Scientific Narrative: Herbicide Resistant Crops and Weed Management

Table of Contents

1. Introduction
2. Background - Terminology
3. Background - Herbicides and Weed Control
4. Background - Molecular Basis of Herbicide Toxicity and Resistance
5. Background - Developing New Herbicides and Herbicide Resistant Crops
6. Background - Impact of Herbicides on Weed Populations
7. ALS Herbicides and ALS Resistant Crops 
8. Glyphosate and Glyphosate Resistant Crops
9. Evolution of Resistance to ALS Herbicides
10. Evolution of Resistance to Glyphosate
11. Managing Herbicide Resistant Weeds with New Herbicide Resistant Crops

I. Introduction and Background

Farmers began using chemical herbicides paired with crop varieties that could resist those herbicides long before breakthroughs in plant biotechnology led to the new generation of herbicide resistant crops. In the U.S., farmers have been using synthetic chemical herbicides to control weeds ever since the first commercial herbicidal compound, 2,4-D [(2,4-dichlorophenoxy) acetic acid], was discovered in the 1940s.  The discovery of 2,4-D launched a concerted effort to discover and synthesize novel herbicides specifically for weed control, and by the late 1950s, farmers were applying herbicides to over 50 % of the corn acreage and 25% of the soybean acreage in the U.S.  By 1975, the percentage had increased to 80% for both crops and grew to more than 90% by 1980. By 1990 virtually all acres of the major commodity crops, such as corn, soybean, cotton, and small grains, as well as some horticultural crops were being treated with herbicides (Duke et al., 1991).

Crop pests respond to the repeated use of any mechanism that attempts to control them by evolving biological tactics to escape control. Not surprisingly then, the widespread use of herbicides can lead to weed populations that are no longer susceptible. The first documented case of a weed evolving resistance in response to repeated use of an herbicide occurred in the mid 1960s (Ryan, 1970). During the 1970s, farmers in the U.S. and Europe began to realize that one class of herbicides (triazines) that had successfully controlled many different weeds was no longer effective against certain populations of 30 weed species (LeBaron and McFarland, 1990; Bandeen et al., 1982). By 1990, weed scientists had evidence that at least 81 weed species contained individuals (biotypes) that had evolved resistance to one or more herbicides; 15 different classes of herbicides were no longer effective against at least one weed species (Holt and LeBaron, 1990). Currently more than 290 biotypes of herbicide resistant weeds occur around the world (Data from the international survey of herbicide resistant weeds can be found at http://www.weedscience.org/in.asp).

As more weeds became resistant to an increasing number of herbicides, agricultural scientists explored new weed management options that relied on the fundamental strategy of using herbicides in conjunction with crop varieties that are resistant to them.  Two general research approaches were available to scientists hoping to provide more options: discover new herbicides or give crop plants new genetic ability to tolerate currently marketed herbicides that were still effective against most weeds.

Discovering new herbicidal compounds, especially those with novel modes of action, had become increasingly difficult and the cost of commercialization, increasingly expensive. As a result, scientists turned their attention to the crop plant itself. They attempted to use traditional crop improvement techniques that had proven so successful in the past, such as selective breeding and mutagenesis. However, they had little success broadening the spectrum of herbicides that a crop could tolerate.

The advent of plant biotechnology in the 1980s-90s provided scientists with new methods for giving crops novel genetic capacities to survive herbicide treatments that previously would have damaged them. The new biotech crop varieties increase the number of weed management options because they allow farmers to choose from a broader array of herbicides that are already on the market.

Irrespective of the method that is used to develop them, the ecological and agronomic concerns associated with the use of herbicides and herbicide resistant crops are the same.  Those concerns include the evolution of resistant weeds, misapplication of the herbicide, herbicide drift, crop injury, carry over between growing seasons, costs, and the need for timely application. This paper focuses primarily on herbicide resistant biotech crops and their impact on weed management. Other issues associated with herbicide resistant biotech crops, such as gene flow from crops to wild relatives or food safety, are discussed in other Science Narratives on this website.

Return to Table of Contents

The scientific literature on herbicide resistant crops sometimes uses different terms to describe similar phenomena.

In this paper we follow the definition for herbicide resistance, as outlined by Heap and Lebaron (2001), in which herbicide resistance is the inherited ability of a plant to survive and reproduce following exposure to the dose of an herbicide typically used in agricultural systems. Note that in this paper, this definition is used for both weeds and crops. In addition, it does not distinguish between the various sources of genetically based resistance. The resistance can be endogenous or innate, before exposure to the herbicide (crops and weeds); created by laboratory modification (crops); or newly evolved after herbicide exposure (weeds).

Return to Table of Contents

3.1   Herbicide Selectivity

All the major commodity crops, such as corn, soy, cotton, wheat and rice, have some inherent resistance to certain herbicides, but not others. For example, the dosage of atrazine that controls many weeds does not affect corn, but greatly affects soybeans. The reverse is true for imazaquin, which belongs to a different class of herbicides than atrazine. Farmers use imazaquin to control weeds in soybean fields, but not in corn.  Therefore, in the absence of any modifications by plant breeders, corn has endogenous resistance to atrazine, and soybeans are naturally resistant to imazaquin. The capacity of an herbicide to harm some plants but not others is known as selectivity.

Herbicide selectivity extends to wild plants, as well. Certain weed species are naturally susceptible to some herbicides, but not others. In general, herbicides that control grasses (monocots) are ineffective against broadleaf weeds (dicots). However, a selective herbicide that effectively controls some broadleaf weeds does not necessarily control all of them; the same is true for those that control grasses. Therefore, when considering the number of available options for managing weeds, it is important to realize that even though there are over 100 herbicides on the market, only some can be used with certain crops, and a subset of those control some weeds, but not others.

In reality, farmers must battle a wide array of broad leaf and grassy weeds simultaneously. In selecting a weed management strategy, farmers choose the most economical herbicides that do not harm the crop, but control those weeds that lower crop productivity or decrease crop quality. If the herbicides a farmer selects control all but one species of weed, the farmer may not have significantly reduced the impact of weeds on crop productivity, because that one species can increase its numbers and successfully compete with the crop for access to essential resources such as water, light and nutrients.  Therefore, farmers must often utilize a combination of weed management techniques, including application of different herbicides, to effectively control weeds in their fields.

3.2   Variable Impacts of Herbicides on Plants

It is useful to think about responses to herbicide treatment as a set of possible responses rather than categorizing the plant’s response as either resistant or susceptible; lives or dies. In this paper, we recognize a continuum of responses based on two attributes that are important to crop productivity and weed resistance evolution: damage to the plant (either weed or crop) and effects on plant fitness/productivity (number of viable seeds).  A plant’s response to a certain dose of herbicide can range from death (and a fitness level of zero) to the other extreme of no damage and no decrease in fitness. Between those two extremes, the detrimental effects of the herbicide on the weed’s fitness and the crop’s productivity can vary considerably. For purposes of this paper, it is important to conceptually distinguish four levels of response to the specific dose of an herbicide:

• Plant dies, therefore weed fitness/crop productivity = zero
• Plant survives, but is damaged and has lower weed fitness/crop productivity
• Plant is damaged but has little or no decrease in weed fitness/crop productivity
• No damage and little or no decrease in weed fitness/crop productivity

Increasing or decreasing the dose of an herbicide can change a plant’s response from one level to another. For example, if a plant, whether crop or weed, dies at a certain dose of an herbicide, it may survive lower doses. However, while the dosage decrease may lessen the herbicide’s impact on one plant, other plants may not respond to the dosage decrease in the same way.

Return to Table of Contents

The physiological effects of herbicide application on plants generally fall into one of four categories:

• Inhibits cell division
• Regulates plant growth
• Inhibits photosynthesis and/or respiration
• Interrupts other essential metabolic processes

Herbicides contain both active and inert ingredients.  The physiological responses listed above describe the plant’s response to the active ingredient.

The physiological impact of virtually all herbicides has been determined and, in some cases, scientists have elucidated the precise molecular basis of the physiological changes that lead to plant death. (For more information on the mode of action of various herbicides, visit the web site of the Herbicide Resistance Action Committee at www.plantprotection.org.)

Herbicide resistant plants achieve resistance through one of the following biochemical mechanisms:

• Deactivation of the herbicide. For example, the resistant plant has a gene encoding an enzyme that breaks down the herbicide into innocuous compounds.
• Reduced sensitivity to the herbicide. For example, the target enzyme of the herbicide is modified in a way that prevents the herbicide from binding, but the target enzyme continues to maintain its normal physiological function.
• Avoidance of the herbicide through uptake and translocation inhibition or sequestration. For example, some cotton varieties have innate resistance to triazine herbicides because the plants preferentially store the herbicide in glands on its stems and leaves, which prevents the herbicide from reaching its target site.

These are the fundamental strategies that plants use to survive exposure to herbicides whether their resistance is endogenous; develops in response to selective pressure due to exposure to the herbicide in the field or lab; or is incorporated in to a crop plant through biotechnology or other means. Endogenous herbicide resistance includes examples of all three mechanisms, and different plants may use different resistance mechanisms to survive exposure to the same herbicide. Most resistance in crops developed under laboratory conditions is based primarily on herbicide deactivation and reduced sensitivity (Duke et al., 1991; Duke, 2005; Dill, 2005). For practical utility, any such efforts must result in crops that retain sufficient overall agronomic properties to be of interest to production agriculture.

4.1   Herbicide resistance through deactivation

Plants that are capable of utilizing this molecular strategy for herbicide resistance have detoxification enzymes that breakdown the herbicide into an innocuous molecule (s) or reduce its toxicity through a chemical alteration that inhibits binding to its target site.

A number of herbicide resistant crops developed through biotechnology utilize the herbicide deactivation strategy. The herbicide bromoxynil affects plants by inhibiting photosynthesis. Scientists discovered a gene for a detoxification enzyme, nitrilase, in the soil bacterium, Klebsiella pneumonie, subspecies ozaenae. Biotech cotton plants that contain the bacterial gene for nitrilase are resistant to bromoxynil, because they are able to convert the herbicide into a non-toxic molecule. The enzyme converts a carbon that is triple bonded to a nitrogen atom (a nitrile) to a carboxyl group, making the molecule harmless to the plant.

Glufosinate tolerant biotech crops utilize the deactivation mechanism, but rather than detoxifying the molecule, they alter its binding ability. The active ingredient of the herbicide glufosinate is phosphinothricine, which binds to the enzyme (glutamine synthase) that normally joins glutamate and ammonia to produce glutamine, an amino acid. By competing with glutamate for the enzyme’s active site, phosphinothricine harms plants through the accumulation of ammonia. Soil bacteria contain an enzyme, known as phosphinothricine acetyltransferase, or PAT, that adds an acetyl group to phosphinothricine. Biotech crops that contain a gene encoding the microbial PAT enzyme are capable of deactivating phosphinothricine via acetylation, which prevents it from binding with glutamine synthase.

4.2   Herbicide resistance through reduced sensitivity

Reduced sensitivity to herbicides can be achieved in number of ways. The molecular target of an herbicide is usually a plant enzyme or a cellular receptor. In the first case, the herbicide works by binding to the enzyme and inhibiting an essential metabolic process, as described above for glufosinate. In the second case, the herbicide binds to the receptor and triggers a cascade of molecular event that leads to plant death. Reduced sensitivity to the herbicide can be achieved by altering the target protein, so that it has less affinity for the herbicide. A second mechanism for reducing sensitivity involves increasing the number of target proteins, so that the number of target sites in a plant exceeds the number of herbicide molecules. As a result, some target proteins remain unaffected by the herbicide and an acceptable level of normal plant function is maintained. Clearly, the effectiveness of this last approach is dependent upon the herbicide dose. 

Examples of crops that are herbicide resistance because of reduced sensitivity are discussed at length below.

Return to Table of Contents

Most herbicides that have been used in agriculture for the past 60 years were developed by generating many compounds in the lab and spraying them on weeds in the greenhouse to identify potential candidates for further development. Those that were herbicidal were then tested on crops to assess the crop’s sensitivity. Once researchers had promising candidates for new herbicidal compounds, they then conducted tests to determine whether the herbicide and herbicide resistant crop performed well under field conditions and to assess any impacts on the environment and human health.

A number of decades ago, the number of chemicals that must be tested to develop a new herbicide began to increase exponentially from less than 1,000 in 1950 to approximately one million today. As a result, companies began to consider how they could develop new weed management options by focusing on modifying the crop, rather than finding a new herbicide. Starting with an herbicide already been proven to be effective and then giving the crop plant the capacity to resist that herbicide is more efficient and allows companies that sell herbicides to expand the applicability of their most effective and safest products.

Prior to the advent of modern biotechnology, this approach was not very productive.  Crop developers were constrained by the need to use plants in the same or very closely related species, and most crop plants have very little genetic diversity for herbicide resistance (Green, 1998). An exception was endogenous tolerance of soybeans to metribuzin and 2,4-D, which varied among different cultivars. As a result, soybean breeders were able to use conventional breeding and selection to increase soybean’s endogenous tolerance to these two herbicides and to combine resistance to both in a cultivar that was resistant to one, but not the other. 

Crop developers have used a number of techniques to create novel herbicide resistance traits. Screening and selection of plants with novel genetic variability led to the development of soybeans tolerant to sulfonylurea herbicides (Sebastian et al., 1989).

In the example above, researchers subjected the seeds to mutagens and tested whole plants for herbicide resistance. Screening whole plants for herbicide resistance takes considerable effort, time and space, because the seeds must be planted and allowed to develop before the herbicide is applied. Recent breakthroughs in plant cell and tissue culture associated with modern biotechnology allow researchers to screen cells and undifferentiated plant callus for herbicide resistance, which increases development efficiency. Treatment of corn tissue cultures with imidazolines followed by recurrent selection created imidazoline resistant callus that could then be regenerated into corn plants and bred into a variety of cultivars (Dyer, 1994).

Crop developers have also accessed novel genetic variability by exploiting the capacity of weeds to evolve resistance. For example, a breeding program was initiated to move triazine resistance found in a weed, Brassica rapa, to canola, a crop closely related to the weed (Hall et al., 1996). Triazine resistant canola was created, but unfortunately, the new resistance trait lowered crop productivity, which reduced the cultivar’s commercial value.

The tools of modern biotechnology reduced the impediments imposed by the requirement to cross breed crops with the same or closely related species. An herbicide resistant gene present in a given organism can now be incorporated into virtually any crop. This flexibility has triggered a revolution in weed management by expanding the spectrum of herbicides a crop can tolerate. 

The first transgenic herbicide resistant crop, bromoxynil-resistant cotton, was commercialized in 1995, and glyphosate-resistant soybeans came to market the following year. Glyphosate resistant soybeans, corn, cotton and canola are now grown on millions of hectares around the world (James, 2005). The molecular basis of glyphosate resistance and the evolution of glyphosate resistant weeds are discussed below.

Return to Table of Contents

6.1   Evolution of Herbicide Resistant Weeds

Weeds will eventually adapt and circumvent most control mechanisms. Herbicideresistant weeds are a well-established aspect of weed control for many herbicide classes including ACCase and ALS inhibitors, dinitroanilines, triazines and other PSII inhibitors. The occurrence of herbicide resistant weeds is dependent on the herbicide program, the weed species present, and the other crop management practices a farmer employs.

Herbicide resistance usually evolves in only one or two weed species in an area, even though a much larger number of weeds are exposed to the same herbicide selection intensity. Nonetheless, weed resistance to herbicides currently affects hundreds of thousands of fields and the most widely used herbicides (Heap, 2005). According to a recent survey, more than 290 types of herbicide resistant weeds are present in agricultural fields around the world (www.weedscience.org/in.asp). Resistant weeds often increase the cost of crop production and limit the effectiveness of herbicides that can be used and the crops that can be grown. Despite these challenges, farmers have used a variety of approaches to limit the impact of weeds on crop productivity. 

Developing new varieties of herbicide resistant crops may improve a farmer’s weed control options, but having access to new varieties will never completely eliminate the problem of weeds evolving resistance to herbicides (Heap and LeBaron, 2001; Owen, 2001).However, by using a wide variety of weed control options (including proper use of herbicides), a farmer can delay the development of weeds resistant to herbicides.

The capacity of weeds to become resistant to an herbicide depends, first and foremost, on the existence of individual weeds (biotypes) that happen to have resistance genes that enable them to both survive and reproduce when exposed to the herbicide. The herbicide itself does not directly cause the genetic change that imparts resistance; the resistance trait appears randomly in different populations of different weed species. It is difficult to predict exactly which weed species will contain biotypes that are resistant to specific herbicides.

A weed species’ inherent rate of mutation is a key factor for predicting new occurrences of resistant weeds, but this rate is difficult to quantify (Shane-Friesen and Hall, 2004; Gressel, 2006). Actual mutation rates have not been measured for any weed under field conditions. Saari et al. (1994) estimated for Arabidopsis that one mutation in a billion may lead to a gene variant that confers herbicide resistance. Other models that researchers use to study resistance evolution are based on a much higher mutation rate of one in a million (Cavan and Moss, 2001). Even a low estimate of mutation frequency can give a high probability of a resistant biotype occurring, because weeds produce vast numbers of seeds each year. Weed seed banks can be as high as 50,000 seeds/m², but even a modest 1,000 seeds/m² represents 10,000,000 seeds/ha.

To date, the resistance gene in all cases in which a weed has become resistant is the result of the weed’s inherent ability of generating gene variability. No weed has become resistant by acquiring resistance genes from herbicide resistant crops. Scientists are able to determine the source of the resistance gene by studying the molecular basis of the resistance. Weeds have evolved biochemically unique resistance strategies not found in crops. Even so, it is still possible for weeds to acquire resistance genes from crops that are close relatives. This is especially true if the resistant phenotype is provided by a single gene, as it is often the case with crops developed through biotechnology. For a discussion of the potential of weed herbicide resistance acquired through gene flow, see the Science Narrative ‘Gene Flow via Pollination‘.

The spread of a resistance phenotype, which leads to a weed population that is not susceptible to the herbicide, depends primarily upon the exposure to the herbicide that the weed is able to tolerate. When an herbicide is applied, most of the susceptible weeds die, while the resistant weeds survive, mature, and produce seed. Even though they may still be few in number, repeated application of the same herbicide continues to increase the proportion of resistant weeds in the population. Thus, when growers say that their "weeds have become resistant," they really mean that the population of resistant weed biotypes, which formerly existed at low numbers, has increased.

In the absence of exposure to the respective herbicide, resistant weed biotypes may be maintained in the populations at low frequencies, if the new resistance phenotype does not significantly reduce the weed’s fitness. If, however, the new resistance gene carries a significant fitness cost, the resistant gene’s frequency will decrease over time in the absence of any selection for the gene, i.e., application of the herbicide. For example, many weed biotypes that evolved resistance to atrazine have lower vigor and fitness and cannot compete with biotypes that are susceptible to atrazine (Stowe and Holt, 1988), because the resistance trait is associated with a less efficient photosynthetic system.

The existence of herbicide resistant weeds is not necessarily an economic or ecological problem. It is an economic problem only if:

• the herbicide the weed tolerates is an economically desirable option, and
• few herbicide options can be used in the crop(s) where the resistant weed occurs.

Herbicide resistance can become an ecological problem if the resistant weed biotype replaces the non-resistant biotype in the weed population. Even then, the shift to an herbicide resistant population of weeds has ecological consequence only if the resistant population cannot be controlled with other herbicides or other control practices. This is rarely the case. Many hundreds of cases of resistant weeds have been documented worldwide, but only under exceptional circumstances has resistance become a limiting factor for crop production. For example in some locations in Australia, biotypes of rigid ryegrass (Lolium rigidum) are resistant to many herbicides in several different classes (Heap, 2005). 

In spite of the evolution of herbicide resistance in weed populations, farmers in the U.S. who grow corn and soybeans continue to have many herbicides and agricultural management options for weed control. Even so, growers must always be concerned about herbicide sustainability and the economic consequences of losing an herbicide due to the evolution of resistant weeds. This is especially the case if the most difficult to control weeds become resistant to more than one herbicide, which is known as cross-resistance.  With very high re-registration costs for older herbicides and high development costs for new herbicides, farmers cannot assume they will continue to have access to new herbicides that replace those that have lost their value.

6.2   Weed Spectrum Shifts

The composition and density of weeds in cropland will change in response to weed control practices, whether or not herbicides are used. A few individual weed plants that are partially resistant or difficult to control with certain practices will survive and reproduce, eventually leading to weed populations that are more difficult to control and a need for the grower to change weed control measures and/or herbicides. The ability to use a greater array of herbicide options with herbicide resistant crops should help to minimize shifts in weed spectrum that occur and will provide more opportunities for effective control if shifts occur.

Not all cases of weed shifts that are driven by herbicide use can be explained by herbicide selective pressure that favors weeds with a genetically based biochemical capacity to survive exposure. Weeds with delayed emergence and slower development are able to avoid exposure to the herbicide (Hilgenfield et al., 2004). Therefore, application of herbicides can favor weed species that emerge late or can select for biotypes within a species that are capable of delaying emergence. In a study with a range of weed species, ivyleaf morning glory (Ipomoea hederacea), and shattercane (Sorghum bicolor) were able to avoid exposure to glyphosate applications due to delayed emergence.

6.3   Integrated Weed Management and Herbicide Resistance

Growers often do not detect resistance until the herbicide resistant biotype comprises 30 to 40% of the weed population.Weed scientists have established the maximum number of various weed species that can be tolerated before significantly decreasing yields – the economic threshold. Although growers are encouraged to wait until the weed reaches the economic threshold before implementing control measures, the wisest strategy for managing weed resistance is to act proactively and not wait until the problem is recognized (Yancy, 2005). Such a strategy centers on reducing the selective pressure imposed by the herbicide. “Selection pressure is the key,” says Dr. Alan York of North Carolina State University. “You have no way of knowing whether that ‘one in a 100 million’ plant is present or not. You must assume it might be present and your resistance management strategy becomes one of reducing selection pressure on that plant.”

Managing agroecosystems to minimize the development of resistant weeds relies upon changing weed control measures regularly. As a result, new control methods continually need to be evaluated and implemented, so that farmers can stay one step ahead of weed resistance problems.

Herbicide resistance or weed spectrum shifts become a problem most often when a single herbicide or herbicides with the same mode of action are used repeatedly on a weed population over several years. Therefore, weed resistance is often associated with monoculture cropping systems. In addition, relying only on herbicides to control weeds and neglecting other methods, such as crop cultivation techniques, accelerates the development of herbicide resistant weeds. Therefore, the risk of development of resistant weeds is highest in corn and soybean agroecosystems, if herbicides, especially those with the same mode of action, are used as the sole tactic for weed management (Owen, 2001; Heap, 2005). 

Herbicides are very important tools for farmers, and they should be used properly to preserve their effectiveness. Any weed management option that reduces herbicide-imposed selection pressure will reduce the rate of resistance development to the herbicide. By adopting options, such as mixtures of herbicides with different mode of action and crop rotation, selection pressure for resistant weeds can be reduced and the usefulness of herbicides preserved. Generally, the best approach to resistance management is to use Integrated Weed Management (IWM) practices. IWM utilizes a range of weed control methods and includes the following tactics:

• Avoid using the same herbicide or herbicides with the same mode of action multiple times per year or year after year.
• Use tank-mixtures consisting of different herbicide types with overlapping weed spectra.
• Use crop rotations because different crops allow different cultural and tillage options that compete much differently with weeds. 
• When using herbicides, use full label rates and tank mix partners.
• Use clean seeds and clean equipment to minimize spread of weed seed.
• Monitor fields after herbicide applications for appearance of resistant weeds.
• Control weeds before they form seed.
• Where practical, use cover crops and other methods to reduce weed seeds in soil.

Return to Table of Contents

II. Developing Herbicide Resistant Crops

The tools of modern biotechnology have provided crop developers with strategies for broadening the spectrum of herbicides that can be used on a crop. In this section, we describe two of the methodologies, both of which rely on modifying the genetic makeup of the crop plant.

In one method, crop developers use biotechnology to select for herbicide resistant plant cells and tissues, rather than waiting for the plant to develop from a seed.  Often, but not always, the cell suspensions and tissue cultures are treated with chemicals before scientists screen them for herbicide resistance and select those with the resistance trait. At other times, keeping cells and tissue in culture encourages the development of new genetic variation, some of which may provide herbicide resistance. 

The second methodology utilizes biotechnology to genetically modify crop plants. For more information on methods used in biotechnology, see the Science Narrative ‘Methods Used in Biotechnology’ on this website.

Return to Table of Contents

ALS-inhibiting herbicides were discovered in 1975 (Stetter, 1994; Shaner and O’Connor, 2000; Tan et al., 2005). They inhibit a plant enzyme called acetolactate synthase (ALS) that is required for the production of essential branched-chain amino acids such as valine, leucine, and isoleucine. This biochemical pathway is not present in humans and animals (Saari and Mauvais, 1996). There are five different chemical classes of ALS herbicides (sulfonylureas (SU), imidazolinones (IMI), triazolopyrimidines (TP), pyrimidinylthiobenzoates (PTB), and sulfonylamino-carbonyl-triazolinones (SCT).

ALS herbicides control a wide spectrum of grass and broadleaf weeds at very low application rates. In addition, they generally have very low mammalian toxicity and possess a favorable environmental profile. Today, about 56 different ALS herbicide active ingredients are marketed with registrations in all major crops. Significant changes in herbicide potency, crop selectivity, and weed control can be made with small chemical alterations within the ALS herbicide class.

Soon after the commercialization of the first ALS herbicides, tissue culture was used to successfully select highly tolerant tobacco lines (Chaleff and Ray, 1984). The development of ALS-resistant corn began in 1982 with tissu

Also in 1980s, breeders introduced genetic variation into soybean seeds and subsequently screened for those that emerged into sulfonylurea resistant plants. One plant showed significant increase in SU resistance, as a result of a single amino acid mutation in the ALS enzyme (Sebastian et al., 1989). The progeny of this plant was utilized in breeding programs to incorporate the trait into a wide range of agronomically useful soybean varieties, which became commercially available in 1993 as STS® soybeans (Sulfonylurea tolerant soybeans).

Shortly thereafter, a further improved source of ALS resistance was discovered in corn and soybeans, a double mutant ALS allele that encodes for a highly resistant form of the ALS enzyme (HRA; highly resistant allele) (Lee et al., 1988; Mazur and Falco, 1989; Bedbrook et al., 1995). The HRA trait is dominant and is expressed throughout the plant, so ALS herbicides do not harm plant roots or shoots.

The two mutations in the double mutant are also amino acid substitutions, but the locations of the substitutions vary slightly depending on the plant source. The two mutations provide different types of resistance to ALS herbicides (Tranel and Wright, 2002), with one giving broad-spectrum resistance to ALS herbicides, while the other one confers resistance to sulfonylureas and triazolopyrimidines. Both mutations together, as present in the double mutant, provide resistance to all five classes of ALS herbicides. When HRA is adequately expressed in plants, high rates of ALS herbicides can be applied without crop injury.

STS® is a registered trademark of DuPont or its affiliates

Return to Table of Contents

Glyphosate is a broad spectrum herbicide that was introduced in the 1970s for management of annual, perennial and biennial herbaceous grasses, sedges, and broadleaves, as well as woody brush and trees (Franz et al., 1996). 

Glyphosate controls plants by inhibiting the enzyme EPSPS (5-enolpyruvylshikimate-3-phosphate synthase). EPSPS is an essential enzyme in the shikimate pathway that ultimately leads to the production of aromatic amino acids (tryptophan, tyrosine, and phenylalanine). The shikimate pathway for synthesizing aromatic amino acids, and therefore the enzyme EPSPS, is found in plants, bacteria and fungi, but not animals.

The structure of glyphosate resembles the structure of the substrate of EPSPS, which is phosoenolpyruvate (PEP). Therefore, glyphosate competes with PEP for the enzyme’s active site and prevents conversion of PEP to the precursor that is required in the synthesis of aromatic amino acids. Aromatic amino acids are essential for many plant processes, such as protein synthesis, cell wall formation, pathogen defense and hormone production.  By preventing the production of aromatic amino acids, glyphosate is toxic to virtually all plants.

There is no significant endogenous tolerance to glyphosate in crop plants or their weedy relatives. Consequently, attempts to conventionally breed most crops for glyphosate resistance have not succeeded (Padgette et al., 1995). Resistance was found in tissue cultures derived from 10 species. However, the resistance had no commercial value because in some cases it was not genetically based; in others, it was not stably inherited.

It is difficult to find glyphosate-resistant EPSPS molecules that still possess sufficient catalytic activity to provide adequate functioning of the shikimate pathway (Kishore et al., 1992; Padgette et al., 1996b). A naturally-occurring multiple mutation in corn EPSPS was discovered and has been utilized to generate glyphosate resistance (Lebrun et al., 1997).  Chemically mutating corn callus also led to an EPSPS with two amino acid changes that provided resistance to glyphosate without impairing EPSPS function (Dill, 2005).   However, multiple copies (amplification) of this double mutant are needed to provide a level of glyphosate tolerance that is acceptable to farmers.

In 1983 scientists discovered a bacterial gene, CP4, which encodes for a highly efficient, glyphosate-resistant form of EPSPS (Padgette et al., 1996a). Using the new, gene-based methods of biotechnology, crop developers were able to deliver the Agrobacterium CP4 gene into crop plants. As a result, biotech crops contain both the glyphosate-susceptible form of EPSPS, encoded by crop genes, and the glyphosate-resistant EPSPS encoded by the microbial CP4 gene. Having the bacterial form of EPSPS allows crops to continue to synthesize aromatic amino acids, even as the plant EPSPS enzyme is inactivated by glyphosate competitive inhibition, because glyphosate does not bind to bacterial EPSPS. The bacterial CP4 gene provides the basis of glyphosate resistance in most biotech crops that are resistant to the herbicide.

Biotech crop developers have utilized two strategies to enhance glyphosate resistance: increasing access of bacterial EPSPS to glyphosate and bolstering EPSPS insensitivity-based resistance with a new molecular mechanism, herbicide deactivation.

• Increased access
  EPSPS is found in chloroplasts. Enhancing the efficiency of the choloroplast transfer proteins (CTP) that are necessary for moving EPSPS into chloroplast has increased the chloroplast concentration of EPSPS molecules that are resistant to glyphosate. This allows the insensitive EPSPS molecule to out compete the plant’s endogenous EPSPS molecules for access to glyphosate.
• Deactivation
  Resistance to glyphosate encoded by the CP4 gene can be enhanced with other novel genes to improve their level of resistance. For example, glyphosate resistant canola contains, in addition to CP4, a bacterial gene that encodes the enzyme glyphosate oxidoreductase (GOX). The microorganism Ochrobactrum anthropi, the source of the GOX gene,  uses the  enzyme to degrade glyphosate to glyoxylate and aminomethylphosphonate (AMPA), two compounds that are non-toxic to plants. This mechanism is most likely the dominant glyphosate degradation pathway that soil microorganisms evolved in response to glyphosate exposure, because they, too, rely on the shikimate pathway to synthesize aromatic amino acids. So, glyphosate resistant canola plants rely on two molecular mechanisms: herbicide deactivation and target insensitivity.

Recently, a new mechanism to detoxify glyphosate in planta, deactivation by acetylation, was developed (Castle et al., 2004). The enzyme, glyphosate acetyltransferase, converts glyphosate into the non-toxic molecule acetylglyphosate, before glyphosate can reach and inhibit the the EPSPS enzyme. The transferase gene was derived from a naturally occurring soil bacterium (Bacillus licheniformis). Researchers used a technology, known as gene shuffling, to optimize both the enzyme’s acetylation efficiency and its specificity before incorporating the gene into plants (Castle et al., 2004; Siehl et al., 2005).

In summary, the mechanisms that have been used to provide glyphosate resistance to crop plants are:

• Target insensitivity - EPSPS molecules that do not bind glyphosate;
• Target insensitivity plus amplification to increase the number of insensitive sites;
• Herbicide deactivation - GOX and glyphosate acetyltransferase;
• Improved transport of insensitive targets to increase access to glyphosate (CTP).

Return to Table of Contents

III. Herbicide Resistant Crops and Weed Control

Regardless of the techniques that are used in developing herbicide resistant crops, concerns related to sustainable weed management and herbicide use remain the same, including the development of resistant weeds. 

In this section, we discuss some of the crop management concerns and opportunities associated with crops that are resistant to the ALS herbicides or glyphosate.

The five classes of ALS herbicides differ chemically, but all bind to the same target site on the ALS enzyme. The primary molecular basis of weeds resistant to ALS is reduction of target site sensitivity, although there are biotypes that are resistant through more rapid detoxification of the herbicide to inactive metabolites. Metabolic tolerance provided by detoxification generally results in lower levels of resistance than decreasing sensitivity to the herbicide through target site modifications.

To date, weed scientists have identified 92 weed species containing biotypes that are resistant to ALS herbicides (Heap, 2005). The large number of resistant biotypes is due, in part, to the relatively large number of amino acid substitutions that can change the ALS enzyme from a sensitive to a resistant form. Five different mutations sites have been identified in naturally occurring resistant weed populations (Bernasconi et al., 1995; Tranel and Wright, 2002). However, not all ALS resistant weeds are resistant to all classes of ALS herbicides (Tranel and Wright, 2002). ALSresistance generally falls into three categories:

• broad resistance to sulfonylureas (SU), imidazolinones (IMI), triazolopyrimidines (TP), and pyrimidinylthiobenzoates (PTB);
• resistance to IMI and PTB only; and
• resistance to SU and TP only.  

Most ALS resistant weeds occur in small and localized areas only. Exceptions include ALS-resistant Kochia (Kochia scoparia) and Russian thistle (Salsola iberica), which are now present in over 60% of the wheat fields in northern United States (Heap and LeBaron, 2001), and ALS resistant water hemp in Illinois. As a consequence, few growers in these areas use ALS herbicides to control these weeds. However, they do apply ALS herbicides to control other weed species in their fields.

Return to Table of Contents

For more than two decades, the evolution of glyphosate resistance was not perceived as a problem (Bradshaw et al., 1997). This opinion was based on the difficulty of discovering a fully functional EPSP synthase that was insensitive to glyphosate; the inability of plant species to enzymatically deactivate glyphosate; the lack of soil activity of glyphosate (thus reducing the selection pressure due to exposure); and the empirical observation that no resistant weeds had appeared after at least 20 years of usage.

The three most likely ways for weeds to develop resistance to glyphosate did not seem likely (Jasieniuk, 1995). Over-production of the EPSP synthase target site did not increase resistance enough for plants to survive glyphosate amounts used in agricultural settings (Kishore and Shah, 1988). The EPSPS modifications that conferred glyphosate resistance in bacteria were inside the enzyme’s active site, which reduced its catalytic efficiency and thus would probably reduce plant fitness (Padgette et al., 1995). No higher plants could be found with even low levels of ability to metabolically inactivate glyphosate (Dyer, 1994).

However, the views about the capacity of weeds to develop glyphosate resistance changed in 1996 when a glyphosate-resistant weed, Lolium rigidum, was discovered in Australia (Pratley et al., 1999; Powles et al., 1998). Since then, glyphosate resistant biotypes in at least 10 other weed species have been confirmed (Nandula et al., 2005), and more than a million hectares are now infested with resistant weeds. Some of these biotypes exhibit resistance at application rates 4 to 13 times higher than susceptible populations.

The most widespread glyphosate-resistant weed is marestail (Conyza canadensis). Glyphosate resistant marestail was confirmed in 2000 (VanGessel, 2001). Marestail produces very large numbers of light, wind dispersed seed and can cross-pollinate, leading to widespread infestation of no-tillage crop and non-crop land. As a result, five years after its first occurrence, it was found on a half-million hectares across the U.S. Midwest, South and Atlantic states (VanGessel, 2001; Heap, 2005). Dose response analysis showed these populations were eight to 13-fold more resistant that susceptible marestail populations.

In 2002 and 2003, farmers, consultants and extension agronomists frequently observed common lambsquarters (Chenopodium album), a significant problem in row crops, as the only surviving weed in soybean fields after glyphosate usage. Farmers and extension agents thought factors other than endogenous resistance, such as an inability of glyphosate to penetrate lambsquarters thick cuticle, might explain persistence after glyphosate expsoure. However, King et al (2004) recently confirmed some common lambsquarters biotypes in Virginia are glyphosate resistant.

Significant populations of common and tall waterhemp (Amarathus sp.) that survived glyphosate application were first observed in fields in Iowa, Illinois, and Missouri (Owen, 2002; Smeda and Schuster, 2002). Studies indicate that plants survived glyphosate rates 2.6 times the label rate and some waterhemp plants have now been classified as glyphosate resistant (Owen and Zelaya, 2005).

The molecular basis for weed resistance to glyphosate is not understood in most cases.  Initial studies of various weed species revealed EPSPS target site insensitivity (Braerson et al., 2002). Differences in translocation and transport to the chloroplast are also important in some weeds (Lorraine-Colwill et al., 2003; Feng et al., 2004). The mechanisms of glyphosate resistance in other weeds appear to be complex and polygenic. 

10.1   Weed Spectrum Shifts and Glyphosate

The effectiveness, economic benefits, and ease of using glyphosate have led to repeated applications, year after year, in areas where glyphosate resistant biotech crops are grown.  This intensive usage has resulted in a high selection pressure for weeds that inherently are difficult to control with glyphosate (Culpepper, 2004). Eventually, this selection pressure can lead the spectrum of weeds in the fields to a shift to those weeds that inherently can tolerate glyphosate. In cotton, the annual grasses, waterhemp (Amaranthus spp.), dayflower (Commelina spp.), morning glory (Ipomoea spp.), and winter annuals are increasingly present in fields treated with glyphosate. In soybeans, some farmers have difficulty controlling waterhemp (Amaranthus spp.), copperleaf (Acalypha spp.), velvetleaf (Abutilon theophrasti), giant ragweed (Ambrosia trifida), marestail (Conyza canadensis), winter annuals, and lambsquarters (Chenopodium spp.) with glyphosate.

Considerable genetic variation in endogenous glyphosate resistance naturally exists in morningglory species (Ipomoea sp.) (Baucom and Mauricio, 2004) and waterhemp (Amaranthus rudis and A. tuberculatus) (Patzoldt et al., 2002).  In some morningglory species, the more tolerant plants often produce fewer seeds, so the resistance trait seems to carry a fitness disadvantage. However, a Kansas study showed that ivyleaf morningglory (Ipomoea hederacea) and large crabgrass (Digitaria sanguinalis) dominated the weed community in a corn-soybean rotation using glyphosate applications (Marshall et al.,  2000). In North Carolina, Coble and Warren (1997) reported similar increases in resistant morningglory species after three years of continuous glyphosate use.  

As predicted, spectrum shifts to weed populations with endogenous glyphosate resistance have occurred more rapidly than evolved resistance in response to glyphosate exposure (Shaner, 2000). For example, in Iowa, common waterhemp (Amaranthus rudis) and velvetleaf (Abutilon theophrasti) became a concern in glyphosate-resistant soybean soon after crop commercialization (Owen, 1997). 

Westra et al. (2004) conducted a 6-year field study involving several crops, crop rotations, and herbicide treatments. In the reduced glyphosate rate regime (0.42 kg/ha), significant weed shifts were observed; common lambsquarters, wild buckwheat (Polygonum convolvulus), Palmer amaranth (Amaranthus palmerii) and other species became the dominating weeds. However, at the labeled rate of glyphosate (0.84 Kg/ha), no significant weed shifts were noted. 

Return to Table of Contents

In most cases where herbicide resistant weed biotypes occur, the same herbicide can continue to have utility due to the spectrum of weeds present. To control the herbicide resistant weed, weed scientists recommend the adding another herbicide to the primary herbicide. For example, in fields with glyphosate resistant marestail, glyphosate could be the base treatment that growers augment another herbicide to manage marestail.

Using mixtures of herbicides with different modes of action and overlapping weed spectrums is a common tactic of farmers to manage the evolution of resistant weeds, because it is highly unlikely that weeds will develop resistance to two herbicides with different modes of action if they are used simultaneously. The potential for developing dual herbicide resistance is more likely if the weed is exposed to two herbicides sequentially.

Mixtures of glyphosate with other herbicides can offer a total weed control option for many years to come, provided the crop is resistant. For example, glyphosate and ALS herbicides could be used on biotech corn and soybean that are resistant to both.

Return to Table of Contents