To determine feeding conditions that maximize observable phenotypes, we started with the existing L4440 vector and strain HT115(DE3) and varied a number of parameters that could affect the efficiency of RNAi. We chose two initial test genes that were easy to assay: gpb-1, for which mutants are embryonic lethal, and unc-22, which results in a post-embryonic uncoordinated phenotype (Unc), as determined by deletion mutants and by RNAi [4,7]. We first tested different methods of induction with isopropylthiogalactoside (IPTG) to see if this would affect the RNAi phenotypes observed. Uninduced bacteria produced no phenotypes, but, somewhat unexpectedly, each presumably stronger method of induction resulted in a lower penetrance of phenotypes, culminating in 0% phenotype from overnight induction in culture (Table 1). The best induction method was to grow bacteria in culture without induction, to seed these bacteria onto plates containing IPTG, and then to incubate overnight at room temperature; with this method, gpb-1 produced 100% dead embryos and unc-22 produced 99% Uncs. To further test this new induction method, we fed two more genes, par-1 and par-3, mutations in either gene
