mass separation distance in the z direction between Zwitt-AuNPs and lipid membranes.

Figure 5. Equilibrated states of Zwitt-AuNP with different protonation degree interacting with lipid membranes, (a) 25%, (b) 50%, (c) 75%, and (d) 100%. Water molecules are not displayed for clarity.

membrane. Obviously, three different interaction modes are presented here, i.e., adsorption on the membrane surface, translocation across the membrane, and encapsulation by the lipid membrane. When the protonation degree is 25%, the surface charge of AuNP is relative low; thus, the electrostatic interaction is not strong enough to make the AuNP pass through the lipid membrane but only to adsorb on the membrane surface (Figure 5a). In our previous work, we demonstrated that the AuNPs with low surface charge density can insert into the membrane core through hydrophobic contacts, which formed between the protruding solvent-exposed lipid tails and the hydrophobic ligands on AuNPs.36 However, in this study, the further penetration of AuNPs is suppressed by the hydrophilic zwitterionic polymer coating. For the Zwitt-AuNPs with 50% and 75% protonation degrees, both of them can step across the lipid membrane and reach the inner leaflet (Figure 5b,c). Meanwhile, it can also be observed that the Zwitt-AuNP with 50% protonation degree shows better translocation performance than the Zwitt-AuNP with 75% protonation degree, presumably due to the latter bearing more positive charges, thus it has a stronger electrostatic attraction with the inner leaflet of membrane, which will increase the difficulty of dropping off from the membrane. In a real biological system, these particles which have crossed the lipid membrane will continue to interact with other membrane constituents and finally enter into the interior environment of cells completely. After increasing the protonation degree to 100%, we can clearly see that the ZwittAuNP is totally wrapped by the curved lipid membrane, and a certain amount of lipid molecules adsorbs on the particle surface through the strong electrostatic attraction (Figure 5d). This formed vesicle-like structure resembles the intermediate state which often happeneds in the endocytosis pathway. In this case, the particle will need to take more time to escape from the lipid membrane and to translocate into the cell interior. Similar results have been reported by da Rocha et al.,53 which show that NPs can be internalized into cells through different mechanisms such as

1 × 100% 1 + 10 pH − pK a

where pKa is the acidity constant of the polymer. From the above equation, we can easily find that the protonation degree increases with the decrease of pH. In this work, four different protonation degree (25%, 50%, 75%, and 100%) are considered, which can qualitatively illustrate the influence of pH value in the environment on the interactions between Zwitt-AuNPs and cell membranes. To avoid unexpected artifacts, polymer chains were evenly selected for protonation. Figure 4 gives the center-of-mass separation distance between Zwitt-AuNPs and lipid membranes over time. It shows that these Zwitt-AuNPs can both quickly adsorb on the membrane surface from the solution driven by electrostatic attraction. All systems reach equilibrium after 150 ns, the final position these AuNPs stayed are evidently different. This indicates that the protonation degree has little effect on the adsorption of Zwitt-AuNPs onto membrane surface, it mainly influences the subsequent translocation process. Figure 5 shows the final configurations of the Zwitt-AuNPs with different protonation degrees interacting with the lipid 14483

DOI: 10.1021/acs.langmuir.7b03544 Langmuir 2017, 33, 14480−14489

Article

Langmuir

Figure 6. PMFs as a function of the center-of-mass separation distance between Zwitt-AuNPs and lipid membranes under different protonation degrees: (a) 25%, (b) 50%, (c) 75%, and (d) 100%.

passive translocation for NPs with low charge density or endocytosis for NPs with high charge density. To illustrate the thermodynamics of the above four systems, PMFs of Zwitt-AuNPs as a function of their z distances from the membrane center were calculated for the four protonation degree cases (Figure 6). At the 25% protonation degree, the Zwitt-AuNP can adhere onto the outer membrane surface spontaneously because there exists a local minimal free energy value on the outer leaflet. However, an even lower free energy minimum appears on the inner leaflet surface due to the potential difference in the asymmetric lipid membrane, and this indicates that the Zwitt-AuNP sticking to the inner leaflet surface should correspond to a more stable state, but an energy barrier (about 15 kcal/mol) prevents the penetration of Zwitt-AuNP into the bilayer center (Figure 6a). In the case of 50% protonation degree, the PMF decreases monotonously with the decrease of the z distance between the Zwitt-AuNPs and the lipid bilayer, and there is a local minimal free energy value on the location close to the inner leaflet surface (Figure 6b). The PMF at 75% protonation degree shows a similar trend (Figure 6c). This means that the Zwitt-AuNPs at 50% and 75% protonation degree can spontaneously cross the lipid membrane due to the enhancement of electrostatic attractions. However, the PMFs also demonstrate that these Zwitt-AuNPs need to overcome a large free energy barrier to further fall off the inner leaflet surface. At the 100% protonation degree, the PMF first decreases and appears a minimum energy value at the bilayer center, which indicates that the full wrapping of the Zwitt-AuNP corresponds to a stable state. However, the further translocation has been prevented because of a large free energy barrier (Figure 6d). The PMF calculations clearly illustrate the interactions between Zwitt-AuNPs and lipid membranes at different protonation degrees from the viewpoint of free energy.

Although previous studies indicated that the penetration efficiency of AuNPs can be improved by increasing their surface charge density,33,36,54 the results of this work demonstrate that increasing the protonation degree of zwitterionic polymer cannot always enhance the transmembrane efficiency of AuNPs. The AuNP can cross the lipid membrane only at moderate protonation degree. As described above, the protonation degree of pH-sensitive groups is usually related to the intrinsic pKa value and the exterior environment pH value. Therefore, the results shown here can provide a strategy to design pH-responsive biocompatible nanocarriers; that is, the pKa of zwitterionic polymer materials should be carefully adjusted according to the pH environment of target sites. Only in this way can faster and more efficient transmembrane delivery be obtained. In general, nanomaterials interacting with plasma membranes can often disrupt the membrane structure, which will result in the production of cytotoxicity. To give a quantitative characterization of the structural changes of the lipid bilayer in this study, we calculated the average order parameter of lipid tails as a function of polymer protonation degree. The average order parameter of lipid tails can characterize the molecular orientation of lipids, which is calculated by the following formula: P2 =

1 (3 cos2 θ − 1) 2

where θ is the angle between the tail bond and the membrane normal. P2 = 1 means that all lipid tails are parallel to the membrane normal, and P2 = 0 represents random distributions of the lipid tail.55 As can be seen from Figure 7, the lipid average order parameter increases slightly at low protonation degree (25%), and this is due to the fact that the adsorption of charged NPs on the membrane surface could induce the ordered reconstruction of lipids.56 With the further increase of protonation degree, the lipid 14484

DOI: 10.1021/acs.langmuir.7b03544 Langmuir 2017, 33, 14480−14489

Article

Langmuir

adsorbs onto the membrane surface and induces the membrane bending inward. In previous studies, it has been demonstrated that the charged thiol monolayer-protected AuNPs adsorbed on the membrane can subsequently penetrate into the membrane core and be stabilized with a snorkeling configuration, in which the hydrophobic ligands contact with the lipid carbon tails while the charged ligands stably bind to the lipid head-groups of both leaflets.36,57,58 However, in this study, the further penetration of AuNP into the hydrophobic membrane core has not been observed; instead, there gradually forms a pore on the membrane surface, which may be caused by the hydrophilicity of polymers that are decorated on the AuNP’s surface. It can be seen that there distributes a number of hydrophilic lipid headgroups on the periphery of this membrane pore. Then the Zwitt-AuNP migrates into the interior of membrane along the hydrophilic pore accompanied by the gradual alleviation of membrane bending. A similar transmembrane process has also been observed in previous simulation works. Lin et al.34 found that a model cell membrane can generate a nanoscale hole to assist the spontaneous translocation of cationic AuNPs as well as HIV-1 TAT peptides to the cytoplasm side under a transmembrane potential. Shimizu et al.59 demonstrated that cationic AuNPs can cross the phospholipid bilayer through a membrane pore which is formed by exerting an external electric field. After the ZwittAuNP crossing the lipid membrane and reaching the inner leaflet, the membrane pore disappears and the lipid bilayer almost recovers to its initial state. As illustrated in Figure 5b, the top views of the system clearly show the whole process from the membrane pore formation to pore closure. This self-healing behavior of lipid membranes has also been demonstrated by some previous simulation studies.34,49,59,60 For the above translocation process, we quantitatively analyzed the change of membrane pore size over simulation time. The calculation method of pore size referred to in the work of Lin and co-workers.34 Briefly, the lipid membrane was first meshed into n square grids on the x−y plane. Then, every grid was checked to see whether lipid head groups exist or not. The grid elements in which lipid head groups that cannot be found are considered membrane pore area, while grid elements that are occupied by lipid head groups are not. As shown in Figure 9a, the membrane pore starts to form at 25 ns, and it expands quickly over the simulation time. At 42 ns, it reaches a maximum size of 32 nm2, which approximately equals the cross-sectional area of the Zwitt-AuNP. At this time, the Zwitt-AuNP just locates in the center of the membrane pore. Then the size of the pore quickly reduces and basically approaches zero after 80 ns, indicating that the pore completely closes once the particle crosses the lipid membrane. As a whole, this pore formation and closure process is relatively quick. To explore the structural characters of the system when the Zwitt-AuNP is passing the membrane pore, the density profiles of each components in the z direction were analyzed, as illustrated in Figure 9b. It clearly shows that there exists some hydrophilic lipid headgroups and water molecules in the pore region, indicating that some mediums such as water and ions can exchange between extracellular fluid and cytosol along this hydrated channel, which may lead to acute cytotoxicity.33,36,61 3.2.2. Full Wrapping Process. The full wrapping process of the Zwitt-AuNP with 100% protonation degree by the lipid membrane is also revealed by a series of trajectory snapshots during the simulation, as depicted in Figure 10. It shows that the Zwitt-AuNP which is initially placed in water quickly adsorbs onto the membrane and results in membrane bending to wrap it

Figure 7. Influence of protonation degree of Zwitt-AuNPs on average order parameter of lipid tails. (The dashed line in the figure represents the average order parameter of lipid tails when there is no Zwitt-AuNPs added.)

average order parameter decreases, indicating that the bilayer structure of the membrane is destroyed during the cellular uptake process; thus, the lipid tails arrange in random order. This shows consistency with previous works which suggest that increasing AuNP’s surface charge density can cause more disruption to the lipid membranes.33,36 However, for AuNPs decorated with pHresponsive zwitterionic polymers, the membrane damage and the concomitant cytotoxicity originated from the zwitterionic-tocationic charge conversion would produce only in the tumor environment and thus will lead to the possibility of targeted tumor therapy. As above, the Zwitt-AuNPs with 50% and 75% protonation degree can directly translocate across the lipid membrane, while the Zwitt-AuNP with 100% protonation degree is fully wrapped by the lipid membrane. In the following text, we will choose the Zwitt-AuNPs with 50% and 100% protonation degrees to examine in greater detail about the dynamics of these two interaction processes, i.e., the direct translocation process and the full wrapping process, respectively. 3.2.1. Direct Translocation Process. Figure 8 presents the representative snapshots of the Zwitt-AuNP with 50% protonation degree interacting with the lipid membrane at different times. At the beginning, the Zwitt-AuNP was placed above the membrane. Driven by electrostatic attraction, it quickly

Figure 8. (a) Time sequence of representative snapshots of interactions between Zwitt-AuNPs with 50% protonation degree and lipid membranes and (b) the corresponding top views of the lipid membrane. Water molecules are not displayed for clarity. 14485

DOI: 10.1021/acs.langmuir.7b03544 Langmuir 2017, 33, 14480−14489

Article

Langmuir

Figure 9. (a) Change of membrane pore size over simulation time. (b) Density profiles of components in the z direction when NPs passing the membrane pore.

Figure 11. (a) Change of membrane size over simulation time. (b) Density profiles of components in the z direction at the equilibrated state.

Moreover, due to the hydrophilicity of the environment around the Zwitt-AuNP, some water molecules are trapped into the membrane core after the wrapping process, which is different from the water exchange along the formed membrane pore. The existence of water in this case could enhance the biological compatibility of the Zwitt-AuNP to some extent. 3.3. Effect of Polymer Chain Length on the Cellular Uptake of Zwitt-AuNPs. It is generally accepted that the polymer chain length can greatly influence the interactions between polymer-grafted NPs and plasma membranes.19,27,29 However, the effect of polymer chain length on the cellular uptake of Zwitt-AuNPs is far from being completely understood. Concerning this issue, four kinds of Zwitt-AuNPs with different polymer chain length (N = 3, 5, 7, and 9) were considered, and the interactions between these Zwitt-AuNPs and the lipid membrane were simulated at 50% and 100% protonation degrees of the polymer, respectively. The simulation results suggest an interesting interplay between polymer chain length and protonation degree. As dipicted in Figure 12, the difference in polymer chain length has little effect on the adsorption of Zwitt-AuNPs; both of them can quickly adhere to the membrane surface which is governed by the electrostatic attraction. However, the subsequent interaction process is strongly related to the polymer chain length and protonation degree. Figure 13 presents the final configurations of these systems at the end of 400 ns of simulation. The behaviors of Zwitt-AuNPs with polymer chain length N = 5 have been described above. From Figure 13, we can see that the interactions between Zwitt-AuNPs with shorter polymer chain length (N = 3) and lipid membranes are similar to that of ZwittAuNPs with polymer chain length N = 5; the particles can translocate across the lipid membrane at 50% protonation degree while they are fully wrapped by the curved membrane at 100% protonation degree. However, for the Zwitt-AuNPs with longer

Figure 10. Time sequence of representative snapshots of interactions between Zwitt-AuNPs with 100% protonation degree and lipid membranes. Water molecules are not displayed for clarity.

partly due to the electrostatic attraction. However, this ZwittAuNP carries more positive charges when compared with the one with 50% protonation degree; thus, more negatively charged lipids gradually spread onto the particle and finally wrap it completely. In this process, the membrane size first reduces rapidly and then tends to equilibrium after 100 ns, as shown in Figure 11a. It should be noted that although the membrane fluctuation is intensified in the wrapping process, no membrane pore is found in the simulation, indicating that the membrane integrity is well preserved. This phenomena is in line with the work conducted by Li et al.,27 who investigated the passive recruitment dynamics of lipids induced by the adsorption of charged NPs and concluded that the electrostatic attraction of NPs plays a crucial role in the wrapping process. In addition, they also reported that, with the increase of electrostatic energy, the charged NPs can even be wrapped by neutral lipid membranes.62 Ding et al.63 demonstrated that the increase of NP’s surface ligand density can obviously improve the wrapping extent of NPs by lipid membranes. Figure 11b displays the density profiles of each component in the z direction. It shows that the distribution curve of lipid tails gets wider and lower, while the distribution of lipid headgroups in the membrane core increases significantly, which is conforming to the formed vesicle-like structure. 14486

DOI: 10.1021/acs.langmuir.7b03544 Langmuir 2017, 33, 14480−14489

Article

Langmuir

present neither NPs translocation nor wrapping when low surface charge densities are used.53 However, as depicted in Figure 13b, the Zwitt-AuNPs with longer polymer chains (N = 7 and 9) are not wrapped by the lipid membrane at 100% protonation degree, and on the contrary, they can cross the lipid membrane and reach the inner leaflet, and the transmembrane efficiency is higher when longer polymer chains are used. Due to the increase of polymer chain length, the Zwitt-AuNPs exhibit a larger effective radius, thus increasing the bending energy of membrane to wrap the particles. However, even though the Zwitt-AuNPs present a strong electrostatic attraction with the lipid membrane in this situation, it is not enough to overcome the bending energy. Instead, an induced pore gradually forms on the membrane surface, and the ZwittAuNPs finally move to the inner leaflet of the lipid membrane through this membrane pore, as demonstrated in Figure 14.

Figure 12. Time sequence of the center-of-mass separation distance in the z direction between Zwitt-AuNPs with different polymer chain length and lipid membranes at 50% and 100% protonation degrees.

Figure 14. Transmembrane pathway of AuNPs coated by zwitterionic polymers with longer chain length (N = 7 and 9) at 100% protonation degree. Water molecules are not displayed for clarity.

Meanwhile, we notice that some lipids are pulled out from the membrane along with the translocation of Zwitt-AuNPs, which will lead to more severe disruption to the membrane structure. It should be noted that the Zwitt-AuNPs with longer polymer chain and higher protonation degree at tumor tissues can not only exhibit more efficient transmembrane delivery but also produce more severe disruption to cancer cells; thus, it is advantageous for these Zwitt-AuNPs to be used in targeted tumor therapy. Hence, based on the discussion above, we think that the increase of the polymer chain length will block the translocation of AuNPs when the protonation degree is not high, while it can improve the transmembrane efficiency of AuNPs at higher protonation degree in a direct translocation way. This interesting interplay between polymer chain length and protonation degree may provide some guidelines for designing and developing pHresponsive biocompatible nanocarriers in targeted tumor therapy.

4. CONCLUSIONS By employing a coarse-grained simulation method, we investigated the interactions between stealth polymer-decorated AuNPs and lipid membranes. Our results show that AuNPs coated by zwitterionic polymers are easier to approach the membrane surface than those grafted with PEG due to the weak electrostatic attraction between them, as demonstrated by the free energy calculation. For AuNPs decorated with pHresponsive zwitterionic polymers, the effect of protonation degree on their interactions with lipid membranes was discussed as well. Simulation results demonstrate that increasing the protonation degree of the zwitterionic polymer cannot always promote the permeability of AuNPs through lipid membranes. Different interaction modes were presented which depend on the protonation degree. The Zwitt-AuNP with low protonation degree just adsorbed on the membrane surface, while those with

Figure 13. Final equilibrated configurations of Zwitt-AuNPs with different polymer chain length interacting with lipid membranes at (a) 50% and (b) 100% protonation degree, respectively. Water molecules are not displayed for clarity.

polymer chain length (N = 7 and 9), they cannot translocate across the lipid membrane and just adhere onto the membrane at 50% protonation degree, and the bending of the membrane induced by the adsorption of Zwitt-AuNPs reduces along with the increase of polymer chain length (Figure 13a). This may be ascribed to the fact that the enhanced hydrophilicity of polymers with longer chain length weakens the electrostatic attraction between Zwitt-AuNPs and lipid membranes. The results are consistent with previous simulation work in that longer ligands 14487

DOI: 10.1021/acs.langmuir.7b03544 Langmuir 2017, 33, 14480−14489

Article

Langmuir

Effects on Protein Adsorption and Cellular Uptake. ACS Nano 2015, 9, 6996−7008. (9) Walkey, C. D.; Olsen, J. B.; Guo, H.; Emili, A.; Chan, W. C. W. Nanoparticle Size and Surface Chemistry Determine Serum Protein Adsorption and Macrophage Uptake. J. Am. Chem. Soc. 2012, 134, 2139−2147. (10) Cao, Z.; Jiang, S. Super-hydrophilic zwitterionic poly(carboxybetaine) and amphiphilic non-ionic poly(ethylene glycol) for stealth nanoparticles. Nano Today 2012, 7, 404−413. (11) Shao, Q.; Jiang, S. Molecular understanding and design of zwitterionic materials. Adv. Mater. 2015, 27, 15−26. (12) Schlenoff, J. B. Zwitteration: coating surfaces with zwitterionic functionality to reduce nonspecific adsorption. Langmuir 2014, 30, 9625−9636. (13) Min, W.; Zhao, D.; Quan, X.; Sun, D.; Li, L.; Zhou, J. Computer simulations on the pH-sensitive tri-block copolymer containing zwitterionic sulfobetaine as a novel anti-cancer drug carrier. Colloids Surf., B 2017, 152, 260−268. (14) Liao, M.; Liu, H.; Guo, H.; Zhou, J. Mesoscopic Structures of Poly(carboxybetaine) Block Copolymer and Poly(ethylene glycol) Block Copolymer in Solutions. Langmuir 2017, 33, 7575−7582. (15) Cheng, G.; Liao, M.; Zhao, D.; Zhou, J. Molecular Understanding on the Underwater Oleophobicity of Self-Assembled Monolayers: Zwitterionic versus Nonionic. Langmuir 2017, 33, 1732−1741. (16) Tatumi, R.; Fujihara, H. Remarkably stable gold nanoparticles functionalized with a zwitterionic liquid based on imidazolium sulfonate in a high concentration of aqueous electrolyte and ionic liquid. Chem. Commun. 2005, 83−85. (17) Yang, W.; Zhang, L.; Wang, S.; White, A. D.; Jiang, S. Functionalizable and ultra stable nanoparticles coated with zwitterionic poly(carboxybetaine) in undiluted blood serum. Biomaterials 2009, 30, 5617−5621. (18) Dai, Q.; Walkey, C.; Chan, W. C. W. Polyethylene Glycol Backfilling Mitigates the Negative Impact of the Protein Corona on Nanoparticle Cell Targeting. Angew. Chem., Int. Ed. 2014, 53, 5093− 5096. (19) Ding, H.-m.; Ma, Y.-q. Design strategy of surface decoration for efficient delivery of nanoparticles by computer simulation. Sci. Rep. 2016, 6, 26783. (20) Yuan, Y. Y.; Mao, C. Q.; Du, X. J.; Du, J. Z.; Wang, F.; Wang, J. Surface charge switchable nanoparticles based on zwitterionic polymer for enhanced drug delivery to tumor. Adv. Mater. 2012, 24, 5476−5480. (21) Mizuhara, T.; Saha, K.; Moyano, D. F.; Kim, C. S.; Yan, B.; Kim, Y. K.; Rotello, V. M. Acylsulfonamide - Functionalized Zwitterionic Gold Nanoparticles for Enhanced Cellular Uptake at Tumor pH. Angew. Chem., Int. Ed. 2015, 54, 6567−6570. (22) Tian, H.; Guo, Z.; Lin, L.; Jiao, Z.; Chen, J.; Gao, S.; Zhu, X.; Chen, X. pH-responsive zwitterionic copolypeptides as charge conversional shielding system for gene carriers. J. Controlled Release 2014, 174, 117−125. (23) Ding, H. M.; Ma, Y. Q. Theoretical and computational investigations of nanoparticle−biomembrane interactions in cellular delivery. Small 2015, 11, 1055−1071. (24) Ding, H. M.; Ma, Y. Q. Computational approaches to cellnanomaterial interactions: keeping balance between therapeutic efficiency and cytotoxicity. Nanoscale Horiz. 2018, DOI: 10.1039/ C7NH00138J. (25) Li, Y.; Kroger, M.; Liu, W. K. Shape effect in cellular uptake of PEGylated nanoparticles: comparison between sphere, rod, cube and disk. Nanoscale 2015, 7, 16631−16646. (26) Oroskar, P. A.; Jameson, C. J.; Murad, S. Simulated Permeation and Characterization of PEGylated Gold Nanoparticles in a Lipid Bilayer System. Langmuir 2016, 32, 7541−7555. (27) Li, Y.; Hu, Y. Computational investigation of the influence of chain length on the shielding effect of PEGylated nanoparticles. RSC Adv. 2014, 4, 51022−51031. (28) Yue, T.; Zhang, X. Molecular understanding of receptor-mediated membrane responses to ligand-coated nanoparticles. Soft Matter 2011, 7, 9104−9112.

high protonation degree were fully wrapped by the curved lipid membrane, which may result in the slowdown of transmembrane rate. Translocation of AuNPs across lipid membranes can only be obtained at moderate protonation degree. In this case, the ZwittAuNP induced the formation of a hydrophilic pore on the membrane surface, along which it crossed the lipid membrane. After the translocation was completed, the membrane pore was gradually closed. In addition, this work presents an interesting interplay between the chain length and the protonation degree of polymers on the interactions between Zwitt-AuNPs and lipid membranes. The increase of polymer chain length will block the translocation of AuNPs when the protonation degree is not high, while it can enhance the transmembrane efficiency of AuNPs at higher protonation degree in a direct translocation way. These findings are of immediate interest to the design and synthesis of pH-responsive nanomaterials based on zwitterionic polymers and can promote their further application in nanomedicine.

■

AUTHOR INFORMATION

Corresponding Author

*Tel./fax: +86 20 87114069. E-mail: [email protected]. ORCID

Libo Li: 0000-0001-7699-4484 Jian Zhou: 0000-0002-3033-7785 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was financially supported by National Natural Science Foundation of China (Nos. 21776093, 91334202, 21376089, and 21506066), the National Key Basic Research Program of China (No. 2013CB733500), Guangdong Science Foundation (Nos. 2014A030312007 and 2014A030310260), and the Fundamental Research Funds for the Central Universities (SCUT-2015ZP033 and SCUT-2017ZD069). We are grateful to the SCUTGrid at South China University of Technology and the ScGrid of the Supercomputing Center, the Computer Network Information Center of the Chinese Academy of Sciences for providing computation time.

■

REFERENCES

(1) Lim, E. K.; Kim, T.; Paik, S.; Haam, S.; Huh, Y. M.; Lee, K. Nanomaterials for Theranostics: Recent Advances and Future Challenges. Chem. Rev. 2015, 115, 327−394. (2) Boisselier, E.; Astruc, D. Gold nanoparticles in nanomedicine: preparations, imaging, diagnostics, therapies and toxicity. Chem. Soc. Rev. 2009, 38, 1759−1782. (3) Li, Z. X.; Barnes, J. C.; Bosoy, A.; Stoddart, J. F.; Zink, J. I. Mesoporous silica nanoparticles in biomedical applications. Chem. Soc. Rev. 2012, 41, 2590−2605. (4) Mura, S.; Nicolas, J.; Couvreur, P. Stimuli-responsive nanocarriers for drug delivery. Nat. Mater. 2013, 12, 991−1003. (5) Mahmoudi, M.; Lynch, I.; Ejtehadi, M. R.; Monopoli, M. P.; Bombelli, F. B.; Laurent, S. Protein−Nanoparticle Interactions: Opportunities and Challenges. Chem. Rev. 2011, 111, 5610−5637. (6) Walkey, C. D.; Chan, W. C. W. Understanding and controlling the interaction of nanomaterials with proteins in a physiological environment. Chem. Soc. Rev. 2012, 41, 2780−2799. (7) Lynch, I.; Dawson, K. A. Protein-nanoparticle interactions. Nano Today 2008, 3, 40−47. (8) Pelaz, B.; del Pino, P.; Maffre, P.; Hartmann, R.; Gallego, M.; Rivera-Fernández, S.; de la Fuente, J. M.; Nienhaus, G. U.; Parak, W. J. Surface Functionalization of Nanoparticles with Polyethylene Glycol: 14488

DOI: 10.1021/acs.langmuir.7b03544 Langmuir 2017, 33, 14480−14489

Article

Langmuir (29) Liang, Q. Penetration of polymer-grafted nanoparticles through a lipid bilayer. Soft Matter 2013, 9, 5594−5601. (30) Ding, H. M.; Ma, Y. Q. Controlling Cellular Uptake of Nanoparticles with pH-Sensitive Polymers. Sci. Rep. 2013, 3, 2804. (31) 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. (32) Tian, W. D.; Ma, Y. Q. Insights into the endosomal escape mechanism via investigation of dendrimer-membrane interactions. Soft Matter 2012, 8, 6378−6384. (33) Lin, J.; Zhang, H.; Chen, Z.; Zheng, Y. Penetration of Lipid Membranes by Gold Nanoparticles: Insights into Cellular Uptake, Cytotoxicity, and Their Relationship. ACS Nano 2010, 4, 5421−5429. (34) Lin, J.; Alexander-Katz, A. Cell Membranes Open “Doors” for Cationic Nanoparticles/Biomolecules: Insights into Uptake Kinetics. ACS Nano 2013, 7, 10799−10808. (35) Yue, T.; Xu, Y.; Li, S.; Zhang, X.; Huang, F. Lipid extraction mediates aggregation of carbon nanospheres in pulmonary surfactant monolayers. Phys. Chem. Chem. Phys. 2016, 18, 18923−18933. (36) Quan, X.; Peng, C.; Zhao, D.; Li, L.; Fan, J.; Zhou, J. Molecular Understanding of the Penetration of Functionalized Gold Nanoparticles into Asymmetric Membranes. Langmuir 2017, 33, 361−371. (37) Dong, J.; Li, J.; Zhou, J. Interfacial and Phase Transfer Behaviors of Polymer Brush Grafted Amphiphilic Nanoparticles: A Computer Simulation Study. Langmuir 2014, 30, 5599−5608. (38) Quan, X.; Peng, C.; Dong, J.; Zhou, J. Structural properties of polymer-brush-grafted gold nanoparticles at the oil-water interface: insights from coarse-grained simulations. Soft Matter 2016, 12, 3352− 3359. (39) Lee, H.; de Vries, A. H.; Marrink, S.-J.; Pastor, R. W. A CoarseGrained Model for Polyethylene Oxide and Polyethylene Glycol: Conformation and Hydrodynamics. J. Phys. Chem. B 2009, 113, 13186− 13194. (40) Zachowski, A. Phospholipids in animal eukaryotic membranes: transverse asymmetry and movement. Biochem. J. 1993, 294, 1−14. (41) Li, Z. L.; Ding, H. M.; Ma, Y. Q. Translocation of polyarginines and conjugated nanoparticles across asymmetric membranes. Soft Matter 2013, 9, 1281−1286. (42) He, X.; Qu, Z.; Xu, F.; Lin, M.; Wang, J.; Shi, X.; Lu, T. Molecular analysis of interactions between dendrimers and asymmetric membranes at different transport stages. Soft Matter 2014, 10, 139−148. (43) He, X.; Lin, M.; Lu, T.; Qu, Z.; Xu, F. Molecular analysis of interactions between a PAMAM dendrimer-paclitaxel conjugate and a biomembrane. Phys. Chem. Chem. Phys. 2015, 17, 29507−29517. (44) 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. (45) Torrie, G. M.; Valleau, J. P. Monte Carlo free energy estimates using non-Boltzmann sampling: Application to the sub-critical LennardJones fluid. Chem. Phys. Lett. 1974, 28, 578−581. (46) Kumar, S.; Rosenberg, J. M.; Bouzida, D.; Swendsen, R. H.; Kollman, P. A. THE weighted histogram analysis method for free-energy calculations on biomolecules. I. The method. J. Comput. Chem. 1992, 13, 1011−1021. (47) Hess, B.; Kutzner, C.; Van Der Spoel, D.; Lindahl, E. GROMACS 4: Algorithms for highly efficient, load-balanced, and scalable molecular simulation. J. Chem. Theory Comput. 2008, 4, 435−447. (48) Humphrey, W.; Dalke, A.; Schulten, K. VMD: visual molecular dynamics. J. Mol. Graphics 1996, 14, 33−38. (49) Gupta, R.; Rai, B. Penetration of Gold Nanoparticles through Human Skin: Unraveling Its Mechanisms at the Molecular Scale. J. Phys. Chem. B 2016, 120, 7133−7142. (50) Gupta, R.; Kashyap, N.; Rai, B. Transdermal cellular membrane penetration of proteins with gold nanoparticles: a molecular dynamics study. Phys. Chem. Chem. Phys. 2017, 19, 7537−7545. (51) Li, Y.; Chen, X.; Gu, N. Computational Investigation of Interaction between Nanoparticles and Membranes: Hydrophobic/ Hydrophilic Effect. J. Phys. Chem. B 2008, 112, 16647−16653.

(52) Zhou, W.; Shao, J.; Jin, Q.; Wei, Q.; Tang, J.; Ji, J. Zwitterionic phosphorylcholine as a better ligand for gold nanorods cell uptake and selective photothermal ablation of cancer cells. Chem. Commun. 2010, 46, 1479−1481. (53) Rocha, E. L. d.; Caramori, G. F.; Rambo, C. R. Nanoparticle translocation through a lipid bilayer tuned by surface chemistry. Phys. Chem. Chem. Phys. 2013, 15, 2282−2290. (54) Allen, C.; Qiu, T. A.; Pramanik, S.; Buchman, J. T.; Krause, M. O.; Murphy, C. J. Research highlights: investigating the role of nanoparticle surface charge in nano−bio interactions. Environ. Sci.: Nano 2017, 4, 741−746. (55) Murtola, T.; Róg, T.; Falck, E.; Karttunen, M.; Vattulainen, I. Transient Ordered Domains in Single-Component Phospholipid Bilayers. Phys. Rev. Lett. 2006, 97, 238102. (56) Wang, B.; Zhang, L.; Bae, S. C.; Granick, S. Nanoparticle-induced surface reconstruction of phospholipid membranes. Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 18171−18175. (57) Simonelli, F.; Bochicchio, D.; Ferrando, R.; Rossi, G. MonolayerProtected Anionic Au Nanoparticles Walk into Lipid Membranes Step by Step. J. Phys. Chem. Lett. 2015, 6, 3175−3179. (58) Van Lehn, R. C.; Atukorale, P. U.; Carney, R. P.; Yang, Y.-S.; Stellacci, F.; Irvine, D. J.; Alexander-Katz, A. Effect of Particle Diameter and Surface Composition on the Spontaneous Fusion of MonolayerProtected Gold Nanoparticles with Lipid Bilayers. Nano Lett. 2013, 13, 4060−4067. (59) Shimizu, K.; Nakamura, H.; Watano, S. MD simulation study of direct permeation of a nanoparticle across the cell membrane under an external electric field. Nanoscale 2016, 8, 11897−11906. (60) Adhikari, U.; Goliaei, A.; Tsereteli, L.; Berkowitz, M. L. Properties of Poloxamer Molecules and Poloxamer Micelles Dissolved in Water and Next to Lipid Bilayers: Results from Computer Simulations. J. Phys. Chem. B 2016, 120, 5823−5830. (61) Verma, A.; Stellacci, F. Effect of Surface Properties on Nanoparticle-Cell Interactions. Small 2010, 6, 12−21. (62) Li, Y.; Gu, N. Thermodynamics of Charged Nanoparticle Adsorption on Charge-Neutral Membranes: A Simulation Study. J. Phys. Chem. B 2010, 114, 2749−2754. (63) Ding, H. M.; Ma, Y. Q. Role of physicochemical properties of coating ligands in receptor-mediated endocytosis of nanoparticles. Biomaterials 2012, 33, 5798−5802.

14489

DOI: 10.1021/acs.langmuir.7b03544 Langmuir 2017, 33, 14480−14489