Herbicides as Weed Control Agents: State of the Art: II. Recent Achievements

Herbicide discovery has faced significant challenges over the past few decades, and weed control innovations are urgently required.

Abstract

In response to changing market dynamics, the discovery of new herbicides has declined significantly over the past few decades and has only seen a modest upsurge in recent years. Nevertheless, the few introductions have proven to be interesting and have brought useful innovation to the market. In addition, herbicide-tolerant or herbicide-resistant crop technologies have allowed the use of existing nonselective herbicides to be extended into crops. An increasing and now major challenge is being posed by the inexorable increase in biotypes of weeds that are resistant to herbicides. This problem is now at a level that threatens future agricultural productivity and needs to be better understood. If herbicides are to remain sustainable, then it is a must that we adopt diversity in crop rotation and herbicide use as well as increase the use of nonchemical measures to control weeds. Nevertheless, despite the difficulties posed by resistant weeds and increased regulatory hurdles, new screening tools promise to provide an upsurge of potential herbicide leads. Our industry urgently needs to supply agriculture with new, effective resistance-breaking herbicides along with strategies to sustain their utility.

Only a few companies are significantly pursuing herbicide discovery in the 21st century. Most of these have combined seed and traits businesses, since fees for traits constitute a considerable part of the income of agrochemical companies today. In concert with a review of the historical perspectives of herbicide research (Kraehmer et al., 2014), we provide here a short description of the current major research activities within the remaining 21st century agrochemical companies. After an overview of the chemicals that have entered the market in the 21st century, we provide a brief summary of the current nature of the weed-resistant herbicide problem. We then go on to summarize breeding-assisted and transgenic approaches toward the improvement of crop selectivity through the delivery of so-called herbicide-tolerant (HT) or herbicide-resistant crops, and conclude with a discussion of the new herbicide discovery screening tools that have been employed since the year 2000 and prospects for the future.

MAJOR CHEMICAL TRENDS AFTER 2000

Several new compounds have entered the herbicide market in recent years. Although not representing new modes of action (MoAs), they have increased the number of tools available for farmers to use to control weeds. Even in known and older herbicidal classes, new, interesting, and marketable molecules have been discovered. For example, and perhaps surprising given the relative age of the class of herbicides, new (after 2000) acetolactate synthase (ALS) inhibitors have provided solutions for farmers that can be regarded as real innovations. One of them is mesosulfuron-methyl (Fig. 1), a sulfonylurea herbicide that, when combined with iodosulfuron-methyl sodium, has broad-spectrum postemergence grass weed control at dose rates of 4.5 to 15 g active ingredient (a.i.) ha⁻¹ (Safferling, 2005).

Figure 1.

The ALS inhibitors mesosulfuron-methyl and iodosulfuron-methyl sodium.

Another very successful new ALS herbicide is thiencarbazone-methyl (TCM; Fig. 2), a compound of the sulfonylaminocarbonyl-triazolinone subgroup. TCM is a broad-spectrum herbicide with a maximum seasonal use rate of 45 g a.i. ha⁻¹ that is able to control a wide range of grasses and broadleaf weeds. Due to its lack of inherent selectivity, utility in a crop is only possible when combined with safeners such as mefenpyr-diethyl for cereals (Veness et al., 2008) or the new safener cyprosulfamide for corn (Zea mays; Santel, 2012). Different safeners thus make TCM a product for different crops. It can be used flexibly for preemergence and postemergence weed control and is a good example of an herbicide that can be used in multiple situations given the right mixture partners. Although TCM controls a wide spectrum of weeds, it has gaps that require mixture partners. For example, isoxaflutole (Fig. 3), a 4-hydroxyphenylpyruvate dioxygenase (HPPD) inhibitor that is used together with the safener cyprosulfamide, provides a good mixture partner for TCM in the preemergence control of weeds in corn. A suitable mixture partner for postemergence applications is another HPPD inhibitor, tembotrione (Fig. 3), that is marketed together with the safener isoxadifen-ethyl. One key selling point of this mixture is that it complements the efficacy of glyphosate and glufosinate in HT corn and provides resistance management options especially against glyphosate-resistant weeds (Müller et al., 2012).

Figure 2.

The ALS inhibitor TCM.

Figure 3.

The HPPD inhibitors isoxaflutole and tembotrione.

A third ALS inhibitor that entered the market after the year 2000 is pyroxsulam (Fig. 4; Wells, 2008). This compound belongs to the ALS subgroup triazolopyrimidine sulfonamides and controls a broad spectrum of annual grass and broadleaf weeds with an application rate of 9 to 15 g a.i. ha⁻¹. Crop selectivity is achieved in wheat (Triticum aestivum), rye (Secale cereale), and triticale varieties (hybrids of wheat and rye; x Triticale Tscherm.-Seys ex Müntzing) in combination with the safener cloquintocet-mexyl. To complete the weed spectrum, pyroxsulam is mixed with other products, such as florasulam (Fig. 4). It is also sold in a mixture with pendimethalin in Europe. Since the first registration of pyroxsulam in Chile in 2007, the compound has taken significant market share, and it has become one of the most important herbicides for cereals in Europe. It is surprising that, without exception, the latest innovative ALS solutions having a significant market impact all depend on safeners for crop selectivity.

Figure 4.

The ALS inhibitors pyroxsulam and florasulam.

The HPPD inhibitor herbicides have included some remarkably successful introductions in recent years, especially in corn but also in other crops (Ahrens et al., 2013). The first HPPD products, pyrazolynate, pyrazoxyfen, and benzofenap (Fig. 5), were introduced to the market in the 1980s and were used in rice (Oryza sativa) production in Japan with very high application rates of up to 4 kg a.i. ha⁻¹ (van Almsick, 2012). The first HPPD inhibitor for corn was sulcotrione (Fig. 6), a triketone with a somewhat lower but still relatively high application rate of 300 to 450 g a.i. ha⁻¹ for postemergence control of mainly broadleaf weeds. The real market success, however, began with the introduction of the second generation of the triketone HPPD inhibitors. Mesotrione (Fig. 6) represented a significant innovation not only because it could be applied at much lower rates than the previous generation but because it could be applied either preemergence or postemergence. Rates of only 70 to 150 g a.i. ha⁻¹ in postemergence treatments and somewhat higher rates of 100 to 225 g a.i. ha⁻¹ in preemergence treatments are sufficient to achieve good control of targeted weeds (Edmunds and Morris, 2012). To complete the spectrum, it is always mixed with other compounds, such as S-metolachlor and atrazine, or alternatively with terbuthylazine in countries where atrazine is no longer registered. Since its introduction into the United States in 2001, mesotrione has been a major success. Sales of mesotrione-based products have been increasing steadily, such that in 2007 it was already among the five best-selling herbicides worldwide (Cheung et al., 2008).

Figure 5.

The first HPPD inhibitors for rice: pyrazolynate, pyrazoxyfen, and benzofenap.

Figure 6.

The HPPD inhibitors sulcotrione and mesotrione.

Isoxaflutole (Fig. 3) was also developed in the late 1990s for preemergence use in corn (Luscombe et al., 1995). Even though the herbicide gives excellent control of selected broadleaf and grass weeds, the necessary application rates to achieve such a broad spectrum were near 100 g a.i. ha⁻¹. Unfortunately, such rates led to problems in crop selectivity from time to time. With lower application rates of 75 g a.i. ha⁻¹, the crop injury problems could be solved, but at that rate significant weed control was lost in certain grass weeds and there was a risk of ending up with only a broadleaf herbicide. Once again, the addition of a safener made the difference, allowing the use of the higher rate. An additional triketone HPPD inhibitor for postemergence control in corn, tembotrione, has recently followed sulcotrione and mesotrione into the market (Fig. 3). It offers a broader weed spectrum than the older compounds and also has outstanding selectivity in combination with the safener isoxadifen-ethyl (Van Almsick et al., 2009). With application rates of 75 to 100 g a.i. ha⁻¹, tembotrione is able to control common grass weeds, such as foxtails (Setaria spp.) and woolly cupgrass (Eriochloa villosa), but also a large number of broadleafed species. This includes a few glyphosate-, ALS-, or dicamba-resistant weeds. Tembotrione is not persistent in soil and, therefore, does not limit crop rotation opportunities for other crops in the following seasons. Another new representative of the HPPD inhibitors is pyrasulfotole (Fig. 7), a pyrazolone compound related to the above-mentioned rice HPPD inhibitors in Figure 5 (Schmitt et al., 2008). With a weed control spectrum limited to broadleaves, the compound (in mixture with bromoxynil or 2-methyl-4-chlorophenoxyacetic acid) is the first, and still only, HPPD inhibitor herbicide in the cereal market. A new MoA for a crop often offers the chance to control weeds that have developed herbicide resistance. This is the case for ALS-resistant Kochia scoparia, for example. Once again, this compound (pyrasulfotole) needed a safener, even though its herbicidal activity on grass weeds is very limited and thus theoretically possesses sufficient tolerance to monocot crops. In combination with mefenpyr-diethyl, a safe postemergence use is possible in all varieties of wheat, barley (Hordeum vulgare), and triticale. A further interesting aspect is the synergism of HPPD inhibitors with a PSII inhibitor such as bromoxynil, which helps to limit the application rate of pyrasulfotole to 25 to 50 g a.i. ha⁻¹ and serves to broaden the broadleaf control spectrum. Mixture partners for additional grass weed control such as fenoxaprop-P-ethyl are required. In this case, the safener mefenpyr-diethyl works for both compounds, fenoxaprop-P-ethyl and pyrasulfotole, even though the MoA of both is completely different (acetyl-coenzyme A carboxylase [ACCase] inhibitor and HPPD inhibitor).

Figure 7.

The HPPD inhibitor pyrasulfotole for cereals.

There are additional compounds that complete the newest generation of HPPD inhibitors (Fig. 8), such as topramezone for corn or tefuryltrione for rice. Other molecules in this class, such as bicyclopyrone for corn and fenquinotrione for rice, are currently in development.

Figure 8.

Further HPPD inhibitors: topramezone, tefuryltrione, bicyclopyrone, and fenquinotrione.

Completely nonselective herbicides are rarely found and developed. One relatively new compound that entered the nonselective market in 2010 is indaziflam (Ahrens et al., 2011; Fig. 9). This herbicide belongs to the so-called alkylazine class and is a cellulose biosynthesis inhibitor, representing a new MoA for this market. From the chemical point of view, the compound represents a high degree of innovation in manufacturing because of the need to synthesize this complicated chiral compound in relatively large quantities. Indaziflam controls weeds in established permanent crops such as tree plantations, perennial crops such as sugarcane (Saccharum officinarum), and turf grasses. With application rates of 73 to 95 g a.i. ha⁻¹, indaziflam provides control of weeds up to 90 d or longer after treatment. When weeds are present at application, the addition of a foliar herbicide such as glyphosate or glufosinate-ammonium is useful due to its limited postemergence activity. To expand the spectrum of weed control, indaziflam can be mixed with a range of other herbicides such as metribuzin and isoxaflutole.

Figure 9.

The cellulose biosynthesis inhibitor indaziflam.

With the success of HT crops and the ease of postemergence applications combined with relatively low herbicide costs, and combined with the perceived advantages of applying an herbicide only when weed growth was observed, the end of residual herbicides was prophesied to have arrived a while ago. The situation has changed recently with the appearance of glyphosate-resistant weeds such as Amaranthus tuberculatus and Amaranthus palmeri. The demonstrated advantages of using preemergence herbicides to reduce the population of these highly competitive weeds that germinate over an extremely long period (Hager et al., 2002) prompted companies to develop new residual products such as saflufenacil (Fig. 10), a protoporphyrinogen IX oxidase inhibitor (Kixor Herbicide Worldwide Technical Brochure, 2008). This compound can be used alone or, more importantly, mixed with glyphosate and applied preplanting for burndown applications. Therefore, saflufenacil is a useful addition to a very important segment of glyphosate-tolerant crops where glyphosate use predominates (Knezevic et al., 2009). The compound controls primarily dicot weeds, but it controls more than 80 of them, including key driver weeds resistant to glyphosate and ALS herbicides. It can be used as a preemergence treatment in corn and sorghum (Sorghum bicolor) to control major dicot or broadleaf weeds without triazine herbicides. It also can be used as a preplanting burndown product for other crops.

Figure 10.

The protoporphyrinogen IX oxidase inhibitor saflufenacil.

Another new trend in weed control is the renaissance of auxinic herbicides (Fig. 11), the class that provided the first modern herbicides (Peterson, 1967). Compounds such as aminopyralid (Masters et al., 2012), aminocyclopyrachlor (Claus and Finkelstein, 2012), and halauxifen-methyl are new representatives of this long-established MoA (Schmitzer et al., 2013). The latter of this group is supposed to enter the market in 2014 (http://newsroom.dowagro.com/press-release/dow-agrosciences-announces-arylex-active-global-commercial-brand-name-new-herbicide). Aminopyralid controls primarily broadleaf weeds, including noxious, poisonous, and invasive plants in rangeland, pasture, and industrial vegetation management sites. It was discovered and registered in the United States for noncrop and turfgrass uses for the control of annual and perennial broadleaves and brush weeds. Halauxifen methyl is also potentially useful as a broadleaf herbicide in selected row crops and a potential mixture partner for cereal portfolios.

Figure 11.

The auxins aminopyralid, aminocyclopyrachlor, and halauxifen-methyl.

The inhibition of ACCase is, like that of ALS, one of the most commercially important MoAs for weed management, even though ACCase inhibitors are active only on grass weeds. The aryloxyphenoxypropionates (fops) and cyclohexanediones (dims) have been present in the marketplace for more than 30 years. The only commercially available phenylpyrazoline ACCase inhibitor with selectivity in cereals is pinoxaden (Fig. 12; Hofer et al., 2006). The herbicide is a postemergence graminicide for a wide range of key annual grass species in cereals at rates of 30 to 60 g a.i. ha⁻¹. It is mixed with the safener cloquintocet-mexyl as mentioned previously. Pinoxaden also shows some activity against several ACCase inhibitor-resistant biotypes but is not active against all of them.

Figure 12.

The ACCase inhibitor pinoxaden.

A new chemical entry within a well-known MoA is pyroxasulfone (Fig. 13), which belongs to a new class of isoxazoline herbicides and is an inhibitor of the synthesis of very-long-chain fatty acids (Nakatani et al., 2012). This new herbicide demonstrates excellent efficacy against a broad range of grass and broadleaf weeds with both preemergence and postemergence activities. It is selective for use on corn, soybean (Glycine max), cereals, and cotton (Gossypium hirsutum) at application rates between 50 and 250 g a.i. ha⁻¹. More importantly, the herbicide provides effective control of trifluralin-, ALS-, and ACCase-resistant Lolium rigidum in Australia and of glyphosate-resistant Amaranthus rudis in the United States. In addition, it has a favorable soil residual profile that allows its application to be extended from the very early preplanting stage through postemergence stages without consequences to following crops. The compound was discovered by Kumiai Chemical Industry and is being developed by several companies for different crops. A second compound of the same class is fenoxasulfone (Fig. 13), which is currently undergoing development as a selective rice herbicide in Japan. Table I summarizes the above-mentioned new herbicides of the 21st century.

Figure 13.

Pyroxasulfone and fenoxasulfone.

Table I. Summary of some new compounds developed for weed control after 2000.

Post, Postemergence application; Pre, preemergence application.

MoA, Target Site	Examples	Use	Launch Date
Auxins	Aminopyralid	Post, rangeland, industrial sites, dicots	2005
	Aminocyclopyrachlor	Post, noncrop, brush control	2010
	Halauxifen-methyl	Post, dicots in cereals	Expected 2014
Cellulose biosynthesis	Indaziflam	Plantations, turf	2010
Acetohydroxyacid synthase or ALS inhibitors	Mesosulfuron-methyl	Post, cereals, grasses	2001
	Thiencarbazone-methyl	Post, cereals and corn	2008
	Pyroxsulam	Post, cereals, grasses	2008
HPPD inhibitors	Topramezone	Post, corn	2006
	Tembotrione	Post, corn	2007
	Pyrasulfotole	Post, cereals, dicots	2008
	Tefuryltrione	Post, rice	2010
	Bicyclopyrone	Post, corn and sugarcane	Unknown
	Fenquinotrione	Not specified	Unknown
Protoporphyrinogen oxidase	Saflufenacil	Pre and Post, various crops	2010
ACCase	Pinoxaden	Post, cereals, grasses	2006
Very-long-chain fatty acid biosynthesis	Pyroxasulfone	Pre and Post, various crops, monocots and dicots	2012
	Fenoxasulfone	Pre and Post, various crops, monocots and dicots	Unknown

HERBICIDE RESISTANCE

Herbicide resistance has been defined in numerous ways (WSSA, 1998; Heap and LeBaron, 2001; HRAC, 2014), but ultimately the definitions agree that a resistant weed is one that survives and reproduces following an herbicide treatment that would normally kill it. The selection of survivors with existing traits that are present in a population at a relatively low frequency is generally considered to be the antecedent to resistance and is set in motion through the intensity of selection pressure (Holt and LeBaron, 1990; Neve et al., 2009; Powles and Yu, 2010; Délye et al., 2013). Survival of an herbicide treatment results in selection of individual plants with the enabling resistance trait or traits, giving them an opportunity to pass these on to future generations. Several factors, including the biology and genetics of the weed species, herbicide chemistry, and its MoA, as well as key agroecosystem characteristics and herbicide application and handling, influence the development of herbicide resistance, which follows evolutionary processes (Darmency, 1994; Jasieniuk et al., 1996; Christoffers, 1999; Powles and Yu, 2010).

The possibility of herbicide resistance was first predicted over 50 years ago (Harper, 1956), just over a decade after the introduction of the first modern commercial herbicide, 2,4-dichlorophenoxyacetic acid, in 1945 (Peterson, 1967). The first occurrences of resistance were reported just one year after Harper’s prophetic publication in two disparate cases (Heap, 2014). We have been living with resistance to an increasing extent ever since. After a relatively quiet period in the 1950s and 1960s, the first big wave of resistance hit the PSII inhibitors, which include the triazines (Herbicide Resistance Action Committee [HRAC] group C1), in the 1970s, which was followed a decade later by the next wave of resistance to ALS inhibitors and ACCase inhibitors beginning in the mid-1980s (Fig. 14). It is often enlightening to revisit discussions of the past, where fears of resistance predominantly to products with longer soil residual activity were characterized as the major issue (Herbicide resistance: a call for industry action, 1990). This has now been eclipsed by fears of resistance to products with little to no soil residual activity and to products that are primarily applied to foliage. A few years after the introduction of crops resistant to glyphosate to the North American market in 1996, the first case of resistance development by the weed Conyza canadensis in a row crop (soybean) was reported (VanGessel, 2001). Glyphosate resistance in weeds, however, was already detected in a population of L. rigidum in Australia as early as 1995 (Pratley et al., 1996). Since then, the number of weed species resistant to glyphosate has been growing steadily (Fig. 14), and this situation has been reviewed extensively (Powles, 2008; Duke and Powles, 2009; Nandula, 2010; Vencill et al., 2012). Much has been said therein about what is driving this phenomenon. Despite awareness of the problem and recognition of the danger of continuing the use of glyphosate as the sole weed-control measure, many farmers have been reluctant to change (Prince et al., 2012), even if studies show long-term benefits to proactive resistance management (Norsworthy et al., 2012).

Figure 14.

Number of resistant species for herbicides with selected sites of action. HRAC codes are in parentheses. Note that PSII inhibitors are combined. (With permission of Dr. Ian Heap, WeedScience.org, 2014.)

The rapid adoption of glyphosate as the single weed-control measure in major American row crops, particularly soybean and cotton, had a profound effect not only on farmers but also on the agrochemical industry. It led to a loss of overall herbicide market value, reduction in the use of other herbicides, and thus directly and indirectly contributed to significant reductions in investment into herbicide discovery (Duke, 2005, 2012). These factors and the loss of intellectual capacity (chemists and biologists) that followed the protracted consolidation of the industry (Rüegg et al., 2007; Duke, 2012) are partially responsible for the lack of introductions of herbicides with new MoAs over the last two decades. The industry is still recovering from this downturn.

Resistance Confirmation

The first indication of resistance to an herbicide in a field is often a report of nonperformance. Many cases of reported resistance are actually weed-control failures due to other causes, attributed usually to either agronomic or climatic factors (Bayer CropScience, unpublished data). Thus, proper testing methods are extremely important to correctly assess whether the lack of expected efficacy was due to an agronomic issue or truly due to resistance. There is a well-accepted approach on how to respond to a field complaint where resistance is suspected. The first step is to record detailed field observations, including the herbicide treatment history; the next step is to properly sample seeds and then to test them in the greenhouse using (preferably) whole-pot assay techniques; and the last, but extremely important, step is interpreting the results in the proper context (Moss, 1999; Beckie et al., 2000; Burgos et al., 2013). It is better to test using more than just one discriminating rate and to generate dose-response curves using several rates in order to determine the resistance factor or index correctly (Moss, 1999). This is particularly important with populations that have non-target-site mechanisms of resistance, especially enhanced metabolism, because these types of resistance impart variable levels of tolerance to herbicides (Beffa et al., 2012). Figure 15 shows an illustration of the variability of enhanced metabolism within populations that exhibit various degrees of enhanced metabolism. The radiochromatograms of four individual Alopecurus myosuroides plants from five populations, two sensitive and three resistant to mesosulfuron-methyl, demonstrated relatively low levels of metabolite formation in the sensitive plants. In a few of the sensitive plants, some metabolism had occurred, as shown by the presence of metabolites. The more rapid degradation of mesosulfuron-methyl in these individual plants results in a higher degree of resistance, which is represented by the presence of greener plants among the dying, yellow plants in the pot. However, in most of the resistant plants, a large number of metabolite peaks are observed with a corresponding decrease in the mesosulfuron-methyl peak. In some of the plants, the relative size of the intact mesosulfuron-methyl peak is very low compared with some of the metabolite peaks, indicating that it has been metabolized extensively and is no longer present at a concentration high enough to injure the weed. In one plant from one of the resistant populations, hardly any metabolism has occurred (Fig. 15, top row, middle radiochromatogram), indicating that this plant is most likely sensitive. The accompanying photograph illustrates that plants exhibiting enhanced metabolism as a resistance mechanism can show a high degree of variability within a population, partly due to the complex polygenic control and the accumulation of resistance alleles over several selection cycles (Busi et al., 2012; Délye, 2013).

Figure 15.

Radiochromatograms of sensitive (S) and resistant (R) A. myosuroides plants incubated with ¹⁴C-labeled mesosulfuron-methyl. Triangles indicate the positions of the chromatographic peak of the intact active substance (mesosulfuron-methyl), and peaks to the left are those of inactive metabolites. (From Beffa et al. [2012].)

Resistance Mechanisms

Weeds have evolved numerous mechanisms of resistance that can be classified broadly into two main types, target site and nontarget site (Powles and Yu, 2010; Beckie and Tardif, 2012). Mutations to the target site that confer resistance have been well studied, whereas nontarget resistance mechanisms remain less clear (Powles and Yu, 2010). The first group of resistance mechanisms, collectively known as target-site resistance (TSR), includes all modifications of proteins targeted by herbicides, including gene coding sequence mutations, gene overexpression, and gene duplication (Powles and Yu, 2010; Délye et al., 2013). TSR generally confers a relatively narrow and generally high level of resistance to weeds within a single MoA, but digressions from this do occur (Powles and Yu, 2010). Alteration of the target site through mutations that modify herbicide binding and thus herbicidal efficacy can usually be effected by a single nucleotide substitution, hence making it relatively easy to select for this type of resistance (Yu and Powles, 2014). There can be differences in resistance expression to a particular target-site mutation between subgroups (chemical classes) within a single MoA, as for example between the aryloxyphenoxyproprionate, cyclohexadione, and phenylpyrazoline classes within the ACCase inhibitors (Yu et al., 2007; Délye et al., 2008). Other types of TSR, such as enhanced enzyme expression or increased gene copy number, can increase the number of active enzymes and thus sufficiently dilute the effective relative concentration of an herbicide, conferring resistance (Gaines et al., 2010; Délye, 2013). The second group of resistance mechanisms is known collectively as non-target-site resistance (NTSR), where processes not directly involving the targeted proteins, such as modification of the herbicide penetration into the plant, decreased rate of herbicide translocation, increased rate of herbicide sequestration, or metabolism, confer resistance (Powles and Yu, 2010; Délye, 2013). NTSR, especially enhanced metabolic resistance, can confer resistance to a much broader range of herbicides (Powles and Yu, 2010; Délye, 2013). They are surmised to develop through an accumulation of different mechanisms and are likely polygenetic (Délye, 2013) and, thus, theoretically more difficult to evolve. The use of lower than full herbicide rates has been implicated in the selection of NTSR through cycles of selection of individuals with slightly enhanced metabolism and can evolve quite rapidly (Neve and Powles, 2005; Busi et al., 2012). Especially threatening for the future are herbicide-degrading cytochrome P450s, glutathione S-transferases, and other enzymes potentially able to detoxify current, relatively new, and future herbicides, even herbicides from new structural classes yet to be discovered (Powles and Yu, 2010). Despite extensive studies and reviews of herbicide resistance, the genetic issues associated with resistance evolution have not yet been investigated extensively (Powles and Shaner, 2001, Gressel, 2002, Busi et al., 2013). Modern molecular biology methods, and in particular new-generation sequencing, are now being implemented. Initial results have identified glutathione S-transferases and cytochrome P450 genes associated with enhanced metabolic resistance (Gaines et al., 2014). This approach not only contributes to helping increase our basic knowledge of this kind of resistance but also has the potential to allow the development of better diagnostic tools. As resistance becomes more complex, accurate and sensitive resistance diagnostics tools can contribute to making the best possible weed management decisions. An integrated approach to relieve selection pressure on herbicides is critical to preserve their usefulness.

Organizations

The HRAC is one of the organizations concerned with herbicide resistance. It is composed of representatives of the agrichemical industry (Table II) whose aim is to manage resistance by fostering a responsible attitude to herbicide use, support and promote research, understand causes of herbicide resistance, communicate effective resistance management strategies, and collaborate with public and private researchers (HRAC, 2014). It is supported financially by member companies and CropLife International and though without a set structure, its members meet regularly at global and regional levels to facilitate communication between industry members. It supports the International Survey of Herbicide-Resistant Weeds (Heap, 2014), a survey of confirmed resistance cases that is a good resource for the current state of resistance. One of the most recognized projects is the classification of herbicide MoAs as embodied in the World of Herbicides poster available online (http://www.hracglobal.com/Portals/5/moaposter.pdf). It is revised periodically to reflect new discoveries. HRAC supports and participates in local, regional, and global research into resistance to understand its causes and effects as well as outreach programs to bring the best and latest knowledge to management programs. Local HRAC organizations tailor their activities to specific issues within each area.

Table II. HRAC global member companies.

BASF
Bayer CropScience
Dow AgrowSciences
DuPont Crop Protection
FMC
Makhteshim Agan
Monsanto
Syngenta Crop Protection
Sumitomo

Other organizations, such as the Weed Science Society of America, the European Weed Research Society, the Asian-Pacific Weed Science Society, and la Asociación Latinoamericana de Malezas (the Latin American Weed Association), are mentioned here as examples of regional institutions sponsoring research and organizing regular conferences, meetings, and workshops on weed resistance.

Management Strategies

Recommendations for best management strategies begin with understanding the biology of the targeted weeds, understanding the situation in the particular field, using a diversified approach including preemergence and postemergence herbicides with multiple MoAs at labeled rates in sequences and mixtures, and the inclusion of nonchemical practices including cultivation where appropriate (Norsworthy et al., 2012; Walsh and Powles, 2014). The inclusion of nonchemical control methods and diversified cropping systems greatly aids consistency in weed control and slows the evolution of resistance (Beckie, 2006; Walsh and Powles, 2014). More research needs to be done in combining chemical and nonchemical methods in order to protect the continued utility of all herbicides. In response to the worsening resistance situation, we must reexamine our thinking about herbicides as the sole weed-control technology to be implemented simply out of convenience. We are facing the loss of many more chemical tools through resistance if we continue to rely exclusively on them. This loss would make weed management in many crops much more difficult and, perhaps, impossible. We must become better at implementing integrated approaches.

Future of Resistance

Acknowledgment of the current status of resistance as a threat to the production of some crops and its continued development in intensity and complexity has led to calls for new herbicide options or a new paradigm in weed control (Tranel et al., 2011). We need to understand it better. In previous years, much of the work by academia and industry focused on the goal of preventing herbicide resistance. Given the economic and other factors driving weed-control decisions on the farm, the situation is changing to resignation that resistance is inevitable and the best result that one can achieve is to delay the onset of resistance (Neve et al., 2011). The solution to resistance has been stated simply: getting farmers to add diversity in their weed-control programs to reduce selection pressure from any one means of control while at the same time keeping populations sufficiently controlled (Beckie, 2006). The key is making this argument compelling to farmers and offering effective, integrated management tools at an affordable cost. Until industry is successful at delivering new weed-control products, we must continue to protect the remaining chemical tools and increase the integration of nonchemical tools. Once a new herbicide is discovered and introduced into the market, all efforts should be made to protect it from the beginning of its market introduction.

HT CROPS

Safener technologies have allowed the introduction of novel weed-control solutions in a number of crops such as cereals, rice, and corn. Safeners for dicot crops such as soybean (Glycine max), canola (a special spring rapeseed form derived from various Brassica spp.), and sugar beet (Beta vulgaris), however, could not be found despite immense screening efforts by many companies. In the past, broad-spectrum, one-shot weed control was only possible with mixtures. The rapid development of breeding and molecular engineering tools at the end of the last century led some agrochemical companies to a completely new approach: the development of HT crops. The first HT crop worldwide was a glyphosate-tolerant soybean from Monsanto, which was deregulated and approved from use by growers in 1994 in the United States and commercialized in 1995 (for a list of other HT crops, see Table III). The first commercial example of herbicide tolerance in crop plants in Europe, bromoxynil tolerance based upon the expression of a bacterial nitrilase gene, was deregulated in 1994 (one month after the U.S. approval for Monsanto’s glyphosate-tolerant soybeans) and entered the market in 1995 (MacKenzie, 1994). It was developed by the French company SEITA. In 1995, several other HT crops received commercial approval in the United States (e.g. bromoxynil tolerance in cotton [Calgene] and glyphosate tolerance in cotton [Monsanto]) and Canada (e.g. glufosinate tolerance in canola [AgrEvo/PGS] and glyphosate tolerance in canola [Monsanto]). At that time, DuPont also was working on a transgenic ALS herbicide tolerance system (James and Krattiger, 1996). From the HT traits mentioned previously, only glyphosate and glufosinate tolerance have gained a significant market share, with glyphosate tolerance being far ahead. The adoption of herbicide tolerance traits took place at an unprecedented speed. In 2012, herbicide-tolerant soybeans and cotton had gained a market share of 81% worldwide. In 2013, 93% of soybean acreage, 85% of corn acreage, and 82% of cotton acreage in the United States has been planted with HT crops (Fernandez-Cornejo et al., 2014). Monsanto, together with the seed company KWS, introduced the glyphosate-tolerant sugar beet line H7-1 in the United States in 2007. Two years later, in 2009, approximately 96% of the U.S. sugar beet area was planted with this genetically modified line (Nehls et al., 2010). The ease of use and its efficacy against a broad range of weeds made glyphosate by far the most widely used herbicide (Powles, 2008).

Table III. Year of commercial introduction of herbicide-tolerant crops.

Year	Crop	Type
1992	Corn	Nontransgenic
1995	Canola, cotton, and soybean	Transgenic
2001	Wheat	Nontransgenic
2002	Rice	Nontransgenic
2003	Sunflower	Nontransgenic
2006	Lentil	Nontransgenic
2011	Winter oilseed rape (Brassica spp.)	Nontransgenic

In the United States, the lack of rotation to other HT crops and limited use of herbicides other than glyphosate can clearly be identified as the major factors contributing to the development of glyphosate resistance. Canadian farmers, on the other hand, have tended to rotate different HT systems, such as glyphosate-, glufosinate-, and to a much lesser degree imidazolinine-tolerant canola, with, as yet, only seven cases of weeds resistant to glyphosate documented in Canada (as of June 20, 2014; Heap, 2014). It is becoming clearer, meanwhile, that weed management should not rely on a single herbicide; rather, it is imperative to rotate and to use mixtures of herbicides with different MoAs. All major agrichemical companies still participating in herbicide research have reacted to this need. In their continuous efforts to control weeds, and especially glyphosate-resistant weeds, one common response from companies has been in the area of HT crops. Many companies have now started to transfer multiple genes conferring tolerance to several classes of herbicides with different MoAs into crops in order to provide farmers with more options within a season. Corn plants tolerant to both glyphosate and glufosinate have been developed recently. Glyphosate- and glufosinate-tolerant cotton plants have already been commercialized. Another platform for corn and soybean combines glyphosate tolerance with tolerance to 2,4-dichlorophenoxyacetic acid. This particular technology is awaiting its approval for commercialization in the United States and is anticipated to enter the market in 2015. The same is true for a platform that combines glyphosate tolerance with tolerance to dicamba, another herbicide from the auxin class, which is also anticipated to enter the market in 2015. A further approach is the combination of tolerance to glyphosate with tolerance to the sulfonylurea class of herbicides (ALS inhibitors). Finally, varieties tolerating HPPD inhibitors plus glyphosate and glufosinate will enter the market a few years from now.

Mutation Breeding of HT Crops

Since the discovery of naturally occurring HT mutant plants and those not involving gene transfers, HT crops have also been conventionally bred. For instance, various herbicide-resistant canola culture systems are currently available. Imidazolinone-tolerant Clearfield canola was achieved through microspore mutagenesis and selection with imazethapyr and conventional breeding. HT crops created by induced mutation and breeding are classified as nongenetically modified (non-GM) crops. Unlike genetically modified HT crops, no heterologous gene transfer is involved. The endogenous target gene is modified/mutated at the natural location in the plant’s genome; thus, position effects can be excluded.

Crop mutants can be created by different means (for review, see Meksem and Kahl, 2010). In brief, the following methods are used for HT mutant selection: selection of spontaneous mutants with the herbicide; chemical mutagenesis and selection with the herbicide; chemical mutagenesis and target sequencing (targeting induced local lesions in genomes); physically induced mutation with ion beams; and molecular breeding technologies.

In Vitro HT Mutant Selection

As a starting point for the selection of HT crop mutants, different plant material can be used in plant cell and tissue culture. Depending on the crop and its properties in tissue culture, the starting materials can be leaves, calluses, suspension cultures, protoplasts, microspore-derived embryos, and immature embryo-derived cultures. However, the prerequisite for successful mutant selection experiments is that the plant material must be rapidly growing with rapidly dividing cells that can be regenerated to fertile plants. The tissue culture approach is advisable when spontaneous mutations are required in the target gene for herbicide tolerance.

In Vivo HT Mutant Selection

Nature itself can be a source of non-GM HT crops with the selection of naturally occurring mutant plants. Another very successful method to isolate HT mutants is the chemical mutagenesis of seeds (e.g. with ethyl methanesulfonate), subsequently growing the seeds into the M1 and M2 generations followed by selection of HT mutants through herbicide applications or identifying mutations/mutants in the target gene through sequencing (targeting induced local lesions in genomes).

ALS-Tolerant Herbicide Systems

The Clearfield system confers tolerance to crops otherwise susceptible to imidazolinone (ALS) herbicides. It consists of two elements: non-GM imidazolinone-tolerant crops (Tan et al., 2005) and the respective imidazolinone herbicides that can be used selectively in the now tolerant crop. Since 1992, the Clearfield technology has been consequently introduced in several crops and launched to the market as shown in Table III. The Clearfield system of BASF is currently marketed as a win-win situation for the farmer and the industry. The advantages for the farmer are more weed control options. As a result, the advantage for the company is that the herbicide active ingredients from this class will be utilized on a much broader scale (Pfenning, 2013).

Novel Weed Control in Non-GM Sugar Beet

In Europe and countries in other regions that do not accept GM crops, there is a strong demand for effective one-pass solutions in all dicot crops due to the lack of selective herbicide innovation (e.g. for sugar beet). Phenmedipham-based products have contributed to reliable weed control in sugar beet for more than 40 years. However, no fundamentally new herbicide active ingredients in sugar beet have come onto the market for many years, unlike in other crops like wheat or corn. Thus, a project to select sugar beet mutants tolerant to ALS herbicides was started in 2001 by Bayer CropScience. The technology is based on the breeding of sugar beet varieties that are tolerant to herbicides in the ALS inhibitor class with broad-spectrum weed control (Hain et al., 2012a, 2012b). A mutant having a naturally occurring amino acid substitution at position 574 in the ALS enzyme, which is involved in the biosynthesis of essential branched-chain amino acids, was selected and used in further breeding. It was very important that these varieties are not a product of transfer to the crop genome from another organism, so that they could be registered in Europe as a non-GM crop. In spring 2012, Bayer CropScience and KWS signed an agreement to jointly develop and commercialize this system for weed control in sugar beet for the global market.

The novel herbicide tolerance trait was selected in Frankfurt approximately 10 years ago using sugar beet cell culture techniques. Out of about 1.5 billon cells tested, one herbicide-tolerant cell was selected and regenerated to produce a sugar beet plant labeled FM12-1, forming the basis for the development of the new weed-control system. The number of cells selected is equivalent to selecting a single sugar beet plant out of 15,000 ha of the crop. Subsequently, the HT trait has been introduced into the elite sugar beet germplasm of KWS by marker-assisted breeding.

Non-GM Sulfonylurea Tolerance in Soybean

At DuPont Pioneer, a soybean line was developed through seed mutagenesis and rounds of selection (Fig. 16) through the application of a sulfonylurea herbicide normally not tolerated by soybean (Sebastian et al., 1989). The mutant line displays a high degree of ALS-based resistance to both postemergence and preemergence applications of a variety of sulfonylurea herbicides (Walter et al., 2014).

Figure 16.

Selection of non-GM HT mutants.

NEW DISCOVERY APPROACHES

In parallel with the development of new herbicide-tolerant crops, new screening tools have been employed for the selection of chemicals with new herbicidal MoAs. With the discovery of new highly potent low-dose herbicide classes, specifically the sulfonylureas in the 1980s, requirements for compound quantities for herbicide testing decreased significantly, from grams to the low milligram scale. Subsequently, in the mid-1990s, combinatorial chemistry, a novel synthesis tool able to produce hundreds of thousands of new chemical entities in a relatively short time, became possible (Smith, 2003; Lindell and Scherkenbeck, 2005; Scherkenbeck and Lindell, 2005; Lindell et al., 2009). As a result, plant pot-based primary screening for lead identification was replaced with novel screening systems able to cope with high numbers of chemicals in a cost-effective way (Ridley et al., 1998). In response to this, all major agrochemical companies introduced high-throughput screening technologies: both in vitro high-throughput biochemical screening and high-throughput in vivo screening (HTVS); consequently, the number of compounds screened reached new heights (Fig. 17). However, soon after the new screening technology was adopted, it was realized that this enormous increase in screening input did not lead to the expected higher number of strong hits and subsequent development projects (Kraehmer, 2012). As a result, screening inputs were lowered again in favor of smaller and more diverse and targeted compound libraries (Fig. 17).

Figure 17.

Evolution of compound input to herbicide screening over time, highlighting the impact of combinatorial chemistry (dotted circle).

From a purely statistical point of view, today an average of 140,000 chemical compounds per indication need to be screened in order to bring one new crop protection product to the market (Phillips McDougall, 2010). This translates into 420,000 compounds needed on average for a continuous product flow in all indications: fungicides, insecticides, or herbicides, including safeners.

Successful agrochemical research requires a constant input of novel chemistry to the screening cascade, because once a chemical compound has proven to be inactive against the tested species, there is usually no reason to test it again. The primary objective is to find herbicidal activity. The next objective is to characterize this activity and potential for crop selectivity. The bar at the screening level is usually set low enough to ensure that activity is found at reasonable use rates. Sources of chemical innovations for herbicide research arise from in-house chemistry research, other indications, life science compound pools, commercial providers, academia, natural products, and others. The huge numbers of chemical compounds being processed require large-scale automated storage and retrieval systems for sample management together with powerful logistics, all serving the individual indications in an efficient way.

Screening is defined as the stepwise assessment of the biological activity of a compound leading to strong candidates for field development testing. The basic principles of compound screening in agrochemical research have been described in several review articles (Giles, 1989; Copping, 2002; Cobb and Reade, 2010). This process can be broken down into two main consecutive steps: lead finding and lead optimization. Usually, although sometimes named differently, the test procedure consists of a primary and secondary screening, followed by field trials (Giles, 1989). As a result of the strongly increased input numbers observed at the end of the last century, high-throughput screening systems were introduced as initial screening tools (Fig. 18). High-throughput screening in agrochemical discovery has been reviewed recently (Tietjen et al., 2005; Drewes et al., 2012). Another approach to discover new herbicidal precepts consists of the systematic analysis of plant gene functions (Lein et al., 2004). This approach is expected to aid the development of new herbicidal target assays. The study of small-molecule metabolite profiles, generally referred to as metabolomics (Kamp et al., 2012), and gene expression profiling (Eckes and Busch, 2012) are valuable tools in this context. Alternatively, new herbicidal leads also may arise from the combination of whole-plant screening with physiological investigations, recently defined as physionomics (Grossmann et al., 2012). All these approaches are covered under the general concept of systems biology, which is a more holistic approach to biological research (Kitano, 2002).

Figure 18.

Modern herbicide screening process based on screening technologies, with numbers of compounds and active ingredient requirements. HTBS, High-throughput biochemical screening.

One big advantage of agrochemical screening over pharmaceutical screening is that agrochemicals can be applied directly to the living target organisms in early screening stages. There is no intrinsic need to start with a model system. Relevant properties like compound uptake, speed of action, and metabolism are covered directly within whole-organism screening. However, hit identification in HTVS is limited to metabolically stable and bioavailable compounds. Any active ingredient with marginal stability or poor bioavailability can rarely be identified with whole-organism screening. In consequence, both techniques, high-throughput biochemical screening and HTVS, due to their complementarity, are required in agrochemical discovery.

In HTVS, the 96-well microtiter plate format is used extensively across the agrochemical industry. Substance requirements to provide information on the herbicidal potential against selected target plants are generally low (Tietjen et al., 2005). There are several criteria to consider concerning the plant species being used in HTVS (e.g. the size of the seeds, germination potential, ease of visual assessment, representation of the species in downstream screening levels, and the relative importance of the test plant with respect to the target markets).

Despite being employed extensively, the use of automated 96-well microtiter plate assessment techniques remains a challenge when screening using HTVS. Innovations in small-scale whole-plant imaging technologies remain very limited compared with other recent developments, such as high-content screening or other areas of image analysis, that have taken place in pharmaceutical research (Haney, 2008). Plant growth assessment using image analysis ideally provides a three-dimensional view, but this makes it very challenging to fulfill all the requirements for fully automated systems applied to continually growing plants. Today, the numbers of sensors for plant phenotyping are numerous: red-green-blue, fluorescence, and near infrared are standard technologies that are used for many assays with image acquisition restricted to a two-dimensional perspective when considering the top view for microtiter plates. The high density of plants grown in a small area restricts the use of available standard technologies. On the other hand, there is good progress in image-analysis software. Powerful hardware and commercial imaging software enable trained users to evaluate herbicide screening trials (also restricted to two dimensions) in totally different ways. The combination of new sensors, time-lapse imaging, and ultra-high-quality images provides much deeper insights than any standard visual assessment technologies or techniques. Software tools like Metamorph by Molecular Devices or the Lemnatec image-analysis platforms are well-established tools to support the screener in plant phenotyping to extract useful information out of images.

With the introduction of high-throughput screening, an enormous increase in test data followed, exceeding 100,000 data points per day on every single technology platform. These experimental raw data need to be stored in appropriate databases and processed for the development of structure-activity relationships. Research at Bayer CropScience, for example, applies ActivityBase from ID Business Solutions as a data management tool and Spotfire DecisionSite for the visualization of screening results (Tietjen et al., 2005). This, together with tailored in-house information technology solutions, permits a rapid correlation of biological results for the high numbers of chemical structures over all screening levels. In this context, it has to be stressed that a close and effective interaction between the individual research departments, specifically chemistry, biology, and biochemistry, is crucial, given the fact that the discovery of a development candidate is primarily based on iterative cycles of syntheses and screening, thus optimizing initial lead compounds, rather than just filtering the right compound from a big substance pool. Finding a product like this (merely by filtering), in fact, is a very rare event.

OUTLOOK

The world is likely facing its biggest challenge ever in our ability to feed the global population. According to the Food and Agriculture Organization of the United Nations, agricultural production must increase by 70% until 2050 to supply 9 billion people with sufficient food (Bruinsma, 2009). Current yield trends suggest that our efforts to raise production are insufficient. A turnaround for yield increases in broad acre crops as the basis for world food security is urgently required. Furthermore, by 2050, more than 6 billion people (67%) will live in urban areas, ranging from an urbanization level of 86% in developed countries to 64% in less developed regions (United Nations, 2012). In view of limited possibilities to expand the area of arable land, high-yielding but sustainable agriculture is the only plausible alternative. Preservation of soil, maintenance of soil fertility, and high water use efficiency, as well as maintaining high levels of local biodiversity, are of key importance for sustainably managing the global bioeconomy.

To fully exploit the maximum yield potential in crop production, weed control with management programs based upon effective herbicides is of utmost importance for sustainable land use. Along with preventing potential yield losses, other measures to help improve soil fertility, including minimizing wind and water erosion as well as enabling an increase of organic matter, are necessary. Today’s modern crop production has all too often abandoned diverse crop rotations and tillage, trading them for monoculture crops and no-till/low-till technologies relying on an effective herbicide technology. However, after several decades of herbicide use, weed resistance to major chemical classes continues to spread further. Currently, more than 60% of the global herbicide market (value) is represented by products from only four MoAs, all of which actually have serious resistance issues (Fig. 19). With an additional three MoAs, together they cover approximately 80% of the global market in value. The situation is similar for resistance to fungicides and insecticides, where approximately 75% of the global market is served by six and four MoAs, respectively (Casida, 2009). Due to the slower growth of weed-control markets during the past few decades, increased costs for discovering and developing new active ingredients as well as the impact of glyphosate crop tolerance adoption in the immense North and South American corn and soybean markets led to significantly reduced efforts in global herbicide discovery. As a consequence of these factors, the herbicide industry has undergone a protracted consolidation process, which continues to impact agriculture.

Figure 19.

Herbicide market 2012 per MoA (without trait fees). The shaded area indicates MoAs with significant resistance problems. VLCFA, Very-long-chain fatty acid biosynthesis; EPSPS, 5-enolpyruvylshikimate 3-phosphate synthase.

The analysis of published patent applications for new active ingredients indicates a striking decrease in the number of applications since 1990 (Fig. 20). It is obvious that only a very small group of companies remain actively engaged in the discovery of novel herbicide candidates. A review of the known development pipeline shows the same tendency. It is estimated that from now until 2020, only five to 10 new herbicides will enter the market. Another indication of the effects of the market forces on herbicide discovery is the number of products introduced by decade. Whereas approximately 50 new active ingredients were introduced per decade in the 1980s and 1990s, only approximately 20 herbicides were launched within the last decade (Fig. 21).

Figure 20.

Active herbicide ingredient patent applications from 1990 until 2012. [See online article for color version of this figure.]

Figure 21.

Active herbicide ingredient introductions by decade from 1945 until 2015. [See online article for color version of this figure.]

Due to an increasing lack of effective herbicide solutions and an increase in multiple resistant populations, weed control has become more complex in order to combat resistant weed species in major broad acre crops. The soaring agroeconomy as well as greater inputs into weed management have induced further growth of the herbicide market, which will exceed 20 billion euros per year by 2020 (http://www.agro.basf.com/agr/AP-Internet/en/function/conversions:/publish/content/news_room/news/downloads/11-03-10-press-release-basf-crop-protection-pipeline-value-jumps-to-2.4-billion-eur.pdf). The demand for new resistance management solutions is rewarding the renewed focus on herbicide discovery. However, the regulatory requirements to develop and register new herbicides are ever increasing, especially in Europe. Consequently, the total cost for the discovery and development of one new herbicidal active ingredient is approaching 200 million euros (Phillips McDougall, 2012). These costs could continue to increase further.

To achieve a sufficient return on investment for those rising research and development costs, the industry requires increasing business opportunities per development candidate. The development of single new herbicides exclusively for smaller or niche crops is economically unfavorable. The threshold for entry of new players in herbicide discovery is extremely high; thus, no new herbicide research-oriented companies are expected to enter the market. However, various research institutes in China are increasingly engaged in herbicide discovery, but thus far they are lacking in internationally focused, integrated weed research entities.

Due to recent extraordinary market dynamics, some companies are once again increasing their engagement in herbicide discovery. Their main strategies are MoA and target identification of herbicidal leads followed by high-throughput screening and/or structure-based design; design of new herbicidal structural scaffolds with validated targets; and agrophore synthesis strategies.

To fully exploit the discovery potential, a fully integrated approach employing all state-of-the-art technologies and scientific approaches is essential. However, based on an analysis and assessment of financial analyst publications as well as patent applications (Fig. 20), a market introduction of significant new herbicide classes having new resistance-breaking MoAs probably can only be expected after 2025.

Therefore, in the near future, all stakeholders must contribute to the sustainability of current herbicides by adopting diversity in crop rotations, herbicide combinations, and sequences as well as other nonchemical measures. The industry has to take responsibility to redouble its herbicide discovery efforts and supply agriculture with new, effective resistance-breaking herbicides.

Glossary

HT

herbicide-tolerant

ALS

acetolactate synthase

TCM

thiencarbazone-methyl

HPPD

4-hydroxyphenylpyruvate dioxygenase

ACCase

acetyl-coenzyme A carboxylase

MoA

mode of action

HRAC

Herbicide Resistance Action Committee

TSR

target-site resistance

NTSR

non-target-site resistance

non-GM

nongenetically modified

HTVS

high-throughput in vivo