A method for detecting and preventing negative RNA interference in preparation of lentiviral vectors for siRNA delivery.

The lentiviral vector is a useful tool for delivery of hairpin siRNA (shRNA) into mammalian cells. However, the efficiency of this system for carrying double-stranded siRNA (dsRNA) has not been explored. In this study we cloned the two forms of siRNA-coding sequence, a palindromic DNA with a spacer loop for shRNA and a double-stranded DNA with opposing Pol III promoters for dsRNA, into lentiviral DNA vectors, and compared their viral vector production yields. Our results indicate that sharply lower titer vector was obtained for dsRNA while much higher titer vector was produced for shRNA, posing a fundamental concern whether siRNA-carrying viral RNA itself is an inherent target of RNAi. Further experimental analyses using packaging cells that either allow or do not allow siRNA transcription indicate that the shRNA-carrying viral RNA is resistant to RNAi but the viral RNA carrier for dsRNA is not, offering a linker of RNAi bias-target secondary structure that causes shRNA vector to evade RNAi degradation. More importantly, the poor yield of dsRNA vector production was restored when a novel packaging cell line was used that blocks the antisense strand from dsRNA duplexes. This method has important implications for the RNAi field, especially for those who are using lentiviral dsRNA and dsRNA libraries for various biological discovery and therapeutic interventions.