Protein trans-splicing as a means for viral vector-mediated in vivo gene therapy

Abstract

Inteins catalyze protein splicing in a fashion similar to how self-splicing introns catalyze RNA splicing. Splitinteins catalyze precise ligation of two separate polypeptides through trans-splicing in a highly specific manner. Here we report a method of using protein trans-splicing to circumvent the packaging size limit of gene therapy vectors. To demonstrate this method, we chose a large dystrophin gene and an adeno-associated viral (AAV) vector, which has a small packaging size. A highly functional 6.3-kb Becker-form dystrophin cDNA was broken into two pieces and modified by adding appropriate split-intein coding sequences, resulting in splitgenes sufficiently small for packaging in AAV vectors. The two split-genes, after codelivery into target cells, produced two polypeptides that spontaneously trans-spliced to form the expected Becker-form dystrophin protein in cell culture in vitro. Delivering the split-genes by AAV1 vectors into the muscle of a mouse model of Duchenne muscular dystrophy rendered therapeutic gene expression and benefits.

FIG. 1.

An illustration of the protein trans-splicing method for overcoming the size limitation of gene delivery vectors. A large gene is converted into two smaller, intein-containing split-genes and delivered into a target cell by means of two vectors. Expression of the two split-genes produces two intein-containing split-proteins that spontaneously undergo protein trans-splicing to produce the mature protein.

FIG. 2.

Application of the protein trans-splicing method to dystrophin and AAV vectors. (A) Schematic illustration of the experimental procedure. The 6.3-kb Becker dystrophin cDNA is split at the site marked with an arrowhead, and the split-intein coding sequences (N and C) are added as illustrated. The recombinant split-genes (N-dys gene and C-dys gene) are each packaged in an AAV vector, which contains inverted terminal repeats (ITRs), a poly(A) site, and the CMV promoter. After infection of a target cell with both split-genes and after gene expression, the resulting split-proteins (N-dys protein and C-dys protein) spontaneously undergo protein trans-splicing to produce the Becker dystrophin protein. (B) Amino acid sequences of the N-dys protein and the C-dys protein, which were generated with the S1067 split site of Becker dystrophin and the Ssp DnaB split-intein. Only sequences near the dystrophin–intein junctions are shown; the lengths of omitted sequences are indicated with numbers in parentheses, and the intein sequences are indicated with boldface letters. Sequences of helix 2 and helix 3 inside the sixth of the 8.5 spectrin-like coiled-coil repeats (rods) of the Becker dystrophin are also marked.

FIG. 3.

Observation of dystrophin production and function, using the protein trans-splicing method. (A) Western blot analysis of dystrophin production. The human 293 cell line was transfected with either the N-dys gene or C-dys gene, or both, as indicated (−, absent; +, present), using a total of 25 μg of plasmid DNA per transfection. Cells were harvested 48 hr after transfection, and ~50 μg of total cellular protein was resolved by polyacrylamide gel electrophoresis under denaturing (SDS) and reducing (DTT) conditions. Western blotting was done to visualize dystrophin, using the monoclonal antibody NCL-DYS3 against the N-terminal part of dystrophin (Anti-N-dys) or the monoclonal antibody NCL-DYS2 against the C-terminal part of dystrophin (Anti-C-dys). The N-dys protein (marked with the letter N), the C-dys protein (marked with the letter C), and mature dystrophin (marked with the letters N-C) have predicted sizes of 135, 111, and 212 kDa, respectively. In lanes 2, 3, and 5, the N-dys and C-dys proteins contain the Ssp DnaB split-intein. In lanes 4 and 6, the N-dys and C-dys proteins contain the Rma DnaB split-intein. (B) Immunofluorescence staining of functional dystrophin. The gastrocnemius muscles of 3-week-old male mdx mice were coinjected with AAV1-N-dys and AAV1-C-dys vectors and examined after 6 months. Each muscle was injected with 40 μl of AAV1-N-dys (1.5 × 10¹¹ vector genomes [VG]) and AAV1-C-dys (0.5 × 10¹¹ VG) at a 3:1 ratio. Cryosections of mdx muscle were stained immunofluorescently either with a rabbit anti-R1R2 antibody against the N-terminal part of dystrophin (panel b) or a rabbit anti-R22R23 antibody against the C-terminal part of dystrophin (panel c). As a negative control, a cryosection of untreated mdx muscle was stained immunofluorescently with the anti-R1R2 antibody (panel a). All panels were then counterstained for cell nuclei with DAPI (blue). Note the lack of centrally localized nuclei in the AAV-treated muscle, indicating the protective effect of the gene vectors.