An acquaintance of mine, on a message board, recently played the Frankenstein gambit in a discussion about the politicization of science. Here’s his quote (modified slightly to improve readability):
You seem to be ignoring the Frankenstein aspect of genetically modified crops in that genes are being inserted that are entirely alien to the organism…the kind of mutation that would never occur in a natural environment. Yes, it’s the point of GMO, a pretty powerful technology that has been harnessed to this point successfully…what might a failure in this technology look like?
At some point before or after this comment, my acquaintance expressed a preference for hybridization to generate new crop varieties, rather than transgenic technology, and also argued that “tampering with nature can be very wrong.” This appeal to nature, essentially the converse of the Frankenstein gambit, seems to be a powerful (if fallacious) argument that can be applied to any new technology (including, at one time, hybridization). Indeed, this appeal is so powerful that many pseudoscientific websites adopt appeals to nature in their very URL (such as Natural News, RawForBeauty, and so on). It’s also why companies are falling all over themselves to get the word “natural” into their product names.
In any event, my main objection to the Frankenstein gambit is not so much the appeal to nature (grating as that is), but rather its reliance on a very superficial understanding of genes. That superficiality is betrayed by the comment that transgenic technology requires the insertion of genes which are “entirely alien to the organism”. Unfortunately, scientists compound this problem.
Genes Make Proteins, Not Organisms
Consider golden rice. This remarkable transgenic crop will save millions of lives by providing a staple food with beta carotene, a nutrient our bodies convert into vital Vitamin A. The original incarnation of golden rice borrowed a gene from the daffodil and a gene from a bacterium in order to alter the nutritive characteristics of rice.
An unfortunate and misleading way of describing this would be to say that scientists put a bacteria gene and a daffodil gene into rice (reread my original description and note the different semantics). This way of phrasing invites the Frankenstein image: a monstrous rice, cobbled together from bits of bacteria and bits of daffodil. Who wants to eat bacteria? (Never mind that we shovel trillions of bacteria down our throats every day.) Who wants to eat daffodil? Yuck!
By calling a gene a daffodil gene, we imply that the gene’s job is to make a daffodil. We imply that the rice now has daffodil-like qualities. But that’s not at all what happens.
Consider an analogy. Jack is a contractor, who’s been hired to build an elementary school. Naturally, this requires purchasing a lot of raw material – bricks, wood, drywall, insulation, pipes, paint, wire, fixtures, tile, shingles, nails, screws, and so on. After the job is done, he has a few hundred bricks left over. He is contracted next to build a private residence for Tom, a home-owner, and he says to Tom, “You’re in luck, I happen to still have a few hundred bricks from the school project, so I’m willing to offer you a discount.” Tom, revolted, complains, “Those are school bricks, Jack. I’m asking you to build me a house. I would go crazy living in a school.”
Genes don’t make organisms, any more than bricks are limited to making a particular kind of structure. Genes make proteins. In fact, they don’t make proteins in isolation; they contain a recipe for making a protein (in the form of the genetic code) in the presence of cellular machinery capable of converting the gene’s instructions into a working protein. This requires the participation of enzymes for transcribing the gene into messenger RNA, organelles like ribosomes for anchoring the growing protein, nucleic acids (in the form of transfer RNAs) to deliver the amino acids to the protein, and any number of additional enzymes for facilitating the process and shaping the final result. Amazingly, a human cell is capable of faithfully (or nearly faithfully) reading the instructions of genes borrowed from virtually any other life form on earth. The implication of this astonishing fact is that genes aren’t proprietary to particular organisms – we share far, far more in common with the source of any “alien” DNA than most people realize.
Rather than refer to one of golden rice’s transgenes as a daffodil gene, it would be more precise to refer to it as a phytoene synthase gene.
The gene doesn’t make daffodils, or the mysterious essence of daffodils: it makes (in the right cellular environment) a protein called phytoene synthase. The protein gets its name from its ability to synthesize phytoene from precursor molecules. In the right environment, this molecule will be further processed to beta carotene. Amazingly, the rice grain is just such an environment. Even though the rice grain lacks phytoene synthase, it contains all of the other enzymes required to make beta carotene (except phytoene desaturase, which is why two transgenes are required for golden rice).
Is there any justification for calling this a daffodil gene rather than a phytoene synthase gene? It wouldn’t seem so. Bricks aren’t just used to build schools, and neither is phytoene synthase just used to build daffodils. This enzyme is found in a myriad organisms; this table shows a partial list.
In fact, today’s golden rice doesn’t make use of the gene from the daffodil. Instead it uses a gene from corn. Both daffodil and corn make beta carotene. Vitamin A is Vitamin A, whether it comes from eating daffodil (not recommended), corn, golden rice, or regular rice sprinkled with vitamin A from a multivitamin pill. The gene from the corn causes rice to make more than 20 times as much beta carotene as the variety using the gene from the daffodil, which is why it was used. How can this be? I presume the genes are slightly different: even though both proteins allow the synthesis of phytoene, the corn’s protein must be slightly different, such that the reaction occurs faster in corn than daffodil.
Don’t let this fact trouble you. The bricks Jack ordered for the school might differ slightly from the bricks he would have ordered for a private home, but that doesn’t necessarily make them incompatible. It might even make them slightly better, like when they overbook your flight and offer to let you sit in First Class rather than Coach. In any event this subtle variation in gene products brings up another important issue: how evolution works.
Evolution Isn’t Directed To A Purpose
My correspondent quoted above makes another interesting assertion. He says that a gene dropped into rice from a daffodil (for example) is “the kind of mutation that would never occur in a natural environment.” On one hand he may simply be saying that daffodils and rice can’t have sex with one another to “naturally” mix their DNA. But if we read his statement literally, he seems to be suggesting that daffodils and rice are so dissimilar – so alien to each other – that you’d never have rice arise with the ability to make beta carotene through natural means.
But that’s just wrong. In fact, rice does make beta carotene. It just doesn’t do so in the grain part of the plant that we eat. Furthermore, rice does make other enzymes required to synthesize beta carotene in the grain, suggesting that a relatively minor mutation could reinstate the synthetic pathway. It may even be that minor mutations caused the rice to lose the ability it once had.
Humans, for example, don’t make the enzymes required to make Vitamin C, though some fairly closely related organisms can (see figure; black lines are Vitamin C producing animals and gray lines are Vitamin C requiring animals). Given that Vitamin C is absolutely vital to survival (it is estimated that about 66 times as many British naval personnel died in the Seven Years War from Vitamin C deficiency – scurvy – than those who died in battle), it is difficult to see that loss as being adaptive. Ask anybody at the risk of dying of scurvy if they wouldn’t like to borrow a gene from a rat or a lemur or a cat or a rabbit that permits their body to synthesize this life-saving substance. The fact that human ancestors had this ability, but modern humans don’t, is an accident of evolution that was possible only because humans evolved the ability to eat a diet rich in Vitamin C. This is probably the only reason that the failure to make endogenous Vitamin C doesn’t effect one’s genetic fitness – or rather didn’t, until man invented sailing ships that could stay at sea for weeks at a time.
Thus, also evolution generally enhances the genetic fitness of a species over generations, it does not always provide individuals with the ideal genetic complement. And, even if it does, changes in the environment or in the milieu of a species can undermine that adaptability. Thus, when young men started sailing during the age of exploration, a new vulnerability – scurvy – became apparent. Thus, when mankind became agrarian and discovered that rice was the easiest, least expensive, most reliable crop in certain parts of the world, Vitamin A deficiency suddenly became a killer. Whereas my friend views the current human genome as something sacrosanct, not to be tampered with, I view it as a halfway decent compromise, generated by a trial and error mechanism which, while a solid foundation, could stand quite a bit of improvement.
How does evolution work? Evolution is a 3-step process. First, you must have variation. Variation includes both different alleles (versions) of the same gene across members of a population, and also includes the emergence of spontaneous mutations through copying errors in the DNA. Next, you must have selection. In the natural world, selection is, essentially, early death: those most-adapted to an environment are more likely to survive than those poorest-adapted. Third, you must have inheritance: children must resemble their parents. Thus, the survivors pass on their genes to the next generation, whereas those who experienced an early death, do not.
Evolution doesn’t just work in the natural world. It works everywhere those 3 factors exist. Why were there no reality TV shows 30 years ago, whereas now every other show is a reality show? Because there was variation – in this case, innovation – reality shows probably starting with MTV’s The Real World began to be added to the variety of TV shows available. Then there was selection – the reality shows attracted viewers, made money, and thrived. Then there was inheritance – the reality shows were renewed for additional seasons, and produced spin offs and copies – and you had an evolution of TV programs. One can tell a similar story about how SUVs suddenly came to dominate American roadways, how certain funny videos go viral, or how certain breeds of dog become popular.
However – and this is a key point – selection can only work on the varieties currently in existence. Consider another analogy. Jenny and I are coauthoring a magazine article together. We decide that Jenny will write the first draft, and I will edit her draft and add my own comments to it. It must be obvious that this very well could result in a very different article than if I wrote the first draft and she edited it. My job will be to select and shape her ideas into something better – but this is a very different process than producing a first draft myself. Evolution works in this way – selection is a wonderful mechanism for increasing the fitness of organisms, but selection can only work on the genes currently in the genome. This is why I say that I view the human genome as having been arrived at through a bit of a slapdash process. We happen to have these particular genes, I suppose we might call them “human genes”, but I don’t attach too much importance to that. Our lack of having certain genes was, in some cases, merely an accident of they way things played out In addition: 1) Many other organisms have many of those same genes, and 2) We have the ability to use other organisms’ genes just fine, as they may use ours just fine. I elaborate briefly each of these points below.
Human Genes (If You Must Call Them That) Are Found Naturally In Other Species As Well
Earlier I pointed out that a good number of plant species have the capacity to produce phytoene synthase, so much so that it was a bit pointless to call the phytoene synthase gene from the daffodil a “daffodil gene”. The same can be said for the human genome as well. Because of evolution, useful variations which cropped up in our ancestors hundreds of millions of years ago are still with us today – and still with many of our evolutionary cousins such as lemurs, or foxes, or pigeons, or squid. Because genes are composed of hundreds or thousands of codons (sets of base pairs that specify amino acid constituents of proteins), the genes are often slightly different in one species or another, but that’s also true comparing one human to another.
But we have so much in common with other animals (indeed, this is why neuroscientists often study other animals – not to learn about monkeys or rats or worms or fruit flies, but to learn about ourselves). GABA, for example, is one of the most important neurotransmitters in the human brain – but you’ll also find it in the nervous systems of most animals. Why? Because many species make the enzyme that converts glutamate into GABA – because we all tend to have the gene that specifies the recipe of a protein that will serve this function. In fact any time you hear about some molecule that’s found in both humans and animals, it is very likely that some gene is responsible for making that molecule, or that some gene is responsible for making an enzyme that catalyzes the synthesis of that molecule.
Thus it is possible to say that we share 98% of our genes with chimpanzees, 85% of our genes with the zebra fish, and 21% of our genes with roundworms.
Consider the discovery that insulin injections can alleviate the symptoms of diabetes. Insulin is a very large molecule and thus not practical to synthesize in a laboratory, so originally the insulin was harvested from the pancreas of cows and pigs. The reason this is even possible is because cows and pigs make insulin too (as do any number of species), and although there may be slight differences in the cow, pig, and human genes and therefore the cow, pig, and human insulin peptide, our cells respond to this animal insulin sufficiently enough that this “alien” insulin saved people’s lives.
I imagine there are any number of people who are squeamish at the thought of genetic modification but are perfectly okay with the idea of injecting cow insulin to treat diabetes. It is difficult to understand why. Surely grinding up cow pancreas, purifying the juice, and injecting it into a human is “unnatural”. Yet we recognize that many useful drugs can be derived from not only animals but plants, and we are used to the idea of putting these foreign substances in our bodies. Why play the Frankenstein gambit when it comes to GMOs?
We Have The Ability To Use The Genes Of Other Species (And Vice Versa)
But that’s not the best part of the insulin story. The best part of the insulin story is that we no longer have to grind up the pancreas of farm animals to supply the diabetics of the world with insulin. If you think carefully about where that insulin is coming from, the solution is obvious. A cow’s pancreas (like ours) contains cells whose job it is to release insulin when the organism eats a caloric meal. (The insulin is a signal to cells all over the body to prepare for the coming rush of energy.) Because the pancreas has to release insulin every meal (3 or more times a day in us, and awful lot more than that in a grazing herbivore like the cow), those cells have to keep making insulin or they will run out. Each time the cells have to make insulin, the insulin gene in the cow’s DNA is copied to a messenger RNA strand, and that messenger RNA’s genetic code for the insulin peptide is “read”, line by line, by transfer RNA molecules that build the growing molecule.
Point is, you don’t need a whole cow to make insulin. You just need certain cells in its pancreas. But actually, you don’t even need that. You just need the gene, plus the cellular machinery for making proteins. Once you have that, you have an insulin factory. Because all organisms share so much in common, the cellular machinery for making proteins exists in pretty much the same form in pretty much any living cell on the planet.
The solution? Rather than kill a bunch of cows, take a copy the human version of the gene for making insulin, and place it into a yeast cell or an E. Coli bacterium. Not only have you made an insulin factory, but you’ve made one that self-replicates. Just feed it, come back tomorrow, and you’ve got thousands of insulin factories. Not only that, but they’re cranking out pure insulin, reducing the possibilities of impurities (from the cow or pig) causing allergic reactions in the diabetic patients.
There’s another name for this insulin factory. It’s a genetically modified organism. (Shhh!)
The Last Question
The only thing I haven’t responded to in my acquaintance’s post is “What might a failure in this technology look like?” My response to this question would be to point out the kinds of failures that won’t happen – failures imagined by the people who don’t understand the technology:
- You won’t be altered in any odd way by eating a genetically modified crop. Although I’ve emphasized the remarkable fact that E. Coli can read our genes, and therefore that we can read the genes of, say, some herbicide-tolerant ear of corn, that only occurs if you put the genes into the nucleus of some cell. Everything we eat contains DNA (except salt), but none of that DNA – “natural” or “modified” – can be incorporated into our cells. We enzymatically break down proteins and nucleic acids before we absorb them. In some cases that’s unfortunate – if someone eating golden rice suddenly gained the ability to make their own Vitamin A, that might not be a bad thing!
- You won’t gain some strange allergy to the genetically modified food you eat – though I wouldn’t be so sure about new varieties that appear from hybridization. Genetically modified crops undergo safety testing – something not required of more “natural” means of creating new varieties. Yet there is no principled reason to require it in one case but not the other.
- We won’t experience some collapse of the food supply because GM crops become so popular we end up with a monoculture problem. There’s nothing inherently different about GM crops from other specialty crops, so such concerns aren’t unique to GM. And in many cases GM crops are adding needed variety – as, for example, in the case of the banana.
- We won’t unleash some kind of superplant or superfish that destroys the ecosystem. The Frankenstein gambit also invites the notion that we are creating Wrath of Khan type super beings that can out-compete their natural rivals. But any farmer can tell you how difficult it is to maintain a thriving farm – our crops survive not because they’ve been bred to be evolutionary winners; they survive because farmers provide them with round the clock tender loving care. Not to mention water, nutrients, pesticides, and, I suppose, scarecrows. Varieties selected or modified to better tolerate farm conditions are unlikely to be especially suited for spreading outside the farm.
- We won’t destroy natural plants through inadvertent cross-breeding with modified plants. Surely some plants will escape the farm, but a GM crop is so minutely different from the standard crops (one or a few genes’ difference) that you wouldn’t be able to distinguish an all-natural variety from a cross-breed. Nor is it likely that the new genes will significantly spread through the native populations.
- We won’t unleash some environmental destruction by creating crops that can withstand pesticide application or which endogenously produces pesticides. On the whole, GM crops significantly improve the environment. Crops that make their own pesticides don’t have to be sprayed, minimizing leakage of pesticides off-farm. Decreased use of pesticides will also minimize crop dusting and use of farm machinery and therefore usage of fossil fuels. As better-able to tolerate the farm environment, more crop can be grown on less farmland, minimizing the encroachment of farmland into native forests. All of these benefits have already been documented.
What fears are left? I don’t know, I can’t get into that head-space very easily. The failures of the technology which are the most plausible are basically the same kinds of problems that occur with all modern agriculture. Growing lots of food puts a stress on the environment. On the whole, modern ag is a godsend to human comfort and longevity, but as with most things that cause major benefits, the industry comes with major trade-offs. The point is that GM doesn’t add anything new to those trade-offs and, if anything, begins to minimize some of the problems that already exist.
It’s been a long time since I read Frankenstein, but wasn’t one of the morals to that story that the monster wasn’t so bad – it was the people’s reaction to the monster that was regrettable? Well, if not, that’s how I would have written it.
P.S. Bonus points to those of you who know that The Last Question is this post’s Asimov Easter Egg. Double bonus points if you read this whole damn article.