Low-Density Lipoproteins and Human Serum Albumin as Carriers of

Subscriber access provided by READING UNIV

Article

Low density lipoproteins and human serum albumin as the carriers of squalenoylated drugs: insights from molecular simulations Semen O. Yesylevskyy, Christophe Ramseyer, Mariia Savenko, Simona Mura, and Patrick Couvreur Mol. Pharmaceutics, Just Accepted Manuscript • DOI: 10.1021/acs.molpharmaceut.7b00952 • Publication Date (Web): 04 Jan 2018 Downloaded from http://pubs.acs.org on January 4, 2018

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.

Molecular Pharmaceutics 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

Molecular Pharmaceutics

Low density lipoproteins and human serum albumin as the carriers of squalenoylated drugs: insights from molecular simulations Semen O. Yesylevskyy 1*, Christophe Ramseyer 2, Mariia Savenko 2, Simona Mura 3, Patrick Couvreur 3.

1

Department of Physics of Biological Systems, Institute of Physics of the National Academy of

Sciences of Ukraine, Prospect Nauky 46, 03028 Kyiv, Ukraine 2

Laboratoire Chrono Environnement UMR CNRS 6249, Université de Bourgogne Franche-

Comté, 16 route de Gray, 25030 Besançon Cedex, France 3

Institut Galien Paris-Sud, UMR 8612, CNRS, Univ Paris-Sud, Université Paris-Saclay, Faculté

de Pharmacie, 5 rue Jean-Baptiste Clément, F-92296 Châtenay-Malabry Cedex, France Keywords: squalenoylated drugs, molecular dynamics, docking, low density lipoproteins, human serum albumin.

Abstract: We have studied the interaction of three clinically promising squalenoylated drugs (gemcitabine-squalene,

adenine-squalene

and

doxorubicin-squalene)

with

low-density

ACS Paragon Plus Environment

1

Molecular Pharmaceutics 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

lipoproteins (LDL) by means of atomistic molecular dynamics simulations. It is shown that all studied squalenoylated drugs accumulated inside the LDL particles. This effect is promoted by the squalene moiety, which acts as an anchor and drives the hydrophilic drugs into the hydrophobic core of the LDL lipid droplet. Our data suggest that LDL particles could be a universal carriers of squalenoylated drugs in the blood stream. Interaction of gemcitabinesqualene with human serum albumin (HSA) was also studied by ensemble of docking simulations. It is shown that HSA could also act as a passive carrier of this bioconjugate. To be noted that the binding of squalene moiety to HSA was unspecific and did not occur in the binding pockets devoted to fatty acids.

Introduction Squalene (SQ) is a natural lipid, precursor of the cholesterol’s biosynthesis, which was recently proposed as a promising biocompatible material for drug delivery purposes. The concept of “squalenoylation” is based on the covalent conjugation of squalene with a drug molecule. The obtained bioconjugates are able to spontaneously form nanoparticles in water without the need of any other excipient

1, 2

. The in vitro and in vivo pharmacological evaluation of these

nanoparticles revealed impressive anticancer activity

3-6

, neuroprotection

7

or antimicrobial

activity 8, depending of the nature of the drug conjugated to the squalene. Once introduced into the blood stream, squalene-based nanoparticles encounter a complex environment composed of various blood cells and plasma components. It was shown in previous studies

9

that, once in contact with human blood, gemcitabine-squalene (SQGem) nanoparticles

mainly distributed between low-density lipoproteins (LDL, ~50 %) and human serum albumin (HSA, ~20 %). The interaction of SQGem and albumin was also observed in cell culture medium


2

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


10

. That is why studying interaction of squalene-based drugs with blood proteins and lipoproteins

are of primary importance for understanding their pharmacokinetics. It was previously shown that the LDL particles were able to incorporate into SQGem at a molecular level 9. The targeting capacity of these SQGem-incorporated LDL particles towards LDL receptors, which are overexpressed in cancer cells, was shown both in vitro and in vivo, thus highlighting the therapeutic importance of this interaction 9. The observed spontaneous interaction between the squalenoylated drugs and the lipoproteins may therefore represent a novel concept in drug delivery, since the squalenoylated bioconjugates utilize endogenous lipoproteins as an “indirect” natural carrier. We have previously shown

9

that in silico modeling of SQGem interaction with LDL was in

agreement with experimental data and could be used to predict the properties of other squalenoylated drugs. Thus, in the current study, we tested the interaction of two other squalenoylated drugs with LDL, ie. adenosine-squalene (SQAde), and doxorubicin-squalene (SQDox), Adenosine (Ade), doxorubicin (Dox) and squalene (SQ) being used for comparison. Propensity of accumulation inside the LDL lipid core was estimated for these compounds and compared with previous data on SQGem and free gemcitabine (Gem). In this study, we also investigated the interaction of SQGem with HSA using advanced docking simulations, which complements recent experimental findings of increased cellular uptake of SQGem by extracellular proteins

10

. Preferable binding sites on the protein surface

were identified, as well as, their relation to known binding pockets of fatty acids, routinely transported by HSA.

Methods


3


Page 4 of 28

LDL simulations Full-size LDL particles are too large to be simulated in a reasonable time in all-atom representation. However, in the current study we were interested in the thermodynamics of interaction of the ligands with LDL lipid droplet, which didn’t require explicit modeling of the whole particle. We constructed a simplified atomistic model, which reproduces the essential thermodynamic properties of the lipid core of LDL particles. We considered a narrow columnlike slice of the LDL particle which contained only 10% of its whole volume. This slice was simulated with periodic boundary conditions which resembles the usual simulation setup for lipid bilayers but with much thicker hydrophobic core (Supplementary Fig. S1). This dramatically reduced the size of the system and made it computationally tractable on atomistic level, while keeping major physical features of the LDL lipid droplet such as thick hydrophobic core and an amphiphilic surface layer. The lipid composition of the system is summarized in Table 1. This composition was used previously in the coarse-grained modeling of full-size LDL 11 and corresponds to typical human LDL particle.

Table 1. Composition of the system used for production simulations. The number of lipid molecules corresponds to 1/10 of the whole LDL particle used in 11. Molecule

Number

POPC

64

18:1 Lyso PC

8

Cholesterol

60

Cholesterol oleate 160


4

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


Glyceryl trioleate

18

Water

4900

The system was designed as follows. The hydrophobic core containing cholesterol, cholesterol oleate and glyceryl trioleate was constructed as a random mixture of components and simulated for 50 ns in NPT conditions to obtain an equilibrium lipid density. After that, two identical surface layers containing POPC and Lyso PC molecules were added to the sides of the core. The surface layers were taken from the pre-equilibrated mixed lipid bilayer of needed composition. The system was solvated at both sides of the surface layers and subject to equilibration for 500 ns. The snapshot of equilibrated system is shown in Fig 1.


5


Page 6 of 28

Figure 1. Snapshot of equilibrated LDL-like slice. POPC lipids are shown in blue, lyso PC in red, cholesterol in orange, cholesterol oleate in gray and glyceryl trioleate in green. Water is not shown for clarity. Single periodic cell is shown.

The slipids force field 12 was used for all lipids. This force field is considered as one of the best all-atom force fields for lipid systems nowadays, which reproduces properties of the lipid bilayer above the melting point remarkably well

13

. Topologies for lyso PC, cholesterol oleate and

glyceryl trioleate were constructed manually by combining pieces of POPC and cholesterol


6

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


topologies. Initial topologies of the pristine drugs Gem, Ade and Dox, their conjugates with squalene SQGem, SQAde and SQDox and pristine squalene were generated by Acpype topology generator

14

. The structures of the drugs were optimized in Gaussian09

15

at the B3LYP/6-

31++G(d) level of theory. The ESP partial charges were computed and added to initial topology. The atom types of squalene tails were adjusted to match slipids force field. All simulations were performed in Gromacs 5.1.2 package

16

. TIP3P water model was used.

All simulations were performed in NPT conditions with the temperature 320 K and the pressure of 1 bar maintained by v-rescale thermostat and Berendsen barostat respectively. The time step of 2 fs was used with all bonds converted to rigid constraints. This choice of parameters was used with great success in our previous works

9, 17

and corresponds to recommended values for

slipids force field, which was tested against available experimental data for various lipid bilayers 12, 13

. The temperature of 320K allows keeping the lipid tails in liquid phase and speeds up

equilibration. In addition, this makes computed potentials of mean force directly comparable with our previous work 9. Potentials of mean force (PMFs) of transferring molecules from water to the core of the LDL particle were computed by umbrella sampling simulations. The center of masses of the ligand was restrained by harmonic potential with the force constant of 2000 kJ·mol-1·nm-2 at different distances from the center of the slice. This resulted in 90 umbrella sampling windows with the step of 0.1 nm along Z axis. Each window was simulated for at least 500 ns and the last 100 ns were used for analysis. PMFs were obtained with the weighted histogram technique

18

as

implemented in Gromacs package.

HSA simulations


7


Page 8 of 28

Binding of SQGem molecules to HSA was assessed by means of ensemble docking simulations, which account for large-scale protein dynamics. The protocol established in our previous studies

19, 20

was used with some modifications. We utilized all-atom MD trajectory of

pre-equilibrated HSA in water described elsewhere 21. The length of trajectory was 100 ns. 200 frames were extracted from trajectory at even intervals and used as a representative ensemble of protein conformations. All protein conformations were aligned by their peptide backbones. SQGem molecule was docked to each of these conformations within the docking volume of 20 nm3 centered sequentially at each solvent-exposed residue of HSA. The protein was kept fixed for each docking run since its flexibility was already accounted by using ensemble of conformations from MD trajectory. This resulted in ~73000 independent docking simulations. Such procedure allows effective scanning of the whole solvent-exposed protein surface in each selected protein conformation for binding with SQGem. Any binding pocket or interdomain cleft, which opens and closes dynamically in the course of thermal protein fluctuations, becomes available for docking which is not possible using static protein structure. Large number of docking simulations allows collecting sufficient statistics for reliable estimates of preferable binding sites of SQGem. The SQGem ligand and the corresponding protein structures were prepared for docking using MGLTools-1.5.6 software. The docking was performed with the Autodock Vina

22

with the

default scoring function. Analysis of the docking results was performed with custom software based on the Pteros 2.0 molecular modeling library 23, 24. For each docking simulation single topranked pose was recorded.

Results


8

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


Accumulation of squalenoylated bioconjugates in LDL particles

Figure 2. Potentials of mean force of transferring individual molecules from water to the lipid core of the LDL particle. The plots are superimposed onto the snapshot of simulated LDL-like slice shown as semi-transparent background and aligned with X axis. POPC lipids are shown in blue, lyso PC in red, cholesterol in orange, cholesterol oleate in gray and glyceryl trioleate in green.


9


Page 10 of 28

Figure 2 shows potentials of mean force of transferring different ligands from water to the center of the LDL particle. It can clearly be seen that very deep (~100 kJ/mol) energy well exists inside the LDL particle for purely hydrophobic SQ molecules. The bottom of this energy well is essentially flat across the whole hydrophobic part of the LDL droplet which suggests that the SQ molecules distribute uniformly inside the core of LDL particle. Linkage of SQ moiety to the hydrophilic drugs decreases the depth of the energy well progressively in a row SQAde SQGem SQDox. The shape of the energy profile in the hydrophobic LDL core becomes rugged for these derivatives. All three derivatives show a local energy well at the interfacial shell of lipids. This interfacial energy well is very pronounced in the case of SQDox, barely visible for SQAde and has an intermediate depth for SQGem. A second local energy well is observed in the center of the LDL particle but it should be interpreted with caution, due to possible artifacts in this region caused by simulation setup (see Limitations below). These results show that the accumulation in the LDL particles is thermodynamically favorable for all three studied squalene derivatives. SQAde accumulates almost uniformly in the whole volume of the LDL particle, when SQGem shows propensity to accumulate more in the interfacial shell of lipids than in its hydrophobic core. SQDox displays strong propensity to accumulate at the interfacial layer with less probable location in the hydrophobic core. Free drugs show drastic difference of interaction with LDL in comparison with their squalene derivatives. No binding of Gem molecules to the LDL particles is observed and the whole energy profile is repulsive. Dox weakly binds to the lipid-water interface of the LDL. Ade displays a slightly stronger binding to the interfacial layer of LDL. In addition Ade has two regions of binding inside the particle which could be reached by overcoming high energy barriers between


10

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


them. However, Ade clearly does not accumulate in the LDL because the free energies of Ade in these regions are the same as in the water phase.

Interaction of SQGem with HSA The statistical analysis of the best docking poses suggests that SQGem binds unspecifically to the vast majority of solvent-exposed amino acids of HSA. There are several sites of preferential binding of SQGem which do not show strong preference in binding score but seems to be more kinetically accessible. The strongest binding energy was -10.8 kcal/mol, while the median binding energy was about -5.5 kcal/mol (Supplementary Fig. S2). The broad distribution of the binding energies suggests unspecific character of the binding, driven by (i) the general hydrophobicity of the protein surface and (ii) the number of non-specific Van der Waals contacts with ligand in particular surface region. The supplementary movie (available online), displaying the distribution of the density of ligand atoms over the protein surface, provides another visualization of this observation. These findings were corroborated by computing the probabilities of SQGem binding to individual HSA aminoacids (Supplementary Fig. S3). The probability of binding was the highest in deep interdomain grooves, thus supporting the hypothesis of an unspecific binding due to weak Van der Waals and hydrophobic interactions. Clustering analysis of aligned best docking poses revealed several well-distinguished clusters of similar conformations, which are located in different parts of the HSA surface (Fig. 3). Such clusters represent the most probable locations of SQGem diffusing from the bulk solution towards the protein surface, thus representing the most kinetically accessible binding spots. It is clearly seen that SQGem molecules fill the grooves on the protein surface where the number of


11


Page 12 of 28

non-specific Van der Waals contacts could be maximized. The definitions of the ligand clusters are provided in Supplementary information. It is important to note that the “clusters” discussed here are not physical aggregates of molecules bound to HSA surface. The term “cluster” is used here in mathematical sense to indicate grouping of individual SQGem molecules which bind in similar positions and orientation as revealed by docking simulations.

Figure 3. Largest clusters of SQGem molecules bound to HSA. The protein is shown in cartoon representation. The SQGem molecules are shown as wireframe. Each distinct cluster of SQGem molecules is marked by different color. Single conformation of HSA is shown for clarity while all other conformations where aligned to it to provide correct relative positions of all docked SQGem molecules.


12

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


We also compared preferential binding spots of SQGem with the binding pockets of fatty acids revealed in crystallographic studies 25. It appeared that SQGem rarely interacted with the known fatty acid binding sites (Fig. 4). It is evident that SQGem molecules do not penetrate as deep into the hydrophobic pockets as small fatty acids which is easily explained by the larger volume of bulky branched SQ tails. The majority of SQGem binding appeared on the shallow grooves of protein surface, which were never occupied by fatty acids in the crystal structures (Fig. 4).

Figure 4. Comparison of the preferable binding positions of SQGem (colored wireframe) and the fatty acids (blue space-fill reresentation) derived from

25

. Protein structure is hidden for clarity.

Each distinct cluster of SQGem molecules is marked by different color.

Discussion


13


Recently

developed

self-assembled

nanoparticles

Page 14 of 28

made

of

various

drugs-squalene

bioconjugates show great potential as anti-cancer, neuroprotective or anti-microbial agents

3-8

.

Assessment of their interaction with blood constituents, such as lipoproteins and serum albumin, is of crucial importance to understand their pharmacodynamics and predict the in vivo fate after intravenous administration. Recently we have disclosed the unique mechanism of transport of SQGem bioconjugates in the blood 9. It was shown that SQGem nanoparticles strongly interacted with LDL particles and single SQGem molecules incorporated and are then transported in the circulation by LDL. This interaction was shown also in silico, in great agreement with experimental data 9. Binding of SQGem with serum albumin was also demonstrated

10

but the relative significance of these two

types of endogenous carriers is not yet understood in details. The interaction of SQGem NPs with LDL was, however, previously shown to be a dynamic process which first involved the adsorption of the small LDLs (22 nm) at the NPs surface as soon as they enter into contact 9. This event was clearly observed in TEM images where shapemodified SQGem NPs were observed in a disassembly process. The fast disassembly of the NPs when in contact with LDL was also demonstrated by FRET 26. Thus, in this study, we focused on the scenario when interaction occurs at the level of individual molecules, which were either released from nanoparticles to the serum or exchanged between nanoparticles and LDL/HSA upon their unspecific contact. Our goal was to check the feasibility of this scenario at the molecular level. In the present study, we generalize our previous observations by systematically investigating the interaction with LDL of three squalene derivatives (SQGem, SQAde and SQDox), as well as the corresponding free drugs (Gem, Ade and Dox) and the free squalene (SQ) moiety by all-atom


14

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


MD simulations. Moreover, for the first time we examined the interaction of SQGem (used as a model bioconjugate) with human serum albumin using ensemble docking simulations. We show here that all studied drugs do not accumulate in LDL particles in their free form. While Ade and Dox could bind to LDL particles transiently, no interaction of free Gem with these lipoproteins occurs. In contrast, the conjugation of the squalene moiety to these drugs drastically changes their behavior. All three squalene derivatives show pronounced propensity of accumulation in the LDL particles, which depends on the nature of the drug itself. Strongest accumulation is observed for SQAde, intermediate for SQGem and weakest for SQDox. Preferable location of squalene derivatives in the LDL particles also differs: SQAde is distributed uniformly across the particle, SQDox accumulates in the interfacial lipid layer while SQAde shows intermediate behavior. Remarkably, the effect of conjugation of the squalene moiety to the drugs is non-additive: whereas the binding of free Gem and Dox to LDL looks similar, the accumulation of their counterparts SQGem and SQDox inside the LDL particles turn to be very different by strength and localization. Pristine squalene accumulates inside LDL particles stronger than any of the drug derivatives, which suggests a simple and universal mechanism of its action. Squalene tail plays a role of nonspecific hydrophobic “anchor”, which drives the hydrophilic drug inside the LDL particle. Large volume of branched non-polar SQ tail leads to very high propensity of incorporation into the LDL hydrophobic core, which is enough to internalize even such large and complex molecule as Dox.


15


Page 16 of 28

Our data suggest that the accumulation in the LDL particles may be a universal transport mechanism for all squalene-based bioconjugates, regardless of the chemical nature of the conjugated drug. Another possible carrier of squalene derivatives in the blood is serum albumin, which is known to transport fatty acids and other hydrophobic compounds by means of dedicated hydrophobic binding pockets on its surface

25

. Serum albumin is an extremely flexible protein with a rich

repertoire of large-scale inter and intra-domain motions. This means that geometry and properties of albumin surface are highly dynamic and the static crystal structure is unlikely to provide an adequate model for binding of such complex molecules as squalene derivatives. We applied the method of ensemble docking with scanning over all solvent-accessible residues to study the binding of SQGem with HSA. The present study highlights that SQGem could bind to a surface of HSA in many different sites in a completely unspecific manner. The binding is governed by the total number of weak Van der Waals contacts of the ligand with the protein surface, which results in a stronger binding in the surface grooves and inter-domain clefts. There is no preferable binding of the squalene moiety in the dedicated binding sites of fatty acids. We suggest that branched squalene tail is too large and bulky to fit into these narrow binding pockets. Our observations do not exclude the possibility of accidental spontaneous binding of squalene moiety to the fatty acids binding sites of albumin but suggest that such events are not dominant since no precipitation was noted after incubation of SQGem with serum containing albumin. It is concluded that HSA could act as an unspecific carrier for SQGem molecules with a binding mechanism being different from one observed for the fatty acids. Currently there is no experimental data concerning the binding of other squalene bioconjugates with HSA, thus


16

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


performing docking simulations for other compounds is postponed until the appearance of appropriate experimental results.

Limitations A certain number of limitations characterize the computational methodology used in this study. First, the change from the spherical symmetry of the real LDL lipid droplet to the bilayer-like symmetry of the model system may lead to inevitable change of lipid packing in the central parts of the hydrophobic slab. Particularly, the relative volume of the central region may be overestimated and the radial packing of the lipid tails in this region is lost. This is the reason why computed PMFs should be considered as semi-quantitative in the central region. However, as the surface of our model keep the same topology of the surface of a spherical LDL lipid droplet, the computed PMFs should be considered as accurate in this region and obtained PMFs could be considered as quantitative starting from 1.5-2 nm from the center of the particle. Thus, such column-like system provides a reasonable compromise between a precise atomistic description of the LDL lipid droplet and the computational tractability. Of note, possible error introduced by the lipid packing in the central region of the particle could be especially important for such molecules as Ade because it can possibly change the PMF in this region from repulsive to slightly attractive. However, such mistake is highly unlikely because lipid packing does not change the overall hydrophobicity of the central part of the system. Even if the PMF for Ade is slightly more attractive in the center of the particle this, will not change the qualitative conclusions about the role of SQ tail, which is the main result of this study.


17


Page 18 of 28

In addition, the ApoB protein was not taken into account, since only the hydrophobic lipid core of the particle was of interest in the present study. Although there is no evidence for specific binding of squalene derivatives to ApoB protein, such possibility cannot be excluded. The influence of ApoB on lipid packing could also distort the PMFs to some extent however these effects are unlikely to change the final results qualitatively. Concerning the interaction of SQGem with albumin, although our ensemble docking simulations take into account protein flexibility and scanned its surface precisely, the inherent limitations of the docking methodology still apply. The absence of explicit water may be crucial for the docking of hydrophobic compounds, the influence of the ligands on the protein dynamics in the case of their binding in the inter-domain clefts is ignored, etc. Finally, the influence of the structure of the drug on the binding to HSA still remains questionable because only SQGem was considered in this study. Additional docking simulations of other squalene-based drugs and dedicated experiments are needed to clarify this question.

Conclusion In this study, we show convincingly that LDL particles may be considered as universal carriers of various squalenoylated drugs (such as SQGem, SQAde and SQDox) in the blood stream. Accumulation of these compounds into LDL is triggered by highly hydrophobic squalene moiety, which acts as an anchor and drives even large hydrophilic drugs into the hydrophobic interior of the LDL lipid droplet. However, accumulation propensity and localization of squalenoylated compounds inside the LDL particles strongly depends on the drug’s chemical structure.


18

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


On the other hand, human serum albumin was found to act as a passive carrier of SQGem, which binds in an unspecific manner to its surface grooves and inter-domain clefts. The binding of squalene moiety to HSA does not occur in the binding pockets for the fatty acids due to its much larger size and volume.

Supporting information The following supplementary files are available free of charge: “supplementary_figs.pdf” – the file with supplementary figures S1, S2 and S3 (PDF). “ligand_distribution.mpg” – the movie, which shows the volume density of ligand atoms around the HSA molecule (MPG). “clusters.zip” – the archive, which contains the PDB files for all best docking poses corresponding to 100 largest clusters as described in the text (ZIP).

Corresponding author * [email protected] Author Contributions PC initiated the study and formulated its goal. SY and CR developed simulation methodology. SY created molecular models of LDL and HSA and wrote analysis software. SY, CR and MS were running simulations and analyzing the results. SM and PC communicated latest experimental findings, participated in discussion and interpretation of the data. The manuscript was written by SY, CR, SM and PC. All authors have given approval to the final version of the manuscript.


19


Page 20 of 28

Funding sources S.Y. and C.R. were supported by the European Union's Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No. 690853 and the NATO Science for Peace and Security programme under the project SPS 985291. The authors declare no competing financial interests. Acknowledgements The computations were performed using HPC resources from GENCI-[TGCC/CINES/IDRIS] (Grants 2015-[c2016077586], 2016-A0020707586), Mésocentre de calcul de Franche-Comté and Centre de Calcul de Champagne-Ardenne ROMEO. References 1.

Couvreur, P.; Stella, B.; Reddy, L. H.; Hillaireau, H.; Dubernet, C.; Desmaële, D.;

Lepêtre-Mouelhi, S.; Rocco, F.; Dereuddre-Bosquet, N.; Clayette, P.; Rosilio, V.; Marsaud, V.; Renoir, J.-M.; Cattel, L. Squalenoyl Nanomedicines as Potential Therapeutics. Nano Letters 2006, 6, (11), 2544-2548. 2.

Desmaële, D.; Gref, R.; Couvreur, P. Squalenoylation: A generic platform for

nanoparticular drug delivery. Journal of Controlled Release 2012, 161, (2), 609-618. 3.

Couvreur, P.; Reddy, L. H.; Mangenot, S.; Poupaert, J. H.; Desmaële, D.; Lepêtre-

Mouelhi, S.; Pili, B.; Bourgaux, C.; Amenitsch, H.; Ollivon, M. Discovery of New Hexagonal Supramolecular Nanostructures Formed by Squalenoylation of an Anticancer Nucleoside Analogue. Small 2008, 4, (2), 247-253. 4.

Maksimenko, A.; Alami, M.; Zouhiri, F.; Brion, J.-D.; Pruvost, A.; Mougin, J.; Hamze,

A.; Boissenot, T.; Provot, O.; Desmaële, D.; Couvreur, P. Therapeutic Modalities of Squalenoyl


20

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


Nanocomposites in Colon Cancer: An Ongoing Search for Improved Efficacy. ACS Nano 2014, 8, (3), 2018-2032. 5.

Maksimenko, A.; Dosio, F.; Mougin, J.; Ferrero, A.; Wack, S.; Reddy, L. H.; Weyn, A.-

A.; Lepeltier, E.; Bourgaux, C.; Stella, B.; Cattel, L.; Couvreur, P. A unique squalenoylated and nonpegylated doxorubicin nanomedicine with systemic long-circulating properties and anticancer activity. Proceedings of the National Academy of Sciences 2014, 111, (2), E217-E226. 6.

Caron, J.; Maksimenko, A.; Wack, S.; Lepeltier, E.; Bourgaux, C.; Morvan, E.; Leblanc,

K.; Couvreur, P.; Desmaële, D. Improving the Antitumor Activity of Squalenoyl-Paclitaxel Conjugate Nanoassemblies by Manipulating the Linker between Paclitaxel and Squalene. Advanced Healthcare Materials 2013, 2, (1), 172-185. 7.

Gaudin, A.; Yemisci, M.; Eroglu, H.; Lepetre-Mouelhi, S.; Turkoglu, O. F.; Dönmez-

Demir, B.; Caban, S.; Sargon, M. F.; Garcia-Argote, S.; Pieters, G.; Loreau, O.; Rousseau, B.; Tagit, O.; Hildebrandt, N.; Le Dantec, Y.; Mougin, J.; Valetti, S.; Chacun, H.; Nicolas, V.; Desmaële, D.; Andrieux, K.; Capan, Y.; Dalkara, T.; Couvreur, P. Squalenoyl adenosine nanoparticles provide neuroprotection after stroke and spinal cord injury. Nat Nano 2014, 9, (12), 1054-1062. 8.

Sémiramoth, N.; Meo, C. D.; Zouhiri, F.; Saïd-Hassane, F.; Valetti, S.; Gorges, R.;

Nicolas, V.; Poupaert, J. H.; Chollet-Martin, S.; Desmaële, D.; Gref, R.; Couvreur, P. SelfAssembled Squalenoylated Penicillin Bioconjugates: An Original Approach for the Treatment of Intracellular Infections. ACS Nano 2012, 6, (5), 3820-3831. 9.

Sobot, D.; Mura, S.; Yesylevskyy, S. O.; Dalbin, L.; Cayre, F.; Bort, G.; Mougin, J.;

Desmaële, D.; Lepetre-Mouelhi, S.; Pieters, G.; Andreiuk, B.; Klymchenko, A. S.; Paul, J.-L.;


21


Page 22 of 28

Ramseyer, C.; Couvreur, P. Conjugation of squalene to gemcitabine as unique approach exploiting endogenous lipoproteins for drug delivery. Nature Communications 2017, 8, 15678. 10.

Bildstein, L.; Marsaud, V.; Chacun, H.; Lepetre-Mouelhi, S.; Desmaele, D.; Couvreur, P.;

Dubernet, C. Extracellular-protein-enhanced cellular uptake of squalenoyl gemcitabine from nanoassemblies. Soft Matter 2010, 6, (21), 5570-5580. 11.

Murtola, T.; Vuorela, T. A.; Hyvonen, M. T.; Marrink, S.-J.; Karttunen, M.; Vattulainen,

I. Low density lipoprotein: structure, dynamics, and interactions of apoB-100 with lipids. Soft Matter 2011, 7, (18), 8135-8141. 12.

Jämbeck, J. P. M.; Lyubartsev, A. P. Another Piece of the Membrane Puzzle: Extending

Slipids Further. Journal of Chemical Theory and Computation 2013, 9, (1), 774-784. 13.

Pluhackova, K.; Kirsch, S. A.; Han, J.; Sun, L.; Jiang, Z.; Unruh, T.; Böckmann, R. A. A

Critical Comparison of Biomembrane Force Fields: Structure and Dynamics of Model DMPC, POPC, and POPE Bilayers. The Journal of Physical Chemistry B 2016, 120, (16), 3888-3903. 14.

Sousa da Silva, A. W.; Vranken, W. F. ACPYPE - AnteChamber PYthon Parser

interfacE. BMC Research Notes 2012, 5, (1), 1-8. 15.

Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman,

J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery Jr., J. A.; Peralta, J. E.; Ogliaro, F.; Bearpark, M. J.; Heyd, J.; Brothers, E. N.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A. P.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, N. J.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.;


22

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


Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, Ö.; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09, Gaussian, Inc.: Wallingford, CT, USA, 2009. 16.

Abraham, M. J.; Murtola, T.; Schulz, R.; Páll, S.; Smith, J. C.; Hess, B.; Lindahl, E.

GROMACS: High performance molecular simulations through multi-level parallelism from laptops to supercomputers. SoftwareX 2015, 1–2, 19-25. 17.

Yesylevskyy, S. O.; Rivel, T.; Ramseyer, C. The influence of curvature on the properties

of the plasma membrane. Insights from atomistic molecular dynamics simulations. Scientific Reports 2017, 7, (1), 1. 18.

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. Journal of Computational Chemistry 1992, 13, (8), 1011-1021. 19.

Yesylevskyy, S. O.; Ramseyer, C.; Pudlo, M.; Pallandre, J.-R.; Borg, C. Selective

Inhibition of STAT3 with Respect to STAT1: Insights from Molecular Dynamics and Ensemble Docking Simulations. Journal of Chemical Information and Modeling 2016, 56, (8), 1588-1596. 20.

Pallandre, J.-R.; Borg, C.; Rognan, D.; Boibessot, T.; Luzet, V.; Yesylevskyy, S.;

Ramseyer, C.; Pudlo, M. Novel aminotetrazole derivatives as selective STAT3 non-peptide inhibitors. European Journal of Medicinal Chemistry 2015, 103, 163-174. 21.

Yesylevskyy, S. O.; Hushcha, T. O. Conformational relaxations of human serum

albumin studied by molecular dynamics simulations with pressure jumps. Biopolymers and Cell 2012, 28, (6), 486-492.


23


22.

Page 24 of 28

Trott, O.; Olson, A. J. AutoDock Vina: Improving the speed and accuracy of docking

with a new scoring function, efficient optimization, and multithreading. Journal of Computational Chemistry 2010, 31, (2), 455-461. 23.

Yesylevskyy, S. O. Pteros: Fast and easy to use open-source C++ library for molecular

analysis. Journal of Computational Chemistry 2012, 33, (19), 1632-1636. 24.

Yesylevskyy, S. O. Pteros 2.0: Evolution of the fast parallel molecular analysis library

for C++ and python. Journal of Computational Chemistry 2015, 36, (19), 1480-1488. 25.

Curry, S.; Brick, P.; Franks, N. P. Fatty acid binding to human serum albumin: new

insights from crystallographic studies. Biochimica et Biophysica Acta (BBA) - Molecular and Cell Biology of Lipids 1999, 1441, (2–3), 131-140. 26.

Cayre, F.; Mura, S.; Andreiuk, B.; Sobot, D.; Gouazou, S.; Desmaële, D.; Klymchenko,

A. S.; Couvreur, P. In Vivo FRET Imaging to Predict the Risk Associated with Hepatic Accumulation of Squalene-Based Prodrug Nanoparticles. Advanced Healthcare Materials, 1700830-n/a.

For Table of Contents Only


24

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


Figure 1. Snapshot of equilibrated LDL-like slice. POPC lipids are shown in blue, lyso PC in red, cholesterol in orange, cholesterol oleate in gray and glyceryl trioleate in green. Water is not shown for clarity. Single periodic cell is shown. 245x406mm (72 x 72 DPI)



Figure 2. Potentials of mean force of transferring individual molecules from water to the lipid core of the LDL particle. The plots are superimposed onto the snapshot of simulated LDL-like slice shown as semitransparent background and aligned with X axis. POPC lipids are shown in blue, lyso PC in red, cholesterol in orange, cholesterol oleate in gray and glyceryl trioleate in green. 248x209mm (300 x 300 DPI)


Page 26 of 28

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


Figure 3. Largest clusters of SQGem molecules bound to HSA. The protein is shown in cartoon representation. The SQGem molecules are shown as wireframe. Each distinct cluster of SQGem molecules is marked by different color. Single conformation of HSA is shown for clarity while all other conformations where aligned to it to provide correct relative positions of all docked SQGem molecules. 119x118mm (220 x 220 DPI)



Figure 4. Comparison of the preferable binding positions of SQGem (colored wireframe) and the fatty acids (blue space-fill reresentation) derived from 25. Protein structure is hidden for clarity. Each distinct cluster of SQGem molecules is marked by different color. 121x141mm (220 x 220 DPI)


Page 28 of 28

Low-Density Lipoproteins and Human Serum Albumin as Carriers of

Recommend Documents