Biofortified » Cody Cobb

Escape! Crop-Specific Gene Flow to Wild Relatives

Cody Cobb — Wed, 24 Feb 2010 01:42:45 +0000

As a molecular biologist, most of my work is done on a bench at or below room temperature. I can count on one hand the number of times I’ve been to a research field because I have more than two fingers. I’ve never taken a course in ecology, and I’ve rarely dealt with full, intact organisms. It is with just such a background that I absorbed a talk by Allison Snow at Rutgers ten days ago.

Snow* is an evolutionary biologist and ecologist who’s been running an interesting experiment on wild radishes for more than a decade now. In the 90′s, when transgenic crops like Bt corn and Roundup Ready soybean were beginning to dominant the market (and the landscape), there were concerns that wild relatives would incorporate the transgenes and spread as superweeds. Corn and soybean, with their lack of compatible relatives in the US, are exempt from this concern. However, as more and more transgenic crops with compatible relatives come down the pipeline (and with some, like canola, already here) there needs to be some hard data on just how easily transgenes can persist and spread in wild populations.

Wild Radish. USDA-NRCS PLANTS Database.

Domestic radish (Raphanus sativus), like nearly all domesticated crops, differs profoundly from its wild relative (Raphanus raphanistrum). The traits that make for a delicious ingredient in a salad often make for a wimpy competitor in the wild. Humans have been cultivating radishes for so long that some alleles are only found in the domestic varieties. In this 10 year long experiment, these crop-specific alleles served as surrogate transgenes in the sense that their presence in wild relatives had to have been the result of successful hybridization of crop and weed. Snow and collaborators chose as their genetic markers two allozymes, glucose-6-phosphate isomerase (GPI) and phosphoglucomutase (PGM), and the gene for petal color (R. sativus has white petals, the dominant allele, and R. raphanistrum has yellow petals, the recessive allele). Even though it’s typical for domestic traits to have a negative impact on fitness in the wild, no a priori assumptions were made about these specific alleles.

Allele and phenotype frequencies for four populations over a decade. Figure from Snow et al., 2010. Click to enlarge.

Four plots, each containing 100 wild radish plants and 100 wild/domestic F1 hybrids, were set up and more or less left to the devices of nature for ten years. Allele frequency at the start was 25% for all three markers, with 50% of all plants white-flowered owing to dominance. By the end of experiment, the percentage of plants with white flowers had dropped to 3-15%. Crop-specific GPI ranged from 5-12%, and crop-specific PGM declined the least with an allele frequency ranging 16-26%. As it turns out, the allele for white petals is linked to delayed flowering, a deleterious trait which seems to explain the precipitous drop in white-flowered plants in the second and subsequent years. Despite this selection pressure, the white flower allele persisted in all populations. The GPI crop allele behaved similarly, declining in frequency but never disappearing. The PGM crop allele was a little different, declining in frequency in only one population and remaining more or less the same in the other three.

Pollen fertility levels. Figure from Snow et al., 2010.

What’s more, the first generation of F1 hybrids suffered from a significant disadvantage: only around 60% of their pollen was viable, compared to 80-95% viable pollen from the neighboring R. raphanistrum. And yet, despite this initial setback, each population eventually regained normal pollen fertility levels (> 70% fertile) while still retaining low-but-not-zero levels of crop-specific alleles.

So what kind of effects in the wild can we expect to see from transgenes based on this study? From the paper’s conclusion we’re warned:

Clearly, crop alleles can persist for many generations following a single hybridization event, and crop-wild hybrids may recover wild-type fitness in later generations. Thus, beneficial or neutral transgenes that recombine independently of deleterious crop alleles may spread and persist indefinitely (Snow et al., 2010).

A relevant example is her 2003 study on Bt sunflower, which found that the Bt transgene in cultivated sunflower (Helianthus annuus), when crossed into wild sunflower (also Helianthus annuus), allows each plant to produce, on average, 55% more seeds relative to non-transgenic controls under field conditions (Snow et al., 2003). Rather frustratingly, follow-up work was halted when the companies sponsoring the study — Pioneer Hi-Bred International and Dow AgroSciences — refused to allow further access to the transgene or the seeds since they decided not to sell Bt sunflowers anyway.

Left: Cody Cobb. Right: Allison Snow. Bottom: Jean Marie Hartman's thumb.

With all this in mind, what are some steps we genetic manipulators and tamperers can take to lower the risk of transgene flow into wild relatives? One thought is to link the transgene of interest with another gene that’s deleterious in the wild but tolerated or even desirable in agricultural situations. Better yet, find two such genes and flank the transgene. Creating such a construct would require a lot more work, not to mention the difficulty of finding appropriate crop-tolerant-but-wild-harmful genes. But then, it’s just a thought. What are yours?

*Astute readers of Tomorrow’s Table might recognize the name from a parenthetical citation on page 110: “For this reason, some ecologists see the application of GE as a way to spare even more land from destruction by enhancing yields (Qaim and Zilberman 2003; Snow et al. 2005).”

Snow, A., Pilson, D., Rieseberg, L., Paulsen, M., Pleskac, N., Reagon, M., Wolf, D., & Selbo, S. (2003). A Bt transgene reduces herbivory and enhances fecundity in wild sunflowers. Ecological Applications, 13 (2), 279-286 DOI: 10.1890/1051-0761(2003)013[0279:ABTRHA]2.0.CO;2

Snow AA, Culley TM, Campbell LG, Sweeney PM, Hegde SG, & Ellstrand NC (2010). Long-term persistence of crop alleles in weedy populations of wild radish (Raphanus raphanistrum). The New phytologist PMID: 20122132

Plastid Engineering

Cody Cobb — Tue, 03 Nov 2009 15:40:08 +0000

Long ago – before you or anyone in your family photo albums were born – a small, unassuming cyanobacterium was busy being engulfed by another cell. The engulfing cell’s intentions were most likely along the lines of “Yum, food!”, but lucky for us the cyanobacterium was not consumed. Instead, it stayed there, establishing a new home inside the confines of its voracious captor. We now know this happy accident was a momentous first step towards a greener, more botanical planet, because our little cyanobacterium was the photosynthetic ancestor to that most remarkable organelle: the chloroplast.

(By law, any discussion of chloroplast origins compels me to mention the similar origin of the mitochondrion. With those requirements now met, let us now continue.)

The focus of this post will be more technological than biological, but there are a few basic facts we need to get out of the way before we can proceed. Briefly:

• Chloroplasts, along with leucoplasts, proteinoplasts, elaioplasts, amyloplasts, statoliths, and chromoplasts, belong to a class of organelles known as plastids. The names of these other plastids aren’t important so long as you realize the chloroplast isn’t the only game in town. That’s why the title of this post is “Plastid Engineering” and not “Chloroplast Engineering.”

• Plastids replicate separately from their host cell, and in any given cell there can be 100 to 1,000 plastids. Moreover, plastids contain multiple copies of their genome (plastome) to the point where a single plant cell may contain 10,000 plastomes. By contrast, the nuclear genome has only one copy (this is manifestly untrue, but we’re talking orders of magnitude here).

• Plastids behave a lot like prokaryotes. Their genome is circular, their proteins aren’t glycosylated (i.e., have sugars attached to them), and they can process polycistronic mRNA (i.e., more than one protein produced from a single mRNA; most eukaryotic genes are monocistronic).

• Over history, most plastid genes have migrated into the nucleus, even though the protein produced might still accumulate in the plastid. Those proteins are instead brought back to the plastid by a specific targeting sequence. Quite a few genes have been lost from the original cyanbacterial ancestor, leaving only 50 to 200 of the original ~3,000 genes in most plastids today. In scientifically and agriculturally important species, these genes have all been sequenced and characterized.

• Plastids are inherited uniparentally, that is, from one parent and not the other. In most flowering plants, only maternal plastids are passed on. In some species, such as pine trees, paternal transmission in the pollen is the norm.

Ideally as you pored over those facts your brain started piecing together the reasons why we would want to tinker with plastid – rather than nuclear — DNA. Uniparental inheritance is a big one: even people who know next to nothing about GM crops know there’s concern about, say, GM corn in one farmer’s field contaminating non-GM corn in their neighbor’s field. Crops with genetically engineered plastids (known by the awesomely retro-sounding name transplastomics) don’t have this problem since plastids aren’t usually found in pollen. Of course plant biology is, technically, a biological science, so there are exceptions that will to be need to be addressed.

Extreme polyploidy is another attractive feature: inserting a gene of interest (GOI) into the chloroplast genome means having up to 10,000 or more copies of that gene per cell. That translates (hah!) into very high levels of protein production indeed. And since most plastid genomes are already well characterized, we can know in advance where our inserted DNA will wind up.

Non-glycosylation differs in usefulness depending on the source of the foreign gene. Plants, mammals, fungi, and insects all have different patterns of glycosylation, with plastids and prokaryotes not participating in the ritual at all. So, proteins normally present in prokaryotes are produced identically in plastids, whereas proteins of eukaryotic origin might be missing structural elements crucial to their function (or the protein might find it does just fine without those extra sugars, you never know).

So what are some limitations and problems with plastid engineering? To answer that question, we must first learn how transplastomic plants are created.

Today, only a few species have had their plastids successfully transformed. The first transplastomic organism was created in 1988 using the unicellular alga Chlamydomonas reinhardtii, notable for having only one large chloroplast. Two years later, stable tobacco transplastomics were created. Since then, varying levels of success have been achieved with potato, tomato, rapeseed, cauliflower, poplar, rice, soybean, and a few others, but only in tobacco is plastid transformation routine.

The first step in plastid transformation is introducing the new genes to the old. Typically this is done by particle bombardment (“biolistics” or the “gene gun”) or polyethylene glycol (PEG) treatment. In the latter, you remove the cell wall of a plant cell to create a protoplast and then subject it to a solution of DNA in PEG, whereas in the former you basically shoot the plant with DNA. Since particle bombardment is the more commonly used of the two, I’ll explain its mechanism.

First you need your gene of interest in a plasmid (a small circle of DNA that contains of a few genes and can be grown in and purified from bacteria). The plasmid will also contain a selectable marker (a gene that confers resistance to antibiotics like spectinomycin, streptomycin, or kanamycin) and a visual marker (green fluorescent protein or a derivative thereof). The GOI, selectable marker, and visual marker will be flanked by sequences taken from the plastid genome, carefully chosen so that the site of homologous recombination (see further reading) does not disrupt the function of normal plastid genes.

Next, the plasmids are expressed to high quantities in bacteria and purified, then adhered to small particles of tungsten or gold, often to less than a millionth of a meter in diameter. A small section of leaf tissue is placed into a low-pressure vacuum chamber and bombarded with a volley of DNA-coated particles, obliterating most of it.

A very small percentage of the remaining tissue will contain transformed plastids at this point. Worse yet, a surviving cell with a transformed plastid will still overwhelmingly contain untransformed plastids. The next steps are the lengthiest and most tedious part of the process, for now the bombarded tissue must be coaxed into regenerating into a wholly new plant while at the same time eliminating any untransformed plastids it may still harbor. Stringent antibiotic regimens are applied to emerging plantlets, and visual inspection of GFP expression reveals areas of transformed plastids. Those areas are then sliced away and grown on their own regenerative media. This process is repeated for about 20 cell divisions before a state of exclusively transformed plastids (homoplasmy) is achieved. Once reached, the plantlets are allowed to grow in the absence of antibiotic selection and set seed at maturity. If the progeny are shown to be homoplasmic, then the line is considered stably transformed.

So you’ve created a transplastomic plant. Now what? Obviously that antibiotic resistance gene is no longer doing you any good, so you’ll have to find a way to get rid of it lest it sap precious metabolic resources and stunt your plant’s growth. And just how certain are we that plastid inheritance is uniparental? What if life, as renowned chaos theorist Ian Malcolm once gravely intoned, finds a way? Shouldn’t we run a few tests to determine the likelihood of plastid-transference via pollen? And what about those really important plants, the cereals? Why are their plastids so difficult to transform?

All important questions, yes, but we’ve already reached 1,200+ words in this primer, so you’ll have to wait for subsequent posts to quench your curiosity!