Environ Sci Pollut Res DOI 10.1007/s11356-015-4224-1

RESEARCH ARTICLE

Depth, soil type, water table, and site effects on microbial community composition in sediments of pesticide-contaminated aquifer Marja K. Mattsson & Xinxin Liu & Dan Yu & Merja H. Kontro

Received: 31 October 2014 / Accepted: 9 February 2015 # Springer-Verlag Berlin Heidelberg 2015

Abstract Microbial community compositions in pesticidecontaminated aquifers have not been studied, although such information is important for remediation and maintaining freshwater sources clean under changing climate. Therefore, phospholipid (PLFAs), glycolipid (GLFAs), and neutral lipid (NLFAs) fatty acids were determined from sand and clay sediments at depths of 0.3–24.8 m, all contaminated with triazines and dichlobenil/2,6-dichlorobenzamide. The portion of fungi and Gram-negative bacteria at 0.3 m was greater than at 0.8 m, where the percentage of Gram-positive bacteria, actinobacteria, and sulfate-reducing bacteria (SRB) increased. In deeper sediments, microbial biomass, activity, and diversity decreased. Clay sediments seemed to serve as a reservoir for slow pesticide elution to groundwater, and their biomarker portion for all bacteria except actinobacteria was greater than in sand sediments. The slow pesticide dissipation seemed to occur in the main groundwater flow zone, resulting in nitrogen release simultaneously with organic matter elution from Responsible editor: Robert Duran Highlights - PLFAs, GLFAs, and NLFAs of pesticide-contaminated aquifer sediments determined. - Gram− and fungi at 0.3 m; Gram+, actinobacteria, and SRB at 0.8 m; deeper all reduced. - Gram−, Gram+, and SRB in clay sediments more than sandy; unaffected by water table. - Organic matter and PLFAs correlated, due to increase in all biomarkers except TBSA. - Slow pesticide dissipation below garden, where increased organic matter and biomass. - Conditions for natural enhanced attenuation of pesticides suggested. Mattsson, previously Talja; Kontro, previously Suutari M. K. Mattsson (*) : X. Liu : D. Yu : M. H. Kontro Department of Environmental Sciences, University of Helsinki, Niemenkatu 73, 15140 Lahti, Finland e-mail: [email protected]

gardening and bank filtration. As a result, microbial biomass, activity, and diversity were increased. This shift in conditions towards that in surface soil may be appropriate for enhanced natural attenuation of pesticides in groundwater sources. Keywords Lipid biomarkers . Groundwater . Sediments . Pesticides . Organic matter . Nitrogen

Introduction The s-triazine pesticides, like atrazine (6-chloro-N-ethylN′-(1-methylethyl)-1,3,5-triazine-2,4-diamine) and simazine (6-chloro-N,N′-diethyl-1,3,5-triazine-2,4-diamine) have been used worldwide in agriculture, and they are still in use, e.g., in USA and China. In soil surface, these herbicides are degraded in a few months by microorganisms, but the biodegradation rate decreases with increasing depth. In groundwater, they typically are persistent (Pearson et al. 2006; Krutz et al. 2010; Talja et al. 2008; Pukkila and Kontro 2014). Although the sale of atrazine, simazine, and dichlobenil (2,6dichlorobenzonitrile) in Finland was banned in 1992, 2004, and 2011, respectively, these pesticides, and 2,6dichlorobenzamide (BAM), a metabolite of dichlobenil are still discovered in groundwater (Pukkila et al. 2009; Talja et al. 2008). BAM is more persistent than its parent compound dichlobenil especially in aquifer sediments (Clausen et al. 2007; Pukkila and Kontro 2014). The microbial attenuation of leachable pesticides in surface soil and subsurface sediments is related to the microbial community composition, which is affected by environmental conditions, like depth, sediment texture, temperature, moisture, oxygen, and nutrients (Albrechtsen and Winding 1992; Kieft et al. 1998; Hoyle and Arthur 2000; Talja et al. 2008; Pukkila

Environ Sci Pollut Res

et al. 2009; Pukkila and Kontro 2014). The subsurface microbial community compositions above and below the water table may vary much depending on fluctuating water and nutrient concentrations. In deep sediments, microorganisms have developed survival strategies under the low nutrient concentration and they grow slowly anaerobically (Balkwill et al. 1989; Albrechtsen and Winding 1992; Kieft et al. 1997, 1998; Krumholz 2000; Ekelund et al. 2001; Holden and Fierer 2005). Phospholipid fatty acid (PLFA) analysis has been used to determine microbial biomass and community structure in wide variety of different environments. Phospholipids are essential cell membrane constituents of living microorganisms and they are rapidly hydrolyzed in dead cells (Zelles 1999; Green and Scow 2000). During the PLFA analysis, fractions for glycolipid fatty acids (GLFAs) and neutral lipid fatty acids (NLFAs) are also separated, but their fatty acid compositions have only seldom been determined from environmental samples. GLFAs are components of cell membranes and their outer surface. Especially that gram-positive actinobacteria are known to contain a variety of glycolipids. NLFAs include storage triacylglycerols, and fatty acids and diacylglycerols released by lipases from phospholipids and triacylglycerols (Nelson and Cox 2008; Bååth 2003). Fatty acids esterified to phospholipids, glycolipids, and neutral lipids can be used as biomarkers for different microbial groups (Zelles 1999). Microbial community compositions in deep sediments of pesticide-contaminated aquifers have not been studied, although such information is a prerequisite for planning remediation practices of contaminated groundwater sources in changing climate conditions. Therefore, PLFAs, GLFAs, and NLFAs were examined from surface soil down to the depth of 24.8 m in an aquifer having atrazine, simazine, hexazinone, dichlobenil, and BAM in sediments and/or groundwater. The railway drilling site is located in groundwater recharge area at the edge of the aquifer. The garden drilling site is above the main groundwater flow zone to groundwater well (Talja et al. 2008). The hypotheses of the study were that depth, soil type (clay, sand), water table, and organic matter affect the microbial community composition and can be related to pesticide dissipation in sediments and groundwater.

Materials and methods Sample collection, and physical and chemical measurements Subsurface sediments were collected in drillings in the aquifer having pesticides in groundwater (Talja et al. 2008). Sediments were collected in Lahti (Finland) next to the railway station by drilling to the depth of 16.7 m, and in the city garden by drilling to the depth of 24.8 m. The drill diameter was 75 mm down to about 10 m, and below that 48 mm. The groundwater table was 15.0 m at the railway station, and 4.6 m

in the city garden. Samples were collected in drilling sites to polyethylene bags, frozen and stored at −20 °C for lipid analyses. The soil dry weight was determined from the weight loss of triplicate samples, heated at 105 °C for 16 h. The soil organic matter was determined from the loss on ignition (550 °C, 4 h), and converted to milligram per gram dry weight. NH4–N and NO3–N of railway sediments at the depth of 13.6 m were determined as has been presented (Talja et al. 2008).

Pesticide analyses Pesticides were extracted from 10 g (dry weight) of sediment by adding 55 ml of methanol/water (3:1 v/v), and 40 μl of propazine (500 ng μl−1) as an internal standard, all in triplicate. Samples were mixed and sonicated (Everest Ultrasonic, Istanbul, Turkey) at 20 °C for 20 min. After shaking overnight (200 rpm) at 21±2 °C, samples were centrifuged for 10 min at 2,500×g (Multifuge® Heraeus, DJB Labcare, Buckinghamshire, UK), and supernatants were evaporated (Christ RVC 2–18, Martin Christ, Osterode, Germany; diaphragm vacuum pump, Alfred Zippe, Wertheim, Germany). The extraction was repeated two more times, and residues from evaporation were solubilized in 2.5 ml of methanol/ water (3:1, v/v). Calibration standards in methanol/water (3:1; v/v) contained atrazine (2.3–92.7 μM), simazine (2.5–99.2 μM), and BAM (2.6–105.2 μM) in six concentrations, and 145 μM of propazine as an internal standard. The linearity of calibration curves was r≥0.998. For HPLC analysis, 600 μl samples and calibration standards were filtrated through a 0.45-μm GHP membrane (Acrodisc®, Gelman Pall Corporation Ltd., NY, USA), and 20 μl was analyzed. HPLC was equipped with Shimadzu Prominence (Shimadzy, Kyoto, Japan) SIL-20 A auto sampler, LC-20AT solvent delivery module, DGU20A5 on-line degasser, and SPD-20 UV/VIS detector, and SunFire column (C18, 3.5 μm, 3.0×150 mm, Waters, MA, USA), and Shimadzu LC Solution software. Atrazine and simazine were measured at the wave length of 225 nm and BAM at 215 nm. The mobile phase was acetonitrile in water at the flow rate of 0.6 ml min−1. The acetonitrile concentration was held at 30 % for 2.5 min, then increased to 65 % in 5 min, and finally decreased back to 30 % in 3.5 min. The concentrations of hexazinone, atrazine, desethylatrazine, deisopropylatrazine, and desethyldeisopropylatrazine in groundwater were determined in Ramboll Analytics Ltd. (Lahti, Finland) using accredited methods. The data is from annual groundwater quality analyses done by waterworks (Lahti Aqua Ltd, Finland), and it is presented as average concentrations measured in 2011 and 2012 from two groundwater monitoring pipes (n=4) upstream and downstream from garden drilling site.

Environ Sci Pollut Res

Lipid analyses The subsurface sediments about 5.0 g in dry weight were extracted by shaking overnight (200 rpm) in 28.8 ml of methanol/chloroform/phosphate buffer (50 mM, pH 7.4) at the ratio of 1:2:0.8 (v/v/v; Bligh and Dyer 1959; Keinänen et al. 2003; Kontro et al. 2006), all in duplicate. The internal standard, 500 μl of dipentadecanoylphosphatidylcholine in methanol (1 mg ml−1) was added, and the flasks were shaken for further 5 min. The solvent phase was separated by centrifuging (2, 500×g), the volume was measured, and the ratio of chloroform/methanol/phosphate buffer was adjusted to 1:1:0.9 (v/v/ v). The samples were centrifuged for 5 min (2,500×g), the solvent layer was separated, and evaporated in a centrifugal evaporator. The dry lipid extract was dissolved in 2×100 μl of chloroform, and applied to a 10 ml Varian column (Varian, Las Vegas, NV, USA). Neutral lipids were eluted with 10 ml of chloroform, glycolipids with 20 ml of acetone, and phospholipids with 10 ml of methanol, followed by evaporation to dryness in the centrifugal evaporator. Internal standards were added to PL, GL, and NL fractions, 30 μl (1 μg μl−1) of tridecanoic acid methyl ester and nonadecanoic acid methyl ester, both in hexane. The fatty acids of fractions were saponified, methylated, and extracted as methyl esters as has been presented (Kontro et al. 2006). The fatty acids were analyzed with a Shimadzu (Duisburg, Germany) model GC-17A gas chromatograph equipped with a mass selective detector (model GCMC-QP5000), automatic sampler (model AOC-17, long), and ZB-5 capillary column (30 m×0.32 mm×0.25 μm; Shimadzu). The carrier gas was helium (1.0 ml min−1), injection split less, injector temperature 280 °C, and detector transfer interface temperature 270 °C. The oven temperature was held at 50 °C for 1 min, then increased 30 °C/min to 160 °C, and further 5 °C/min to 270 °C. The mass spectra were recorded at electron energy of 70 eV and emission current of 75 mA. The ion source temperature was 215 °C. The fatty acid profiles of the shallowest layers were analyzed in SCAN and SIM modes, and the deeper sediments with low quantities of fatty acids were only analyzed by SIM. Fatty acids were identified on the basis of their mass spectra and retention times. Conversion coefficients were calculated to multiply SIM areas to the corresponding SCAN areas as has been presented (Kontro et al. 2006), and then percentage fatty acid profiles were calculated. Internal standards were used to calculate the quantities of fatty acids. Calculations An estimate of microbial biomass was calculated using following conversion factors: bacteria contain 100 μmol of PLFAs g−1 dry weight; and 1 g of bacteria is equivalent to 2.0×1012 cells on dry weight basis (Balkwill et al. 1988). The Shannon–Weaver diversity index H was calculated from

PLFAs as H=− Σpi(lnpi), and the Simpson diversity index D from D=1- Σpi, where pi is the peak area under the ith peak as a proportion of the total peak area (Shannon and Weaver 1949; Hedrick et al. 2000). The following indices of specific microbial groups were used: saturated straight-chain fatty acids (14:0, 16:0, 17:0, 18:0, 20:0), general microbial biomass; iso- and anteiso-branched fatty acids (i-14:0, i-15:0, a15:0, i-16:0, i-17:0, a-17:0, i-18:0, i-19:0), Gram-positive bacteria; middle-branched acids, sulfate-reducing bacteria (SRB, 10-Me-16:0) and actinobacteria (10-methyloctadecanoic acid, tuberculostearic acid, TBSA); monounsaturated straight-chain fatty acids (16:1ω9c, 16:1ω7c, 16:1ω7t, 16:1ω5c, 18:1ω9c, 18:1ω7c, 18:1ω7t, 18:1ω5c), Gram-negative and other bacteria; cyclopropane acids (cy-17:0ω9c, cy-17:0ω7c, cy19:0ω7c), Gram-negative and lactic acid bacteria; and 18:2ω6c, fungi (Zelles 1999). Principal component analysis (PCA), Pearson two-tailed correlation analyses, Kruskal– Wallis test (K-W, p