3.1. Glyphosate Dose–Response Experiment
A dose–response experiment was carried out to confirm that population GR was resistant to glyphosate and to determine the resistance status of CC04. Glyphosate was unable to completely control GR across all rates applied (
Figure 1 and
Figure 2), thus confirming its status as resistant. Both CC04 and GS showed similar responses to glyphosate and were completely controlled at rates higher than 570 g a.e. ha
−1 in experiment 2 (
Figure 2), while there was low survival at rates of 1140 and 2280 g a.e. ha
−1 in experiment 1 (
Figure 1).
A three-parameter log-logistic model was used to determine the LD
50 values for the three
L. rigidum populations. The susceptible population GS displayed an LD
50 of 526 and 542 g a.e ha
−1 in Experiment 1 and 2, respectively (
Table 2), which was similar to the population of unknown resistance status (CC04) which had LD
50 values 0.6 and 1 times that of GS for experiment 1 and 2, respectively. The average rate of survival for CC04 was the same as GS at each glyphosate dose in experiment 2. The LD
50 for these two populations was higher than is often seen for susceptible populations of
L. rigidum [
5,
7,
10,
22], which could be due to environmental effects. Temperature is known to alter the effectiveness of glyphosate [
23,
24,
25] and higher temperatures have previously been shown to increase the resistance level of
L. rigidum [
25]. However, temperature data for this experiment were not compared with data from studies showing lower LD
50 values, so the effect of data on resistance in this study cannot be confirmed. The low rate of survival of populations GS and CC04 at 1140 and 2280 g a.e. ha
−1, as well as the greater LD
50 in comparison to other susceptible populations of
L. rigidum, may indicate that these populations are developing resistance in some members of the population. However, as population CC04 was collected from the roadside it is unlikely to have been regularly sprayed with glyphosate, and as such would have no selection pressure to evolve glyphosate resistance.
Population GR was confirmed by the dose–response experiment to be resistant with 6.1 and 4.4 times greater LD
50 values than GS in Experiment 1 and 2, respectively (
Table 2). GR also maintained 100% survival up to 570 g a.e. ha
−1 in experiment 1 and 84% survival at the same rate in Experiment 2. The LD
50 values of GR are higher than those seen in other studies for resistant
L. rigidum [
5,
7], which may also indicate that there was an increased level of resistance for all populations due to environmental factors. However, Yu et al. [
10] also showed a similar LD
50 in
L. rigidum as seen for GR in Experiment 2, and Collavo and Sattin [
22] have shown far greater levels of resistance in a
L. rigidum population.
As with the LD
50 values, the two susceptible populations (CC04 and GS) had similar GR
50 values to one another for both experiments (
Table 2). Glyphosate had less of a growth-reducing effect on both populations in Experiment 1 with GR
50 values of 406 and 506 g a.e. ha
−1 for population CC04 and GS, respectively, while GR
50 values in Experiment 2 were 276 and 288 g a.e. ha
−1, respectively. This indicates that while these two populations are not completely controlled at the standard glyphosate rate, any surviving populations will be stunted at rates of 285 g a.e. ha
−1 and above. The resistant population (GR) had the lowest growth reduction with a GR
50 of 2265 and 2413 g a.e. ha
−1 in experiment 1 and 2, respectively. As such, population GR was observed to be 4.5 and 8.4 times more resistant to glyphosate than GS in Experiment 1 and 2, respectively (
Table 2). Overall, dry weight decreased with increasing glyphosate dose for all populations. In Experiment 2, no plants survived for populations CC04 and GS at rates of 1140, 2280, and 4560 g a.e. ha
−1, and as such, there was complete inhibition of new growth. In Experiment 1, however, there was still low survival at rates of 1140 and 2280 g a.e. ha
−1, which allowed new growth in some plants.
These results show that rates of glyphosate up to 4560 g a.e. ha−1 are insufficient to control the summer population GR. As such, it suggests that alternative weed control options should be considered and demonstrates that GR-resistant L. rigidum could become an issue in summer cropping systems, such as cotton (Gossypium hirsutum L.). The results also show that a population initially thought to be susceptible can survive the application of the standard rate used in Australia.
3.2. EPSPS Sequencing
Mutations in the gene sequence of the target gene can cause TSR when the mutation changes the amino-acid sequence so that the enzyme’s function is not inhibited by the herbicide [
26]. Target-site resistance to glyphosate occurs when there is a mutation in the sequence of the
EPSPS gene [
27]. Mutations have been found in multiple codons of the
EPSPS gene in
L. rigidum but are mostly found in the Pro-106 codon [
5,
7,
10,
28,
29].
Sequence data for each population revealed no missense mutations in any of the populations. As such, no amino acid substitutions were observed in any population, suggesting that the resistance to glyphosate observed in population GR was likely due to NTSR mechanisms and further confirming the resistance status of populations CC04 and GS as susceptible to glyphosate. Silent mutations were observed in two codons for some of the plants analyzed (
Table 3), however, these mutations did not result in any amino acid substitutions and, therefore, are unlikely to confer any resistance. For populations GR and GS, the plants showed either GGC or GGG at codon 98, which both result in the amino acid glycine. For all populations, the plants showed either GCG or GCA at codon 103, which both result in the amino acid alanine. These results confirm the susceptibility to glyphosate observed in the dose–response experiment for populations CC04 and GS but do not identify the resistance mechanism that confers resistance in population GR. As such, further work is required to understand the mechanisms behind the resistance of GR to glyphosate.
Target-site resistance cannot be ruled out as a cause of resistance in this population as this type of resistance also includes changes to the gene copy number. An increase in the gene copy number of the
EPSPS gene has been previously linked with glyphosate resistance due to the increased number of EPSPS enzymes synthesized [
13,
30]. It should also be noted that target-site mutations may be present in other plants of the population. Further testing of a larger batch of plants may identify target-site mutations.
Due to the lack of target-site mutations in the resistant population (GR), further experiments are required to determine whether NTSR mechanisms have caused the resistance. NTSR due to altered glyphosate translocation has been previously reported in
L. rigidum [
31]. However, NTSR can also be due to the degradation of the herbicide by detoxification [
14,
32]. The resistance observed in GR could also be due to a combination of these mechanisms or altogether new mechanisms that have yet to be identified.
3.3. Alternative Herbicide Response
Eight herbicides were tested on all three populations to determine the effectiveness of alternative herbicides to glyphosate (
Table 4 and
Table 5). There was a significant interaction effect on both survival % and dry weight per plant between population and herbicide in both Experiment 1 (
p ≤ 0.001) and Experiment 2 (
p ≤ 0.001).
Paraquat was the most effective herbicide, controlling 100% of plants for all populations and in both experiments (
Table 4). As such, there was no new dry matter production for any of these plants. Paraquat is a fast-acting non-selective herbicide that diverts electrons from Photosystem I and leads to a chain of reactions that ruptures cell membranes [
33]. The complete control of all populations in this study suggests that this herbicide would be an effective substitute for glyphosate or could be used following glyphosate to kill any surviving weeds after the initial herbicide application. It could also be used in rotation with glyphosate. Paraquat resistance has previously been identified in
L. rigidum [
10,
34]. Busi and Powles [
35] also demonstrated the development of paraquat resistance when exposed to repeat selection under low doses of glyphosate. As such, there is the possibility for both the GR and GS
L. rigidum populations to evolve paraquat resistance. While resistance to both glyphosate and paraquat is rare, it has been previously observed [
10].
The enzyme acetyl co-enzyme A carboxylase (ACCase) is a critical component of the fatty acid biosynthesis pathway in plants [
36]. In this study, the effectiveness of five ACCase-inhibiting herbicides was tested on the three
L. rigidum populations (
Table 4 and
Table 5). Clethodim application controlled most plants across all three populations and over both experiments, with survival ranging from 0–8%. Corresponding to this was a large decline in dry weight per plant, ranging from 67–100% reductions in comparison to controls. Clethodim has been extensively used in Australia to control
L. rigidum plants in dicotyledonous crops, which has led to many populations developing resistance to clethodim [
37]. As such, while this is an effective alternative to glyphosate, it should be used cautiously in populations without clethodim resistance to prevent further cases of resistance.
Population GR was the most resistant of the three populations to the ACCase-inhibiting herbicides propaquizafop, haloxyfop, and pinoxaden, ranging from 28–51% survival. Additionally, these three herbicides exhibited less of a growth-reducing effect on the surviving plants of GR than the other populations (5–36% reduction compared to control for GR, 83–100% for CC04, and 32–100% for GS). Therefore, while these herbicides can control some plants of the GR population, they may not be a good option for this population. Multiple herbicide resistance to both propaquizafop and haloxyfop, in addition to glyphosate and paraquat resistance, has occurred in a population of
L. rigidum [
10]. As such, the use of these herbicides may lead to increased resistance in this population. Pinoxaden was less effective in experiment 1 compared to Experiment 2 for populations CC04 and GS, with survival of 36 and 20% in Experiment 1 for CC04 and GS, respectively, compared with 0 and 8% in Experiment 2 for CC04 and GS, respectively. Propaquizafop and haloxyfop completely controlled CC04 in both experiments, while GS possessed some survival (0 to 16%).
While these herbicides may not be highly effective for controlling the GR population, they could be a good option to use in rotation with glyphosate to slow the development of herbicide resistance in susceptible populations. These ACCase-inhibiting herbicides are recommended for controlling grasses in broadleaf crops and, as such, they should only be used in fallow fields for cereal crops. Pinoxaden, on the other hand, has been observed to control grass weeds in wheat (
Triticum aestivum L.) and barley (
Hordeum vulgare) [
38], although it may need to be supplied in a mixture with other modes of action herbicides [
39].
Both GR and GS had low survival under the acetolactate synthase (ALS), inhibiting herbicide combination imazamox + imazapyr, while CC04 had the highest survival in both experiments. Imazamox + imazapyr had a larger growth-reducing effect in Experiment 2 than in Experiment 1, with 81–100% growth reductions compared to the control in Experiment 2, while CC04 and GS ranged from 34–67% reduction in Experiment 1 and GR produced 14% greater dry weight per plant compared to the control. This study showed a range of effectiveness for imazamox + imazapyr, indicating differences due to population and environment. Environmental differences between years, such as temperature and humidity, may have caused the difference in control between experiments; however, we did not measure these parameters. As such, while there is potential for this herbicide to control L. rigidum, it may be ineffective in some circumstances.
When sprayed with pyroxsulam + halauxifen, an ALS-inhibiting herbicide and a synthetic auxin herbicide, respectively, most plants survived for all populations across both experiments (
Table 4). While survival ranged from 68–97% across populations, pyroxsulam + halauxifen had a moderate growth-reducing effect, where dry weight per plant was reduced by 33–68% (
Table 5). The last herbicide tested, glufosinate, also showed a low ability to control each population across both experiments, with a survival rate ranging from 76–96%. Despite high survival rates, there was a moderate to large growth reduction for each population (ranging from 59–75%), except for GS in experiment 1 (12% reduction). Pyroxsulam + halauxifen and glufosinate were not effective alternatives to glyphosate for controlling these populations of
L. rigidum, regardless of glyphosate resistance status. As some plants were killed and the surviving plants showed reductions in growth, these herbicides may reduce the number of seeds produced from infestations of
L. rigidum but cannot be relied upon to control this weed.
3.4. Alternative Herbicide Dose–Response Experiment
Three herbicides were selected following the alternative herbicide response experiment to further elucidate their ability to control the three populations. Propaquizafop was selected due to its ability to control the two glyphosate-susceptible populations, CC04 and GS, but not GR. Imazamox + imazapyr was chosen due to its reduced control of the glyphosate-susceptible population CC04 compared with GR. Glufosinate was chosen due to its low efficacy to control any of the three populations. A dose–response experiment was carried out for these three herbicides to further identify their effectiveness for controlling glyphosate-susceptible and resistant L. rigidum.
Propaquizafop was effective in controlling most plants above 15 g a.i. ha
−1 for CC04 and GS, with only a few plants surviving at higher concentrations (
Figure 3). However, some GR plants were able to survive at concentrations of up to 120 g a.i. ha
−1. Resistance to propaquizafop was previously observed by Yu et al. [
10] in
L. rigidum with higher rates of survival (>50%) at applications rates up to 200 g ha
−1. The low rate of survival in the GR population at 120 g a.i. ha
−1 may indicate that this population is also developing resistance to Propaquizafop. Imazamox + Imazapyr controlled most plants of each population at concentrations above 24.75 g a.i. ha
−1 and completely controlled all populations at the highest concentration of 99 g a.i. ha
−1; however, a low rate of survival was observed at 49.5 g a.i. ha
−1 (
Figure 4). While a low rate of plants did survive at these concentrations in Experiment 2, the biomass was reduced. This low rate of survival may be an indication of the populations developing resistance. Broster et al. [
40] identified populations of
L. rigidum with resistance to imazamox + imazapyr, as well as several populations developing resistance (survival rates between 10–19%) with plants sprayed with 48 g a.i. ha
−1. Glufosinate was only effective at rates of 1500 g a.i. ha
−1 and above in Experiment 1 but showed greater effectiveness on each population at lower concentrations in Experiment 2 (
Figure 5). Population CC04 was completely controlled at rates of 1500 g a.i. ha
−1 and above. All populations had greater survival at lower application rates in experiment 1, which may be due to temperature differences at the time of each experiment (higher temperatures in experiment 1 due to delayed sowing). Glufosinate has shown varying effectiveness under different temperatures in
Raphanus raphanistrum, although it was less effective under lower temperatures rather than higher temperatures [
41]. The survival of some
L. rigidum plants to 1500 g a.i. ha
−1 may indicate the presence of some glufosinate-resistant plants.
A three-parameter log-logistic model was used to determine the LD
50 values for CC04, GR, and GS. Populations CC04 and GS both had similar LD
50 values for the propaquizafop treatment, with values of 7.3 and 8.5 g a.i. ha
−1 for CC04 and 8.8 and 8.7 g a.i. ha
−1 for GS in Experiment 1 and 2, respectively (
Table 6). For the imazamox + imazapyr treatment, population GS had the lowest LD
50 values of 9.7 and 10.2 g a.i. ha
−1, with LD
50 values for CC04 increasing to 13.9 and 13.9 g a.i. ha
−1 for Experiment 1 and 2, respectively (
Table 7). Glufosinate was less effective in the first experiment, where populations GS, CC04, and GR had LD
50 values of 1665, 901, and 1273 g a.i. ha
−1, respectively, compared with 439, 366, and 569 g a.i. ha
−1, respectively, in Experiment 2 (
Table 8).
In addition to survival %, dry weight measurements were also recorded (
Figure 6,
Figure 7 and
Figure 8), and GR
50 values were calculated based on dry weight per plant (
Table 6,
Table 7 and
Table 8). Population CC04 had the lowest GR
50 value when treated with propaquizafop, at 7.6 and 8.9 g a.i. ha
−1 for Experiment 1 and 2, respectively. GR
50 values varied across experiments for GS and GR populations treated with propaquizafop, with values of 10.4 and 53.9 g a.i. ha
−1 for GS in Experiments 1 and 2, respectively, and 596.0 and 34.4 g a.i. ha
−1 for GR in Experiments 1 and 2, respectively. The effect of herbicide on biomass reduction also differed between experiments for plants treated with imazamox + imazapyr. Population CC04 was less sensitive to Imazamox + Imazapyr than GS and GR, with GR
50 values of 52.0 and 8.1 g a.i. ha
−1 for Experiment 1 and 2, respectively. GS and GR each had the same GR
50 value of 6.5 g a.i. ha
−1 in Experiment 2, although GS had a higher GR
50 value of 23.5 g a.i. ha
−1 in Experiment 1 compared with 15.2 g a.i. ha
−1 for population GR. Population CC04 was also less sensitive to glufosinate, in terms of the effect on biomass, than populations GS and GR, with GR
50 values of 686 and 610 g a.i. ha
−1 for Experiment 1 and 2, respectively. GR
50 values were lowest for population GS, at 324 and 425 g a.i. ha
−1 for Experiment 1 and 2, respectively. GR
50 values for population GR ranged from 460 to 600 g a.i. ha
−1 across Experiment 1 and 2, respectively.