Antibiotic Resistance Genes in Interconnected Surface Waters as Affected by Agricultural Activities

Abstract

Pastures have become one of the most important sources of antibiotic resistance genes (ARGs) pollution, bringing risks to human health through the environment and the food that is grown there. Another significant source of food production is greenhouse horticulture, which is typically located near pastures. Through waterways, pasture-originated ARGs may transfer to the food in greenhouses. However, how these pasture-originated ARGs spread to nearby waterways and greenhouses has been much less investigated, while this may pose risks to humans through agricultural products. We analyzed 29 ARGs related to the most used antibiotics in livestock in the Netherlands at 16 locations in an agricultural area, representing pastures, greenhouses and lakes. We found that ARGs were prevalent in all surface waters surrounding pastures and greenhouses and showed a similar composition, with sulfonamide ARGs being dominant. This indicates that both pastures and greenhouses cause antibiotic resistance pressures on neighboring waters. However, lower pressures were found in relatively larger and isolated lakes, suggesting that a larger water body or a non-agricultural green buffer zone could help reducing ARG impacts from agricultural areas. We also observed a positive relationship between the concentrations of the class 1 integron (intl1 gene)-used as a proxy for horizontal gene transfer-and ARG concentration and composition. This supports that horizontal gene transfer might play a role in dispersing ARGs through landscapes. In contrast, none of the measured four abiotic factors (phosphate, nitrate, pH and dissolved oxygen) showed any impact on ARG concentrations. ARGs from different classes co-occurred, suggesting simultaneous use of different antibiotics. Our findings help to understand the spatial patterns of ARGs, specifically the impacts of ARGs from pastures and greenhouses on each other and on nearby waterways. In this way, this study guides management aiming at reducing ARGs' risk to human health from agricultural products.

Keywords: agricultural area; antibiotic resistance genes; co-occurrence; environmental pollution; greenhouse; intl1 gene; pasture.

Figure 1

Sampling locations. The colors indicate the location types, including pasture (yellow), greenhouse (green) and lake (blue). Each location is highlighted by a circle with a diameter of 500 m to indicate the zone of immediate influence on the sampling location.

Figure 2

ARG concentrations. (a) Average of the relative concentrations [log¹⁰(x*10⁸+1) transformed, copies/16S rRNA gene] of each ARG in all 48 samples, with error bars on the top showing the standard errors (SE). Colors indicate ARGs, each color family represents an antibiotic class, sulfonamide (yellow), trimethoprim (red), tetracycline (blue), beta-lactamase (green) and multidrug (grey). (b) Relative ARG concentrations (copies/16S rRNA gene) of each sample. Row names indicate the land use type of the sampling location (P, pasture; G, greenhouse; L, lake) followed by the number of the sequence of locations and replicates, e.g., P1 (the first pasture location)-1 (the first sampling replication). Colors indicate ARGs in the same way as in (a).

Figure 3

Total absolute ARG concentrations (copies/mL water) (a) and total relative ARG concentrations (copies/16S rRNA gene copy) (b). Colors indicate the location types, including pasture (yellow), greenhouse (green) and lake (blue).

Figure 4

(a) The total relative ARG concentration of each antibiotic class (copies/16S rRNA gene), including sulfonamide, trimethoprim, tetracycline, beta-lactamase and multidrug. (b) The total relative ARG concentration of each antibiotic mechanism (copies/16S rRNA gene), including protection, deactivate and efflux. Significant differences between land use types from the Dunn′s test as the post-hoc test of a Kruskal–Wallis test are shown by asterisks on the top. Note that the ranges of the different y-axes are different. (*, p < 0.05; **, p < 0.01). (c) NMDs diagram of land use types based on the ARGs composition. Colors and shapes indicate the land use types.

Figure 5

The relative concentration of the intl1 gene (copies/16S rRNA gene) for each ARG class ((a), sulfonamide; (b), trimethoprim; (c), tetracycline; (d), beta-lactamase; (e), multidrug). The ARG concentration was log¹⁰(x*10⁶) transformed except for sulfonamide to meet normal distributions. Significant relations are shown by the blue lines indicating the linear regression line for each data set.

Figure 6

The relative concentration of the intl1 gene (copies/16S rRNA gene) for each ARG mechanism ((a), deactivate; (b), protection; (c), efflux). The ARG concentration was log¹⁰(x*10⁶) transformed except for protection to meet normal distributions. Significant relations are shown by the blue line indicating the linear regression line.