Most commercially grown genetically modified (GM) crops are engineered to produce foreign proteins, but new ones are increasingly engineered to produce RNA of a special kind - double-stranded RNA (dsRNA) - that aims to interfere with the expression of a specific gene, usually to silence the gene(Table 1).
Table 1 GM crops with dsRNA commercially approved or the approval pipeline
| Flav Savr tomato
||Monsanto||Withdrawn from market|
| High oleic acid soybean lines G94-1, G94-19 and G168
Withdrawn from market
| New Leaf Y and New Leaf Plus Potato
||Dupont Pioneer|| FSANZ* approved 2001
Withdrawn from market
| High oleic acid soybean lind DP-305423-1
||Dupont Pioneer||FSAMZ* approved 2010|
| Herbicde tolerant, high oleic acid soybean Line MON87705
| Golden mosaic virus resistant pinto bean
||Brazil approved 2011|
| Papaya ringspot virus resistant papaya
||Hawaii University|| USA 1996, Canada 2003, Japan 2011
|Altered grain starch wheat|| CSIRO*
|| Approved for field trials & feeding experiment
*CSIRO Commonwealth Scientific and Industrial Research Organization *Embrapa Brazilian Agricultural Research Corporation
*FSANZ Food Standards Australia New Zealand
The ability of dsRNA to interfere with gene expression was known since the 1980s; and the biochemistry of the phenomenon - referred to as RNA interference (RNAi) - was worked out in the roundworm Caenorhabditis elegans in the late 1990s . The same RNAi pathway has been identified since in practically all plant and animal kingdoms . DsRNA includes siRNA (short-inhibitory RNA), miRNA (microRNA), shRNA (short hairpin RNA) etc., all intermediates leading to RNA interference of protein synthesis. This can happen at transcription, or at translation. Typically, dsRNA originates from a long RNA molecule with stretches of complementary base sequences that base pair to form a stem ending in a non-base-paired loop. This stem-loop structure is then processed into a shorter dsRNA, and one strand, the guide strand does the job of interfering. It binds to a mRNA (messenger RNA) molecule in the cytoplasm by complementary base-pairing to prevent the mRNA from being translated into protein. Alternatively, the guide strand targets and chemically modifies DNA sequences in the nucleus by adding methyl groups to the DNA, and cause modification of histone proteins associated with the DNA. The nuclear pathway is known to inhibit transcription and to seed the formation of heterochromatin, an inactive, non-transcribed region of chromosomes.
Actually, dsRNA genetic modification has been used before. The first GM crop to be commercialized, the Flav Savr tomato, created with 'antisense' technology to delay ripening, is now known to act via dsRNA (Table 1).
Interestingly, the gene silencing effect of dsRNA can become inherited (either indefinitely, or through two or more generations) in cells and organisms that are not genetically modified, but simply exposed to the dsRNA for a period of time. It can happen via methyl groups added to the DNA, or the modification of histones, without changing the base sequence of the DNA in the genome [3, 4]. This is another example of the inheritance of acquired characters now known to occur through many different mechanisms (see  Epigenetic Inheritance - What Genes Remember and other articles in the series, SiS 41) that makes genetic modification all the more hazardous.
Obvious dangers of dsRNA ignored by regulators
DsRNA genetic modification has large implications on safety based on what is already known (see below): DsRNA is stable, it resists digestion and may enter the bloodstream; its role in modifying gene expression is universal and acts across kingdoms; toxicity to animals have been amply demonstrated and exploited in targeting pests; although the intended target is a specific gene, many off-target effects have been identified; finally, plant dsRNA has been found circulating in the human bloodstream where it can be taken up into cells and tissues to interfere with the expression of genes. Consequently, animals including human beings eating the GM food containing dsRNA could well be harmed.
However, regulators are ignoring and dismissing the findings despite repeated warnings from scientists. Jack Heinemann at the University of Canterbury, Christchurch, in New Zealand and his colleagues have had the same experience as independent scientists everywhere with their national regulators; in Heinemann's case, the Food Standards Australia New Zealand (FSANZ). FSANZ has approved for use as human food at least 5 GM products with modification to produce dsRNA (see Table 1), in blatant disregard of evidence brought to their notice again and again. Heinemann and colleagues call it aptly  "regulation by assumption", and show how the same applies to regulatory agencies in the US and in Brazil.