Interactions between Cationic Lipid Bilayers and Model Chromatin

School of Biological Sciences, Nanyang Technological University, 60 Nanyang Drive, Singapore 637551. ∥ National Synchrotron ... Publication Date (We...
0 downloads 0 Views 4MB Size
Download PDF
pubs.acs.org/Langmuir © 2010 American Chemical Society

)

Interactions between Cationic Lipid Bilayers and Model Chromatin Dan Lundberg,†,‡,^ Nikolay V. Berezhnoy,§,^ Chenning Lu,§,# Nikolay Korolev,§ Chun-Jen Su, Viveka Alfredsson,‡ Maria da Grac-a Miguel,† Bj€orn Lindman,*,‡ and Lars Nordenski€old*,§

Department of Chemistry, University of Coimbra, Rua Larga, 3004-535 Coimbra, Portugal, ‡Department of Physical Chemistry, Center for Chemistry and Chemical Engineering, Lund University, P.O. Box 124, 221 00 Lund, Sweden, §School of Biological Sciences, Nanyang Technological University, 60 Nanyang Drive, Singapore 637551, and National Synchrotron Radiation Research Center, Hsinchu Science Park, Taiwan. ^ These authors contributed equally to this work. # Present address: Department of Cell Biology, Biological and Biomedical Sciences Program, Harvard Medical School, Boston, Massachusetts 02115 )

†

Received April 13, 2010. Revised Manuscript Received June 21, 2010 Complexes formed in mixtures of cationic liposomes of varying charge density and nucleosome core particles (NCPs) or nucleosome arrays have been characterized. Under most of the conditions studied, the lipids and NCPs or arrays formed lamellar structures similar to those obtained with the liposomes and pure DNA. Thus, the dissociation of DNA from the NCP or nucleosome array and the formation of a DNA-lipid complex is thermodynamically favored, which can likely be ascribed mainly to the gain in entropy on release of the small counterions. Only at very low liposome charge densities are there indications that the NCPs/arrays do not dissociate upon interaction with the lipid bilayers. The reported results can serve as a valuable reference point in investigations of biologically more relevant systems.

Introduction In the eukaryotic cell nucleus, which can be regarded as the control center of the eukaryotic cell, DNA is compacted by histone proteins into the nucleoprotein complex chromatin. As different sections of the genome are made accessible for expression over time and in different locations and cell types, the structure of chromatin is heterogeneous and highly dynamic.1 A minor fraction of the nuclear material, e.g., 3.2% of the dry weight in rat liver nuclei, consists of lipids.2 The nuclear lipids are largely ascribed a structural function as crucial components of the double-membrane nuclear envelope, which separates the nuclear environment from the surrounding cytoplasm. Over the last few decades, it has been recognized, however, that they have several additional functions. For instance, there are indications of direct membrane-chromatin interactions during mitosis.3,4 Furthermore, growing attention has been given to the nonmembrane endonuclear lipid pool. A detailed understanding of the different roles and molecular organization of this fraction of lipids is limited, but it has been found that a significant fraction of the endonuclear lipids are colocalized with decondensed chromatin domains and there are several indications of functional relationships with gene expression.2 An increased understanding of the interactions between lipids and chromatin, particularly the possible influence of nuclear lipids in controlling the structure and organization of chromatin, is motivated from a biophysical point of view. In addition, the competitive association and phase behavior in systems of lipids, DNA, and protein is of general physicochemical interest. In this work, we approach the problem by investigating the interplay *To whom correspondence should be addressed. E-mail: [email protected], [email protected]. (1) Wolffe, A. P. Chromatin: Structure and Function, 3rd ed.; Academic Press: San Diego, CA, 1998; p 447. (2) Ledeen, R. W.; Wu, G. J. Lipid Res. 2004, 45, 1–8. (3) Antonin, W.; Franz, C.; Haselmann, U.; Antony, C.; Mattaj, I. W. Mol. Cell 2005, 17, 83–92. (4) Kutay, U.; Hetzer, M. W. Curr. Opin. Cell Biol. 2008, 20, 669–677.

12488 DOI: 10.1021/la1014658

Figure 1. Illustrations of (a, b) a nucleosome core particle (NCP) and (c) the nucleosome array used in this study. (a, b) DNA atoms are shown as spheres with charged phosphate groups in color; the histone octamer is presented as schematic secondary structure, and each of the eight histones is colored differently. Approximate dimensions of the NCP are indicated. The crystal structure of the NCP (1KX57) and the Chimera software8 were used. (c) The 12-177-601 nucleosome array is schematically shown with DNA (red) wrapped around the histone octamer core (green) with histone tails shown in blue.

between model membranes in the form of liposomes of varying lipid composition with model chromatin. The chromatin models used, which are schematically depicted in Figure 1, consist of either individual nucleosome core particles (NCPs) (Figure 1a,b), which constitute the fundamental structural units of chromatin, or arrays of connected NCPs (Figure 1c); both of these are assembled from recombinant DNA and protein components and are very homogeneous and essentially monodisperse.5,6 The NCP is composed of a histone octamer protein complex, around (5) Luger, K.; Rechsteiner, T. J.; Richmond, T. J. Methods Enzymol. 1999, 304, 3–19. (6) Dorigo, B.; Schalch, T.; Bystricky, K.; Richmond, T. J. J. Mol. Biol. 2003, 327, 85–96.

Published on Web 07/01/2010

Langmuir 2010, 26(15), 12488–12492

Lundberg et al.

Letter

which DNA of 147 base pairs (bp) is coiled into a left-handed superhelix. Roughly half of the negative charges on DNA are neutralized by the net positive charge of the histones, which leaves the NCP with a net charge of approximately -150e. In chromatin, the core particles are linked by short fragments of DNA in a beads-on-a-string arrangement, often referred to as the 10 nm fiber, which is further organized into higher-level arrangements of debated structure. The nucleosome array used in this work consists of 12 nucleosomes with a DNA repeat length of 177 bp (hence comprising 30 bp of linker DNA between each NCP). Aqueous mixtures of NCPs or nucleosome arrays and multicomponent liposomes are complex systems, the behavior of which will be controlled by an intricate balance between several competing interactions. In an initial stage, it is valuable to tackle a reduced system, and the choice of lipids took some consideration. A large body of work has been performed on the characterization of complexes formed by DNA and cationic lipids (CL).9-14 DNA-CL systems constitute a good reference point, hence our initial work was performed with cationic lipid membranes; specifically, the herein-presented experiments were performed with liposomes prepared from the cationic 1,2-dioleoyl-3-trimethylammonium-propane (DOTAP) and the zwitterionic 1,2-dioleoylsn-glycero-3-phosphocholine (DOPC) in different proportions. An understanding of systems based on CL provides a good basis for the investigation of systems containing more biologically relevant lipids. It is found that mixtures of DNA and liposomes prepared from zwitterionic lipids with or without the addition of anionic lipids in the presence of divalent ions can form multilamellar aggregates with DNA that are very similar to those formed in corresponding systems with cationic lipids.15,16 In this work, results from a multitechnique investigation of the behavior of mixtures of NCPs or nucleosome arrays and DOTAP/DOPC liposomes are presented. A rather wide range of compositions, with respect to both the ratio of net anionic charges on the chromatin to the total cationic charge from the lipid mixture and the fraction of DOTAP in the liposomes, was considered.

Results and Discussion Visual Inspection for the Identification of Phase Separation. Because the NCPs and nucleosome arrays carry a high negative net charge, one can expect a strong electrostatic attraction between these species and cationic liposomes. It was anticipated that, similarly to what is observed in mixtures of DNA and cationic liposomes, there is a strong propensity for associative phase separation driven by the entropic gain from the release of the small counterions. For the herein discussed rather complex system, which includes several polyionic species (with the NCP (7) Davey, C. A.; Sargent, D. F.; Luger, K.; Maeder, A. W.; Richmond, T. J. J. Mol. Biol. 2002, 319, 1097–1113. (8) Pettersen, E. F.; Goddard, T. D.; Huang, C. C.; Couch, G. S.; Greenblatt, D. M.; Meng, E. C.; Ferrin, T. E. J. Comput. Chem. 2004, 25, 1605–1612. (9) Safinya, C. R. Curr. Opin. Struct. Biol. 2001, 11, 440–448. (10) Lasic, D. D.; Strey, H.; Stuart, M. C. A.; Podgornik, R.; Frederik, P. M. J. Am. Chem. Soc. 1997, 119, 832–833. (11) R€adler, J. O.; Koltover, I.; Salditt, T.; Safinya, C. R. Science 1997, 275, 810– 814. (12) R€adler, J.; Koltover, I.; Jamieson, A.; Salditt, T.; Safinya, C. R. Langmuir 1998, 14, 4272–4283. (13) Mel’nikov, S. M.; Sergeyev, V. G.; Mel’nikova, Y. S.; Yoshikawa, K. J. Chem. Soc., Faraday Trans. 1997, 93, 283–288. (14) Mel’nikova, Y. S.; Mel’nikov, S. M.; Lofroth, J. E. Biophys. Chem. 1999, 81, 125–141. (15) Liang, H.; Harries, D.; Wong, G. C. L. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 11173–11178. (16) McManus, J. J.; R€adler, J. O.; Dawson, K. A. Langmuir 2003, 19, 9630– 9637.

Langmuir 2010, 26(15), 12488–12492

Figure 2. Complexes of NCP and liposomes (100% DOTAP, CR = 1, 0.57 mM total charges) visualized by (a) DIC and (b-d) fluorescence microscopy, demonstrating colocalization on the optical length scale. The components are individually labeled with different fluorescent probes to allow color coding: (b) red, lipids; (c) blue, DNA; and (d) green, histone proteins. The scale bar represents 10 μm.

being constituted by DNA and 4  2 protein units) that can potentially combine in a range of different ways, it is not trivial to predict the phase behavior. The first step in the characterization of the system was accomplished by a visual inspection of samples with different compositions. Indeed, macroscopic phase separation is observed as visible solid particles down to very low (millimolar) concentrations of cationic liposomes for both NCPs and chromatin arrays. This is particularly evident in mixtures that are charge neutral with respect to the macromolecular components (i.e., samples with a charge ratio (CR) of 1 with respect to the net negative charge of the NCP/array to the net positive charge of the liposomes) and for highly charged liposomes. In unbalanced mixtures, with an excess of either NCPs/arrays or liposomes, there is at low concentrations typically the formation of stable dispersions; this is found particularly for samples with excess liposomes. Fluorescence Microscopy for the Assessment of Component Colocalization. Because the investigated system, as briefly discussed in the previous section, contains several components, it is not unlikely that the separated solid material consists of multiple phases. To estimate the extent of colocalization of the components, fluorescence and differential interference contrast (DIC) microscopy experiments were performed on NCP-based samples prepared with individually labeled DNA, histones, and liposomes. Examples of results from an experiment are shown in Figure 2. The microscopy experiments show that there is good colocalization of the components within the length-scale resolution of the technique, which is estimated to be approximately 500 nm, for samples in a wide range of compositions (5-100% DOTAP and CR = 0.5-2.0). However, it can be noted that in some samples the histones appears to be partially expelled from the complexes; the protein then appears at the surface of particles with perfect colocalization of DNA and lipid. An example is given in the Supporting Information (Figure S2). Expulsion of the histones from the NCPs is, as will be discussed below, consistent with the results from other techniques. Small-Angle X-ray Scattering for the Identification of Internal Structure. To obtain information on the inner structure of the solid phases formed, synchrotron small-angle X-ray scattering (SAXS) was applied. Solid phases were precipitated from rather concentrated solutions of liposomes and NCPs/arrays (in the ranges of 0.4-29 mM and 0.32-16 mM with respect to net DOI: 10.1021/la1014658

12489

Letter

Lundberg et al.

Figure 3. (a) SAXS spectra for samples containing NCP with CR = 1 and liposomes with different charge densities. (b) The spectrum at 50% DOTAP from part a on a magnified scale. (c) Lamellar repeat distance, L, as a function of the fraction of DOTAP in the liposomes for samples with NCP or DNA. (d) SAXS spectrum for 5% DOTAP and CR = 1 represented as I  q as a function of q. (Spectral intensities are given in arbitrary units in parts b and d.

negative charges for NCPs and nucleosome arrays, respectively), in varying ratios. Reference experiments with cationic liposomes and the corresponding DNAs were also performed. Figure 3a shows the SAXS spectra obtained for a series of NCP samples with a constant CR of 1 and a varying fraction of DOTAP in the liposomes. Results for samples with other values of CR (0.25-2.0) are very similar (data not shown). There is a clear gradual change in the appearance of the spectra with variation in liposome composition. However, for a majority of the samples one can see well-defined diffraction peaks, which reveal the formation of lamellar ordered structures; the peaks positioned close to 0.1 and 0.2 A˚-1 follow the expected relative positions of 1:2 and some spectra also show higher-order peaks. The corresponding characteristic distances are in the range of 60-70 A˚. Thus, the SAXS data on samples with NCP and liposomes with a high fraction of DOTAP are similar to those obtained for complexes formed with cationic liposomes and DNA alone.11 In such complexes, the components generally form highly ordered structures with the DNA double helices sandwiched between parallel lipid bilayers. There is often, because of the high stiffness of DNA, a high directional correlation between DNA chains, i.e., they lie practically parallel between the bilayers. Furthermore, there are clear indications from NMR experiments17 as well as computer simulations18 that in complexes prepared from net-cationic mixed vesicles, the cationic component is substantially concentrated in the vicinity of the DNA chains. Figure 3b shows one spectrum from Figure 3a (CR = 1.0, 50% DOTAP) on a more expanded scale, illustrating the lamellar Bragg peaks up to fourth order and also the expected DNA-DNA correlation peak. As illustrated in Figure 3c, a comparison between the observed characteristic distances calculated from the reflections observed in Figure 3a and those from the corresponding reference samples prepared with cationic liposomes and 147 bp DNA does indeed show that for liposomes of high charge density the lamellar repeat distances are practically identical to those found for the reference series. Also, the corresponding samples with chromatin arrays or array template DNA (12  177 bp DNA) give spectra with similar characteristics (Supporting Information, Figure S3). The findings from SAXS thus suggest that when the NCPs encounter the highly charged cationic liposomes, the histones are expelled from (17) Leal, C.; Sandstr€om, D.; Nevsten, P.; Topgaard, D. Biochim. Biophys. Acta 2008, 1778, 214–228. (18) Dias, R. S.; Pais, A. A. C. C.; Linse, P.; Miguel, M. G.; Lindman, B. J. Phys. Chem. B 2005, 109, 11781–11788.

12490 DOI: 10.1021/la1014658

the NCPs or the array and a lamellar DNA-lipid complex with a structure practically identical to that found in the absence of the histones is formed. From the SAXS results, it is not possible to make any detailed suggestions about the location of the histones. The increase in the repeat distance (decrease in q) observed with a decreasing fraction of DOTAP in the liposomes (Figure 3c) is consistent with previous findings for corresponding DNAliposome complexes.11 With liposomes of lower charge density, there is for all CR investigated, as exemplified by the data for CR = 1 in Figure 3c, a reproducible tendency for a difference in repeat distance, on the order of 5 A˚ larger, for complexes prepared with NCPs/arrays as compared to those prepared with DNA alone. These findings can be taken as an indication that a certain fraction of the histone proteins are incorporated into the lamellar DNA-lipid structure. It is also possible that ternary complexes coexist with DNA-lipid complexes and larger domains of expelled proteins (compare to the discussion of the fluorescence microscopy results above). In general, spectra of samples based on liposomes of low charge density are dominated by form factor contributions, and the peaks are broad and of low intensity. The aforementioned finding that samples with all CR investigated gives results very similar to those shown in Figure 3a-d, suggests that excess NCPs or lipids have a very limited influence on the structure of the formed complexes. This can be explained by a high cooperativity in their formation. A particularly interesting finding from the SAXS experiments is obtained with the lowest investigated liposome charge density where there is a weak but rather well-defined peak at low q. This peak is more clearly observable in Figure 3d where the uppermost spectrum from Figure 3a is plotted as I  q versus q. The position of this peak corresponds to a characteristic distance of about 160-170 A˚, which is approximately the distance expected for a situation with intact NCPs sandwiched between lipid bilayers. Importantly, this peak is not observed in the relevant reference sample with lipids and DNA only (data not shown), and its presence can be taken to suggest that intact NCPs take part in an ordered structure together with the lipid bilayers. The presence of Bragg peaks close to 0.1 and 0.2 A˚-1 reveal the coexistence of a small fraction of multilamellar complexes with DNA. The outcome of mixing NCPs/nucleosome arrays can thus be separated into two regimes: With liposomes of high charge density, the thermodynamically favorable situation is an expulsion of the histones from the initial NCP/array and the formation of complexes very similar to those formed by cationic liposomes Langmuir 2010, 26(15), 12488–12492

Lundberg et al.

Figure 4. Cryo-TEM images of aggregates formed by mixtures of NCPs, nucleosome arrays, or DNA and liposomes of different composition. (a) NCP and 50% DOTAP, CR = 0.15. (b) Array and 100% DOTAP, CR = 0.3. (c) Array and 10% DOTAP, CR = 0.3 (which corresponds to a DNA-lipid charge ratio of 0.5). (d) Array DNA and 10% DOTAP, CR = 0.5. The total lipid concentration is 0.56 mM in all samples.

and DNA alone. With low charge density, however, there are indications of an at least partial retention of NCPs/arrays. In the former case, the driving force for formation is likely dominated by the increase in entropy from an efficient release of the small counterions. A contributing driving force could also be the relief of unfavorable bending of the DNA around the histone core. In the latter case, with a lower lipid net charge, one can expect a weaker electrostatic attraction between the liposome surface and the NCPs/arrays. This could partially be ascribed to a larger entropic penalty associated with the segregation of the cationic lipid molecules (in order to concentrate these adjacent to the DNA) in lipid bilayers of lower charge density. Cryo-Transmission Electron Microscopy (Cryo-TEM). To obtain further information on the structure of the investigated complexes, cryo-TEM was used. To allow for sufficient transmission of electrons, it is necessary that the investigated particles are small. By preparing samples with “unbalanced” compositions at low concentrations, it is possible, as briefly discussed above, to obtain a dispersion of small complexes. Because the results from SAXS suggest that excess lipid or NCPs do not greatly affect the structure of the complexes, a comparison of results from the two techniques should be relevant despite the differences in sample conditions. Examples of cryo-TEM images are shown in Figure 4. There is a polydispersity in size and a variation in particle structure among the samples, but some definite trends can be identified. Samples prepared with liposomes of high charge density clearly give lamellar structures (Figure 4a,b). The multilamellar features of these images are similar to previous cryo-EM observations for CL-DNA complexes (cf. Figure S2 in Supporting Information).10,19,20 Although the conditions, as discussed above, are rather different, the findings from these images are consistent with the conclusions from the SAXS experiments; there are close similarities in repeat (19) Huebner, S.; Battersby, B. J.; Grimm, R.; Cevc, G. Biophys. J. 1999, 76, 3158–3166. (20) Weisman, S.; Hirsch-Lerner, D.; Barenholz, Y.; Talmon, Y. Biophys. J. 2004, 87, 609–614.

Langmuir 2010, 26(15), 12488–12492

Letter

distances as well as in the particle appearance for samples prepared with NCPs/arrays (Figure 4a,b) compared to those prepared with DNA only (Figure S2). Furthermore, also in agreement with SAXS data, complexes formed with liposomes containing 50% DOTAP (Figure 4a,b) generally show somewhat more “diffuse” structures than those formed with liposomes composed of DOTAP only. Additional findings from the cryoTEM experiments include the observation that complexes formed with liposomes of high charge density seem to be built up of subunits of about 100 nm size that are clustered into larger units (Figure 4a,b), as well as indications that for samples prepared with liposomes of low charge density the lamellar order increases with time (from 15 min to 24 h after sample preparation; data not shown). In some samples, one can observe regions completely lacking lamellar order. These domains could tentatively consist of expelled histones. It can be noted that in SAXS spectra obtained for dispersed samples comparable to those shown in Figure 4a,b only one peak, occurring at practically the same q as the first reflection of the corresponding spectra shown in Figure 3a, is usually observed. The absence of additional reflections can be ascribed to a low signal-to-noise ratio. Similar to the SAXS results, some of the most interesting findings from the cryo-TEM experiments are obtained with liposomes of low charge density. Figure 4c,d shows images of samples prepared with liposomes containing 10% DOTAP and a nucleosome array (Figure 4c) or DNA only (Figure 4d). With DNA only, one can clearly see that multilamellar structures with the expected interlamellar spacing are formed. However, with complete chromatin arrays or NCPs (mixed with liposomes in a DNA-lipid charge ratio close to that in the DNA-lipid reference samples; see caption of Figure 4), which give practically identical results, no multilamellar structures are formed; one can observe only a very limited tendency for liposome-liposome attachment. These observations show that, at low DOTAP content, the formation of a lipid-DNA complex is not favored, which in turn gives indirect evidence for intact or at least not completely disrupted nucleosomes. The observed clustering of liposomes (particularly the connection of liposomes via flat contact surface, as can be observed, for instance, in the lower right corner of Figure 4c) is likely mediated by dissociated DNA present in a minor amount. This finding supports the results from SAXS for samples with liposomes of low charge density.

Concluding Remarks Complexes formed in aqueous mixtures of NCPs or nucleosome arrays and cationic liposomes have been characterized. Samples prepared with liposomes containing even quite low fractions of cationic lipid form lamellar structures similar to those formed by the corresponding mixtures of cationic lipids and DNA alone. Thus, the dissociation of DNA from the NCPs/arrays and the formation of a DNA-lipid complex is thermodynamically favored. This outcome is, similar to the case of corresponding mixtures of lipids and DNA alone, likely driven by electrostatic interactions and counterion release. Importantly, however, results from SAXS (on macroscopic precipitates) and cryo-TEM (on dilute dispersed complexes) suggest that with low liposome charge densities the NCPs/arrays do not dissociate upon interaction with the lipid bilayers. From the present results, it is not possible to estimate to what extent the NCP/array structure is distorted in the mixed assemblies. The herein discussed results serve as a good basis for continuing studies, when approaching systems of more direct biological relevance. DOI: 10.1021/la1014658

12491

Letter

Acknowledgment. We are indebted to Mr. Abdollah Allahverdi for the preparation of nucleosome arrays. The plasmid constructs for the overexpression of histone proteins and 147 bp and 12  177601 bp DNA were kindly provided by Dr. Curt Davey (Nanyang Technological University, Singapore) and Prof. Timothy Richmond (ETH, Switzerland). We thank Dr. Li Hoi Yeong (Nanyang Technological University, Singapore) for assistance with microscopy measurements and Dr. Curt Davey (Nanyang Technological University, Singapore) and Dr. U-Ser Jeng (NSRRC, Hsinchu, Taiwan) for helpful discussions. This work was financially supported by the Singapore Agency for Science Technology and Research (A*STAR) (LN), an MOE ARC-Tier 2 grant (LN), the Hans Werthen Foundation, the sixth EU framework program

12492 DOI: 10.1021/la1014658

Lundberg et al.

as part of a EU-STREP project with an NEST program (NEONUCLEI, contract 12967), the Linnaeus Center of Excellence on Organizing Molecular Matter through the Swedish Research Council, and Fundac-~ao para a Ci^encia e a Tecnologia (contract number SFRH/BPD/48522/2008) (D.L.). Synchrotron X-ray scattering on solid precipitated samples was performed at the SAXS beamline at NSRRC (Hsinchu, Taiwan). Synchrotron SAXS studies of dispersed samples were made at the SAXS beamlines at EMBL (Hamburg, Germany) and Maxlab (Lund, Sweden). Supporting Information Available: Materials and methods and supporting figures. This material is available free of charge via the Internet at http://pubs.acs.org.

Langmuir 2010, 26(15), 12488–12492