Development of gene transfer methods that can precisely deliver therapeutic genes to the localized or targeted tissue(s) would be highly beneficial in developing new gene therapy approaches and may also extend animal models for studying in vivo gene function and regulation at molecular levels in the selected tissues. We investigated lipid- and AAV-mediated gene transfer in rabbit cornea using a lamellar flap-technique. The goals of this study were to (1) analyze methods for in situ gene transfer into keratocytes, (2) identify efficient and suitable vectors for gene transfer into keratocytes, and (3) characterize times of first detectable expression, localization and duration of transgene expression in keratocytes with different vectors. A lamellar flap was produced in the rabbit cornea with a microkeratome. Recombinant adeno-associated viral vector (rAAV) expressing either beta-galactosidase (rAAV-beta-gal) or chloramphenicol acetyltransferase (rAAV-CAT) reporter genes, or plasmid-cationic lipid complexes expressing CAT (pMP6-CAT) or beta-galactosidase (pTR-beta-gal) were applied beneath the lamellar flap for two minutes. The flap was repositioned and eyelids sutured overnight. Corneas were removed at 4hr, 12hr, 36hr, 3 days, 7 days, or 10 days after application and either fixed in 2% formaldehyde, cryosectioned and stained for beta-galactosidase activity or homogenized and measured for CAT levels by ELISA. Corneas infected with rAAV-beta-gal vector showed positive beta-gal staining in the center and periphery of the flap interface in whole corneas and corneal beds at 3, 7, and 10 days, but not at earlier time points. Corneas treated with pTR-beta-gal plasmid vector showed positive beta-gal expression at the interface at 4, 12 and 36hr, but not at 3 or 7 days. The posterior surface of the lamellar interface where the vector was applied showed more expression than the overlying anterior surface with both plasmid and viral vectors. The level of gene expression was less with plasmid vector than viral vector monitored using beta-gal staining. CAT-ELISA confirmed expression of the CAT reporter gene with either the plasmid or rAAV vector. These results demonstrate that foreign genes can be introduced into keratocytes with plasmid or viral vectors using a lamellar flap to gain access to the stroma. The expression profile of the reporter genes depended on the vector. Transfection of keratocytes with plasmid vectors produced rapid expression of the reporter genes, but for a short duration. Reporter gene expression following transduction by rAAV vector was delayed several days, but was at higher levels and for a longer duration. This is the first report to demonstrate selective gene transfer into keratocytes and would be highly useful in studying function and regulation of genes in vivo and may eventually furnish a tool for the treatment of corneal dystrophies.