Inexpensive, serotype-independent protocol for native and bioengineered recombinant adeno-associated virus purification

Recombinant adeno-associated virus (AAV) is a valuable and often used gene therapy vector. With increased demand for highly purified virus comes the need for a standardized purification procedure that is applicable across many serotypes and includes bioengineered viruses. Currently cesium chloride banding or affinity chromatography are the predominate forms of purification. These approaches expose the final purified virus to toxic contaminants or are highly capsid dependent and may require significant optimization to isolate purified AAV. These methods may also limit crude viral lysate processing volume resulting in a significant loss of viral titer. To circumvent these issues, we have developed an AAV purification protocol independent of toxic compounds, supernatant volume and capsid moiety. This purification method standardizes virus purification across native serotype and bioengineered mosaic capsids.


BACKGROUND
Recombinant and pseudotyped adeno-associated virus (AAV) is the vector of choice for many laboratories because of its ability to maintain in vivo gene expression throughout an animal's life span. Naturally isolated serotypes 1-9 have been the most intensely studied and used in the laboratory setting. Each of these capsids displays a significantly different tissue tropism panel, although none appear to be target specific [1][2][3][4]. These off-target effects can dilute transduction and manifest confounding results. These observations have led investigators to successfully engineer mosaic capsids that refine the vectors targeting ability [5].
One significant hurdle for incorporating AAV into laboratory experiments remains rapid and facile vector purification across many serotypes. Current purification methods generally involve gradient centrifugation or affinity column purification. Cesium chloride and iodixanol gradients are used in many laboratories as they are independent of capsid binding moieties. While gradient purification is useful, it has been shown to significantly reduce infectivity. In addition, the associated health risks and toxicity may introduce confounding experimental results if not completely removed. Affinity column chromatography is also a favored method of vector isolation. This method relies on capsid binding motifs and often requires significant resources to optimize. However, considering the massive potential for capsid design and manipulation, purification techniques requiring the presence of a specific moiety are simply not practical.
Here, we detail an inexpensive purification method independent of gradient centrifugation, capsid moiety and has been used by us with success for high titer (>2 × 10 13 vector genomes [vg]/ml) viral-mediated in vivo heart transduction studies [6]. This protocol incorporates polyethylene glycol 8000 precipitation of empty and full AAV capsids and two spin purification steps based on the sedimentation coefficient for AAV2 capsids. The sedimentation coefficients of empty and full capsids are ~72 s and ~138 s respectively. As proof-of-concept, we used our protocol to purify pseudotyped AAV2/6 and AAV2/41. AAV2/41 is a mosaic capsid created through DNA shuffling experiments and contains elements from capsids AAV 1, 6, 7, and 8 [11]. While several previously published protocols for AAV purification incorporate PEG to precipitate virus [13], all focus on cell pellet processing and discard culture media at harvest. In our experience, this translates into a significant loss of titer (upwards of 80%) due to HEK cells detaching during viral production. For this reason, we developed our protocol to capture AAV virions in both cell pellet and cell media concomitantly. Transfecting adherent HEK293 cells is still the most popular method of AAV production and we have developed our protocol to reflect primary research (non-clinical) needs.
All viruses contain a constitutively expressed luciferase cassette. The protocol is focused on a rapid and inexpensive rAAV production/ purification technique that offers several benefits over existing methods. These include ease of use by incorporating Cell Factory transfection, whole preparation processing reducing loss of titer, high titers, no gradient or chloroform extraction required over existing protocols [7][8][9][10]. This protocol is written here to be highly accessible to the non-specialist.

Recipes
All cell culture, transfections, related media handling and the final viral suspension should be performed in a biosafety cabinet using sterile technique. Plasmid preparations, lysate filtrations and centrifuge tube transfers may be performed at the bench.

Cell culture media
For For cell seeding media, add 100 ml FBS, 10 ml L-Glutamine, 10 ml Antibiotic-Antimycotic and 10 ml 200 mM MgCl 2 to 870 ml DMEM. The serum exchange media is DMEM supplemented with L-Glutamine, Antibiotic/Antimycotic, 1x MEM Non-Essential Amino Acids only.

Plasmid DNA
Plasmid DNA should be prepared previous to transfection and stored at -20°C for up to one year. The most efficient method of preparing large amounts (5-10 mg) is by using giga-scale purification kits. Qiagen provides an extremely fast, pure and high concentration DNA stock in their Plasmid Plus Giga Kit. Plasmid stocks should be suspended to 1.0 mg/ml. TIP: Plasmid DNA must have a 260 nm/280 nm absorbance ratio between 1.80 and 1.85. Plasmid stocks that fall out of this range cause extremely large aggregates during transfection that are not readily taken up by cells. In addition, all plasmids should be maintained in E. coli strains engineered for low recombination events, as the AAV2 ITR's display significant recombination with the bacterial genome. To ensure minimal recombination, ITR's should be digested as described previously [12,13].

Calcium phosphate precipitation reagents
Transfection Reagent A. Transfection reagent A is made up fresh at the time of transfection. In a 250 ml conical tube combine 50.4 ml ddH 2 O, 6.0 ml 2.5 M CaCl 2 and AAV plasmids. Our dual plasmid system uses 1.2 mg shuttle vector to 2.4 mg pDGM6.
Transfection Reagent B. Transfection reagent B can be made up previously and stored at room temp for up to six months. Combine 50 mM BES, 1.5 mM Na 2 HPO 4 and 280 mM NaCl 2 .
TIP: pH must be between 7.00 and 7.05.

× citric saline cellular detachment solution
For a 1 × working stock, dilute 20 ml 10 × solution to 200 ml in Hyclone molecular biology grade water.

Precipitation reagent stock solution (PEG/NaCl) (5 ×)
This precipitation solution should be made in advance and can be stored at room temp for up to three months. Dissolve 400 g PEG8000 and 146.1 g NaCl 2 in 1.0 L ddH 2 O.

Pellet suspension buffer
Dissolve 14.61 g NaCl in 1.0 L Hyclone Water for a final concentration of 250 mM.

Animals
C57Bl6 mice were used in this study to determine in vivo functionality of purified vector. All experiments were in complete compliance with Institutional animal guidelines.

4.2.
Transfer crude lysate/PEG mixture into two 500 ml conical tubes. The tubes can accommodate a maximum volume of ~650 ml (Crude Lysate in Fig. 2).

4.7.
Clarify pellet suspension by centrifugation at 10,000 × g and 4°C for 10 min.

4.8.
Transfer the viral containing aqueous layer to a Beckmann Ultra Clear SW32 tube and spin at 149,000 × g/4°C for 3 h. Discard aqueous layer ensuring minimal pellet disruption (Aqueous Layer B in Fig. 2).

ANTICIPATED RESULTS
We developed a new AAV purification protocol and validated it using pseudotyped AAV2/6 and AAV2/41. AAV2/41 displayed similar results to AAV2/6 (data not shown). Figure one compares AAV2/6 heparin sulfate column and our PEG/Spin method based on vector genome recovery normalized to cell growth area. A 1.0-L AAV Cell Factory transfection was split equally and processed accordingly. Our purification protocol captured almost 30 × more POL Scientific vector genomes per cm 2 , indicating equivalent titers may be gleaned from considerably less cell culture. Aliquots at various steps of our protocol were collected to identify total protein loss and capture. Protein tracking experiments illustrate a step-wise progression to viral concentration and purity, and need not be performed for general quality control (QC) validation. To that end, we have instituted a three-tier QC process involving SDS-PAGE/ Coomassie staining to assess capsid capture and purity, qRT-PCR to identify vector genome titer and a bioactivity transduction assay to ensure transgene expression. Because these assays are common and dependent on individual AAV applications, their protocols fall outside the scope of this article. Vector genome titers were not performed on protein tracking samples as the highly variable nature of media and buffers confounds qRT-PCR results. Figures two and three illustrate quantitated total protein tracking and visualization respectively. Taken together, these figures demonstrate most contaminating proteins are removed during the crude pellet and final centrifugation steps.
The final purity of the AAV suspensions is visualized in Figure 4. Gels were imaged using a Li-Cor Odyssey that provides Coomassie resolutions equivalent to silver staining. As observed, very little cellular proteins are carried over to the final suspension. Also, dialysis may be implemented if further purification is required. It should be noted that lane smearing and non-capsid bands are most likely partially denatured viral aggregates.
To confirm in vivo functionality, we delivered a CMV luciferase cassette to C57BL/6 mice at 2.0 × 10 12 vg of AAV2/6 or AAV2/41 (Fig.  5). Virus was delivered systemically via tail vein. Both of these preparations successfully delivered the reporter cassette to murine tissues and demonstrated expression patterns consistent with published literature. Variable luciferase expression levels are attributed to different capsid infectivity efficiencies and are not reflective of our process.
In conclusion, this rapid PEG/Spin protocol for AAV purification produced significantly higher yields than traditional column chromatography. Both wild-type and bioengineered virions were able to be captured and functioned equivalent to column purified AAV. This protocol represents an inexpensive, efficient and scalable method for AAV purification in the laboratory setting that is independent of capsid serotype or moiety.

TROUBLESHOOTING
Possible problems and their troubleshooting solutions are listed in Table 1. • Suspension of PEG/NaCl and final pellet must be carried out in high osmotic buffers to prevent aggregation.