Quantitative estimation of fitness cost associated with glyphosate resistance in Echinochloa colona

A population of Echinochloa colona from the northern region of Western Australia
evolved glyphosate resistance after 10 years of glyphosate selection and the expression
of associated fitness costs in this population was studied. When assessing fitness costs
associated with resistance, it is crucial to ensure that glyphosate-susceptible (S) and -
resistant (R) individuals share a similar genetic background, except for the resistance
alleles. A plant cloning technique followed by phenotypic identification and glyphosate
selection was the experimental approach employed to identify S and R phenotypes
within the segregating glyphosate-resistant E. colona population. For selection of R
plants, seedlings were treated with 2,160 g ha^-1 of glyphosate (Roundup 540 g L^-1) at the
two- to three-leaf stage and the survivors were allowed to produce seeds. For selection
of S plants, 300 g ha^-1 of glyphosate (Roundup 540 g L^-1) was applied and the clones of
the killed plants were used for seed production. Plants from these seeds served as the
materials for whole-plant dose response and resistance mechanism studies as well as
evaluation of fitness costs associated with glyphosate resistance in this E. colona
population. Estimations of LD₅₀ (dose causing 50% mortality) and GR₅₀ (dose causing
50% reduction in growth rate) showed an eightfold glyphosate resistance in this
population. More shikimate accumulated in leaves of the S phenotype than in the R
phenotype after glyphosate treatment. Glyphosate resistance was not found to be due to
any resistance endowing target site EPSPS gene mutation. It is also unlikely that the
resistance mechanism is due to amplified EPSPS gene expression or enhanced
glyphosate metabolism as no differences were found between the S and R phenotypes.
Similarly, there were no differences in leaf glyphosate uptake and translocation at the
whole plant level between the S and R phenotypes. However, there is evidence
suggesting a reduced cellular glyphosate uptake in the R phenotype.

The glyphosate resistance mechanism in this R phenotype of E. colona remains to be elucidated and
non-target-site based resistance mechanisms (such as vacuole sequestration and/or
membrane uptake) should be explored. Seeds of the S and R phenotypes kept at warmer
temperatures (after-ripening dry storage at 15-35C) or on the ground surface outdoors,
germinated better than those seeds kept for the same period (14 days) at lower
temperature (8C). Light and longer dry after-ripening times increased the rate of seed
dormancy release and germination of both S and R phenotypes. Equal decline in
seedling emergence with increasing soil burial depth was observed in both S and R
phenotypes. Thus, the S and R phenotypes exhibited similar characteristics of seed
dormancy release and germination, and seedling emergence. Under non-competitive
conditions, plants from the S phenotype showed greater leaf area, vegetative
aboveground biomass and root biomass as well as higher relative growth rate compared
to plants of the R phenotype during the vegetative growth stage. At the reproductive
stage, plants of the S phenotype also produced more reproductive biomass than plants
from the R phenotype. A physiological fitness cost of 31 (seed mass) or 22 (seed
number) was associated with the R phenotype. The results clearly suggest there is an
evolutionary trade-off between glyphosate resistance and vegetative and reproductive
growth in resistant E. colona plants when grown in glyphosate-free environments.
Plants of the S phenotype were also stronger competitors with a rice crop than plants of
the R phenotype, and allocated more resources to reproductive organs than the R
phenotype. An ecological fitness cost of 60 was expressed in the R phenotype under
interspecific (vs rice) competitive conditions. Similarly, a greater reduction in seed mass
and number was exhibited by R plants when competing with S plants. In addition, a
significant decline of 14% in the phenotypic frequency of glyphosate resistance was
estimated after two generations in field conditions under no glyphosate selection.
Overall, results from the present investigation demonstrate that there is a significant
fitness cost associated with glyphosate resistance in the studied E. colona population. It
is possible to speculate that under no further glyphosate selection, the frequency of
glyphosate resistance is likely to remain low or even be eliminated in agroecosystems.