Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Purification of adenovirus and adeno-associated virus: comparison of novel membrane-based technology to conventional techniques


Adenovirus (Ad) and Adeno-associated virus (AAV) are efficient gene delivery systems; manipulation of the wild-type genome allows their use as vectors for the overexpression of desirable transgenes. Generation and purification of such viral vectors can be labour intensive, costly and require specialized equipment, but a new generation of membrane-mediated ion exchange kits for purification of recombinant virus may facilitate this process. Here, we examine the yields, transgene expression and purity of preparations of Ad and AAV purified using commercially available kits in comparison to other established techniques for purification of recombinant viral vectors. We demonstrate comparable results for Ad and AAV respectively in all parameters investigated, with a substantial reduction in purification time for the kit-based technology. Such approaches are attractive methods for small-scale purification of recombinant Ad and AAV viral vectors.


The main barrier in the progress of clinical gene therapy is the absence of a safe and efficient vector system with regulatable and tissue-specific characteristics. However, progress has been made in vector development, although the optimal vector has not yet been discovered. Furthermore, deleterious effects have also been documented.1, 2 Ultimately, the refinement of gene therapy based treatment may provide the potential cure for genetic disorders, cancer and AIDS. The predominant viral vectors used for experimental gene delivery are retrovirus, adenovirus (Ad) and adeno-associated virus (AAV). Challenges associated with the use of viral vectors in clinical trials are production of sufficient quantities of clinical grade material and the biosafety of the viral vector. Production and purification methods for viral vectors vary both in terms of methodology and virus yield and quality. Here, we assess two commercially available kits for the purification of Ad and AAV vectors in comparison to established and conventional purification techniques, examining viral yield, successful expression of the transgene and quality of the final viral preparation.

Ad vectors are capable of transducing both replicating and nonreplicating cells, resulting in high levels of transgene expression and can accommodate large amounts of foreign DNA. Most vectors used to date are of first generation: E1 and E1/E3 deleted vectors. From a safety perspective, the introduction of replication-deficient adenoviral vector to humans has minimal side effects, with efficient transgene expression. However, the duration of the expression is short, an inflammatory reaction may occur and readministration is not possible due to an immune response. To minimize toxicity and reduce immunological response, advances in Ad production have focused on modifying viral production and developing new methodologies.

The standard method for purification of adenoviral vectors is based on using a cesium chloride (CsCl) density gradient combined with ultracentrifugation. Two rounds of centrifugation are performed on the virus, and the purified virus is then extracted. However, more recently, column chromatography is becoming the method of choice for the large-scale production and purification of recombinant adenovirus (rAd). Studies have focused on the development of column chromatography technology by comparing anion exchange, size exclusion, hydrophobic interaction and metal chelating resins.3 These results have shown that column chromatography offers a viable alternative to ultracentrifugation. Further improved high performance liquid chromatography (HPLC) methods have been described4, 5, 6 coupled with the use of DEAE-Fractogel and zinc affinity chromatography.7

Currently, there are commercially available kits, which use membrane-based technology as an alternative to double CsCl centrifugation. Adenopure™ is based on membrane absorber technology, which allows the binding of ion exchange functional groups to the inner surface of synthetic microporous membranes in a syringe-filter format.8 The purification time is substantially reduced to 3 h, in comparison to the traditional CsCl, which takes 1.5 days to purify. Key areas of interest are the purity of isolated Ad and the time the procedure takes. The purity and sterility of Ad are an essential outcome of any purification strategy and the ability to produce an adenoviral vector in 3 h with similar purity and sterility to that of the double CsCl centrifugation is highly favourable.

Production and purification of Ad involves many different steps. Ad (serotype 5) is first propagated in HEK293 cell line, as 293 cells contain the full E1 region of Ad5, making these cells suitable for the generation and growth of helper-independent rAd.9 Following cell lysis, purification of Ad is usually undertaken by CsCl gradient purification, a procedure that takes 1.5 days. The yield of virus may be assessed by plaque assay (determining plaque-forming units (PFU)) and optical density (OD, determining viral particles). A sensitive PCR-based method of detecting wild-type virus contamination has developed,10 based on the detection of adenoviral E1 DNA.

AAV vectors are being widely assessed as a viable alternative to traditional Ad vector technologies due to their wide ranging tropism, ability to infect nondividing cells, low immunogenicity and low levels of integration into the host genome. The 4.7 kb AAV genome consists of single-stranded DNA encoding replication (rep) and structural capsid (cap) proteins flanked by inverted terminal repeat (ITR) sequences that are essential for integration, rescue of the provirus after integration and viral packaging. The ITRs also contain the origin of replication. A recombinant AAV (rAAV) vector is generated by deleting the rep and cap sequences and inserting the gene of interest with appropriate regulatory elements between the ITRs, the only cis-acting elements required in the system. Virus particles are generated by cotransfecting this plasmid with other genetic sequences encoding the deleted rep and cap proteins in addition to Ad helper sequences such as E2a and E4. Conventional methods for AAV purification have relied on the isolation of virus by density gradient (CsCl or ioxidanol), which alone does not generate entirely pure preparations and which indeed may be cytotoxic. Additional steps were incorporated to further purify the rAAV sample, such as HPLC11 and affinity chromatography with heparin12 or mucin13 sepharose. The later generations of ion exchange resins14 have provided a further alternative to methods for AAV purifiation. However, these processes are usually laborious and intensive, may require specialized equipment and are often difficult to scale up in order to generate quantities of rAAV required for in vivo studies.

In this report, rAAV purification by a standard iodixanol gradient and heparin column affinity chromatography method is compared to that performed by a commercially available kit (ViraKit™).15 Various assays were performed to demonstrate the efficacy of the ViraKit™ in comparison to the standard technique of iodixanol gradient separation and heparin column affinity chromatography for AAV2-paraoxonase 1 (AAV2-PON1). Determination of viral titre, demonstration of viability, expression and function of PON1 transgene, in-gel staining and turbimetric endotoxin assays to demonstrate purity of the viral preparations was performed.

An advantage of both the Adenopure™ kit and ViraKit™ over the described conventional protocols was an appreciable reduction in time and labour-intensive steps required for the purification of Ad-Manganese Superoxide Dismutase (Ad-MnSOD) and AAV2-PON1, respectively, making these kits viable and attractive methods for small-scale purification of rAd and AAV viral vectors.



Time of preparation

The Adenopure™ purification took approximately 3 h, whereas the CsCl purification step required almost 1.5 days.

Volume of preparation

We compared the volume of Ad-MnSOD produced by both methods. Final volume of Ad-MnSOD purified from the Adenopure kit ranged from 3.8 to 4.2 ml, compared to a final volume of 1.4 ml for CsCl purification (Table 1).

Table 1 Volume (ml); optical density (OD): viral particles per ml (VP/ml), plaque-forming units (PFU)-plaque assay and ratio of VP:PFU for all four preparations of Ad-MnSOD


A higher physical particle titre in the CsCl purified Ad-MnSOD was shown (Table 1), while all three of the Adenopure kit AdMnSOD showed a lower physical titre (5–8.25 × 1011 VP/ml for Adenopure kit preps, with 1.07 × 1012 VP/ml for CsCl prep). This was also found to be the case with the biological titre (Table 1). The CsCl viral vector had a titre of 4.33 × 1010 PFU/ml, whereas Adenopure virus preparations demonstrated titres in the range of 1 × 1010 to 3.33 × 1010.

Total yield

The total PFU yield ranged from 4.2 × 1010 to 1.26 × 1011 PFU for the Adenopure preparations. AdCsCl preparation had a total yield of 6 × 1010 PFU (Table 1).

Particle to IU ratio

The Adenopure MnSOD 3 had a ratio of 15:1, as compared with the CsCl Ad-MnSOD, with a ratio of 23:1. Adenopure MnSOD1 and 2 showed higher particle to PFU ratios of 49:1 and 66:1, respectively (Table 1).

Replication competent Ad

RCA detection by PCR was negative (Figure 1).

Figure 1

PCR detection of wild-type adenovirus sequences. Using E1 and E2 region primers, amplification of viral DNA (extracted by DNase and phenol/chloroform) over 30 cycles was performed. Lane 1: 100 bp ladder; Lane 2: negative control; Lane 3: positive control with E1 (upper band) and E2 regions; Lane 4: AdenoPure MnSOD1; Lane 5: AdenoPure MnSOD2; Lane 6: AdenoPure MnSOD 3; Lane 7: Ad-MnSOD CsCl; Lane 8: 100 bp ladder.

Transgene expression

MnSOD expression was detected in Coronary Artery smooth muscle cells (CASMC) by Western blot and immunofluorescence. By Western blotting, the level of expression obtained was similar for both purification methods (Figure 2). Fluorescence immunostaining was also demonstrated, with significant MnSOD expression in all four viral preparations (Figure 3).

Figure 2

Immunofluorescence detection of MnSOD in CASMC transduced with Ad-MnSOD. Cells were grown in 100 mm plates and infected at MOI 100 from each viral preparation. Staining for MnSOD protein was performed after 48 h, detection was achieved with FITC-labelled secondary antibody. (a, c, e, g, i) Fluorescent microscope view, and (b, d, f, h, j) under normal microscope view. (a and b) Nontransduced CASMC; (c and d) Ad-MnSOD CsCl; (e and f) Ad-MnSOD AdPure 1; (g and h) ViraKit (Prep. 2; Ad-MnSOD AdPure 2); and (i and j) Ad-MnSOD AdPure 3.

Figure 3

Western blot for MnSOD expression in CASMC. Protein from 6 × 100 mm2 plates of CASMC infected at MOI 100 for 72 h was extracted by cell lysis and 20 μg of protein from each preparation was electrophoresed, blotted and probed for human MnSOD expression. Lane 1: nontransduced; Lane 2: AdNull; Lane 3: Ad-MnSOD CsCl; Lane 4: AdenoPure MnSOD 1; Lane 5: AdenoPure MnSOD 2; Lane 6: AdenoPure MnSOD 3.


All preparations were analysed by silver and coomassie staining. Similar levels of purity in all four preparations were observed with coomassie blue staining, but a difference was seen with silver staining; in comparison to AdenoPure 1 and AdenoPure 2 preparations, the CsCl preparation contains fewer background bands (Figure 4). However, the AdenoPure 3 preparation demonstrated fewer contaminant bands than the CsCl preparation (Figure 4).

Figure 4

Silver (a) and coomassie blue (b) staining of Ad-MnSOD viral proteins to demonstrate purity. 5 × 108 PFU/ml of virus from all four preparations was electrophoresed on a 12% SDS-PAGE gel under denaturing conditions and stained with silver nitrate (a) or coomassie blue (b). Lane 1: AdenoPure MnSOD1; Lane 2: AdenoPure MnSOD 2; Lane 3: AdCsCl1; Lane 4: AdenoPure MnSOD 3.


Endotoxin levels are described in Table 2. Both purification methods gave comparably low endotoxin levels: 3.38 EU/ml Ad-MnSOD CsCl and 3.39–3.99 EU/ml AdPure MnSOD1-3.

Table 2 Endotoxin analysis of Ad-MnSOD preparations

Adeno-associated virus

Time of preparation

The ViraKit™ purification procedure took approximately 2.5 h and the iodixanol gradient and heparin affinity column purification method was performed in 5 h.

Titre and total yield

Yields of AAV2-PON1 from both techniques of purification were closely comparable. Estimation of viral DNase resistant particles (drp) by dot blot analysis demonstrated that both purification techniques generated approximately equal amounts of AAV2-PON1 (Figure 5 and Table 3): 3.88 × 1010 and 4.24 × 1010 (preparation 1), 2.82 × 1010 and 2.65 × 1010 (preparation 2) and 2.38 × 1011 and 2.12 × 1011 (preparation 3) for ioxidanol gradient/heparin column chromatography and ViraKit™, respectively, purified from six 150 mm plates each.

Figure 5

AAV DNA dot blot. Serial dilutions of pAAV2-PON1 plasmid standard were prepared ranging from 80 to 0.3125 ng. All viral DNA samples were carried out in duplicate from each of the three preparations: (a) standard iodixanol/heparin column purification, (b) ViraKit™ purification; 10 μl of preparations 1 and 2 and 1 μl of preparation 3 were assessed.

Table 3 Yields of AAV2-PON1 from each preparation

Transgene expression and activity

Western blotting for PON1 secreted into the medium demonstrates good expression levels of PON1 for cells infected for 48 h with 1–10 μl of each viral preparation (Figure 6), with no detectable PON1 for nontransduced cells. Evidence of PON1 protein expression was supported by immunofluorescence (Figure 7), demonstrating expression of PON1 protein in cells infected with the various preparations of AAV2-PON1.

Figure 6

Western blot for PON1 expression. Media from HEK 293T cells (treatment for 48 h with 10 μl of each virus preparation) were electrophoresed, blotted and probed for human PON1 expression. Lanes 1 and 2: AAV2-PON1 Prep. 1; Lanes 3 and 4: AAV2-PON1 Prep. 2; Lanes 5 and 6: AAV2-PON1 Prep. 3. Lanes 1, 3, 5: iodixanol/heparin column purification, Lanes 2, 4, 6: ViraKit™ purification, Lane 7: nontransduced, Lane 8: purified PON1 protein.

Figure 7

Immunofluorescence detection of PON1 in HEK 293T cells transduced with AAV2-PON1. Cells were grown in six-well plates and infected with 1 μl each viral preparation. Staining for PON1 protein was performed after 48 h, detection was achieved with FITC-labelled secondary antibody. (a and b) iodixanol/heparin column purification Prep. 1; (c and d) ViraKit™ purification Prep. 1; (e and f) iodixanol/heparin column purification Prep. 2; (g and h) ViraKit™ Prep. 2; (i and j) iodixanol/heparin column purification Prep. 3; (k and l) ViraKit™ Prep.

An assay to demonstrate fully functional PON1 protein by measuring arylesterase activity was performed (Table 4). For each purification method, comparable rates of phenylacetate hydrolysis were achieved (10.23±0.22 to 26.09±5.53 U/mg/min) compared to nontransduced cells (2.26±0 to 2.56±0.09 U/mg/min) demonstrating that the PON1 protein produced is functional.

Table 4 Arylesterase activity of PON1 in 293T cells to demonstrate functional activity of the purified virus


Silver and coomassie blue staining of 12% SDS-PAGE gels containing 2.1–2.5 × 1010 total drp AAV2-PON1 (Preparation 3) were performed to assess contaminating cellular proteins (Figure 8); due to the relatively lower viral yield of preparations 1 and 2 there was insufficient material to be visible by SDS-PAGE. Viral proteins 1, 2 and 3 (VP1, VP2 and VP3) are clearly visible as arrowed (Figure 8) with no contaminating proteins evident at this level of detection. Purification of AAV2-PON1 by the standard iodixanol and heparin column protocol versus the ViraKit™ demonstrated comparable purity of the viral preparation with respect to contaminating cellular proteins.

Figure 8

Silver and coomassie blue staining of AAV2-PON1 viral proteins to demonstrate purity. Representative blots are shown; 50 μl (2.1–2.5 × 1010 drp) of virus from preparation 3 was electrophoresed on a 12% SDS-PAGE gel under denaturing conditions and stained with silver nitrate (a) or coomassie blue (b). Virus was purified by (1) iodixanol gradient and heparin column or (2) ViraKit™.


Aliquots of each sample were assessed for endotoxin content (Cape Cod Associates). Preparations 1 and 2, by both purification techniques, demonstrated comparably low endotoxin levels (3.78–5.70 EU/ml; Table 5). Preparation 3 was seen to have unusually high levels of endotoxin, possibly due to in vitro contamination during the 48 h transfection process. The endotoxin level of AAV2-PON1 purified by iodixanol and heparin column was three-fold lower (82.9 EU/ml) than that purified by the ViraKit™ (240 EU/ml) for Preparation 3, suggesting the former to be a superior purification method should there be contaminating endotoxin.

Table 5 Endotoxin analysis of AAV2-PON1 preparations

Discussion and conclusion

With an increasing number of clinical trials using viral vectors for gene transfer to humans, there is a need for the development of methods for the production and purification of large volumes of vector with high purity and sterility. For Ad, this necessitates virus purification procedures other than the traditional CsCl density gradient centrifugation, which is not suitable for large-scale virus production.7 Efforts to develop new technologies for purification of Ad have been studied, including new scaleable methodologies in virus production and purification.16 Chromatographic technology has been shown effective in adenoviral purification, using a modification of an anion exchange HPLC,4 resulting in highly purified preparations from low-titre crude lysates. Historically, purification of AAV has also been problematic with the same limitations as outlined for Ad.

The purpose of this study was to compare a commercially available membrane-based method for purification of Ad and AAV vectors with standard CsCl and iodixanol gradient with heparin column purification, for Ad and AAV isolation, respectively. Purification of Ad vectors by CsCl gradient has been the standard method for many years. This method takes approximately 1.5 days and requires equipment, such as an ultracentrifuge and storage at −80°C. In comparison, the Adenopure membrane purification method takes approximately 3 h and requires only a low speed centrifuge and can be stored at −20°C. Equally, the membrane-based technique for AAV purification takes relatively less time at 2.5 h compared to gradient and affinity chromatography purification, which takes 5 h.


For Ad, there was a notable difference in the volume produced (Table 1) from both purification methods. The Adenopure kit produced titres in the range of 5 × 1011 to 8.25 × 1011 VP/ml, whereas the cesium method had a titre of 1 × 1012 VP/ml (Table 1). Combining the total volume, with the PFU titre, gives the total viral yield obtained from each of the preparations. The total yield from AdenoPure 3 (Table 1) was nearly two-fold higher than the CsCl preparation and AdenoPure 1, while it was almost three-fold higher than that of AdenoPure2.

In addition to a high total yield, the AdenoPure kit preparation had a more favourable particle:PFU ratio. This ratio can be used as an indicator of the potential for generation of an immune response. Under FDA regulations, the recommended ratio is less than 30:1.17 Out of the three preparations from the Adenopure kit, two of the preparations had ratios under 50:1, with only AdPure2 over 50:1. Adenopure 3 preparation and the CsCl preparation met the FDA requirements, with AdenoPure 3 having the lowest ratio of 15:1. This indicates that the Adenopure kit can produce a relatively high titre virus with a satisfactory VP:PFU ratio. It is interesting to note that for the CsCl purification, 30 × 100 mm plates of 293 cells was required for the standard method, but for the Adenopure method, only five 150 mm plates were required. The same volume of starting viral stock was used for both methods, but the Adenopure kit ultimately produced a more favourable total PFU yield.

Another issue in adenoviral production is the occurrence of RCA in a population of replication deficient Ad. RCA can emerge as a result of recombination events that restore replication competency or from carry over of a replication-competent intermediate vector construct or contamination during batch processing.18 RCA virus was not detected by a PCR-based method with either purification techniques. Furthermore, successful Ad-MnSOD transduction of CASMC was demonstrated, with efficient transgene expression.

For in vivo work, high-purity viral vector preparations are required to avoid adverse reactions. For the cesium preparation, there are few contaminating cellular protein bands, indicating a high-purity preparation (Figure 4). For the AdenoPure kit preparations, variable results were observed. AdenoPure preparations 1 and 2 have more background protein bands than that of Adenopure 3 and AdCsCl. The AdenoPure 3 preparation demonstrated a lower level of contaminating cellular protein bands, similar to the cesium chloride gradient preparation. The amount of virus used for electrophoresis was based on PFU/ml rather than VP/ml. Because AdenoPure 3 had the lowest Particle:PFU ratio, this could explain why this preparation appears to be the purest in both silver staining and coomassie blue staining. Of the three preparations from the Adenopure kit, two of the preparations had comparable background protein levels with the cesium preparation, with AdenoPure 3 having a more favourable protein content.

The endotoxin levels observed in all three kit preparations were similar to the cesium preparation. In summary, we have shown the ability of the Adenopure™ kit to purify a large volume (up to 4.2 ml) and a high yield of Ad (up to 3.33 × 1010 PFU/ml).

Adeno-associated virus

Methods for the generation and optimal purification of rAAV are continuously being evaluated. Classically, as for Ad purification, CsCl2 was the method of choice but the lengthy purification process in addition to scale-up difficulties and associated toxicity for AAV2 in particular, led to the quest for alternative methods for rAAV purification. The observation that AAV2 (and now other serotypes) binds cellular heparin sulphate proteoglycans19 led to the development of novel methods for purifying AAV2 through heparin affinity chromoatography by HPLC and FPLC.11, 12, 13 Resin ion exchange chromatography technology has also been described to be effective for AAV purification,14, 20, 21 adapted also to membrane-mediated ion exchange (see below). Additionally, immunoaffinity chromoatography using an antibody (A20), which recognizes only assembled AAV2 structures, demonstrated reasonable success in purifying AAV2 by its coupling to Hi-Trap sepharose and use in an immuno-affinity column for AAV2 purification.22 Tropism of different AAV serotypes for various cellular receptors and antigens has led to novel methods for viral purification.23, 24, 25 These techniques are valuable and demonstrate success but may require special equipment (HPLC instrument and ultracentrifuges) and often are lengthy procedures. In contrast, ion exchange on a membrane matrix appears to lend itself to both time efficiency and generation of comparable yields of virus to that purified by alternative more laborious techniques26 without the need for specialized and expensive equipment as has been demonstrated here.

A critical factor in the generation of rAAV is the efficiency of plasmid transfection into the virus-producing HEK 293T cells, depending on the mode of transfection, viability and state of the cells, quality of DNA and length of transfection time. Viral titres differ between preparations as is seen in this study, which generated AAV2-PON1 at titres of 2.65 × 1010 drp to 2.38 × 1011 drp for the same size experiment performed on different days. The focus of this study was to compare the lengthier iodixanol gradient and heparin column purification method with that of Virapur's ViraKit™. Based on the recovery of virus and viral titres seen in Figure 5 and Table 3, respectively, it is evident that both protocols were closely comparable in their viral yield in terms of DNAse-resistant particles implying that the ViraKit™ method of purfication is adequate and indeed superior in terms of time efficiency, incurring little loss of viral particles compared to the iodixanol and heparin column method.

Following transduction of cells with AAV2-PON1, expression studies demonstrated that the transgene was capable of producing PON1 protein in virally transduced 293T cells as shown by Western blotting and immunofluorescence (Figures 6 and 7). Furthermore, this protein was shown to be functionally viable by arylesterase activity of PON1 catalysing the hydrolysis of phenylacetate (Table 4). AAV2-PON1-transduced cells demonstrated a four- to 10-fold increase in arylesterase activity relative to nontransduced cells, indicating functionality.

Similar to the result seen for Ad purity, further analyses of each rAAV preparation demonstrated comparable purity, with both purification methods generating AAV2-PON1 free of contaminating cellular proteins, within the detectable levels of gel staining assays (Figure 8). Silver stained and coomassie blue stained gels depicted three bands representing viral particles 1, 2 and 3 (VP1, VP2, VP3) of the AAV2-PON1 preparation 3 purified by standard iodixanol and heparin column affinity chromatography (Figure 8a(1) and b(1)) and by the ViraKit™ (Figure 8a(2) and b(2)). There was no evidence of contaminating cellular proteins. Endotoxin analysis of each preparation was performed to determine the level of endogenous pyrogens, which could be detrimental during subsequent in vivo experiments with purified AAV2-PON1. As shown in Figure 8, preparations 1 and 2 demonstrated comparably low levels of endotoxin (3.78–5.70 EU/ml). However, the final preparation showed a high content of endotoxin for both purification methods, with that for the ViraKit™ being significantly higher than for the standard method of rAAV purification. The unexpectedly high levels of endotoxin were not eliminated completely by either purification step. However, it is seen that purification of AAV2-PON1 by iodixanol gradient may provide a relatively cleaner preparation with a three-fold increase in endotoxin levels in the sample purified by ViraKit™. Providing the starting cell lysate contains relatively low levels of endotoxin, both purification methods have been shown to be equally efficient in generating rAAV of acceptable purity as seen from the values obtained for preparations 1 and 2.

An important factor in generating rAAV is the time frame involved in the purification process of rAAV from the transfected cell lysate. In this regard, the ViraKit™ is superior. The relatively lengthier iodixanol gradient method entails a centrifugal spin for 1 h followed by extraction of the appropriate band. A labour-intensive step ensues wherein the virus-containing fraction is applied to an equilibrated heparin sulphate column followed by column washes, and such columns have pressure constraints and this can be a slow and laborious step. Finally, elution of bound virus is performed. In comparison, the ViraKit™ is faster and less labour intensive. Total approximate purification times are 5 h for the standard iodixanol gradient and heparin column chromatography and 2.5 h for the ViraKit™.

In conclusion, both methods produce comparable yields of AAV2-PON1 with a fully functional transgene. Purity of the viral extract was also examined and was deemed equal for both techniques in terms of contaminating cellular proteins. We demonstrated that the ViraKit™ method was not as successful in eliminating endogenous endotoxin compared to the standard method of purification, from a preparation with high starting levels of endotoxin (preparation 3). However, two other preparations (preparations 1 and 2) demonstrated comparable purity from this viewpoint. Further studies will be undertaken to determine suitability of these viral preparations in vivo. An advantage of both the Adenopure™ kit and ViraKit™ over the described conventional protocols was an appreciable reduction in time and labour required for purification of Ad-MnSOD and AAV2-PON1, respectively, with comparable yields and purity to standard protocols, making these kits viable and attractive methods for small-scale purification of rAd and AAV viral vectors.

Materials and methods

Cell culture

HEK 293 (ATCC) and HEK 293T (ATCC) cells were grown in DME medium (Cambrex) with 4.5 g/l glucose (Cambrex) supplemented with 10% fetal bovine serum (FBS, Cambrex), 5 mM L-glutamine (Cambrex) and 5 mM penicillin/streptomycin (Cambrex). Coronary Artery Smooth Muscle Cells (CASMC) were grown in Med 231, with SMG supplement and PSA (Cascade Biologics). The cells were maintained at 37°C with 5% CO2 and 95% humidity.

Ad propagation

For the traditional CsCl method of viral purification, HEK293 cells were seeded into 30 × 100 mm2 plates. Once the cells reached a confluency of about 80%, they were infected with a high titre stock of Ad-MnSOD. In total, 100 μl of high titre stock was added to each plate, along with 1.9 ml of DME medium supplemented with 2% FBS. Plates were incubated at 37°C for 2 h, with periodic gentle agitation. After 2 h, 8 ml of DME medium supplemented with 10% FBS was added to the plates. The cells were harvested at 36 h postinfection, when the full cytopathic effect was observed.

For the Adenopure™ kit method of purification, the cells were grown and infected in a similar procedure as above but with the following modifications: (1) the cells were seeded into 5 × 150 mm2 plates; (2) 400 μl of high titre stock was added to each plate along with 19.6 ml of DME medium supplemented with 2% FBS and were incubated for 4 days postinfection, until full cytopathic effect was observed.

Ad purification

Cells and medium were collected and placed in 500 ml Beckman centrifuge tubes, spun at 1200 r.p.m. for 20 min, in a Beckman Coulter Centrifuge (Avanti™ J-30 I) and the cell pellet was resuspended in 43 ml of the supernatant with the remainder being discarded. Three freeze–thaw cycles were performed on the suspension, using a dry-ice ethanol bath and a 37°C water bath. The suspension was centrifuged at 2000 r.p.m. for 20 min and the 43 ml of supernatant was collected, containing the released virus. In an ultracentrifuge tube (Beckman SW41), 2.5 ml of 1.34 g/ml CsCl was overlaid with 2 ml of 1.43 g/ml CsCl. This was overlaid with 7 ml of viral supernatant and centrifuged at 35 000 r.p.m. for 2 h at 18°C. Two bands were visible in each tube, with the upper band representing defective viral particles. The lower band was removed and the volume was increased to 5 ml with 1.34 g/ml CsCl solution. A second ultracentrifuge spin was performed at 35 000 r.p.m. for 18 h. The resultant single band was extracted and virus desalted using a sephadex PD-10 column (Amersham Biosciences). The purified Ad-MnSOD fraction was collected and the virus was stored in 10% glycerol at −70°C. Total purification time was 1.5 days.

Purification of Ad-MnSOD using the Adenopure™ kit was performed as outlined in the kit instructions. Firstly, the cells and medium were collected, subjected to three freeze–thaw cycles as previously described and centrifuged at 2000 r.p.m. for 5 min at room temperature. The supernatant was filtered through a 0.2 μm filter unit and treated with 25 U/μl benzonase at 37°C for 30 min. A × 10 dilution buffer (kit component) was added, adding a ninth of the overall volume of the benzonase-treated filtered lysate. The Adenopure virus-binding module (VBM) was equilibrated with 30 ml of equilibration wash buffer using the provided Adenopump. The entire viral suspension was loaded and passed through the VBM, followed by 50 ml of equilibration wash buffer. The VBM was then reattached to a 10 ml syringe, containing 3 ml of elution buffer. In total, 15 drops of elution buffer was passed over the module and collected followed by the syringe/module being placed on its side for 5 min at room temperature. The remaining elution buffer was passed over the module and the purified Ad-MnSOD was collected and stored at −20°C. Total purification time was 3 h.

AAV2-PON1 preparation by Ca3(PO4)2 transfection

HEK 293T cells were plated the day before transfection onto 150 mm plates (six for each preparation) so that at the time of transfection, cells were approximately 70% confluent. A DNA precipitate was prepared by mixing a total of 500 μg pAAV2-PON1 and pDG22 plasmids at a 1:1 molar ratio, with 250 mM CaCl2 and added slowly to 2 × HBS while vortexing, to create a suspension. Following the addition of prewarmed DME medium supplemented with 10% FCS, the suspension was added to the prepared HEK 293T cells (concentrations of DNA and CaCl2/HBS described are for 10 plates) and incubated at 37°C with 5% CO2 for 48 h. Plates from each preparation were divided so that half were purified by the standard iodixanol/heparin column method and half by the ViraKit™ for a total of six plates each and cells harvested by scraping.

AAV2 purification

For iodixanol gradient and heparin column chromatography purification, the following protocol was followed.27 Cells were pelleted by centrifugation at 2000 r.p.m. for 10 min and resuspended in 15 ml of 150 mM NaCl, 50 mM Tris-HCl (pH 8.5). Three rounds of freeze–thawing were performed on the cell lysate in an ethanol/dry ice bath and 50 U/ml benzonase (Sigma) was added and incubated for 1 h at 37°C. The lysate was then centrifuged at 5000 r.p.m. for 20 min and supernatant retained and transferred to an Optiseal ultracentrifuge tube (Beckman). An iodixanol gradient was established with 15, 25, 40 and 60% iodixanol (Optiprep, Axis-Shield); the 25 and 60% fractions contained phenol red (Sigma) so that the 40% fraction, which contains the AAV2-PON1, was easily visualized. Ultracentrifugation of the gradient was performed in a Beckman ultracentrifuge (rotor type 70Ti) at 69 000 r.p.m. for 1 h at 18°C. The 40% fraction (<3 ml) was removed using a 21G needle and applied to a prewashed 1 ml Heparin HP column (Amersham Biosciences). The column was washed in 1 × PBS-MK (1 × PBS, 1 mM MgCl2, 2.5 mM KCl) and virus was eluted in 6–7 ml 1 × PBS-MK, 1 M NaCl. The viral preparation was desalted by repeated application to an ultrafiltration device (Amicon) and centrifuging at 3300 r.p.m. at RT until 1 ml remained, followed by resuspension in 1 × PBS-MK. A final volume of 500 μl of each viral preparation in 1 × PBS-MK was recovered and the virus was stored at −70°C in this solution.

Purification of AAV2-PON1 by the ViraKit™ (Virapur, LLC) was performed as outlined in the kit instructions. Cells and growth medium were harvested and freeze/thawed as described. The suspension was centrifuged for 30 min at 2800 r.p.m. and the supernatant was retained and incubated with 50 U/ml benzonase at 37°C for 30 min. The ViraKit™ filters were assembled and the viral suspension added to the prewashed filtration vessel, retaining the flow through. The filtered supernatant was then diluted in the provided Dilution Buffer and passed through the ViraKit™ filter device. Retaining the viral particles, the smaller filter was washed in Wash Buffer and 3 ml Elution Buffer (kit components) was used to elute the virus. The viral preparation was diluted in 1 × PBS-MK and washed and concentrated by applying to an Amicon Ultra filtration device as described. A final volume of 500 μl of each viral preparation in 1 × PBS-MK was recovered and the virus was stored at −70°C in this solution.

Viral vector titration

Ad-MnSOD titration

Physical viral particles were determined by measuring OD at a wavelength of A260 nm, diluted in 20 mM sodium phosphate, 0.5% SDS, pH 7.2. The conversion factor 1.1 × 1012 particles per absorbance unit at 260 nm were used to calculate particle number.28

Infectious units were determined by PFU assay. Viral vector was diluted from 10−8 to 10−11 using basal DME medium. In total, 1 ml of each serial dilution was added in triplicate to HEK 293 cells and incubated at 37°C for 2 h. The medium was then replaced with 3 ml of 2% sea plaque agar (Cambrex)/DME medium supplemented with 2% FCS. The plates were incubated at 37°C. Every 2–3 days, cells were overlaid with fresh agar/medium mixture and at day 13 the viral titre was quantified by counting the number of plaques for each dilution.

AAV2-PON1 titration

Viral titre was determined as previously described.25, 27 For each preparation, 1–10 μl of virus was treated with 1 U/μl DNase1 (Invitrogen) for 1 h and 100 μg proteinase K (Invitrogen) at 37°C for 1.5 h. Viral DNA was phenol–chloroform extracted and precipitated in the presence of 40 μg glycogen (Sigma). DNA pellets were resuspended in 400 μl of 0.4 M NaOH/10 mM EDTA pH 8.0, heated to 100°C for 5 min and cooled on ice. In parallel, known concentrations of PON1-plasmid DNA were prepared ranging from 0.3125 to 80 ng and treated similarly. Both samples and standards were applied to nylon membrane using a dot blot apparatus (BioRad), washed further with 0.4 M NaOH/10 mM EDTA pH 8.0 and the membrane was dried briefly. Prehybridization was performed using reagents from the ECL Direct Labelling and Detection System (Amersham Biosciences) for 1 h. A DNA probe was prepared by denaturing and labelling 100 ng PON1 plasmid DNA with HRP (ECL Direct Labelling and Detection System) and added to the prehybridization solution overnight at 42°C. Washes were performed in 1 × SSC/6 M urea/0.4% SDS at 42°C and the membrane was incubated in ECL reagent and exposed to BioMax X-ray film (Kodak). The viral particle titre was determined by comparing intensity of the signals relative to the standards by densitometric analysis. A quantitative titre was determined by multiplying the molecular weight of the AAV2-PON1 plasmid by the estimated density of the dot relative to standards, correcting for volume. A final correction for double-stranded DNA (standards) versus single-stranded DNA (AAV) gave the viral titre in drp/μl.

Ad wild-type detection by PCR

Adenoviral vector DNA was extracted from 10 μl of virus by treatment with 40 U DNase1 followed by incubation in 50 μg Proteinase K at 37°C for 1 h. Phenol–chloroform extraction was performed and viral DNA was precipitated with ammonium acetate and isopropanol followed by resuspension of viral DNA pellet in Tris-EDTA (pH 8.0) buffer. Viral DNA was amplified by PCR, using primers (Invitrogen) to E1 and E2 regions, for 30 cycles. Amplified DNA was electrophoresed on a 1.8% agarose gel.

Assessment of transgene expression

Western blot

For MnSOD, CASMCs were transduced in 100 mm plates with Ad-MnSOD (MOI 100) and incubated for 72 h at 37°C. Whole protein was extracted by standard techniques and total protein concentration determined by BCA protein assay (Pierce). A total of 15 mg of protein was electrophoresed on a 12% SDS-PAGE gel and resolved proteins were transferred onto a 0.2 mm nitrocellulose membrane (BioRad) using a semidry electrophoretic transfer system (BioRad). Following overnight blocking in 5% nonfat milk, the membrane was incubated with primary anti-MnSOD antibody (1:2000, polyclonal rabbit anti-human MnSOD, Stressgen) at 37°C for 30 min. Secondary antibody (1:4000, HRP-conjugated goat anti-rabbit, Stressgen) incubation was performed for 1 h at RT. All washes were performed with PBS-0.05% Tween-20. Specific MnSOD protein was detected by enhanced chemiluminescence (ECL, Amersham Life Sciences).

For PON1, HEK 293T cells were transduced for 48 h with 10 μl of AAV2-PON1 and 50 μl cell medium was electrophoresed on a 15% denaturing gel by SDS-PAGE with purified PON1 as positive control. Protein was transferred onto nitrocellulose membrane (BioRad) by semidry blotting and membrane incubated in 5% solid milk/PBS overnight at 4°C. Primary mouse anti-human PON1 antibody (1:2000, clone 3C6.33; gift from Dr D Draganov, U Michigan) in 5% milk/PBS was applied for 3 h at 4°C. Secondary HRP-conjugated sheep anti-mouse antibody (1:5000, Amersham Biosciences) in 5% milk/PBS was applied for 1 h at RT. All washes were performed in PBS, 0.05% Tween 20. PON1 protein was detected using ECL autoradiography.


For MnSOD, CASMCs were infected (MOI 100) with Ad-MnSOD and incubated at 37°C for 72 h. Cells were fixed with 4% paraformaldehyde (PFA), blocked with 5% FBS in PBS and permeabilized using 1% triton X-100. The cells were incubated in polyclonal rabbit anti-human MnSOD antibody (1:100, Stressgen) for 30 min at 37°C, washed three times and then incubated with secondary FITC-conjugated goat anti-rabbit antibody (1:100, Chemicon) for 30 min at 37°C, followed by washes as before. The cells were visualized by fluorescence microscopy.

For PON1, HEK 293T cells were seeded at 2 × 105 cells per well in a six-well plate (Falcon) and infected with 1 μl of each viral preparation for 72 h. The cells were fixed in situ with 4% PFA at 4°C for 10 min, followed by methanol at 4°C for 3 min. Cells were incubated in 0.5% bovine serum albumin (BSA) blocking solution for 15 min followed by an incubation in primary mouse anti-human PON1 antibody (1:100) for 1.5 h at 37°C and secondary FITC-labelled sheep anti-mouse antibody (1:50, Dako) for 1.5 h at RT. All washes were performed with PBS. Cells were visualized by fluorescence microscopy using the analySISB programme (Olympus).

Assessment of PON1 protein activity

Arylesterase activity was determined by using phenylacetate as a substrate. Briefly, 106 HEK 293T cells treated with 10 μl AAV2-PON1 for 48 h were sheared through a fine guage needle. Hydrolysis of 1 mmol/l phenylacetate took place in the presence of 0.9 and 20 mmol/l Tris-HCl pH 8 at 22°C. Absorbance at 270 nm was read immediately the following addition of phenylacetate and every 2.5 min thereafter for 7.5 min. Nontransduced cells were used in a similar manner as a negative control. Rate of arylesterase activity was calculated using the molar extinction coefficient of 1310 M−1 cm−1 and expressed as units of phenylacetate hydrolyzed/mg protein/min.

Assessment of purity and sterility

Volumes corresponding to 5 × 108 pfu of Ad or 2.1–2.5 × 1010 drp of AAV were electrophoresed on 12% SDS-PAGE precast gels (Pierce). For coomassie blue staining, the gels were washed three times with ddH2O and incubated with 0.25% coomassie blue stain (BDH) overnight, with gentle agitation. The gel was washed repeatedly with ddH2O and scanned using HP Precision Scan. Silver staining was performed with a silver stain kit (BioRad), according to the manufacturer's instructions and gels were visualized as before. Endotoxin levels for both Ad and AAV vectors were assessed by the PyroTurb ES Kinetic Assay (Associates of Cape Cod, Liverpool).


  1. 1

    Hacein-Bey-Abina S et al. LMO2-associated clonal T cell proliferation in two patients after gene therapy for SCID-X1. Science 2003; 302: 415–419.

    CAS  Article  Google Scholar 

  2. 2

    Lehrman S . Virus treatment questioned after gene therapy death. Nature 1999; 401: 517–518.

    CAS  Article  Google Scholar 

  3. 3

    Huyghe BG et al. Purification of a type 5 recombinant adenovirus encoding human p53 by column chromatography. Hum Gene Ther 1995; 6: 1403–1416.

    CAS  Article  Google Scholar 

  4. 4

    Blanche F et al. An improved anion-exchange HPLC method for the detection and purification of adenoviral particles. Gene Therapy 2000; 7: 1055–1062.

    CAS  Article  Google Scholar 

  5. 5

    Shabram PW et al. Analytical anion-exchange HPLC of recombinant type-5 adenoviral particles. Hum Gene Ther 1997; 8: 453–465.

    CAS  Article  Google Scholar 

  6. 6

    Klyushnichenko V, Bernier A, Kamen A, Harmsen E . Improved high-performance liquid chromatographic method in the analysis of adenoviral particles. J Chromatogr B Biomed Sci Appl 2001; 755: 27–36.

    CAS  Article  Google Scholar 

  7. 7

    Vellekamp G et al. Empty capsids in column-purified recombinant adenovirus preparations. Hum Gene Ther 2001; 12: 1923–1936.

    CAS  Article  Google Scholar 

  8. 8

    Adenopure Instruction Manual

  9. 9

    Graham FL, Smiley J, Russel WC, Narin R . Characteristics of a human cell line transformed by DNA from human adenovirus type 5. J Gen Virol 1997; 36: 59–72.

    Article  Google Scholar 

  10. 10

    Zhang WW, Koch PE, Roth JA . Detection of wild-type contamination in a recombinant adenoviral preparation by PCR. Biotechniques 1995; 18: 444–447.

    CAS  Google Scholar 

  11. 11

    Clark KR, Liu X, McGrath JP, Johnson PR . Highly purified recombinant adeno-associated virus vectors are biologically active and free of detectable helper and wild type viruses. Hum Gene Ther 1999; 10: 1031–1039.

    CAS  Article  Google Scholar 

  12. 12

    Zolotukhin S et al. Recombinant adeno-associated virus purification using novel methods improves infectious titer and yield. Gene Therapy 1999; 6: 973–985.

    CAS  Article  Google Scholar 

  13. 13

    Auricchio A et al. Isolation of highly infectious and pure adeno-associated virus type 2 vectors with a single-step gravity-flow column. Hum Gene Ther 2001; 12: 71–76.

    CAS  Article  Google Scholar 

  14. 14

    Brument N et al. A versatile and scalable two-step ion-exchange chromatography process for the purification of recombinant adeno-associated virus serotypes-2 and -5. Mol Ther 2002; 6: 678–686.

    CAS  Article  Google Scholar 

  15. 15

  16. 16

    Green AP et al. A new scalable method for the purification of recombinant adenovirus vectors. Hum Gene Ther 2002; 13: 1921–1934.

    CAS  Article  Google Scholar 

  17. 17

    McIntyre, 2001-Adenovirus Reference Material Working Group is a consortium of industry, government and academic researchers to manufacture adenovirus reference standards and codify analytical methods under the auspices of the FDA. Information is available at

  18. 18

    Zhu J et al. Characterization of replication-competent adenovirus isolates from large-scale production of a recombinant adenoviral vector. Hum Gene Ther 1999; 10: 113–121.

    CAS  Article  Google Scholar 

  19. 19

    Summerford C, Samulski RJ . Membrane-associated heparan sulfate proteoglycan is a receptor for adeno-associated virus type 2 virions. J Virol 1998; 72: 1438–1445.

    CAS  PubMed  PubMed Central  Google Scholar 

  20. 20

    Gao G et al. Purification of recombinant adeno-associated virus vectors by column chromatography and its performance in vivo. Hum Gene Ther 2000; 11: 2079–2091.

    CAS  Article  Google Scholar 

  21. 21

    Smith RH, Ding C, Kotin RM . Serum-free production and column purification of adeno-associated virus type 5. J Virol Methods 2003; 114: 115–124.

    CAS  Article  Google Scholar 

  22. 22

    Grimm D, Kern A, Rittner K, Kleinschmidt JA . Novel tools for production and purification of recombinant adeno-associated virus vectors. Hum Gene Ther 1998; 9: 2745–2760.

    CAS  Article  Google Scholar 

  23. 23

    Rabinowitz JE et al. Cross-dressing the virion: the transcapsidation of adeno-associated virus serotypes functionally defines subgroups. J Virol 2004; 78: 4421–4432.

    CAS  Article  Google Scholar 

  24. 24

    Kaludov N et al. Adeno-associated virus serotype 4 (AAV4) and AAV5 both require sialic acid binding for hemagglutination and efficient transduction but differ in sialic acid linkage specificity. J Virol 2001; 75: 6884–6893.

    CAS  Article  Google Scholar 

  25. 25

    Walters RW et al. Binding of adeno-associated virus type 5 to 2, 3-linked sialic acid is required for gene transfer. J Biol Chem 2001; 276: 20610–20616.

    CAS  Article  Google Scholar 

  26. 26

    Davidoff AM et al. Purification of recombinant adeno-associated virus type 8 vectors by ion exchange chromatography generates clinical grade vector stock. J Virol Methods 2004; 121: 209–215.

    CAS  Article  Google Scholar 

  27. 27

    Harris JD, Beattie SG, Dickson JG . Novel tools for production and purification of recombinant adeno-associated viral vectors. Methods Mol Med 2003; 76: 255–267.

    CAS  PubMed  Google Scholar 

  28. 28

    Maizel Jr JV, White DO, Scharff MD . The polypeptides of adenovirus. I. Evidence for multiple protein components in the virion and a comparison of types 2, 7A, and 12. Virology 1968; 36: 126–136.

    CAS  Article  Google Scholar 

Download references


We thank Science Foundation Ireland (SFI), Health Research Board Ireland and Higher Education Authority for their financial support and also Martina Harte (NUIG) and Vecteurotrain, for their guidance in establishing the adenoviral purification protocol. We thank Dr T Athanasopoulos, Dr J Harris and Dr G Dickson for guidance and advice with iodixanol gradient and heparin column purification of AAV, Dr D Grimm and Dr J Kleinschmidt for plasmid pDG and Dr N Madigan for cloning of PON1. We also thank Dr D Draganov for anti-hPON1 antibody.

Author information



Corresponding author

Correspondence to P M Strappe.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Duffy, A., O'Doherty, A., O'Brien, T. et al. Purification of adenovirus and adeno-associated virus: comparison of novel membrane-based technology to conventional techniques. Gene Ther 12, S62–S72 (2005).

Download citation


  • adenovirus
  • adeno-associated virus
  • purification
  • ion exchange

Further reading


Quick links