rAAV2 traffics through both the late and the recycling endosomes in a dose-dependent fashion

Abstract

Inefficient trafficking of recombinant adeno-associated virus type-2 (rAAV2) to the nucleus is a major barrier for transduction. Using imaging and subcellular fractionation techniques, we evaluated the extent of rAAV2 movement through the late (Rab7) and recycling (Rab11) endosomes. Following rAAV2 infection of HeLa cells, immunoisolation of HA-Rab7- or HA-Rab11-tagged endosomes and intracellular colocalization of Cy3-labeled rAAV2 with EGFP-Rab7 or EGFP-Rab11 markers demonstrated dose-dependent trafficking of rAAV2 through the recycling and late endosomal compartments. At low multiplicities of infection (m.o.i. 100 genomes/cell), rAAV2 predominantly trafficked to the Rab7 compartment. In contrast, rAAV2 predominantly trafficked to the recycling endosome at 100-fold higher m.o.i. siRNA studies inhibiting either Rab7 or Rab11 demonstrated that reducing Rab11 protein levels more significantly inhibited rAAV2 transduction on a per genome basis compared to inhibition of Rab7. Dose-response curves, comparing the m.o.i. of AV2Luc infection to relative transduction, also supported the hypothesis that viral movement through the Rab11 compartment at high m.o.i. is more competent for transgene expression ( approximately 100-fold) than virus that moves through the Rab7 compartment at low m.o.i. These findings suggest that strategies to shunt viral movement from the late to the recycling endosome may be effective at increasing viral transduction for gene therapy.

FIG. 1

Endosomal fractionation using iodixanol gradient ultracentrifugation. HeLa cells were infected with AV2Luc for 30 min and PNS was generated for subcellular fractionation on an iodixanol gradient. The PNS was bottom loaded into the gradient and fractions (Fr) were collected from the top (Fr 1) to the bottom (Fr 11) of the tube following centrifugation. (A) Density of each collected fraction. (B) Glucosaminidase activity (an enzymatic maker for late endosomes) was assayed for each fraction and the relative activity is plotted. (C) Quantification of AV2Luc genomes in each fraction using TaqMan-PCR. Reconstitution experiments spiking purified virus into mock-infected iodixanol fractions demonstrated that the sensitivity of the TaqMan-PCR assay was 1000 viral genomes. All experimental values were also within the linear range of this assay. (D) Purified AV2Luc was mixed with mock-infected HeLa cell PNS and was separated by iodixanol gradient centrifugation. Fractions were then quantified for the content of viral genomes using TaqMan-PCR. (E) Western blots for Rab7, Rab11, and biotin-labeled transferrin (Tf) in various iodixanol gradient fractions. 10 mM biotin-Tf was added to HeLa cell cultures 30 min prior to PNS preparation and iodixanol gradient centrifugation. Peroxidase-labeled avidin was used to detect biotin-Tf in the gradient fractions.

FIG. 2

Immunoisolation of Rab7 and Rab11 endosomes using ectopically expressed HA-tagged Rab’s. HeLa cells were transfected with (A) HA-Rab7 or (B) HA-Rab11 expression plasmids and PNS was generated at 48 h posttransfection. The PNS was then separated on iodixanol gradients and the peak endosomal fractions (2-4) were combined and subjected to immunoisolation with HA-antibody-coated magnetic Dynabeads. Western blots for HA, Rab7, and Rab11 were performed to assess the purity and recovery of each HA-Rab compartment. PNS, postnuclear supernatant loaded onto iodixanol gradients; F2-4, the combined peak endosomal fractions used for immunoisolation; B, endosomal fraction bound to HA-antibody-coated magnetic Dynabeads; UB, unbound fraction that was not absorbed to HA-antibody-coated magnetic Dynabeads. An equal percentage of each sample was loaded in each lane following concentration of endomembranes by centrifugation at 100,000g for 1 h.

FIG. 3

Quantification of rAAV2 viral genomes in Rab7 and Rab11 endosomal compartments using immunoisolation. HeLa cells were transfected with HA-Rab7 or HA-Rab11 expression plasmids 48 h prior to infection with AV2FVIII for 30 min at an effective m.o.i. of (A) 100 genomes/cell or (B) 10,000 genomes/cell. PNS was then generated and subjected to iodixanol gradient centrifugation. The peak endosomal fractions (2-4) were combined and used for immunoisolation with anti-HA bound magnetic Dynabeads. Viral genome quantification was then performed using TaqMan-PCR on a fraction of the starting PNS, combined endosomal fractions 2-4, and anti-HA immunoabsorbed endosomes. Values represent the mean (FSEM) genome copies per cell from three independent experiments.

FIG. 4

Colocalization of Cy3-AAV2 with EGFP-Rab7 and EGFP-Rab11 following infection of HeLa cells. HeLa cells were transfected with EGFP-Rab7 or EGFP-Rab11 expression plasmid. At 24 h posttransfection, cells were cooled to 4°C and incubated with Cy3-AAV2 for 60 min at 4°C at an effective m.o.i. of (A) 100 genomes/cell and (B) 10,000 genomes/cell. The cells were then shifted to 37°C and incubated for 30 min. After being washed four times with cold PBS, the cells were treated for 10 min with nocodazole (0.5 AM) on ice to disaggregate clustering of endosomal compartments. The samples were then fixed with 4% paraformaldehyde for 10 min and mounted in Vecta-Shield. Images were acquired with a Yokogawa CSU-100 confocal unit under a 63×oil objective. The boxed regions are enlarged in the lower image in each group, and arrows depict vesicles with colocalized Cy3-AAV2 and EGFP-Rab signals. (C) Quantification of Cy3-AAV2 and EGFP-Rab colocalization was then performed by counting the number of vesicles positive for EGFP-Rab11, EGFP-Rab7, and/or Cy3-AAV2. Results depict the mean (FSEM) numbers of vesicles per cell in each group from a total of 24 individual cells quantified. All vesicles were quantified within each cell. The height of the green, red, and yellow bars indicates the mean number of EGFP-, Cy3-, and colocalized EGFP/Cy3-positive endosomes per cell in each group.†Significant difference between groups for colocalized EGFP/Cy3-positive endosomes using the Student t test (P < 0.001).

FIG. 5

Inhibition of Rab7 and Rab11 with siRNA significantly reduces rAAV2 transduction of HeLa cells. HeLa cells were transfected with Rab7, Rab11, or control scrambled siRNA’s for 24 h. Transfected cells were then infected with AV2Luc at an effective m.o.i. of 100 genomes/cell for 24 h, after which the relative luciferase activity was measured for each experimental group. (A) Cell lysates were assessed for the level of Rab inhibition at 30 h posttransfection with each of the siRNA’s by Western blotting with anti-Rab7, anti-Rab11, and anti-actin antibodies. (B) The degree of targeted Rab protein inhibition was quantified using the Odyssey 2 infrared scanner from Li-cor Biosciences, Inc., and Rab protein signals were normalized to anti-actin signals for each sample. Results depict the mean (FSEM, N = 3) relative ratios of Rab to actin signal in each group. (C) Luciferase activity of siRNA-transfected HeLa cells at 24 h postinfection with AV2Luc. Results depict the mean (FSEM, N = 3) relative luciferase activity in each sample. (D) TaqMan-PCR analysis of cellular associated rAAV genomes following rAAV2 binding and/or endocytosis for the indicated times and temperature with cells transfected with the indicated siRNA’s. Analyses were performed using an effective viral m.o.i. of 100 genomes/cell. (E) HeLa cells were transfected with the indicated plasmid expression constructs. At 48 h posttransfection, cells were infected with AV2Luc at an effective m.o.i. of 100 or 10,000 genomes/cell as indicated. At 24 h postinfection, luciferase transgene expression was assessed (mean ± SEM, N = 6). †Significant difference between groups using the Student t test (P < 0.001).

FIG. 6

Nonlinear rAAV2 transduction at increasing titers of infection correlates with dose-dependent movement of virus through the Rab7 and Rab11 compartments. HeLa cells were infected at the indicated m.o.i. with AV2Luc virus. As with all other assays, virus was prebound at 4°C for 1 h, followed by a 37°C infection for 24 h in the continued presence of virus. Cells were harvested for luciferase assay at 24 h postinfection. Results depict the mean (FSEM) for N = 5 independent experimental points at each titer of virus. Two linear phases of transduction are indicated in brackets with the linear correlation coefficient as shown.

FIG. 7

Schematic summary of changes in rAAV2 trafficking through the Rab7 and Rab11 compartments as a consequence of increasing the multiplicity of infection. (Left) When HeLa cells are infected with an effective m.o.i. of 100 genomes/cell, rAAV2 predominantly traffics to the Rab7 compartment. (Right) In contrast, when HeLa cells are infected with an effective m.o.i. of 10,000 genomes/cell, rAAV2 predominantly traffics to the Rab11 compartment.