Based on invited presentation at 1st Forum of Development and Environmental Safety, under the theme “Food Safety and Sustainable Agriculture 2014”, 25 - 26 July 2014, Beijing
Genetic determinism has long dominated scientific thought, education and public understanding of evolution and biology in the West, misguiding philosophy, medicine, politics, public policies, and society at large.
The central dogma of molecular biology describes the linear flow of genetic information from DNA (deoxyribose nucleic acid) to RNA (ribosenucleic acid) to protein, with each protein performing a functional role in the organism . The concept rules out the effects of the environment on the organisms’ function and heredity. It has provided motivation and justification for biotechnologies such as genetically modified crops, where the thinking is that insertion of one gene into a crop or animal will not have consequences for the rest of the genome, or the organism as a whole and those exposed to it.
Over the last few decades there has been accumulating evidence that the central dogma of molecular biology is outdated and overly simplified. There is a paradigm shift occurring in our understanding of our intricate and complex relationship with the environment, supported by mounting work in the field of epigenetics. A few out of many examples include work showing that the in utero environment can influence the development of off-spring as well as their heath prospects long into adulthood, including life-expectancy and stress-related illnesses , while malnutrition of either fathers at time of conception of mothers at certain points during pregnancy, can affect metabolism of grandchildren [3-5]. Learned behavioural traits are found to pass down the generations [6, 7]; and chemicals can have transgenerational effects [8, 9].
There are many players involved in the complex intercommunication necessary for organism survival that can have transgenerational effects, including the natural genetic engineering of the genome itself in response to environmental cues (see  Evolution by Natural Genetic Engineering, SiS 63). Organisms themselves are able to swap DNA with unrelated species of organisms in what is now acknowledged as horizontal gene transfer. We know that the levels to which genes are switched on and off through chromatin based mechanisms, something arguably as important, or more important than the genetic sequence itself, are often passed down the generations. Non-coding RNAs have also been revealed as major players in epigenetic regulation of gene expression that can also be inherited (see  Artificial versus Natural Genetic Modification and Perils of GMOs and  Ban GMOs Now, special I-SIS for further details). I shall concentrate on epigenetic effects due to RNA interference.
There are now around 50 known classes of RNAs most of which are non- protein-coding . Non-coding RNAs are transcribed from vast regions of the genome dubbed ‘junk’ DNA not so long ago, as less than 2 % of the human genome codes for proteins. Non-coding RNAs are strongly conserved between vertebrates, invertebrates and plants, and appear to be involved in most biological and metabolic processes from editing DNA and RNA, acting as enzymes, and controlling gene expression.
One mechanism of RNA-mediated epigenetics is the RNA interference (RNAi). RNAi is a natural epigenetic process highly conserved in plants, vertebrates as well as invertebrates in which short non-coding double-stranded RNAs (dsRNAs) regulate gene expression in a sequence-specific manner, normally to down-regulate individual genes, or set of genes. Short dsRNAs includes siRNA (short-inhibitory RNA), miRNA (microRNA), shRNA (short hairpin RNA) etc., are all intermediates leading to RNA interference of protein synthesis. The underlying mechanism of RNAi gene regulation relies on the sequence complementarity of the small RNA molecules to its target mRNA of a given gene, resulting either in degradation of the target RNA, or translational repression of the protein product which is often the case when there is incomplete complementarity to the target sequence. 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 an 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 (see  New GM Nightmares with RNA, SiS 58). The nuclear pathway is known to inhibit transcription and to seed the formation of heterochromatin, an inactive, non-transcribed region of chromosomes. Short dsRNAs can target and regulate hundreds or even thousands of genes, with synthetics RNAs having an estimated 10 % error rate , despite being designed to target specific genes.
Two interesting examples of RNAi-mediated epigenetic inheritance include the Kit paramutated mouse, which was the first mouse model of RNA-mediated inheritance. Paramutation is the transfer of an epigenetic state from one allele to another of the same locus such that heterozygous mutants for the Kit paramutation can transfer its epigenetic state to the wild-type locus, which is then passed down to wild-type offspring. Homozygosity for the Kit paramutation is lethal, while heterozygotes have distinctive white feet and tail tips. Breeding these heterozygotes with wild-type mice results in some wild-type offspring with white feet and tails. Proof of the role of RNA in this non-Mendelian form of inheritance was shown through injecting RNA from the Kit heterozygotes into fertilized mouse oocytes also recapitulated the phenotype .
Another example of RNA-mediated epigenetic heritability is the recent finding that an RNA-mediated antiviral response can pass down in C. elegans worms. This is based on the inheritance of the siRNA machinery, showing that cytoplasmic inheritance of the RNAs was required for the trait to be passed on . Over 80 generations inherited this trait, showing the long-lasting effects of epigenetic inheritance in certain cases.
With all the studies mentioned showing the wide-reaching effects of short dsRNAs in mediating organism function and inheritance, it is critical within the context of GM crops that utilize RNAi technology to assess the possible exposure routes to humans as well as non-target organisms. Recent work shows that miRNAs from plant foods survive digestion and even go on to mediate genes in the body following consumption. A 2012 study analysed global miRNA levels in humans and 5 other mammalian species following consumption of rice. They found a selective uptake of 30 miRNAs and when they investigated this further in mice models, they found that one of the miRNAs, mi168a went on to mediate expression of the liver gene (LDLRAP1), leading the authors to speculate whether dsRNAs are indeed a nutrient  (see  How Food Affects Genes, SiS 53). Other studies have since confirmed the presence of exogenous RNA from food in humans, including rice, corn, barley, tomato, soybean, wheat, cabbage, grapes and carrot . Further, miRNAs have been discovered in human plasma and other body fluids including human breast milk, which is stable in conditions commonly assumed to degrade RNA including freeze thaw cycle, high and low acid conditions (as would be found in the stomach), boiling, and extended storage, all of which destroys synthetic RNA [21-24].
Monsanto have attempted to discredit Zhang’s work, most recently in a petition to the USDA for deregulation of MON87411 which contains dsRNA directed against the DvSnfy7 gene of the Western corn rootworm . DvSnfy7 is the corn rootworm’s version of the gene coding for Snf7, a protein that is well-conserved from yeast to human and several studies have revealed its role in endosomal sorting and additional functional roles in biological processes such as viral budding, cell division and regulation of gene transcription. Their petition cited a study led by Witwer, which failed to detect plant miRNA . This study used two animals only, compared to that of Zhang’s team, which included 10 women, 11 men (plus pooled serum from 10 extra individuals) along with 6 animals of 5 additional mammalian species. The Witwer study only looked at 7 plant miRNAs, whereas Zhang’s team looked at global RNA levels. Monsanto cited a second study which again failed to detect plant RNA in mammals after assessing only a few miRNAs . Further, a communication from Monsanto published in Nature Biotechnology claims to have conducted their own experiments and failed to detect exogenous rice miRNA in mice . However, failure to detect anything is not proof that the exogenous RNA is not present in the mouse. The easiest way to find nothing is to do the experiment badly. As pointed out by Zhang in response to the publication , their detection of plant miRNA even in rice, the positive control, was well below the expected levels, and hence minimizes the chance for detecting anything in the mice. The technical issues surrounding their quantitative PCR experiments, including lack of appropriate controls and absence of raw data make it impossible to judge the quality of the experiment and the findings. They clearly know the implications of this work and are trying their best, as they have done in the past, to discredit any scientist who exposes the potential dangers of their products.
I have submitted a comment to the USDA on the behalf of ISIS pointing out the above limitations as well as the risks (see below) omitted by Monsanto in their petition.
All the work done on dsRNAs and the RNAi pathway clearly indicate that any nucleic acids introduced into our foods, can survive digestion and end up in our blood and organs. This means there is a worrying possibility that these dsRNAs will interact with human genes and alter their expression, which may even be passed down the generations. Previous work using RNAi technologies for gene therapy experiments in mice found that with 49 different miRNAs introduced into mice, 36 were severely toxic; 23 were lethal in every case, killing the animals within two months, showing the potential for toxicity despite the designing of sequence-specific dsRNAs. Bioinformatics tools show lethality/toxicity of RNA-based technologies despite designing dsRNAs to specific sequences, there is around a 10 % error rate, suggesting it is impossible to create a dsRNA to target one specific gene (see  New GM Nightmares with RNA, SiS 58) and  RNA Interference “Complex and Flexible”, SiS 59) and ). MicroRNAs do not require complete complementarity to their target mRNA sequence, they only need a match of 7 bases in a row to bind, making the potential for off-target effects unavoidable. Without assessing the potential for off-target effects, it makes no sense whatsoever to approve such crops. It is impossible to predict the effects that each individual dsRNA will have on any given organism. There are also species-specific effects of RNAi, making it difficult to predict their toxicity from one organism to another.
Transgenic crops have the additional problem of expressing these transgenes in an artificial manner, interfering with the natural processes of the fluid genome and in different species. Further, as in the case of the GM wheat DIR093 generated to have altered starch content, the RNA sequences that have been inserted are present with both the matching and inverted repeat on the same strand, which does not occur naturally (see  for summary of miRNA GM crop risks). So for regulators to assert that as we eat RNAs in food all the time that RNAs are all safe to eat is both ignorant and misleading. This could be said of proteins, but prions disease has proved that we cannot make such assumptions, especially if it is out of its natural context or not naturally occurring. Sequence-independent effects are also common sources of toxicity for oligonucleotide therapeutics, which is length-dependent with increasing toxicity at lengths about 30 nucleotides, well below the length of the dsRNA DvSnf7 (240 nucleotides long) contained in MON87411 and DR093 [33, 34].
The effects of any dsRNA in GM crops are completely unpredictable and could impact generations to come. Without thorough testing of likely off-target effects, novel DNA/RNA molecules, and sequence independent effects, we cannot allow these crops to be approved. From the evidence existing to date, it appears unlikely that RNAi can really be specific to one intended target of one particular species. Keeping up to date with the ever evolving field of epigenetics is of upmost importance in identifying such risks.
Article first published 28/08/14
Got something to say about this page? Comment
There are 3 comments on this article so far. Add your comment above.
Nina Galang Comment left 29th August 2014 21:09:05
Dear Mae Wan-Ho, Thank you for your very informative and educational explanations and your continuing work on GMOs. i wrote sometime ago that i would like to purchase one of your 'black' paintings. i was answered that you were out of the country and you would get back to me when you returned. Do you still have paintings that have not been bought? Nina (Miriam College, Philippines)
Ken Lys Comment left 30th August 2014 19:07:39
The USA handing Monsanto a free pass to alter (and then own) the genome of EVERY food plant and food animal in the world is simply insanity. We are still easily decades away from understanding how life fully works through the DNA of an unaltered single species genome, and much further yet at understanding the long term implications of mixing genes cross-species. This is brand new technology and there has been enough discovered already to put a halt to this until our safety is assured - but that likley won't work given the profit Monsanto has already tasted and whatever US political motivations support this infant technology. Sadly, it will take a catastrophy in our environment or our health to wake us up.
Alice Cho Comment left 31st August 2014 15:03:26
What's the future of agriculture and food from Western Australia when its wheat & barley plant-breeding program (InterGrain) is now jointly owned by the state and federal governments and Monsanto? What's the chances of getting outcomes in the public interest from this commercial shareholding?