Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Pesticide Resistance: Strategies and Tactics for Management. 1986. National Academy Press, Washington, D.C. Modes and Genetics of Herbicide Resistance in Plants JONATHAN GRESSEL Herbicide resistance is becoming an increasing problem through- out the world, but one that can be managed with the right tools and by understanding how plants develop resistance to the various her- bicides. Population genetics models can help scientists to discern, broadly, why resistance occurs or will occur in some situations and not in others (i.e., why resistance has not developed in monoculture, monoherbicide wheat, but why it has developed in corny. Genetics and molecular biology allow scientists to understand the details of resistance development and the types of inherited resistance: nuclear with dominance, recessiveness, monogenic, polygenic, organelle, and gene duplication. Herbicides act on plants in different ways. By understanding all the processes, better methods artful strategies of delaying or managing resistance to herbicides can be devised. INTRODUCTION The idea of weeds becoming resistant to herbicides is not new. Warnings about the possibility of weeds evolving resistance were issued soon after the phenoxy herbicides were introduced (Abel, 1954~; however, as no confirmed cases of resistance to phenoxy herbicides occurred, the warnings were ig- nored even after the first triazine-resistant weeds appeared. In Europe and the United States triazine resistance has become a serious problem: at least 42 species have resistant biotypes. Six weed species are resistant to paraquat; one weed species each is resistant to diclofop-methyl and trifluralin. All evolved from sensitive biotypes in agricultural situations (Figure 11. For example, more than 75 percent of Hungary's (the Eastern block's major 54 -
HERBICIDE RESISTANCE IN PD4NTS Trifluralin /Eleusine o c 100 Q 80 60 40 20 Alrazine/Seneclo At- ~1 0 0.28 1.12 4.48 55 loon Al I ~ v v I, I. ~ I l l 1- All I I I I 1_ ;~ _ I,, Paraquat/Erigeron W Diclofop-Me/Lolium ~ I ~,: ~ ' ~L . ~ . . 41 1 1 ~1~ 1 o ( ) 0.28 Q56 1.12 2.24 .187 .375 .75 kg /ha 1.5 ~.5 1 2 4 8 16 11,~ 1 1 H~ \Atrazine/Siolonum \nigrum populations \\\ e \ Atrazine/ \ ~ ~/ from' \\\ \ Amaranthus be Legarno \b~\ \ blitoides . ~ V Montereale \ ~~ _ Vivaro Ha ~\ I ~ he Roveredo in \ \ `0 ~ _ Piano \ ~ | (Italy) ~ To- \ 1 is, , ~ W7 ~1 1 ~1 ~ 1 1 - 3 ~ ~ 10 ~ 30 4' 0.1 0.3 l 25 kg /ha FIGURE 1 Dose response curves of the wild type and the herbicide-resistant weeds that have evolved. (Box near base of each graph denotes the recommended agricultural rates for each herbicide; concentrations are on log scales.) A: Resistanee of Eleusine indica, the fifth worst weed in the world (Holm et al., 1977), evolved in South Carolina, after about 10 years of tr~fluralin use as the sole herbicide in monoculture cotton. Dose response curves vary among separately evolved resistant biotypes. (Figure plotted from tabular data in Mudge et al., 1984.) B: Resistanee of Lolium rigidum to diclofop-methyl in legume fields receiving six applications in four years. (Redrawn from data of Heap and Knight, 1982.) C: Tolerance of Erigeron philadelphicus (=Conyza philadelphicus) to paraquat. Multiple yearly applications of paraquat were used as the sole herbicide in mulberry plantations. (Redrawn from data of Watanabe et al., 1982.) D: Resistanee of Senecio vulgaris to atrazine appeared in a nursery where atrazine and simazine were used once or twice annually for 10 years. Data measured as survival after preemergence treatment. (Plotted from tabular data in Ryan, 1970.) E: Variable response of s-tr~azine- resistant accessions of Solanum nigrum. Seeds of the resistant biotypes were gathered from the four isolated places listed (in Northern Italy) and were assayed in pot tests. (Plotted from tabular data in Zanin et al., 1981.) F: The appearance of atrazine-resistant Amaranthus blitoides. Monoculture maize fields were treated for 17 years with atrazine. A 1 m2 area was found with this accession in Hungary. (Plotted from unpublished data supplied by Dr. P. Solymosi, Plant Protection Inst., Budapest, 1982.) maize-growing area) agricultural land is infested with triazine-resistant pig- weeds (Hartmann, 1979; Solymosi, 19811; s-triazine can be used only in mixtures. Tolerancei to herbicides continues to increase (LeBaron and Gres ~Tolerance is defined as any decrease in susceptibility, compared with the wild type. Resistance is complete tolerance to agriculturally used levels of a herbicide (LeBaron and Gressel, 1982).
56 MECHANISMS OF RESISTANCE TO PESTICIDES set, 1982), and resistance to other herbicides may soon appear in the field (Gressel, 1985~. The problem has been compounded because of the low price of atrazine and its superior season-long control of weeds in corn when compared with the phenoxy herbicides. It is 20 percent cheaper to treat corn with atrazine than with 2,4-D (Ammon and Irla, 1984~. Thus farmers use more atrazine per year, and many stop rotating crops and herbicides. Resistance to the triazines and other herbicides has appeared in the agricultural areas of mon- oculture, monoherbicide use. The potential economic risk is great: while it now costs cat $12/ha to treat sensitive weeds with atrazine, if all major corn weeds become resistant the alternative treatments would cost cat $125/ha (Ammon and IrIa, 1984~. A second problem involving resistant weeds is "problem soils." Repeated applications of herbicides can create problem soils when soil-applied her- bicides can no longer control susceptible weeds. In such soils herbicides are degraded more quickly than in nonproblem soils (Kaufman et al., in press). For example, the rate of EPTC degradation more than doubles in soils that receive multiple treatments of EPTC, and there is a 50-fold increase in degradation in soils with a 12-year history of repeated diphenamid applica- tions (Kaufman et al., in press). The problem becomes greater because the microbial enzymes degrading these pesticides often have a broad specificity that leads to cross-resistances within herbicides and between groups of her- bicides and some other pesticides (Kaufman et al., in press). It is possible to conceive of the use of herbicide "extenders" that would act by inhibiting the specific soil microorganisms or the degradative enzymes' systems. By analogy it is possible to conceive the scientific feasibility of doing this from the effective specific inhibition of ammonia-oxidizing bacteria by nitrapyrin. In this chapter we will look at the basic genetics, biochemistry, and phys- iology of resistance so that we can make recommendations that will delay the appearance and spread of resistance to our most cost-effective herbicides. POPULATION GENETICS Simple population genetics models suggest why there has been no resis- tance to phenoxy herbicides in monoculture, monoherbicide wheat and why resistance to triazines, especially in corn, has become so widespread (Gressel and Segel, 1978, 1982~. These models, along with common sense and a closer look at crop and weed ecologies and agronomy, can help us develop strategies to delay resistance. The appearance of resistance depends on characteristics of the different weeds and herbicides, which can be mathematically integrated into models. If a gene or genes for resistance do not exist at some low frequency in the population, resistance will never appear in that species unless introduced by
HERBICIDE RESISTANCE IN PLANTS 57 genetic engineering. When resistant biotypes are grown in competition with susceptible (wild-type) biotypes of the same species without herbicides, their seed yield is often about one-half that of the wild type (Radosevich and Holt, 1982; Gressel, 1985; Gressel and Ben-Sinai, in press). Fitness will decrease the rate of enrichment for resistance when nonpersistent herbicides such as 2,4-D are used, since much of the season will be available for the remaining susceptible individuals to exert their superiority. This competition between fit and unfit biotypes is especially fierce when seedlings are established. Persistence of herbicides interrelates not only with fitness but also with dormancy characteristics that separate weeds from crops and from other pests the spaced germination of weed seeds. Weeds germinate not only throughout the season, but also over many seasons. Susceptible weed seeds can germinate after a rapidly degraded herbicide has disappeared; they then produce more seeds before the season is over, considerably lowering the effective selection pressure. Selection pressure is a result of "effective kill," which is not the same as the "knock down" after herbicide treatment. Ef- fective kill is a measure of the number of surviving seeds or propagules at the end of a season, not after treatment. Every time we enrich for resistant individuals by using a herbicide the resistant seeds are diluted by a seed bank of susceptible seeds from previous years. These seeds exert a buffering effect and delay the appearance of resistance. The first weed reported to evolve triazine resistance, Senecio vulgaris (Ryan, 1970), does not have an appreciable seed bank. The inter- action of selection pressure, herbicide persistence, and seed bank on the rates of enrichment for resistance can be modeled to visualize how each parameter affects the rate at which resistance should appear. Similar modeling has been done for the evolution of insecticide resistance (Georghiou and Taylor, 1977), for fungicide resistance (Delp, 1981), and for resistance of cancer cells to antitumor drugs (Goldie and Coldman, 1979~. In our model (Gressel and Segel, 1978, 1982) the factors governing the rates of evolution of herbicide-resistant weeds, including the effects of the seed bank, are expressed in the equation: Nn = No (1 + faln~n, where Nn is the proportion of resistants in a population in the nth year of continued treatment of a herbicide, and No is the initial frequency of resistant individuals in the field before herbicide treatment. No is a steady state achieved by natural mutation to resistance, lowered by the fitness of a biotype. The factor in parentheses governs the rate of increase of resistance. The overall fitness f (measured without the presence of herbicide) is that of the resistant compared with the susceptible biotype. With triazine resistance f is usually between 0.3 and 0.5 (Gressel, 1985; Gressel and Ben-Sinai, in press). Se- lection pressure (~) is defined as the proportion of the resistant propagules
58 MECHANISMS OF RESISTANCE TO PESTICIDES I o-o Resistance visible in the field~- 30%R ~ f I ~i ~, ,' L,' , , ~, I , ~ , low in (n o 10 6 2 lo-8 o 1 o-lo 10-4 r . t<~40~ ~; //C~\O' /~ So Beforel o iXoCOt~ 116°~- / /C f=Fitness of resistants e-Average seed longevity EK= Effective kil I , I I I I I I I I I I 1 1 10 20 SEASONS REPEATED TREaTt\/lENT 30 lolo 8 tn z lo6 ~ _ 104 ~ of _ 100 _ O Z FIGURE 2 Effects of various combinations of selection pressure ~ (measured as effective kill, EK) fitness and of soil seed-bank longevity (n) on the rates of enrichment of herbicide- resistant individuals over many seasons of repeated treatment. The values are plotted for fitnesses that would develop after the herbicide degrades. With the persistent treading herbicide the fitness (f = 1.0; f = 0.8) would be high, as the fitness differential has no time to become apparent. With the phenoxy type, fitness differentials (f = 0.6; f = 0.4) will have time to be influential. Resistance (R) would become apparent in the field only when more than 30 percent of the plants are resistant. The scale on the right indicates the increase in resistance from any unknown initial frequency of resistant weeds in the population; whereas the scale on the left starts from a theoretically expected frequency of a recessive monogene character in a diploid organism. (Plotted from equations in Gressel and Segel, 1978.) divided by the proportion of susceptibles at the end of the season. For example, if all resistants remain and all but 5 percent of the susceptibles are lost, ax = 1/0.05 = 20. Selection pressure and fitness are divided by n, approximately the half-life of seed in soil. In weeds that germinate imme- diately, such as Senecio, n = 1. With most weed species, n is between two and five years. An increase in n depresses the rate at which resistance will increase. The interrelationships are clearer when we use the equation to generate hypothetical lines from different scenarios (Figure 21. In Figure 2 we arbi- trarily started in year zero from a frequency of 10- in, but the frequency scale can be moved to fit any initial field frequency. More important are the slopes showing the ratio at which enrichment occurs. The slopes show that we always enrich herbicide-resistant individuals when we treat with herbicides.
HERBICIDE RESISTANCE IN PLANTS 59 It takes many years for the frequency of resistant weeds to become no- ticeable (i.e., more than the 1 to 10 percent that remain after a herbicide treatment). Thus, we do not realize we are enriching for herbicide resistance until it is upon us. Note in Figure 2 the rates of enrichment for the triazine versus the phenoxy herbicides. The selection pressures are estimated, since the proper ecological studies have not yet been done. The phenoxy herbicides have a much lower selection pressure because their shorter soil persistence allows late-season weed germination. Even without this late-season germination the actual ef- fective selection pressure of the phenoxies is lower than the triazines. A midseason survey of North Dakota wheat fields showed that the best control of Amaranthus retroJ?exus, Chenopodium album, and Brassica campestris with a phenoxy herbicide gave 0.2 plants (probably all "escaped" suscep- tibles) of each weed per m2 (Dexter et al., 19811. Considering the plasticity of these weeds, there would be hundreds of seeds per m2 for a good stand of susceptible weeds the following year. The population genetics models (Figure 2) can also predict what happens when a monoherbicide culture is not used. In the model the number of weed generations that are treated affects enrichment. If it takes 10 years with no herbicide rotation to obtain resistance, it would take 20 and 30 years in 1 in 2 or 1 in 3 herbicide rotations, respectively. Indeed, all s-triazine resistance has come from monoherbicide cultures. If these theories are true, triazine resistance should have developed in the U.S. corn belt, where corn with atrazine has been grown in a one- in two- year rotation. This has not happened, but it may still be too soon to expect resistance to appear, or it may be that rotation is a more potent tool to decrease the rate of resistance than previously thought. Herbicide mixtures (atrazine plus an acetamide) are also used widely in the U.S. corn belt. No triazine resistance has appeared where such mixtures are used. Gressel and Segel (1982) and Gressel (in press [a]) provide theoretical analyses of the effects of such mixtures. BIOCHEMICAL AND PHYSIOLOGICAL MODES OF RESISTANCE s-Triazines The s-triazine herbicides, as well as many phenyl-urea and uracil herbi- cides, inhibit photosynthetic electron transport on the reducing side of pho- tosystem II in leaf plastics. These herbicides loosely bind to the thylakoids. Death occurs from release of free radicals, or chlorophyll photo-bleaching, or starvation for photosynthate. The first sign of damage is an immediate rise in chlorophyll fluorescence (Figure 3A). These herbicides also inhibit photosynthesis in the crops where they are
60 MECHANISMS OF RESISTANCE TO PESTICIDES 3 q 2 In 11 o A.lchenopodium leoves B ATRAZINE I S+ atrazine _ _ _ \ _- ~/usceptible R+atrazine ~ Res stant 0 100120 180 0.05 0.1 0.15 Illumination (see] Free Atrazine (,u M) cP 1.5= 1.0 m .c 0.5 0 c FIGURE 3 Special properties of most evolved atrazine-resistant weeds. A: Lack of in- creased chlorophyll fluorescence due to treatment with atrazine. Whole leaves of Chen- opodium album R (resistant) and S (susceptible) biotypes were treated with 15 ,uM atrazine 2h before scanning. (Redrawn from Ducruet and Gasquez, 1978.) The scan of the R biotype with atrazine is similar to the scans of R and S biotypes without atrazine. B: Specific loss of atrazine binding site in thylakoids from triazine resistant weeds. Binding of ~4C-atrazine to susceptible and resistant chloroplast membranes was measured. (Re- drawn from Pfister and Arntzen, 1979.) used: corn and orchards. Corn, however, is unique; it has high levels of a glutathione-S-transferase (GST) that conjugates glutathione to atrazine and simazine and detoxifies them before they can do lasting damage. In orchards simazine binds to the upper layer of soil; it does not reach the roots of trees, but lethally partitions into weed seedlings growing through this layer. The herbicide-resistant weeds that appear in orchards and corn fields, however, do not have the enhanced rate of atrazine degradation as appears in corn. Instead, the plastics of these weeds are resistant because the triazines did not bind to thylakoids (Figure 3B) (Arntzen et al., 1982; Gressel, 19851. The simplest field test for this type of resistance is to use a field fluorometer modified from the designs of Ducruet and Gasquez (19781. One takes a fluorescence reading on a leaf, applies atrazine, and later takes another reading (Ahrens et al., 1981; All and Souza-Machado, 19814. Fluorescence in the resistant biotypes will not change, but it will increase in the susceptible types (Figure 3A). The levels of resistance in weeds with the plastic-type triazine resistance are quite variable. Most evolved biotypes have the type of resistance shown in Figure ID; saturating doses of atrazine, many times the levels used in agriculture, have no effect on the weed. Some resistant biotypes are inhibited differently by such rates (Figure 1E); marginally resistant biotypes have similar reactions at normal concentrations (Figure IF). Weed germination in the last probably occurs after some of the atrazine has been biodegraded. Triazine tolerance and resistance evolve differently even in the same spe
HERBICIDE RESISTANCE IN PINTS TABLE 1 Differences in Inherent Tolerance to Atrazine Rate for 90-100% Necrosis Species Chenopodium album, Amaranthus retroflexus 0.03 kg/ha 0.1 kg/ha Poa pratensisa, Digitaria sp., Stellaria media 0.3 kg/ha Echinochloa crus-galli, Avena fatuaa, Bromus inermisa, Agropyron repensa, Alopecuris pratensisa, Sinapis arvenis, Datura stramonium, soybean, Chrysanthemum segetum NOTE: The rate used by farmers in corn varies between 2.2 and 4.4 kg/ha. aMembers of subfamily Poaceae. SOURCE: Data from a commercial screen provided by P. F. Bocion, Dr. Maag Ltd., Dielsdorf, Switzerland (1984). 61 cies. For example, one population of Senecio vulgaris slowly increased in tolerance to sublethal triazine doses (Holliday and Putwain, 1980), yet an- other population "suddenly" evolved plastic resistance to high levels of atrazine (Scott and Putwain, 19811. The biochemical reasons for the increases in tolerance could not be discerned (Gressel et al., 1983b). Similarly, some populations of Echinochloa crus-galli have slowly increased in tolerance (Grignac, 1978), while other Echinochloa biotypes evolved plastic and non- plastid resistance (Gressel et al., 1982b). Tolerance to triazines also varies among species (Table 11. The first species to evolve resistance were those with the greatest inherent susceptibility to atrazine; selection pressure was higher with fewer nonresistant escapees. Higher levels of atrazine are needed to control some species, especially the Poaceous grasses, which possess higher levels of the GST that conjugates atrazine to glutathione. An interesting development for managing resistance is the use of a tridi- phane, an herbicide "extender" that inhibits GST in the Poaceae; thus, much lower levels of atrazine need to be used (Lamoureux and Rusness, 1984~. Lowering the triazine levels should decrease the rate at which dicots evolve triazine resistance (i.e., the slopes in Figure 2 would be less acute) but should not affect the rate that resistance evolves in the Poaceae. If, however, the triazine rates applied are not reduced when tridiphane is used, triazine-re- sistant grasses should evolve more rapidly. Paraquat The mode of tolerance to paraquat has been studied in two systems: Lolium perenne and Conyza bonariensis (= C. Iinefolia). Paraquat, at the levels
62 MECHANISMS OF RESISTANCE TO PESTICIDES used in agriculture, seems to be a specific acceptor of electrons from pho- tosystem I of photosynthesis. The electrons are transferred from paraquat to oxygen, giving rise to highly reactive oxygen radicals that rapidly cause membrane damage due to lipoxidation. Paraquat reacts with photosystem I in the paraquat-resistant weeds, but damage is minimal. The tolerance has been correlated with a 50 percent higher level of superoxide dismutase in Lolium and a three-fold higher level in Conyza (Harvey and Harper, 1982~. Superoxide dismutase foes hydrogen peroxide from the oxygen radicals. As peroxide is also toxic the resistant species must have sufficient levels of other enzymes to further detoxify the peroxide. These enzymes probably are in the plastics where the peroxide is formed and the first membrane lipox- idation occurs. Diclofop-methyl Wheat detoxifies diclofop-methyl and is thus resistant (Shimabukuro et al., 19791. It is not yet known if the resistant biotype of Lolium rigidum has evolved this system or some other mode of resistance. Trifluralin There is no information thus far on the mode of dinitro-aniline resistance that has evolved in Eleusine, nor is there adequate information on modes of selectivity in the species on which they act. CROSS -RESISTANCE The appearance of cross-resistances to totally unrelated groups of insec- ticides is even more disturbing because of the unpredictability of such re- sistances to compounds with totally different modes of action. Fortunately? with herbicides cross-resistance has been more logical and thus more pre- dictable. Triazines The weeds that evolved plastic-level resistance to atrazine and simazine are resistant to all s-triazine herbicides and to some, but not all, asymmetric triazines (triazinones) such as metribuzin. Initially all the plastic-level tria- zine-resistant weeds were thought to be susceptible to diuron, a phenyl-urea herbicide with a similar mode of action as the triazines. Until triazine resis- tance occurred the phenyl-ureas were believed to have a totally identical binding site with the triazines (Pfister and Arntzen, 1979; Arntzen et al., 1982~. Triazine-resistant biotypes, however, were found to have different cross-tolerances to the various phenyl-urea and uracil herbicides (Table 21.
HERBICIDE RESISTANCE IN PINTS TABLE 2 Cross-Resistances to Photosystem II Herbicides of Resistant Biotypes 63 Amaranthus A. Chenopodium Brassica Chlamydomonas Herbicidea retro,flexusb hybridusC albums e campestrise reinhardiif (inhibitory dose for R versus inhibitory dose for S) Atrazine2511,000.0201.0 501.050.0 Metr~buzin1,500260.0100.0 45.01,740.0 Diuron42.61.2 0.714.5 Chloroxuron7941.46.2 Bromacil2,0002.024.0 -106.0 aAtrazine is a symmetric tr~azine; metribuzin is an asymmetric tr~azine (tr~azinone); diuron and chloroxuron are phenyl-ureas; bromacil is a uracil derivative. All act on photosystem II and bind to wild-type thylakoids when they are competitive with each other for overlapping or alloster~c binding siteks). SOURCE: bOettmeier et al. (1982); Poster and Arntzen (1979) (incorrectly designated therein as A. retroflexus); ~Arntzen et al. (1982); eThiel and Boger (1984);fJanatlova and Wildner (1982). A more complete comparison table may be found in Gressel (1985). This, along with the data depicted in Figure lD-lF, suggests that the mu- tations can be at different loci in each of the biotypes, which was further borne out by the molecular biology. So far all triazine-resistant weed biotypes are susceptible to diuron, even if not to other phenyl-urea herbicides, but this need not continue (Table 21. There is probably a spectrum or continuum of binding sites that can be mutated in organisms that gives varying cross- specificities of herbicides affecting photosystem II. Plants seem to have more substrate specificity of GSTs than found in mammalian (liver) systems. Three different GST systems in corn are sub- strate-specific for three herbicide groups: chloro-s-triazines (atrazine and simazine), acetamides (e.g., alachlor), and thiocarbamates (e.g., EPTC) (Mozer et al., 1983~. The GST for atrazine is usually at a high constitutive level, but it probably can be induced to higher levels (Jachetta and Rados- evich, 19811. The GST for alachlor can vary, but can be increased greatly by the protectant flurazole (Mozer et al., 19839. The GST of EPTC can be induced to higher levels, which has been correlated with resistance, by a dichloracetamide-type protectant (Lay and Casida, 19761. No cross-protec- tion has been found in corn systems; induction of protection to one herbicide group does not grant protection to the others. Cross-protection has not been checked in the Poaceous weeds. Paraquat The biochemical nature of tolerance suggests that there should be ways to chemically induce tolerance (Lewinsohn and Gressel, 1984) and that there should be cross-tolerance of paraquat-resistant species with other herbicides
64 MECHANISMS OF RESISTANCE TO PESTICIDES and xenobiotics (Gressel et al., 1982a). There is no perfect cross-tolerance within the bipyridillium group: the paraquat-resistant conyzas are partially tolerant to diquat (M. Parham, I.C.I. Bracknell, United Kingdom, personal communication, 1981; Watanabe et al., 19821. Cross-resistance has positive effects, as seen in the following examples. A Lolium perenne biotype, which evolved sulfur dioxide tolerance downwind from a coal-fired power plant (Horsman et al., 1978), had a modicum of tolerance to paraquat. The paraquat-resistant Conyza bonariensis is tolerant to sulfite (which releases SO2) and to oxyfluorfen (a diphenyl-ether herbicide causing photoenergized membrane lipoxidation). Some ozone-tolerant to- bacco varieties are also paraquat-tolerant. It might be possible, therefore, to design protectants that will guard against herbicides from more than one group as well as protect against environmental pollutants, such as sulfur dioxide and possibly ozone. It is also apparent that if a farmer were to rotate the use of two herbicides such as paraquat and oxyfluorfen, the final effect on enrichment for resistance would be the same as using a single herbicide. Trifluralin The Eleusine biotype, selected for by repeated trifluralin treatments, is resistant to all other dinitroaniline-type herbicides but not to herbicides in six other chemical types (Mudge et al., 19841. Diclofop-methyl The Lolium biotype that is tolerant to diclofop-methyl is not cross-tolerant to oxyfluorfen, as might be expected from its different mode of action. The diclofop-methyl-tolerant material, however, was tolerant to fluazifop-butyl and chlorazifop-propynil, diphenyl-ether herbicides that probably possess similar modes of action as diclofop-methyl (I. Heap and R. Knight, Waite Institute, Adelaide, Australia, personal communication, 1984~. As diphenyl- ether herbicides are being developed with selectivity to different crops, they may be considered for use without herbicide rotation. This Lolium biotype can be used to further study cross-tolerances to ascertain which diphen ether rotations are not really rotations (i.e., whether cross-tolerance occurs). GENETICS AND MOLECULAR BIOLOGY OF RESISTANCE During the few years in which herbicide resistance has appeared and has been studied, we have reports of possibilities of all types of inheritance: nuclear with dominance, recessiveness, monogenic and polygenic, and or- ganelle inherited. There are even cases, studied only in tissue culture, of possible gene duplications (Gressel, in press [al). The discussions that follow
HERBICIDE RESISTANCE IN PLANTS 65 are concerned with only those cases where genetics has been studied in weeds or where the data obtained bear on what is expected to happen in weeds. s-Triazines and Other Photosystem ll Herbicides Triazine tolerance and resistance can be inherited in many ways. The GST that degrades atrazine is inherited as a single dominant gene in corn (Shi- mabukuro et al., 19711. Thus, only one parent of each inbred line used in hybrid seed production needs to bear the trait. The increased levels of tol- erance to triazines that evolved in Senecio vulgaris are inherited polygenically with a low heritability (Holliday and Putwain, 1980~. The plastic-level re- sistance to triazines is maternally inherited, most probably on the plastome (chloroplast genome) (Souza-Machado, 1982~. This has many implications for the appearance and spread of triazine resistance. Once resistance has appeared in weeds it cannot spread by pollen, only by seed. This should considerably slow the spread of resistance. Each plastic has more than one DNA molecule, and each cell has more than one plastic; however, a mutation in a single plastome DNA molecule can create resistance. Most known plastome-mutant plants are a result of using mutagens. From the mode of action, triazine-resistant mutations should be the equivalent of recessive; all thylakoids must not bind triazines, oth- erwise lethal products would be produced. The natural rate of recessive mutations resulting in mutant plants is very low; most plastome-DNA spe- cialists refuse to guess their actual natural frequency. Two factors seem to converge to quicken the natural evolution of popu- lations of triazine-resistant weeds. The first is population genetics. The sec- ond may be a nuclear gene, a plastome mutator, that increases the frequency of plastome mutations. This gene has been found in only four species (Arntzen and Duesing, 19831. Original triazine-resistant plants from which populations evolved probably were in a subpopulation that had a plastome mutator. Therefore, a given mutant is more likely to appear in a population of mutagen- treated plants than plants without mutagen. The selection pressure of triazine treatments enriches for triazine-resistant plants (which are almost always less fit than the wild type) and stabilized resistance in the population. The plastome mutator, which causes other plastome mutations, drains the population and is slowly bred out by actual hybrid selection. It is easier to use unicellular algae with one chloroplast for basic studies on the selection, inheritance, and molecular biology of resistance than to use weeds. For example, resistance to phenyl-urea and uracil-type herbicides is maternally inherited in the green alga Chlamydomonas (Galloway and Mets, 19841; therefore, we can get mutants to other photosystem II-inhibiting her- bicides. This has implications to proposed uses of the other herbicides as mixtures or in sequence with triazines. Population genetics theory states that
66 MECHANISMS OF RESISTANCE TO PESTICIDES whenever we treat with atrazine, we enrich for resistant alleles. If triazine- resistant alleles are found in plants with plastome mutators, enrichment for triazine-resistant individuals also enriches for individuals carrying the plas- tome mutator. The plastome mutator should increase mutation frequency for all plastome mutants including resistance to phenyl-ureas and uracil herbi- cides. Thus, enrichment for triazine resistance should also carry enrichment for resistance to other photosystem II-inhibiting herbicides, which may not be beneficial. For example, if it took 10 years to get triazine-resistant weed populations in a given orchard, a diuron-resistant population would appear in even less time, if diuron is the replacement for atrazine. If atrazine and diuron are used together, as has been proposed for roadside weed control, or used in rotation, each could help enrich for resistance to the other by coenriching for the plastome mutator. If diuron is used for roadside weed control where atrazine resistance has occurred, the rapid appearance of diuron resistance is expected even though there is no cross-resistance between diuron and atrazine. Cross-resistance between high levels of atrazine and diuron may be precluded on molecular grounds (Table 31. Large spans of railroad rights- of-way in the United States and Europe and roadsides in Europe and Israel (Gressel et al., 1983a) are covered with recently evolved triazine-resistant weeds. Adequate long-term recommendations are needed for weed control along these roadways and the new areas where resistant biotypes continually appear. The involvement of a peculiar protein in membranes of the plastics (thy- lakoids) may be responsible for susceptibility or resistance to photosystem II herbicides (Arntzen et al., 1982; Arntzen and Duesing, 1893; Gressel, 1985~. Unlike most membrane proteins this protein (often called "the 32 kD" protein) has a very high turnover rate, which is under positive photo- control, suggesting important plastic functions. This protein also is one of the most highly conserved proteins in biology. It should be very important in plastic functions and mutations in structure should negatively affect photosynthesis and thus growth potential (Radosevich and Holt, 1982; Gres- sel, 19851. Mutations in the plastic-coded gene for this protein confer resis- tance to photosystem II-inhibiting herbicides (Table 31. Transversions at different places in the sequence lead to different resistances and cross-resis- tances. Unfortunately, only two triazine-resistant weeds have been sequenced; they do not differ in amino acid transversion. The Italian Solanum nigrum biotypes (Figure 1E), however, might have different transversions than the French biotype sequenced because of the different dose-response curves. Amaranthus blitoides with its marginal resistance (Figure 1F) and A. retroflexus with its different cross-tolerance to chloroxuron (Table 2) should have different trans- versions from the A. hybridus sequence (Table 31. These algal mutations
67 :~ .~ ca In o 4 - o 4 - C~ ._ Cal Pa .s o .e o ._ o IS: EM ._ Cal o · _ Cal Cal ._ Cal .s ¢ Cal [L) cat ._ m ~ O O au lo: C) Cal 4-§ i~ C .- C ~ ;> 5 ~ ca ca .O 4) Cal ~ ~5 lo: d. 4_ ~ I_ ~ 5 ~ o ~ .C .b -o ~ C) C C C C C ::^ ~ r5~3 .5 ., .~ .S C ~U. C ~C~ ._ . :~ o ~ ~ ~ ~ I.E C~ C~ ~D a, C~ ~3 m O O a a 3 ~ ~5 ~ E ~ E E ~tE 8EE3E c.,£ a 5 ~ c ~ca ~ a g~ o.c,~ ct ~ ~ =, ~ ~ D ~ | e a C e a, 3 ,,, ~ 5 3 ~ - -, G~ ~ O ~ -E ~ 8 ~ ,, E ~x . - ._ co ct c, 0 0 c~ au c~ c~ ~i ._ 3
68 MECHANISMS OF RESISTANCE TO PESTICIDES TABLE 4 Single Nuclear-Gene Herbicide Resistance Herbicidea Site of Action Mode of Resistance Inheritance Reference picloram auxin type nondegradation dominant Chaleff (1980) phenmedipham PSII unknown recessive Radin and Carlson (1978) bentazon PSII unknown recessive Radin and Carlson (1978) chlorsulfuron acetolactate modified enzyme dominant Chaleff and synthase Ray (1984) atrazine PSII degradation dominant Shimabukuro et al. (1971) glyphosate EPSP synthase modified enzyme Comani et al. (1983) aResistant plants (except atrazine and glyphosate) were selected in the laboratory using tissue culture techniques with tobacco. Atrazine was in corn, and glyphosate in bacteria. also have different degrees of fitness loss, which has implications on the biotechnological uses of mutants in this gene for conferring atrazine resistance in crops (Gressel, in press [al). Other Herbicides The genetics of other herbicide-resistant weeds have not been reported to date. Tolerance of Lolium perenne to paraquat and of Senecio to atrazine have polygenic inheritance (Faulkner, 1982~. Faulkner (1982) and Gressel (1985) have reviewed the inheritance of herbicide resistances in crops. LESSONS FROM BIOTECHNOLOGY There are compelling commercial reasons for biotechnologically conferring cost-effective herbicide resistance to crop species (Gressel, in press [a]~. The ease with which nuclear monogenic mutants resistant to many herbicides have been obtained in the laboratory (Table 4) is cause to pause and consider the implications for weed control practices. Resistances can be dominant or recessive, with the genetics clearly related to mode of action. When resistance is due to degradation of the herbicide or to overcoming a herbicide-caused metabolic blockage of a vital pathway, resistance is dominant (Table 4~. Phenmedipham is thought to act on photosystem II similarly to atrazine and diuron, and it competes with them (Tischer and Strottmann, 19771. Presumably resistance is due to a nonbinding of the herbicide, since the mutation is recessive (Radin and Carlson, 19781. If the mutation was dom
HERBICIDE RESISTANCE IN PC4NTS 69 inant, part of the thylakoids in a heterozygote would suicidally bind the herbicide. Bentazon is also a photosystem II, electron-transport inhibitor; thus, the reasons for resistance are similar to those for phenmedipham. It has been easy to obtain bacterial mutants with a modified enol-pyruvate- shikimate-phosphate-synthase (EPSP synthase), the enzyme thought to be the sole target of glyphosate. Should one then expect to obtain glyphosate resistance in the field as its use increases? It depends: glyphosate is an "ephemeral" herbicide; it affects only those plants on which it is sprayed. This lack of persistence should give the herbicide the selection pressure needed for long field-life. With paraquat, a similar ephemeral herbicide, lack of soil persistence can be compensated for by the persistence of the farmers. Most paraquat resistance happened when the farmers sprayed about 10 times a year. If farmers do the same with glyphosate, they can expect resistant weeds. Chlorsulfuron and other sulfonyl-urea herbicide-resistant mutants are easily obtained and regenerated to resistant plants (Chaleff and Ray, 19841. Re- sistance in tobacco is from a single dominant gene that modifies aceto-lactate synthase, the sole enzyme target of this group. A new imidazole-type her- bicide affects the same enzyme site, but no data are available on cross- resistance. The specific sites affected on the enzyme may be different, as with atrazine and diuron, although neither are reversed by pyruvate, one of the substrates. Chlorsulfuron, at the rates used for weed control in wheat, has long soil persistence, rivalling that of the triazines. The models (Figure 2) predict that if sulfonyl-ureas are used without rotation or are not mixed with other herbicides, resistance will rapidly appear. The initial gene fre- quency of sulfonyl-urea resistant mutants in weed populations should be many orders of magnitude higher than triazine-resistant mutants; therefore, resistance should appear in a few years of widespread monoculture. There also may be enrichment for soil organisms that degrade chlorsulfuron, as in the problem soils. Once we know whether resistance is dominant or recessive we can estimate the initial frequency in the population and plug this information into Figure 2. The frequency for dominant mutations in diploid species should be 10-5 to 10-7. The frequency for recessive mutations should be 10-~° to 1o-~4 according to theory, but classical theory may be wrong because of somatic recombinations, and the frequencies may be 10-7 to 10-9 (Williams, 19761. Even these orders of magnitude differences between dominant and recessive will affect the time until resistance appears (Figure 21. CONCLUSION If good, cost-effective herbicides are judiciously used (only where and when needed, and in rotations and in mixtures), costly resistances can be considerably delayed. To make educated recommendations one must know
70 MECHANISMS OF RESISTANCE TO PESTICIDES the modes of action, cross-actions and cross-resistances, genetics, and mo- lecular biology of the weeds and herbicides. The basic sciences have helped us understand the nature of excesses in agronomic practices and resistance and have given us information on how to slow down the process. We must learn from this short history. Since each herbicide and resistance may have very different properties, we must have this basic information, otherwise knowledgeable extrapolations are hard to make. "Spray and pray" must become a concept of the past if we wish to keep the most effective herbicides in our arsenal to fight the continual battle against loss of yields caused by weeds. ACKNOWLEDGMENTS I thank the many scientists around the world who have supplied yet un- published data for use in this review. The author's own work on oxidant resistance is supported in part by the Israel Academy of Sciences and Hu- manities program in basic research. The author is the Gilbert de Botton professor of plant science. REFERENCES Abel, A. L. 1954. The rotation of weedkilling. P. 249 in Proc. Br. Weed Cont. Conf., Cliftonville, Margate, England, 1953. Ahrens, W. H., C. J. Arntzen, and E. W. Stoller. 1981. Chlorophyll fluorescence assay for the determination of triazine resistance. Weed Sci. 29:316. All, A., and V. Souza-Machado. 1981. Rapid detection of triazine resistant weeds using chlorophyll fluorescence. Weed Res. 21:191. Ammon, H. U., and E. Irla. 1984. Bekampfung resistenter Unkrauter in Mais-Erfahrungen mit mechanischen und chemischen Verfahren. Die Grune 112:12. Arntzen, C. J., and J. H. Duesing. 1983. Chloroplast encoded herbicide resistance. Pp. 273-294 in Advances in Gene Technology: Molecular Genetics of Plants and Animals, K. Downey, R. W. Voellmy, F. Ahmad, and J. Schultz, eds. New York: Academic Press. Arntzen, C. J., K. Pf~ster, and K. E. Steinback. 1982. The mechanism of chloroplast triazine resistance: Alterations in the site of herbicide action. Pp. 185-213 in Herbicide Resistance in Plants, H. LeBaron and J. Gressel, eds. New York: John Wiley and Sons. Chaleff, R. S. 1980. Further characterization of picloram tolerant mutants of Nicotiana tabacum. Theor. Appl. Genet. 58:91. Chaleff, R. S., and T. B. Ray. 1984. Herbicide resistant mutants from tobacco cell cultures. Science 223:1148. Comai, L., L. C. Sen, and D. M. Stalker. 1983. An altered aroA gene product confers resistance to the herbicide glyphosate. Science 221:370. Delp, C. J. 1979. Resistance to plant disease control agents: How to cope with it. Pp. 253-261 in Proc. Symp. 9th Int. Cong. Plant Prot., Vol. 1, T. Kommedahl, ed. Minneapolis, Minn.: Burgess. Dexter, A. G., J. D. Nalewaja, D. D. Rasmusson, and J. Buchli. 1981. Survey of wild oats and other weeds in North Dakota: 1978 and 1979. N. D. Res. Rep. No. 79. Fargo: North Dakota State Extension Service. Ducruet, J. M., and J. Gasquez. 1978. Observation of whole leaf fluorescence and demonstration of chloroplastic resistance to atrazine in Chenopodium album L. and Poa annua L. Chemosphere 8:695.
HERBICIDE RESISTANCE IN PLANTS 71 Erickson, J. M., M. Rahire, P. Bennoun, P. Delepelaire, B. Diner, and J. D. Rochaix. 1984. Herbicide resistance in Chlamydomonas reinhardtii results from a mutation in the chloroplast gene for the 32 kD protein of photosystem II. Proc. Natl. Acad. Sci. 81:3617. Faulkner, J. S. 1982. Breeding herbicide-tolerant crop cultivars by conventional methods. Pp. 235- 256 in Herbicide Resistance in Plants, H. M. LeBaron and J. Gressel, eds. New York: John Wiley and Sons. Galloway, R. E., and L. J. Mets. 1984. Atrazine, bromacil and diuron resistance in Chlamydomonas. Plant Physiol. 74:469. Georghiou, G. P., and C. E. Taylor. 1977. Operational influences in the evolution of insecticide resistance. J. Econ. Entomol. 70:653-658. Goldie, J. H., and A. J. Coldman. 1979. A mathematical model for relating drug sensitivity of tumors to their spontaneous mutation rate. Cancer Treat. Rep. 63:1727. Goloubinoff, P., M. Edelman, and R. B. Hallick. 1984. Chloroplast-coded atrazine resistance in Solanum nigrum: psbA loci from susceptible and resistant biotypes are isogenic except for a single codon change. Nucleic Acids Res. 12:9489-9496. Gressel, J. 1985. Herbicide tolerance and resistance: Alteration of site of activity. Pp. 159-189 in Weed Physiology, Vol. 2, S. O. Duke, ed. Boca Raton, Fla.: CRC Press. Gressel, J. In press(a). Biotechnologically conferring herbicide resistance in crops: The present realities. ln Molecular Form and Function of the Plant Genome, L. Van Vloten-Doting, G. S. P. Groot, and T. C. Hall, eds. New York: Plenum. Gressel, J. In press(b). Strategies for prevention of herbicide resistance in weeds. In Rational Pesticide Use, K. J. Brent and R. Atkin, eds. Cambridge: Cambridge University Press. Gressel, J., and G. Ben-Sinai. In press. Low intra-specific competitive fitness in a triazine resistant, nearly nuclear-isogenic line of Brass~ca napus. Plant Sci. Lett. 38. Gressel, J., and L. A. Segel. 1978. The paucity of plants evolving genetic resistance to herbicides: Possible reasons and implications. J. Theor. Biol. 75:349-371. Gressel, J., and L. A. Segel. 1982. Interrelating factors controlling the rate of appearance of resistance: The outlook for the future. Pp. 325-348 in Herbicide Resistance in Plants, H. M. LeBaron and J. Gressel, eds. New York: John Wiley and Sons. Gressel, J., G. Ezra, and S. M. Jain. 1982a. Genetic and chemical manipulation of crops to confer tolerance to chemicals. Pp. 79-91 in Chemical Manipulation of Crop Growth and Development, J. S. McLaren, ed. London: Butterworth. Gressel, J., H. U. Ammon, H. Fogelfors, J. Gasquez, Q. O. N. Kay, and H. Kees. 1982b. Discovery and distribution of herbicide-resistant weeds outside North America. Pp. 32-55 in Herbicide Resistance in Plants, H. M. LeBaron and J. Gressel, eds. New York: John Wiley and Sons. Gressel, J., Y. Regev, S. Malkin, and Y. Kleifeld. 1983a. Characterization of an s-triazine resistant biotype of Brachypodium distachyon. Weed Sci. 31:450. Gressel, J., R. H. Shimabukuro, and M. E. Duysen. 1983b. N-dealkylation of atrazine and simazine in Senecio vulgaris biotypes, a major degradation pathway. Pestic. Biochem. Physiol. 19:361. Grignac, P. 1978. The evolution of resistance to herbicides in weedy species. Agro-Ecosystems 4:377. Hartmann, F. 1979. The atrazine resistance of Amaranthus retroflexus L. and the expansion of resistant biotype in Hungary. Novenyvedelem 15:491. Harvey, B. M. R., and D. B. Harper. 1982. Tolerance to bipyridylium herbicides. Pp. 215-234 in Herbicide Resistance in Plants, H. M. LeBaron and J. Gressel, eds. New York: John Wiley and Sons. Heap, J., and R. Knight. 1982. A population of ryegrass tolerant to the herbicide diclofop-methyl. J. Aust. Inst. Agri. Sci. 48:156. Hirschberg, J., and L. McIntosh. 1983. Molecular basis of herbicide resistance in Amaranthus hybridus. Science 222:1346. Hirschberg, J., A. Bleecker, D. J. Kyle, L. McIntosh, and C. J. Arntzen. 1984. The molecular basis of triazine-herbicide resistance in higher plant chloroplasts. Z. Naturforsch. 39c:412.
72 MECHANISMS OF RESISTANCE TO PESTICIDES Holliday, R. J., and P. D. Putwain. 1980. Evolution of herbicide resistance in Senecio vulgaris: Variation in susceptibility to simazine between and within populations. J. Appl. Ecol. 17:779. Holm, L. G., D. L. Plucknett, J. V. Pancho, and J. P. Herberger. 1977. World's Worst Weeds. Honolulu: University Press of Hawaii. Horsman, D. A., T. M. Roberts, and A. D. Bradshaw. 1978. Evolution of sulphur dioxide tolerance in perennial ryegrass. Nature (London) 276:493-494. Jachetta, J. J., and S. R. Radosevich. 1981. Enhanced degradation of atrazine by corn. Weed Sci. 29:37. Janatkova, H., and G. F. Wildner. 1982. Isolation and characterization of metribuzin-resistant Chlamydomonas reinhardtii cell. Biochim. Biophys. Acta 682-227. Kaufman, D. D., Y. Katan, D. F. Edwards, and E. G. Jordan. In press. Microbial adaptation and metabolism of pesticides. In Agricultural Chemicals of the Future. BARC symposium, No. 8., J. L. Hilton, ed. Totowa, N.J.: Rowman and Allanheld. Lamoureux, G. L., and D. L. Rusness. 1984. Glutathione-S-transferase as the basis of Dowco 356 (tridiphane) synergism of atrazine. Am. Chem. Soc. Meet. Abstr. 181. Lay, M. M., and J. E. Casida. 1976. Dichloracetamide antidotes enhance thiocarbamate sulfoxide detoxification by elevating corn root glutathione content and glutathione-S-transferase activity. Pestic. Biochem. Physiol. 6:442. LeBaron, H. M., and J. Gressel, eds. 1982. Herbicide Resistance in Plants. New York: John Wiley and Sons. Lewinsohn, E., and J. Gressel. 1984. Benzyl viologen mediated counteraction of diquat and paraquat phytotoxicities. Plant Physiol. 76:125. Mozer, T. J., D. C. Tiemeier, and E. G. Jaworski. 1983. Purification and characterization of corn glutathione-S-transferase. Biochemistry 22:1068. Mudge, L. C., B. J. Gossett, and T. R. Murphy. 1984. Resistance of goosegrass (Eleusine indica) to dinitro-aniline herbicides. Weed Sci. 32:591. Oettmeier, W., K. Masson, C. Fedtke, J. Konze, and R. R. Schmidt. 1982. Effect of different photosystem II inhibitors on chloroplasts isolated from species either susceptible or resistant toward s-triazine herbicides. Pestic. Biochem. Physiol. 18:357. Pfister, K., and C. J. Arntzen. 1979. The mode of action of photosystem II-specific inhibitors in herbicide resistant weed biotypes. Z. Naturforsch. 34c:996. Radin, D. N., and P. S. Carlson. 1978. Herbicide resistant tobacco mutants selected in situ recovered via regeneration from cell culture. Genet. Res. 32:85. Radosevich, S. R., and J. S. Holt. 1982. Physiological responses and fitness of susceptible and resistant weed biotypes to triazine herbicides. Pp. 163-184 in Herbicide Resistance in Plants, H. M. LeBaron and J. Gressel, eds. New York: John Wiley and Sons. Ryan, G. F. 1970. Resistance of common groundsel to simazine and atrazine. Weed Sci. 18:614. Scott, K. R., and P. D. Putwain. 1981. Maternal inheritance of simazine resistance in a population of Senecio vulgaris. Weed Res. 21: 137. Shimabukuro, R. J., D. S. Frear, R. Swanson, and W. C. Walsh. 1971. Glutathione conjugation: An enzymatic basis for atrazine resistance in corn. Plant Physiol. 47:10. Shimabukuro, R. J., W. C. Walsh, and R. A. Hoerauf. 1979. Metabolism and selectivity of diclofop- methyl in wild oat and wheat. J. Agric. Food Chem. 27:615. Solymosi, P. 1981. AzAmaranthus retroflexus triazine resistenciujanak oroklodese. Novenytermeles 30:57. Souza-Machado, V. 1982. Inheritance and breeding potential of triazine tolerance and resistance in plants. Pp. 257-274 in Herbicide Resistance in Plants, H. M. LeBaron and J. Gressel, eds. New York: John Wiley and Sons. Thiel, A., and P. Boger. 1984. Comparative herbicide binding by photosynthetic membranes from resistant mutants. Pestic. Biochem. Physiol. 22:232. Tischer, W., and H. Strotmann. 1977. Relationship between inhibitor binding by chloroplasts and inhibition of photosynthetic electron transport. Biochim. Biophys. Acta 460:113.
HERBICIDE RESISTANCE IN PLANTS 73 Watanabe, Y., T. Honma, K. Ito, and M. Miyahara. 1982. Paraquat resistance in Erigeron phi- ladelphicus. Weed Res. (Japan) 7:49. Williams, K. L. 1976. Mutation frequency at a recessive locus in haploid and diploid strains of a slime mould. Nature (London) 260:785. Zanin, G., B. Vecchio, and J. Gasquez. 1981. Indagini sperimentali su popolazioni di dicotiledoni resistenti alit atrazine. Riv. Agron. 5:196. Zurawski, G., H. J. Bohnert, P. R. Whitfeld, and W. Bottomley. 1982. Nucleotide sequence of the gene for the M,32,000 thylakoid membrane protein from Spinacia oleracea and Nicotiana debueyi predicts a totally conserved primary translation product of M,38,950. Proc. Natl. Acad. Sci. 79:7699.