Amaranthus tuberculatus (Mq. ex DC) J.D. Sauer: potential for selection of glyphosate resistance
by Ian A. Zelaya and Micheal D. K. Owen

blueline.jpg (1822 bytes)
September 30, 2002This paper was presented at the 13th Australian Weeds Conference held in Perth, Australia  on September 8-13.  The paper is published in the Papers and Proceedings  on pages 630-633.

Summary  Reports alleging inconsistent common waterhemp (Amaranthus tuberculatus (Mq. ex DC) J.D. Sauer) control indicated that the species demonstrated an inherent variability to glyphosate.  Correlation of tissue shikimic acid and phenotype in the field supported the tenet that the response was attributable, at least in part, to differences in glyphosate inhibition of 3-phosphoshikimate 1-carboxyvinyltransferase (EPSPS; EC
    Whole plant dose responses of the Everly, Iowa A. tuberculatus and a pristine population from Paint Creek, Ohio indicated that the Everly biotype demonstrated more variability to glyphosate than the unselected population.  Isolation of resistant and susceptible plants through recurrent selection resulted in a 1.7 and 3.5 fold increase in population divergence in the first (F1) and second (F2) filial generations, respectively.  While the selection method has increased the frequency of resistant individuals within the population, significant segregation for glyphosate efficacy was still apparent in the selected material. This limited segregation suggested that the response to glyphosate observed in A. tuberculatus may be governed by a polygenic event.  At present we are attempting to reduce the genetic variability in A. tuberculatus and investigate possible resistance mechanisms in asexually propagated plantlets. Characterization of the mechanism(s) of glyphosate resistance may be important in developing strategies to mitigate potential future problems.

Keywords  asexual reproduction, recurrent selection, log-logistic analysis, shikimic acid, resistance.


    Evolution of glyphosate resistance in plants is uncommon, despite of the prolonged and ubiquitous use of the herbicide worldwide (Caseley and Copping 2000). The rareness of this event was attributed to the limited metabolism of glyphosate in plants, the short-half life in the environment and the unique biochemical characteristics of the herbicide, and the fact that engineering glyphosate resistance in crops required of complex molecular modifications (Bradshaw et al. 1997). Nonetheless, inter- and intra-specific variability to glyphosate may enable the selection of preexistent individuals within a population that demonstrate an improved fitness to glyphosate and result in weed population shifts (Baylis 2000).
    In higher organisms, glyphosate resistance was engineered by amplification (Goldsbrough et al. 1990), enhanced transcription (Klee et al. 1987), or increased half-life of EPSPS (Holländer-Czytko et al. 1992) and by point mutations in EPSPS associated with an increased apparent dissociation constant (Ki) for glyphosate (Kiglyphosate) (Padgette et al. 1991). An alternative resistance mechanism comprises glyphosate decarboxylation to
a-aminomethylphosphonic acid (AMPA), a less-toxic compound to plants. The enzyme glyphosate oxido-reductase (GOX) mediates metabolism to AMPA by oxidation of the N-Ca bond in glyphosate (Barry et al. 1992). Concomitant expression of CP4-EPSPS and GOX instigated high resistance levels in glyphosate-resistant crops (Mannerlöf et al. 1997).
    Recently, glyphosate resistance has been reported in Italian ryegrass (Lolium multiflorum Lam.) (Perez and Kogan 2002), rigid ryegrass (Lolium rigidum Gaudin) (Pratley et al. 1999), goosegrass (Eleusine indica (L.) Gaertner) (Lee and Ngim 2000), and horseweed (Conyza canadensis (L.) Cronq) (Van Gessel 2001).  To date no cases of glyphosate resistant weed populations have been reported in the Mid West US.  However, some Iowa producers have reported inconsistent control of A. tuberculatus with glyphosate in glyphosate-resistant crops.  Two isolated incidents in Everly and Badger, Iowa suggested that A. tuberculatus plants were differentially affected by glyphosate.  Plant samples were collected from Everly, Iowa to conduct an exhaustive assessment of the incident and evaluate the potential for glyphosate resistance.


Evaluation of glyphosate efficacy  Ninety-seven plants (53 ♀: 44 ♂) collected from the Everly field were grown to maturity and crossed without pollination restrictions. Glyphosate efficacy was evaluated in the progeny at the whole-plant and seedling levels.  Seedling assessment comprised germinating 20 seeds per well in a 24 well cell culture cluster plate in 32 mM, 10 mM, 3.2 mM, 1 mM, 0.32 mM, 0.1 mM, and 0.032 mM glyphosate.  Seeds were grown at 30 C and 14 hrs light and 20 C 10 hrs dark conditions.  Seedling germination, hypocotyl, and radicle length were recorded two weeks after establishment.
    Whole plant assessment comprised three transplanted seedlings per 12 cm diameter pot in a peat:perlite:loam soil-mix (1:2:1) which were placed under natural light, and supplemented with 16 hrs of 600-1000
mmol-2 m s-1 PPFD.  When 10-12 cm, plants were arranged in a complete randomized block design with three replications and sprayed with 0, 0.21, 0.52, 0.83, 1.25, and 1.37 kg ae ha-1 glyphosate.  Plants were harvested two weeks after application and visual estimate of herbicide and biomass recorded.  Plant biomass was dried at 35 C for 48 hrs; dry apices of treated plants were pooled for shikimic acid determination.

Recurrent selection  A recurrent selection process was implemented to isolate resistant and susceptible individuals within the Everly population.  This selection was conducted at the seedling level to expedite the recurrent selection process and deal with the prolific seed productivity of A. tuberculatus.  Each generation was selected for susceptibility and resistance according to parameters determined by the seedling dose response (Table 1). Resistant individuals were considered those surviving the selection media, while susceptible seedlings comprised those demonstrating herbicide injury at low herbicide doses.  Selected seedlings were rescued from the media by rinsing in distilled water, treating with Rootone (TechPac, Lexington, KY 40504), and transplanted in the potting soil-mix.  Mature plants were intercrossed in plastic tents where pollen contamination was less than 0.01% (Brenner and Widrlechner 1998). Because all 12 plants surviving 10 mM glyphosate differentiated to male plants (Table 1), F1-R individuals were backcrossed to female plants previously selected at 3.2 mM glyphosate.


Table 1. Dose parameters for the recurrent selection, frequency of resistant (R) and susceptible (S) individuals and sex ratios of selected A. tuberculatus populations.




sex ratio


3.2 mM


7 ♀: 6 ♂


0.032 mM


24 ♀: 29 ♂


10 mM


0 ♀: 12 ♂


0.01 mM


16 ♀: 21 ♂


8 mM


7 ♀: 11 ♂


a = percentage relative to population
b = parental population
= not estimated

Asexual reproduction  Perpetuation of specific genotypes was conduced asexually by cutting the main stem of plants, treating with Rootone, and providing 95% relative humidity and 16 hrs 1100 mmol-2 m s-1 PPFD.  This process not only preserved and facilitated seed increase, but also instigated the re-differentiation of floral meristems back to vegetative development.

Shikimic acid assay  Determination of shikimic acid was conducted spectrophotometrically with procedures modified from Cromartie and Polge (2000).  Ten ml of the supernatant extract were mixed with 0.5% periodic acid, 0.5% sodium meta-periodate (w/w), incubated at 37 C for 30-45 min, and quenched with 1 M NaOH:0.056 M Na2SO3 (3:2 v/v).  Finally, absorbance was detected at 380 nm.  Standard curves were constructed with shikimic acid (Fisher Scientific International) at a range of 1 to 10,000 mM.

Statistical analysis  Plant biomass and shikimic acid data was subjected to analysis of variance, with mean separation and correlation coefficients determined by Fisher’s least significant difference and Spearman correlation analysis, respectively (SAS 1996).  In addition, biomass data was used to calculate GR50, standard error, and 95% upper and lower confidence interval limits according to the log-logistic analysis (Seefeldt et al. 1995).


A. tuberculatus inherent variably to glyphosate   Most reported cases of inconsistent A. tuberculatus control are attributed to application problems, sub-lethal dosage, or delayed glyphosate applications with respect to plants phenology.  Nevertheless, our visual assessment of the Badger, Iowa A. tuberculatus population suggested that A. tuberculatus plants were differentially affected by glyphosate.  Adjacent plants within the crop row that received a comparable glyphosate dose, displayed significantly different phenotypes.  These differences were consistent with shikimic acid determinations made from the apices of putative susceptible and resistant phenotypes and from untreated A. tuberculatus plants.  Putative susceptible plants consistently accumulated at least five folds more shikimic acid than the resistant phenotype (Figure 1). In addition, accumulation of shikimic acid in untreated A. tuberculatus plants was undetectable. While patterns of shikimic acid do not provide conclusive evidence with respect to the mechanism of resistance, the data supports the tenet that EPSPS in these plants was differentially inhibited by glyphosate.
    Evaluations of A. tuberculatus seed from the Everly, Iowa population and other populations from agricultural environments suggested that the species was inherently variable to glyphosate, and thus may provide the genetic basis for the evolution of individuals with increased fitness to glyphosate.  Genetic diversity is a common trait in Amaranthaceae (Chan and Sun 1997).  Because A. tuberculatus is an obligate outcrosser, genetic recombination during mitosis could explain the diversity observed in the species.


Figure 1. Shikimic acid accumulation in apex tissue of A. tuberculatus plants demonstrating different levels of field injury, two weeks after glyphosate application.  Samples were segregated in susceptible (100 %, 60-90%, and 20-50%) and resistant (0-5%) and a control (not sprayed). Data represents the mean of six plants.

    The recurrent selection increased the frequency of resistant individuals within the Everly, Iowa population; this corresponded to a 3.5 fold resistance increase after two generations of selection (Table 2).  With the exception of backcross of F1-R and F2-R (F2B-R), the selection procedure also reduced the variability to glyphosate as estimated by the standard error of the log-logistic analysis.  Furthermore, the variability of selected material was comparable to that of the pristine Ohio population (Table 2).  Regardless of attempts to isolate stable-homogeneous resistant and susceptible populations, significant segregation for glyphosate efficacy was still apparent after two generations of recurrent selection. Variable responses of Amaranthaceae to glyphosate have been documented (Baylis 2000).  Reports indicate that individual A. tuberculatus plants can tolerate field rates of at least 6.72 kg of glyphosate per ha (Smeda 2000).

Potential for selection of glyphosate resistance in A. tuberculatus  The lack of segregations during recurrent selection suggests that the variability to glyphosate in A. tuberculatus may be attributed to a polygenic character of the trait.  Thus, recombination of resistant and susceptible alleles during mitosis would result in progeny with different allele combinations and may explain the phenotype observed at the whole plant level.  Alternatively, variable response to glyphosate could be explained by the differential expression of EPSPS polymorphs. At least two EPSPS isoforms with similar structural and kinetic properties exist in maize (Forlani et al. 1994).  In addition, the expression of EPSPS in plants was tissue-specific and developmentally regulated (Benfey et al. 1990).  To date, no clear resistance mechanism has been elucidated from the reported cases of resistant weeds.  However, studies in field bindweed (Convolvulus arvensis L.) suggest that cellular and metabolic processes may concomitantly act to determine glyphosate tolerance in plants (Westwood et al. 1997).

Table 2. Dose-response parameters for the pristine, parental, and selected A. tuberculatus populations.




95% CIc





























a = kg ae/ha that reduced 50% biomass accumulation
= standard error of the log-logistic analysis
= 95% upper and lower confidence intervals (kg ae/ha)
= resistance to susceptibility ratio
= not applicable as material was unselected

    Our current data suggests that the potential for the evolution of glyphosate resistance in A. tuberculatus is considerable.  Nevertheless, several cycles of selection may be required to isolate resistant individuals.  Reported cases of glyphosate resistance in L. multiflorum, L. rigidum, E. indica, and C. canadensis required of several years of persistent glyphosate selection (Pratley et al. 1999, Lee and Ngim 2000, Van Gessel 2001).  This requirement for prolonged selections may be attributed to a low frequency of resistant individuals within the population or to a physiological penalty associated with the resistance trait.  Point mutations in the binding domain of EPSPS may result in a kinetically less efficient enzyme (Padgette et al. 1991).
    While no cases of glyphosate resistant weeds have been reported in the Mid West US, increases in the area planted with glyphosate-resistant crops may enhance the selection pressure for glyphosate and instigate weed population shifts.  Farmers typically rely on multiple glyphosate applications to meet weed control expectations.  Thus frequent and ubiquitous use of glyphosate in field crops will likely provide enough selection pressure for the evolution of glyphosate resistant weed populations. 

Current work

Evaluations of mechanisms of resistance  The recurrent selection of A. tuberculatus was based on the seedling assay.  This approach elucidated resistant and susceptible individuals, however the method has potential for misidentification of phenotypes.  Thus, a three level selection approach will be conducted to isolate plant material suitable for translocation, metabolism, and EPSPS DNA sequence analysis.  Plants will be characterized by their response to glyphosate at the seedling and whole-plant level and their shikimic acid accumulation patterns.  The mechanisms of resistance will be investigated in asexually-propagated resistant, susceptible, and pristine A. tuberculatus plantlets.


This investigation was possible by the economic support of Monsanto and Syngenta companies.  The authors wish to thank Dr. Jonathan Gressel for his comments and recommendations to this investigation.  


Barry, G., Kishore, G., Padgette, S., Taylor, M., Kolacz, K., Weldon, M., Re, D., Eichholtz, D., Fincher, K. and Hallas, L. (1992). Inhibitors of amino acid biosynthesis: strategies for impairing glyphosate tolerance to crop plants. In 'Biosynthesis and Molecular Regulation of Amino Acids in Plants,' eds. Singh, B.K., Flores, H.E. and Shannon, J.C., pp. 139-45. (American Society of Plant Physiologists, Vol. 7, Rockville, MD).

Baylis, A.D. (2000). Why glyphosate is a global herbicide: strengths, weaknesses and prospects. Pest Management Science 56,299-308.

Benfey, P.N., Takatsuji, H., Ren, L., Shah, D.M. and Chua, N.H. (1990). Sequence requirements of the 5-enolpyruvylskikimate-3-phosphate synthase 5'-upstream region for tissue-specific expression in flowers and seedlings. Plant Cell 2,849-56.

Bradshaw, L.D., Padgette, S.R., Kimball, S.L. and Wells, B.H. (1997). Perspectives on glyphosate resistance. Weed Technology 11,189-98.

Brenner, D.M. and Widrlechner, M.P. (1998). Amaranthus seed regeneration in plastic tents in greenhouses. Plant Genetic Resources Newsletter 116,1-4.

Caseley, J. and Copping, L. (2000). Twenty-five years of increasing glyphosate use: the opportunities ahead. Pest Management Science 56,297.

Chan, K.F. and Sun, M. (1997). Genetic diversity and relationships detected by isozyme and RAPD analysis of crop and wild species of Amaranthus. Theoretical and Applied Genetics 95,865-73.

Cromartie, T.H. and Polge, N.D. (2000). An improved assay for shikimic acid and its use as monitor for the activity of sulfosate. WSSA Abstract 40,121.

Forlani, G., Parisi, B. and Nielsen, E. (1994). 5-enol-pyruvyl-shikimate-3-phosphate synthase from Zea mays cultured cells: purification and properties. Plant Physiology 105,1107-14.

Goldsbrough, P.B., Hatch, E.M., Huang, B., Kosinsky, W.G., Dyer, W.E., Herrmann, K.M. and Weller, S.C. (1990). Gene amplification in glyphosate tolerant tobacco cells. Plant Science 72,53-62.

Holländer-Czytko, H., Sommer, I. and Amrheim, N. (1992). Glyphosate tolerance of cultured Corydalis sempervirens cell is acquired by an increased rate of transcription of 5-enolpyruvylshikimate 3-phosphate synthase as well as by a reduced turnover of the enzyme. Plant Molecular Biology 20,1029-36.

Klee, H.J., Muskopf, Y.M. and Gasser, C.S. (1987). Cloning of an Arabidopsis thaliana gene encoding 5-enolpyruvylshikimate-3-phosphate synthase: sequence analysis and manipulation to obtain glyphosate-tolerant plants. Molecular and General Genetics 210,437-42.

Lee, L.J. and Ngim, J. (2000). A first report of glyphosate-resistant goosegrass (Eleusine indica (L) Gaertn) in Malaysia. Pest Management Science 56,336-9.

Mannerlöf, M., Tuvesson, S., Steen, P. and Tenning, P. (1997). Transgenic sugar beet tolerant to glyphosate. Euphytica 94,83-91.

Padgette, S.R., Re, D.B., Gasser, C.S., Eichholtz, D.A., Frazier, R.B., Hironaka, C.M., Levine, E.B., Shah, D.M., Fraley, R.T. and Kishore, G.M. (1991). Site-directed mutagenesis of a conserved region of the 5-enolpyruvylshikimate-3-phosphate synthase active site. The Journal of Biological Chemistry 266,22364-9.

Perez, A. and Kogan, M. (2002). Glyphosate-resistant Italian ryegrass (Lolium multiflorum) in Chile. WSSA Abstract 42,21.

Pratley, J., Urwin, N., Stanton, R., Baines, P., Broster, J., Cullis, K., Schafer, D., Bohn, J. and Krueger, R. (1999). Resistance to glyphosate in Lolium rigidum. I. Bioevaluation. Weed Science 47,405-11.

SAS, Institute, The SAS system for windows. SAS Institute, Cary, NC, (1996).

Seefeldt, S.S., Jensen, J.E. and Fuerst, E.P. (1995). Log-logistic analysis of herbicide dose-response relationships. Weed Technology 9,218-27.

Smeda, R.J. (2000). Insensitivity of a common waterhemp population to glyphosate. NCWSS Proceedings 55,90.

VanGessel, M.J. (2001). Glyphosate-resistant horseweed from Delaware. Weed Science 49,703-5.

Westwood, J.H., Yerkes, C.N., DeGennaro, F.P. and Weller, S.C. (1997). Absorption and translocation of glyphosate in tolerant and susceptible biotypes of field bindweed (Convolvulus arvensis). Weed Science 45,658-63. 

 Prepared by Micheal D. K. Owen, extension weed management specialist, Department of Agronomy, Iowa State University

For more information contact:
ISU Extension Agronomy
2104 Agronomy Hall
Ames, Iowa 50011-1010
Voice: (515) 294-1923
Fax: (515) 294-9985
This site designed and managed by Brent Pringnitz.
Submit questions or comments here.  

Copyright © 1996-2003, Iowa State University, all rights reserved  

Common chemical and trade names are used in this publication. The use of trade names is for clarity by the reader. Inclusion of a trade name does not imply endorsement of that particular brand of herbicide and exclusion does not imply nonapproval.