| Collection Mémoires et thèses électroniques |
|
| Accueil À propos Nous joindre |
| Chapter III Purification and characterization of retrovirus vector particles by rate zonal ultracentrifugation | ||
|---|---|---|
|
|
|
Table des matières
María de las Mercedes Segura 1,2, Alain Garnier 2, and Amine Kamen 1*
1- Biotechnology Research Institute, NRC, 6100 Royalmount Avenue, Montreal, Quebec, Canada, H4P 2R2; phone: 514-496-2264 ; fax: 514-496-7251; e-mail: amine.kamen@cnrc-nrc.gc.ca
2- Department of Chemical Engineering, Centre de Recherche sur la Fonction, la Structure et l’Ingénierie des Protéines, Université Laval, Québec, Canada, G1K 7P4
Accepted by the Journal of Virological Methods, December 6th 2005
L’ultracentrifugation jusqu'à équilibre sur gradient de sucrose demeure la technique la plus largement utilisée pour la purification des rétrovirus. Cependant, les préparations de virus purifiés obtenues avec cette méthode standard contiennent habituellement des quantités considérables de vésicules membranaires contaminantes. De plus, les solutions de sucrose sont très visqueuses et hyperosmotiques compromettant l’intégrité et la fonctionnalité des particules rétrovirales. Dans le but de surmonter ces problèmes, une technique alternative de purification a été développée, utilisant l’ultracentrifugation zonale transitoire et l’iodixanol pour générer un gradient. Les particules rétrovirales recombinantes ont été produites en utilisant des cellules 293-GPG en suspension en présence de 10% FBS. Les surnageants concentrés ont été purifiés par sédimentation transitoire dans un gradient d’iodixanol continu de 10 à 30%. Les particules virales localisées dans les fractions médianes du gradient étaient intactes et actives. En utilisant cette stratégie, des niveaux élevés de purification ont été obtenus, sans contamination avec des vésicules membranaires cellulaires comme indiqué par les études de traitement à la subtilisine. Le niveau de pureté des préparations rétrovirales est plus grand que 95% comme le montre l’analyse par électrophorèse sur gel de polyacrylamide en présence de SDS et par chromatographie sur tamis moléculaire. Les particules purifiées ont des dimensions et des formes homogènes selon les résultats de microscopie électronique à coloration négative. De plus, de grandes quantités de particules rétrovirales défectueuses produites dans les cellules 293-GPG peuvent être séparées des particules rétrovirales fonctionnelles en utilisant cette stratégie de purification.
Sucrose equilibrium density ultracentrifugation remains the most widely used technique for retrovirus purification. However, purified virus preparations obtained by this routine method usually contain considerable amounts of contaminating cell membrane vesicles. In addition, sucrose solutions are highly viscous and hyperosmotic which jeopardizes the integrity and functionality of the retrovirus particle. In order to overcome these limitations, an alternative purification technique using rate zonal ultracentrifugation and iodixanol as gradient medium was developed. Recombinant retrovirus particles were produced by 293-GPG packaging cells grown in suspension in the presence of 10% FBS. Concentrated supernatants were purified by rate zonal sedimentation on a 10-30% continuous iodixanol gradient. Virus particles were recovered intact and active from the central fractions of the gradient. By using this strategy, high levels of purification were achieved, with no evident contamination with cell membrane vesicles as indicated by subtilisin treatment studies. The level of purity of the retrovirus preparation is over 95% as shown by SDS-PAGE analysis and size-exclusion chromatography. Purified particles appear homogenous in size and morphology according to negative stain electron microscopy. In addition, large amounts of defective retrovirus particles produced by 293-GPG packaging cells can be separated from functional retrovirus particles using this purification strategy.
Highly purified virus preparations are required for characterization studies, immunological studies, gene transfer purposes and as “gold standards” for downstream processing. These preparations are difficult to obtain, particularly in cases where viruses occur at low titers and the viral particles are unstable, which is the case for retroviruses.
Mature retrovirus particles are composed of the cleavage products of three precursor polyproteins: Gag, Gag-Pol and Env-protein. Retrovirus Gag structural proteins alone represent about ¾ of the total virion protein content. The Env-proteins make up the vast majority of the remaining protein mass while the viral enzymes, derived from proteolytic cleavage of the Gag-Pol fusion polyprotein, are represented to a lesser extent. Additionally, even after successive rounds of equilibrium density centrifugation, purified virus preparations show several faint bands of cellular polypeptides detectable by SDS-PAGE. While some of these cellular proteins are an integral part of the virus (Ott, 1997; Ott, 2002), others are associated with cell membrane vesicles from broken or intact cells that have a density similar to that of the virions and thus co-purify with them by equilibrium (isopycnic) density centrifugation (Bess et al., 1997; Gluschankof et al., 1997).
Centrifugation processes are widely used for the isolation of viral particles. Typically, retrovirus particles are first separated from the bulk of contaminating serum proteins present in the growth medium by high speed centrifugation. The resulting pellet is resuspended in a small volume of buffer allowing simultaneous purification and concentration of the virions. However, this method lacks resolving capacity and is usually coupled with density gradient ultracentrifugation. The most widely used gradient media for virus purification are sucrose and cesium chloride (CsCl). Both media are hyperosmotic at the densities used to band retrovirus particles. Sucrose solutions are very viscous and thus require longer sedimentation times for separation. Moreover, the high viscosity of sucrose has been associated with loss of surface structures and thus loss of infectivity upon purification. Iodixanol (OptiPrepTM) offers several advantages over these two media. Iodixanol can be diluted in iso-osmotic buffers to form iso-osmotic solutions that help preserve retrovirus particle integrity and functionality (Dettenhofer and Yu, 1999; Moller-Larsen and Christensen, 1998). In addition, it is less viscous than sucrose resulting in shorter processing times. Moreover, this medium, originally designed as an X-ray contrast solution, is non-toxic to cells and allows for subsequent viral infectivity assays directly without need for removal.
Large amounts of contaminating host protein-laden membrane vesicles, either microvesicles or exosomes, are found in density-gradient purified virus preparations. Between 2 to 4 fold more cell membrane vesicles than virions were found by electron microscopy in HIV-1 purified preparations from lymphoid cells (Bess et al., 1997; Gluschankof et al., 1997). Previous studies have also shown that a significant amount of VSV-G vesicles are released by 293 cells expressing VSV-G into the culture medium (Abe et al., 1998). Complete removal of contaminating cell membrane vesicles is difficult to accomplish since these particles show important similarities in morphology, composition and physical characteristics with the virions. A possible way to remove these cellular vesicles is to employ immunoaffinity chromatography provided that a surface protein is found to be exclusively incorporated into either the virions or the vesicles. Taking advantage of the differential incorporation of CD45 into HIV-1 and cell membrane vesicles (Esser et al., 2001), Trubey et al. (2003) developed an immunoaffinity approach to deplete selectively membrane vesicles from density-purified retrovirus preparations. However, non-hematopoietic cells (i.e. HEK 293) are not expected to express CD45, limiting the usefulness of this technique. The separation of defective and functional virus particles poses an even more serious challenge. Although well documented for other types of viruses and viral vectors, a method that allows the separation of defective from functional retrovirus particles has not yet been described in the literature.
To date most highly purified retrovirus preparations used in retrovirus characterization studies were obtained by equilibrium ultracentrifugation on sucrose density gradients. Using this technique, retrovirus particles are isolated from a band at a density of ~1.16 g/mL, corresponding to 35% (w/w) sucrose. However, the large number of cell membrane vesicles that co-purify with the virus when this standard technique is used greatly complicates the identification and quantification of cellular proteins incorporated into retrovirus particles and it is undesirable for gene therapy applications. The goal of this work was to develop a centrifugation procedure that yielded highly purified active retrovirus preparations free of contaminating cell membrane vesicles and thus, suitable for the analysis of the exterior virus particle proteins. Since these vesicles show a wider range of size (50-500nm), higher levels of purification could be achieved by rate zonal ultracentrifugation. In this case, separation of viral particles from contaminants is based on size and density, in contrast to the standard equilibrium ultracentrifugation procedure in which irrespective of the size, particles are separated according to their buoyant density alone. Iodixanol was selected as gradient medium because of the advantages it offers over sucrose mentioned above. This iodinated medium showed to be effective for the purification of Moloney murine leukaemia virus (MoMLV) Gag retrovirus-like particles (Hammarstedt et al., 2000) and other retroviruses, including HIV-1 and HTLV-1 (Dettenhofer and Yu, 1999; Moller-Larsen and Christensen, 1998).
A novel, efficient and reproducible method for the purification of functional MoMLV-derived retrovirus particles that employs a combination of membrane filtration and rate zonal ultracentrifugation is described in this work. This strategy resulted in highly purified virus preparations with no evident contamination with cell membrane vesicles and proved to separate defective from functional retrovirus particles.
2.1 Retrovirus vector, cell lines and virus stock preparatio n
The retrovirus used in this study is a MoMLV-derived vesicular stomatitis virus glycoprotein (VSV-G) pseudotyped retrovirus vector produced by 293-GPG packaging cells (Ory et al., 1996). This cell line, a generous gift from Dr. J. Galipeau (Lady Davis Institute for Medical Research, Montreal, QC, Canada), was stably transfected to generate a retrovirus vector encoding a fusion protein between the herpes simplex virus thymidine kinase protein (TK) and the green fluorescent protein (GFP) (Paquin et al., 2001). Suspension-adapted cultures were grown in calcium free DMEM medium supplemented with 10% foetal bovine serum (FBS; HyClone, Logan, UT) and tetracycline (1 µg/mL; Fisher Scientific, Nepean, ON, Canada) until the cell density reached 2×106 cells/mL, at which point VSV-G expression was induced by entirely removing tetracycline from the culture medium. Retrovirus supernatants were harvested every 24 h during 7 days as previously described (Segura et al., 2005). Infective particles were detectable at day 3 post-induction and the maximum peak of virus production was detected on day 5 post-induction (1×107 IVP/mL). The quality of harvested supernatants during vector production varies in terms of the viral titer as well as DNA concentration. The latter was found to increase continuously during production (Segura et al., 2005). Since the purification performance can be affected by DNA viscosity, a representative feed generated by pooling equal volumes of supernatants harvested from day 4 to day 7 post-induction was used for the purification experiments. Sequential microfiltration and ultrafiltration steps were selected for clarification and concentration of virions from crude supernatants (Figure 14). This strategy resulted in concentrated and partially purified virus stocks with minimal losses of infective particles as previously described (Segura et al., 2005). Briefly, supernatants were filtered using a dual HT Tuffryn® polysulfone membrane (0.45/0.2 µm) capsule filtration device to remove any cells and cell debris. Clarified permeates were subjected to ultra/diafiltration processes using an OmegaTM polyethersulfone membrane disc filter (MWCO = 300 kDa, Pall Life Sciences, Mississauga, ON, Canada) and a 2 L stirred cell ultrafiltration unit (Amicon® 2000; Millipore, Etobicoke, ON, Canada). Retrovirus supernatants (1 L) were concentrated 20-fold under constant nitrogen pressure (30 psi) and tip speed (33.5 cm/sec). Retrovirus-enriched retentate was diafiltered against cold storage buffer (150 mM NaCl in 20 mM Tris-HCl buffer, pH 7.5). The diafiltration step was repeated 3 times with 100 mL buffer in discontinuous mode. Virus stocks were aliquoted and stored at -80ºC prior to ultracentrifugation. A removal of 30% of the total proteins and 65% of small molecular weight DNA fragments was achieved during the ultrafiltration step (Segura et al., 2005). Target 143B cells used for virus titration were graciously provided by M. Caruso (Centre de recherche en cancérologie de l’Université Laval, Québec, QC, Canada). These cells were maintained in DMEM (GIBCO-BRL, Gaithersburg, MD) medium supplemented with 10% FBS at 37°C.
2.2 Retrovirus vector titer determination
Quantification of infective particles by flow cytometric analysis has been previously reported (Segura et al., 2005). Briefly, 143B target cells in 6-well plates were exposed to 1 mL aliquots of serial dilutions of virus in DMEM containing 8 µg/mL of polybrene during 3 h at 37ºC. After the addition of 1 mL of DMEM containing 20% FBS, cells were further incubated for 48 h at 37°C under 5% CO2 atmosphere. Transduced cells were washed with PBS, detached with trypsin-EDTA, fixed with 2% formaldehyde and resuspended in 1 mL of PBS. Samples were then subjected to FACS analysis and viral titers were calculated as previously described (Segura et al., 2005).
2.3 Protein and DNA analysis
Total protein concentration was determined by the Bradford Protein Assay (Bio-Rad Laboratories, Hercules, CA) according to the manufacturer’s instructions using Bovine Serum Albumin (BSA) as the standard. Cellular double-stranded DNA was detected and quantified using the PicoGreen® dsDNA Quantitation Kit (Molecular Probes, Eugene, OR). The microassay was performed according to the manufacturer’s instructions using lambda DNA as the standard. Before electrophoresis, virus samples were mixed 3:1 with 4x NuPage® sample buffer (Invitrogen Life Technologies, Burlington, ON, Canada) containing 50 mM DTT and heated at 70°C for 10 min. Proteins were fractionated by electrophoresis on Novex® 4-12% Tris-Glycine gradient gels (InvitrogenTM) or 4-20% Tris-HCl Ready Gels (Bio-Rad) and run under reducing conditions (SDS-PAGE). Mark12 MW standard (InvitrogenTM) was applied to each gel to determine molecular weights. Protein bands were visualized by silver staining (Silver Stain Plus Kit, Bio-Rad) or by staining with Coomassie brilliant blue R250. For Western blot, proteins were transferred to a Protran® nitrocellulose membrane (Schleicher & Schuell Bioscience Inc., Keene, NH) using a Mini Trans-Blot® transfer cell (Bio-Rad). Viral proteins bound to the membrane were identified by indirect immunostaining using a monoclonal antibody against the C-terminus of VSV-G envelope protein (mouse Mab anti-VSV-G; Roche Diagnostics, Indianapolis, IN), a polyclonal antibody raised against wild type MoMLV (goat antiserum to MoMLV; ATCC VR-1522AS-Gt) or a polyclonal antibody for host proteins (goat anti-HEK 293 cell HRP conjugate, Cygnus Technologies Inc, Southport, NC). Primary antibodies were incubated overnight at room temperature. Appropriate secondary antibodies coupled to horseradish peroxidase (Jackson Immunochemicals, Mississauga, ON, Canada) allowed visualization by chemiluminescence using the BM Chemiluminescence Blotting Substrate (Roche Diagnostics, Indianapolis, IN). Images were obtained with a Kodak Digital Science Image Station 440cf equipped with Kodak Digital Science 1D image analysis software version 3.0 (Eastman Kodak Co., Rochester, NY).
2.4 Rate zonal ultracentrifugation
A 10-30% continuous iodixanol gradient was made in a 25 × 89 mm Beckman UltraClearTM tube (Beckman Instruments, Inc., Palo Alto, CA) using a two-chamber gradient maker. For this purpose, two solutions, 10% and 30% iodixanol (w/v) in 20 mM Tris-HCl, 1 mM EDTA, 0.85 w/v NaCl (pH 7.5) were prepared using a 60% (w/v) sterile stock solution of OptiPrepTM (AXIS-SHIELD PoC AS, Oslo, Norway). Retrovirus particles were isolated by carefully layering 3 mL of the concentrated virus stock on top of a 34 mL continuous iodixanol gradient (Figure 14). Ultracentrifugation was carried out in a Beckman ultracentrifuge using a SW28 rotor and spinning at 100,000×g for 4 h at 4ºC. Fifteen fractions of 2.5 mL (corresponding to a bandwidth of 5 mm) were collected by puncturing the bottom of the tube. The density of each fraction was determined by weight of 200 μL aliquots on an analytical balance. Gradient fractions were immediately titrated or kept at -80°C for subsequent analysis. Concentration of a pool of virus-containing fractions (6 to 9) was required for virus detection in certain experiments as specified. Concentration was carried out by ultrafiltration using a 10 mL stirred cell unit and a YM-100 kDa membrane at 10 psi and low tip speed (20 cm/sec). Retentates were diafiltered against PBS and the final retentate volume was adjusted to achieve a final concentration of 10-fold.
2.5 Size-exclusion chromatography
Size-exclusion chromatography was used to remove iodixanol and evaluate the purity of the virus preparation (Figure 14). Sepharose CL-4B gel (Amersham Biosciences, Piscataway, NJ) was packed into a XK 16/40 glass column (1.6 cm diameter x 30 cm bed height). A 10-fold concentrated pool of iodixanol purified virus-containing fractions was passed through a 0.45 µm GHP Acrodisc® Syringe Filter (Pall Life Sciences) and 2.5 mL of feed were loaded onto the PBS equilibrated column. Size-exclusion chromatography was performed at room temperature using a low-pressure liquid chromatography system (GradiFrac; Amersham Biosciences, Uppsala, Sweden) and monitoring protein elution by UV absorbance at 280 nm. Elution was performed with PBS at a linear flow rate of 24 cm/h and 1.5 mL fractions were collected.
2.6 Enzymatic treatments
The purity of virions was further evaluated by the use of enzymatic treatments. Benzonase® and subtilisin digestions were employed to rule out the presence of contaminating nucleic acids and cell membrane vesicles respectively. Digestion with Benzonase® of a 10-fold concentrated pool of iodixanol purified virus-containing fractions was performed by adding 30 µL per mL of sample of a working solution containing 10,000 U/mL of the endonuclease in 20 mM Tris-HCl, 1 mM MgCl2.6H2O, 20 mM NaCl (pH 8) buffer. The sample was incubated at room temperature with gentle shaking for 30 min and filtered using 0.45 µm GHP Acrodisc® Syringe Filter (Pall Life Sciences) before being loaded onto the size exclusion chromatography column. Subtilisin treatment of a 10-fold concentrated pool of iodixanol purified virus-containing fractions was carried out as previously described (Ott et al., 1995) using a 5X stock solution containing 25 mg/mL of subtilisin (Fluka Biochemika, Buchs, Switzerland) in 100 mM Tris-HCl, 5 mM CaCl2 (pH 8) for a final concentration of 5 mg of protease/mL. The reaction was stopped by the addition of 5 μg/mL phenylmethylsulfonyl fluoride (Fluka BioChemika). Digested retrovirus preparations were then subjected to dead end filtration and size exclusion chromatography as previously described in section 2.5. Chromatographic and electrophoretic profiles of treated and untreated samples were then compared.
2.7 Electron microscopy
Retrovirus morphology, size and total virus particle counts were determined by negative stain electron microscopy (NSEM). Particles in suspension were mixed with a known concentration of latex beads, mounted onto grids, negatively stained with uranyl acetate and examined in a transmission electron microscope at the Armand-Frappier Institute (Laval, QC, Canada) as previously described (Alain, 1997; Alain et al., 1987).
Harvested supernatants from 293 GPG cell cultures were pooled (day 4 to day 7 post-induction) and subjected to a series of downstream processing steps. Microfiltration and ultrafiltration were selected for clarification and concentration of crude supernatants. Retrovirus particles were isolated by carefully layering 3 mL of the 20-fold concentrated virus stock on top of a 10-30% continuous iodixanol gradient and spinning at 100,000×g for 4 h at 4ºC. Size-exclusion chromatography was further introduced to remove iodixanol and evaluate the purity of the pool of virus-containing fractions.
3.1 Purification of retrovirus particles by rate zonal ultracentrifugation
Although seldom used for the purification of retrovirus particles, rate zonal ultracentrifugation seemed to be a promising tool for the separation of the virions from closely related particles. For the method to be efficient, retrovirus particles must be layered on top of a continuous density gradient as a narrow band and allowed to move down through the gradient, but not reach equilibrium. In this way, virions can be separated from particles with the same or very similar densities but different size such as some cell membrane vesicles.
For 3 out of 11 runs performed, 15 gradient fractions per run were collected and thoroughly analyzed (Figure 15A). The density gradient formed was continuous from 1.00 to 1.16 g/mL (Figure 15B). The virus banded approximately halfway down the gradient (~1.08 g/mL). According to FACS analysis, viral activity was detected in the 4 central fractions (6 to 9) of the 15 fractions collected. The recovery of infective viral particles in these 4 virus-containing fractions was 36.5 ± 3% (n=3). Only 0.1% of the total proteins originally loaded on top of the gradient were recovered in these fractions. Retrovirus particles were well separated from nucleic acids as well as the bulk of contaminating serum proteins that remained at the top of the tube (Figure 15B). A low amount of infective viral particles remained at the top of the centrifugation tube, possibly associated with protein aggregates.
(A) Gradient at the end of the ultracentrifugation run. The virus was collected by dripping 2.5 mL fractions from the bottom of the centrifugation tube. Infective virus particles were consistently detected in fractions 6 to 9. (B) Analysis of the 15 gradient fractions collected. The percentage of GFP + cells upon virus titration (fraction dilution 1/10), protein concentration (μg/mL), DNA concentration (μg/mL) and density (g/mL) from 3 runs was determined. Values represented are the mean ± standard deviation of these 3 runs.
Electrophoretic analyses confirmed the presence of the virus in fractions 6 to 9 (Figure 16). Bands corresponding to the major viral proteins, including VSV-G (68 kDa), capsid protein (p30CA), matrix protein (p15MA) and a ~10 kDa polypeptide migrating with the dye front (p12 MoMLV and the nucleocapsid protein p10NC), are visible in the 4 central lanes of the silver stained gel (Figure 16A). This result was confirmed by Western blot analyses using anti-virus antibodies (Figure 16B). A very faint band corresponding to the Gag polyprotein (65 kDa) was detected using a polyclonal antibody raised against wild type MoMLV (Figure 16B). The low levels of this protein detected in the purified preparation indicate that the vast majority of particles produced are mature virions. A ~23 kDa band was also recognized by this polyclonal antibody. The intensity of the viral bands increased continuously from fractions 6 to 9, suggesting that most viral particles banded on the top fraction. The isolation of retrovirus particles in these 4 central fractions of the gradient was consistent as demonstrated by SDS-PAGE in the 11 centrifugation runs (data not shown).
Most serum proteins remained at the top of the centrifugation tube in fractions 10 to 15. Cross-reactivity of the polyclonal antibody with BSA and other serum proteins in the top fractions was observed in this and past experiments (Segura et al., 2005). Furthermore, separation of particles with a higher sedimentation coefficient than the virus in the bottom fractions (1 to 5) was observed. These particles show a protein profile that somewhat differs from the profile of the virus.
A) SDS-PAGE. Gradient fractions (1 to 15) were analyzed on 4-20% gradient polyacrylamide silver stained gels (Bio-Rad). Fractions 1 to 9 were loaded undiluted (25 uL/lane). The protein concentration in fractions 10-15 was adjusted to load 250 ng of protein per lane. B) Western Blot. Fractions were fractionated as described in Figure 16A and transferred to a Protran® nitrocellulose membrane. Viral proteins bound to the membrane were analyzed by indirect immunostaining using a polyclonal anti-MoMLV antibody or a monoclonal anti-VSV-G antibody. Virus-containing fraction lanes are delimited by solid black squares (
). All major viral proteins were observed in fractions 6 to 9.
3.2 Size-exclusion chromatography
Size-exclusion chromatography, typically used as a polishing step for retrovirus particles (Segura et al., 2005; Transfiguracion et al., 2003), was used in this work as an analytical tool to evaluate the purity of the virus-containing fractions and remove iodixanol. Iodixanol strongly absorbs at 244 nm. Since the virus has a very high molecular weight (~216,000 kDa) it was excluded from the gel pores thereby eluting in the void volume (Vo), while iodixanol molecules were retained inside the gel pores and eluted later. A single virus peak eluting at the expected retention volume (Vo = 23 mL) was observed (Figure 17). Iodixanol molecules (retention volume, Vr = 54 mL) were well separated from the virus peak. Similar chromatographic profiles were obtained for the virus samples treated with Benzonase® and subtilisin prior to chromatography (data not shown). There was no significant difference in the peak height after these enzymatic treatments indicating that the preparations are free of nucleic acid and cell membrane vesicle contamination.
Due to iodixanol interference at 280 nm, purity determination directly from the chromatogram could not be performed. Instead, retrovirus purity in a 10-fold concentrated pool of virus-containing fractions was evaluated by comparing electrophoretic profiles and band intensities before and after size exclusion chromatography (Figure 18). The Gag protein bands are clearly visible by Coomassie blue staining (Figure 18A). This gel shows 2 prominent bands, one corresponding to the capsid protein, p30CA, and another one that contains the remaining Gag proteins in a single non-resolved band (p15MA, p12 and p10NC) migrating with the dye front at the bottom of the gel (Figure 18A). A faint ~68 kDa band corresponding to the VSV-G envelope protein is also visible. These viral-encoded proteins alone account for >90% of the total protein content as assessed by visual inspection. No significant differences in the electrophoretic profile of the virus samples before and after size exclusion chromatography were detected in this gel.
Silver staining revealed the presence of several additional bands in the pool of purified virus fractions (Figure 18A). It should be noted that some protein bands, including the VSV-G band and a ~23 kDa band, were disproportionately more stained by silver than Coomassie blue. The electrophoretic patterns obtained before and after size exclusion chromatography remained identical using this sensitive staining method. However, the intensity of the VSV-G envelope protein band (68 kDa) is markedly reduced after size exclusion chromatography. A ~135 kDa band (dimer form also recognized by the anti-VSV-G Mab, Figure 18B) was eliminated by size exclusion chromatography as well. The authors have no convincing hypothesis to explain this loss of VSV envelope protein during chromatography. Finally, the intensity of a ~194 kDa band is also reduced after size exclusion chromatography.
A pool of iodixanol-containing fractions eluting from the size exclusion chromatography column was also analyzed by SDS-PAGE. No protein bands were detected in these fractions, with the exception of a single ~30 kDa band, nonreactive to anti-MoMLV antibodies, when Benzonase® treated virus samples were loaded onto the column (data not shown). The authors believe this band correspond to the Benzonase® nuclease used for the treatment since the enzyme consists in two subunits of 30 kDa each.
Purity determination of rate zonal purified retrovirus. 2.5 mL of a 10-fold concentrated pool of virus-containing fractions were loaded onto a Sepharose CL-4B column (60 mL bed volume). The virus was eluted at 24 cm/h in PBS buffer. Elution was monitored at 280 nm. The virus was recovered in the void volume of the column (Vo) whereas iodixanol molecules eluted later.
A) SDS-PAGE. Rate zonal purified samples before (-) and after (+) size exclusion chromatography were analyzed on 4-12% gradient polyacrylamide gels (InvitrogenTM) stained with Coomassie brilliant blue R-250 or silver stain. A wide range molecular weight standard lane (InvitrogenTM) is shown (MW). B) Western blot. Samples were fractionated as described in Figure 18A and transferred to a Protran® nitrocellulose membrane. Viral proteins bound to the membrane were analyzed by indirect immunostaining using a polyclonal anti-MoMLV antibody or a monoclonal anti-VSV-G antibody. No major differences in the protein profiles before and after size exclusion chromatography were observed demonstrating the high purity of the retrovirus preparation obtained by rate zonal centrifugation.
3.3 Subtilisin treatment analysis
Subtilisin is a serine protease extensively used by HIV-1 researchers to study the location of cellular proteins in retrovirus particles (Ott, 1997; Ott et al., 1995; Ott et al., 1996). This non-specific protease is able to digest all extravirion proteins including external virus proteins, proteins simply adhered to the virus exterior and, interestingly, proteins found in contaminating cell membrane vesicles. Interior virus proteins are protected from digestion by the viral lipid envelope. The loss in protein content after subtilisin treatment is considered an indication of the amount of cell membrane vesicles contaminating a virus preparation. To evaluate the potential contamination of rate zonal purified virions with cell membrane vesicles, a pool of virus-containing fractions was submitted to subtilisin digestion. Protease-treated virions were then separated from the digestion mixture and other protein fragments by size exclusion chromatography and analyzed by SDS-PAGE and Western blot.
Electrophoretic analysis of treated and untreated virus preparations revealed that most proteins are resistant to subtilisin digestion (Figure 19) and thus should be located in the interior of the virions. This demonstrates that most proteins observed in purified preparations are associated with virions and not with contaminating cell membrane vesicles. The complete removal of VSV-G surface protein (Figure 19B) indicates that the enzymatic treatment was efficient. An intensely silver-stained 23 kDa band was also digested by subtilisin (Figure 19A). On the other hand, a ~21 kDa fragment was generated by the protease treatment. This fragment was specifically recognized by the anti-VSV-G Mab and therefore could be a cleavage product of the VSV-G that remains on the viral membrane. Western blot using polyclonal anti-MoMLV antibody provided a further indication of purity, since this antibody, that cross-reacts with BSA and other serum proteins, did not detect any protein susceptible to degradation by subtilisin. The presence of host cell derived proteins in the virions was investigated using a polyclonal antibody against 293 HEK proteins. Three proteins (~32, 50 and 71 kDa) resistant to subtilisin treatment and thus located inside the virions were detected in the purified fractions using this antibody. Very faint bands co-migrating with the 50 and 71 kDa polyproteins are observed in lane C (HEK 293 control provided by Cygnus technologies). The origin of the 32 kDa band is less clear since although it reacted to the anti-293 HEK host cellular protein antibody, no corresponding band could be visualized in the control lane and the possibility that p30CA cross-reacted with the antibody can not be completely discarded.
Electron microscopy studies
Rate zonal purification efficiency was also monitored by transmission electron microscopy (Figure 20). Low magnification electron micrographs of the starting material (59100×) showed a significant amount of protein aggregates (Figure 20A). Typical type-C retrovirus particles were observed. The background became clearer after rate zonal ultracentrifugation (Figure 20B), thus indicating an efficient removal of contaminating proteins. Numerous roughly spherical particles in a size range of 80-120 nm in diameter were observed in the purified preparation. The particles appeared uniform in shape, except for the presence of some broken particles. Fractions 6 to 9 were further analyzed separately by electron microscopy in an attempt to explain the broad distribution of virus particles along the gradient. No significant difference in size or morphology of the virions in the different gradient fractions was observed (data not shown). Using a higher magnification (207000×), the round condensed core centered in the middle of the particle, characteristic of mature type-C retrovirus particles, was clearly observed (Figure 20C). Retrovirus particles appeared intact following treatment with subtilisin (Figure 20D) demonstrating that the protease digestion eliminates exterior virion proteins without affecting the integrity of the retrovirus particles as previously shown for HIV-1 (Ott et al., 1995).
A) SDS-PAGE. Rate zonal purified virus particles before (-) and after (+) subtilisin treatment were isolated by size exclusion chromatography and analyzed on 4-12% gradient polyacrylamide (InvitrogenTM) silver stained gels. A wide range molecular weight standard lane (InvitrogenTM) is shown. B) Western Blot. Samples were fractionated as described in Figure 19A and transferred to a Protran® nitrocellulose membrane. Proteins bound to the membrane were analyzed by indirect immunostaining using a polyclonal anti-MoMLV antibody, a monoclonal anti-VSV-G antibody or a polyclonal against 293 HEK host cellular proteins antibody. Lane C corresponds to the HEK 293 control provided by Cygnus Technologies. No significant loss in protein content after subtilisin treatment is observed (except for the VSV-G protein as expected and other faint bands that possibly correspond to exterior virus proteins) indicating that the preparations are free of contaminating cell membrane vesicles. Most proteins found in purified preparation, both virus-encoded and cellular proteins, are resistant to subtilisin digestion and thus are part of the virions. Complete removal of VSV-G was observed after subtilisin digestion demonstrating the efficiency of the treatment.
Figure 20 Transmission electron micrographs of negatively stained virions using uranyl acetate staining.
The bars represent 100 nm. (A) The 20-fold concentrated supernatant used as feed for rate zonal ultracentrifugation contained intact retrovirus particles (magnification 59100×). A significant amount of contaminating protein aggregates is visualized as a grey background. (B) Pool of purified rate zonal fractions containing uniformly shaped virus particles, free of protein aggregates (magnification 59100×). (C) Retrovirus particles purified by rate zonal sedimentation at a 207000× magnification. (D) Subtilisin digested retrovirus particles purified by rate zonal sedimentation at a 207000× magnification.
3.5 Separation of defective from functional retrovirus particles
Analysis of the gradient fractions by SDS-PAGE showed that retrovirus particles were distributed over a wide range of densities (from 1.06 to 1.09 g/mL) occupying almost 1/3 of the gradient (Figure 16A). The amount of viral proteins in these fractions increased continuously from fraction 6 to 9. However, the highest amount of infective particles was found in fractions 7 and 8. Curiously, the fraction that showed the highest amount of viral proteins (fraction # 9) also contained the lowest amount of infective viral particles (Figure 16A). This result suggests that incomplete defective viral particles band at lower densities (1.06 g/mL) and thus, can be separated from complete functional particles that move further down through the gradient using this centrifugation method.
To confirm this hypothesis, the distribution of infective viral particles and total virus particles counted by negative stain electron microscopy was compared in the purified gradient fractions. Although fractions in the middle of the gradient presented the highest infective virus titers, negative stain electron microscopy analysis confirmed that the number of total retrovirus particles increased continuously from the bottom (fraction # 6) to the top (fraction # 9) of the ultracentrifugation tube (Figure 21). Most retrovirus particles were isolated in the top fraction 9. However, this fraction showed 1 log less infectivity than the other fractions, reflecting the poor quality of the particles banding at lower densities. In fact, fraction 9 could be excluded from the purified pool without significantly affecting the recovery of infective particles. By doing so, the overall quality of the virus preparation could be improved.
Figure 21 Distribution of infectious virus particles and total virus particles throughout the gradient.
The concentration of total virus particles (VP)(
) and infectious virus particles (IVP)(
) in fractions 6 to 9 is plotted. Viral titer values shown are the mean ± standard deviation of triplicate samples. The mean density (g/mL) for each fraction is indicated in parenthesis.
The isolation of MoMLV-derived retrovirus vector particles by rate zonal ultracentrifugation is here described. The high level of purity achieved by this method was demonstrated using a variety of techniques including electrophoretic analysis, size-exclusion chromatography, subtilisin digestion and electron microscopy. Analysis by Coomassie blue stained SDS-PAGE gels is particularly valuable to evaluate the overall composition of the virus particles since Coomassie blue staining intensity is proportional to the amount of protein in each band whereas silver staining presents a high degree of protein-to-protein variability and its intensity is dependent on each polypeptide sequence and degree of glycosylation. Considering the overall composition of the retrovirus particle, analysis of the Coomassie blue stained gel indicates that the level of purity of the retrovirus preparation is greater than 95%. Although several additional minor polypeptide bands could be visualized by silver staining, these proteins were clearly associated with the particles as indicated by size exclusion chromatography and subtilisin analyses. The inability of size exclusion chromatography to further purify the pool of virus-containing fractions strongly suggests that the unidentified polypeptides visualized by electrophoresis are an integral part of the virions. Analysis of subtilisin digestion strengthens this conclusion since this treatment did not result in a significant reduction of the protein content indicating that most of these bands correspond to internal components of the virions. Some of these subtilisin resistant protein bands were identified as host proteins using an anti-HEK 293 cell polyclonal antibody. A few others were recognized by a polyclonal antibody raised against wild type MoMLV and could be incomplete Gag cleavage products and viral enzymes. Finally, NSEM studies provide further evidence that supports the level of purity observed by electrophoretic analysis. Transmission electron micrographs of rate zonal purified fractions showed homogenous fields of similar sized particles free of protein aggregates.
The identification of host cell derived proteins incorporated into the retrovirus envelope has been complicated by the presence of contaminating cell membrane vesicles containing a similar array of proteins on their membrane. HIV-1 researchers have previously attempted the use of rate zonal sedimentation for the removal of cell membrane vesicles from virus preparations (Dettenhofer and Yu, 1999). In their protocol, viral particles were loaded on top of a 6-18% iodixanol step gradient, centrifuged for 1.5 h at 250,000 × g and recovered in the 4 bottom fractions of the tube. However, the ability of this protocol to separate efficiently vesicles from virions has been questioned (Trubey et al., 2003). The authors believe that recovering the virus from the middle instead of the bottom fractions of the tube could play a role in the efficiency of separation. In fact, the separation of retrovirus vector particles from contaminating particles present in the bottom fractions observed in this work would not have been possible if the virus would have been allowed to reach the bottom of the tube. By setting the separation conditions in order to collect the virus from the middle of the gradient, the authors have shown that retrovirus preparations could be free of any significant cell membrane vesicle contamination. Thus, the virus preparations obtained by rate zonal ultracentrifugation using the described method are suitable for the study of both interior and exterior virus particle proteins.
The recovery of infective particles (37%) is reasonable considering the instability of the retrovirus particles. Moreover, it is comparable to recoveries reported for MoMLV-derived vectors using other purification methods (Kuiper et al., 2002; Transfiguracion et al., 2003; Williams et al., 2005; Ye et al., 2004). The study of the distribution of infective viral particles and total virus particles throughout the gradient is presented here for the first time to show that defective retrovirus particles can be separated from functional particles by centrifugation methods in a similar way to what was shown for other types of wild-type viruses and viral vectors (Laughlin et al., 1979; Ruchti et al., 1991; Toth et al., 1982; Vellekamp et al., 2001; Zhou et al., 1994). Retrovirus vector supernatants were found to be contaminated with a high amount of defective virus particles. Most of these particles remain in the top fraction 9 after centrifugation and their concentration decrease towards the bottom fractions. Since no significant difference in the particle size was found by electron microscopy in the 4 iodixanol purified virus-containing fractions, the authors believe that these particles slightly differ in their densities. A possibility is that the defective particles on the top fraction contain incomplete genomic RNA species while functional particles, containing complete genomic RNA species, would migrate further down the gradient. Separation of incomplete defective viral vector particles from complete functional viral particles is highly desirable for gene therapy applications since this procedure should facilitate the production of retrovirus vector preparations with optimal purity, potency and safety. Additionally, this procedure could enable the study of the biological and physicochemical properties of different retrovirus populations.
In conclusion, rate zonal centrifugation is a reliable, fast and efficient tool for the purification of retrovirus vector particles. The method renders highly purified preparations with reasonable infectivity recoveries and no evident contamination with cell membrane vesicles. Moreover, it allows for the separation of incomplete defective retrovirus particles. Using this method, the maximum volume of vector stock that can be processed per run using SW28 rotors is 18 mL of 20-fold concentrated supernatant. The concentration step was strategically placed before ultracentrifugation to allow the processing of higher volumes of supernatant per run (360 mL). The authors encourage the use of this method for other types of retrovirus particles including HIV-1 and lentiviral vectors. The technique may be particularly attractive for investigators either attempting to identify host proteins incorporated into the virus membrane or interested in studying differences between defective and functional virus populations. The authors are currently using this technique to obtain highly purified retrovirus vector preparations for the identification of proteins on the virus surface that play a role in the unspecific virus attachment to target cells.
The authors wish to thank Dr. David Ott for helpful comments and discussions and Normand Arcand and Alice Bernier for careful review of this manuscript, and technical assistance. The help of Robert Alain with NSEM, Lucie Bourget with flow cytometry analysis and Andre Migneault with image files is gratefully appreciated. This work was financially supported by NSERC and the Canadian Stem Cell Network.
Abe, A., Miyanohara, A. and Friedmann, T., 1998. Enhanced gene transfer with fusogenic liposomes containing vesicular stomatitis virus G glycoprotein. J Virol 72, 6159-63.
Alain, R., 1997. Quantitation of virus particles by negative stain electron microscopy. Microscopy Today 97, 20.
Alain, R., Nadon, F., Seguin, C., Payment, P. and Trudel, M., 1987. Rapid virus subunit visualization by direct sedimentation of samples on electron microscope grids. J Virol Methods 16, 209-16.
Bess, J.W., Jr., Gorelick, R.J., Bosche, W.J., Henderson, L.E. and Arthur, L.O., 1997. Microvesicles are a source of contaminating cellular proteins found in purified HIV-1 preparations. Virology 230, 134-44.
Dettenhofer, M. and Yu, X.F., 1999. Highly purified human immunodeficiency virus type 1 reveals a virtual absence of Vif in virions. J Virol 73, 1460-7.
Esser, M.T., Graham, D.R., Coren, L.V., Trubey, C.M., Bess, J.W., Jr., Arthur, L.O., Ott, D.E. and Lifson, J.D., 2001. Differential incorporation of CD45, CD80 (B7-1), CD86 (B7-2), and major histocompatibility complex class I and II molecules into human immunodeficiency virus type 1 virions and microvesicles: implications for viral pathogenesis and immune regulation. J Virol 75, 6173-82.
Gluschankof, P., Mondor, I., Gelderblom, H.R. and Sattentau, Q.J., 1997. Cell membrane vesicles are a major contaminant of gradient-enriched human immunodeficiency virus type-1 preparations. Virology 230, 125-33.
Hammarstedt, M., Wallengren, K., Pedersen, K.W., Roos, N. and Garoff, H., 2000. Minimal exclusion of plasma membrane proteins during retroviral envelope formation. Proc Natl Acad Sci U S A 97, 7527-32.
Kuiper, M., Sanches, R.M., Walford, J.A. and Slater, N.K., 2002. Purification of a functional gene therapy vector derived from Moloney murine leukaemia virus using membrane filtration and ceramic hydroxyapatite chromatography. Biotechnol Bioeng 80, 445-53.
Laughlin, C.A., Myers, M.W., Risin, D.L. and Carter, B.J., 1979. Defective-interfering particles of the human parvovirus adeno-associated virus. Virology 94, 162-74.
Moller-Larsen, A. and Christensen, T., 1998. Isolation of a retrovirus from multiple sclerosis patients in self-generated Iodixanol gradients. J Virol Methods 73, 151-61.
Ory, D.S., Neugeboren, B.A. and Mulligan, R.C., 1996. A stable human-derived packaging cell line for production of high titer retrovirus/vesicular stomatitis virus G pseudotypes. Proc Natl Acad Sci U S A 93, 11400-6.
Ott, D.E., 1997. Cellular proteins in HIV virions. Rev Med Virol 7, 167-180.
Ott, D.E., 2002. Potential roles of cellular proteins in HIV-1. Rev Med Virol 12, 359-74.
Ott, D.E., Coren, L.V., Johnson, D.G., Sowder, R.C., 2nd, Arthur, L.O. and Henderson, L.E., 1995. Analysis and localization of cyclophilin A found in the virions of human immunodeficiency virus type 1 MN strain. AIDS Res Hum Retroviruses 11, 1003-6.
Ott, D.E., Coren, L.V., Kane, B.P., Busch, L.K., Johnson, D.G., Sowder, R.C., 2nd, Chertova, E.N., Arthur, L.O. and Henderson, L.E., 1996. Cytoskeletal proteins inside human immunodeficiency virus type 1 virions. J Virol 70, 7734-43.
Paquin, A., Jaalouk, D.E. and Galipeau, J., 2001. Retrovector encoding a green fluorescent protein-herpes simplex virus thymidine kinase fusion protein serves as a versatile suicide/reporter for cell and gene therapy applications. Hum Gene Ther 12, 13-23.
Ruchti, F., Siegl, G. and Weitz, M., 1991. Identification and characterization of incomplete hepatitis A virus particles. J Gen Virol 72 ( Pt 9), 2159-66.
Segura, M., Kamen, A., Trudel, P. and Garnier, A., 2005. A novel purification strategy for retrovirus gene therapy vectors using heparin affinity chromatography. Biotechnol Bioeng 90, 391-404.
Toth, M., Tarodi, B. and Beladi, I., 1982. Preparative separation of intact incomplete and empty adenovirus type 2 particles. Acta Virol 26, 217-20.
Transfiguracion, J., Jaalouk, D.E., Ghani, K., Galipeau, J. and Kamen, A., 2003. Size-exclusion chromatography purification of high-titer vesicular stomatitis virus G glycoprotein-pseudotyped retrovectors for cell and gene therapy applications. Hum Gene Ther 14, 1139-53.
Trubey, C.M., Chertova, E., Coren, L.V., Hilburn, J.M., Hixson, C.V., Nagashima, K., Lifson, J.D. and Ott, D.E., 2003. Quantitation of HLA class II protein incorporated into human immunodeficiency type 1 virions purified by anti-CD45 immunoaffinity depletion of microvesicles. J Virol 77, 12699-709.
Vellekamp, G., Porter, F.W., Sutjipto, S., Cutler, C., Bondoc, L., Liu, Y.H., Wylie, D., Cannon-Carlson, S., Tang, J.T., Frei, A., Voloch, M. and Zhuang, S., 2001. Empty capsids in column-purified recombinant adenovirus preparations. Hum Gene Ther 12, 1923-36.
Williams, S.L., Eccleston, M.E. and Slater, N.K., 2005. Affinity capture of a biotinylated retrovirus on macroporous monolithic adsorbents: towards a rapid single-step purification process. Biotechnol Bioeng 89, 783-7.
Ye, K., Jin, S., Ataai, M.M., Schultz, J.S. and Ibeh, J., 2004. Tagging retrovirus vectors with a metal binding peptide and one-step purification by immobilized metal affinity chromatography. J Virol 78, 9820-7.
Zhou, L., Yu, Q. and Luo, M., 1994. Characterization of two density populations of feline calicivirus particles. Virology 205, 530-3.
© Maria Delas Mercèdes Segura, 2006
|
|
|
|
| Chapter II A novel purification strategy for retrovirus gene therapy vectors using heparin affinity chromatography |
|
Chapter IV Heparin-binding properties of MoMLV-based retrovirus vectors and its implications for gene therapy |
