SEARCH

SEARCH BY CITATION

Keywords:

  • ACCase mutation;
  • after-ripening;
  • dark stratification;
  • fatal germination;
  • light requirement;
  • P450 metabolism;
  • weed management

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • 1
    Quantification of fitness differences between herbicide-resistant and herbicide-susceptible weeds permits better prediction of herbicide resistance, and the design of weed management strategies to exploit those traits that result in reduced ecological performance. Reported here is the first attempt to compare the germination and seedling emergence characteristics of one herbicide-susceptible and two herbicide-resistant phenotypes from a single weed population.
  • 2
    A series of experiments was carried out in controlled conditions to study seed germination and emergence in Lolium rigidum phenotypes possessing target-site and non-target-site mechanisms of herbicide resistance. Lines composed of herbicide-susceptible (S), metabolism-based (P450) and target site-based (ACCase) resistant individuals, isolated from a single multiple-resistant population (SLR31), were used in the comparative experiments.
  • 3
    No major differences in seedling emergence were found among the phenotypes when exposed to an alternating 25/15 °C cycle with a 12-hourly photoperiod. However, the absence of light associated with soil burial (1–8 cm) markedly inhibited total germination and seedling emergence in the ACCase phenotype compared with the S and P450 phenotypes.
  • 4
    Germination at constant temperatures was also inhibited in seeds of the ACCase phenotype, which showed the highest base temperature (Tb) for germination and required more time to reach 50% emergence (tE50) than the S and P450 phenotypes.
  • 5
    Seedling emergence from deep burial (8 cm) in soil promoted significantly higher fatal germination in the S and P450 phenotypes compared with the ACCase phenotype. Despite fatal germination, the S phenotype produced greater emergence from deep burial than both herbicide-resistant phenotypes.
  • 6
    Synthesis and applications. This study demonstrates differential germination and emergence responses to light and thermal environments between herbicide-susceptible phenotypes and phenotypes possessing target-site (ACCase) and non-target site (P450) herbicide resistance within a single L. rigidum population. Shallow seed burial (1 cm) by any cultivation tool will potentially inhibit seedling recruitment of the ACCase phenotype in contrast with the S and P450 phenotypes, which require less restricted conditions to germinate and emerge.

Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Almost 50 years after Harper (1956) predicted that weed species would evolve resistance to herbicides, modern cropping agriculture contends with the existence of herbicide-resistant populations in 171 weedy plant species world-wide (Heap 2004). In Australia, although persistent herbicide use has resulted in the evolution of herbicide-resistant populations in 27 weedy plant species, the dominant problem is with Lolium rigidum Gaudin. This allogamous annual grass weed is very widespread, often at high densities, genetically diverse and, following persistent herbicide selection, has evolved multiple herbicide resistance to selective and non-selective herbicides across at least nine different modes of action (Burnet et al. 1994; Powles et al. 1998; Preston & Powles 2002a).

A major feature of herbicide resistance in Australian L. rigidum populations is the ability to accumulate several resistance mechanisms within individuals and within populations (i.e. multiple resistance) (Tardif & Powles 1994; Preston et al. 1996). The most studied L. rigidum population (SLR31) exhibits, at the population level, multiple resistance to aryloxyphenoxypropionate (APP), cyclohexanedione (CHD), sulphonylurea (SU), imidazolinone (IM), dinitroaniline, carbamate and chloroacetamide herbicides (Christopher, Powles & Holtum 1992; Burnet, Barr & Powles 1994; Holtum et al. 1994; Tardif & Powles 1994; McAlister, Holtum & Powles 1995). Each of these herbicide classes inhibits different and specific target enzymes. APP and CHD herbicides inhibit fatty acid biosynthesis by inhibiting acetyl-CoA carboxylase (ACCase herbicides), whereas SU and IM herbicides inhibit branched-chain amino acid biosynthesis by inhibiting acetolactate synthase (ALS herbicides) (reviewed by Preston & Mallory-Smith 2001).

It is well established that both target-site and non-target-site mechanisms confer resistance to ACCase and ALS herbicides in the L. rigidum SLR31 population and that individuals may possess one or more of these resistance mechanisms (reviewed by Hall, Holtum & Powles 1994; Preston & Powles 2002b). Approximately 15% of SLR31 individuals are target-site resistant to ACCase herbicides because of a resistant ACCase (Tardif & Powles 1994) endowed by a point mutation changing isoleucine to leucine at position 1781 of the ACCase gene (X. Zhang & S. B. Powles, unpublished data). A larger proportion of the SLR31 population is resistant to ACCase, ALS and other herbicides, because of enhanced rates of herbicide metabolism mediated by the cytochrome P450 mono-oxygenase enzyme complex (Christopher et al. 1991; Christopher, Preston & Powles 1994; Tardif & Powles 1994; Preston & Powles 1998).

A major challenge in predicting the wider effects of mutations that endow herbicide resistance is determining whether these mutations have pleiotropic effects on plant fitness. Understanding which biological attributes play a major role in determining fitness and how these interplay with environmental factors is essential information for predicting the impact of herbicide resistance in weed populations. Any determination of the relative ecological fitness of herbicide-resistant and susceptible phenotypes must assess traits that contribute to success throughout the entire life cycle (e.g. seed germination, seedling survival, relative growth rate).

Knowledge of the resistance-endowing mechanism and/or mutation is essential to assess fitness of herbicide-resistant and -susceptible plants because different resistance genes are likely to confer different pleiotropic effects (Roux, Gasquez & Reboud 2004). Furthermore, herbicide-resistant individuals should be compared with susceptible individuals with a similar genetic background (Bergelson & Purrington 1996). A specific proline–serine mutation of the psbA gene, which confers resistance to triazine herbicides in many species, endows a reduction in photosynthetic efficiency and thus a fitness penalty (reviewed by Gronwald 1994; Preston & Mallory-Smith 2001). Studies with Arabidopsis thaliana containing a specific mutation of the ALS gene (proline 197–serine), which confers resistance to certain ALS herbicides, reveal that this mutation results in a fitness penalty (Purrington & Bergelson 1997; Roux, Gasquez & Reboud 2004). In other studies with field-evolved weed biotypes resistant to ACCase or ALS herbicides, in which the genetic background could not be controlled, fitness penalties have not been identified (Thompson, Thill & Shafii 1994; Gill, Cousens & Allan 1996; Wiederholt & Stoltenberg 1996; Poston, Wilson & Hines 2002; Rashid et al. 2003).

In the above-mentioned studies, the resistant populations probably only possessed one resistance mechanism, and in some studies the specific mechanism and/or gene mutation(s) endowing resistance were unknown. Increasingly, herbicide-resistant weed populations exhibit multiple herbicide resistance because of possession of multiple herbicide-resistance mechanisms/genes. Until now, no attempt has been made to determine plant fitness associated with non-target herbicide-resistance mechanisms. We examined the relative fitness of individuals within a multiple herbicide-resistant population that exhibits two different resistance mechanisms. We report on the dynamics of seed germination and seedling emergence under different thermal and light environments and reveal an unexpected difference between biotypes.

Materials and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

plant material and selection of susceptible and resistant phenotypes

The multiple herbicide-resistant L. rigidum population SLR31 originates from Bordertown, South Australia (140°46′E, 36°18′S). Since first collected from the field, several generations have been produced under controlled conditions. The P450 and ACCase phenotypes exhibit differential responses to the ACCase-inhibiting APP and CHD herbicides. For instance, the resistant ACCase confers resistance to certain CHD herbicides (e.g. sethoxydim), while those individuals in the population possessing only enhanced herbicide metabolism (P450) are susceptible to these CHD herbicides but resistant to diclofop-methyl (Tardif & Powles 1994). Based on this difference, susceptible (S) and P450- and ACCase-based resistant individuals could be identified within the SLR31 population (Fig. 1).

image

Figure 1. Overview of the herbicide selection protocol carried out to isolate the three phenotypes possessing different mechanisms of resistance in SLR31. (1) Selection with diclofop-methyl allowed identification of herbicide susceptible plants (S). (2) Individuals susceptible to sethoxydim were identified as enhanced metabolism-based resistant plants (P450), whereas sethoxydim-resistant plants were confirmed as target-site based resistant mutants (ACCase) (see text for details).

Download figure to PowerPoint

In 2001, 500 SLR31 seeds were sown in 40 plastic containers on 0·7% (w/v) agar and incubated at a 12-hourly alternating 25/15 °C cycle with a 12-h photoperiod. Light was provided by two 40-W fluorescent white tubes giving a photon flux density of 20 µmol m−2 s−1. After 10 days, 300 randomly chosen seedlings were transplanted into 10 trays (28 × 33 × 5 cm) (30 per tray) containing a standard potting mixture (50% peat moss : 50% river sand) and were maintained under glasshouse conditions. At the three to four tiller stage, plants were excavated from trays and one tiller per plant (one clone) carefully excised. These clones were trimmed to 1 cm of shoot material, repotted and maintained under glasshouse conditions. The parent plants were also repotted (five per 25-cm diameter pot) and allowed to continue growing outdoors in normal winter growing conditions. Parent plants and their respective clones were numbered to ensure subsequent identification. When the clones reached the two to three leaf stage, they were sprayed with 450 g ha−1 diclofop-methyl (Hoegrass EC 375 g a. i. L−1 plus 0·25% wetting agent BS1000). Herbicides were applied in a laboratory spray cabinet, with a 2-nozzle boom travelling at 3·6 km h−1 and delivering a volume of 100 L ha−1 water at 200 kPa. Thirty days after diclofop-methyl treatment, 6% of the 300 clones had died and were classified as the herbicide-susceptible (S) phenotype within the SLR31 population. The surviving 94% were sprayed with 186 g ha−1 sethoxydim (Sertin EC 1 L plus 1% wetting agent Hasten). Diclofop-methyl-resistant plants that were susceptible to sethoxydim were classified as the P450 phenotype. The remaining fraction of the SLR31 population (28%) that survived exposure to both diclofop-methyl and sethoxydim was classified as the ACCase phenotype. It is emphasized that this ACCase phenotype also possesses enhanced P450 metabolism (see the Results).

Following herbicide application and the classification of SLR31 clones as herbicide-susceptible (S), target-site (ACCase) or metabolic (P450) resistant, their corresponding parent plants were identified. Plants of each phenotype (n = 18 for S, n = c. 80 for P450 and n = c. 80 for ACCase) were repotted into 25-cm diameter pots (five plants per pot when possible) and grown to maturity in a common environment (experimental garden). The smaller sample size for bulk-crossed S plants was unavoidable as only 18 S clones (6% of 300) were identified. As a result, the expression of the S phenotype may have been compared in relatively fewer genetic backgrounds. We believed that this would not significantly bias estimates of relative fitness. Similar flowering time was observed for the three phenotypes. Each phenotype was independently bulk-crossed by surrounding all pots containing that phenotype with pollen-proof enclosures. Mature seed was collected from senescing plants at the end of the growing season. After threshing and seed cleaning, seed water content was evaluated by drying a seed sample (n = 150) of each phenotype at 103 °C for 17 h (Steadman, Crawford & Gallagher 2003b). Expressed on a dry weight basis, 9–9·5% moisture content was found in all three phenotypes. Seeds were placed inside aluminium foil bags (one bag for each phenotype) with minimal airspace to maintain water content, and stored at constant 20 °C.

characterization of selected s, p450 and accase phenotypes

Seedlings of the S, P450 and ACCase phenotypes were sprayed with appropriate herbicides to confirm their resistance status. Seeds were germinated in plastic containers as described previously. Similar-sized seedlings were transplanted into 18-cm pots (20 pot−1) containing a standard potting mix and kept outdoors during the normal growing season for L. rigidum. For each treatment, three (diclofop-methyl and sethoxydim) or four (chlorsulphuron) replicates were used. When plants reached the two to three leaf stage, they were sprayed with diclofop-methyl (APP) (EC 375 g L−1, BS1000 adjuvant 0·25%) at 93·75, 187·5, 375, 750, 1500 and 3000 g a.i. ha−1, sethoxydim (CHD) (EC 186 g L−1, Hasten adjuvant 1%) at 12, 23, 47, 93, 186, 372, 744, 1488, 2976 and 5952 g ha−1 and chlorsulphuron (SU) (WDG 750 g kg−1, BS1000 adjuvant 0·1%) at 0, 4, 8, 15, 30, 60, 120 and 240 g ha−1. Dose–response curves were obtained by assessing plant mortality following diclofop-methyl or sethoxydim treatments. For chlorsulphuron-treated plants, above-ground biomass of green plants was harvested 5 weeks after treatment and dried at 70 °C for 72 h then weighed. These values were used to calculate biomass per phenotype as a percentage of the untreated control. A two-way analysis of variance (anova) was performed to determine principal effects of phenotype and herbicide on plant survival and biomass. Plant biomass (% of control) and survival (%) values were angular-transformed (y= arcsine √x) to increase normality and variance homogeneity. The GR50 (herbicide rate required to cause 50% growth reduction) was estimated for the phenotypes under chlorsulphuron selection by regression analysis using the logistic model:

  • image(eqn 1)

where B is the biomass as a percentage of the mean control at herbicide rate x, a is the maximum biomass (%) and b is the slope at GR50.

seedling emergence: effect of burial depth

Following seed collection and storage for 9 months in the conditions previously described, mean seed weight was measured for each of the three phenotypes and was identical (mean seed weight 2·2 ± 0·34 mg, n = 300). Twenty seeds of each phenotype (mean weight 1·9–2·5 mg ± 15%) were seeded in a standard potting mixture in 9-cm diameter pots at six seeding depths (0, 1, 2, 4, 6 and 8 cm). Pots were arranged in a completely randomized design in a controlled environment room at alternating 25/15 °C with a 12-h photoperiod. Air and soil temperature was monitored hourly by data loggers at the soil surface and at 8 cm depth, respectively (Gemini Data Loggers, NSW, Australia). Pots were watered to field capacity twice a day. A photon flux density of 280 µmol m−2 s−1 was provided by six metal halide lamps (1500 W) and two high-pressure sodium lamps (1500 W). The experiment was conducted twice with four and six replicates used for the first and second experiment, respectively. Seedling emergence was evaluated daily until no further emergence was recorded. Buried seeds were considered emerged when the coleoptile was visible at the soil surface. Seeds sown on the soil surface (0 cm) were considered emerged when coleoptiles were 3 mm long. When no further emergence was recorded, substrate from the pots was sieved to determine if unemerged seeds remained ungerminated or had germinated but failed to reach the soil surface (fatal germination). Any ungerminated seeds were tested for viability using 1% 2,3,5-triphenyl-tetrazolium chloride solution (TZ) for 24 h in the dark at 30 °C (Steadman, Crawford & Gallagher 2003b).

seedling emergence: regression model and statistical analysis

Cumulative seedling emergence values for each phenotype were fitted to a functional three-parameter sigmoidal model using least-squares non-linear regression (SigmaPlot V6, SPSS Science). The model used was:

  • image(eqn 2)

where E is cumulative emergence over time x, a is the maximum emergence (%), tE50 is the time to reach 50% of final seedling emergence (days) and b indicates the slope around tE50. Parameter b provides an indication of the distribution of the response (emergence) over time (i.e. low b-values indicate that a high proportion of the population germinates around tE50).

When no statistically significant differences were found by anova (P < 0·05) between regression models from individual experiments, data were combined. Emergence values were angular transformed. Parameter estimates were compared by one-way anova and means separated using Tukey's HSD multiple comparison test (α = 5%). A two-way anova was performed to determine the effect of phenotype and depth of burial on seedling mortality by fatal germination. Mortality values were angular transformed and means separated using Tukey's HSD multiple comparison test (α = 5%).

seed germination: effect of light regimes

Germination of fresh seeds of each phenotype was tested immediately after collection from mature plants and seeds were subsequently stored at constant 20 °C. Germination was tested on three occasions during the following 16 months. Fifty seeds of each phenotype were incubated at 25/15 °C on 6-mm deep solidified 0·7% (w/v) agar water in 9-cm diameter Petri dishes under three light regimes: 14 days with 12-h light (14 dL); 14 days in darkness (14 dD) or 14 days in darkness followed by 14 days exposure to 12-hourly light (14 dD + 14 dL). There were three or four replicates per treatment depending on seed availability. Exclusion of light was achieved by wrapping the plates in aluminium foil. At the end of the experiment ungerminated seeds were evaluated for viability using TZ.

seed germination: statistical analysis

All data were analysed using the SAS Version 8·02 statistical package (SAS Institute, Cary, NC). Three- and two-way anova were performed on germination data. The main effects (phenotype, light regime and time) and interactions were tested using the general linear model (GLM) procedure. Angular transformations were performed on germination percentage to meet the assumptions of anova. Additional comparisons of means were performed by Tukey's HSD multiple range test. A probability level of P < 0·05 was used to delineate significant main and interaction treatment differences.

To control for the difference in total incubation time between the 14-day regime (14 dL) and the 28-day treatment (14 dD + 14 dL), germination in the 14 dL treatment was assessed for a further 14 days. An anova was conducted to compare total germination over 28 days in the two treatments. A significant difference would indicate an effect of dark imbibition of seeds on subsequent germination in the light.

seed germination: effect of thermal regimes

Seeds stored at 20 °C for approximately 9 months were germinated on solidified 0·7% (w/v) agar water (6-mm depth) in 9-cm diameter Petri dishes. Fifty seeds of each phenotype were incubated at constant 5, 8, 12, 15 and 20 °C, and 12-hourly alternating 25/15 °C (three replicates per treatment). All temperature treatments had a 12-h photoperiod with 20 µmol m−2 s−1 provided by 40-W fluorescent white light tubes. The number of germinated seeds was counted on a daily basis until no further germination was observed. Seeds were classified as germinated when a protruded radicle was visible. Seeds were removed from the plate once they germinated and ungerminated seeds were assessed for viability using TZ. Incubating temperatures were monitored and recorded once per hour until the end of the experiment using portable data loggers.

seed germination: regression model, calculation of base temperature and statistical analysis

Cumulative germination was fitted to the sigmoidal model described in equation 2. Least-squares non-linear regression curves were generated with SigmaPlot (V6, SPSS Science). Parameter estimates describing the percentage of seed germination (G) were a maximum germination (%), tG50 (the time to reach 50% final germination, in days) and b (slope around tG50). A regression model was used to calculate the number of days required to reach each decile from 10% to 90% of germinated seeds for each constant temperature and for each phenotype by rearranging equation 2. Germination rates for each decile were calculated as the inverse of the time (i.e. 1/t) required to reach that decile of germination and were plotted against the incubation temperature. The base temperature (Tb) (temperature at which the rate of germination is zero) was calculated from a linear regression of germination rate and incubation temperature. Thermal time required for 50% of germination of each phenotype was calculated as:

  • image(eqn 3)

where θ is the accumulated thermal time (measured in degree days, °Cd) up to day n required to reach 50% of germination, Ti is the incubation temperature and Tb is the estimated base temperature.

To comply with the assumptions of anova, daily percentages of germination and Tb values were angular and log-transformed (y = log x), respectively. A one-way anova was performed to compare parameters obtained from the sigmoidal model and Tb values between L. rigidum phenotypes. Means were separated using Tukey's HSD multiple comparison test (α = 5%). Because of unequal variances, values for thermal time to 50% germination were compared with a non-parametric test (Kruskal–Wallis).

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

characterization of selected s, p450 and accase phenotypes from slr31

Herbicide dose–response experiments confirmed significant differences in the herbicide-resistance profiles of the three phenotypes within the SLR31 population (P < 0·0001; Fig. 2a–c). As expected, the S phenotype was confirmed as susceptible to sethoxydim (Fig. 2a), diclofop-methyl (Fig. 2b) and chlorsulphuron (GR50 28·2 ± 5·4) (Fig. 2c). The P450 phenotype was confirmed as resistant to chlorsulphuron (GR50 87·3 ± 24·7) (Fig. 2c) and diclofop-methyl (Fig. 2b) and, as expected, susceptible to sethoxydim (Fig. 2a). The ACCase phenotype was resistant to chlorsulphuron (GR50 120·8 ± 31·3), diclofop-methyl and sethoxydim. Resistance to all three herbicides in the ACCase phenotype confirmed that this phenotype had both a resistant ACCase and enhanced P450 metabolism. The dose–response relationships (Fig. 2) confirmed that the cloning technique had enabled isolation of herbicide-susceptible plants (S), plants with a non-target-site resistance mechanism (P450 metabolism) and plants with both the ACCase mutation and the P450 mechanism of resistance (ACCase). These three phenotypes within the SLR31 population differed in herbicide-resistance mechanisms but otherwise shared a common genetic background.

image

Figure 2. Plant survival and above-ground biomass produced by phenotypes S (circles), P450 (squares) and ACCase (triangles) under selection with increasing rates of sethoxydim (a), diclofop-methyl (b) and chlorsulphuron (c). Field rates are 186, 375 and 20 g a. i. ha−1 for sethoxydim, diclofop-methyl and chlorsulphuron, respectively.

Download figure to PowerPoint

seedling emergence: effect of burial depth

There were no differences in soil temperature (25·5 °C ± 0·3/14·5 °C ± 0·15) between the soil surface and 8-cm depth. Non-linear regressions (equation 2) provided a significant fit (P < 0·001, R2 = 0·51–0·97) to emergence data for all combinations of phenotype and seeding depth. Small differences were found in the final proportion of emerged seedlings (parameter a) between phenotypes (P < 0·05) when seeds were on the soil surface (Fig. 3a and Table 1). Seedlings of the ACCase phenotype had the lowest final emergence (59%) compared with phenotypes S (65%) and P450 (73%).

image

Figure 3. Dynamics of seedling emergence fitted to sigmoidal model of the S (circles), P450 (squares) and ACCase (triangles) phenotypes at different sowing depths when incubated at 25/15 °C (12 h/12 h) with a regime of 12 hourly light/dark. Vertical bars are SEM values. Experiments were performed twice for 0–4-cm depths.

Download figure to PowerPoint

Table 1.  Parameter estimates obtained from sigmoidal models (y = a/1 + exp(–(xtE50)/(b))) that describe the dynamics of seedling emergence over time for the S, P450 and ACCase phenotypes at increasing sowing depths at 25/15 °C (12 h/12 h). Seeds on the soil surface were exposed to a 12 h light/darkness period. Different superscript letters indicate significant differences between similar parameters within each sowing depth, according to Tukey's HSD test (α = 5%). (1) and (2) represent parameter values from the first and second experiments, respectively, where emergence responses were not identical
Depth (cm)Phenotype atE50b
SP450ACCaseSP450ACCaseSP450ACCase
0   65·5a73·0b59·1c   5·1a5·2a4·9a   1·48a1·74a1·43a
1   59·3a66·3b31·0c   4·1a4·1a4·8b   0·55a1·01a0·67a
2(1)69·9a/ (2)55·6b57·0b36·1c(1)4·2a/(2)4·9b4·6b4·8b(1)0·79a/(2)0·35b0·60ab0·36bc
4   59·6a57·0a26·6b   5·4a5·5a6·4b   0·29a0·38a0·43a
6   51·5a49·5a25·2b   6·7a6·5a7·3a   0·72a0·51a0·73a
8   33·6a20·0b26·5b   8·4a8·0a8·4a   1·13a0·55a0·80a

When seeds of the three phenotypes were buried from 1-cm to 6-cm depth, the ACCase phenotype consistently displayed less emergence than the S and P450 phenotypes (Fig. 3b–e and Table 1). There were no differences in total emergence (parameter a) between the S and P450 phenotypes for seeds buried up to 6 cm (Fig. 3b–e and Table 1). However, when seeds were buried at 8-cm depth, the P450 phenotype had a similar final proportion of emerged seedlings to the ACCase phenotype, and both phenotypes had lower final emergence than the S phenotype (Fig. 3f and Table 1). There was no clear relationship between the distribution of seedling emergence around tE50 (parameter b) and depth (Fig. 3 and Table 1). The mean values of the slopes around tE50 did not differ between phenotypes in any of the depths (Fig. 3 and Table 1). Values of tE50 increased with depth of burial (Table 1). The ACCase phenotype exhibited significantly greater tE50 values than the S and P450 phenotypes at 1-cm, 2-cm and 4-cm depths, but there were no differences in tE50 at depths greater than 4 cm (Table 1).

In order to determine if differences in final seedling emergence between phenotypes were a result of germination and subsequent mortality prior to emergence (fatal germination), or to differences in germination alone, soil was excavated from pots and sieved to determine the fate of seeds and seedlings. Seedling mortality because of fatal germination was observed at 6- and 8-cm depths. The effect of depth on the fatal germination significantly interacted with phenotype (P = 0·017). Approximately 10% seedling mortality was observed for all phenotypes at 6-cm depth. However, significantly (P < 0·05) more seedlings of the S (39%) and P450 (54%) phenotypes suffered fatal germination when emerging from 8-cm depth compared with the ACCase phenotype (10%).

Almost 100% germination was observed when buried seeds of all phenotypes that did not germinate were retrieved from the soil and exposed to ideal germination conditions (25/15 °C) with 12-hourly light. This suggests that light is an essential requirement for germination in the ACCase phenotype but not in the S or P450 phenotypes.

seed germination: effect of light regimes

As light appeared essential for germination of ACCase-phenotype seeds, but not for the S or P450 phenotypes, detailed germination and after-ripening studies were conducted. Phenotype and after-ripening (storage) time interacted to influence germination in the dark (P = 0·02). Strong dormancy was evident in fresh seeds tested at time 0 (immediately after seed collection), with less than 10% germination for all three phenotypes (Fig. 4a). Dormancy was released over time for both the S and P450 phenotypes, with up to 60% germination in darkness after 480 days of storage at 20 °C. However, even following 480 days of after-ripening, only 20% of the ACCase phenotype germinated.

image

Figure 4. Final germination of the S (open bars), P450 (shaded bars) and ACCase (black bars) phenotypes at different after-ripening times at a constant 20 °C and under three light treatments: 14 days without (14 dD) (a), with exposure to 12-hourly light (14 dL) (b), and dark-imbibed for 14 days followed by exposure for 14 days to 12-hourly light (14 dD + 14 dL) (c) at 12-hourly 25/15 °C. Different letters indicant significant differences in germination mean values, separated by Tukey's HSD test (α = 5%). Vertical bars denote SEM values.

Download figure to PowerPoint

More seeds germinated when tested in light (14 dL) than in darkness (Fig. 4b). Germination of all phenotypes increased markedly after 258 days storage, but fewer seeds of the ACCase phenotype germinated than S and P450 phenotypes (Fig. 4b).

Imbibition of seeds of the three phenotypes in darkness followed by a period of light (14 dD + 14 dL) increased germination compared with that obtained in dark or light alone (14 dD, 14 dL) at all after-ripening times (Fig. 4c vs. Fig. 4a). Thus, it appears that a period of dark imbibition followed by light accelerates dormancy release in seeds of all phenotypes. In order to eliminate the potential confounding effect of the additional incubation time when 14 dD + 14 dL and 14 dL treatments were compared, seeds in the latter treatment were allowed to germinate for an additional 14 days in light. Germination of fresh seed (time 0) of all phenotypes was greater in the 14 dD + 14 dL light regime (64%) compared with seeds maintained in germinable conditions in the light for 28 days (43%) (P < 0·0001), confirming that dark imbibition promoted dormancy release. In after-ripened seeds (258 and 480 days), which were already > 80% germinable in the light, dark imbibition followed by light (14 dD + 14 dL) did not further increase the level of germination in relation to seeds incubated for 14 days in light (14 dL) (Fig. 4c vs. Fig. 4b).

The seed germination results presented in Fig. 4a support and explain the seedling emergence patterns observed in Fig. 3. In comparison with the ACCase phenotype, significantly greater proportions of seeds of the S and P450 phenotypes were able to germinate in darkness and therefore emerge from soil (from 6 cm). The ACCase phenotype germinated poorly in darkness and therefore few seedlings emerged from below the soil surface. Failure to germinate in darkness resulted in fewer losses as there was little fatal germination in seeds of the ACCase phenotype at depth. These experiments therefore reveal a clear difference in germination response between seeds of the phenotypes with different herbicide-resistance mechanisms existing within a single population.

seed germination: effect of thermal regimes

Sigmoidal models (equation 1) were fitted (P < 0·001, R2 = 0·86–0·99) to cumulative germination data for all combinations of phenotype and incubation temperature. When seeds were placed at alternating 25/15 °C with a 12-h photoperiod, the three phenotypes showed a similar germination dynamic (Fig. 5a), although the ACCase phenotype reached a slightly lower final germination (parameter a) (91%) compared with S (94%) and P450 (96%) (P < 0·05) (Table 2). However, when seeds of the three phenotypes were germinated with a 12-h photoperiod at constant temperatures, the ACCase phenotype exhibited consistently lower germination than the S and P450 phenotypes (Fig. 5b–f and Table 2). This difference occurred across a wide range of temperatures. The percentage of final germinated seeds achieved by the ACCase phenotype ranged from 60% at 20 °C to 18% at 5 °C (Fig. 5b–f and Table 2). Final germinations at 20 °C, 15 °C and 12 °C were not significantly different between the S and P450 phenotypes (Fig. 5b–d and Table 2), but at 8 °C and 5 °C germination of the P450 phenotype was less than the S phenotype, although still greater than the ACCase phenotype (Fig. 5e–f and Table 2). The sigmoidal models (equation 2) provided another biological parameter that estimated the time (days) required to achieve 50% germination. As expected, at warmer temperatures germination of all phenotypes occurred more rapidly, as indicated by lower tG50 values. The ACCase phenotype exhibited a higher tG50 than the S and P450 phenotypes at constant 8 °C and 12 °C and alternating 25/15 °C, indicating that it germinated more slowly at these temperatures (Table 2).

image

Figure 5. Dynamics of germination fitted to sigmoidal model of the S (circles), P450 (squares) and ACCase (triangles) phenotypes under different fluctuating and constant temperature regimes and a 12 h light/darkness period. Vertical bars are SEM values. Seeds were after-ripened at a constant 20 °C for 9 months before conducting the experiment to ensure seed dormancy was minimized.

Download figure to PowerPoint

Table 2.  Parameter estimates obtained from sigmoidal models (y = a/1 + exp(–(xtG50)/(b))) that describe the dynamics of germination over time for the S, P450 and ACCase phenotypes under different temperatures. Different superscript letters indicate significant differences between similar parameters within the same temperature condition, according to Tukey's HSD test (α = 5%)
Temperature (°C)Phenotype atG50b
SP450ACCaseSP450ACCaseSP450ACCase
Constant
560·6a41·0b18·0c10·7a10·6a11·1a0·86a1·37a1·20a
883·7a72·0b36·0c 6·6a 6·7a 7·7b0·70a0·73a1·05a
1288·2a85·1a73·0b 4·4a 4·9b 5·9c0·60a0·71ab0·95b
1579·3a79·5a41·3b 3·4a 3·5a 3·6a0·39a0·47a0·38a
2092·3a89·6a63·5b 2·1a 2·1a 2·2a0·30a0·31a0·45a
Alternating
25/1594·3a96·3a91·3b 1·9a 2·0b 2·2c0·29a0·30a0·32a

The relationship between rate of germination to each decile of germination (10–90%) for all phenotypes and incubation temperature was approximately linear (R2 > 0·90). Tb of the three phenotypes was significantly different, with the P450 (4 °C ± 0·4) intermediate between the S (2·9 °C ± 0·1) and ACCase (5·6 °C ± 0·4) phenotypes (P < 0·001). The accumulated thermal time above Tb required to achieve 50% germination was similar for phenotypes S (36·8 °Cd), P450 (35·8 °Cd) and ACCase (40·1 °Cd) (P = 0·37).

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The timing of seed germination and seedling emergence is one of the most important events that determines the success or failure of annual plant species invading, infesting and increasing in cropping agro-ecosystems (Forcella et al. 2000). This study, conducted with a single, multiple herbicide-resistant L. rigidum population originating from one field, uncovered significant differences in germination and emergence responses between herbicide-susceptible individuals and individuals possessing a target-site resistant ACCase mechanism vs. a non-target metabolism mechanism (P450). The ACCase phenotype shows deeper environmentally conditional dormancy that leads to different patterns of seed germination, seedling emergence and establishment compared with individuals in the SLR31 population that possess enhanced herbicide metabolism (P450-based) or are herbicide susceptible.

Thermal and light conditions were shown to be important factors that determined differential responses between the phenotypes. A fluctuating temperature regime of 25/15 °C supplemented with light (continuous 14 days either with or without pre-imbibition in darkness) proved to be an optimal condition for germination of all phenotypes, confirming previous studies on seed germination in L. rigidum (Gill, Cousens & Allan 1996; Steadman, Bignell & Ellery 2003a; Steadman 2004). Under this optimal environment there were no major differences in germination between the three phenotypes (Figs 3a, 4b and 5a, c). However, in suboptimal environments (either constant temperature or darkness) a greater proportion of individuals comprising the ACCase phenotype was dormant compared with herbicide-susceptible or P450-based resistant plants (Figs 4a and 5b–f). This reduced germination of the ACCase phenotype in darkness resulted in a lower proportion of seedling emergence when seeds were below the soil surface (Fig. 3b–f). Even after 16 months of after-ripening, 80% of the ACCase-resistant individuals still required light to germinate (Fig. 4a). The light requirement for germination found in the majority of individuals comprising the ACCase phenotype is a heritable plant trait (data not shown). A period of imbibition in darkness (dark stratification) resulted in increased germination in light and enabled high levels of germination of fresh seeds of all phenotypes, confirming that this alternative dormancy release mechanism is a common response in L. rigidum populations (Steadman 2004) regardless of herbicide-resistance status. Light receptors are involved in seed germination responses for many weed species (Casal & Sanchez 1998). Whether different ratios of active and inactive forms of phytochrome mediate the differential germination responses between the SLR31 phenotypes merits future investigation.

Emergence of buried seeds in non-dormant populations is the result of the key processes of germination and seedling elongation rate, which in turn depend on Tb for germination, soil temperature, soil impedance and water potential (Ψ), among other factors (Forcella et al. 2000). A high proportion of the ACCase-resistant seeds was unable to germinate when exposed to either constant temperature supplemented with light or fluctuating temperature in darkness. Those seeds of the ACCase phenotype that were able to germinate under fluctuating temperatures (25/15 °C) in darkness (i.e. buried in soil) produced seedlings that took longer to reach 50% emergence (tE50) when buried at 1-cm, 2-cm and 4-cm depths, compared with the herbicide-susceptible and P450-based resistant individuals (Table 1). The higher Tb of the ACCase phenotype, together with unknown factors such as seed imbibition and seedling elongation rates, could have interacted to account for this higher tE50. Phenotypes possessing higher tE50 values, leading to delayed and poor seedling establishment, may be at an ecological disadvantage as rapid occupation of biological space by seedlings with an annual growth habit is crucial to capture light and avoid shading from competitors.

Previous reports have indicated that Tb for germination may vary between subpopulations or fractions of a population (Pritchard et al. 1999). The ACCase-resistant phenotype not only possessed the highest value of Tb but also the lowest fraction of the seed population capable of germinating at constant temperatures or in darkness. At the lowest temperature (5 °C), the herbicide-susceptible phenotype (S) had higher total germination compared with the P450- and ACCase-resistant phenotypes. Germination responses among the phenotypes at 5 °C are explained by the variation in their Tb: more thermal time was accumulated by those phenotypes (S and P450) with Tb below 5 °C, while the ACCase phenotype with Tb similar to the incubation temperature was unable to accumulate thermal units. Defining the level of seed dormancy of each phenotype as the ‘temperature-inducible fraction of the population’ (Ghersa, Benech-Arnold & Martinez-Ghersa 1992), we can conclude that there was a relationship between the level of dormancy and the Tb.

ecological significance of results

One of the biological features of Australian L. rigidum populations is the production in late spring of large amounts of seed. A high proportion of these seeds are dormant at maturity, preventing germination following any sporadic rainfall events over the ensuing long, hot, dry summer. A temperature-driven dormancy release during the summer ensures that a high proportion of the population is capable of germination in autumn, when rainfall and temperature become favourable for seedling survival and growth (Steadman, Bignell & Ellery 2003a; Steadman, Crawford & Gallagher b). This study, conducted with one population, has revealed that herbicide-susceptible and P450-based resistant individuals within this population will germinate under a range of environmental conditions. However, individuals with target-based ACCase resistance have much higher levels of seed dormancy and therefore much lower germination and seedling emergence of buried seed (without light). If the required light and fluctuating thermal requirements are met for the ACCase phenotype, then identical germination and emergence will occur for all three phenotypes within the population. Preliminary experiments on dormancy release of the three phenotypes when after-ripened under field temperatures confirm the results of this study (data not shown).

The ecological consequences of the differential germination and emergence responses between the three phenotypes are complex to predict. It is the ecological environment that will ultimately determine the success of the phenotypes during germination, emergence and establishment. For instance, the light requirement for germination in the ACCase phenotype may be beneficial if viewed as a depth-sensing mechanism to prevent fatal germination of buried seeds at deep soil layers, ensuring a longer seed-bank life and perhaps enabling emergence after early-season weed control practices. This could be interpreted as endowing greater fitness on the ACCase phenotype over the S and P450 phenotypes. However, the ungerminated buried ACCase seeds will be exposed to losses as a result of disease and predation or loss of viability (Harper 1977). Additionally, the low germinability of shallow-buried ACCase seeds could be a clear fitness disadvantage where these individuals could otherwise successfully emerge and establish. The high proportion of fatal germination observed at 8-cm depth for the S phenotype should be balanced with the fact that at 8-cm depth more seedlings of the S phenotype actually emerged compared with both resistant P450 and ACCase phenotypes (Fig. 3f). At the population level, the outcome of all these potential trade-offs will depend on ecological scenarios, which ultimately will decide whether any of the phenotypes have advantages or disadvantages.

The results of this study highlight the importance of (i) conducting investigations on the ecological responses of herbicide-resistant and susceptible populations in a wide range of environmental conditions, (ii) discriminating herbicide-resistant populations by their mechanism of resistance and (iii) ensuring minimal differences in genetic background between plants under comparison.

implications for resistance management

Knowledge of the agro-ecological behaviour of herbicide-resistant plants is generally lacking and yet may be very useful in implementing successful management (Powles & Matthews 1992). Given the results reported here, management that combines early season cultural and chemical practices to optimize the ecological fitness differential between susceptible and resistant phenotypes can be envisaged. Where there is no soil cultivation during the fallow period (summer) seeds of the S, P450 and ACCase phenotypes produced in the preceding growing season will be at or close to the soil surface. In autumn, they will experience temperature and light conditions conducive to germination and emergence (> 75%) of all phenotypes. However, seed burial as a result of any shallow cultivation will mean little germination of ACCase-resistant individuals during early autumn. Taking advantage of the relatively minor differences in germination and emergence of susceptible and P450 metabolic resistant phenotypes will require burial of seeds at depths greater than 6 cm, thus requiring a deeper soil cultivation. Additionally, similar germination responses to dark stratification observed for both resistant phenotypes could be exploitable by alternating exposure of seeds to darkness and light conditions during rainfall events during summer or early autumn. Then, soil cultivation will potentially promote and control almost 90% germination of the resistant phenotypes of SLR31.

The implications of our results for the management of herbicide-resistant populations depends on their incorporation into mechanistic models that contemplate thermal and light requirements for germination and fatal germination as parameters for predicting the timing and proportion of herbicide-susceptible and target-and non-target-site resistant phenotypes that germinate, emerge and establish in field conditions. The broader application and adoption of the management practices discussed here will depend on whether the responses associated with the ACCase mutation and P450 metabolism are consistent with the same mutations and mechanisms in other resistant L. rigidum populations.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

WAHRI is funded by the Australian Grains Research and Development Corporation (GRDC). An International Postgraduate Research Scholarship supported M. M. Vila-Aiub during this research at The University of Western Australia.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Bergelson, J. & Purrington, C.B. (1996) Surveying patterns in the cost of resistance in plants. American Naturalist, 148, 536558.
  • Burnet, M.W.M., Barr, A.R. & Powles, S.B. (1994) Chloroacetamide resistance in rigid ryegrass (Lolium rigidum). Weed Science, 42, 153157.
  • Burnet, M.W.M., Hart, Q., Holtum, J.A.M. & Powles, S.B. (1994) Resistance to nine herbicide classes in a population of rigid ryegrass (Lolium rigidum). Weed Science, 42, 369377.
  • Casal, J.J. & Sanchez, R.A. (1998) Phytochromes and seed germination. Seed Science Research, 8, 317329.
  • Christopher, J.T., Powles, S.B. & Holtum, J.A.M. (1992) Resistance to acetolactate synthase-inhibiting herbicides in annual ryegrass (Lolium rigidum) involves at least two mechanisms. Plant Physiology, 100, 19091913.
  • Christopher, J.T., Powles, S.B., Lijegren, D.R. & Holtum, J.A.M. (1991) Cross-resistance to herbicides in annual ryegrass (Lolium rigidum). II. Chlorsulfuron resistance involves a wheat-like detoxification system. Plant Physiology, 95, 10361043.
  • Christopher, J.T., Preston, C. & Powles, S.B. (1994) Malathion antagonizes metabolism-based chlorsulfuron resistance in Lolium rigidum. Pesticide Biochemistry and Physiology, 49, 172182.
  • Forcella, F., Benech-Arnold, R.L., Sanchez, R. & Ghersa, C.M. (2000) Modelling seedling emergence. Field Crops Research, 67, 123139.
  • Ghersa, C.M., Benech-Arnold, R.L. & Martinez-Ghersa, M.A. (1992) The role of fluctuating temperatures in germination and establishment of Sorghum halepense. Regulation of germination at increasing depths. Functional Ecology, 6, 460468.
  • Gill, G.S., Cousens, R.D. & Allan, M.R. (1996) Germination, growth, and development of herbicide resistant and susceptible populations of rigid ryegrass (Lolium rigidum). Weed Science, 44, 252256.
  • Gronwald, J.W. (1994) Resistance to photosystem II inhibiting herbicides. Herbicide Resistance in Plants. Biology and Biochemistry (eds S. B.Powles & J. A. M.Holtum), pp. 2760. Lewis Publishers, Boca Raton, FL.
  • Hall, L.M., Holtum, J.A.M. & Powles, S.B. (1994) Mechanisms responsible for cross resistance and multiple resistance. Herbicide Resistance in Plants. Biology and Biochemistry (eds S. B.Powles & J. A. M.Holtum), pp. 243261. Lewis Publishers, Boca Raton, FL.
  • Harper, J.L. (1956) The evolution of weeds in relation to resistance to herbicides. British Crop Protection Conference, 1, 179188.
  • Harper, J.L. (1977) Population Biology of Plants. Academic Press, London, UK.
  • Heap, I.M. (2004) International Survey of Herbicide Resistant Weeds. WeedSmart, Corvallis, OR. http://www.weedscience.com, accessed on 24/02/04.
  • Holtum, J.A.M., Häusler, R.E., Devine, M.D. & Powles, S.B. (1994) Recovery of transmembrane potentials in plants resistant to aryloxyphenoxypropanoate herbicides: a phenomenon awaiting explanation. Weed Science, 42, 293301.
  • McAlister, F.M., Holtum, J.A.M. & Powles, S.B. (1995) Dinitroaniline herbicide resistance in rigid ryegrass (Lolium rigidum). Weed Science, 43, 5562.
  • Poston, D.H., Wilson, H.P. & Hines, T.E. (2002) Growth and development of imidazolinone-resistant and -susceptible smooth pigweed biotypes. Weed Science, 50, 485493.
  • Powles, S.B. & Matthews, J.M. (1992) Multiple herbicide resistance in annual ryegrass (Lolium rigidum): a driving force for the adoption of integrated weed management. Resistance’91: Achievements and Developments in Combating Pesticide Resistance (eds I.Denholm, A. L.Devonshire & D. W.Hollomon), pp. 7587. Elsevier, London, UK.
  • Powles, S.B., Lorraine-Colwill, D.F., Dellow, J.J. & Preston, C. (1998) Evolved resistance to glyphosate in rigid ryegrass (Lolium rigidum) in Australia. Weed Science, 46, 604607.
  • Preston, C. & Mallory-Smith, C.A. (2001) Biochemical mechanisms, inheritance, and molecular genetics of herbicide resistance in weeds. Herbicide Resistance and World Grains (eds S.B.Powles & D.L.Shaner), pp. 2360. CRC Press, Boca Raton, FL.
  • Preston, C. & Powles, S.B. (1998) Amitrole inhibits diclofop metabolism and synergises diclofop-methyl in a diclofop-methyl-resistant biotype of Lolium rigidum. Pesticide Biochemistry and Physiology, 62, 179189.
  • Preston, C. & Powles, S.B. (2002a) Evolution of herbicide resistance in weeds: initial frequency of target site-based resistance to acetolactate synthase-inhibiting herbicides in Lolium rigidum. Heredity, 88, 813.
  • Preston, C. & Powles, S.B. (2002b) Mechanisms of multiple resistance in Lolium rigidum. Agrochemical Resistance. Extent, Mechanism, and Detection (eds J. M.Clark & I.Yamaguchi), pp. 150160. ACS Symposium Series 808. Oxford University Press, Washington, DC.
  • Preston, C., Tardif, F.J., Christopher, J.T. & Powles, S.B. (1996) Multiple resistance to dissimilar herbicide chemistries in a biotype of Lolium rigidum due to enhanced activity of several herbicide degrading enzymes. Pesticide Biochemistry and Physiology, 54, 123134.
  • Pritchard, H.W., Steadman, K.J., Nash, J.V. & Jones, C. (1999) Kinetics of dormancy release and the high temperature germination response in Aesculus hippocastanum seeds. Journal of Experimental Botany, 338, 15071514.
  • Purrington, C.B. & Bergelson, J. (1997) Fitness consequences of genetically engineered herbicide and antibiotic resistance in Arabidopsis thaliana. Genetics, 145, 807814.
  • Rashid, A., Newman, J.C., O'Donovan, J.T., Robinson, D., Maurice, D., Poisson, D. & Hall, L.M. (2003) Sulfonylurea herbicide resistance in Sonchus asper biotypes in Alberta, Canada. Weed Research, 43, 214220.
  • Roux, F., Gasquez, J. & Reboud, X. (2004) The dominance of the herbicide resistance cost in several Arabidopsis thaliana mutant lines. Genetics, 166, 449460.
  • Steadman, K.J. (2004) Dormancy release during hydrated storage in Lolium rigidum seeds is dependent on temperature, light quality and hydration status. Journal of Experimental Botany, 55, 929937.
  • Steadman, K.J., Bignell, G.P. & Ellery, A.J. (2003a) Field assessment of thermal after-ripening time for dormancy release prediction in Lolium rigidum seeds. Weed Research, 43, 458465.
  • Steadman, K.J., Crawford, A.D. & Gallagher, R.S. (2003b) Dormancy release in Lolium rigidum seeds is a function of thermal after-ripening time and seed water content. Functional Plant Biology, 30, 345352.
  • Tardif, F.J. & Powles, S.B. (1994) Herbicide multiple-resistance in a Lolium rigidum biotype is endowed by multiple mechanisms: isolation of a subset with resistant acetyl-CoA carboxylase. Physiologia Plantarum, 91, 488494.
  • Thompson, C.R., Thill, D.C. & Shafii, B. (1994) Growth and competitiveness of sulfonylurea-resistant and -susceptible Kochia (Kochia scoparia). Weed Science, 42, 172179.
  • Wiederholt, R.J. & Stoltenberg, D.E. (1996) Similar fitness large crabgrass (Digitaria sanguinalis) accessions resistant or susceptible to acetyl-coenzyme A carboxylase inhibitors. Weed Technology, 10, 4249.