How do genetically engineered foods compare with conventional foods in nutritional quality? This is an important issue, and one for which there will probably be much research in the future, as crops that are engineered specifically for improved nutritional quality are marketed. However, the potential for engineering nutritional traits has so far taken a back seat to goals such as pest resistance and herbicide tolerance. Thus, there have been only a few studies to date comparing the nutritional quality of genetically modified foods to their unmodified counterparts.

The central question for GE crops that are currently available is whether plant breeders have accidentally changed the nutritional components that we associate with conventional cultivars of a crop. The isoflavone content of soybeans is a nutritional component that has been investigated because of its potential for preventing disease. Some isoflavones are converted by the body into phytoestrogens, which are believed to help prevent heart disease, breast cancer, and osteoporosis. Soy is added to many health foods to increase dietary isoflavones. People who eat soy products for the health benefits are interested in whether RoundupReady soybeans, which have been genetically engineered to allow them to survive being sprayed with Roundup herbicide, contain the same amount of isoflavones as conventional soybeans.

Soybean pods. Source: Mark L. Tucker, USDA/ARS

Lappé et al. (1999) in their study of herbicide-tolerant soybeans (abstract available at http://www.biotech-info.net/phytoestrogens.html) reported that the RoundupReady soybeans they studied contained 12-14% less isoflavones than unmodified soybeans. If RoundupReady soybeans have reduced isoflavones, they would be less "healthy" for us than conventional soybeans.

On the other hand, it has also been suggested that RoundupReady soybeans contain unnaturally high amounts of isoflavones, enough to cause uterine cancer in mice. (See http://www.europabio.org/pages/module_04.asp#10 for a brief discussion of this research.) If this were true, Roundup Ready soybeans would be "unhealthy" for the opposite reason: too much instead of too little of the isoflavones that we hope will prevent cancer and heart disease.

Research published by employees of Monsanto (Padgette et al., 1996; Taylor et al., 1999), the maker of Roundup, suggests that RoundupReady soybeans contain the same amounts of isoflavones that conventional soybeans do.

Who is right? The definitive study has not been done, but a comparison of available results reveals only small differences. The following table shows the results for two important isoflavones tested in three different experiments. A detailed explanation is given beneath the table. These results should be interpreted cautiously. Because environmental conditions vary from place to place and from year to year, soybeans grown at different locations in different years are likely to contain different amounts of nutrients due to natural conditions. In many cases it will not be possible to attribute changing levels of nutrients to introduced genes or to treatments during the growing season.

Explanation of the table: For each important isoflavone, the first line in the table shows the average amount present in a conventional cultivar and in the genetically engineered "twin" containing the RoundupReady (RR) trait. The second line in the table, shaded green, shows how much the RoundupReady cultivar differed from the conventional cultivar. A value of -3 indicates that the RoundupReady cultivar averaged 3 micrograms less of the isoflavone than the conventional cultivar, while a value of +61 indicates that the RoundupReady cultivar averaged 61 micrograms more of the isoflavone. The table shows that the RoundupReady cultivar had higher levels of isoflavone in some experiments and lower levels in other experiments. The third line in the table shows the lowest and highest values of isoflavone measured in the crop in that experiment. The fourth line, shaded in green, shows the numerical spread from high to low values, calculated by subtracting the low value from the high value in the third line. A comparison of the second line with the fourth line gives an idea of how much the conventional and RoundupReady cultivars vary in the context of how much the individual plants of each cultivar vary from one another.

A comparison of the two experiments done by Monsanto in 1992 and 1993 is an example of the kind of variation in isoflavone levels that can occur from year to year. Monsanto used the same conventional and RoundupReady varieties in both years. Both varieties had lower levels of genistein and daidzein in 1993 than in 1992. This is not an unusual result. Wine afficionados know that the weather can influence the quality of grapes, causing "good" years and "bad" years for wine.

Monsanto's researchers did not spray the RoundupReady soybeans in 1992 and found essentially the same levels of isoflavones in the conventional and RoundupReady lines. In the context of results that varied by about 1,000 micrograms from low to high values, an average drop of 3 or 13 micrograms is essentially zero change. However this experiment did not show how much isoflavone the RoundupReady plants would produce when they were grown under the intended conditions, that is, when they were sprayed with Roundup.

Monsanto's researchers did spray the RoundupReady plants in 1993 and saw a small increase in the level of both important isoflavones in the sprayed plants. As in 1992, the isoflavone levels vary over nearly 1,000 micrograms from lowest sample to highest sample in each cultivar, so average increases of 57 and 61 micrograms in the RoundupReady seeds are quite small in relation to the variation normally seen from sample to sample.

In Monsanto's experiments, the conventional and genetically engineered cultivars were grown side by side in the field, so the plants received very nearly the same conditions of irrigation and weather. This is what researchers try to achieve when they compare the results from two cultivars.

The experiments done by Lappé et al. also compared a conventional cultivar with its genetically engineered "twin," but did not compare cultivars grown side by side in the same field. The seeds were ordered from a commercial seed provider and may have been grown by different farmers. The researchers assumed, although they did not know for certain, that the RoundupReady plants had been sprayed with Roundup during seed production, thus providing a test of how Roundup application affects the isoflavone levels in the seeds. The experiments done by Lappé et al. tested two different pairs of cultivars, while the Monsanto experiments tested only one pair.

The Lappé study found a much smaller variation in isoflavone levels from sample to sample than the Monsanto studies showed. The Lappé study found reduced levels of both isoflavones in the RoundupReady cultivars. The reductions for genistein are more extreme than the changes found by the Monsanto studies, but these reductions did not exceed the range of variation in either cultivar. Additional comments by Lappé and his co-workers on their results and Monsanto's results are available at http://www.biotech-info.net/cetos_response.html.

The USDA and the University of Missouri (Duke et al., 2003) also studied the levels of diadzein and genistein in RoundupReady soybeans that were sprayed with Roundup, sprayed with other herbicides, or weeded by hand. For one soybean cultivar in Mississippi, soybeans that were sprayed twice with Roundup had higher levels of diadzein than soybeans that were weeded by hand. The row showing the different levels of diadzein found in these treatments is highlighted in the table below. The amounts shown are in micrograms of isoflavone per gram of plant tissue. The levels of genistein showed no differences. A second soybean cultivar in Missouri showed no differences in levels of diadzein or genistein between plants sprayed with Roundup or weeded by hand. However, there may be questions about whether these results settle the issue. In both locations, the hand-weeded plants showed traces of Roundup, probably the result of spray drifting from nearby plots. Ideally, the hand-weeded plots would be entirely free from the treatment that is being tested.

Another study involving herbicide use on beans is often cited as evidence that Roundup may affect the level of isoflavones in seeds. This study (Sandermann and Wellmann, 1988) was done on Phaseolus vulgaris (string beans or kidney beans), a different species from soybeans, which are Glycine max. The beans were not genetically engineered to be tolerant to Roundup. Indeed, no RoundupReady plants were available commercially in 1988. Few details of the experiment were provided, but Sandermann and Wellmann appear to have sprayed the plants with Roundup and then measured isoflavones before the plants died. They reported that they found increased levels of isoflavones in all plant tissues, "leaves as well as fruit", that were exposed to Roundup. They did not show data comparing the isoflavone levels of treated and untreated plants. While interesting as an indicator of issues that warrant further investigation, the Sandermann and Wellman study is not useful as evidence of what actually occurs in RoundupReady soybeans.

The studies completed so far seem to indicate that isoflavone levels in soybeans change in response to several factors. The application of Roundup may have an effect, but it is unclear whether the resulting change is an increase or a decrease in isoflavone levels. The magnitude of the change appears to be small or moderate in comparison with natural variation in isoflavone levels. Because isoflavones may be an important component in the effort to prevent disease via diet, further research on therapeutic levels of isoflavones in the human diet, the levels of isoflavones in soybeans, and the effect of Roundup on these levels is warranted. Additional evidence may clarify the arguments for and against Roundup applications as a risk factor in soybean cultivation.

Other than the issue of isoflavone levels, the nutritional value of RoundupReady soybeans seem to be the same as that of conventional soybeans. Experiments with rats, chickens, catfish and dairy cows (Hammond et al., 1996) indicate that animals eat the same amount of food and gain weight whether the soybean diet is conventional or transgenic. This is important information because many farmers use soybeans as a source of protein for their animals.

Industry studies submitted in support of applications for permission to sell transgenic crops indicate that the nutritional components that are commonly tested are similar in transgenic foods and conventional foods.

For criticism of the research that has been done thus far on the nutritional value of transgenic foods, see the article by Arpad Pusztai, one of the figures in a controversial episode concerning the nutritional value of GM potatoes (http://www.actionbioscience.org/biotech/pusztai.html).