Onion cells reveal combined toxic effects of atrazine and glyphosate

Two of the world’s most widely used herbicides—glyphosate and atrazine—can damage cells and DNA in a classic plant bioindicator even at concentrations that fall within current environmental guidelines, according to a new open-access study published in the journal Ecotoxicology. The research, led by Karyne Marriel Moreira and Tatiana da Silva Souza at the Federal University of Espírito Santo in Brazil, together with colleagues at the Federal University of Juiz de Fora, is among the most comprehensive assessments to date of how these agrochemicals behave when they occur together, as they routinely do in agricultural waterways.

Glyphosate, which kills weeds by blocking the synthesis of essential aromatic amino acids, is the single most heavily applied herbicide in Brazil and, increasingly, across the globe. Atrazine, a triazine compound that shuts down photosynthesis by inhibiting Photosystem II, ranks third among the country’s most commercialized agrochemicals and is a go-to option for controlling glyphosate-resistant weeds. Because the two are often tank-mixed and sprayed over the same fields, they frequently turn up together in rivers, streams, and reservoirs. Previous surveys cited in the study report glyphosate in Brazilian surface waters at concentrations ranging from below detection limits up to 500 micrograms per liter, with extreme values of 360 to 3,700 micrograms per liter recorded in agricultural streams. Atrazine was detected in nearly 11 percent of sampled Brazilian waters, peaking at 3.3 micrograms per liter. Elsewhere, the picture can be even more alarming: agricultural runoff in Nigeria has carried glyphosate at up to 25.2 milligrams per liter, while monitoring in Argentina documented a staggering 105,000 micrograms per liter.

To test what such exposure means for living organisms, the team turned to the common onion, Allium cepa, a plant long favored in environmental monitoring because its large chromosomes and rapidly dividing root-tip cells make cytotoxic and genotoxic damage easy to detect under the microscope. Onion seeds were germinated in Petri dishes containing solutions of glyphosate alone, atrazine alone, or one of five herbicide mixtures. The concentrations were chosen deliberately: glyphosate was tested at 62.5, 125, 250, 500, and 1,000 micrograms per liter, and atrazine at 0.25, 0.5, 1, 2, and 4 micrograms per liter—values anchored to the maximum levels permitted in Brazilian freshwater under CONAMA Resolution 357/2005 and in drinking water under Ministry of Health Ordinance 888/2021, with the highest doses simulating critical contamination scenarios. The five mixtures, labeled M1 through M5, paired each glyphosate level with its corresponding atrazine concentration.

Rather than relying solely on the traditional onion-root assay, the researchers layered on a suite of modern analytical tools, examining fourteen distinct endpoints across multiple levels of biological organization. Germination rate and root length served as macroscopic indicators of phytotoxicity. Cytogenetic analysis of Feulgen-stained root meristems—6,000 cells per treatment—quantified the mitotic index and the frequency of chromosomal abnormalities such as micronuclei, chromosomal breaks, C-metaphases, and anaphase bridges. Flow cytometry of propidium iodide–stained nuclei, run on a CytoFLEX instrument with 10,000 events recorded per sample, mapped the distribution of cells across the G₁, S, and G₂/M phases of the cell cycle and tracked a series of cell-death proxies: the sub-G₁ fraction, fluorescence intensity, forward and side scatter, and the coefficient of variation of the G₁ DNA peak. Finally, Evans Blue uptake measured plasma membrane integrity, while the reduction of 2,3,5-triphenyltetrazolium chloride (TTC) to red formazan by mitochondrial dehydrogenases served as a readout of respiratory activity and cellular viability.

The results paint a picture of toxicity that operates largely below the threshold of visible harm. Germination was significantly reduced by glyphosate at 125 and 250 micrograms per liter, by atrazine at 4 micrograms per liter, and by the most concentrated mixture, M5. Root length dropped only for glyphosate at 250 and 500 micrograms per liter. Beyond these macroscopic measures, however, the damage was widespread. Intermediate and high concentrations of both herbicides depressed the mitotic index, and every one of the five mixtures significantly increased the frequency of chromosomal abnormalities relative to the control. Micronuclei and chromosomal breaks were the dominant lesions, hallmarks of both clastogenic damage—direct DNA strand breaks—and aneugenic effects stemming from errors in chromosome segregation. Mixture M5 produced the highest frequency of abnormalities of any treatment in the study.

The flow cytometry data proved especially revealing. Several treatments shifted the cell-cycle distribution, consistent with checkpoint activation in response to DNA damage: when atrazine was present at just 0.25 micrograms per liter, cells accumulated in G₁, an arrest pattern that allows DNA repair enzymes time to act before replication proceeds. At higher concentrations and for the strongest mixtures, the pattern reversed—fractions of G₁, S, and G₂/M nuclei fell together while the sub-G₁ population expanded, indicating that damaged cells had abandoned repair and entered cell-death pathways. Across the board, fluorescence intensity of G₁ nuclei declined, forward and side scatter shrank, and the coefficient of variation of the G₁ peak rose, the latter serving as a sensitive indicator of genomic instability and uneven DNA distribution between daughter cells. M5 produced stronger cytometric disruptions than either herbicide alone.

Mitochondrial assays added a metabolic dimension to the toxicity profile. In most experimental groups, treated root cells failed to reduce TTC efficiently, revealing impairment of the mitochondrial respiratory chain even where the Evans Blue assay showed plasma membranes largely intact. This dissociation suggests that mitochondrial dysfunction is an early event, preceding membrane rupture, and is consistent with apoptosis-like cell death—matching the elevated sub-G₁ fractions and diminished nuclear integrity seen cytometrically. Reduced energy metabolism, the authors note, can compromise cell growth, mitosis, and DNA repair, creating a feedback loop that amplifies genetic damage.

Across the full matrix of fourteen endpoints and fifteen treatments, 113 of 210 endpoint–treatment combinations—53.8 percent—differed significantly from the untreated control. The pattern of sensitivity was strikingly uneven. Cytometric parameters, particularly G₁ fluorescence intensity and the G₁ coefficient of variation, along with the mitochondrial TTC assay, responded most consistently. Germination, root elongation, and membrane integrity proved the least sensitive. This hierarchy carries a practical warning: a water sample might pass a routine germination test yet still harbor compounds capable of quietly disrupting cell division and genome stability in exposed organisms.

Perhaps the most consequential finding concerns the mixtures themselves. Most combinations did not exceed the toxicity of the individual herbicides, suggesting that glyphosate and atrazine act through largely independent mechanisms or that cellular stress-response pathways become saturated, capping any additive effect. But mixture M5—whose glyphosate component of 1,000 micrograms per liter and atrazine component of 4 micrograms per liter fall within the range actually measured in agricultural surface waters—outperformed both single compounds on several biomarkers, including chromosomal abnormalities, the G₁ fraction, the sub-G₁ death signal, nuclear size, and genomic instability metrics. The authors caution that their design was not intended to formally model mixture interactions, but the results imply that heavily contaminated water bodies pose an elevated risk to non-target organisms, echoing earlier work by Bordin and colleagues, who observed heightened chromosomal damage when the two herbicides were combined.

The study also situates its findings within a growing body of evidence that these herbicides harm aquatic life at environmentally realistic doses. Prior work has shown atrazine inducing micronuclei, chromosomal bridges, and losses in onion cells at concentrations as low as 1 to 2 micrograms per liter, while glyphosate exposures between 0.1 and 1,000 micrograms per liter have produced cytogenetic damage in the same bioindicator. In fish, environmentally relevant glyphosate concentrations have been linked to DNA strand breaks, micronucleus formation in erythrocytes, oxidative imbalance, and immune suppression.

Crucially, the authors emphasize that genotoxic substances are considered to lack a safe threshold of exposure. The fact that most mixture effects were “only” equivalent to those of the individual herbicides offers no reassurance; the damage was still there, occurring at concentrations regulators currently deem acceptable. The team argues that their integrated approach—wedding the classical onion bioassay to flow cytometry, membrane, and mitochondrial biomarkers—detects subtle cellular insults long before plants show visible symptoms, and they call for such sensitive endpoints to be incorporated into routine ecotoxicological assessment. As tank-mixing of herbicides becomes ever more routine in modern agriculture, the water draining from the world’s fields may be quietly taxing the genomes of the organisms living downstream, one dividing cell at a time.