William Saletan exposes his ignorance of the basic mechanisms of molecular biology

In the second of a two-part series, Claire Robinson points out scientific errors in the journalist William Saletan’s latest attack on her and other GMO critics

In part 1 of this series I explained how the journalist William Saletan, in an attack on me and other GMO critics, had avoided confronting facts and scientific evidence on GMOs. Crucially, Saletan exposed his ignorance of the basic mechanisms of molecular biology, using false and misleading claims about these technical topics to suggest that GMOs are safe.

It started with Saletan taking exception to my point that “the GM gene insertion and subsequent tissue culture processes used in genetic engineering create mutations. These can result in biochemical changes in the plant, which in turn could make the plant unexpectedly toxic or allergenic.”

Saletan countered: “Mutations are hardly unique to GMOs. They’re ubiquitous, especially in plant breeding. You’re no more likely to get a toxic mutation from a GMO than from a non-GMO.”

These are bold assertions that require evidence to back them up. So I followed up his two linked references. The first leads to a blog page on the pro-GMO website Academics Review.

The problem with the Academics Review page is that it misleads the public, apparently deliberately. It confuses a discussion of mutations with references to changes that happen as a result of conventional breeding. Most of these changes are not caused by mutations; nor are they equivalent to mutations. Saletan does not appear to realize this and his expert sources have not enlightened him otherwise.

What are mutations?

Mutations are permanent changes in the DNA base sequence, which can result in damage to the structure and function of one or more genes. Such damage can be caused by errors in DNA replication, exposure to radiation or chemical mutagens, or the GM transformation process (insertion of the foreign transgene cassette into the DNA of the plant cell and growing of the plant cells in tissue culture).

Mutations are often harmful to the living organisms in whose DNA they occur. For that reason, living organisms have DNA repair mechanisms to correct mutations and minimize their impacts. Also for that reason, chemicals known to cause mutations are banned or restricted by regulators worldwide.

In contrast, the vast majority of the changes that happen in traditional breeding – such as differences in DNA due to reshuffling (“recombination”) of intact genes and changes in gene expression – are not mutations. They do not share the often harmful effects of mutations. They are simply variations on a theme arising from combinations of chromosomes from the two parent plants. This is a good thing. If it didn’t happen, every marriage would be the equivalent of breeding with siblings.

Are mutations “ubiquitous” in plant breeding?

Contrary to Saletan’s claim, mutations are not “ubiquitous” in plant breeding. Conventional plant breeding can produce heritable mutations, but it seems to be a rare event. A plant breeding textbook by Denis Murphy, Professor of Biotechnology at the University of Glamorgan, UK, says, “In the normal course of events, the appearance of a spontaneous germline [heritable] mutation is relatively rare.”

Mutation breeding does not equal conventional breeding

Another falsehood of the Academics Review page linked to by Saletan is that it equates mutation breeding, or mutagenesis, with conventional breeding. It says that conventional breeding, “accepted as safe by all”, causes “much more genetic disruption” than GM: “Conventional plant breeding shuffles DNA more than transgene insertion does… The magnitude of these unpredictable and unexpected changes far exceeds the more limited number of changes created by transgene insertion… [yet these] have never been linked with an adverse outcome.” Academics Review gives three references to support this claim, two of which (Batista and colleagues 2008; Shirley and colleagues 1992) focus on mutation breeding.

This is a favourite trick of GMO proponents. Instead of comparing GM with safe conventional breeding, they compare GM with mutation breeding, another risky and imprecise technology. And because mutation breeding is used, albeit on a relatively small scale, by conventional plant breeders, GMO proponents conflate mutation breeding with conventional breeding – but without actually specifying that they are talking about mutation breeding. Then they falsely conclude that GM is “safer than conventional breeding”. What they are really saying is that GM is safer than mutation breeding. This may or may not be true: there appears to be no toxicological research directly comparing the two, so it is not possible to conclude which is safer.

The trick of conflating mutagenesis and mutation breeding with conventional breeding is dishonest. While mutagenesis is sometimes used in conventional breeding, it is not in itself conventional breeding. Mutagenesis consists of bombarding plant genomes with radiation or chemical mutagens in the hope that a mutation will occur that is beneficial to the plant breeder. Plants carrying the desired mutation can then be crossed with other plant varieties, using conventional breeding.

Mutagenesis produces a lot of defective plants, which likely accounts for its marginal importance in crop breeding. One textbook on plant breeding explains the high risk of mutation breeding: “Invariably, the mutagen kills some cells outright while surviving plants display a wide range of deformities.” Two plant breeding textbooks conclude that most such induced mutations are harmful and lead to unhealthy and/or infertile plants. Mutation breeding of food crops snuck under the radar of regulators before people understood the risks.

Significantly, an expert committee of the US National Research Council concluded that genetic engineering was more likely to cause unintended changes than all other crop development methods except mutation breeding.

None of the three papers cited by Academics Review tested for any “adverse outcome” such as toxicity. So Academics Review’s conclusion that no such outcomes were found is disingenuous. And none of these papers provide any evidence whatsoever for Saletan’s claim that “You’re no more likely to get a toxic mutation from a GMO than from a non-GMO.”

In linking to the Academics Review page that falsely portrays mutation breeding as the same as conventional breeding, Saletan is once again misleading his readers.

How likely is a “toxic mutation” from a GMO and a non-GMO?

A number of animal feeding studies have found toxic effects from GM crops. Thanks to the suppression of research finding problems with GMOs, these studies have not been followed up. So we don’t know if the cause of the toxicity was a pesticide the GMO was grown with (e.g. Roundup), a pesticide that was built into the plant’s cells (e.g. Bt toxin), or a “toxic mutation” in the GMO. And we still have no idea how likely it is that a GMO will produce a “toxic mutation”.

In the case of virtually all commercially available GM crops, bits of DNA are cobbled together from a wide variety of sources. The most widely planted GM crop, Roundup Ready soy, contains DNA from soybeans, a plant virus, a soil bacterium, and a petunia plant.

Apart from the diverse sources of DNA used in first-generation GM plants and the mutagenic effects of the GM gene insertion, the sheer speed with which radical changes are made to our food crop plants adds to the risk. GMOs are not adequately tested for toxicity and allergenicity before being released into global food markets.

Conventional breeding, in contrast, has a long history of safe use. Breeders have drawn upon a familiar gene pool over millennia. While toxic mutant plants might theoretically be produced, they would have been selected out of the breeding programme over time. On the whole, in conventional breeding, genes remain intact and function consistently. Otherwise saved seed would not produce the same crop year upon year.

Plant breeders know which toxins to look for

Conventional breeding does not seem to produce new toxins. There is a possibility that a conventionally bred plant might produce elevated levels of toxins normally present at low levels, but plant breeders know what they are and what to look for. For example, breeders know that some potatoes can contain harmful levels of toxic glycoalkaloids, so they check for that in a new variety.

Even without access to labs, humans have found ways of testing their crops for likely toxins. The Aymara people of Bolivia, who have many centuries’ experience of potato breeding and cultivation, can assess the levels of glycoalkaloids by taste. Their judgement seems to be so accurate that they can spot glycoalkaloid content at or above toxic levels.

In contrast, with GM, new toxins can appear unexpectedly. Because GM crop developers do not know what to look for, they are not reliably screened out. The researcher Dr Arpad Pusztai was shocked when he discovered that the GM potatoes developed by Prof John Gatehouse and the biotech company Axis Genetics had toxic effects on rats. Mysteriously, the glycoalkaloid content of the GM potatoes was less than that of the non-GM parent lines. The source of the toxicity is still unknown. What’s more, it never will be known, as Pusztai’s data were confiscated in the panic that erupted in the UK scientific establishment after Pusztai’s findings were made public.

Chinese research fails to show what Saletan claims it does

Saletan’s evidence for his claim that “You’re no more likely to get a toxic mutation from a GMO than from a non-GMO” is a paper by Chinese researchers. The paper reports the results of an analysis comparing a GM rice line with a rice line produced by conventional breeding and marker assisted selection (MAS). Both rice lines were developed to resist bacterial leaf blight.

MAS is a modern biotechnology that allows researchers to select genes of interest for their breeding programmes. It can be used to speed up the development of GM or non-GM crops. MAS with conventional breeding enjoys wide support, including from organic farming bodies, with concerns mostly being focused on patent ownership.

The Chinese researchers appear to be annoyed by the public support enjoyed by MAS with conventional breeding and the contrasting antipathy towards GM. They rather waspishly note that GM remains controversial in spite of the “immense opportunities” it offers “to curtail the severe threat of food shortage”. Their paper appears to be an attempt to show that the public’s dislike of GM is unjustified.

To that end, the researchers compared a GM rice line and a line produced by conventional breeding and MAS with the parent line, focusing on the transcriptome, or gene expression pattern, of each rice line. They found that the GM rice line had 43% fewer gene expression differences than the line produced by conventional breeding and MAS, when compared with the parent line. Oddly, in view of this 43% difference, the researchers concluded that the two lines were “substantially equivalent”.

Does this paper show, as Saletan apparently believes, that “You’re no more likely to get a toxic mutation from a GMO than from a non-GMO”?

No, it does not.

Differentially expressed genes do not equal mutations. Nor are differences in expression always caused by mutations. “Expression” in the case of the Chinese paper was only a measure of the first step in making a protein from the gene, a step called transcription. The transcriptome is the population of RNA molecules produced by a genome at any given time. The transcriptome responds as a network. Changes in gene expression can be caused by mutations or by non-mutational factors, such as epigenetic influences (which switch genes on and off and affect how cells read genes). In the case of epigenetic factors, the DNA remains unchanged in sequence but the gene expression changes.

What can the Chinese paper tell us about GMO safety? Should we be reassured by the fact that the GM rice line had 43% fewer gene expression differences than the line produced by conventional breeding and MAS? It seems not. According to Prof Jack Heinemann of the School of Biological Sciences, University of Canterbury, New Zealand, “The number of differences between transcriptomes is largely irrelevant. What those changes are is the relevant factor. It is like confusing a statistic such as: ‘The number of vehicles on one road is 10 times the number of vehicles on another, so the latter is safer’ – but not realising that the vehicles on the latter are high-speed and heavy trucks and the vehicles on the former are bicycles.”

In other words, the transcriptomic changes from conventional breeding may be like bicycles, whereas the changes that brought about from genetic engineering could be like high-speed, heavy trucks. They may move faster and do more damage.

Also, Prof Heinemann says, there is no correlation between the number and type of transcriptome changes and changes in proteins, the kinds of molecules that could lead to an unexpectedly toxic or allergenic crop: “Many times the increase in transcription of a gene results in the same or less protein and vice versa.”

So what is the value of transcriptomics data? Prof Heinemann explains: “The value lies in telling us what differences there are, not necessarily in telling us how many differences there are.”

But that’s just the beginning of the process. Prof Heinemann continues: “Then proper risk assessment science is informed by the actual differences and can investigate whether newly expressed genes, or genes expressed at higher or lower levels, results in a hazard.”

That’s a journey that the Chinese researchers do not embark upon. Surprisingly, few researchers do. And crucially, they do not address the question of toxicity.

As a revealing footnote, the Chinese researchers found that the “differences in gene expression” in the MAS line did not result in any adverse outcomes and it performed as well as the GM. So perhaps Saletan can tell us: why bother with GM?

Here, in sum, is what we can say about the Chinese paper:
1. It doesn’t look at mutations, but gene expression
2. It doesn’t look at toxicity
3. No predictions about whether a newly developed crop is likely to be toxic can be made on the basis of the number of changes in gene expression in the new crop as compared with the parent crop
4. The conventionally bred rice with MAS performed as well as the GM.

In other words, the paper has nothing to do with what Saletan thinks it’s about – and it provides evidence for the lack of need for GM.

Saletan is not to blame for his ignorance on this technical topic. I am no more an expert in molecular biology than he is. I have to consult scientists and try to translate what they tell me into lay language, and I suspect he has to go through the same process. But what is reprehensible is Saletan’s lack of critical attitude to what his sources tell him. He appears to have consulted people who are trying to pull the wool over the public’s eyes rather than remove it – and has simply accepted what they tell him at face value. Slate readers deserve better than this bad science.