Exogenous RNAi triggers such as shRNAs ideally exert their activities exclusively via the antisense strand that binds and silences designated target mRNAs. However, in principle, the sense strand also possesses silencing capacity that may contribute to adverse RNAi side effects including off-target gene regulation. Here, we address this concern with a novel strategy that reduces sense strand activity of vector-encoded shRNAs via codelivery of inhibitory tough decoy (TuD) RNAs. Using various shRNAs for proof of concept, we validate that coexpression of TuDs can sequester and inactivate shRNA sense strands in human cells selectively without affecting desired antisense activities from the same shRNAs. Moreover, we show how coexpressed TuDs can alleviate shRNA-mediated perturbation of global gene expression by specifically derepressing off-target transcripts carrying seed matches to the shRNA sense strand. Our combination of shRNA and TuD in a single bicistronic gene transfer vector derived from Adeno-associated virus (AAV) enables a wide range of applications, including gene therapies. To this end, we engineered our constructs in a modular fashion and identified simple hairpin design rules permitting adaptation to preexisting or new shRNAs. Finally, we demonstrate the power of our vectors for combinatorial RNAi strategies by showing robust suppression of hepatitis C virus (HCV) with an AAV expressing a bifunctional TuD against an anti-HCV shRNA sense strand and an HCV-related cellular miRNA. The data and tools reported here represent an important step toward the next generation of RNAi triggers with increased specificity and thus ultimately safety in humans.espite the recent advances in gene-engineering technologies (1), RNAi remains one of our most versatile and powerful tools for studying and manipulating gene expression in eukaryotic organisms (2). Part of its attraction stems from the simplicity with which it is triggered, namely, by introducing small, interfering double-stranded RNAs that mimic processing products of endogenous microRNAs (miRNAs) and thus engage the cellular RNAi silencing machinery. One specific variant is shRNA that consists of a stem composed of an antisense (or guide) strand that is complementary to a target mRNA and a sense (or passenger) strand that ideally is inert and merely completes the double-stranded molecule. The two strands are linked by a loop that is eliminated by the cellular RNase III enzyme Dicer during shRNA maturation, before the resulting small RNA duplex is loaded into an RNA-induced silencing complex (RISC). After removal of the sense strand, the remaining single-stranded antisense arm of the shRNA directs the RISC to a fully complementary target mRNA, which then is cleaved. The ability to express shRNAs from RNA polymerase II or III promoters offers sundry choices for cell-or tissue-specific and ectopically regulated RNAi applications ranging from genome annotation to therapeutic silencing. Moreover, the compatibility of shRNA expression cassettes with the plethora of ...