A genetic interference phenomenon in the nematode Caenorhabditis elegans has been described in which expression of an individual gene can be specifically reduced by microinjecting a corresponding fragment of double-stranded (ds) RNA1. One striking feature of this process is a spreading effect: interference in a broad region of the animal is observed following the injection of dsRNA into the extracellular body cavity. Here we show that C. elegans can respond in a gene-specific manner to dsRNA encountered in the environment. C. elegans normally feed on bacteria, ingesting and grinding them in the pharynx and subsequently absorbing bacterial contents in the gut. We find that Escherichia coli bacteria expressing dsRNAs can confer specific interference effects on the nematode larvae that feed on them.
Three C. elegans genes were used for this analysis. For each gene, a corresponding dsRNA was expressed in E. coli by inserting a segment of the coding region into a plasmid vector designed for bidirectional transcription by bacteriophage T7 RNA polymerase. The dsRNA segments used for these experiments were the same as those used previously1. We then observed the results of feeding these bacteria to C. elegans, and compared the effects with those of loss-of-function mutations and to animals microinjected with dsRNA.
The C. elegans gene unc-22 encodes an abundant muscle filament protein2. Null mutations3 or injection of unc-22 dsRNA1 produce a characteristic and uniform twitching phenotype in which the animals can sustain only transient muscle contraction. When wild-type animals were fed bacteria expressing a dsRNA segment from unc-22, 85% exhibited a weak but distinct twitching phenotype characteristic of partial loss of function for the unc-22 gene.
The gene fem-1 encodes a late component of the C. elegans sex-determination pathway4,5. Null mutations4 or injection of dsRNA1 prevent the production of sperm and lead euploid (XX) animals to develop as females (wild-type XX animals develop as hermaphrodites). When wild-type animals were fed bacteria expressing dsRNA corresponding to fem-1, 43% exhibited a spermless (female) phenotype and were sterile.
We then assessed the ability of dsRNA to interfere with a transgene target. When animals expressing a green fluorescent protein (GFP) transgene were fed bacteria expressing dsRNA corresponding to the gfp reporter1,6, a decrease in GFP fluorescence was observed in about 12% of the population (Fig. 1).
The effects of bacteria carrying different dsRNAs were fully gene specific: dsRNAs from fem-1 and gfp produced no twitching; dsRNAs from unc-22 and gfp did not produce females; and dsRNAs from unc-22 and fem-1 did not reduce GFP expression. These interference effects were evidently mediated by dsRNA, as bacteria expressing only the sense or antisense strand (for gfp or unc-22) caused no evident phenotypic effects (data not shown).
As with injected dsRNAs, the effects of feeding dsRNA to C. elegans are reversible and do not reflect a stable genetic change, as transfer of affected animals back to non-engineered bacterial food led within a generation to loss of the Unc-22 or faint-GFP phenotypes. From an engineering perspective, bacterial-mediated delivery of dsRNA is less effective than direct microinjection. This is evident from the frequency and severity of the interference phenotypes discussed above, and from observations that, for several genes known to be inhibited by injected dsRNA, bacterially mediated interference was marginal or not evident. Differences in susceptibility could reflect resistance of some cells or stages to the consequences of ingested dsRNA.
This work provides an example of RNA-mediated transfer of information between organisms and between species. It is not yet known whether such RNA-mediated interference-transfer mechanisms participate in natural ecological interactions, such as antiviral defence or communication during biological symbiosis.
Fire, A. et al. Nature 391, 806–811 (1998).
Benian, G., L'Hernault, S. & Morris, M. Genetics 134, 1097–1104 (1993).
Brenner, S. Genetics 77, 71–94 (1974).
Doniach, T. & Hodgkin, J. A. Dev. Biol. 106, 223–235 (1984).
Spence, A., Coulson, A. & Hodgkin, J. Cell 60, 981–990 (1990).
Chalfie, M., Tu, Y., Euskirchen, G., Ward, W. & Prasher, D. Science 263, 802–805 (1994).
Studier, F. W., Rosenberg, A. H., Dunn, J. J. & Dubendorff, J. W. Methods Enzymol. 185, 60–89 (1990).
About this article
Cite this article
Timmons, L., Fire, A. Specific interference by ingested dsRNA. Nature 395, 854 (1998). https://doi.org/10.1038/27579
This article is cited by
Structural insight into the human SID1 transmembrane family member 2 reveals its lipid hydrolytic activity
Nature Communications (2023)
BMC Genomics (2022)
The gut efflux pump MRP-1 exports oxidized glutathione as a danger signal that stimulates behavioral immunity and aversive learning
Communications Biology (2022)
Nature Reviews Genetics (2022)
Nature Cell Biology (2022)