A mutation in the herbicide target site acetohydroxyacid synthase produces morphological and structural alterations and reduces fitness in Amaranthus powellii



  • • We investigated the effect of a herbicide resistance-conferring mutation on fitness in Amaranthus powellii.
  • • Morphological and histological observations were made. Growth and leaf appearance were recorded for six resistant and six susceptible populations. The competitiveness of a susceptible population was compared with that of a resistant population using a replacement series experiment.
  • • Leaves of the resistant plants were distorted and much smaller than those of susceptible plants. Additionally, they exhibited an abnormal morphological and structural pattern consisting of a mosaic of heterogeneous areas in the same leaf blade. The roots and stems had similar structures in susceptible and resistant plants, but the former were up to four times more developed. The resistant plants were slower to develop and produced 67% less biomass and 58% lower leaf area than susceptible plants. Under competitive conditions, one susceptible population outperformed one resistant population by 7–15 times.
  • • The Trp574Leu acetohydroxyacid synthase (AHAS) mutation appears to have considerable pleiotropic effects on the early growth and development of the plants which, in competitive conditions, greatly reduce fitness.


Acetohydroxyacid synthase (AHAS, EC, also called acetolactate synthase, ALS) is a plastidic enzyme that is found in plants, fungi and bacteria (Duggleby & Pang, 2000). It is a key enzyme in the synthesis of the branched-chain amino acids valine, leucine and isoleucine. AHAS catalyses the decarboxylation of pyruvate and its condensation with another pyruvate, or with an α-ketobutyrate. The first reaction produces acetolactate, the precursor of valine and leucine, while the second reaction produces acetohydroxybutyrate leading to isoleucine synthesis (Chipman et al., 1998). AHAS is the target site of many classes of commercial herbicides including the sulfonylureas (SUs), imidazolinones (IMIs), pyrimidinyloxybenzoates (POBs) and triazolopyrimidines (TPs). Inhibition of AHAS in plants leads to death from branched-chain amino acid starvation (Shaner & Singh, 1997). AHAS-inhibiting herbicides are widely used because of their high weed control efficacy, the large range of crops to which they can be applied, their low use rates and their low mammalian toxicity (Saari et al., 1994).

A large number of weed species populations are resistant to AHAS inhibitors because of selection in the field. Laboratory selection has also produced a number of resistant mutants in species such as Arabidopsis thaliana, maize (Zea mays) and tobacco (Nicotiana tabacum) (Tranel & Wright, 2002). In the great majority of cases, point mutations in the AHAS gene confer resistance by encoding various amino acid substitutions. In weeds, six mutation sites have been documented encoding amino acid substitutions at Ala122, Pro197, Ala205, Asp376, Trp574, and Ser653 (Tranel & Wright, 2002; Corbett, 2004). These substitutions prevent inhibitor binding while yielding an enzyme that is still functional because the inhibition site is separate from the enzyme catalytic site (Pang et al., 2003). Depending on the substitution, the level of protection will vary, as will the number of inhibitor classes affected (Tranel & Wright, 2002). For example, the Ala122Val substitution confers resistance to the IMIs and POBs only (Bernasconi et al., 1995; Siehl et al., 1996) while the Trp574Leu substitution endows plants with very high resistance to all inhibitor classes (Bernasconi et al., 1995; Foes et al., 1998; Sibony & Rubin, 2003a).

Although a herbicide-resistant plant is expected to have reduced fitness (Bergelson & Purrington, 1996), most studies have failed to detect penalties on fitness components associated with AHAS inhibitor resistance (Holt & Thill, 1994; Tranel & Wright, 2002). For example, differences between resistant and susceptible populations of Lactuca serriola in growth rate and of Kochia scoparia in biomass accumulation did not affect their competitiveness (Alcocer-Ruthling et al., 1992; Thompson et al., 1994; Christoffoleti et al., 1997). Similar results were reported with AHAS inhibitor-resistant Amaranthus hybridus and Amaranthus retroflexus (Poston et al., 2002; Sibony & Rubin, 2003b). As a result, it is generally accepted that resistance-conferring mutations in AHAS may have subtle effects on plant growth and development, but they do not consistently reduce plant fitness (Holt & Thill, 1994).

In contrast to AHAS-inhibiting herbicides, resistance to the triazine herbicides (photosystem II inhibitors) has been consistently associated with a significant fitness penalty (Holt & Thill, 1994; Gressel, 2002). The large impact of the triazine resistance trait also translates into severe reductions in growth and biomass accumulation and is generally observable throughout the vegetative phase (Holt & Thill, 1994). Failure to observe such striking fitness costs in AHAS inhibitor-resistant plants influenced the conclusions of researchers. Indeed, even when experimental results show differences in traits that may have an impact on fitness, they are viewed with skepticism (Kuk et al., 2004).

There are nevertheless some reports that resistance to AHAS inhibitors negatively affects fitness components in the absence of herbicides. When grown in the field under noncompetitive conditions, Arabidopsis plants transformed with the chlorsulfuron resistance allele Csr-1 of AHAS (Pro197Ser substitution; Haughn et al., 1988) produced 26–37% fewer seeds than herbicide-susceptible plants, despite no differences in total biomass accumulation (Bergelson et al., 1996; Purrington & Bergelson, 1997; Roux et al., 2004). Because the genetic background of the lines compared was controlled, these studies provided a strong indication that the Pro197Ser mutation has an influence on fitness despite having no effect on vegetative growth.

Lack of genetic background control has been suggested as one of the main reasons why many studies have failed to document an effect on fitness of AHAS inhibitor-resistance mutations (Bergelson & Purrington, 1996; Cousens et al., 1997; Gressel, 2002). It is also possible that fitness differences may be prevalent at certain stages of the life cycle of the plant and, if that stage is not measured in an experiment, then it is possible that differences may not be detected (Radosevich et al., 1997; Gressel, 2002). Many experiments have used transplants of resistant and susceptible lines and this likely prevented the detection of differences that would be expressed earlier in the development of the plants (Gressel, 2002).

One way to reduce the impact of various genetic backgrounds is to compare a large number of resistant and susceptible lines (Cousens et al., 1997). Simply comparing one resistant and one susceptible population can be problematic as it is always possible that the susceptible line has a naturally weak genotype, whereas the resistant comes with a ‘stronger’ set of genes. Therefore, a negative effect of the resistance gene might be masked by other ‘stronger’ genes. Hence, comparisons amongst several resistant and susceptible lines would provide a better estimate of the resistant allele effect (Cousens et al., 1997).

Amaranthus powellii is an annual weed that is very abundant in Eastern Canada and parts of the USA (Costea et al., 2004). Because it reduces yields in many crops such as maize and soybeans, it has been the target of many AHAS-inhibiting herbicides including thifensulfuron (SU) and imazethapyr (IMI) (Costea et al., 2004). As a result, many populations of A. powellii from Ontario have developed resistance to the AHAS inhibitors (Ferguson et al., 2001). We have reported that various mutations in the AHAS gene confer resistance in these populations (Diebold et al., 2003; McNaughton et al., 2005). During the course of previous experiments, we noticed that some resistant A. powellii populations exhibited distorted growth, even in the absence of herbicides. All these populations had the Trp574Leu amino acid substitution in the AHAS enzyme (Corbett, 2004; McNaughton et al., 2005). Our hypothesis is that the growth distortion of resistant A. powellii is a consequence of the herbicide resistance-conferring mutation and that it negatively affects the biomass accumulation, phenology and competitive ability of the plants, ultimately reducing their fitness. Our specific objectives were to describe more precisely the morphological and anatomical basis of the growth distortion. We also wished to characterize the growth and development of resistant plants and finally determine the impact of competition on growth and seed production of resistant vs susceptible plants.

Materials and Methods

Plant material

Six resistant (R) and six susceptible (S) Amaranthus powellii S. Watson populations originating from Southern Ontario, Canada, were chosen for the experiments (Table 1). The S populations are controlled by the AHAS inhibitors IMI and SU, whereas the R populations survive application of high rates of these herbicides (Ferguson et al., 2001; F. J. Tardif & P. J. Smith, unpublished results). The R populations were chosen from a collection of 13 different populations with the Trp573Leu substitution (Corbett, 2004; McNaughton et al., 2005). Six populations were selected from these 13 that had adequate numbers of viable seeds and that originated as far from each other as possible. Collection in the field was typically from five to 20 mature plants, which were air-dried at room temperature before seed separation and cleaning. The seeds were stored at 4°C before they were used in the experiments. Resistance was confirmed by treating three- to four-leaf stage seedlings with IMI and SU using a precision cabinet sprayer. Herbicide doses were selected that totally killed the susceptible population but caused no injury to resistant plants, as determined by full dose–response curves. The AHAS Trp574Leu substitution is present in all the R populations and its presence was confirmed by DNA sequencing of at least three individuals or by subjecting them to mutation-specific endonuclease tests (Corbett, 2004).

Table 1. Origin of the Amaranthus powellii populations used in the study
PopulationStatusOrigin1Coordinates (latitude, longitude)
  • 1

    All locations refer to the closest municipality or township in Ontario, Canada where the populations were sampled.

  • R, resistant; S, susceptible.

12SHullett Township43°37′ N, 81°32′ W
41SHarrow42°2′ N, 82°55′ W
53SHarrow42°2′ N, 82°55′ W
83SRidgetown42°26′ N, 81°53′ W
84STalbotville42°47′ N, 81°11′ W
86SHarrow42°2′ N, 82°55′ W
29RBrigden42°48′ N, 82°17′ W
50RGlencoe42°43′ N, 81°43′ W
60RBecher42°35′ N, 82°23′ W
64RRidgetown42°26′ N, 81°53′ W
70RKerwood42°57′ N, 81°37′ W
73RCorruna42°59′ N, 82°24′ W


Seeds were incubated in Petri dishes on 0.6%[weight/volume (w/v)] agar under a temperature regime of 40, 15 and 40°C for 10, 2 and 12 h, respectively, in the dark. This was done to ensure even germination across all the populations as some of them had a higher percentage of dormant seeds. Petri dishes with germinated seeds were transferred to a growth room for 5 d with a 16-h photoperiod at 25°C and an 8-h scotoperiod at 20°C. Light was supplied by a mixture of fluorescent tubes and incandescent bulbs for a photosynthetic photon fluence rate (PPFR) of approx. 400 µmol m−2 s−1 over the waveband 400–700 nm. Seedlings were then transplanted to plastic pots (15 cm diameter) containing a growth medium based on Canadian sphagnum peat moss (55–65%), complemented with perlite, dolomite, calcite, vermiculite and wetting agent (Mix #4 Aggregate Plus; Sun Gro Horticulture Inc., Vancouver, BC, Canada). There was one seedling per pot and five replications of each population. The seedlings were placed in a controlled-environment growth cabinet programmed for a 16-h photoperiod at 25°C and an 8-h scotoperiod at 15°C. The PPFR at the top of the pots was approx. 500 µmol m−2 s−1 provided by incandescent bulbs and fluorescent tubes. Pots were watered with a nutrient solution containing nitrogen (N), phosphorus (P), potassium (K), calcium (Ca), magnesium (Mg) and chelated micronutrients every 2 or 3 d. The experiment was arranged as a completely randomized design with five replications.

Histological observations were made on all plants 28 d after transplanting. Hand and microtome transversal sections were taken through the roots, stems and leaves. The principal roots were sectioned immediately below the collar. The stems were sectioned through the internodes below the seventh node. Leaves from the ninth and tenth nodes were used for the study of the lower epidermis and the structure of the mesophyll and vascular tissue. Leaf blade strips and epidermal peels were sampled from the middle regions situated between the third and fourth secondary veins. Two distinct zones of the lamina of R populations (Fig. 1b) were sampled separately. The epidermal peels were cleared with chloral-hydrate and stained with toluidine blue for 4 s. Samples of stem, root and leaf blade were fixed in 2.5% glutaraldehyde for 3 h, dehydrated in a series of ethanol baths and embedded in LR white resin (SPI Supplies/Canada, Toronto, ON, Canada). Transverse sections of 5–10 µm were stained for 5 s with toluidine blue. Observations and measurements were taken with a standard brightfield microscope. Measurements were made on 20 sections per vegetative organ or morphological type of leaf, on each of the 60 plants.

Figure 1.

Leaves of susceptible (S; a) and resistant (R; b) Amaranthus powellii plants. The sampled leaves were fourth leaves and were collected 28 d after emergence when the plants were at the 16–18-leaf stage. Bars, 1 cm. W, wrinkled leaf zones; Y, yellow zones.

Growth and development

Seeds of the same populations were germinated, planted and grown as above. Once in the growth cabinet, the number of leaves on the main stem was counted at 3-d intervals until 28 d after transplanting when they were sampled for biomass and leaf area. Plants were separated into leaves (leaf blades and petioles) and stems. After measuring the total leaf area with an LI-3000 area meter (LI-COR Inc., Lincoln, NB, USA), all plant parts were combined and dried in a forced-air oven at 80°C for 2 d and weighed. There were five replications of each population and the experiment was conducted three times.

Data were analysed using proc glm (SAS Institute Inc., Cary, NC, USA) with replication-in-time as a random variable. All data satisfied the assumptions of independent and normally distributed variance and data transformation was not required. S and R populations were compared using a contrast. Leaf appearance over time was determined by regression analysis using raw data.

Determination of competitive ability

The herbicide-susceptible population 41 (S41) was randomly selected among all susceptible populations for this experiment. Because herbicide-resistant population 29 (R29) was more productive than the other resistant populations, it was selected so as to not obtain an overestimated measure of fitness. Both of these populations had low-level seed dormancy (< 15%) and had the same germination and emergence rates (F. J. Tardif, unpublished results). Seeds were placed 5 mm deep in the same growth medium as above, in 30-cm-diameter, 4-l pots. In order to ensure the presence of one plant at each position, three seeds were buried in each position. Seeds were spaced 1.3 cm apart in a grid with 24 positions in the center of the pot according to a 2-4-6-6-4-2 pattern. Populations were grown either as pure stands or in mixtures at ratios of 24 : 0, 18 : 6, 12 : 12, 6 : 18, or 0 : 24 (S41 to R29). They were placed in a glasshouse with the temperature set at 25°C during the day and 18°C at night. Natural daylight was supplemented by high-pressure sodium lamps for a maximum PPFR of 1200 µmol m−2 s−1 for 16 h. Three days after planting, seedlings were thinned to one per position by cutting the hypocotyl of extra plants at soil level with fine scissors. Pots were watered as needed and fertilized once a week with 50 ml of a 5% solution of 20-20-20 (N-P-K) fertilizer. The above-ground biomass of R29 and S41 plants was determined 63 d after planting. Resistant and susceptible plants were sampled separately by cutting stems at soil level, placing them in paper bags, and drying them at 70°C for 2 d before weighing biomass. The relative crowding coefficient (RCC) (Novak et al., 1993) was calculated according to the formula:

image(Eqn 1)

(DB, dry biomass; s, S41 plants; r, R29 plants; superscript ratios, proportion of S41 to R29; N, number of mixed plantings; here N = 3.)

Therefore, an RCC value of 1 indicates that the two populations have equal competitiveness, while values greater than 1 show superior competitiveness of S41 over R29. RCC values were compared to 1 using a χ2 test. In addition to total above-ground biomass, seeds of plants grown at the 24 : 0, 12 : 12, and 0 : 24 ratios were separated from the inflorescence by threshing. They were cleaned through metal sieves and debris separated using a blown-air seed cleaner before weight measurement. There were four replications of each treatment and the experiment was later repeated. Run 1 and run 2 were conducted in 2003.


The general appearance of the R plants was altered compared with that of the S plants. The differences were obvious and noticeable shortly after emergence. While the cotyledons of R and S plants were similar and emerged at the same time, greater differences were observed in the young leaves. The leaves of the S plants were fully extended and uniformly colored, whereas the leaves of the R plants were distorted and smaller, with discolored and intensely green-colored areas on the same leaf blade (Fig. 1). The distorted appearance of the leaves of the R plants was obvious from the first leaf up to the tenth. After the 10-leaf stage, the symptoms slowly diminished in intensity although they were still obvious on many plants (data not shown).

Leaf morphology

The leaf shape of the S plants tended to be fairly homogeneous and ranged from ovate to elliptic (Fig. 1a). This was relatively consistent from leaf 2 to leaf 10. The fourth leaf had maximum and minimum lengths of 7.5 and 15 cm and maximum and minimum widths of 4.5 and 7 cm. In contrast, the shape of the leaves of R plants was more variable as they were narrow-ovate, narrow-elliptic, lanceolate, obovate or spathulate (Fig. 1b). There tended to be more variation in width (0.7–2 cm) than in length (5–9 cm). Furthermore, leaves of R plants usually comprised two different morphological regions: (1) a yellow flat zone in the lower half of the leaf blade, and (2) a dark-green intensely wrinkled or convoluted area, in the middle to the upper half of the leaf blade (Fig. 1b). The symptomology present on the fourth leaf (Fig. 1) was similar to that observed on leaves 2 to 10. As mentioned above, the symptoms tended to gradually disappear after the 10-leaf stage.

Anatomy of leaves

Susceptible populations  The C4 structure of the leaves was similar in all the S populations. The tertiary bundles were surrounded by a sheath (Kranz anatomy) formed by a layer of tightly packed, thick-walled cells containing many large chloroplasts (Fig. 2). The upper and lower epidermises were more or less parallel and the mesophyll was dorsiventral, with one palisade parenchyma layer and several layers of spongy parenchyma cells. The vascular bundles were collateral and those of the main and secondary veins consisted of both primary and secondary tissues (Fig. 2). The leaf blades were 180–201 µm thick in the S populations (Table 2, Fig. 2). Thickness of the palisade parenchyma ranged between 40 and 50 µm, while that of the spongy parenchyma ranged between 25 and 33 µm (Table 2). Tertiary veins were 83–93 µm thick and the thickness of bundle sheaths varied between 34 and 38 µm (Table 2). For all these variables, population S83 always had the smallest values, while population S84 consistently had the highest values; the differences were not, however, significant except for the thickness of the spongy parenchyma and of the bundle sheath cells. The stomatal apparatus was anomocitic and guard cells were in the same plane as other epidermal cells (Fig. 2). The density of stomata in the lower epidermis was between 72 and 76 stomata mm−2 and did not vary among the S populations (Table 2, Fig. 3).

Figure 2.

Cross-sections of representative leaf laminas of (a) susceptible (S; population 41) and (b, c) resistant (R; population 29) Amaranthus powellii plants, with (b) showing the wrinkled zone and (c) the yellow zone of a resistant plant. The images of leaves on the left indicate the sampled area. In each population, the tenth leaf was sampled 28 d after emergence. Bars on the left, 1.5 cm; bars on the right, 35 µm. Arrowheads, guard cells; p, palisade parenchyma; bs, bundle sheath; sp, spongy parenchyma; ue, upper epidermis; le, lower epidermis.

Table 2. Anatomical measurements of leaves of susceptible (S) and resistant (R) Amaranthus powellii populations
PopulationThickness (µm)Stomatal density1 (number mm−2)
Leaf bladePalisadeSpongy parenchymaTertiary veinsBundle sheath cells
  • Values are given as mean ± standard error.

  • For all populations, the ninth or tenth leaf was sampled when the plants were at the 16–18-leaf stage of growth, 28 d after emergence.

  • 1

    Stomatal density was measured on the lower epidermis of leaves.

Susceptible populations
S12198 ± 1846 ± 3.229 ± 1.9 92 ± 5.6  37 ± 1.9 73 ± 5
S41198 ± 1846 ± 3.029 ± 1.8 92 ± 5.6  37 ± 1.3 72 ± 6
S53198 ± 1846 ± 3.629 ± 1.9 91 ± 6.8  36 ± 2.1 72 ± 4
S83180 ± 1740 ± 4.025 ± 1.6 83 ± 5  34 ± 1.5 75 ± 8
S84201 ± 1950 ± 4.933 ± 2.1 93 ± 5.5  38 ± 1.9 76 ± 6
S86199 ± 1848 ± 4.130 ± 1.9 91 ± 4.84  38 ± 1.7 76 ± 5
Resistant populations; wrinkled zones
R29221 ± 2059 ± 5.141 ± 3.6101 ± 8.1  38 ± 1.6125 ± 11
R50206 ± 2052 ± 4.738 ± 3.7 96 ± 7.99  37 ± 1.6118 ± 8
R60205 ± 2052 ± 4.537 ± 3.1 95 ± 8.1  38 ± 1.6118 ± 8
R64200 ± 1951 ± 4.833 ± 3.1 93 ± 8.3  36 ± 1.52116 ± 8
R70225 ± 2159 ± 5.641 ± 3.8102 ± 9.3  40 ± 2.1126 ± 12
R73218 ± 2155 ± 5.040 ± 3.6 97 ± 8.66  38 ± 1.8125 ± 10
Resistant populations; yellow zones
R29120 ± 1023 ± 2.621 ± 2.2 45 ± 5.1  21 ± 1.8 66 ± 5
R50104 ± 1020 ± 3.120 ± 1.9 45 ± 5.2  19 ± 1.5 62 ± 5
R60 98 ± 1018 ± 2.118 ± 1.7 42 ± 4.718.5 ± 2.1 54 ± 5
R64 98 ± 1018 ± 2.217 ± 2.1 42 ± 4.818.5 ± 1.8 55 ± 4
R70101 ± 1121 ± 2.420 ± 1.7 44 ± 5.119.6 ± 2.1 62 ± 5
R73120 ± 1122 ± 2.622 ± 1.6 44 ± 4.121.1 ± 2.2 65 ± 6
Figure 3.

Density of stomata in the lower epidermis of the tenth leaf of (a) susceptible (S) Amaranthus powellii plants (population 41) and (b, c) resistant (R) plants (population 29) for the (b) wrinkled and (c) yellow zones 28 d after emergence. Arrowheads, mother guard cells and developing stomata. Bars, 25 µm.

Resistant populations  The six R populations showed similar structural patterns, which consisted of the presence of a mosaic of histologically heterogeneous areas in the same leaf blade. In the dark-green wrinkled zones, the upper and lower epidermises were sinuous (Fig. 2), and distinct secondary convoluted microareas were sometimes formed. The cells of the lower epidermis were smaller than in leaves of the S populations (Fig. 3). The mesophyll and the veins were somewhat more developed and the leaves appeared to be thicker than in the S plants (Table 2, Fig. 2). The thickness of the leaf blade in these zones ranged between 200 and 225 µm, the palisade parenchyma was between 51 and 59 µm thick, and the spongy parenchyma ranged between 33 and 41 µm (Table 2). While the tertiary veins were thicker than in S leaves, ranging between 83 and 93 µm, the bundle sheaths were only marginally thicker (36–40 µm) (Table 2). The adaxial epidermis had raised stomata (Fig. 2), which were at a higher density (116–126 stomata mm−2) than in the epidermis of susceptible plants (Table 2, Fig. 3).

The yellow leaf blade zones of R plants were much thinner than the wrinkled zones, and also much thinner than the leaf blade zones of the S plants (Fig. 2). The abaxial epidermal cells were much smaller than in the wrinkled areas. Their anticlinal walls were not undulated as in the wrinkled areas or as in the epidermal cells of S plants (Fig. 3). The mesophyll was poorly differentiated and the tertiary veins underdeveloped. The bundle sheath cells were small, with thin walls, and the vascular tissue was represented only by a few primary elements (Table 2, Fig. 2). Chloroplasts were visible only in hand sections and were located only in the bundle sheath cells. The thickness of the leaf blade in the yellow zone was about half that of the leaf blade in the wrinkled zone and ranged between 98 and 120 µm. All other tissues were also c. 50% thinner than the same tissues in S plants or in the wrinkled zones of R plants (Table 2). Stomata were in the same plane with other epidermal cells but occurred at much lower densities (54–66 stomata mm−2) than in other leaf types. There were only a few fully developed stomata and numerous guard mother cells and immature stomata (Fig. 3).

Anatomy of roots and stems

While the structures of roots and stems in S and R populations were essentially the same, important quantitative differences were observed between the two groups. The roots and stems of S populations were up to 4 times thicker than those of the R populations; the root diameter of R plants ranged from 1.82 to 4.98 mm, while that of S plants was 6.79–8.59 mm (Tables 3 and 4, Fig. 4).

Table 3. Anatomical measurements on the roots of susceptible (S) and resistant (R) Amaranthus powellii populations
PopulationRoot diameter (mm)Supplemental secondary zones (mm)Central secondary tissue (mm)Differentiated cambial zones (no.)
  1. Values for root diameter, supplemental secondary zones and central secondary tissue are given as mean ± standard error.

  2. Thicknesses of the principal root, phelloderm and cortex, supplemental secondary zones and central secondary tissue and the number of differentiated cambial zones were measured in six herbicide-susceptible and six herbicide-resistant populations of A. powellii.

Susceptible populations
S127.0 ± 0.102.3 ± 0.031.7 ± 0.023
S417.3 ± 0.112.4 ± 0.031.7 ± 0.033
S537.1 ± 0.112.3 ± 0.031.7 ± 0.033
S836.7 ± 0.091.9 ± 0.021.6 ± 0.033–4
S848.6 ± 0.133.0 ± 0.021.8 ± 0.033–4
S868.1 ± 0.132.8 ± 0.031.8 ± 0.023–4
Resistant populations
R295.0 ± 0.091.2 ± 0.021.0 ± 0.022
R503.6 ± 0.080.9 ± 0.020.8 ± 0.022
R601.8 ± 0.040.2 ± 0.010.6 ± 0.021–2
R641.9 ± 0.030.3 ± 0.010.6 ± 0.021–2
R702.4 ± 0.050.5 ± 0.010.7 ± 0.022
R734.3 ± 0.071.1 ± 0.020.9 ± 0.022
Table 4. Thickness of the stem, cortex, cambial zone, and central cylinder in six herbicide-susceptible (S) and six herbicide-resistant (R) populations of Amaranthus powellii
PopulationStem diameter (mm)Cortex (µm)Cambial zone (µm)Central cylinder (mm)
  1. Values are given as mean ± standard error.

Susceptible populations
S126.3 ± 0.07414 ± 4.5626 ± 1.15.4 ± 0.06
S416.5 ± 0.07408 ± 4.126 ± 1.15.2 ± 0.05
S536.1 ± 0.06401 ± 4.024 ± 1.35.2 ± 0.05
S835.6 ± 0.06372 ± 4.025 ± 1.84.8 ± 0.05
S847.3 ± 0.07421 ± 4.731 ± 2.16.3 ± 0.06
S867.0 ± 0.07416 ± 5.131 ± 1.96.1 ± 0.06
Resistant populations
R293.7 ± 0.05321 ± 5.012 ± 2.13.1 ± 0.04
R502.9 ± 0.03233 ± 3.810 ± 0.82.3 ± 0.04
R602.3 ± 0.03208 ± 3.410 ± 0.81.9 ± 0.03
R642.3 ± 0.03208 ± 3.710 ± 0.81.9 ± 0.02
R702.5 ± 0.03219 ± 4.310 ± 0.82.1 ± 0.03
R733.7 ± 0.04321 ± 3.612 ± 0.93.1 ± 0.03
Figure 4.

Cross-section of roots and stems of Amaranthus powellii plants at 28 d after transplanting. Roots of resistant (R) plants (a) are much smaller than those of susceptible (S) plants (b). The R plants have only two supplementary cambia while the S plants have three. Stems of R plants (c) are also much less developed than those of S plants (d). Large arrowheads, supplemental cambia in the roots and the first cambium in the stems; small arrowheads, intrafascicular cambia of collateral bundles in stem bundles. Bars, 250 µm.

The secondary structure of roots and shoots of Amaranthus spp. is characterized by the centrifugal development of supplemental successive cambial zones. In the roots, additional cambia are more or less concentric and continuous (Costea & DeMason, 2001). The differentiation of both primary and secondary tissues was considerably delayed and reduced in R plants compared with the S plants. The roots of S populations at 28 d after transplanting had already three or four supplemental active cambia, while the R plants had only one or two (Table 4, Fig. 4). In addition, cambia were less active in the roots of R plants compared with S plants. The thickness of these additional secondary zones ranged between 0.29 and 1.20 mm in R and between 1.91 and 2.99 mm in S populations. The central secondary tissue produced by the first vascular cambium was also much thinner in R plants (0.6–1.0 mm) than in S plants (1.6–1.8 mm).

Stems of R plants were much thinner than those of S plants. The diameter at the base of the stem ranged between 5.6 and 7.3 mm in S plants and between 2.3 and 3.7 mm in R plants (Table 4). Similarly, cortex thickness ranged from 372 to 421 µm in S plants and from only 208 to 321 µm in R plants. The central cylinder was more developed in S plants (4.8–6.3 mm in diameter) than in R plants (1.9–3.1 mm in diameter) (Table 4). The cambial zone and the interfascicular cambia developed faster and were more active in the stems of S plants than in R plants (Fig. 4). As a result, the first cambial zone was c. 2.5 times thicker in S plants (25–31 µm) than in R plants (10–12 µm) (Table 4).

Growth of susceptible and resistant populations

Over a period of 28 d after transplanting, the S plants always had more leaves than the R plants (Fig. 5). The greatest difference was between day 16 and day 20, when the S plants had on average three more leaves than the R plants. The rate of leaf appearance in the R plants increased thereafter such that, at day 28, there was no difference in leaf number between the S and R plants (Fig. 5). At 28 d after transplanting, the R populations had 58% lower leaf area (LA) than the S populations (Fig. 6) and accumulated 67% less above-ground biomass than the S plants (Fig. 7). On average, the LA of S plants (270 cm2) was significantly greater than that of the R plants (114 cm2). The S populations accumulated 3 times more biomass than the R populations: the average above-ground biomass of the S populations was 3.3 g per plant whereas it was 1.1 g per plant for the R populations. LA and biomass varied less among S populations than among R populations (Figs 6 and 7). After 28 d, population R29 produced twice as much biomass and LA as the averages of the five other R populations. At this time, the R plants had much less leaf surface area than the susceptible plants, mainly because of the smaller size of their leaves.

Figure 5.

Leaf numbers of susceptible (S; open squares) and resistant (R; closed squares) Amaranthus powellii plants from 8 to 28 d after transplanting. Each point is the average of six populations over three repeated experiments with five replications each. Regression lines were drawn with the following equations: y = 0.1435x1.4341 (r2 = 0.976) for S and y = 0.0252x1.9291 (r2 = 0.954) for R, where x is the number of days.

Figure 6.

Leaf area of susceptible (S) and resistant (R) Amaranthus powellii populations 28 d after transplanting. Each bar is the average of three experiments with five replications each. Error bars indicate standard error of the mean (n = 15).

Figure 7.

Aerial dry biomass of susceptible (S) and resistant (R) Amaranthus powellii populations 28 d after transplanting. Each bar is the average of three experiments with five replications each. Error bars indicate standard error of the mean (n = 15).

Determination of competitive ability

Populations S41 and R29 were used in order to test whether the differences in growth between S and R plants would affect their competitiveness. After 63 d, in the absence of interpopulation competition, the S41 plants produced 10 and 30% more above-ground biomass than the R29 plants in runs 1 and 2, respectively (Fig. 8). The difference between the two runs was likely attributable to the time of the year at which these experiments were conducted. When the two populations were mixed, the S plants always outperformed the R plants (Fig. 8). The competitive advantage of the S41 plants was such that, even when they were grown in a proportion of 6 : 18 (S41:R29), they still produced c. 4–5 times more biomass than the R29 plants (Fig. 8). The RCC values were 15.0 [standard error (SE) 3.0] in run 1 and 6.8 (SE 0.9) in run 2 and were significantly different than 1 (P < 0.0001). This indicates that the S41 plants were much more competitive than the R29 plants. The rapidity of growth of the S41 plants in the early stages of development helps them outcompete the R29 plants, even when they are present in much lower proportions. This difference in competitive ability also had a great impact on the seed production of the R29 and S41 plants. Seed biomass was determined for the monocultures and for the 12 : 12 ratio. In monoculture, R29 plants produced c. 30% less seed biomass than S41 plants (Fig. 9). However, when grown at a ratio of 12 : 12, S41 produced the majority of the seed biomass. Under competition, R29 plants produced only 2.5 and 9.4% of the total seed biomass in runs 1 and 2, respectively. This indicates that the reduced competitiveness of the R plants in the early stages of their growth negatively affects their reproductive ability.

Figure 8.

Replacement series diagrams for above-ground dry biomass of Amaranthus powellii populations S41 (open squares) and R29 (closed squares) 54 d after planting. Error bars represent standard error of the mean (n = 4).

Figure 9.

Seed biomass of susceptible (S41, white bars) and resistant (R29, black bars) Amaranthus powellii plants 54 d after planting. A. powellii plants were grown in monoculture or mixed in equal proportions. Each bar is the average of four replications and error bars represent standard error of the mean (n = 4).


The anatomy, morphology and growth pattern of the R plants were considerably different from those of the S populations during the early stages of development. As all R plants had the mutation coding for the AHAS Trp574Leu substitution and the S plants had not, our results suggest pleiotropic effects on plant growth and development. Assuming that this association is real, this would be the first report of this particular mutation having pleiotropic effects on plant growth. Others have observed that mutations in AHAS coding for amino acid substitution at Pro197 have a pleiotropic effect on seed amino acid content or on germination behavior (Dyer et al., 1993). However, there has been no detailed report of any herbicide resistance-conferring mutation in AHAS having extended pleiotropic effects on anatomy and morphology. Moreover, these alterations severely impair growth and negatively impact competitive ability and fitness.

The AHAS Trp574Leu substitution has been observed in other weed species (Tranel & Wright, 2002). There has, however, been no report of any growth disruption linked with this mutation in these species. It is possible that the pleiotropic effect that is observable in A. powellii is stronger or less transient than in other species. Whether such a large impact of this mutation is present in other species would be worth examining.

The leaves of R plants were characterized by their smaller size, which was reflected in smaller LA compared with S plants (Figs 1 and 6). Furthermore, leaf blades of R plants had two distinct morphological zones: a dark-green, wrinkled area in the upper half of the leaf blade, and a yellow zone in the lower half. Leaf blade segments, between certain secondary and tertiary veins from the distal half, grew abaxially and formed extensive convoluted areas.

The yellow zones had chlorophyll only in the bundle sheaths along the tertiary veins, and this was not the result of chlorosis (a reduction in the level of chlorophyll) but rather the consequence of retarded development. This means that those tissues remain in a juvenile developmental state and so chlorophyll was not synthesized. The structure of the basal yellow zones corresponded to that of typical immature leaves 10–20 mm long. The lack of differentiation and development of the mesophyll and the vascular tissue, the morphology of the epidermal cells and the presence of many guard mother cells and immature stomata were all signs that the tissues were in a juvenile state. At the other extreme, the histology of the wrinkled/convoluted areas suggested a tendency to overdevelopment. The wrinkled appearance of these zones is attributable to the fact that the intense localized growth of the lamina was uncoordinated with the slower development and spacing of the vascular framework of secondary and tertiary veins. The stomata are more numerous, the mesophyll is thicker and more intensely green, and the vascular tissue is better developed. Variable-sized transitional areas (having a range of intermediate characteristics that sometimes approach the structure of typical leaves) separate the two developmentally contrasting structures.

The normal ontogeny of dicot leaves produces leaves that are a 3D mosaic of different cell ages (Nelson & Langdale, 1989). In contrast to monocots, where the cell divisions are polarized, in the dicot leaves additional anticlinal divisions may occur in all directions (Nelson & Langdale, 1989; Nelson & Dengler, 1997). However, this developmental heterogeneity is discrete and the new cells are soon integrated into the structural and functional leaf continuum. The vascular system in normal leaves represents the spatial framework for the development of the bundle sheath and mesophyll cells. After the midvein and secondary veins develop, the network of tertiary and minor veins advances in the basipetal direction (reviewed by Dengler & Nelson, 1999). In R populations an abnormal medley of developmentally contrasting regions is formed. The basipetal development of leaves is not completed and the basal areas remain juvenile and underdeveloped. The wrinkled areas may be interpreted as an attempt to compensate for the functional reduction of LA in the yellow regions. In any case, the integrated and coordinated development that is present in normal leaves is lost in R leaves. This polarization of morphogenesis in over- and underdeveloped areas is a striking abnormality, which cannot occur without negative fitness effects. The mutation interferes with the normal C4 expression in developing leaves (see Wang et al., 1992), but the precise mechanism remains to be discovered. All R population leaves from the first to the tenth showed extreme symptomology, as exemplified by leaf 4 (Fig. 1) and leaves 9 or 10 (Fig. 2, Table 2). This symptomology gradually became less pronounced after the 10-leaf stage.

The difference in competitiveness that is caused by the slower early development and altered morphology strongly reduces biomass accumulation and reproduction when the R plants are grown in competition. This lack of competitiveness resulted in the R29 plants producing much less above-ground and seed biomass when competing with S41(Figs 8 and 9). Therefore, the herbicide resistance trait appears to have a profound negative effect on fitness in the absence of herbicides. This means that the proportion of R plants in a mixed population might be maintained at a very low level compared with wild-type herbicide-susceptible plants when AHAS-inhibitor selection pressure is discontinued. It is, however, important to note that the experiment was conducted in the glasshouse, under conditions of high fertility and water availability. This means that competition likely was mainly for light. In such cases, it is possible that the differences in size and growth that we observed in the early phases of development, and that we assume are critical for the outcome of competition between S and R plants, might have been exaggerated compared with the situation in the field. Under natural field conditions, plants are subjected to potentially high variations in temperature and water availability. In addition, competition from crop plants can be higher than intraspecific competition. Other biotic and abiotic factors such as cultivation, insect predation or diseases could also influence the outcome of competition. Under circumstances of competition mostly for light such as occurred in the glasshouse, a small initial size difference could result in a large difference in the competition outcome (Weiner, 1986). It might be logical to assume that, under field conditions, the difference in competitiveness between R and S plants might be significantly lower than the difference we observed under glasshouse conditions. However, in Ontario, Amaranthus spp. are mostly found in well-fertilized agricultural fields and are emerging at times when moisture is available (Rajcan et al., 2004). It would therefore be reasonable to assume that, under normal conditions, competition would be mostly for light and therefore our results, although obtained under artificial conditions, can still provide a realistic estimate of the differences that would occur in the field.

Comparing six populations of each type allowed us to ensure that plants were behaving similarly as a group. For both the total biomass and the LA, the R and S plants responded as two distinct groups for the period of 28 d that followed emergence. This indicated that the pleiotropic effect induced by the mutation was high enough to overcome any genetic background variation that existed among the different populations. The possibility exists that the pleiotropic effects we have observed were not caused by the mutation but were only correlated with it. There is a possibility, although it is faint, that only one mutation event occurred in a plant that already had this particular growth pattern and that it later spread into the different fields where our populations were collected. Gene flow in Amaranthus spp. is most likely to occur by seeds carried by field equipment. All the populations (susceptible and resistant) were from an area of Ontario that is generally under a maize–soybean (Zea maysGlycine max) rotation. In some circumstances, growers may have over-relied on AHAS-inhibiting herbicides as they extended the soybean part of the rotation. This has resulted in a high incidence of AHAS-inhibitor resistance in Amaranthus spp. populations (Ferguson et al., 2001). However, it is worth noting that this similar selection pressure in very distinct fields has resulted in at least five different mutations being selected (Ala122Thr, Ala205Val, Asp376Glu, Trp574Leu and Ser653Thr) depending on the populations (Corbett, 2004; McNaughton et al., 2005). This shows that multiple selection events took place in separate fields.

Of all the R populations, R29 was the most productive 28 d after transplanting (Figs 6 and 7). Yet, when put in competition with S41, its competitiveness and fitness were dramatically impacted (Figs 8 and 9). This is because the delayed growth that occurs in the first few days after emergence is likely to play a determinant role in the outcome of the competitive interaction. Population S41 is able to grow faster, produce more LA and therefore capture a greater share of the available resources.

As seeds were put directly in the soil, we were able to detect interactions between R29 and S41 plants that occurred immediately after emergence. This initial interaction is critical as we observed that symptomology of the R plants was most pronounced in the early stages of growth and tended to gradually disappear after the 10-leaf stage. Many other studies that have used transplants of pregerminated seedlings lack the ability to measure the full scope of the competitive interaction between R and S plants.

The link between the altered AHAS enzyme caused by the Trp574Leu substitution and the morphological/anatomical symptoms observed in the R populations is not clear. This mutation produces a very highly resistant enzyme, which allows plants to survive high rates of AHAS inhibitors (Tranel & Wright, 2002). The impact of this substitution on other enzyme properties has been little investigated, especially in plants. In yeast, the Trp586Leu substitution (equivalent to Trp574Leu in plants), while conferring high resistance to SU herbicides, also increased c. 5-fold the specific activity and increased the Km (Michaelis–Menten constant) of pyruvate by almost 9-fold (Pang et al., 2003). A tobacco AHAS with a Trp574Phe substitution generated by site-directed mutagenesis and expressed in Escherichia coli had high-level resistance to SU, IMI and TP herbicides (Chong et al., 1999). This substitution had a huge impact on substrate affinity as it increased apparent Km by c. 20-fold and reduced the maximum velocity Vmax by more than half. While these substitutions are different from the Trp574Leu substitution present in the A. powellii populations we have studied, they illustrate the crucial role of Trp574, not only on herbicide binding but also on the overall functioning of the enzyme. An enzyme with a substituted amino acid at this position still has catalytic activity, but it clearly differs from wild-type enzymes. It is therefore likely that Trp574Leu substitution in A. powellii also results in a functioning but somewhat impaired enzyme. Whether this is the case, and what eventual link this would have with the growth disruption that we have observed, remain to be determined.

Our results are consistent with the hypothesis that the Trp574Leu mutation comes at a high cost to the plants that carry it. This fitness penalty was attributable to reduced competitiveness in the early phases of development and was expressed through important anatomical and morphological disruption. The reduced biomass accumulation and leaf expansion and delayed leaf appearance were critical when the plants were in competition with wild-type plants. It is also striking that, in monoculture, the R29 plants had their productivity reduced by only 10–30% relative to the S41 plants. While this reduction was important, it was much less than the difference we observed in the early stages of development, from emergence to the 10-leaf stage. Whether such a difference in competitiveness would manifest to the same extent in the field remains to be determined. If such a difference was to exist, in the absence of selection pressure from the herbicide, S plants would quickly overcome R plants and dominate the population.


This work was partly funded by the Natural Science and Engineering Research Council of Canada and the Ontario Ministry of Agriculture and Food through research grants to FJT. The authors thank Mr Peter Smith for his expert technical help and three anonymous reviewers for very insightful comments on an earlier version of this manuscript.