Herbicide resistant weeds are among the greatest threats facing modern agriculture. Glyphosate, the active ingredient in the commercial herbicide RoundUp, is the most widely applied herbicide worldwide, and 43 weed species have evolved resistance to glyphosate to date. My research addressed four key questions about the strength, mechanisms, and fitness effects of glyphosate resistance using Conyza canadensis, a representative glyphosate-resistant weed, and Arabidopsis thaliana, a model plant species. In northcentral Ohio and southern Iowa, maternal biotypes (individual plants) of Conyza canadensis, also known as horseweed or marestail, were collected from 74 locations (both agricultural and non-agricultural) in both states. Dose-response experiments were used to categorize biotypes into resistance categories based on 80% survival at 0x only (Susceptible, S), and up to 1x (equivalent to 840 g ae ha-1; R1), 8x (R2), 20x (R3), and 40x (R4). Extreme glyphosate resistance (R4 biotypes) was quite common in both states and both habitats, and non-agricultural habitats served as a refuge for R4 biotypes.
Glyphosate resistance mechanisms are generally presumed to impose a fitness cost due to possible trade-offs in resource allocation to plant defense, growth, and reproduction. Nine S, eight R1, and nine R4 biotypes originally collected in Iowa were grown in a common garden experiment in Iowa over two years and two sites to test for fitness effects. Based on early rosette sizes and days to bolting, nested ANOVAs showed that R4 biotypes grew as large as, if not larger than, both S and R1 biotypes, and bolted significantly earlier. Furthermore, R1 and R4 biotypes were less likely to display disease symptoms than S biotypes. Glyphosate resistance in horseweed appears not to impose an early growth penalty, and possibly no lifetime fitness cost.
Specific mechanisms of glyphosate resistance in horseweed include altered translocation and vacuolar sequestration. Recently, a Canadian study documented the first report of a target-site point mutation of a proline to serine substitution at position 106 of the EPSPS2 gene in horseweed. 24 Ohio biotypes and 20 Iowa biotypes were screened for this Pro-106-Ser mutation. All 19 biotypes scored as S, R1, or R2 lacked the Pro-106-Ser mutation, and all 25 biotypes scored as R3 or R4 possessed the Pro-106-Ser mutation. The association between the point mutation and extremely resistant horseweed biotypes is noteworthy.
Model species, such as Arabidopsis thaliana, may be useful for understanding the fitness effects of other glyphosate resistance mechanisms such as the over-production of 5-enolpyrvulshikimate-3-phospahte synthase (EPSPS), which has been documented in 13 weed species. Transgenic Arabidopsis lines that overexpress (OX) a native EPSPS gene were created, as were empty vector lines (EV). Six OX lines and seven EV lines were grown in greenhouse fitness experiments in the absence of glyphosate. Two of the OX lines produced significantly more seeds per plant compared to the wild-type and EV lines, and in no case was a fitness cost detected. Taken together, my studies show that glyphosate resistance can arise via several mechanisms and even high levels of resistance do not appear to incur a fitness cost.