Axial Growth and Fusion of Liposome Regulated by Macromolecular

Apr 15, 2015 - The images were analyzed by ImageJ software (public-domain image-processing software). .... Note that the smooth tubing determined by S...
1 downloads 0 Views 3MB Size
Download PDF
Article pubs.acs.org/Langmuir

Axial Growth and Fusion of Liposome Regulated by Macromolecular Crowding and Confinement Yun Liu,† Lin Zhu,† Jingfa Yang,‡ Jianbo Sun,† Jiang Zhao,*,‡ and Dehai Liang*,† †

Beijing National Laboratory for Molecular Sciences and the Key Laboratory of Polymer Chemistry and Physics of Ministry of Education, College of Chemistry and Molecular Engineering, Peking University, Beijing, 100871, China ‡ Beijing National Laboratory for Molecular Sciences, Institute of Chemistry, Chinese Academy of Sciences, Beijing 100190, China S Supporting Information *

ABSTRACT: The endomembrane system, including the endoplasmic reticulum, Golgi apparatus, lysosomes, and endosomes, is located in the crowded intracellular environment. An understanding of the cellular structure and functions requires knowledge of how macromolecular crowding and confinement affect the activity of membrane and its proteins. Using negatively charged liposome and the peptide K3L8K3 as a model system, we studied the aggregation behavior of liposome in a matrix of polyacrylamide and hyaluronic acid. Without matrix, the liposomes form spherical aggregates in the presence of K3L8K3. However, they orient in one dimension and fuse into a tube up to 40 μm long in the matrix. The growth of the tube is via end-to-end connection. This anisotropic growth is mainly due to the macromolecular confinement provided by the polymer network. The study of the interactions between liposome and peptide in the crowded environment helps to reveal the mechanism of membrane-related processes in vivo.

■

INTRODUCTION The densely crowded (5%−40% volume fraction) and heterogeneous medium in a living cell can modulate the kinetics and equilibrium of the reactions taking place in it.1,2 Theoretical study and computer simulations3,4 have revealed two fundamental effects: macromolecular crowding and macromolecular confinement. Both effects have been confirmed by experiments.5,6 The nature of the crowding and confinement is the excluded volume interaction, with the former favoring the compacted conformation 7 and the latter favoring the conformations having a shape complementary to that of the confining volume.8 The macromolecular crowding and confinement not only regulate the association, the denaturation, and the conformational isomerization of proteins in cell, they also serve as the driving force for the helical structure of DNA9,10 and the organizations of cells.11 Recent progress shows that the macromolecular crowding is able to change the structure of the protein containing flexible regions.12 In addition to the crowding and confinement effect, the chemical interaction between protein and the medium also accounts for the stability and dynamics of protein in a cell.13,14 Not only the biopolymers but also the endomembrane system, including the nuclear envelope, endoplasmic reticulum, Golgi apparatus, lysosomes, endosomes, and various kinds of vesicles and vacuoles, is located inside the crowded cell interior. Even the cell itself is surrounded by an extracellular matrix. Therefore, the behavior of these membrane-bound organelles and even cells should be regulated by macromolecular crowding and confinement. For example, the fission and fusion of lipid vesicles, which are responsible for the cargo delivery inside the cell,15,16 should behave differently in a crowded environment. © XXXX American Chemical Society

The function and process of the membrane are closely related to the membrane proteins, which play a key role in the transportation of water and solute through the cell membrane, as well as in signal transduction and cellular responses.17 Therefore, the study of the membrane and its protein under crowded environment should shed light on the mechanism of membrane-related processes in vivo. It should also help to develop drug (gene) delivery vehicles, because cell uptake and endosome escape are two main hurdles in the delivery pathway.18 The effect of macromolecular crowding and confinement on the behavior of membrane (together with protein) is complex within living cells. Herein, we chose a liposome and peptide as the model system and studied their behavior in a crowded environment. The liposome is composed of 1,2-dipalmitoyl-snglycero-3-phosphocholine (DPPC) and 1,2-dipalmitoyl-snglycero-3-phospho-(1′-rac-glycerol) (DPPG). DPPC and DPPG have the same melting temperature of 41 °C,19 and they are miscible at the studied temperature range from 25 to 50 °C. The charge density of the liposome can be tuned by varying the molar ratio of DPPC/DPPG. A bola-typed cationic peptide K3L8K3 is employed to interact with the liposome. In aqueous solution, K3L8K3 is able to deform the membrane and cause aggregation, fusion, and even leakage of the liposome.20 Polyacrylamide (PAM) with Mw = 5.1 × 106 g/mol is used as the crowding agent. The radius gyration (Rg) of PAM is 101 Received: December 2, 2014 Revised: March 7, 2015

A

DOI: 10.1021/la504699y Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir nm. The overlap concentration C* is calculated to be 2.0 × 10−3 g/mL according to the following equation:

■

C* = M w /[NA(4π /3)R g 3]

containing 1 mol % of both probes is mixed with blank liposome (no probes) at a mole ratio of 1:9. The time point of mixing is set as t0. The fusion or semifusion of liposomes results in an increase of fluorescence intensity It. The degree of fusion is calculated according to

(1)

EXPERIMENTAL SECTION

Ft (%) = 100 ×

Materials. DPPC (≥99%) was purchased from Sigma. DPPG (sodium salt), 1,2-dipalmitoyl-sn-glycero-3-phosphoethanolamine-N(7-nitro-2-1,3-benzoxadiazol-4-yl) (NBD−DPPE, ammonium salt), and 1,2-dipalmitoyl-sn-glycero-3-phosphoethanolamine-N-(lissamine rhodamine B sulfonyl) (Rh−DPPE, ammonium salt) in chloroform were purchased from Avanti Polar Lipids, Inc. The peptide K3L8K3 (Ac-KKKL LLLL LLLK KK-amide) (purity ≥98%) was purchased from GL Biochem Ltd. (Shanghai, China) and used as received. A stock solution of K3L8K3 at 1.00 mg/mL was prepared by dissolving the corresponding peptide powder in Milli-Q water, followed by sonication for 5−10 s. Liposome Preparation. A large unilamellar vesicle (LUV) containing DPPC and DPPG was prepared by the extrusion method. In brief, DPPC and DPPG at the desired molar ratio, together with the fluorescence-labeled lipids Rh−DPPE or NBD−DPPE at known concentration, were dissolved in chloroform and dried by a rotary evaporator in a round-bottom flask to obtain a thin film on its wall. The solvent was then removed under reduced pressure. The dry film was hydrated with water. The final concentration of liposome was 1.0 mM. The liposome solution was extruded through a polycarbonate membrane (pore size 100 nm) by using an Avanti miniextruder (Avanti Polar Lipids, Inc.) 21 times to obtain the LUV. Preparation of Liposome and Polymer Complex System. Polyacrylamide (PAM, Mw = 5.1 × 106) at 4% or hyaluronic acid (HA, Mw = 1.3 × 106) at 2% was prepared by dissolving the powdered sample in water. A known amount of liposome solution was then added into the polymer solution, followed by dilution to the desired concentrations. The liposome and polymer complex system was mixed thoroughly and kept at 4 °C before use. Aggregation of Liposome in Polymer Solution. The aggregation of liposome in polymer solution was observed by total internal reflection fluorescence microscopy (TIRFM). The glass surface was pretreated by Sigmacote (Sigma) to prevent sample adsorption. The excitation laser beam (Melles Griot, 532 nm) was introduced to the sample surface from the bottom of the substrate by an oil-immersion objective lens (100×, numerical aperture = 1.45). The laser beam was totally reflected at the surface by tuning the incident angle. The fluorescence of liposome in polymer solution was excited by the evanescent wave above the surface. The fluorescence images were recorded by an electron-multiplying charge-coupled device camera (Andor DV887 EMCCD) with the exposure time of 0.1s. The premixed polymer solution with liposome was kept at 50 °C for 10 min before the addition of K3L8K3. To record the entire process of aggregation, a movie was shot at different times. The images were analyzed by ImageJ software (public-domain image-processing software). Scanning Electron Microscopy (SEM). The polymer solution with liposome was first heated to 50 °C to reach equilibrium, followed by the addition of K3L8K3. The sample was then kept at 50 °C for 1 h before dipping into liquid nitrogen to freeze the morphology. After being dried in a lyophilizer, the sample was cut and the fresh surface was studied by SEM. Transmission Electron Microscopy (TEM). TEM was conducted on a JEM-100CX transmission electron microscope. Fifteen microliters of the sample was dropped on a copper grid covered with a carbon film support (T10023, Beijing Xinxing Braim Technology C., Ltd.). After 30 s, the excess solution was removed by filter paper from the edge of the film. The sample was then kept for at least 24 h to ensure that it was completely dried. Membrane Fusion. The Förster resonance energy transfer (FRET) between probes of NBD−DPPE and Rh−DPPE was applied to determine the degree of membrane fusion (or semifusion). The detailed procedure can be found elsewhere.20 In brief, the liposome

It − I 0 I∞ − I0

(2)

with I0 being the residual fluorescence intensity of the mixture and I∞ being the maximum fluorescence intensity as determined by using the liposome containing 0.1 mol % of each probe.

■

RESULTS AND DISCUSSION The liposome of DPPC/DPPG is prepared by the extrusion method, and the diameter is about 160 nm at 50 °C (Figure S1, Supporting Information). The liposome is labeled with 0.1% Rhodamine B, and its movement in aqueous solution can be monitored by TIRFM. The liposome itself exhibits simple diffusion behavior (Figures S2 and S3, Supporting Information) in PAM at concentrations ranging from 0.5% to 1.5%, above which the movement is barely visible. No aggregation of liposome is observed at the studied conditions. However, the situation is quite different in the presence of K3L8K3. The liposomes form aggregates with K3L8K3, and the morphology of the aggregates is dependent on the PAM matrix. Without PAM, the liposomes of DPPC/DPPG (1:1) form spherical aggregates in the presence of K3L8K3 in 40 min at 50 °C (Figure 1A). The

Figure 1. Snapshots of the DPPC/DPPG = 1:1 liposome after being mixed with K3L8K3 for 40 min: (A) 50 °C with no PAM, (B) 50 °C in 1.0% PAM, (C) 25 °C in 1.0% PAM, and (D) 50 °C in 1.0% PAM with 150 mM NaCl. The dimension of each panel is 50 × 50 μm. The total concentration of lipids is 20 μM, and the peptide to lipid molar ratio is 1.56.

spherical shape is driven by surface energy, since the complex formed by liposome and peptide is hydrophobic after charge neutralization. Interestingly, the liposome and peptide assemble into a fibrous structure in 1.0% PAM (Figure 1B) under the same conditions. The morphology of the assembly in polymer matrix is also dependent on temperature. No fibrous but a spherical structure is observed as the temperature decreases to 25 °C (Figure 1C). Since the liposome of DPPC/DPPG is in the liquid phase at 50 °C but in the solid phase at 25 °C, the formation of different morphologies indicates that the phase of liposome also plays certain role during the assembly of liposome in the crowded environment. Since both liposome B

DOI: 10.1021/la504699y Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir and peptide are charged and contain hydrophobic moieties, the assembly of liposome in the presence of K3L8K3 is driven by multiple interactions, including electrostatic interaction, hydrophobic interaction, and possibly hydrogen bonding. This process is similar to the complexation of oppositely charged polyelectrolytes, which is sensitive to ionic strength.21 As expected, adding 150 mM NaCl in 1.0% PAM results in a significant increase in the length of the fibrous structure, with the value reaching about 40 μm (Figure 1D). We attribute it to the enhanced aggregation of liposome with K3L8K3 due to the screening of the long-range electrostatic repulsions.21 The growth of the fiber at the early stage is closely monitored by TIRFM. One typical growth pattern is shown in Figure 2. A

Figure 3. Effect of PAM concentration (A) and the content of DPPG (B) on the growth of liposome in the presence of K3L8K3. Panels C and D compare the fusion curves of liposome in pure water and in 1.0% PAM. All the experiments are conducted at 50 °C. The peptide to lipid molar ratio is 1.56.

Clearly, polymer matrix is crucial for the liposomes to form fibers in the presence of K3L8K3. However, the liposomes do not form fibers at 25 °C (Figure 1C). The property of the liposome itself, therefore, should also play a certain role. We prepared liposomes with different DPPC/DPPG ratios and compared their behaviors in 1.0% PAM at 50 °C. As shown in Figure 3B, the growth of the fiber is the strongest for the liposome containing equal amounts of DPPC and DPPG. Since the lipid is heavily demixed in the presence of K3L8K3 under such conditions,20 the fiber formed by liposomes should be related to the phase separation of lipids in the presence of peptide. Note that the fiber is not formed by K3L8K3 itself. On one hand, K3L8K3 stays as individual random coils in aqueous solution.20 Its behavior should be similar in PAM solution in that PAM is hydrophilic and its volume content is only about 1.0%. On the other hand, we have tested the interaction of liposome with peptides possessing higher charge density or capacity to assemble in aqueous solution (by varying the sequence or length), but none of them can induce the liposome to assemble into a fiber at the studied conditions (Figure S6, Supporting Information). Our previous study indicated that K3L8K3 not only caused an aggregation of liposome but also a fusion of membrane in aqueous solution.20 To evaluate the effect of macromolecular crowding and confinement on the fusion process, we studied the fusion activity of the DPPG/DPPC (1:1) liposome with and without PAM by measuring the FRET between two fluorescence probes, NBD−DPPE and Rh−DPPE.23 Parts C and D of Figure 3 compare the fusion curves with and without 1.0% PAM. Clearly, the macromolecular crowding and confinement significantly enhanced the fusion process. Without PAM, the fusion is very fast. It reaches about 50% within 5 s and becomes stable with time (Figure 3C), while in 1.0% PAM the first data point already jumps to more than 50% and gradually increases with time (Figure 3D). The value reaches about 220% at the plateau, which is much larger than unity. This is possible because the 100% value is calibrated by using separately prepared liposome containing 0.1 mol % of each

Figure 2. Snapshot of the DPPC/DPPG = 1:1 liposome in 1.0% PAM at 50 °C. The peptide to lipid molar ratio is 1.56. The time points are marked in the corresponding panel. The dimension of each panel is 20 × 20 μm.

spherical particle A and a fibrous structure B are getting close to each other. Particle A is an individual liposome, as determined from the diffusion coefficient (1.7 × 10−13 m2/s in 1.0% PAM). It connects to fiber B via attachment at one end. A movie of the aggregation process is provided in the Supporting Information (Movie 1). If a particle hits the side of the fiber, no connection occurs (Figure S4, Supporting Information). The growth via end-to-end connection ensures the formation of a fiber instead of a branched structure. The macromolecular crowding and confinement effect is generally related to the concentration of the crowding agents.3 Figure 3A compares the time dependence of the fiber length (denoted as length/diameter ratio) at different PAM concentrations. The snapshots at different time points are included in the Supporting Information (Figure S5). The fiber growth rate (6/min) in 0.5% PAM is about 9 times faster than that (0.7/min) in 1.0% PAM. The length of the fiber after reaching equilibrium is also twice as long in 0.5% PAM. The pore size (ξ) of the network provide by PAM at concentration C can be calculated by ξ = Rg(C/C*)−0.76.22 Results show that the pore size in 0.5% PAM is 50 nm, while it decreases to 30 nm in 1.0% PAM. The pore size in both cases is smaller than the diameter of the liposome (160 nm). However, the network of PAM is flexible. It is able to tune the pore size to accommodate the diffusion and aggregation of liposome in the presence of K3L8K3. Clearly, the pore size at 0.5% PAM is closer to the size of the liposome, leading to a faster diffusion rate and a large aggregate after reaching equilibrium. C

DOI: 10.1021/la504699y Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

Figure 4. SEM images of liposome in 1.0% PAM after 50 min: (A) without K3L8K3 and (B and C) with K3L8K3. Temperature: 50 °C. The peptide to lipid molar ratio is 1.56.

individual liposome. Semifusion of membrane at this stage is also possible. If this is the case, the fiber is like a “necklace”, which transforms into a smooth tubing during the freeze-drying process in SEM or during the absorption process in TEM. The macromolecular confinement accounts for the axial growth via an end-to-end connection. According to Doi and Edwards,24 the entangled polymer chains in concentrated solutions confine the movement of an individual chain in a “tube”. Once two spherical liposomes bind together, the anisotropic structure will experiences the tube effect. It orients along the axis of the tube, which also facilitates the fusion or semifusion of the liposomes. The tube model also explains the growth of the fiber via end-toend connection. Only a liposome entering from the end of the tube can merge into the existing fiber. The macromolecular crowding provided by the polymer matrix is able to enhance the aggregation of the liposomes. However, the enhancement may not be prominent, since the concentration of the polymer matrix is only 1%, far lower than that (5%−40%) in living medium.

probe and then mixed with 1.0% PAM. The fusion process in 1.0% PAM is also accompanied by phase separation (Figure 3B), which could enlarge the distance between the two probes, enhancing the fluorescence intensity. To determine the detailed structure of the fibers formed by liposomes in PAM, the sample was freeze-dried, and the freshly peeled surfaces were studied by SEM. As shown in Figure 4A, only individual liposome is observed if no K3L8K3 is presented. They are evenly distributed in the matrix. However, the liposomes align and fuse into a long tubing in the presence of K3L8K3 (Figure 4B,C). The length of the tubing can reach about 10 μm. Note that the smooth tubing determined by SEM is in the dry state. The morphology may be distorted during the sample preparation process. To confirm the formation of tubing by liposome in polymer matrix, we further studied the behavior of liposome with K3L8K3 in 0.05% hyaluronic acid (HA). The anisotropic growth is also observed by TIRFM (Figure S7, Supporting Information). Since negatively charged HA has a low affinity for the carbon grid, TEM is a good choice to determine the detailed structure assembled by liposome and peptide. As shown in Figure 5A,B, tubes of varying length are observed in

■

CONCLUSIONS In conclusion, the macromolecular crowding and confinement regulate not only the kinetics and the equilibrium of the reactions involving liposomes but also the structure and morphology formed by liposomes. The regulation is also affected by the properties of the liposome as well as those of the participating reagents, such as protein or peptide. Our findings help to reveal the life processes in vivo. Since membranes in living cells experience the crowding environment from both inside and outside, the membrane activities, such as endocytosis, budding, and fission, might also be regulated by the heterogeneous crowding media.

■

ASSOCIATED CONTENT

* Supporting Information S

Size distribution of liposome, extra snapshots of liposome in PAM and HA, names and sequences of the other peptides, and a movie showing the aggregation of liposomes. This material is available free of charge via the Internet at http://pubs.acs.org.

Figure 5. TEM images of liposome with K3L8K3 in 0.05% hyaluronic acid after 50 min. The peptide to lipid molar ratio is 1.2. (A and B) Views at different spots. Panel C schematically shows the orientation and fusion of liposomes.

■

these images. Combining the results on temperature effect, phase separation, kinetics of fusion, and TEM and SEM together, a cartoon (Figure 5D) is drawn to schematically show the mechanism of axial growth. K3L8K3 has the capacity to induce not only an aggregation of liposome but also a demixing of DPPC and DPPG if the temperature is above the melting point.20 The formation of domains rich in either DPPC or DPPG is facilitated when the DPPC/DPPG molar ratio is close to unity. A fusion at the connecting point results in the formation of tubing with a diameter close to that of an

AUTHOR INFORMATION

Corresponding Authors

*J.Z. email: [email protected]. *D.L. e-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The financial support from the National Natural Science Foundation of China (21174007) and the National Basic D

DOI: 10.1021/la504699y Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

(23) Struck, D. K.; Hoekstra, D.; Pagano, R. E. Use of Resonance Energy-Transfer To Monitor Membrane-Fusion. Biochemistry 1981, 20 (14), 4093−4099. (24) Doi, M.; Edwards, S. F. The Theory of Polymer Dynamics; Oxford University Press: Oxford, UK, 1986.

Research Program of China (973 Program, 2012CB821500) is greatly acknowledged.

■

REFERENCES

(1) Ellis, R. J.; Minton, A. P. Cell BiologyJoin the Crowd. Nature 2003, 425 (6953), 27−28. (2) Medalia, O.; Weber, I.; Frangakis, A. S.; Nicastro, D.; Gerisch, G.; Baumeister, W. Macromolecular Architecture in Eukaryotic Cells Visualized by Cryoelectron Tomography. Science 2002, 298 (5596), 1209−1213. (3) Zhou, H.-X.; Rivas, G.; Minton, A. P. Macromolecular Crowding and Confinement: Biochemical, Biophysical, and Potential Physiological Consequences. Annu. Rev. Biophys. 2008, 37, 375−397. (4) Elcock, A. H. Models of Macromolecular Crowding Effects and the Need for Quantitative Comparisons with Experiment. Curr. Opin. Struct. Biol. 2010, 20 (2), 196−206. (5) Snoussi, K.; Halle, B. Protein Self-Association Induced by Macromolecular Crowding: A Quantitative Analysis by Magnetic Relaxation Dispersion. Biophys. J. 2005, 88 (4), 2855−2866. (6) Tang, Y. C.; Chang, H. C.; Roeben, A.; Wischnewski, D.; Wischnewski, N.; Kerner, M. J.; Hartl, F. U.; Hayer-Hartl, M. Structural Features of the GroEL-GroES Nano-Cage Required for Rapid Folding of Encapsulated Protein. Cell 2006, 125 (5), 903−914. (7) Minton, A. P. Excluded Volume as a Determinant of Macromolecular Structure and Reactivity. Biopolymers 1981, 20 (10), 2093−2120. (8) Minton, A. P. Confinement as a Determinant of Macromolecular Structure and Reactivity. Biophys. J. 1992, 63 (4), 1090−1100. (9) Nakano, S.-i.; Miyoshi, D.; Sugimoto, N. Effects of Molecular Crowding on the Structures, Interactions, and Functions of Nucleic Acids. Chem. Rev. 2014, 114 (5), 2733−2758. (10) Snir, Y.; Kamien, R. D. Entropically Driven Helix Formation. Science 2005, 307 (5712), 1067−1067. (11) Marenduzzo, D.; Finan, K.; Cook, P. R. The Depletion Attraction: An Underappreciated Force Driving Cellular Organization. J. Cell Biol. 2006, 175 (5), 681−686. (12) Akabayov, S. R.; Akabayov, B.; Richardson, C. C.; Wagner, G. Molecular Crowding Enhanced ATPase Activity of the RNA Helicase eIF4A Correlates with Compaction of Its Quaternary Structure and Association with eIF4G. J. Am. Chem. Soc. 2013, 135 (27), 10040− 10047. (13) Wang, Y.; Sarkar, M.; Smith, A. E.; Krois, A. S.; Pielak, G. J. Macromolecular Crowding and Protein Stability. J. Am. Chem. Soc. 2012, 134 (40), 16614−16618. (14) Monteith, W. B.; Pielak, G. J. Residue Level Quantification of Protein Stability in Living Cells. Proc. Natl. Acad. Sci. U. S. A. 2014, 111 (31), 11335−11340. (15) Pucadyil, T. J.; Schmid, S. L. Conserved Functions of Membrane Active GTPases in Coated Vesicle Formation. Science 2009, 325 (5945), 1217−1220. (16) Bonifacino, J. S.; Glick, B. S. The Mechanisms of Vesicle Budding and Fusion. Cell 2004, 116 (2), 153−166. (17) Beevers, A. J.; Dixon, A. M. Helical Membrane Peptides to Modulate Cell Function. Chem. Soc. Rev. 2010, 39 (6), 2146−2157. (18) Aliabadi, H. M.; Landry, B.; Sun, C.; Tang, T.; Uludag, H. Supramolecular Assemblies in Functional siRNA Delivery: Where Do We Stand? Biomaterials 2012, 33 (8), 2546−2569. (19) Cevc, G. Phospholipids Handbook; Marcel Dekker, Inc.: New York, 1993. (20) Xia, Y.; Sun, J.; Liang, D. Aggregation, Fusion, and Leakage of Liposomes Induced by Peptides. Langmuir 2014, 30 (25), 7334−7342. (21) Zheng, C.; Niu, L.; Pan, W.; Zhou, J.; Lv, H.; Cheng, J.; Liang, D. Long-Term Kinetics of DNA Interacting with Polycations. Polymer 2014, 55 (10), 2464−2471. (22) Kohli, I.; Mukhopadhyay, A. Diffusion of Nanoparticles in Semidilute Polymer Solutions: Effect of Different Length Scales. Macromolecules 2012, 45 (15), 6143−6149. E

DOI: 10.1021/la504699y Langmuir XXXX, XXX, XXX−XXX