Maturation of Nanodiscs to HDL-like Particles via Lecithin-
Cholesterol Acyl Transferase
Chinonso C. Opara
Abstract High Density Lipoprotein (HDL) particles are spherical structures, which play a fundamental role in the reverse cholesterol transport (RCT) pathway, where cholesterol is carried to the liver for excretion. Apolipoprotein (Apo) A-1 is the primary protein component of HDL and plays a critical role in the RCT pathway. As lipid and cholesterol are transported from cells, they are encircled by Apo A-1, forming discoidal pre--HDL particles, which subsequently undergo cholesterol-dependent maturation via lecithin-cholesterol acyl transferase (LCAT), forming spherical HDL. Man-made HDL have many applications in nanotechnology, including intracellular delivery of drug cargo. However, the large size distribution of HDL made from full length Apo A-1 impedes drug delivery precision. We use a shorter construct of Apo A-1, as a membrane scaffold protein for nanodiscs which resemble pre--HDL particles, but differ in their more uniform size range. We hypothesize that treatment of nanodiscs containing cholesterol with LCAT will result in spherical HDL-like particles with a more narrow size window. Current data suggests that LCAT-mediated maturation of cholesterol-containing nanodiscs yields HDL-like particles with a narrow size distribution. The insight gained from this study serves to establish a baseline for using matured nanodiscs as drug delivery vehicles with a uniform size range.
Introduction High density lipoprotein (HDL) particles are known as the good cholesterol because they aid in the excretion of cholesterol from the body via the reverse cholesterol transport (RCT) pathway. The RCT pathway begins with the cellular expulsion of cholesterol and fatty acids, which are subsequently encircled by apolipoprotein (Apo) A-1 the primary protein component of HDL. This forms discoidal pre--HDL particles, which mature into spherical HDL. This is accomplished by the enzyme, lecithin-cholesterol acyl transferase (LCAT), which esterifies cholesterol to fatty acids within the discoidal pre--HDL particles, resulting in a more hydrophobic lipid assembly, leading to sphere formation. Subsequently, mature HDL are taken up by the liver, resulting in cholesterol excretion by the bile 1 . While HDL is utilized in the body for transporting cholesterol, researchers are currently developing them as a drug delivery platform by exploiting the hydrophobic core for encapsulating drug cargo 2,3 . However, man-made HDL constructed from full length Apo A-1 results in a population of particles with a large size window 4 . This is problematic for drug delivery, as it generates uncertainty in dosage (Figure 1).
Figure 1. HDL particles of different sizes carrying varying amounts of drug cargo.
Conversely, by using a shorter construct (MSP1D1) of Apo A-1 as a membrane scaffold protein (MSP), nanodiscs which resemble discoidal pre--HDL particles can be made, via a procedure that mimics the first stage of the RCT pathway (Figure 2). This results in a solution of particles with a narrow size distribution 5 , compared with nascent and man-made discoidal pre--HDL made from full length Apo A-1 6,7 .
1 Groen et al., The Ins and Outs of Reverse Cholesterol Transport. 2 Murakami et al., Intracellular Drug Delivery by Genetically Engineered High- Density Lipoprotein Nanoparticles. 3 Zhang and Chen, Recombinant High Density Lipoprotein Reconstituted with Apolipoprotein AI Cysteine Mutants as Delivery Vehicles for 10- Hydroxycamptothecin. 4 Shahzad et al., Targeted Delivery of Small Interfering RNA Using Reconstituted High-Density Lipoprotein Nanoparticles. 5 Denisov et al., Directed Self-Assembly of Monodisperse Phospholipid Bilayer Nanodiscs with Controlled Size. 6 Vedhachalam et al., Influence of Apolipoprotein (Apo) A-I Structure on Nascent High Density Lipoprotein (HDL) Particle Size Distribution. 7 Scanu et al., Degradation and Reassembly of a Human Serum High-Density Lipoprotein. Evidence for Differences in Lipid Affinity among Three Classes of Polypeptide Chains.
Figure 2. Nanodisc preparation: membrane scaffold protein combines with lipids dissolved in detergent to form mixed micelles. Nanodiscs are formed upon removal of the detergent.
Subsequent maturation of cholesterol-containing nanodiscs has shown to yield spherical HDL-like particles, with three or more nanodiscs combining to form two or more HDL-like particles 8 .
Figure 3. HDL-like particle preparation: LCAT esterifies cholesterol to fatty acids in the nanodiscs. Three or more nanodiscs combine to form two or more HDL-like particles.
Nanodisc maturation may also result in HDL-like particles with a uniform size range. In addition, with further manipulation of the MSP and lipid, the HDL-like particles could be tuned to locate specific cell types for targeted drug delivery.
8 Silva et al., Structure of Apolipoprotein A-I in Spherical High Density Lipoproteins of Different Sizes. Results and Discussion Effect of Cholesterol and LCAT Concentration on Formation of HDL- like Particles Nanodisc formulations containing varying amounts of cholesterol (0%, 5%, 7.5%, 10%, or 12.5% - mole) were incubated without or with LCAT (2.5 mg/mL or 5 mg/mL). Size exclusion chromatography (SEC) of the incubation mixtures suggests HDL-like particle formation is proportional to nanodisc cholesterol content and LCAT concentration (Figure 4). The peaks around 19.5 minutes are residual lipid aggregates resulting from formation of the nanodiscs, which elute around 25 minutes. The peaks around 23 minutes become more pronounced as nanodisc cholesterol content and LCAT concentration are increased. The biggest difference in broadness of the 23 minute peak is between the incubation mixture containing 0% cholesterol nanodiscs and no LCAT (top left panel) and the incubation mixture containing 12.5% cholesterol nanodiscs and 5 mg/mL LCAT (bottom right panel). This indicates that the HDL-like particles elute around 23 minutes and that their formation is augmented with increasing nanodisc cholesterol content and LCAT concentration.
Figure 4. SEC traces of the incubation mixtures. Nanodiscs elute around 25 minutes. HDL-like particles are represented by the shoulder around 23 minutes, directly before the nanodiscs. HDL-like particle formation increases with increasing LCAT concentration and nanodisc cholesterol content, as is shown by the broadening of the 23 minute peak.
However, HPLC traces containing no LCAT also possess a slight peak around 23 minutes. Rather than being HDL-like particles, these peaks are potentially lipid aggregates with diameters close to that of the HDL-like particles. The former would counter the mechanism of nanodisc maturation, where LCAT esterifies cholesterol to fatty acids within the nanodiscs. Yet, the lipid aggregates eluting around 19.8 minutes do not increase in the presence of LCAT. This proposes that the lipid aggregates eluting around 23 minutes also do not intensify when LCAT is added, suggesting that the broadening of the 23 minute peak is accomplished by way of maturation of the nanodiscs to HDL-like particles via LCAT.
Resulting Size Distributions from LCAT Incubation Mixtures In order to assess the size of the nanodiscs, peak III corresponding to the nanodiscs from the incubation mixture containing the 0% cholesterol nanodisc formulation and no LCAT was collected via SEC and subjected to static light scattering (SLS) (Figure 5). The nanodiscs displayed a polydispersity of 8.6% around 8.5 nm, which is similar to the literature value of ~9.7 nm 9 .
Figure 5. Static light scattering size distribution histogram of peak III from the incubation mixture containing the 0% cholesterol nanodisc formulation and no LCAT. The nanodiscs display a polydispersity of 8.6% around 8.5 nm.
9 Denisov et al., Directed Self-Assembly of Monodisperse Phospholipid Bilayer Nanodiscs with Controlled Size. From the incubation mixture containing the 12.5% cholesterol nanodisc formulation and 5 mg/mL LCAT, peaks II and III corresponding to the nanodiscs and HDL-like particles, respectively were collected through SEC and analyzed by SLS (Figure 6). Because the representative peaks of nanodiscs and HDL-like particles were collected, in the histogram below, two size distributions were expected: a population of nanodiscs around 8.5 nm and a population of HDL-like particles with a slightly larger diameter than the nanodiscs. Instead size distributions around 75 nm and 260 nm were observed. This is potentially due to confounding of the data by larger particles in the solution. In the future, this analysis should be repeated for the purpose of obtaining an accurate diameter for the HDL-like particles.
Figure 6. Static light scattering size distribution histogram of peaks II and III from the incubation mixture containing the 12.5% cholesterol nanodisc formulation and 5 mg/mL LCAT. Two distributions, one for nanodiscs and another for HDL-like particles, around 10 nm were expected. Rather, distributions around 75 nm and 260 nm were observed, possibly due to confounding of the data by larger particles in solution.
Visual Comparison of Nanodiscs and HDL-Like Particles Electron microscopy (EM) was used to visualize the nanodiscs and HDL-like particles. Peak I from the incubation mixture containing the 0% cholesterol nanodisc formulation and no LCAT representing nanodiscs (Figure 7) and peaks II and III from the incubation mixture containing the 12.5% cholesterol formulation and 5 mg/mL LCAT (Figure 8) representing both nanodiscs and HDL-like particles, respectively were selected for EM analysis. When the nanodiscs are viewed from their sides, they look like two adjacent rods with a dark band (MSP) in between them. This is consistent with their structure, as two MSPs are wrapped around the hydrophobic core. When viewed from the top, the dark band disappears, while the nanodiscs remain rod-like in appearance.
Figure 7. Electron microscopy image of peak III (purified for nanodiscs) from the incubation mixture containing the 0% cholesterol nanodisc formulation and no LCAT. When viewed from their sides (within red circle), the nanodiscs appear rod-like with a dark band, representing the membrane scaffold protein, through their middle. When viewed from the top (within blue circle), the dark bands are no longer visible and the nanodiscs still appear rod-like. Size bar = 100 nm.
From the EM, the HDL-like particles appear to have a narrow size distribution, which will have to be confirmed by SLS. They also appear more circular than the nanodiscs, which is also consistent with their spherical shape. However, from the 2D image, it is not possible to fully discern the HDL-like particles from the nanodiscs. A 3D reconstruction of the EM image will be needed in order to provide more spatial detail in regards to the structural differences between the two particles.
Figure 8. Electron microscopy image of peaks II and III (purified for HDL-like particles and nanodiscs) from the incubation mixture containing the 12.5% cholesterol formulation and 5 mg/mL LCAT. The HDL-like particles (within yellow circle) are similar in manifestation to the top view of the nanodiscs, but appear more circular, from the 2D image, than the nanodiscs. Size bar = 100 nm.
The data presented from this work propose that discoidal nanodiscs can be converted to spherical HDL-like particles with a narrow size distribution via LCAT. HDL-like particle formation is proportional to nanodisc cholesterol content and LCAT concentration. In the future, adjustment of nanodisc cholesterol content will be needed in order to further optimize HDL-like particle formation. This work establishes a baseline for utilizing matured nanodiscs as a platform for intracellular drug delivery.
Materials and Methods Lipids Lipids were purchased in powder form and dissolved in 100% chloroform. 1-Palmitoyl-2-oleoylphosphatidylcholine (POPC) was purchased from Avanti Polar Lipids and cholesterol was purchased from Sigma. Lipid films were prepared by dispensing the desired amount of lipid using a Hamilton glass syringe into glass culture tubes. Excess solvent was removed via nitrogen stream followed by overnight vacuum desiccation. Lipids containing varying amounts of cholesterol (POPC + cholesterol 0%, 5%, 7.5%, 10%, and 12.5% - mole) were made by adding appropriate amounts of cholesterol dissolved in chloroform into POPC dissolved in chloroform. The resulting lipid film was obtained, as described above.
POPC Phosphate Assay POPC concentration was assessed via phosphate assay. 1 and 2 L aliquots of POPC stock solution dissolved in chloroform were dispensed at the bottom of glass culture tubes. Phosphate standards (0 to 228 mM) were treated likewise. Phosphate moieties were cleaved in a fume hood above 200 O C in 8.9 N sulfuric acid. Subsequently, hydrogen peroxide was added to completely oxidize the remaining organic material. Samples were heated again above 200 O C to remove remaining peroxide. Next, the phosphorous-molybdate complex was formed through addition of 2.5% ammonium molybdate and 10% ascorbic acid and subsequent heating at 100 O C to allow the blue hue to results. When cooled to room temperature, the samples were transferred to a clear flat- bottom microplate and the absorbance was read at 820 nm.
MSP1D1 Expression The expression vectors were purchased from Addgene and the protein was expressed in BL21 Star (DE3) cells (Invitrogen) using the pET expression system. A starting culture was inoculated from freshly streaked plates into 10 mL of Luria Broth (LB) with 30 mg/L kanamycin and allowed to grow for 8-12 hours at 37 O C and 220 rpm. The starting cultures was then diluted into 100 mL of LB with kanamycin and allowed to grow for another 8-12 hours. The cultures were subsequently added to 900 mL LB with kanamycin in 2.8 L fernbach flasks. One liter cultures were grown for 2-3 hours, or until the OD was 0.6. Then, the cultures were induced with Isopropyl -D-1-thiogalactopyranoside (IPTG, Sigma) at a final concentration of 1mM. Before being harvested by centrifugation at 8000xg for 10 minutes, the cells were then allowed to grow for 2.5-3 more hours. Prior to purification, cell pellets were stored wet at -80 O C.
MSP1D1 Purification MSP1D1s were purified via the N-terminal poly-histidine utilizing a 20 mL column of Ni-NTA fast-flow resin (Qiagen). Prior to each round of purification, the column was regenerated, stripping the nickel with 50 mM EDTA, washing with water and 20% ethanol, and then recharging with 100 mM NiBr 2 . The cell pellet from 6 L was thawed on ice prior to resuspension in 50 mL of 20 mM phosphate buffer, 1 mM phenylmethlysulfonyl fluoride (PSMF) and 5 g/mL of deoxyribonuclease I, pH 7.4. The cells were lysed by sonication 30 second bursts, followed by 30 seconds on ice, for a total of six minutes. Subsequently, the lysate was clarified by centrifugation at 30,000xg for 30 minutes. The clarified lysate was loaded on the column through gravity and the flow-through was reloaded to obtain maximal binding. Next, the column was washed using 5 column volumes each of wash 1 (40 mM tris, 300 mM NaCl, 1% Triton X-100, pH 8.0), wash 2 (40 mM Tris, 300 mM NaCl, 50 mM Na cholate, 20 mM imidazole, pH 8.0) and wash 3 (40 mM tris, 300 mM NaCl, 50 mM imidazole, pH 8.0) before elution with elution buffer (40 mM Tris, 300 mM NaCl, 400 mM imidazole, pH 8.0). Five column volumes (100 mL) were then collected in 10 mL aliquots. The elution was pooled and dialyzed against DFB containing 0.5 mM EDTA. The concentration was ascertained via UV-vis absorbance at 280 nm and the protein was concentrated to a final concentration between 150-170 M, aliquoted, and stored at -20 O C. Purity was determined by SDS-PAGE.
Nanodisc Preparation The nanodiscs were made by resuspending lipid films, containing varying amounts of cholesterol (0%, 5%, 7.5%, 10%, and 12.5% - mole) in disc forming buffer (DFB, 20 mM Tris, 100 mM NaCl, pH 7.4) and cholate. The solutions were heated at 37 O C to assist solvation, and then repeatedly sonicated and vortexed until they became clear. MSP1D1 was added at a POPC:MSP1D1 ratio of 65:1. The final concentration of cholate in the incubation was 20mM, which is 2-fold greater than the critical micelle concentration, in order to ensure complete solubilization of the components. The solution was then allowed to incubate at 4 O C with constant agitation for 1 hour. Detergent removal was accomplished by addition of prewashed Amberlite XAD-2 at 0.5 g/mL with constant agitation at 4 O C for 4 hours. The resulting nanodisc solutions were removed from the Amberlite XAD-2 utilizing a 25 gauge needle stored at 20 O C.
LCAT Preparation The mammalian LCAT construct (CHO-hLCATH6) was expressed and purified by Michael J. Dabrowski in our lab. The Chinese Hamster Ovary cells were maintained in DMEM F12 media with 10% heat-inactivated fetal bovine serum, 0.01% pen-strep, 2mM glutamine and vitamins. Upon attaining 95% competency, the cells were washed in Hanks solution and incubated for 72 hours in PFX-CHO serum-free media with the above mentioned supplements. 10mM sodium butyrate was added after 24 hours to boost expression. After 48 hours, the media was collected and centrifuged to remove cell debris. The supernatant, containing the expressed LCAT, was subsequently loaded onto a His60 column for purification.
LCAT Reactions LCAT reactions were performed by incubating 2 M nanodiscs containing varying amounts of cholesterol (0%, 5%, 7.5%, 10%, or 12.5% - mole), 5 mM -mercapthoethanol, and LCAT at two different concentrations (2.5 mg/mL or 5 mg/mL). After 4 hours, the reaction mixtures were stored at 4 O C.
Size Exclusion Chromatography Size exclusion chromatography (SEC) was performed on a calibrated Superdex 200 10/300 column, at an isocratic flow rate of 0.5 mL/min using DFB as the mobile phase.
Static Light Scattering Static light scattering was performed on a DynoPro NanoStar dynamic light scattering instrument from Wyatt Technology. For each sample 30 acquisitions lasting 10 seconds were taken, producing an averaged spectrum. Samples were prepared by collecting fractions eluting at 23.3-27.8 minutes (peak III from the 0% cholesterol nanodisc formulation containing no LCAT) and 21.3-28.3 minutes (peaks II and III from the 12.5% cholesterol nanodiscs formulation containing 5 mg/mL LCAT) from the SEC column and concentrated using micro spin- columns.
Electron Microscopy Electron microscopy was performed by James Williams and Jamie Ebner at the University of Washington Electron Microscopy Suite in the Department of Biochemistry. Samples were prepared by collecting fractions eluting at 23.3-27.28 (peak III from the 0% cholesterol nanodisc formulation containing no LCAT) minutes and 21.3-28.3 (peaks II and III from the 12.5% cholesterol nanodiscs formulation containing 5 mg/mL LCAT) minutes from the SEC column and concentrated using micro spin-columns. Samples were diluted 1 to 20 and then stored at 4 O C prior to spotting on freshly glow-discharged carbon grids. Subsequently, the grids were stained with Nano-W (Nanoprobes, provided by Dr. Kelly Lee in the Department of Medicinal Chemistry) and allowed to dry for at least 1 hour before EM analysis.
References Denisov, I. G., Y. V. Grinkova, A. A. Lazarides, and S. G. Sligar. Directed Self-Assembly of Monodisperse Phospholipid Bilayer Nanodiscs with Controlled Size. Journal of the American Chemical Society 126, no. 11 (March 1, 2004): 347787. doi:10.1021/ja0393574. Groen, AlbertK, RonaldPJ Oude Elferink, HenkjanJ Verkade, and Folkert Kuipers. The Ins and Outs of Reverse Cholesterol Transport. Annals of Medicine 36, no. 2 (January 1, 2004): 135 45. doi:10.1080/07853890310020635. Murakami, Tatsuya, Wassana Wijagkanalan, Mitsuru Hashida, and Kunihiro Tsuchida. Intracellular Drug Delivery by Genetically Engineered High-Density Lipoprotein Nanoparticles. Nanomedicine 5, no. 6 (August 1, 2010): 86779. doi:10.2217/nnm.10.66. Scanu, A, E Cump, J Toth, S Koga, E Stiller, and L Albers. Degradation and Reassembly of a Human Serum High-Density Lipoprotein. Evidence for Differences in Lipid Affinity among Three Classes of Polypeptide Chains. Biochemistry 9, no. 6 (March 17, 1970): 132735. Shahzad, Mian MK, Lingegowda S Mangala, Hee Dong Han, Chunhua Lu, Justin Bottsford-Miller, Masato Nishimura, Edna M Mora, et al. Targeted Delivery of Small Interfering RNA Using Reconstituted High-Density Lipoprotein Nanoparticles. Neoplasia (New York, N.Y.) 13, no. 4 (April 2011): 30919. Silva, R. A. Gangani D., Rong Huang, Jamie Morris, Jianwen Fang, Elena O. Gracheva, Gang Ren, Anatol Kontush, W. Gray Jerome, Kerry-Anne Rye, and W. Sean Davidson. Structure of Apolipoprotein A-I in Spherical High Density Lipoproteins of Different Sizes. Proceedings of the National Academy of Sciences 105, no. 34 (August 26, 2008): 1217681. doi:10.1073/pnas.0803626105. Vedhachalam, Charulatha, Palaniappan Sevugan Chetty, Margaret Nickel, Padmaja Dhanasekaran, Sissel Lund-Katz, George H. Rothblat, and Michael C. Phillips. Influence of Apolipoprotein (Apo) A-I Structure on Nascent High Density Lipoprotein (HDL) Particle Size Distribution. Journal of Biological Chemistry 285, no. 42 (October 15, 2010): 3196573. doi:10.1074/jbc.M110.126292. Zhang, Xinbo, and Baosheng Chen. Recombinant High Density Lipoprotein Reconstituted with Apolipoprotein AI Cysteine Mutants as Delivery Vehicles for 10-Hydroxycamptothecin. Cancer Letters 298, no. 1 (December 1, 2010): 2633. doi:10.1016/j.canlet.2010.05.023.
Acknowledgments I would like to thank my faculty mentor, William Atkins in the Department of Medicinal Chemistry, for his advising and for his provision of supplies and equipment for the project. I would also like to give thanks to Michael Dabrowski and Wynton McClary from the Atkins lab for their guidance in making nanodiscs, performing necessary assays and for providing the LCAT. I would in addition like to thank James Williams and Jamie Ebner from the Kelly Lee lab, also in the Department of Medicinal Chemistry, for their support in collecting the electron microscopy images. Also, I would like to thank the University of Washington Ronald E. McNair Program for allowing me the opportunity to publish in this journal.
Chinonso C. Opara Medicinal Chemistry, William Atkins chinonsoopara@gmail.com
My research interests center around drug delivery. I am attending Vanderbilt Medical School to receive an M.D.