Atrazine

Metribuzin and cyanazine were not detected in any samples. Bromoxynil and fenoxyprop ethyl were detected in < 5% of dust samples and were excluded from analyses. We grouped the remaining herbicides two ways for our analyses. First, we summed the concentrations of the six herbicides used almost exclusively in agriculture (acetochlor, alachlor, atrazine, bentazone, fluazifop-p-butyl, metolachlor; called here agricultural herbicides). Second, we evaluated the four detected herbicides that were used on ≥15% acres of corn and/or soybeans in Iowa in all three pesticide-use reporting years, 1985, 1990, and 1995 (atrazine, dicamba, metolachlor, trifluralin; called long-term–use herbicides). The herbicides accounting for the highest treated acreage of corn and soybeans, respectively, were atrazine (67% of corn acres treated in 1995) and trifluralin (30% of soybean acres treated in 1995). We also evaluated the concentration of individual herbicides that were detected in at least 5% of the samples (Table 1).
The descriptive statistics of the herbicide concentrations were based on the observed concentrations and the concentrations from one imputation, whereas the percent detections did not include any imputed values over the detection limit. We calculated percent detections and concentrations of the agricultural herbicide group separately for homes with and without an agricultural worker and by the location of residences within and outside of towns.
For each home, we determined the acreage of corn and soybeans within 750 m of the residence for each crop map year (1998–2000). We chose 750 m because primary pesticide drift from ground and aerial spraying—the most common methods of application of herbicides to corn and soybeans in Iowa—occurs within this distance (AgDRIFT Task Force 2002; Woods et al. 2001 (link)). We also determined the acres of corn and soybeans within “donut-shaped” buffer zones of < 100, 101–250, 251–500, and 501–750 m from homes. The acres of each crop type in the buffer zones changed little over the 3 years; therefore, we averaged the acreage across the 3 years. Because the average acreage of corn and soybean fields within each buffer zone was highly correlated [r ranged from 0.78 (100 m) to 0.90 (501–750 m)], we evaluated the summed acres of corn and soybeans in all analyses.
We conducted two main types of analyses. First, we used logistic regression to compute odds ratios (ORs) and 95% confidence intervals (CIs) of detecting one or more herbicides in each herbicide group in relation to the crop acreage anywhere within 750 m of the home and in relation to acres of crops within each buffer zone. Second, we used linear regression to model the logarithm of the concentration in relation to crop acreage anywhere within 750 m of the home and in relation to acres within each buffer zone. Linear regression models were run for each of the five data sets that contained the measured and imputed concentrations. Final parameter estimates and CIs were determined using the SAS procedure MIANALYZE (version 8.02; SAS Institute Inc., Cary, NC, USA). We computed likelihood ratio tests comparing the significance of nested models using models based on one imputed value. All analyses were evaluated for confounding by the presence of a current or past farmer in the home and whether the home was inside or outside a town. The ORs for herbicide detections were adjusted for agricultural employment because adjustment resulted in a change of 10% or more.

A model expanding on previous^{5 (link)} and classic^{29 ,30 (link)} work was used to investigate agrochemical effects—of atrazine, chlorpyrifos, and fertilizer—on human schistosomiasis transmission intensity. The model includes snail population dynamics of Bulinus truncatus, the intermediate host of Schistosoma haematobium, subject to logistic population growth and the influence of predation. We focus on S. haematobium for our modeling because it is the predominant schistosome species found in the village in Senegal where the epidemiological data used to parameterize our model were collected. The population dynamics of generalist predators (P) are included, subject to an agrochemical-sensitive mortality rate, μ_P,q, that reflects chlorpyrifos toxicity to the predator population as estimated by the mesocosm experiments and previous work^{35 (link)}. Because crayfish and prawns are generalist predators and will switch to other resources when snail densities are low, the model assumes that predator population dynamics are independent of snail density. Also included is a parameter representing agrochemical enhancement of the snail carrying capacity, φ_N,q, which models the snail population response to bottom-up effects caused by algal stimulation by atrazine and fertilizer as estimated in the mesocosm experiments and other experiments examining the same agrochemicals and outcomes. Additional model state variables represent susceptible, exposed and infected snails (S, E, and I, respectively) and the mean worm burden in the human population (W). The number of mated female worms, M, is estimated assuming a 1:1 sex ratio and mating function, γ(W, k), as in ref.^{53 (link)}. The per capita snail predation rate by predators, modeled as a Holling type III functional response as in ref.^{54 (link)}, ψ, and the total snail population, N, are shown separately for clarity. Parameter values, definitions and reference literature are listed in Supplementary Table 9. $$\frac{\mathrm{d}S}{\mathrm{d}t}=f_N\left(1-\frac{N}{{\phi }_N*{\phi }_{N,q}}\right)\left(S+E\right)-{\mu }_{N}S-{P\psi S}^n-\beta MS$$ $$\frac{\mathrm{d}E}{\mathrm{d}t}=\beta MS-{\mu }_{N}E-{P\psi E}^n-\sigma E$$ $$\frac{\mathrm{d}I}{\mathrm{d}t}=\sigma E-\left({\mu }_N+{\mu }_I\right)I-{P\psi I}^n$$ $$\frac{\mathrm{d}W}{\mathrm{d}t}=\lambda I-\left({\mu }_H+{\mu }_W\right)W$$ $$\frac{\mathrm{d}P}{\mathrm{d}t}=f_P\left(1-\frac{P}{{\phi }_P}\right)\left(P\right)-\left({\mu }_P+{\mu }_{P,q}\right)P$$ $$M=0.5WH\gamma \left(W,k\right)$$ $$\psi =\frac{\alpha }{1+{\alpha T}_h{N}^n}$$ $$N=S+E+I$$

For identification of histone adducts with atrazine, the acquired UHPLC–Q-Exactive-Orbitrap-MS raw data files were converted to MGF files using Raw Converter software. LC–MS data was then preprocessed with the open-source software ProteinProspector to search for histone adducts with atrazine. Taking into consideration that under ESI ionization multicharged ions were obtained for peptides, only ions with charges of + 2 and + 3 were selected. The mass tolerance was set to 5 ppm for precursor and 10 ppm for fragment ions. Trypsin/Glu-C was specified as the cleavage enzyme and maximum missing cleavage was set at 1. Methionine oxidation was specified as variable modifications. In the “User Defined Variable Modifications” parameter, the elemental composition of C₈H₁₃N₅ (179.1171 m/z) from atrazine was selected for potential adduct to amino acid residue of Cys.
For the time-dependent adduct formation study and the concentration-dependent adduct formation study, the acquired UHPLC–Q-Exactive-Orbitrap-MS raw data (in full-scan mode) was analyzed by TraceFinder 5.0 (Thermo Fisher Scientific, Mississauga, ON, Canada). The potential atrazine-modified peptide and non-modified peptide ions were selected as target ions, and the peak areas in extracted chromatograms were compared.

Calf thymus whole histones and human histone H3.3 (expressed in E. coli) were purchased from Millipore-Sigma (Oakville, ON, Canada). Sequencing-grade modified trypsin and Glu-C were purchased from Promega (Madison, WI, USA). Atrazine (analytical standard grade) was purchased from Sigma-Aldrich (Oakville, ON, Canada). A concentration of 0.5 M phosphate-buffered saline (PBS, pH = 7.4) was purchased from Fisher Sci. (Ottawa, ON, Canada). HPLC-grade acetonitrile (ACN), dimethyl sulfoxide (DMSO), ammonium bicarbonate (NH₄HCO₃), formic acid, and a BCA protein assay kit were from Millipore-Sigma. Water was purified on a Milli-Q system (Millipore, Billerica, MA, USA). Asn-Asn-Asn peptide (Asn₃) was synthesized by Biomatik Corporation (Kitchener, ON, Canada).
Atrazine spike solution was prepared by dissolving an amount of Atrazine in DMSO and stored at 4 °C until further use. Calf thymus whole-histone solution was prepared freshly on the day of study by dissolving an amount of histone in 50 mM PBS (pH = 7.4).

Field experiments in 2020, 2021, and 2022 evaluated early-season corn growth impacted by various nutritional levels. The nutritional levels were induced by different nutrient application rates, which constituted the treatments (Table 1). Twelve treatments were assessed in 2021 and 2022, and eight treatments in 2020. Every year, Treatments T1 to T4 had similar N, P, and K (referred hereafter as primary nutrients) application rates, except that T2 received additional Mg, Ca, and S (referred hereafter as secondary nutrients), and T3 received additional B, Zn, Mn, Fe, Cu, and Mo (referred hereafter as micronutrients). Treatment T4 received both secondary and micronutrients. Treatments T5 to T8 received similar N, P, and K rates, but the rates were greater than those in treatments T1 to T4. Additionally, treatment T1 corresponded to T5, T2 to T6, T3 to T7, and T4 to T8 regarding secondary and micronutrient application rates. Treatment T10 received primary, secondary, and micronutrients at greater rates than T8, whereas treatment T12 was a control, receiving no nutrient application. Treatments T9 and T11 were not assessed in 2020 but varied in 2021 and 2022. Treatment T9 received all primary, secondary, and micronutrients in 2021 and only N and P in 2022. Treatment T11 received N, P, K, and S in 2021 and N, P, K, S, and Zn in 2022. Different nutrient rates were used yearly based on initial nutrient levels, previous crops, and observations of the results from preceding years. Fertilizers as the primary sources of the different nutrients are listed in Table 7.
The treatments were laid out in a randomized complete block design with four replications, except for treatment T12 in 2020, which had only one replication embedded in the fourth replication. Each experimental plot was 9.1 m long and 5.5 m wide. Different research fields at the experimental site were used yearly; thus, treatment effects over the three years were not additive. The previous crop for the years 2020 and 2022 was peanut (Arachis hypogaea) and for the year 2021 it was cotton (Gossypium hirsutum). The field was prepared by harrowing, running a field cultivator, and making 20 cm deep furrows that were 91.4 cm apart. Pioneer hybrid P1870R was used every year and it was planted at 88,958 seeds ha⁻¹ on 7 April, 12 April, and 25 March in 2020, 2021, and 2022, respectively. Glyphosate [N-(phosphonomethyl) glycine], Atrazine (2-chloro-4-ethyl amino-6-isopropyl amino-s -triazine), and Prowl [N-(1-ethyl propyl)-3,4-dimethyl-2,6-dinitrobenzenamine] were applied at manufacturer recommended rates to control weeds in the corn plots between the V3 and V4 growth stage. The plots were provided with supplemental water through overhead irrigation whenever the soil moisture reached 50% of the water-holding capacity of the soil.