Water Research 36 (2002) 4785–4794

Atrazine mineralization potential in two wetlands

Kristen L. Andersona, Kevin A. Wheelerb, Jayne B. Robinsonb,Olli H. Tuovinena,*

aDepartment of Microbiology, Ohio State University, 484 West 12th Avenue, Columbus, OH 43210-1292, USAbDepartment of Biology, University of Dayton, 300 College Park, Dayton, OH 45469-2320, USA

Received 1 January 2002; accepted 1 April 2002

Abstract

The fate of atrazine in agricultural soils has been studied extensively but attenuation in wetland systems has received

relatively little attention. The purpose of this study was to evaluate the mineralization of atrazine in two wetlands in

central Ohio. One was a constructed wetland, which is fed by Olentangy River water from an agricultural catchment

area. The other was a natural fen (Cedar Bog) in proximity to atrazine-treated cornfields. Atrazine mineralization

potential was measured by 14CO2 evolution from [U-ring-14C]-atrazine in biometers. The constructed wetland showed

70–80% mineralization of atrazine within 1 month. Samples of wetland water that were pre-concentrated 200-fold by

centrifugation also mineralized 60–80% of the added atrazine. A high extent of atrazine mineralization (75–81%

mineralized) was also associated with concentrated water samples from the Olentangy River that were collected

upstream and downstream of the wetland. The highest levels of mineralization were localized to the top 5 cm zone of the

wetland sediment, and the activity close to the outflow at the Olentangy wetland was approximately equal to that near

the inflow. PCR amplification of DNA extracted from the wetland sediment samples showed no positive signals for the

atzA gene (atrazine chlorohydrolase), while Southern blots of the amplified DNA showed positive bands in five of the

six Olentangy wetland sediment samples. Amplification with the trzD (cyanuric acid amidohydrolase) primers showed a

positive PCR signal for all Olentangy wetland sediment samples. There was little mineralization of atrazine in any of the

Cedar Bog samples. DNA extracted from Cedar Bog samples did not yield PCR products, and the corresponding

Southern hybridization signals were absent. The data show that sediment microbial communities in the Olentangy

wetland mineralize atrazine. The level of activity may be related to the seasonality of atrazine runoff entering the

wetland. Comparable activity was not observed in the Cedar Bog, perhaps because it does not directly receive

agricultural runoff. Qualitatively, the detection of the genes was associated with measurable mineralization activity

which was consistent with the differences between the two study sites.r 2002 Elsevier Science Ltd. All rights reserved.

Keywords: Atrazine; Biodegradation; Herbicide degradation; PCR amplification; Southern hybridization; Wetland system

1. Introduction

Atrazine (2-chloro-4-(ethylamino)-6-(isopropylami-

no)-s-triazine) is a relatively common contaminant in

groundwater and sometimes exceeds the maximum

contaminant level of 3 mg/l established for drinkingwater. Surface waters with agricultural watersheds have

also been reported to contain detectable levels of

atrazine. The major non-point sources of atrazine in

water are subsurface runoff and field tile drainage in

agricultural sites. Because atrazine is a regulated

compound in drinking water, its attenuation in ground

and surface water is an important aspect of the

environmental fate of this chemical. Elucidation of

abiotic and biological degradative pathways of atrazine

in the environment is also crucial in view of its endocrine

disruptor effects [1]. In soils, the most important

mechanism for the attenuation of atrazine involves

microbial degradation and mineralization. In surface

waters and wetlands, the transformations of atrazine are

*Corresponding author. Tel.: +1-614-292-3379; fax: +1-

614-292-8120.

E-mail address: tuovinen.1@osu.edu (O.H. Tuovinen).

0043-1354/02/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved.

PII: S 0 0 4 3 - 1 3 5 4 ( 0 2 ) 0 0 2 0 9 - 9

poorly understood although these systems receive

agrochemicals through surface runoff and drainage.

To date, there is little evidence that native microbial

populations in natural or constructed wetlands can

actively and substantially mineralize atrazine. Ro and

Chung [2] monitored atrazine biotransformation in

spiked wetland sediments. The samples were taken from

a diked wetland constructed to treat sugar mill waste-

water. The initial concentration of atrazine, 10mg/l,

decreased to o10 mg/l within 10 weeks of aerobic

incubation, and subsequent spikes of atrazine were

depleted in 3 weeks. Atrazine concentrations were

measured by HPLC and the extent of mineralization

could not be established from the biodegradation data.

Kao et al. [3] reported atrazine attenuation in the water

column and sediments of a natural wetland system that

received agricultural runoff following a storm event.

Parallel mesocosm experiments with wetland sediments

indicated accelerated biodegradation of atrazine only in

the presence of an external carbon source (sucrose).

Mineralization potential was not established [3]. Wet-

land mesocosm studies have demonstrated the biode-

gradation of atrazine in water columns [4], with losses of

about 70% as determined by gas chromatography, but

the extent of mineralization and unavailability due to

sorption to sediments remain unknown.

A study using sediment and groundwater samples

from a freshwater wetland showed that approximately

4.5% of atrazine was mineralized under aerobic condi-

tions and less than 2% under anaerobic conditions after

68 weeks of incubation [5]. DeLaune et al. [6] reported

that atrazine was relatively persistent in anaerobic

sediments of a swamp-forest wetland that received

runoff from adjacent atrazine-treated agricultural fields,

but the biodegradation could be enhanced upon the

introduction of aerobic conditions. Mineralization of

atrazine was not investigated in the study. Runes et al.

[7] used bioaugmentation with an atrazine spill-site soil

to enhance an otherwise slow mineralization of atrazine

in sediment microcosm and rhizosome mesocosm

samples from a constructed wetland receiving irrigation

runoff from a container nursery.

There are several known genes that encode the

proteins in the atrazine degradative pathway. The most

commonly studied is atzA, which encodes the atrazine

chlorohydrolase. A survey of five geographically distinct

atrazine mineralizing isolates showed that each isolate

contained an atzA gene X99% identical to the gene

found in the original host, Pseudomonas ADP [8]. Shapir

et al. [9] tested soil samples from several agricultural

fields that had different histories of atrazine application.

Regardless of previous application history, most sam-

ples contained organisms that could mineralize atrazine

to a certain extent and the atzA gene could be amplified

from these soils [9]. Shapir et al. [10] amplified a PCR

product with atzA primers from DNA extracted from

sand/gravel lysimeters that had been amended with

secondary wastewater effluent or partially composted

sludge. The data suggested that there was a positive

association between the presence of atzA and atrazine

mineralization, even if the extent was only 1% [10].

Another gene of interest in this work is trzD, which

encodes the cyanuric acid amidohydrolase that catalyzes

the ring cleavage of cyanuric acid. This compound is

believed to be the last aromatic intermediate in the

degradative pathway, but to date the gene and the

enzyme have been characterized only in a melamine-

degrading Pseudomonas sp. [11]. Data based on hybri-

dization signals [12] indicated that, in agricultural soil

samples, the gene trzD was not dominant and its

detection did not reflect the potential mineralization of

atrazine. However, molecular ecological approaches

using atzA, trzD, and other genes encoding enzymes of

the atrazine degradative pathways have not previously

been applied to study mineralization of atrazine in

wetland systems.

In view of the lack of knowledge on atrazine

mineralization in wetlands, the purpose of this study

was to evaluate the mineralization of atrazine in water

and sediment samples retrieved from a constructed

wetland. It was hypothesized that genes such as atzA

and trzD from the degradative pathways are associated

with active mineralization of atrazine in native microbial

populations. This relationship may be consistent only in

circumstances that involve the presence of atrazine or

metabolites in the wetland and may be analogous to the

atrazine history effect described for agricultural soils

[13,14].

The wetland, Olentangy River Wetland Research

Park of The Ohio State University (http://kh465a.ag.o-

hio-state.edu/ORW.html), receives water from the

Olentangy River, which has an upstream agricultural

watershed. Detectable levels of atrazine, up to 13mg/l,have been periodically detected in the Olentangy River

and in the wetland water column in the past (http://

www.epa.state.oh.us/ddagw/pestsbw2.pdf). The atzA

and trzD genes from the mineralization pathways were

targeted for PCR amplification and Southern blotting

using DNA extracted from the Olentangy wetland

samples. A natural wetland, Cedar Bog, fed by water

from springs was also selected for this work. This

wetland, a nature preserve in Ohio, receives no surface

runoff from the agricultural fields in the vicinity.

2. Materials and methods

2.1. Sampling

Sediment and water samples were retrieved from sites

near the inflow and outflow of the experimental wetland

1 at Olentangy River Wetland Research Park, Colum-

K.L. Anderson et al. / Water Research 36 (2002) 4785–47944786

bus, OH (Fig. 1). This site is a constructed wetland that

receives Olentangy River water through the inflow and

feeds back to the river via the outflow. The Olentangy

River has an agricultural watershed upstream of the

wetland and receives non-point source runoff from these

fields. The site is located at 831108100W longitude and

401105900N latitude and was constructed in 1993. The

hydraulic retention time was between 4 and 5 d in 1999.

The retention time in the wetland varies with the flow

velocity of the Olentangy River and can be as long as 3

weeks during periods of low flow of the river. The

chemical, physical, and biological characteristics of the

Olentangy wetland have been summarized in several

publications and annual reports (http://kh465a.ag.ohio-

state.edu/ORW.html).

Olentangy wetland sediments and overlying water

were sampled with a 4� 140 cm acrylic column. Sedi-

ment samples were pooled or sectioned to measure the

potential for atrazine mineralization. Three replicate

sediment samples were mixed to form a composite.

Individual sediment cores were removed by plunging

from the column sampler and then sectioned into three

approximately equal parts (about 5 cm each). Composite

and sectioned sediment samples were taken in April and

August 2000. All sediment samples were taken under

water-saturated conditions. Water content varied be-

tween samples, with the inlet composite sample having a

water content of 65% while the outlet composite was

34%. Inlet sections had water contents of 63% (0–5 cm),

47% (5–10 cm), and 51% (10–15 cm). The outlet section

water contents were 50%, 35% and 26% for the same

sections. The variability in the water content reflects

sediment compaction and clay content. Further chemi-

cal or mineralogical analyses were not within the scope

of this study.

Wetland water samples were taken in August 2000

and Olentangy River samples were taken in October

2001. Surface water samples were collected as grab

samples. Initial measurements of atrazine mineralization

showed very low activity (o1%). Subsequently, watersamples were concentrated 200 fold by centrifugation at

10,000g for 20min. The pellets were resuspended in the

Fig. 1. Sampling sites at the Olentangy wetland and Cedar Bog. The boardwalks that provide access to the sites at the Olentangy

wetland are also indicated. The inflow marks the site where water is pumped into the cell from the Olentangy River and the outflow

marks the pipe that allows water to return to the river. Sampling sites are indicated with open squares.

K.L. Anderson et al. / Water Research 36 (2002) 4785–4794 4787

mineral salts solution for mineralization experiments.

No effort was made to standardize the concentrated

samples with respect to suspended solids, which varied

with hydraulic flow conditions and rainfall. All sediment

and water samples were stored at 41C before use.

The Cedar Bog site (Fig. 1) is an undisturbed alkaline

fen in Urbana, OH, fed by springs. The alkalinity is due

to Ca, Mg-bicarbonates, causing carbonate precipitation

surrounding and downstream from the springs. The site

is located at 83147.50W longitude and 40130N latitude.

This wetland was sampled at two forested sites. For

Cedar Bog site 1, the samples were collected from a

riparian zone approximately 30m from a cornfield.

Samples for Cedar Bog site 2 were collected along the

banks of a small ditch draining into the bog. The

adjacent upstream cornfield had been treated annually

with atrazine for at least the past 10 years, but there is

no direct access of surface runoff to the Cedar Bog

sampling sites. Five subsamples were taken at each site

for a composite in May and June 2000. The top 15 cm of

soil was sampled each time. Neither sampling site in the

Cedar Bog has received any direct application of

atrazine in the past.

2.2. Mineralization studies

Mineralization of [U-ring-14C]-atrazine under aerobic

conditions was measured in biometers, which consisted

of 60ml serum bottles equipped with suspended 2ml

vials to trap evolved 14CO2 [13]. Each biometer received

5.5 g (wet weight) aliquots of sediment samples or 5ml

of water. Each sample was tested in duplicate biometers.

Two inlet composite sediment samples were autoclaved

at 1211C for 20min and used as a sterile control. Each

biometer received 0.065mmol [U-ring-14C]-atrazine (spe-cific activity 1.54mCi/mmol; Sigma-Aldrich Co., St.

Louis, MO) with a total concentration of 2.55mg

atrazine/kg wet weight sediment and 2.8mg/l for water

samples. Trace mineral salts (per liter of 100� stock

solution): MgSO4 � 7H2O, 50mg; CaSO4, 200mg; FeS-O4 � 7H2O, 1 g; MnSO4 �H2O, 20mg; CuSO4, 20mg;ZnSO4 � 7H2O, 20mg; CoSO4 � 7H2O, 10mg; Na-

MoO4 � 2H2O, 5mg; H3BO3, � 5mg; nitrilotriacetate,3 g) were also added (100ml of 100X) to fulfill any tracemetal requirements. Mineralization of atrazine was

monitored as evolution of 14CO2, which was trapped

in 1ml of 0.5M KOH. The KOH was collected and

replaced at intervals. The alkaline trapping solution

was mixed in 10ml scintillation fluid (Scintiverse BD,

Fisher Scientific, Pittsburgh, PA) in a vial and counted

in a scintillation counter. The counting efficiency was

determined using an external standard and was deter-

mined to be 97%. Mineralization was monitored for up

to 38 d.

2.3. Data analysis

The data were analyzed by calculating the means and

standards deviations from the replicate biometers.

Cumulative atrazine mineralization was calculated by

adding the sample percentage of carbon dioxide evolved

to the total from the preceding time course. A first-order

rate expression was used to compare the time courses of

atrazine mineralization between different samples. First-

order kinetic parameters are commonly used to assess

environmental biodegradation of pesticides, based on

the premise that degradation or mineralization is limited

by the concentration of the pesticide [15]. Thus, the

kinetic parameters determined in this study can be

compared with those previously reported for various soil

and other agricultural and environmental conditions.

Half-lives and rate constants were determined by fitting

the data to P ¼ Pmaxð1� e�ktÞ; where P is the observed

amount of 14CO2 evolved (%), Pmax the maximum

extent of mineralization, t the time (d), and k the rate

constant (d�1). Half-life (t1=2) values were calculated

using their respective rate constants as t1=2 ¼ ln 2k�1:The first-order fits were evaluated by calculating the

corresponding r2 values (coefficient of determination).

While atrazine mineralization could not be adequately

described with non-linear first-order rate expression in

samples where the evolution of 14CO2 remained

relatively low, additional kinetic analysis to seek the

best-fitting rate expressions was beyond the scope of this

study.

2.4. Mass balances

After the final sampling, the samples were dried for

approximately 24 h at 901C. The dried soil samples were

crushed to a uniform particle size and 0.25 g samples

(dry weight) were used for analysis. The samples were

combusted in an oxidizer (Biological Oxidizer OX-400,

R.J. Harvey Instrument Corporation, Patterson, NJ)

and the evolved 14CO2 was trapped in scintillation fluid

(14C Cocktail, R.J. Harvey Instrument Corporation).

The cocktail was transferred to a vial for liquid

scintillation counting. For mass balance estimates, the

amount of radioactivity remaining in the sample was

added to the cumulative 14CO2 from the mineralization

studies.

2.5. DNA isolation and PCR amplification

DNA was extracted from 0.5 g sectioned sediment

samples using the alternative lysis method in the

UltraClean Soil DNA Kit (MoBio Laboratories, Solana

Beach, CA). DNA was stored in 10mM Tris pH 8 at

�201C. Positive controls were amplified from plasmid

constructs transformed into Escherichia coli DHa strains

K.L. Anderson et al. / Water Research 36 (2002) 4785–47944788

(pMD4 for atzA and pJK204 for trzD). The Surzycki

et al. [16] protocol was used for plasmid isolation.

Primers for atzA were based on the 528 bp internal

sequence of atzA [17]. The trzD primers were generated

using a computer software package (GeneRunner,

http://www.generunner.com) and were based on an

internal 500 bp sequence of the gene, which was

sequenced by Karns [11]. The trzD forward primer was

50-TCGTTCAGGTCAAGTGCC-30 and the reverse

primer was 50-ATCGTCCAGCATCGTGTG-30. The

PCR mixture (final volume: 50ml) contained 0.2mgtemplate DNA, 1X PCR buffer (30mM Tricine, 0.1%

gelatin, 1.0% Thesit (polyoxyethylene 9 lauryl ether),

and 5mM b-mercaptoethanol), 1.0mM primers,

0.55mM dNTPs, 0.2mg BSA, 3.0mM MgCl2, and 3–5

U Taq polymerase. The atzA gene was amplified in a

PTC-200 DNA engine thermocycler (MJ Research,

Waltham, MA) using the following cycle: denaturation

at 921C for 5min, 40 cycles of 921C for 1.5min, 551C for

1min, 721C for 2min, and a final extension step at 721C

for 5min. The trzD gene was amplified with similar

cycles with the exception of the extension step which was

performed at 581C. 10 ml samples of all PCR reactionswere separated on a 1.0% agarose gels in TAE buffer.

The bands were visualized by soaking the gel for 1 h in a

10�4 dilution of SYBR Green 1 DNA stain (Molecular

Probes, Eugene, OR) in TAE.

The detection limit for atzA and trzD was determined

to be approximately 100 copies per PCR reaction using

plasmid DNA as the template. However, the detection

limits for atzA and trzD when amplified from sediment

DNA samples cannot be defined because of unknown

and perhaps variable efficiency of DNA extraction.

2.6. Southern hybridization analysis

The DNA probes used in Southern blot hybridiza-

tions were prepared by PCR amplification of plasmids

pMD4 and pJK204 containing the cloned atzA and trzD

genes. The amplified fragments were subsequently

purified using a PCR gene clean kit (Qiagen, Valencia,

CA). Approximately 50 ng of the fragment DNA was

boiled for 5min followed by a 5-min incubation on ice to

denature the double stranded templates. DNA probes

were labeled with fluorescein-dUTP using the Gene

Images Random-prime kit (Amersham Pharmacia Bio-

tech, Piscataway, NJ). PCR products were separated by

agarose gel electrophoresis and were transferred to

Hybond-N+ membranes. Hybridization and all washes

were performed at 581C. Hybridization signals were

detected using an anti-fluorescein antibody conjugated

with alkaline phosphatase (CDP-Star detection kit,

Amersham Pharmacia Biotech).

3. Results and discussion

Atrazine mineralization in the Olentangy wetland was

initially evaluated using composite sediment samples. As

shown in Fig. 2, the extent of mineralization after 38 d of

incubation was approximately 43% at the inlet and 47%

at the outlet. Autoclaved sediment samples showed no

Time (Days)

0 5 10 15 20 25 30 35 40

14C

O2

Evo

lved

(%)

0

20

40

60

80

100

Olentangy InletOlentangy OutletSterile ReferenceCedar Bog 1Cedar Bog 2

Fig. 2. Mineralization data from the April, 2000 Olentangy wetland composite samples and Cedar Bog sediment samples. Standard

deviations may be smaller than the points.

K.L. Anderson et al. / Water Research 36 (2002) 4785–4794 4789

mineralization activity (14CO2 evolution o1%). Thisinitial screening of 14CO2 evolution from the ring-

labeled atrazine indicated that an atrazine-mineralizing

microbial community was present in both segments of

the Olentangy wetland.

In general, atrazine mineralization in composite

samples appeared to approach a sustained linear phase.

However, the rate constant and half-life for the

composite inlet sample (April, 2000) indicated a reason-

able first-order fit (r2 > 0:988), suggesting that the rate ofmineralization was accelerating into a nonlinear phase.

This linear phase may be analogous to an extended lag

period preceding accelerated mineralization.

The sediment zone closest to the water interface (0–

5 cm) showed consistently higher levels and faster

mineralization than the underlying zones at both

the inlet and outlet (Fig. 3). These data indicated

differences in the distribution of atrazine-minera-

lizing organisms by depth, with the highest activity

located at the sediment–water interface. The surface of

the sediment is the initial sink for atrazine in the inflow

river water. It is noteworthy that the composite inlet

sediment sample yielded a lower extent of mineralization

as compared to the individual sediment fractions within

a similar period of incubation. This difference may

reflect variability such as clay content and cell density

14C

O2

Evo

lved

(%)

0

20

40

60

80

100

0-5 cm5-10 cm10-15 cm

Time (Days)

0 5 10 15 20 25 30 35

0

20

40

60

80

0 5 10 15 20 25 30 35

(D)(C)

(A) (B)

Fig. 3. Mineralization data for the Olentangy wetland sediment sections: (A) April, 2000 Inlet, (B) August, 2000 Inlet, (C) April, 2000

Outlet, (D) August, 2000 Outlet. Standard deviations were smaller than the symbols.

K.L. Anderson et al. / Water Research 36 (2002) 4785–47944790

and distribution among the core samples retrieved for

the composite.

Figs. 3A and B show the 14CO2 evolution data for the

inlet sections sampled in April and August, respectively,

and Figs. 3C and D show the corresponding data for the

outlet. At the inlet, the extent of atrazine mineralization

in the 0–5 and 5–10 cm zones were approximately 15%

lower in August than April at day 28 (Fig. 3A and B).

However, the rate of mineralization was faster in the

August samples. At the outlet, almost no activity was

present at depths of 5–15 cm in August, unlike the April

samples (39 and 21% compared to 1%).

Table 1 lists kinetic parameters of atrazine miner-

alization derived from the first-order rate expression.

Half-lives were in the order of 5–7 d in the most active

samples. Kinetic parameters were not calculated for

samples where mineralization was low and linear over

the experimental time course.

There are complex reasons underlying these differ-

ences in the kinetics and extent of mineralization in

April and August. During this period, changes took

place, for example, in ambient temperature, hydraulic

flow rate, suspended solids, plant and microbial com-

munity diversity, and sediment organic carbon content.

The role of such complex factors in influencing atrazine

mineralization has yet to be determined.

Mass balance estimates accounted for 85–105% of the14C initially added in these experiments. These mass

balances revealed no major experimental errors or

unaccounted losses of 14CO2 that would have skewed

the calculation of the relative 14CO2 evolution.

Because the activity was so high in the sediment

sections, the atrazine mineralization potential of wetland

water samples was also studied. Atrazine, a pre-

emergent herbicide, is normally applied to agricultural

soils in Ohio at the end of April or beginning of May.

Atrazine would largely be washed from the fields during

rain events in the summer. In river water, atrazine is

mostly associated with suspended solids which include

soil particles from runoff and microorganisms. Biometer

experiments with August samples of wetland water

column showed no mineralization activity. However,

when water samples were first concentrated (200 fold)

before 14CO2 evolution measurements in biometers,

atrazine mineralization was readily detected in the

suspended solids in the inlet water (Fig. 4A).

The Olentangy River was grab sampled about 10m

upstream and 50m downstream of the wetland inlet and

outlet, respectively. The biometer data showed active

mineralization in concentrated water samples, and there

were no differences in the mineralization activity

between the two sites of sampling of river water

(Fig. 4B). In comparison, the rates of mineralization

differ within the wetland, being faster at the inlet

segment. This difference suggests that the capacity for

mineralization in the wetland decreases with distance

from the inlet. Unlike the river, the Olentangy wetland

has low flow conditions (for the 200m distance between

the inlet and the outlet), and sometimes virtual stagna-

tion, which plays a part in the sedimentation of atrazine-

mineralizing microorganisms entering the wetland with

influent river water.

The atzA gene was not detected in DNA from the

Olentangy wetland samples using PCR with atzA

specific primers. However, positive signals were detected

when PCR products were resolved by agarose gel

electrophoresis and subsequently analyzed by Southern

hybridization with the 523 bp PCR product of atzA

(data not shown). These data suggested that atzA was

amplified from DNA but the product was present at

such low levels that it could not be visualized on the gel.

Further attempts were not made to enhance amplifica-

tion of atzA by varying PCR conditions or by nested

PCR. In contrast, the trzD PCR product (500 bp) was

detected in all sections by PCR amplification as well as

by Southern hybridization (Fig. 5).

The mineralization of atrazine was low in samples

from Cedar Bog (Fig. 2). When fitted into the first-order

rate expression, the time course curves of atrazine

mineralization yielded kinetic parameters, but the

corresponding r2 values indicated that these kinetic data

were not of a first order (Table 1). No further kinetic

analysis was undertaken due to a low extent of

mineralization (p13%). No PCR products were ob-

tained with DNA from the Cedar Bog samples using

primers designed for atzA and trzD. Southern hybridi-

zation of the PCR gels failed to detect any signals when

Table 1

Kinetic parameters of atrazine mineralization in the wetland

samples

Sample Date k (d�1) t1=2 (d) r2

Inlet, 0–5 cm April 0.0640 10.8 0.8773

Inlet, 5–10 cm April 0.0644 10.8 0.8970

Inlet, 10–15 cm April 0.0208 33.4 0.9320

Outlet, 0–5 cm April 0.0166 41.9 0.8938

Outlet, 5–10 cm April 0.0142 48.8 0.9556

Outlet, 10–15 cm April 0.0083 83.7 0.8777

Inlet-composite April 0.0166 41.7 0.9883

Outlet-composite April 0.0137 50.6 0.8973

Inlet, 0–5 cm August 0.1458 4.6 0.8554

Inlet, 5–10 cm August 0.1044 6.3 0.8614

Inlet, 10–15 cm August 0.0673 10.3 0.8771

Outlet, 0–5 cm August 0.0233 29.7 0.8703

Outlet, 5–10 cm August 0.0005 1303 0.9832

Outlet, 10–15 cm August 0.0004 1734 0.9286

Inlet-water August 0.0389 17.8 0.5888

Outlet-water August 0.0002 3084 0.0990

Cedar Bog 1 May 0.0053 131 0

Cedar Bog 2 May 0.0045 154 0

K.L. Anderson et al. / Water Research 36 (2002) 4785–4794 4791

hybridized with PCR products of authentic atzA and

trzD. Thus, the Cedar Bog samples showed a low

potential for atrazine mineralization. Since the assay

system used in this study was based on 14CO2 evolution

rather than a decrease in the concentration of atrazine, it

is possible that atrazine degradation exceeded miner-

alization in Cedar Bog samples. The lack of PCR

products and Southern hybridization signals does not,

however, support this possibility.

Detection of the atzA gene in the Olentangy wetland

sediment samples was ambiguous because its presence

could only be confirmed by Southern hybridization.

While the data suggest that Southern hybridization was

more sensitive in detecting PCR products than SYBR

Green 1, the actual detection level of Southern

hybridization is unknown. For trzD, there was good

agreement with the detection of the PCR product and its

Southern hybridization signal. Both genes may index

atrazine mineralization. The atzA gene encodes a

chlorohydrolase at the beginning of the upper pathway.

The trzD gene is involved in the lower pathway and

encodes the cyanuric acid amidohydrolase that catalyzes

the ring cleavage of the heterocyclic s-atrazine. The PCR

product of trzD was detected in DNA in all Olentangy

wetland sediment samples examined (Fig. 5B). Because

trzD was more prevalent than atzA, it is plausible that

14C

O2

Evo

lved

(%)

0

20

40

60

80

100

InletOutlet

Time (Days)

0 5 10 15 20 25 30 35

0

20

40

60

80

UpstreamDownstream

(A)

(B)

Fig. 4. (A) Atrazine mineralization from concentrated wetland water samples taken in August 2000. Both duplicates of the inlet water

reached a similar Pmax value, but one had a 2 d longer lag period before the onset of mineralization. (B) Atrazine mineralization for

concentrated Olentangy River water. Samples were taken in October 2001. Standard deviations, if not shown, were smaller than the

symbols.

K.L. Anderson et al. / Water Research 36 (2002) 4785–47944792

trzD and atzA represent genes from different pathways

of atrazine metabolism in native bacterial communities.

These results indicate that atzA, as originally character-

ized in Pseudomonas ADP, may not be the dominant

chlorohydrolase gene in this wetland system. The

recently identified trzN, a triazine hydrolase gene from

a Nocardioides soil isolate, is supportive of this conclu-

sion [18]. The corresponding amino acid sequence

showed insignificant homology to AtzA and the trzN

gene could not be amplified using atzA-specific primers

[18,19].

In conclusion, the Olentangy River receives agricul-

tural runoff and contains low mg/l levels of atrazine. Inthis work, the river water was shown to contain atrazine-

mineralizing microorganisms, believed to originate from

agricultural watershed runoff. The water from the river

feeds the Olentangy wetland cell, where the mineraliza-

tion activity was found to be more pronounced near the

inlet rather than the outlet. Thus, the mineralization of

atrazine in the wetland can be attributed to the actions

of microorganisms introduced from the river inflow as

well as by microbial populations already established in

the sediment zones. To our knowledge, this is the first

report of active mineralization of atrazine in a con-

structed wetland system without deliberate inoculation

with atrazine-mineralizing organisms.

Acknowledgements

We thank M. de Souza for the E. coli construct

containing atzA and J.S. Karns for the E. coli construct

containing trzD. Partial funding for this work was

received from the US Department of Agriculture, Grant

No. 98-35107-6388. Olentangy Wetland Research Park

publication No. 02-004.

References

[1] Hayes TB, Collins A, Lee M, Mendoza M, Noriega N,

Stuart AA, Vonk A. Hermaphroditic, demasculinized

frogs after exposure to the herbicide atrazine at low

ecologically relevant doses. Proc Nat Acad Sci USA

2002;99:5476–80.

[2] Ro KS, Chung KH. Atrazine biotransformation in

wetland sediment under different nutrient conditions-II:

aerobic. J Environ Sci Health 1995;A30:121–31.

[3] Kao CM, Wang JY, Wu MJ. Evaluation of atrazine

removal processes in a wetland. Wat Sci Technol

2001;44(11–12):539–44.

[4] Moore MT, Rodgers Jr JH, Cooper CM, Smith Jr S.

Constructed wetlands for mitigation of atrazine-associated

agricultural runoff. Environ Pollut 2000;110:393–9.

[5] Larsen L, J�rgensen C, Aamand J. Potential mineraliza-tion of four herbicides in a ground water-fed wetland area.

J Environ Qual 2001;30:24–30.

[6] DeLaune RD, Devai I, Mulbach C, Crozier C, Lindau

CW. The influence of soil redox conditions on atrazine

degradation in wetlands. Agric Ecosyst Environ

1997;66:41–6.

[7] Runes HB, Jenkins JJ, Bottomley PJ. Atrazine degradation

by bioaugmented sediment from constructed wetlands.

Appl Microbiol Biotechnol 2001;57:427–32.

[8] de Souza ML, Seffernick J, Martinez B, Sadowsky MJ,

Wackett LP. The atrazine catabolism genes atzABC are

widespread and highly conserved. J Bacteriol

1998;180:1951–4.

[9] Shapir N, Goux S, Mandelbaum RT, Pussemier L. The

potential of soil microorganisms to mineralize atrazine as

predicted by MCH-PCR followed by nested PCR. Can J

Microbiol 2000;46:425–32.

Fig. 5. (A) PCR products of DNA isolated from wetland

samples, amplified with the trzD primers. (B) Southern

hybridization of the PCR products with a trzD probe. All

DNA samples from Olentangy wetland sediments are from the

sampling in April, 2000. The Cedar Bog samples were taken in

May and June 2000. (Lane 1) 100 bp ladder; (lane 2) pJK204

(positive control); (lane 3) Inlet 10–15 cm; (lane 4) Inlet 5–

10 cm; (lane 5), Inlet 0–5 cm; (lane 6) Outlet 10–15 cm; (lane 7)

Outlet 5–10 cm; (lane 8) Outlet 0–5 cm; (lane 9) Cedar Bog site 1

sampled in May; (lane 10) Cedar Bog site 1 sampled in June;

(lane 11) Cedar Bog site 2 sampled in May; (lane 12) Cedar Bog

site 2 sampled in June; (lane 13) no DNA (negative control);

(lane 14) pJK204 (positive control).

K.L. Anderson et al. / Water Research 36 (2002) 4785–4794 4793

[10] Shapir N, Mandelbaum RT, Fine P. Atrazine mineraliza-

tion by indigenous and introduced Pseudomonas sp. Strain

ADP in sand irrigated with municipal wastewater and

amended with composted sludge. Soil Biol Biochem

2000;32:887–97.

[11] Karns JS. Gene sequence and properties of an s-triazine

ring-cleavage enzyme from Pseudomonas sp. Strain

NRRLB-12227. Appl Environ Microbiol 1999;65:3512–7.

[12] Ostrofsky EB, Robinson JB, Traina SJ, Tuovinen OH.

Effect of cyanuric acid amendment on atrazine miner-

alization in surface soils and detection of the s-triazine

ring-cleavage gene trzD. Soil Biol Biochem 2001;33:

1539–45.

[13] Ostrofsky EB, Traina SJ, Tuovinen OH. Variation in

atrazine mineralization rates in relation to agricultural

management practice. J Environ Qual 1997;26:647–57.

[14] Yassir A, Lagacherie B, Houot S, Soulas G. Microbial

aspects of atrazine biodegradation in relation to history of

soil treatment. Pestic Sci 1999;55:799–809.

[15] Soulas G, Lagacherie B. Modelling of microbial degrada-

tion of pesticides in soils. Biol Fertil Soils 2001;33:551–7.

[16] Surzycki SJ, Surzycki JA, Wagner R. Techniques in

molecular biology workshop: DNA sequencing, PCR,

and RFLP analysis. Laboratory Manual, University of

Indiana, Bloomington, IN, 1997.

[17] de Souza ML, Wackett LP, Boundy-Mills KL, Mandel-

baum RT, Sadowsky MJ. Cloning, characterization and

expression of a gene region from Pseudomonas sp. strain

ADP involved in the dechlorination of atrazine. Appl

Environ Microbiol 1995;61:3373–8.

[18] Mulbry WW, Zhu H, Nour SM, Topp E. The triazine

hydrolase gene trzN from Nocardioides sp. strain C190:

cloning and construction of gene-specific primers. FEMS

Microbiol Lett 2002;206:75–9.

[19] Topp E, Mulbry WM, Zhu H, Nour SM, Cuppels D.

Characterization of s-triazine herbicide metabolism by a

Nocardioides sp. isolated from agricultural soils. Appl

Environ Microbiol 2000;66:3134–41.

K.L. Anderson et al. / Water Research 36 (2002) 4785–47944794