A recent breakthrough in agricultural biotechnology is the introduction of RNAi-mediated strategies in pest control.  However, the off-target effects of RNAi pest control are still not fully understood.  Here, we studied the off-target effects of two insecticidal siRNAs in both target and non-target insects.  The results revealed that off-target effects of insecticidal siRNAs occur widely in both target and non-target insects.  We classified the expression-changed genes according to their homology to the siRNA-targeted gene, related KEGG pathways with the siRNA-targeted gene and continuous matches with siRNAs.  Surprisingly, the unintended significant changes in gene expression levels did not strictly match with the number of contiguous nucleotides in the siRNAs.  As expected, the expression of small portions of the homologous and KEGG-related genes were significantly changed.  We calculated the Shannon entropy of the transcriptome profile of the insects after injecting them with insecticidal siRNAs.  Though hundreds of genes were affected in their expression levels post siRNA-treatment, the Shannon entropy of the transcriptome remained unchanged, suggesting that the transcriptome expression was balanced.  Our results provide evidence that siRNAs cross-reacted with individual genes in non-target species, but did not have significant effects on the integrity of the transcriptome profiles in either target or non-target species on a genomic scale.  The metric we proposed can be used to estimate the off-target effects of insecticidal siRNAs, which might be useful for evaluating the safety of RNAi in pest control.  

The normalization of quantitative real-time PCR (qPCR) is important to obtain accurate gene expression data, and the most common method for qPCR normalization is to use reference genes. However, reference genes can be regulated under different conditions. qPCR has recently been used for gene expression study in Laodelphax striatellus, but there is no study on validation of the reference genes. In this study, five new housekeeping genes (LstrTUB1, LstrTUB2, LstrTUB3, LstrARF and LstrRPL9) in L. striatellus were cloned and deposited in the GenBank with accession numbers of JF728809, JF728810, JF728811, JF728807 and JF728806, respectively. Furthermore, mRNA expressions of the five genes and β-actin were measured by qPCR with insect samples of different instar at nymph stage, and the expression stabilities were determined by the software geNorm and NormFinder. As a result, ARF and RPL9 were consistently more stable than β-actin, while three TUB genes were less stable than β-actin. To determine the optimal number of reference genes used in qPCR, a pairwise variations analysis by geNorm indicated that two references ARF and RPL9 were required to obtain the accurate quantification. These results were further confirmed by the validation qPCR experiment with chitinase gene as the target gene, in which the standard error of the mRNA quantification by using binary reference ARF-RPL9 was much lower than those by ARF, RPL9 or β-actin alone. Taken together, our study suggested that the combination of ARF-RPL9 could replace β-actin as the reference genes for qPCR in L. striatellus.