Is Transformation Really that Different from Backcrossing?
GUEST POST: Benjamin Edge
Benjamin Edge (@edgeben) is a former wheat breeder for Pioneer Hi-Bred, International, a DuPont Company, and for Clemson University. He has released 10 PVP protected wheat varieties and is a co-inventor of record for 5 wheat variety patents. He has taught classes in plant breeding, biology, and computer technology.
Transformation, the insertion of genes into an organism through the use of a ‘gene gun’ or a bacterial vector, is a tool used by plant breeders to introduce new traits to a crop when there is not enough readily useful variation present in the crop they are trying to improve. Transformation results in what we commonly refer to as genetically modified organisms, or GMOs. While some consider this a risky technology, transformation is actually very similar in effect to what conventional breeders do when they find a gene of interest in a wild relative, and use backcrossing to incorporate that gene into an adapted variety.
Backcrossing is a VERY effective tool of conventional plant breeders (Briggs and Knowles, 1977). Once you find a trait you are interested in, you can move that trait from a wild relative (closely related species) or from any member of the species you are working with into an adapted variety with great repeatability (reproduced or repeated easily). Backcrossing is used when you have a well adapted variety, say plant A, with high yield, large seeds, and strong stems, but with some weakness, such as susceptibility to a disease. If you find a plant, say plant B, with disease resistance, but poor yield, small seeds, and weak stems, you can use backcrossing to incorporate that disease resistance trait into plant A, what we call introgression of the trait.
With backcrossing, we call plant A the recurrent parent, because it will be used repeatedly (recurrently) in a series of crosses. We call plant B the donor parent, as it will donate the trait of interest (in this case disease resistance) to the recurrent parent. If we just cross the donor to the adapted recurrent parent once, we get half the genes of the new progeny coming from the donor and half from the recurrent parent. This is not likely to be a desirable outcome, because the donor may have many undesirable qualities or traits. We could let the progeny segregate (producing plants with various combinations of genes possessed by each of the parents), and select out individual plants from the resulting progeny that have better combinations of those genes. Theoretically, it would be possible to select a single plant that had all its genes from the adapted (recurrent) parent and just the one gene of interest from the donor. But the odds of that occurring are very remote. The two parents may vary for thousands of genes, and if they differ for n genes, the number of possible gene combinations that could occur would be 2n. It would not take very many different genes before it would be practically impossible to grow out enough plants to find that ‘one in a billion’ combination you want. If two plants differ for only twenty genes, there would be over one million possible combinations (220 = 1,048,576) of genes in the segregating progeny.
Another problem is that linkage (where two genes are located close to each other on the same chromosome) reduces the chance that all possible gene combinations will occur. So when you cross parent A with that wild donor line, parent B, any plant you select that has the trait you are interested in probably carries many undesirable genes on either side of the trait you want. These flanking genes are said to be linked to the trait, and segregate (separation and random mixing of chromosomes during meiosis) with it, rather than segregating independently of each other. The closer the genes are to the gene of interest, the harder it can be to get them to segregate independently (what we call “breaking the linkage”). These additional linked genes may cause yield drag, disease susceptibility, poor food quality, or poor agronomic characteristics.
In backcrossing, you take the original F1 progeny of your cross (AxB, or AB), before they segregate, and cross them back (backcross) to the adapted recurrent parent line A (it is recurring, or repeating as the parent). Then you take the progeny from that cross and cross it back to the recurrent parent again, and repeat, as many times as needed. Each time you backcross like this, you reduce the genetic contribution from the donor parent by half. Eventually, after 7 backcrosses or so, you can recover the recurrent parent type along with the trait of interest with maybe just a few genes from the donor parent coming along for the ride. But it can be extremely difficult to eliminate ALL the other genes from the donor parent because of linkage. After 7 backcrosses, you can recover 99.6% of the recurrent parent, absent the effects of linkage. After 13 backcrosses, you recover 99.99% of the recurrent parent, absent linkage.
As you get closer to 100% recurrent parent, you can see that the rate of progress slows down (Figure 1). If the linked genes are located very close to the gene for the trait you are interested in from parent B, it can be almost impossible to recover the donor gene without the sometimes bad flanking genes. That is one of the few drawbacks of backcrossing. The main weakness of backcrossing is that the performance of the end product is limited by the performance of the recurrent parent. But, if you have a well adapted, popular variety, with few poor qualities, backcrossing can fix a weakness without having to start from scratch.
Percentage of donor parent in a backcross is reduced with each cycle of
backcrossing and the percentage of recurrent parent increases each cycle.
Recurrent parent is 99.6% of genetic composition by BC7 and 99.99% by BC12
Backcrossing is one of the few tools available to breeders where two breeders could start with the same two parental lines and come out with two finished varieties that are essentially the same. We say that backcrossing is both repeatable and predictable. You can make the same backcross several different times and get essentially the same results (repeatability), and you know from the start what the results should be, the recurrent parent with an added gene (predictability).
Transformation, on the other hand, skips directly to inserting the single gene you are interested in, with no linkage effects.
So you could say that transformation is essentially a glorified backcrossing method as far as its effect, but with much greater precision of only introgressing (introducing) the trait you are interested in. In practice, once a transformation event is deregulated, or approved, the process of introgressing the gene into other elite lines is done through backcrossing combined with another biotechnology tool called marker assisted selection (MAS).
When breeders using transformation backcross the transformed line to other varieties, they use the same backcrossing technique that conventional breeders use (usually with only 3-4 backcrosses), because it is such a powerful breeding tool. With MAS, using a gene linked to the gene of interest as a means of tracking it, breeders can be fairly certain to only transfer the specific gene for the trait (along with the marker) to other lines. But even if a few genes from the transformed parent are linked at this point, they are from an adapted variety, so any negative effects of those genes are likely to be small.
Briggs, F.N. and P.F. Knowles. Introduction to Plant Breeding. 1977. Reinhold Publishing Corporation.