Genetics instability of wtAAV2 genome and AAV promoter activities in the Baculovirus/Sf9 cells system

Abstract

The human Adeno-Associated Virus serotype 2 (wtAAV2) is a common non-pathological virus and its recombinant form (rAAV) is widely used as gene therapy vector. Although rAAVs are routinely produced in the Baculovirus/Sf9 cell system, wtAAV2 has never been studied in this context. We tried to produce wtAAV2 in the baculovirus/Sf9 cell system hypothesizing that the wtAAV2 may be considered as a normal recombinant AAV transgene. Through our attempts to produce wtAAV2 in Baculovirus/Sf9, we found that wtAAV2 p5 promoter, which controls the expression of large Rep proteins in mammalian cells, was active in this system. p5 promoter activity in the baculovirus/Sf9 cell system leads to the expression of Rep78 that finally excises wtAAV2 genome from the baculovirus genome during the earliest phases of baculovirus stock production. Via p5 promoter expression kinetics and strand specific RNA-Seq analysis of wtAAV2, rAAV and Rep2/Cap2 cassettes in the baculovirus context we could demonstrate that wtAAV2 native promoters, p5, p19 and p40 are all active in the context of the baculovirus system and lead to the expression of different proteins and peptides. In addition, this study has proven that the baculovirus brings at least some of the helper functions needed in the AAV replication/life cycle.

Fig 1. PCR at the baculovirus Tn7 site after clonal selection with lysis plaque assay.

A. PCR bands from two baculoviruses carrying the γ-SGC transgene cassette (3.8 kb length). B. PCR from three clones of baculovirus carrying the wtAAV2 genome after lysis plaques. Bands are only one kb in length, meaning that wtAAV2 genome has been excised from the baculovirus backbone. C. PCR on three empty (Bac-to-Bac, Invitrogen) baculoviruses after lysis plaque displayed a 325 bp in length bands. D. PCR on empty bacmid DNA and on bacmid carrying the wtAAV2 genome. The gap between the PCR bands on row B and the second one on row D demonstrate the excision phenomenon.

Fig 2. Fluorescence microscopic observation of Sf9 cells transfected with bacmid containing the Rep78eGFP cassette and expressing GFP.

A, B and C, 30, 55 and 120 hours post transfection of bacmid containing the Rep78eGFP fusion protein under the native p5 promoter. GFP is visible after 30 hours post transfection, and the combination of baculovirus replication followed by cell infection led to near 100% of the cells being GFP positive 120 hours post transfection. D. 120 hours post transfection with the bacmid containing the construct Rep78eGFP delta ITR.

Fig 3. Microscopy observation of infected Sf9 cells 5 days post-infection.

Three different baculovirus (carrying the Rep78-eGFP ITR construct) lysis plaque clones have been used to infect T75 cell flasks (panels 1–3). Despite obvious signs of baculoviruses infection, only in one flask eGFP expression could be observed, meaning that the ITR-Rep78eGFP-ITR construct has been excised from the baculovirus backbone before the reinfection in most of the cases.

Fig 4. Western blot against Rep proteins, five days post transfection with different bacmid constructs (as indicated above the rows).

Large Rep78 proteins are always detectable, excepted for the p5eGFP construction. Small Rep52 proteins are largely less detectable than large Rep proteins. The gap between Rep size can be explained by the eGFP fusion. p35 proteins represent the overall baculovirus proteins amount.

Fig 5. Kinetics of GFP expression by FACS analysis during the seven days following the bacmids transfections.

The positive control (blue dots): bacmid with the eGFP protein under the control of CMV promoter. Negative control: cells transfected without bacmid DNA, showing no GFP fluorescence, as for the Bac-to-Bac construct (without any GFP). GFP expression levels were highly similar between the p5eGFP construction and the Rep78eGFP (without ITR), excepted 5 days post transfection. In comparison the Rep78eGFP, within the ITRs context, showed a higher level of GFP expression than the same construct without ITR. This highlights the influence of ITRs on Rep expression via p5 and/or p19 promoters (n = 3). The bars represent the standard deviation.

Fig 6. Western blot revealed GFP expression 5 days post-transfection.

Sf9 cells were transfected with the different bacmid constructs as indicated at the top of the rows. GFP protein was not detectable, as expected, for the WT AAV2 baculovirus. Rep78eGFP and Rep52eGFP were detected at the expected sizes, respectively. GFP driven by the p5 promoter appeared at 27 kDa, and with the Rep constructs a 32 kDa band was clearly detected.

Fig 7. Fluorescent microscopy of transfected Sf9 cells 48 hours post-transfection.

A. Cells in the have been transfected with a bacmid DNA carrying the ITR-p5-Rep78eGFP-ITR in fluorescent microscopy, light microscopy and merged. B. Cells transfected with a plasmid carrying the same construct. eGFP expression is only visible in a baculovirus context.

Fig 8. RNA-Seq reads alignment on WT AAV2 positive strand genome 48 and 94 hours post transfection.

Peaks of transcription were seen for the sites of the p5, p19 and p40 promoters, with p40 representing the highest level of reads.

Fig 9. RNA-Seq strand specific reads alignments on Rep/Cap cassette, 48 and 94 hours post transfection.

In dark red the reads correspond to rep sequence alignments on the negative strand and in black, alignments of the cap sequence 48 hours post transfection. Pink color represents the rep mapping 94 hours post transfection and the dark blue the cap alignment 94 hours post transfection. This Rep/Cap cassette used leaky scanning mechanisms to produce all needed proteins. Light blue dots represent the stack of reads aligning to the cassette at specific points. At both time-points we can clearly detect p10 and polh promoters, but also p19 and p40 transcription. Contrary to the WT AAV RNA-Seq the negative strand of both Rep and Cap genes is only less transcribed.

Fig 10. WT AAV2 genome in baculovirus RNA-Seq (strand specific).

Dark green represents the reads aligned on the coding strand. Dark grey represents the negative strand. Contrary to the studies performed by Stutika et al. (23), the WT AAV negative strand is highly transcribed in the baculovirus context.